Erdas image analysis for arcgis

Geographic Imaging by Leica Geosystems GIS & Mapping
Using Image Analysis for ArcGIS
Julie Booth-Lamirand

Using the Image Analysis Extension for ArcGIS
Copyright © 2003 Leica Geosystems GIS & Mapping, LLC
All rights reserved.
Printed in the United States of America.
The information contained in this document is the exclusive property of Leica Geosystems GIS & Mapping, LLC. This work is protected under United States copyright law
and other international copyright treaties and conventions. No part of this work may be reproduced or transmitted in any form or by any means, electronic or mechanical,
including photocopying and recording, or by any information storage or retrieval system, as expressly permitted in writing by Leica Geosystems GIS & Mapping, LLC. All
requests should be sent to Attention: Manager of Technical Documentation, Leica Geosystems GIS & Mapping, LLC, 2801 Buford Highway NE, Suite 400, Atlanta, GA,
30329-2137, USA.
The information contained in this document is subject to change without notice.
CONTRIBUTORS
Contributors to this book and the On-line Help for Image Analysis for ArcGIS include: Christine Beaudoin, Jay Pongonis, Kris Curry, Lori Zastrow, Mladen Stojic′ , and
Cheryl Brantley of Leica Geosystems GIS & Mapping, LLC.
U. S. GOVERNMENT RESTRICTED/LIMITED RIGHTS
Any software, documentation, and/or data delivered hereunder is subject to the terms of the License Agreement. In no event shall the U.S. Government acquire greater than
RESTRICTED/LIMITED RIGHTS. At minimum, use, duplication, or disclosure by the U.S. Government is subject to restrictions set forth in FAR §52.227-14 Alternates I,
II, and III (JUN 1987); FAR §52.227-19 (JUN 1987), and/or FAR §12.211/12.212 (Commercial Technical Data/Computer Software); and DFARS §252.227-7015 (NOV
1995) (Technical Data) and/or DFARS §227.7202 (Computer Software), as applicable. Contractor/Manufacturer is Leica Geosystems GIS & Mapping, LLC, 2801 Buford
Highway NE, Suite 400, Atlanta, GA, 30329-2137, USA.
ERDAS, ERDAS IMAGINE, and IMAGINE OrthoBASE are registered trademarks. Image Analysis for ArcGIS is a trademark.
ERDAS® is a wholly owned subsidiary of Leica Geosystems GIS & Mapping, LLC.
Other companies and products mentioned herein are trademarks or registered trademarks of their respective trademark owners.

III
Contents
Contents
Contents iii
Foreword vii
Getting started
1 Introducing Image Analysis for ArcGIS 3
Learning about Image Analysis for ArcGIS 10
2 Quick-start tutorial 11
Exercise 1: Starting Image Analysis for ArcGIS 12
Exercise 2: Adding images and applying Histogram Stretch 14
Exercise 3: Identifying similar areas in an image 18
Exercise 4: Finding areas of change 22
Exercise 5: Mosaicking images 30
Exercise 6: Orthorectification of camera imagery 33
What’s Next? 38
3 Applying data tools 39
Using Seed Tool Properties 40
Image Info 45
Options 47
Working with features
4 Using Data Preparation 55
Create New Image 56
Subset Image 58
Mosaic Images 63
Reproject Image 66

USING IMAGE ANALYSIS FOR ARCGISIV
5 Performing Spatial Enhancement 69
Convolution 70
Non-Directional Edge 75
Focal Analysis 77
Resolution Merge 79
6 Using Radiometric Enhancement 83
LUT Stretch 84
Histogram Equalization 87
Histogram Matching 91
Brightness Inversion 93
7 Applying Spectral Enhancement 95
RGB to IHS 96
IHS to RGB 99
Vegetative Indices 101
Color IR to Natural Color 104
8 Performing GIS Analysis 107
Information versus data 108
Neighborhood Analysis 109
Thematic Change 111
Recode 114
Summarize Areas 120
9 Using Utilities 123
Image Difference 124
Layer Stack 126
10 Understanding Classification 129
The Classification Process 130

CONTENTS V
Classification tips 132
Unsupervised Classification/Categorize Image 134
Supervised Classification 138
Classification decision rules 140
11 Using Conversion 143
Conversion 144
Converting raster to features 145
Converting features to raster 147
12 Applying Geocorrection Tools 149
When to rectify 150
Geocorrection property dialogs 153
SPOT 158
The Spot Properties dialog 160
Polynomial transformation 161
The Polynomial Properties dialog 168
Rubber Sheeting 169
Camera Properties 171
IKONOS, QuickBird, and RPC Properties 173
Landsat 177
Glossary 183
References 201
Index 205

USING IMAGE ANALYSIS FOR ARCGISVI

VII
Foreword
An image of the earth’s surface is a wealth of information. Images capture a
permanent record of buildings, roads, rivers, trees, schools, mountains, and other
features located on the earth’s surface. But images go beyond simply recording
features. Images also record relationships and processes as they occur in the real
world. Images are snapshots of geography, but they are also snapshots of reality.
Images chronicle our earth and everything associated with it; they record a specific
place at a specific point in time. They are snapshots of our changing cities, rivers,
and mountains. Images are snapshots of life on earth.
The data in a GIS needs to reflect reality, and snapshots of reality need to be
incorporated and accurately transformed into instantaneously ready, easy-to-use
information. From snapshots to digital reality, images are pivotal in creating and
maintaining the information infrastructure used by today’s society. Today’s
geographic information systems have been carefully created with features,
attributed behavior, analyzed relationships, and modeled processes.
There are five essential questions that any GIS needs to answer: Where, What,
When, Why, and How. Uncovering Why, When, and How are all done within the
GIS; images allow you to extract the Where and What. Precisely where is that
building? What is that parcel of land used for? What type of tree is that? The new
extensions developed by Leica Geosystems GIS and Mapping, LLC use imagery
to allow you to accurately address the questions Where and What, so you can then
derive answers for the other three.
But our earth is changing! Urban growth, suburban sprawl, industrial usage and
natural phenomena continually alter our geography. As our geography changes, so

USING IMAGE ANALYSIS FOR ARCGISVIII
does the information we need to understand it. Because an
image is a permanent record of features, behavior,
relationships, and processes captured at a specific moment in
time, using a series of images of the same area taken over
time allows you to more accurately model and analyze the
relationships and processes that are important to our earth.
The new extensions by Leica Geosystems are technological
breakthroughs which allow you to transform a snapshot of
geography into information that digitally represents reality in
the context of a GIS. Image Analysis™ for ArcGIS and
Stereo Analyst® for ArcGIS are tools built on top of a GIS to
maintain that GIS with up-to-date information. The
extensions provided by Leica Geosystems reliably transform
imagery directly into your GIS for analyzing, mapping,
visualizing, and understanding our world.
On behalf of the Image Analysis for ArcGIS and Stereo
Analyst for ArcGIS product teams, I wish you all the best in
working with these new products and hope you are
successful in your GIS and mapping endeavors.
Sincerely,
Mladen Stojic′
Product Manager
Leica Geosystems GIS & Mapping, LLC

3
1Introducing Image Analysis for ArcGIS
Image Analysis for ArcGIS™ is primarily designed for natural resource and
infrastructure management. The extension is very useful in the fields of forestry,
agriculture, environmental assessment, engineering, and infrastructure projects
such as facility siting and corridor monitoring, and general geographic database
update and maintenance.
Today, imagery of the earth’s surface is an integral part of desktop mapping and
GIS, and it’s more important than ever to have the ability to provide realistic
backdrops to geographic databases and to be able to quickly update details
involving street use or land use data.
Image Analysis for ArcGIS gives you the ability to perform many tasks:
• Import and incorporate raster imagery into ArcGIS.
• Categorize images into classes corresponding to land cover types such as
vegetation.
• Evaluate images captured at different times to identify areas of change.
• Identify and automatically map a land cover type with a single click.
• Find areas of dense and thriving vegetation in an image.
• Enhance the appearance of an image by adjusting contrast and brightness or by
applying histogram stretches.
• Align an image to a map coordinate system for precise area location.
• Rectify satellite images through Geocorrection Models.
IN THIS CHAPTER
• Updating a database
• Categorizing land cover and
characterizing sites
• Identifying and summarizing
natural hazard damage
• Identifying and monitoring urban
growth and changes
• Extracting features automatically
• Assessing vegetation stress
1Introducing Image Analysis
for ArcGIS

USING IMAGE ANALYSIS FOR ARCGIS4
Updating databases
There are many kinds of imagery to choose from in a wide range of scales, spatial, and spectral resolutions, and map accuracies. Aerial
photography is often the choice for map updating because of its high precision. With Image Analysis for ArcGIS you are able to use imagery
to identify changes and make revisions and corrections to your geographic database.
Airphoto with shapefile of streets

INTRODUCING IMAGE ANALYSIS FOR ARCGIS 5
Categorizing land cover and characterizing sites
Transmission towers for radio-based telecommunications must all be visible from each other, must be within a certain range of elevations,
and must avoid fragile areas like wetlands. With Image Analysis for ArcGIS, you can categorize images into land cover classes to help
identify suitable locations. You can use imagery and analysis techniques to identify wetlands and other environmentally sensitive areas.
The Classification features enable you to divide an image into many different classes, and then highlight them as you wish. In this case the
areas not suitable for tower placement are highlighted, and the placement for the towers can be sited appropriately.
Classified image for radio towers

Identifying and summarizing natural hazard damage
When viewing a forest hit by a hurricane, you can use the mapping tools of Image Analysis for ArcGIS to show where the damage occurred.
With other ArcGIS tools, you can show the condition of the vegetation, how much stress it suffers, and how much damage it sustained in
the hurricane.
Below, Landsat images taken before and after the hurricane, in conjunction with a shapefile that identifies the forest boundary, are used for
comparison. Within the shapefile, you can see detailed tree stand inventory and management information.
The upper two pictures show the area in 1987 and in 1989 after Hurricane Hugo. The lower image features the shapefile.

