FRUIT DETECTION USING
MORPHOLOGICAL IMAGE
PROCESSING TECHNIQUE
Done By Ahmad Taweel
1
1. Outline
• Introduction
• Image Acquisition
• Image Preprocessing
• Color Feature
• Texture Feature
• Result
• Conclusion
2
2. Introduction
• Quality inspection
• Manual
• Not accurate
• Time consuming
• Expensive
• Fruit quality grading is mandatory condition
• New techniques in fruit quality assessment is necessary
3
2. Introduction
• Locating fruit from tree made
• Harvesting more efficient
• Analysis more easier
• Food appearance  Evaluation of quality of fresh food
• Packinghouses demanded a system
• Detect fruit skin defects
4
3.1 Image Acquisition
• Our first step here is to collect sample images of fruits
which are going to be decided the number of fruit present
on the tree.
• Store all the images in .Jpg format
• Resolution 429*322 pixels
5
3.1 Image Acquisition
6
3.2 Image Preprocessing
• The Aim of this process is to
• Improve image data
• Enhance some of its features
• Approaches:
• Image Enhancement
• Noise remove
7
3.2.1 Image Enhancement
• Improve visibility
• Remove any flickering
• Improve contrast
• Approaches:
• Spatial domain
• Frequency domain
8
3.2.2 Noise Remove
• To remove noise we use Masking
• Consist of:
• Filter mask
• Point to point
• Filter values are predefined
• Used:
• Classify
• Grade
9
3.2 Image Preprocessing
10
3.3 Feature Extraction
• Methods:
• Color Feature
• Color space conversion method
• Texture Feature
• Canny edge detection method
• Dilation method
11
3.4 Image Segmentation
• Method:
• Segmentation Partition clustering algorithm
• K partition each partition represent a cluster
• Drawback
• Poor result when a point is close to the center of another cluster
• Overlapping
12
3.4 Image Segmentation
13
3.5 Canny Edge Detection
• What is Edges?
• Significant transitions in pixels
• Canny Edge Detection ?
• Multi-stage algorithm
• Subject to noise in the image
• How does it work?
• Remove the noise
• Canny Edge Detection
14
3.5 Canny Edge Detection
15
3.6 Morphology (Dilation)
• Morphological Image Processing relay only on the relative
ordering of pixel values, not on their numerical values
• Dilation adds pixels to the boundaries of objects in an
image
• The number of pixels added depends on the size and
shape of the structuring element
• Applied:
• Gray scale images
• Binary image
16
3.6 Morphology (Dilation)
17
RESULT (1/2)
• Counting objects is done
• Manually
• By costly electronic systems
• Image Processing using MATLAB
• Effective
• Quick
• Low cost
• No costly equipment
18
RESULT (2/2)
• Accuracy depends on size of the disk structuring element
• Big objects  more counting accuracy
• Small objects  less counting accuracy
• Accuracy can be increased by separating conglutination
among the objects
19
20
CONCLUSION (1/2)
• Quality inspection is done
• Manually
• highly inconsistence in accuracy
• Time consuming
• Boring
• Expensive
21
CONCLUSION (2/2)
• Image Processing algorithms
• Automatically count
• Distinguish fruits
• Packinghouses can use this system to guarantee quality
of products.
22
Thank You
23

Fruit detection using morphological

  • 1.
    FRUIT DETECTION USING MORPHOLOGICALIMAGE PROCESSING TECHNIQUE Done By Ahmad Taweel 1
  • 2.
    1. Outline • Introduction •Image Acquisition • Image Preprocessing • Color Feature • Texture Feature • Result • Conclusion 2
  • 3.
    2. Introduction • Qualityinspection • Manual • Not accurate • Time consuming • Expensive • Fruit quality grading is mandatory condition • New techniques in fruit quality assessment is necessary 3
  • 4.
    2. Introduction • Locatingfruit from tree made • Harvesting more efficient • Analysis more easier • Food appearance  Evaluation of quality of fresh food • Packinghouses demanded a system • Detect fruit skin defects 4
  • 5.
    3.1 Image Acquisition •Our first step here is to collect sample images of fruits which are going to be decided the number of fruit present on the tree. • Store all the images in .Jpg format • Resolution 429*322 pixels 5
  • 6.
  • 7.
    3.2 Image Preprocessing •The Aim of this process is to • Improve image data • Enhance some of its features • Approaches: • Image Enhancement • Noise remove 7
  • 8.
    3.2.1 Image Enhancement •Improve visibility • Remove any flickering • Improve contrast • Approaches: • Spatial domain • Frequency domain 8
  • 9.
    3.2.2 Noise Remove •To remove noise we use Masking • Consist of: • Filter mask • Point to point • Filter values are predefined • Used: • Classify • Grade 9
  • 10.
  • 11.
    3.3 Feature Extraction •Methods: • Color Feature • Color space conversion method • Texture Feature • Canny edge detection method • Dilation method 11
  • 12.
    3.4 Image Segmentation •Method: • Segmentation Partition clustering algorithm • K partition each partition represent a cluster • Drawback • Poor result when a point is close to the center of another cluster • Overlapping 12
  • 13.
  • 14.
    3.5 Canny EdgeDetection • What is Edges? • Significant transitions in pixels • Canny Edge Detection ? • Multi-stage algorithm • Subject to noise in the image • How does it work? • Remove the noise • Canny Edge Detection 14
  • 15.
    3.5 Canny EdgeDetection 15
  • 16.
    3.6 Morphology (Dilation) •Morphological Image Processing relay only on the relative ordering of pixel values, not on their numerical values • Dilation adds pixels to the boundaries of objects in an image • The number of pixels added depends on the size and shape of the structuring element • Applied: • Gray scale images • Binary image 16
  • 17.
  • 18.
    RESULT (1/2) • Countingobjects is done • Manually • By costly electronic systems • Image Processing using MATLAB • Effective • Quick • Low cost • No costly equipment 18
  • 19.
    RESULT (2/2) • Accuracydepends on size of the disk structuring element • Big objects  more counting accuracy • Small objects  less counting accuracy • Accuracy can be increased by separating conglutination among the objects 19
  • 20.
  • 21.
    CONCLUSION (1/2) • Qualityinspection is done • Manually • highly inconsistence in accuracy • Time consuming • Boring • Expensive 21
  • 22.
    CONCLUSION (2/2) • ImageProcessing algorithms • Automatically count • Distinguish fruits • Packinghouses can use this system to guarantee quality of products. 22
  • 23.