Identifying and monitoring urban growth and changes
Cities grow over time, and images give a good sense of how they grow, and how remaining land can be preserved by managing that growth.
You can use Image Analysis for ArcGIS to reveal patterns of urban growth over time.
Here, Landsat data spanning 21 years was analyzed for urban growth. The final view shows the differences in extent of urban land use and
land cover between 1973 and 1994. Those differences are represented as classes. The yellow urban areas from 1994 represent how much
the city has grown beyond the red urban areas from 1973.
The top two images represent urban areas in red, first in 1974 and then in 1994. The bottom image shows the actual growth.

Extracting features automatically
Suppose you are responsible for mapping the extent of an oil spill as part of a rapid response effort. You can use synthetic aperture radar
(SAR) data and Image Analysis for ArcGIS tools to identify and map the extent of such environmental hazards.
The following image shows an oil spill of the northern coast of Spain. The first image shows the spill, and the second image gives you an
example of how you can isolate the exact extent of a particular pattern using Image Analysis for ArcGIS.
Images depicting an oil spill off the coast of Spain and a polygon grown in the spill using Seed Tool.

Assessing vegetation stress
Crops experience different stresses throughout the growing season. You can use multispectral imagery and analysis tools to identify and
monitor a crop’s health.
In these images, the Vegetative Indices function is used to see crop stress. The stressed areas are then automatically digitized and saved as
a shapefile. This kind of information can be used to help identify sources if variability in growth patterns. Then, you can quickly update
crop management plans.
Crop stress shown through Vegetative Indices

Learning about Image Analysis for ArcGIS
If you are just learning about geographic information systems
(GISs), you may want to read the books about ArcCatalog and
ArcMap: Using ArcCatalog and Using ArcMap. Knowing about
these applications will make your use of Image Analysis for
ArcGIS much easier.
If you’re ready to learn about how Image Analysis for ArcGIS
works, see the Quick-start tutorial. In the Quick-start tutorial, you’ll
learn how to adjust the appearance of an image, how to identify
similar areas of an image, how to align an image to a feature theme,
as well as finding areas of change and mosaicking images.
Finding answers to questions
This book describes the typical workflow involved in creating and
updating GIS data for mapping projects. The chapters are set up so
that you first learn the theory behind certain applications, then you
are introduced to the typical workflow you’d apply to get the results
you want. A glossary is provided to help you understand any terms
you haven’t seen before.
Getting help on your computer
You can get a lot of information about the features of Image
Analysis for ArcGIS by accessing the online help. To browse the
online help contents for Image Analysis for ArcGIS, click Help
near the bottom of the Image Analysis menu. From this point you
can use the Table of contents, index, or search feature to locate the
information you need. If you need online help for ArcGIS, click
Help on the ArcMap toolbar and choose ArcGIS Desktop Help.
Contacting Leica Geosystems GIS &
Mapping
If you need to contact Leica Geosystems for technical support, see
the product registration and support card you received with Image
Analysis for ArcGIS. You can also contact Customer Support at
404/248-9777. Visit Leica Geosystems on the Web at
www.gis.leica-geosystems.com.
Contacting ESRI
If you need to contact ESRI for technical support refer to “Getting
technical support” in the Help system’s “Getting more help”
section. The telephone number for Technical Support is 909-793-
3744. You can also visit ESRI on the Web at www.esri.com.
Leica Geosystems GIS & Mapping
Education Solutions
Leica Geosystems GIS & Mapping Division offers instructor-based
training about Image Analysis for ArcGIS. For more information,
got to the training Web site located at www.gis.leica-
geosystems.com. You can follow the training link to Training
Centers, Course Schedules, and Course Registration.
ESRI education solutions
ESRI provides educational opportunities related to GISs, GIS
applications, and technology. You can choose among instructor-led
courses, Web-based courses, and self-study workbooks to find
educational solutions that fit your learning style and pocketbook.
For more information, visit the Web site www.esri.com/education.

11
2Quick-start tutorial
Now that you know a little bit about the Image Analysis for ArcGIS extension and
its potential applications, the following exercises give you hands-on experience in
using many of the extension’s tools. By working through the exercises, you are
going to use the most important components of the Image Analysis for ArcGIS
extension and learn about the types of problems it can solve.
In Image Analysis for ArcGIS, you can quickly identify areas with similar
characteristics. This is useful for identification in cases such as environmental
disasters, burn areas or oil spills. Once an area has been defined, it can also be
quickly saved into a shapefile. This avoids the need for manual digitizing. This
tutorial will show you how to use some Image Analysis for ArcGIS tools and give
you a good introduction to using Image Analysis for ArcGIS for your own GIS
needs.
IN THIS CHAPTER
• Starting Image Analysis for
ArcGIS
• Adjusting the appearance of an
image
• Identifying similar areas in an
image
• Finding areas of change
• Mosaicking images
• Orthorectifying an image
2

Exercise 1: Starting Image Analysis for ArcGIS
In the following exercises, we’ve assumed that you are using
a single monitor or dual monitor workstation that is
configured for use with ArcMap and Image Analysis for
ArcGIS. That being the case, you will be lead through a
series of tutorials in this chapter to help acquaint you with
Image Analysis for ArcGIS and further show you some of the
abilities of Image Analysis for ArcGIS.
In this exercise, you’ll learn how to start Image Analysis for
ArcGIS and activate the toolbar associated with it. You will
be able to gain access to all the important Image Analysis for
ArcGIS features through its toolbar and menu list. After
completing this exercise, you’ll be able to locate any Image
Analysis for ArcGIS tool you need for preparation,
enhancement, analysis, or geocorrection.
This exercise assumes you have already successfully
completed installation of Image Analysis for ArcGIS on your
computer. If you have not installed Image Analysis for
ArcGIS, refer to the installation guide packaged with the
Image Analysis for ArcGIS CD, and install now.
Starting Image Analysis for ArcGIS
1. Click the Start button on your desktop, then click
Programs, and point to ArcGIS.
2. Click ArcMap to start the application.
Adding the Image Analysis for ArcGIS
extension
1. If the ArcMap dialog opens, keep the option to create a
new empty map, then click OK.
2. In the ArcMap window, click the Tools menu, then click
Extensions.
2
1
1

QUICK-START TUTORIAL 13
3. In the Extensions dialog, click the check box for Image
Analysis Extension to add the extension to ArcMap.
Once the Image Analysis Extension check box has been
selected, the extension is activated.
4. Click Close in the Extensions dialog.
Adding toolbars
1. Click the View menu, then point to Toolbars, and click
Image Analysis to add that toolbar to the ArcMap
window.
The Image Analysis toolbar is your gateway to many of the
tools and features you can use with the extension. From the
Image Analysis toolbar you can choose many different
analysis types from the menu, choose a geocorrection type,
and set links in an image.
4
3
1

Exercise 2: Adding images and applying Histogram Stretch
Image data, displayed without any contrast manipulation,
may appear either too light or too dark, making it difficult to
begin your analysis. Image Analysis for ArcGIS allows you
to display the same data in many different ways. For
example, changing the distribution of pixels allows you to
alter the brightness and contrast of the image. This is called
histogram stretching. Histogram stretching enables you to
manipulate the display of data to make your image easier to
visually interpret and evaluate.
Add an Image Analysis for ArcGIS theme of
Moscow
1. Open a new view. If you are starting this exercise
immediately after Exercise 1, you should have a new,
empty view ready.
2. Click the Add Data button .
3. In the Add Data dialog, select moscow_spot.tif, and
click Add to draw it in the view. The path to the example
data directory is ArcGISArcTutorImageAnalysis.
4. Click Add to display the image in the view.
The image Moscow_spot.tif appears in the view.
Apply a Histogram Equalization
Standard deviations is the default histogram stretch applied
to images by Image Analysis for ArcGIS. You can apply
histogram equalization to redistribute the data so that each
display value has roughly the same number of data points.
More information about histogram equalization can be found
in chapter 6 “Using Radiometric Enhancement”.
1. Select moscow_spot.tif in the Table of contents, right
click your mouse, and select Properties to bring up
Layer Properties.
2. Click the Symbology tab and under Show, select RGB
Composite.
3. Check the Bands order and click the dropdown arrows
to change any of the Bands.
3
4

You can also change the order of the bands in your current
image by clicking on the color bar beside each band in the
Table of contents. If you want bands to appear in a certain
order for each image that you draw in the view, go to
ToolsOptionsRaster in ArcMap, and change the Default
RGB Band Combinations.
4. Click the dropdown arrow and select Histogram
Equalize as the Stretch Type.
5. Click Apply and OK.
6. Click the Image Analysis menu dropdown arrow, point
to Radiometric Enhancement, and click Histogram
Equalization.
7. In the Histogram Equalization dialog, make sure
moscow_spot.tif is in the Input Image box.
8. The Number of Bins will default to 256. For this
exercise, leave the number at 256, but in the future, you
can change it to suit your needs.
9. Navigate to the directory where you want your output
images stored, type a name for your image, and click
Save. The path will appear in Output Image.
You can go to the Options dialog, accessible from the Image
Analysis toolbar, and enter the working directory you want to
use on the General tab of the dialog. This step will save you
time by automatically bringing up your working directory
whenever you click the browse button to navigate to it in
order to store an output image.
5
1 3
2
4
6

10. Click OK.
The equalized image will appear in your Table of
contents and in your view.
This is the histogram equalized image of Moscow.
Apply an Invert Stretch to the image of
Moscow
In this example, you apply the Invert Stretch to the image to
redisplay it with its brightness values reversed. Areas that
originally appeared bright are now dark, and dark areas are
bright.
1. Select the equalized file in the Table of contents, and
right-click your mouse. Click Properties and go to the
Symbology tab.
2. If you want to see the histograms for the image, click the
Histograms button located in the Stretch box.
3. Check the Invert box.
7
9
8
10
1
2
3
4

This is an inverted image of Moscow_spot.tif.
You can apply different types of stretches to your image to
emphasize different parts of the data. Depending on the
original distribution of the data in the image, one stretch may
make the image appear better than another. Image Analysis
for ArcGIS allows you to rapidly make those comparisons.
The Layer Properties Symbology tab can be a learning tool to
see the effect of stretches on the input and output histograms.
You’ll learn more about these stretches in chapter 6 “Using
Radiometric Enhancement”.

Exercise 3: Identifying similar areas in an image
With Image Analysis for ArcGIS you can quickly identify
areas with similar characteristics. This is useful for
identification of environmental disasters or burn areas. Once
an area has been defined, it can also be quickly saved into a
shapefile. This action lets you avoid the need for manual
digitizing. To define the area, you use the Seed Tool to point
to an area of interest such as a dark area on an image
depicting an oil spill. The Seed Tool returns a graphic
polygon outlining areas with similar characteristics.
Add and draw an Image Analysis for
ArcGIS theme depicting an oil spill
1. If you are starting immediately after the previous
exercise, clear your view by clicking the New Map File
button on your ArcMap tool bar. You do not need to
save the image. If you are beginning here, start ArcMap
and load the Image Analysis for ArcGIS extension.
2. Click the Add Data button.
3. In the Add Data dialog, select radar_oilspill.img, and
click Add to draw it in the view.
This is a radar image showing an oil spill off the
northern coast of Spain.
Create a shapefile
In this exercise, you use the Seed Tool (also called the
Region Growing Tool). The Seed Tool grows a polygon
graphic in the image that encompasses all similar and
contiguous areas. In order to use the Seed Tool, you will first
need to create a shapefile in ArcCatalog and start editing in
order to enable the Seed Tool. After going through these
steps, you can point and click inside the area you want to
highlight, in this case an oil spill, and create a polygon. The
polygon enables you to see how much of an area the oil spill
covers.
1. Click the Zoom In tool, and drag a rectangle around the
black area to see the spill more clearly.
1 2

2. Click the ArcCatalog button. You can store the shapefile
you’re going to create in the example data directory or
navigate to a different directory if you wish.
3. Select the directory in the Table of contents and right
click or click File, point to New, and click Shapefile.
4. In the Create New Shapefile dialog, name the new
shapefile oilspill, and click the Feature Type dropdown
arrow and select Polygon.
5. Check Show Details.
6. Click Edit.
7. In the Spatial Reference Properties dialog, click Import,
and select radar_oilspill.img and click Add from the
Browse for Dataset dialog that will pop up containing
the example data directory.
9. Click OK in the Create New Shapefile dialog.
10. Select the oilspill shapefile, and drag and drop it in the
ArcMap window. Oilspill will appear in the Table of
contents.
11. Close ArcCatalog.
1
2
3
9
4
65

Draw the polygon with the Seed Tool
1. Click the Image Analysis dropdown arrow, and click
Seed Tool Properties.
2. Type a Seed Radius of 10 pixels in the Seed Radius text
box.
3. Uncheck the Include Island Polygons box.
The Seed Radius is the number of pixels surrounding the
target pixel. The range of values of those surrounding
pixels is considered when the Seed Tool grows the
polygon.
4. Click OK.
5. Click the Editor toolbar button on the ArcMap toolbar to
display the Editor toolbar.
6. Click Editor on the Editor toolbar in ArcMap, and select
Start Editing.
8
7
2
3
4
1

7. Click the Seed Tool and click a point in the center of the
oil spill. The Seed Tool will take a few moments to
produce the polygon.
This is a polygon of an oil spill grown by the Seed Tool.
If you don’t automatically see the formed polygon in the
image displayed in the view, click the refresh button at the
bottom of the view screen in ArcMap.
You can see how the tool identifies the extent of the spill. An
emergency team could be informed of the extent of this
disaster in order to effectively plan a clean up of the oil.
6
5
6

Exercise 4: Finding areas of change
The Image Analysis for ArcGIS extension allows you to see
changes over time. You can perform this type of analysis on
either continuous data using Image Difference or thematic
data using Thematic Change. In this exercise, you’ll learn
how to use Image Difference and Thematic Change. Image
Difference is useful for analyzing images of the same area to
identify land cover features that may have changed over time.
Image Difference performs a subtraction of one theme from
another. This change is highlighted in green and red masks
depicting increasing and decreasing values.
Find changed areas
In the following example, you are going to work with two
continuous data images of the north metropolitan Atlanta,
Georgia, area—one from 1987 and one from 1992.
Continuous data images are those obtained from remote
sensors like Landsat and SPOT. This kind of data measures
reflectance characteristics of the earth’s surface, analogous to
exposed film capturing an image. You will use Image
Difference to identify areas that have been cleared of
vegetation for the purpose of constructing a large regional
shopping mall.
Add and draw the images of Atlanta
and load the Image Analysis for ArcGIS extension.
3. Press the Shift or Ctrl key, and click on
atl_spotp_87.img and atl_spotp_92.img in the Add Data
dialog.
4. Click OK.
With images active in the view, you can calculate the
difference between them.
Compute the difference due to
development
1. Click the Image Analysis dropdown arrow, click
Utilities, and click Image Difference.

2. In the Image Difference dialog, click the Before Theme
dropdown arrow, and select Atl_spotp_87.img.
3. Click the After Theme dropdown arrow, and select
Atl_spotp_92.img. 4. Choose As Percent in the Highlight Changes box.
5. Click the arrows to 15 in the Increases more than box.
6. Click the arrows to 15 in the Decreases more than box.
7. Navigate to the directory where you want to store your
Image Difference file, type the name of the file, and
click Save.
8. Navigate to the directory where you want to store your
Highlight Change file, type the name of the file, and
click Save.
9. Click OK in the Image Difference dialog.
The Highlight Change and Image Difference files
appear in the Table of contents and the view.
1
9
3
8
4
6
5
7
2

Highlight Change shows the difference in red and green
areas.
10. In the Table of contents, click the check box to turn off
Highlight Change, and check Image Difference to
display it in the view.
The Image Difference image shows the results of the
subtraction of the Before Theme from the After Theme.
Image Difference calculates the difference in pixel values.
With the 15 percent parameter you set, Image Difference
finds areas that are at least 15 percent increased than before
(designated clearing) and highlights them in green. Image
Difference also finds areas that are at least 15 percent
decreased than before (designating an area that has increased
vegetation or an area that was once dry, but is now wet) and
highlights them in red.
Close the view
You can now clear the view and either go to the next portion
of this exercise, Thematic Change, or end the session by
closing ArcMap. If you want to shut down ArcMap with
Image Analysis for ArcGIS, click the File menu, and click
Exit. Click No when asked to save changes.
Using Thematic Change
Image Analysis for ArcGIS provides the Thematic Change
feature to make comparisons between thematic data images.
Thematic Change creates a theme that shows all possible
combinations of change and how an area’s land cover class
changed over time. Thematic Change is similar to Image
Difference in that it computes changes between the same area
at different points in time. However, Thematic Change can
only be used with thematic data (data that is classified into
distinct categories). An example of thematic data is a
vegetation class map.
This next example uses two images of an area near Hagan
Landing, South Carolina. The images were taken in 1987 and
1989, before and after Hurricane Hugo. Suppose you are the
forest manager for a paper company that owns a parcel of
land in the hurricane’s path. With Image Analysis for
ArcGIS, you can see exactly how much of your forested land
has been destroyed by the storm.

Add the images of an area damaged by
Hurricane Hugo
button on your ArcMap toolbar. You do not need to save
the image. If you are beginning here, start ArcMap and
load the Image Analysis for ArcGIS extension.
2. Open a new view and click Add Data.
3. Press either the Shift key or Ctrl key, and select both
tm_oct87.img and tm_oct89.img in the Add Data
dialog. Click Add.
This view shows an area damaged by Hurricane Hugo.
Create three classes of land cover
Before you calculate Thematic Change, you must first
categorize the Before and After Themes. You can access
Categorize through Unsupervised Classification, which is an
option available from the Image Analysis dropdown menu.
You’ll use the thematic themes created from those
classifications to complete the Thematic Change calculation.
1. Click the dropdown arrow in the Layers section of the
Image Analysis toolbar to make sure tm_oct87.img is
active.
2. Click the Image Analysis dropdown arrow, point to
Classification, and click Unsupervised/Categorize.
3. Click the Input Image dropdown arrow to make sure
tm_oct87.img is in the text box.
4. Click the arrows to 3 or type 3 in the Desired Number of
Classes box.
5. Navigate to the directory where you want to store the
output image, type the file name (use
unsupervised_class_87 for this example), and click
Save.
6. Click OK in the Unsupervised Classification dialog.
3
4
5
6
1

Using Unsupervised Classification to categorize continuous
images into thematic classes is particularly useful when you
are unfamiliar with the data that makes up your image. You
simply designate the number of classes you would like the
data divided into, and Image Analysis for ArcGIS performs a
calculation assigning pixels to classes depending on their
values. By using Unsupervised Classification, you may be
better able to quantify areas of different land cover in your
image. You can then assign the classes names like water,
forest, and bare soil.
7. Click the check box of tm_oct87.img so the original
theme is not drawn in the view. This step makes the
remaining themes draw faster in the view.
Give the classes names and assign colors
to represent them
1. Double-click the title unsupervised_class_87.img to
access the Layer Properties dialog.
2. Click the Symbology tab.
3. Verify that Class_names is selected in the Value Field.
4. Select Class 001, and double-click Class 001 under
Class_names. Type the name Water.
5. Double-click the color bar under Symbol for Class 001,
and choose blue from the color palette.
Class_names. Type the name Forest.
and choose green.
Class_names. Type the name Bare Soil.
and choose a tan or light brown color.
5
3
4
10
2

Categorize and name the areas in the post-
hurricane image
1. Follow the steps provided for the theme tm_oct87.img
on pages 25 and 26 under “Create three classes of land
cover” and “Give the classes names and assign colors to
represent them” to categorize the classes of the
tm_oct89.img theme.
2. Click the box of the tm_oct89.img theme so that it does
not draw in the view.
Recode to permanently write class names
and colors to a file
After you have classified both of your images, you need to do
a recode in order to permanently save the colors and class
names you have assigned to the images. Recode lets you
create a file with the specific images you’ve classified.
1. Click the Image Analysis dropdown arrow, point to GIS
Analysis, and click Recode.
2. Click the Input Image dropdown arrow to select one of
the classified images.
3. The Map Pixel Value through Field will read <From
view>. Leave this as is.
4. Click the browse button to bring up your working
directory, and name the Output Image.
5. Click OK.
5
4
3
2
1

Now do the same thing and perform a recode on the other
classified image you did of the Hugo area. Both of the images
will have your class names and colors permanently saved.
Use Thematic Change to see how land
cover changed because of Hugo
1. Make sure both recoded images are checked in the Table
of contents so both will be active in the view.
2. Click the Image Analysis dropdown arrow, point to GIS
Analysis, and click Thematic Change.
3. Click the Before Theme dropdown arrow and select the
87 classification image.
4. Click the After Theme dropdown arrow, and select the
89 classification image.
5. Navigate to the directory where you want to store the
Output Image, type the file name, and click Save.
6. Click OK.
7. Click the check box of Thematic Change to draw it in
the view.
8. Double-click the Thematic Change title to access Layer
Properties.
9. In the Symbology tab, double-click the symbol for was:
Class 002, is now: Class 003 (was Forest, is now Bare
Soil) to access the color palette.
10. Click the color red in the color palette, and click Apply.
You don’t have to choose red, you can use any color you
like.
11. Click OK.
You can see the amount of destruction in red. The red
shows what was forest and is now bare soil.
Add a feature theme that shows the
property boundary
Using Thematic Change, the overall damage caused by the
hurricane is clear. Next, you will want to see how much
damage actually occurred on the paper company’s land.
1. Click Add Data.
2. Select property.shp, and click Add.
5
4
3
6

Thematic Change image with the property shapefile
Make the property transparent
1. Double-click on the property theme to access Layer
Properties.
2. Click the Symbology tab, and double-click the color
symbol.
3. In the Symbol Selector, click the Hollow symbol.
4. Click the Outline Width arrows, or type the number 3 in
the box.
5. Click the Outline Color dropdown arrow, and choose a
color that will easily stand out to show your property
line.
6. Click OK.
7. Click Apply and OK on the Symbology tab.
The yellow outline clearly shows the devastation within
the paper company’s property boundaries.
5
4
3
6

Exercise 5: Mosaicking images
Image Analysis for ArcGIS allows you to mosaic multiple
images. When you mosaic images, you join them together to
form one single image that covers the entire area. To mosaic
images, simply display them in the view, ensure that they
have the same number of bands, then select Mosaic.
In the following exercise, you are going to mosaic two
airphotos with the same resolution.
Add and draw the images
and load the Image Analysis for ArcGIS extension with
a new map.
3. Press the Shift key and select Airphoto1.img and
Airphoto2.img in the Add Data dialog. Click Add.
4. Click Airphoto1.img and drag it so that it is at the top of
the Table of contents.
The two airphotos display in the view. The Mosaic tool
joins them as they appear in the view: whichever is on
top is also on top in the mosaicked image.
Zoom in to see image details
1. Select Airphoto1.img, and right-click your mouse.
2. Click Zoom to raster resolution.
The two images are displayed at a 1:1 resolution. You
can now use Pan to see how they overlap.
3. Click the Pan button, then maneuver the images in the
view.

This illustration shows where the two images overlap.
4. Click the Full Extent button so that both images display
their entirety in the view.
Use Mosaic to join the images
1. If you want to use some other extent than Union of
Inputs for your mosaic, you must first go to the Extent
tab in the Options dialog and change the Extent before
opening Mosaic Images. After opening the Mosaic
Images dialog, you cannot access the Options dialog.
However, it is recommended that you keep the default of
Union of Inputs for mosaicking. 2. Click the Image Analysis dropdown arrow, point to Data
Preparation, and click Mosaic Images.
3. Click the Handle Images overlaps dropdown arrow and
choose Use Order Displayed.
3
4
1

4. If you want to automatically crop your images, check
the box, and use the arrows or type the percentage by
which to crop the images.
5. Choose Brightness/Contrast as the Color Balancing
option.
6. If you have changed the extent to something other than
Union of Inputs, check this box, but for this exercise you
will need to leave the extent set at Union of Inputs and
the box unchecked.
7. Navigate to the directory where you want to save your
files, type the file name, and click Save.
8. Click OK.
The Mosaic function joins the two images as they
appear in the view. In this case Airphoto1 is mosaicked
over Airphoto2.
3
5
7
6
8
4

Exercise 6: Orthorectification of camera imagery
The Image Analysis for ArcGIS extension for ArcGIS has a
feature called Geocorrection Properties. The function of this
feature is to rectify imagery. One of the tools that makes up
Geocorrection Properties is the Camera model.
In this exercise you will orthorectify images using the
Camera model in Geocorrection Properties.
Add raster and feature datasets
and load the Image Analysis for ArcGIS extension with
a new map.
3. Hold the Shift key down and select both ps_napp.img
and ps_streets.shp in the Add Data dialog. Click Add.
4. Right click on ps_napp.img and click Zoom to Layer.
The images are drawn in the view. You can see the
fiducial markings around the edges and at the top.
Select the coordinate system for the image
This procedure defines the coordinate system for the data
frame in Image Analysis for ArcGIS.
1. Either select Layers in the Table of contents and right
click, or move your cursor into the view and right click.
2. Select Properties at the bottom of the menu to bring up
the Data Frame Properties dialog.
3. Click the Coordinate System tab.
4. In the box labeled Select a coordinate system, click
Predefined.
7
5
3
6
4

5. Click Projected Coordinate Systems, and then click
Utm.
6. Click NAD 1927, then click NAD 1927 UTM Zone
11N.
7. Click Apply, and click OK.
Orthorectifying your image using
Geocorrection Properties
1. Click the Model Types dropdown arrow, and click
Camera.
2. Click the Geocorrection Properties button on the toolbar
to open the Camera dialog.
3. Click the Elevation tab, and select File to use as the
Elevation Source.
4. Navigate to the ArcGIS ArcTutor directory, and choose
ps_dem.img as the Elevation File.
5. Click the Elevation Units dropdown arrow and select
Meters.
6. Check Account for Earth’s curvature.
7. Click the Camera tab.
8. Click the Camera Name dropdown arrow, and select
Default Wild.
9. In the Principal Point box, enter -0.004 for X and 0.000
for Y.
10. Enter a Focal Length of 152.804.
11. Click the arrows, or type 4 for the number of Fiducials.
12. Click in the Film X and Film Y box where the number
of Fiducials will reduce to 4.
2
1
4
5
6
3

13. Type the following coordinates in the corresponding
fiducial spaces. Use the Tab key to move from space to
space.
1. -106.000 106.000
2. 105.999 105.994
3. 105.998 -105.999
4. -106.008 -105.999
14. Name the camera in the Camera Name box.
15. Click Save to save the camera information with the
Camera Name.
16. Click Apply and move to the next section.
Fiducial placement
1. Click the Fiducials tab, and make sure the first fiducial
orientation is selected.
2. Click the Green fiducial, and the software will take you
to the approximate location of the first fiducial
placement. Your cursor has become a crosshair.
3. Click the Fixed Zoom In tool, and zoom in until you can
see the actual fiducial, and click the crosshair there. The
software will take you to each of the four points where
you can click the crosshair in the fiducial marker.
When you are done placing fiducials, make sure to click
Apply then OK to close. You can then right click on the
image in the Table of contents, and click Zoom to Layer. You
will notice that both the image and the shape file are now
displayed in the view. To look at the root mean square error
(RMSE) on the fiducials tab, you can reopen the Camera
Properties dialog. The RMSE should be less than 1.0. Now,
it is time to rectify the images.
8
11
10
9
7
14
12
16
15
1
2
3

After placing fiducials, both the image and the shapefile
are shown in the view for rectification.
Placing links
1. Click the Add Links button.
2. Looking closely at the image and shapefile in the view,
and using the next image as a guide, line up where you
should place the first link. Follow the markers in the
next image to place the first three links. You will need to
click the crosshair on the point in the image first and
then drag the cursor over to the point in the shapefile
where you want to click.
Your first link should look approximately like this:
3. Place links 2 and 3.

After placing the third link, your image should look
something like this:
4. Zoom to the upper left portion of the image, and place a
link according to this next image.
5. Zoom to the lower left portion of the image, and place a
link according to the previous image.
Your image should warp and become aligned with the streets
shapefile. You can use the Zoom tool to draw a rectangle
around the aligned area and zoom in to see it more clearly.
Now take a look at the RMS Error on the Links tab of Camera
Properties. You can go to Save As on the Image Analysis
menu and save the image if you wish.

What’s Next?
This tutorial has introduced you to some features and basic
functions of Image Analysis for ArcGIS. The following
chapters go into greater detail about the different tools and
elements of Image Analysis for ArcGIS, and include
instructions on how to use them to your advantage.

39
3Applying data tools
You will notice when you look at the Image Analysis menu that there are three
choices called Seed Tool Properties, Image Info, and Options. All three aid you in
manipulating, analyzing, and altering your data so you can produce results that are
easier to interpret than they would be with no data tool input.
• Seed Tool Properties automatically generates feature layer polygons of similar
spectral value.
• Image Info gives you the ability to apply a NoData Value and recalculate
statistics.
• Options lets you change extent, cell size, preferences, and more.
IN THIS CHAPTER
• Seed Tool Properties
• Image Info
• Options
3

Using Seed Tool Properties
As stated in the opening of the chapter, the main function of Seed
Tool Properties is to automatically generate feature layer polygons
of similar spectral value. After creating a shapefile in ArcCatalog,
you can either click in an image on a single point, or you can click
and drag a rectangle in a portion of the image that interests you.
You can decide which method you wish to use before clicking the
tool on the toolbar, or you can experiment with which method looks
best with your data.
In order to use the Seed Tool, you must first create the shapefile for
the image you are using in ArcCatalog. You will need to open
ArcCatalog, create a new shapefile in the directory you want to use,
name it, choose polygon as the type of shapefile, and then use Start
Editing on the Editor toolbar in ArcMap to activate the Seed Tool.
Once you are finished and you have grown the polygon, you can go
back to the Editor toolbar and select Stop Editing.
The band or bands used in growing the polygon are controlled by
the current visible bands as set in Layer Properties. If you only have
one band displayed, such as the red band, when you are interested
in vegetation analysis, then the Seed Tool only looks at the statistics
of that band to create the polygon. If you have all the bands (red,
green, and blue) displayed, then the Seed Tool evaluates the
statistics in each band of data before creating the polygon.
When a polygon shapefile is being edited, a polygon defined using
the Seed Tool is added to the shapefile. Like other ArcGIS
graphics, you can change the appearance of the polygon produced
by the Seed Tool using the Graphics tools.
Controlling the Seed Tool
You can use the Seed Tool simply by choosing it from the Image
Analysis toolbar and clicking on an image after generating a
shapefile. The defaults usually produce a good result. However, if
you want more control over the parameters of the Seed Tool, you
can open Seed Tool Properties from the Image Analysis menu.
Seed Tool dialog
Seed Radius
When you use the simple click method, the Seed Tool is controlled
by the Seed Radius. You can change the number of pixels of the
Seed Radius by opening the dialog from the Image Analysis menu.
From this dialog, you select your Seed Radius in pixels. The Image
Analysis for ArcGIS default Seed Radius is 5 pixels.
The Seed Radius determines how selective the Seed Tool is when
selecting contiguous pixels. A larger Seed Radius includes more
pixels to calculate the range of pixel values used to grow the
polygon, and typically produces a larger polygon. A smaller Seed
Radius uses fewer pixels to determine the range. Setting the Seed
Radius to 0.5 or less restricts the polygon to growing over pixels
with the exact value as the pixel you click on in the image. This can
be useful for thematic images in which a contiguous area might
have a single pixel value, instead of a range of values like
continuous data.

APPLYING DATA TOOLS 41
Island Polygons
The other option on the Seed Tool Properties dialog is Include
Island Polygons. You should leave this option checked for use with
Find Like Areas. For single feature mapping where you want to see
a more refined boundary, you may want to turn it off.

Preparing to use the Seed Tool
Go through the following steps to activate the Seed Tool and
generate a polygon in your image.
1. Open ArcCatalog and make sure your working directory
appears in ArcCatalog, or navigate to it.
2. Click File, point to New, and click Shapefile.
3. Rename the New_Shapefile.
4. Click the dropdown arrow and select Polygon.
5. Check Show Details.
6. Click Edit.
1
2
9
4
65
3

7. Click Select, Import, or New to input the coordinate
system the new shapefile will use. Clicking Import will
allow you to import the coordinates of the image you are
creating the shapefile for.
8. Click Apply and OK in the Spatial Reference Properties
dialog.
9. Click OK in the Create New Shapefile dialog.
10. Close ArcCatalog and click the dropdown arrow on the
Editor toolbar.
11. Select Start Editing.
8
7
11

Using the Seed Tool
These processes will take you through steps to change the
Seed Radius and include Island Polygons. For an in-depth
tutorial on using the Seed Tool and generating a polygon, see
chapter 2 “Quick-start tutorial”.
Changing the Seed Radius
1. Click the Image Analysis dropdown arrow, and click Seed
Tool Properties.
2. Type a new value in the Seed Radius text box.
3. If you need to enable Include Island Polygons, check the
box.
4. Click OK.
After growing the polygon in the image with the Seed Tool, go
back to the Editor toolbar, click the dropdown arrow, and click
Stop Editing.
2
1
3
4

Image Info
When analyzing images, you often have pixel values you need to
alter or manipulate in order to perceive different parts of the image
better. The Image Info feature of Image Analysis for ArcGIS lets
you choose a NoData Value and recalculate the statistics for your
image so that a pixel value that is unimportant in your image can be
designated as such.
You can apply NoData to a single layer of your image instead of to
the entire image if you want or need to do so. When you choose to
apply NoData to single layers, it is important that you click Apply
on the dialog before moving to the next layer. You can also
recalculate statistics (Recalc Stats) for single bands by choosing
Current Band in the Statistics box on the Image Info dialog. It is
important to remember that if you click Recalc Stats while Current
Band is selected, Image Info will only recalculate the statistics for
that band. If you want to set NoData for a single band, but
recalculate statistics for all bands, you can choose All Bands after
setting NoData in the single bands, and recalculate for all.
The Image Info dialog is found on the Image Analysis menu. When
you choose it, the images in your view will be displayed on a
dropdown menu under Layer Selection. You can then type the pixel
value that you wish to give the NoData pixels in your image. The
Statistics portion of the dialog also features a dropdown menu so
you can designate the layer for which to calculate NoData. This
area of the dialog also names the Pixel Type and the Minimum and
Maximum values. When you click Recalc Stats, the statistics for the
image are recalculated using the NoData Value, and you can close
the image in the view, then reopen it to see the NoData Value
applied. The Representation Type area of the dialog will
automatically choose Continuous or Thematic depending on what
kind of image you have in your view. If you find that a file you need
to be continuous is listed as thematic, you can change it here.
NoData Value
The NoDataValue section of the Image Info dialog gives you the
opportunity to label certain areas of your image as NoData. In order
to do this, you assign a certain value that no other pixel in the image
has to the pixels you want to classify as NoData. You will want to
do this when the pixel values in that particular area of the image are
not important to your statistics or image. You have to assign some
type of value to those pixels to hold their place, so you need to come
up with a value that's not being used for any of the other pixels you
want to include. Using 0 does not work because 0 does contain
value. Look at the Minimum value and the Maximum value under
Statistics on the Image Info dialog and choose your NoData value
to be any number between the Minimum and Maximum.
Sometimes the pixel value you choose as NoData will already be
used so that NoData matches some other part of your image. This
problem becomes evident when the image is displayed in the view
and there are black spots or triangles where it should be clear, or
perhaps clear spots where it should be black. Also remember that
you can type N/A or leave the area blank so that you have no
NoData assigned if you don't want to use this option.

Using the Image Info dialog
1. Click the Image Analysis dropdown arrow, and click Image
Info.
2. Click the Layer Selection dropdown arrow to make sure the
correct image is displayed.
3. Click the Statistics dropdown arrow to make sure the layer
you want to recalculate is selected.
4. Choose All Bands or Current Band.
5. Type the NoDataValue in the box.
6. Make sure the correct Representation Type is chosen for
your image.
7. Click Recalc Stats.
9. Close the image and re-open to view the results visually.
7
2
1
5
3
6
4
8

Options
You can access the Options dialog through the Image Analysis
menu. Through this dialog, you can set an analysis mask as well as
setting the extent, cell size, and preferences for future operations or
a single operation. It’s usually best to leave the options set at what
they are, but there may be times you want or need to change them.
When you’re mosaicking images, you can go to the Extent tab on
the Options dialog in order to set the extent at something other than
Union of Inputs, which it automatically defaults to when
mosaicking. The default extent is usually Intersection of Inputs. It
is recommended that you leave the default Union of Inputs when
mosaicking, but you can change it. If you do so, you will need to
check the Use Extent from Analysis Options box on the Mosaic
Image dialog. You can use the Options dialog with any Image
Analysis feature, but you may find it particularly useful with the
Data Preparation features that will be covered in the next chapter.
The Options dialog has four tabs on it for General, Extent, Cell
Size, and Preferences. On the General tab, your output directory is
displayed, and the Analysis mask will default to none, but if you
click the dropdown arrow, you can set it to any raster dataset. If you
want to store your output images and shapefiles in one working
directory, you can navigate to that directory or type the directory
name in the Working directory box. This will allow your working
directory to automatically come up every time you click the browse
button for an output image. The Analysis Coordinate System lets
you choose which coordinate system you would like the image to
be saved with—the one for the input or the one for the active data
frame. Finally, you can select whether or not to have a warning
message display if raster inputs have to be projected during analysis
operation.
The Image Analysis Options dialog
Extent
The Extent tab lets you control how much of a theme you want to
use during processing. You do this by setting the Analysis extent.
The rest of the tab will become active when Same as Display, As
Specified below, and Same as Layer "......" (whatever layer is active
in the view) are chosen. Same as Display refers to the area currently
displayed in the view. If the view has been zoomed in on a portion
of a theme, then the functions would only operate on that portion of
the theme. When you choose Same as Layer, all of the information
in the Table of contents for that layer is considered regardless of
whether or not they are displayed in the view. As Specified below
lets you fill in the information for the extent. You can also click the
open file button on the Extent tab to choose a dataset to use as the
Analysis extent. If you click this button, you can navigate to the
directory where your data is stored and select a file that has extents
falling within the selected project area.

The other options on the Analysis extent dropdown list are
Intersection of Inputs and Union of Inputs. When you choose
Intersection (which is the default extent for all functions except
Mosaic), Image Analysis for ArcGIS performs functions on the
area of overlap common to the input images to the function.
Portions of the images outside the area of overlap are discounted
from analysis. Union is the default setting of Analysis extent for
mosaicking. When the extent is set to Union of Inputs, Image
Analysis for ArcGIS uses the union of every input theme. It is
highly recommended that you keep this default setting when
mosaicking images.
When you choose an extent that activates the rest of the Extent tab,
the fields are Top, Right, Bottom, and Left. If you are familiar with
the data and want to enter exact coordinates, you can do so in these
fields. Same as Display and As Specified Below activate the Snap
extent to field where you can choose an image to snap the Analysis
mask to.
The Extent tab on the Options dialog
Cell Size
The third tab on the Options dialog is Cell Size. This is for the cell
size of images you produce using Image Analysis for ArcGIS. The
first field on the tab is a dropdown list for Analysis cell size. You
can choose Maximum of Inputs, Minimum of Inputs, As Specified
below, or Same as Layer ".....". Choosing Maximum of Inputs
yields an output that has the maximum resolution of the input files.
For example, if you use Image Difference on a 10 meter image and
a 20 meter image, the output is a 20 meter image.
The Minimum of Inputs option produces an output that has the
minimum resolution of the input files. For example, if you use
Image Difference on a 10 meter image and a 20 meter image, the
output is a 10 meter image.
When you choose As Specified below, you can enter whatever cell
size you wish to use, and Image Analysis for ArcGIS will adjust the
output accordingly.
If you choose Same as Layer "....", indicating a layer in the view,
the cell size reflects the current cell size of that layer.
The Cell Size field will display in either meters or feet. To choose
one, click View in ArcMap, click Data Frame Properties, and on the
General Tab, click the dropdown arrow for Map Units and choose
either Feet or Meters.
The Number of Rows and Number of Columns fields should not be
updated manually as they will update as analysis properties are
changed.

The Cell Size tab on the Options dialog
Preferences
It is recommended that you leave the preference choice to the
default of Bilinear Interpolation, but you can change it to Nearest
Neighbor or Cubic Convolution if your data requires one of those
choices. Bilinear Interpolation is a resampling method that uses the
data file values of four pixels in a 2 × 2 window to calculate an
output data file value by computing a weighted average of the input
data file values with a bilinear function.
The Nearest Neighbor option is a resampling method in which the
output data file value is equal to the input pixel that has coordinates
closest to the retransformed coordinates of the output pixel.
The Cubic Convolution option is a resampling method that uses the
data file values of sixteen pixels in a 4 × 4 window to calculate an
output data file value with a cubic function.
The Preferences tab on the Options dialog

Using the Options dialog
The following processes will take you through the parts you
can change on the Options dialog.
The General Tab
1. Click the Image Analysis dropdown arrow, and click
Options.
2. Navigate to the Working directory if it’s not displayed in
the box.
3. Click the dropdown arrow and select the Analysis mask if
you want one, or navigate to the directory where it is
stored.
4. Choose the Analysis Coordinate System.
5. Check or uncheck the Display warning box according to
your needs.
6. Click the Extent tab to change Extents or OK to finish.
3
2
1
6
4
5

The Extent Tab
1. Click the dropdown arrow for Analysis extent, and
choose an extent, or navigate to a directory to choose a
dataset for the extent.
2. If the coordinate boxes are on, you can type in
coordinates if you know the exact ones to use.
3. If activated, click the dropdown arrow, and choose an
image to Snap extent to, or navigate to the directory
where it is stored.
4. Click the Cell Size tab, or OK.
2
1
4
3

Cell Size tab
1. Click the dropdown arrow, and choose the cell size, or
navigate to the directory where it is stored.
2. If activated, type the cell size you want to use.
3. Type the number of rows.
4. Type the number of columns.
5. Click the Preferences tab or OK.
The Preferences tab has only the one option of clicking the
dropdown arrow and choosing to resample using either
Nearest Neighbor, Bilinear Interpolation, or Cubic
Convolution.
1
2
3
4
5

Working with features
Section 2

55
4Using Data Preparation
When using the Image Analysis for ArcGIS extension, it is sometimes necessary
to prepare your data first. It is important to understand how to prepare your data
before moving on to the different ways Image Analysis for ArcGIS gives you to
manipulate your data. You are given several options for preparing data in Image
Analysis for ArcGIS.
In this chapter you will learn how to:
• Create a new image
• Subset an image
• Mosaic images
• Reproject an image
IN THIS CHAPTER
• Create New Image
• Subset Image
• Mosaic Images
• Reproject Image
4

Create New Image
The Create New Image function makes it easy to create a new
image file. It also allows you to define the size and content of the
file as well as choosing whether or not the new image type will be
thematic or continuous.
Choose thematic for raster layers that contain qualitative and
categorical information about an area. Thematic layers lend
themselves to applications in which categories or themes are used.
They are used to represent data measured on a nominal or ordinal
scale, such as soils, land use, land cover, and roads.
Continuous data is represented in raster layers that contain
quantitative (measuring a characteristic on an interval or ratio
scale) and related, continuous values. Continuous raster layers can
be multiband or single band such as Landsat, SPOT, digitized
(scanned) aerial photograph, DEM, slope, and temperature.
With this feature, you also get to choose the value of columns and
rows (the default value is 512, but you can change that) and you
choose the data type as well. The data type determines the type of
numbers and the range of values that can be stored in a raster layer.
The Number of Layers allows you to select how many layers to
create in the new file.
The Initial Value lets you choose the number to initialize the new
file. Every cell is given this value.
When you are finished entering your information into the fields,
you can click OK to create the image, or Cancel to close the dialog.
Data Type
Minimum
Value
Maximum
Value
Unsigned 1 bit 0 1
Unsigned 2 bit 0 3
Unsigned 4 bit 0 15
Unsigned 8 bit 0 255
Signed 8 bit -128 127
Unsigned 16 bit 0 65,535
Signed 16 bit -32,768 32,767
Unsigned 32 bit
Signed 32 bit -2 billion 2 billion
Float Single

USING DATA PREPARATION 57
Creating a new image
1. Click the Image Analysis dropdown arrow, point to Data
Preparation, and click Create New Image.
2. Navigate to the directory where the Output Image should
be stored.
3. Choose Thematic or Continuous as the Output Image
Type.
4. Type or click the arrows to enter how many Columns or
Rows if different from the default number of 512.
5. Click the dropdown arrow to choose the Data Type.
6. Type or click the arrows to enter Number of Layers.
7. Type or click the arrows to enter the Initial Value.
8. Click OK.
1
3
5
7
8
4
6
2

Subset Image
This function allows you to copy a portion (a subset) of an input
data file into an output data file. This may be necessary if you have
an image file that is much larger than the particular area you need
to study. Subset Image has the advantage of not only eliminating
extraneous data, but it also speeds up processing as well, which can
be important when dealing with multiband data.
The Subset Image function works on multiband continuous data to
separate that data into bands. For example, if you are working with
a TM image that has seven bands of data, you may wish to make a
subset of bands 2, 3, and 4, and discard the rest.
The Subset Image function can be used to subset an image either
spatially or spectrally. You will probably spatially subset more
frequently than spectrally. To subset spatially, you first bring up the
Options dialog, which allows you to apply a mask or extent or set
the cell size. These options are used for all Image Analysis for
ArcGIS functions including Subset Image. Spatial subsets are
particularly useful if you have a large image and you only want to
subset part of it for analysis. You can use the Zoom In tool to draw
a rectangle around the specific area you wish to subset and go from
there. If you wish to subset an image spectrally, you do it directly
in the Subset Image dialog by entering the desired band numbers to
extract from the image.
Following are illustrations of a TM image of the Amazon as it
undergoes a spectral subset.
This feature is also accessible from the Utilities menu.
The Amazon TM image before subsetting
Amazon TM after a spectral subset

The next illustrations reflect images using the spatial subsetting
option.
The image of the Pentagon before spatial subsetting
In order to specify the particular area to subset, you click the Zoom
In tool, draw a rectangle over the area, open the options dialog, and
select Same As Display on the Extent tab. The rectangle is defined
by Top, Left, Bottom, and Right coordinates. Top and Bottom are
measured as the locations on the Y-axis and the Left and Right
coordinates are measured on the X-axis. You can then save the
subset image and work from there on your analysis.
The Options dialog
The Pentagon subset image after setting the Analysis Extent in Options

Subsetting an image spectrally
1. Click Add Data to add the image to the view.
2. Double-click the image name in the Table of contents to
open Layer Properties.
3. Click the Symbology tab in Layer Properties.
4. Click Stretched in the Show panel.
5. Click the Band dropdown arrow, and select the layer you
want to subset.
1
6
534

Preparation, and click Subset Image.
8. Click the Input Image dropdown arrow, and click the file
you want to use, or navigate to the directory where it is
stored.
9. Using a comma for separation, type the band numbers
you want to subset in the text box.
10. Type the file name of the Output Image, or navigate to
the directory where it should be stored.
11. Click OK.
10
8
7
9
11

Subsetting an image spatially
1. Click the Add Data button to add your image.
2. Click the Zoom In tool, and draw a rectangle over the
area you want to subset.
3. Click the Image Analysis menu, and click Options.
4. Click the Extent tab.
5. Click the Analysis extent dropdown arrow, and select
Same As Display.
7. Click the Image Analysis dropdown arrow and click Save
As, and save the image in the appropriate directory.
3
7
5
6
4
2
1

Mosaic Images
Mosaicking is the process of joining georeferenced images together
to form a larger image. The input images must all contain map and
projection information, although they need not be in the same
projection or have the same cell sizes. Calibrated input images are
also supported. All input images must have the same number of
layers. You can mosaic single or multiband continuous data, or
thematic data.
It is extremely important when mosaicking to arrange your images
in the view as you want the output theme to appear before you
mosaic them. Image Analysis for ArcGIS mosaics images strictly
based on their appearance in the view. This allows you to mosaic a
large number of images without having to make them all active.
It is also important that the images you plan to mosaic contain the
same number of bands. You cannot mosaic a seven band TM image
with a six band TM image. You can, however, use Subset Image to
subset bands from an existing image and then mosaic regardless of
the number of bands they originally contained.
You can mosaic images with different cell sizes or resolutions.
When this happens you can consult the settings in the Image
Analysis Options dialog for Cell Size. The Cell Size is initially set
to the maximum cell size so if you mosaic two images, one with a
4-meter resolution and one with a 5-meter resolution, the output
mosaicked image has a 5-meter resolution. You can set the Cell
Size in the Options dialog to whatever cell size you like so that the
output mosaicked image has the cell size you selected.
The Extent tab on the Options dialog will default to Union of Inputs
for mosaicking images. If, for some reason, you want to use a
different extent, you can change it in the Options dialog and check
the Use Extent from Analysis Options box on the Mosaic Images
dialog. It is recommended that you leave it at the default of Union
of Inputs.
Another Options feature to take note of is the Preferences tab. For
mosaicking images, you should resample using Nearest Neighbor.
This will ensure that the mosaicked pixels do not differ in their
appearance from the original image. Other resampling methods use
averages to compute pixel values and can produce an edge effect.
When you apply Mosaic, the images are processed using whatever
stretch you’ve specified in the Layer Properties dialog. During
processing, each image is fed through its own lookup table, and the
output mosaicked image has the stretch built in, and should be
viewed with no stretch. This allows you to adjust the stretch of each
image independently to achieve the desired overall color balance.
With the Mosaic tool you are also given a choice of how to handle
image overlaps by using the order displayed, maximum value,
minimum value, or average value.
Choose:
Order Displayed — replaces each pixel in the overlap area with the
pixel value of the image that is on top in the view.
Maximum Value — in order to replace each pixel in the overlap
area with the greater value of corresponding pixels in the
overlapping images.
Minimum Value — replaces each pixel of the overlap area by the
lesser value of the corresponding pixels in the overlapping images.
Average Value — replaces each pixel in the overlap area with the
average of the values of the corresponding pixels in the overlapping
images.

The color balancing options let you choose between balancing by
brightness/contrast, histogram matching, or none. If you choose
brightness/contrast, the mosaicked image will be balanced by
utilizing the adjustments you have made in Layer Properties/
Symbology. If you choose Histogram Matching, the input images
are adjusted to have similar histograms to the top of the image in
the view. Select None if you don’t want the pixel values adjusted.

How to Mosaic Images
1. Add the images you want to mosaic to the view.
2. Arrange images in the view in the order that you want
them in the mosaic.
Preparation, and click Mosaic Images.
4. Click the Handle Image Overlaps by dropdown arrow,
and click the method you want to use.
5. If you want the images automatically cropped, check the
box, and enter the Percent by which to crop the images.
6. Choose the Color Balance method.
7. Check the box if you want to use the extent you set in
Analysis Options.
be stored.
9. Click OK.
For more information on mosaicking images, see chapter 2
“Quick-start tutorial’’.
2
3
5
7
4
6
8

Reproject Image
Reproject Image gives you the ability to reproject raster image data
from one map projection to another. Reproject Image, like all
Image Analysis for ArcGIS functions, observes the settings in the
Options dialog so don’t forget to use Options to set Extent, Cell
Size, and so on if so desired.
ArcMap has the capability to reproject images on the fly by simply
setting the desired projection and choosing View/Data Frame
Properties and selecting the Coordinate System tab. The desired
projection may then be selected. After you select the coordinate
system, you apply it and go to Reproject Image n Image Analysis
for ArcGIS.
At times you may need to produce an image in a specific projection.
By having the desired output projection specified in the Data Frame
Properties, the only things you need to specify in Reproject Image
are the input and output images.
Before Reproject Image
Here is the reprojected image after changing the Coordinate System
to Mercator (world):
After Reproject Image

How to Reproject an Image
1. Click Add Data, and add the image you want to reproject
to the view.
2. Right-click in the view, and click on Properties to bring up
the Data Frame Properties dialog.
3. Click on the Coordinate System tab.
4. Click Predefined and choose whatever coordinate
system you want to use to reproject the image.
2
4
1
5
3

Preparation, and click Reproject Image.
7. Click the Input Image dropdown arrow and click the file
stored.
be stored.
9. Click OK.
6
7
8
9

69
1Performing Spatial Enhancement
Spatial Enhancement is a function that enhances an image using the values of
individual and surrounding pixels. Spatial Enhancement deals largely with spatial
frequency, which is the difference between the highest and lowest values of a
contiguous set of pixels. Jensen (1986) defines spatial frequency as “the number of
changes in brightness value per unit distance for any part of an image.”
There are three types of spatial frequency:
• zero spatial frequency — a flat image, in which every pixel has the same value
• low spatial frequency — an image consisting of a smoothly varying gray scale
• high spatial frequency — an image consisting of drastically changing pixel
values such as a checkerboard of black and white pixels
The Spatial Enhancement feature lets you use convolution, non-directional edge,
focal analysis, and resolution merge to enhance your images. Depending on what
you need to do to your image, you will select one feature from the Spatial
Enhancement menu. This chapter will focus on the explanation of these features as
well as how to apply them to your data.
This chapter is organized according to the order in which the Spatial Enhancement
tools appear. You may want to skip ahead if the information you are seeking is
about one of the tools near the end of the menu list.
IN THIS CHAPTER
• Convolution
• Non-Directional Edge
• Focal Analysis
• Resolution Merge
5

Convolution
Convolution filtering is the process of averaging small sets of pixels
across an image. Convolution filtering is used to change the spatial
frequency characteristics of an image (Jensen 1996).
A convolution kernel is a matrix of numbers that is used to average
the value of each pixel with the values of surrounding pixels. The
numbers in the matrix serve to weight this average toward
particular pixels. These numbers are often called coefficients,
because they are used as such in the mathematical equations.
Applying convolution filtering
Apply Convolution filtering by clicking the Image Analysis
dropdown arrow, and choosing Convolution from the Spatial
Enhancement menu. The word filtering is a broad term, which
refers to the altering of spatial or spectral features for image
enhancement (Jensen 1996). Convolution filtering is one method of
spatial filtering. Some texts use the terms synonymously.
Convolution example
To understand how one pixel is convolved, imagine that the
convolution kernel is overlaid on the data file values of the image
(in one band) so that the pixel to be convolved is in the center of the
window. To compute the output value for this pixel, each value in the
convolution kernel is multiplied by the image pixel value that
corresponds to it. These products are summed, and the total is
divided by the sum of the values in the kernel, as shown in this
equation:
integer [((-1 × 8) + (-1 × 6) + (-1 × 6) +
(-1 × 2) + (16 × 8) + (-1 × 6) +
(-1 × 2) + (-1 × 2) + (-1 × 8))/
: (-1 + -1 + -1 + -1 + 16 + -1 + -1 + -1 + -1)]
= int [(128-40) / (16-8)]
= int (88 / 8) = int (11) = 11
2 8 6 6 6
2 8 6 6 6
2 2 8 6 6
2 2 2 8 6
2 2 2 2 8
Kernel
-1 -1 -1
-1 16 -1
-1 -1 -1
Data

PERFORMING SPATIAL ENHANCEMENT 71
When the 2 × 2 set of pixels near the center of this 5 × 5 image is
convolved, the output values are:
The kernel used in this example is a high frequency kernel. The
relatively lower values become lower, and the higher values
become higher, thus increasing the spatial frequency of the image.
Convolution formula
The following formula is used to derive an output data file value for
the pixel being convolved (in the center):
1 2 3 4 5
1 - - - - -
2 - 11 5 - -
3 - 0 11 - -
4 - - - - -
5 - - - - -
Where:
fij = the coefficient of a convolution kernel at
position i,j (in the kernel)
dij = the data value of the pixel that corresponds to
fij
q = the dimension of the kernel, assuming a square
kernel (if q = 3, the kernel is 3 × 3)
F = either the sum of the coefficients of the kernel,
or 1 if the sum of coefficients is zero
V = the output pixel value
Source: Modified from Jensen 1996; Schowengerdt 1983
The sum of the coefficients (F) is used as the denominator of the
equation above, so that the output values are in relatively the same
range as the input values. Since F cannot equal zero (division by
zero is not defined), F is set to 1 if the sum is zero.
V
fijdij
j 1=
q
∑
 
 
 
 
i 1=
q
∑
F
-----------------------------------=

Zero sum kernels
Zero sum kernels are kernels in which the sum of all coefficients in
the kernel equals zero. When a zero sum kernel is used, then the
sum of the coefficients is not used in the convolution equation, as
above. In this case, no division is performed (F = 1), since division
by zero is not defined.
This generally causes the output values to be:
• zero in areas where all input values are equal (no edges)
• low in areas of low spatial frequency
• extreme in areas of high spatial frequency (high values become
much higher, low values become much lower)
Therefore, a zero sum kernel is an edge detector, which usually
smooths out or zeros out areas of low spatial frequency and creates
a sharp contrast where spatial frequency is high, which is at the
edges between homogeneous (homogeneity is low spatial
frequency) groups of pixels. The resulting image often consists of
only edges and zeros.
Zero sum kernels can be biased to detect edges in a particular
direction. For example, this 3 × 3 kernel is biased to the south
(Jensen 1996).
-1 -1 -1
1 -2 1
1 1 1
High frequency kernels
A high frequency kernel, or high pass kernel, has the effect of
increasing spatial frequency.
High frequency kernels serve as edge enhancers, since they bring
out the edges between homogeneous groups of pixels. Unlike edge
detectors (such as zero sum kernels), they highlight edges and do
not necessarily eliminate other features.
When a high frequency kernel is used on a set of pixels in which a
relatively low value is surrounded by higher values, like this...
...the low value gets lower. Inversely, when the high frequency
kernel is used on a set of pixels in which a relatively high value is
surrounded by lower values...
-1 -1 -1
-1 16 -1
-1 -1 -1
BEFORE AFTER
204 200 197 - - -
201 106 209 - 10 -
198 200 210 - - -

...the high value becomes higher. In either case, spatial frequency is
increased by this kernel.
Low frequency kernels
Below is an example of a low frequency kernel, or low pass kernel,
which decreases spatial frequency.
This kernel simply averages the values of the pixels, causing them
to be more homogeneous. The resulting image looks either more
smooth or more blurred.
BEFORE AFTER
64 60 57 - - -
61 125 69 - 188 -
58 60 70 - - -
1 1 1
1 1 1
1 1 1
Convolution With High Pass
Convolution with High Pass

Apply Convolution
Spatial Enhancement, and click Convolution.
2. Click the Input Image dropdown arrow, and click a file, or
navigate to the directory where the file is stored.
3. Click the Kernel dropdown arrow, and click the kernel you
want to use.
4. Choose Reflection or Background Fill.
be stored.
6. Click OK.
1
3
4
5
6
2
Applying Convolution
Reflection fills in the area beyond the edge of the of the image
with a reflection of the values at the edge. Background fill uses
zeros to fill in the kernel area beyond the edge of the image.
Convolution allows you to perform image enhancement
operations such as averaging and high pass or low pass filtering.
Each data file value of the new output file is calculated by
centering the kernel over a pixel and multiplying the original
values of the center pixel and the appropriate surrounding pixels
by the corresponding coefficients from the matrix. To make
sure the output values are within the general range of the input
values, these numbers are summed and then divided by the sum
of the coefficients. If the sum is zero, the division is not
performed.

Non-Directional Edge
The Non-Directional Edge function averages the results of two
orthogonal first derivative edge detectors. The filters used are the
Sobel and Prewitt filters. Both of these filters are based on a
calculation of the 1st derivative, or slope, in both the x and y
directions. Both use orthogonal kernels convolved separately with
the original image, and then combined.
The Non-Directional Edge is based on the Sobel zero-sum
convolution kernel. Most of the standard image processing filters
are implemented as a single pass moving window (kernel)
convolution. Examples include low pass, edge enhance, edge
detection, and summary filters.
For this model, a Sobel filter has been selected. To convert this
model to the Prewitt filter calculation, the kernels must be changed
according to the example below.
1 0 1–
2 0 2–
1 0 1–
vertical
1– 2– 1–
0 0 0
1 2 1
horizontal
Sobel=
1 0 1–
1 0 1–
1 0 1–
vertical
1– 1– 1–
0 0 0
1 1 1
horizontal
Prewitt=
Image of Seattle before applying Non-Directional Edge
After Non-Directional Edge

Using Non-Directional Edge
Spatial Enhancement, and click Non-Directional Edge.
3. Choose Sobel or Prewitt.
4. Choose Reflection or Background Fill.
6. Click OK.
1
3
4
5
6
2
Using Non-Directional Edge
In step 4, reflection fills in the area beyond the edge of the
image with a reflection of the values at the edge. Background
fill uses zeros to fill in the kernel area beyond the edge of the
image.

Focal Analysis
The Focal Analysis function enables you to perform one of several
types of analysis on class values in an image file using a process
similar to convolution filtering.
This model (Median Filter) is useful for reducing noise such as
random spikes in data sets, dead sensor striping, and other impulse
imperfections in any type of image. It is also useful for enhancing
thematic images.
Focal Analysis evaluates the region surrounding the pixel of
interest (center pixel). The operations that can be performed on the
pixel of interest include:
• Standard Deviation — measure of texture
• Sum
• Mean — good for despeckling radar data
• Median — despeckle radar
• Min
• Max
These functions allow you to select the size of the surrounding
region to evaluate by selecting the window size.
An image before Focal Analysis
After Focal Analysis is performed

Applying Focal Analysis
Spatial Enhancement, and click Focal.
3. Click the Focal Function dropdown arrow, and click the
function you want to use.
4. Click the Neighborhood Shape dropdown arrow, and click
the shape you want to use.
5. Click the Neighborhood Definition dropdown arrow, and
click the Matrix size you want to use.
7. Click OK.
1
3
4
6
7
2
5
Focal Analysis Results
Focal Analysis is similar to Convolution in the process that it
uses. With Focal Analysis, you are able to perform several
different types of analysis on the pixel values in an image file.

Resolution Merge
The resolution of a specific sensor can refer to radiometric, spatial,
spectral, or temporal resolution. This function merges imagery of
differing spatial resolutions.
Landsat TM sensors have seven bands with a spatial resolution of
28.5 m. SPOT panchromatic has one broad band with very good
spatial resolution—10 m. Combining these two images to yield a
seven-band data set with 10 m resolution provides the best
characteristics of both sensors.
A number of models have been suggested to achieve this image
merge. Welch and Ehlers (1987) used forward-reverse RGB to IHS
transforms, replacing I (from transformed TM data) with the SPOT
panchromatic image. However, this technique is limited to three
bands (R,G,B).
Chavez (1991), among others, uses the forward-reverse principal
components transforms with the SPOT image, replacing PC-1.
In the above two techniques, it is assumed that the intensity
component (PC-1 or I) is spectrally equivalent to the SPOT
panchromatic image, and that all the spectral information is
contained in the other PCs or in H and S. Since SPOT data does not
cover the full spectral range that TM data does, this assumption
does not strictly hold. It is unacceptable to resample the thermal
band (TM6) based on the visible (SPOT panchromatic) image.
Another technique (Schowengerdt 1980) additively combines a
high frequency image derived from the high spatial resolution data
(i.e., SPOT panchromatic) with the high spectral resolution Landsat
TM image.
The Resolution Merge function uses the Brovey Transform method
of resampling low spatial resolution data to a higher spatial
resolution while retaining spectral information:
Brovey Transform
In the Brovey Transform, three bands are used according to the
following formula:
DNB1_new = [DNB1 / DNB1 + DNB2 + DNB3] ×
[DNhigh res. image]
[DNhigh res. image]
[DNhigh res. image]
Where:
B = band
The Brovey Transform was developed to visually increase contrast
in the low and high ends of an image’s histogram (i.e., to provide
contrast in shadows, water and high reflectance areas such as urban
features). Brovey Transform is good for producing RGB images
with a higher degree of contrast in the low and high ends of the
image histogram and for producing visually appealing images.
Since the Brovey Transform is intended to produce RGB images,
only three bands at a time should be merged from the input
multispectral scene, such as bands 3, 2, 1 from a SPOT or Landsat
TM image or 4, 3, 2 from a Landsat TM image. The resulting
merged image should then be displayed with bands 1, 2, 3 to RGB.

Resolution Merge
Spatial Enhancement, and click Resolution Merge.
2. Click the High Resolution Image dropdown arrow, and
click a file, or navigate to the directory where the file is
stored.
3. Click the Multi-Spectral Image dropdown arrow, and click
a file, or navigate to the directory where the file is stored.
be stored.
5. Click OK.
1
3
4
5
2
Using Resolution Merge
Use Resolution Merge to integrate imagery of different spatial
resolutions (pixel size).

The following images display the Resolution Merge function:
High Resolution Image Multi-Spectral Image
Resolution Merge

83
1Using Radiometric Enhancement
Radiometric enhancement deals with the individual values of the pixels in an
image. It differs from Spatial Enhancement, which takes into account the values of
neighboring pixels.
Radiometric Enhancement consists of functions to enhance your image by using
the values of individual pixels within each band. Depending on the points and the
bands in which they appear, radiometric enhancements that are applied to one band
may not be appropriate for other bands. Therefore, the radiometric enhancement of
a multiband image can usually be considered as a series of independent, single-
band enhancements (Faust 1989).
IN THIS CHAPTER
• LUT (Lookup Table) Stretch
• Histogram Equalization
• Histogram Matching
• Brightness Inversion
6

LUT Stretch
LUT Stretch creates an output image that contains the data values
as modified by a lookup table. The output is 3 bands.
Contrast stretch
When radiometric enhancements are performed on the display
device, the transformation of data file values into brightness values
is illustrated by the graph of a lookup table.
Contrast stretching involves taking a narrow input range and
stretching the output brightness values for those same pixels over a
wider range. This process is done in Layer Properties in Image
Analysis for ArcGIS.
Linear and nonlinear
The terms linear and nonlinear, when describing types of spectral
enhancement, refer to the function that is applied to the data to
perform the enhancement. A piecewise linear stretch uses a
polyline function to increase contrast to varying degrees over
different ranges of the data.
Linear contrast stretch
A linear contrast stretch is a simple way to improve the visible
contrast of an image. It is often necessary to contrast-stretch raw
image data, so that they can be seen on the display.
In most raw data, the data file values fall within a narrow range—
usually a range much narrower than the display device is capable of
displaying. That range can be expanded to utilize the total range of
the display device (usually 0 to 255).
Nonlinear contrast stretch
A nonlinear spectral enhancement can be used to gradually increase
or decrease contrast over a range, instead of applying the same
amount of contrast (slope) across the entire image. Usually,
nonlinear enhancements bring out the contrast in one range while
decreasing the contrast in other ranges.
Piecewise linear contrast stretch
A piecewise linear contrast stretch allows for the enhancement of a
specific portion of data by dividing the lookup table into three
sections: low, middle, and high. It enables you to create a number
of straight line segments that can simulate a curve. You can
enhance the contrast or brightness of any section in a single color
gun at a time. This technique is very useful for enhancing image
areas in shadow or other areas of low contrast.
A piecewise linear contrast stretch normally follows two rules:
1. The data values are continuous; there can be no break in the
values between High, Middle, and Low. Range specifications
adjust in relation to any changes to maintain the data value
range.
2. The data values specified can go only in an upward,
increasing direction.
The contrast value for each range represents a percentage of the
available output range that particular range occupies. Since rules 1
and 2 above are enforced, as the contrast and brightness values are
changed, they may affect the contrast and brightness of other
ranges. For example, if the contrast of the low range increases, it
forces the contrast of the middle to decrease.

USING RADIOMETRIC ENHANCEMENT 85
Contrast stretch on the display
Usually, a contrast stretch is performed on the display device only,
so that the data file values are not changed. Lookup tables are
created that convert the range of data file values to the maximum
range of the display device. You can then edit and save the contrast
stretch values and lookup tables as part of the raster data image file.
These values are loaded into the view as the default display values
the next time the image is displayed.
The statistics in the image file contain the mean, standard deviation,
and other statistics on each band of data. The mean and standard
deviation are used to determine the range of data file values to be
translated into brightness values or new data file values. You can
specify the number of standard deviations from the mean that are to
be used in the contrast stretch. Usually the data file values that are
two standard deviations above and below the mean are used. If the
data has a normal distribution, then this range represents
approximately 95 percent of the data.
The mean and standard deviation are used instead of the minimum
and maximum data file values because the minimum and maximum
data file values are usually not representative of most of the data. A
notable exception occurs when the feature being sought is in
shadow. The shadow pixels are usually at the low extreme of the
data file values, outside the range of two standard deviations from
the mean.
Varying the contrast stretch
There are variations of the contrast stretch that can be used to
change the contrast of values over a specific range, or by a specific
amount. By manipulating the lookup tables as in the following
illustration, the maximum contrast in the features of an image can
be brought out.
This figure shows how the contrast stretch manipulates the
histogram of the data, increasing contrast in some areas and
decreasing it in others. This is also a good example of a piecewise
linear contrast stretch, which is created by adding breakpoints to the
histogram.

Apply LUT Stretch Class
Radiometric Enhancement, and click LUT Stretch.
2. Click the Input Image dropdown arrow, and click the file
stored.
be stored. Set the output type to TIFF.
4. Click OK.
1
3
4
2
LUT Stretch Class
LUT Stretch Class provides a means of producing an output
image that has the stretch built into the pixel values to use with
packages that have no stretching capabilities.

Erdas image analysis for arcgis

Erdas image analysis for arcgis

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Viewers also liked

Viewers also liked (20)

Similar to Erdas image analysis for arcgis

Similar to Erdas image analysis for arcgis (20)

Recently uploaded

Recently uploaded (20)

Erdas image analysis for arcgis