CAR RECOGNITION SYSTEM
USING
MATLAB
Project Supervisor
Sir Umer Javed
Group Members
Sania Arif (1547)
Namra Afzal (1528)
Laraib Mumtaz
(1522)
Batch F11
BSEE Faculty of Engineering and Technology
IIUI
WHY DID WE CHOOSE THIS PROJECT?
 Identification of stolen cars
 Smuggling of Cars
 Invalid license plates
 Usage of cars in terrorist attacks/illegal activities
 Applications in traffic systems (highway electronic toll
collection, red light violation enforcement, border and
customs checkpoints, etc.).
2
AIM
 We intended to develop a system in MATLAB which
can perform detection as well as recognition of Car
Number plate
 The objective of this project is to recognize car
number plate using serial communication.
3
WORK DIVISION
 Matlab code ( Namra Afzal 1528 /BSEE /FET F11)
 Microcontroller interfacing (Sania Arif / 1547 /BsEE
FET f11)
 Hardware ( Laraib Mumtaz/ 1522/ BSEE /FET F11)
4
TIME DIVISION
Week 1
 Studied project
 Project Planning
Week 2
matlab coding
Week 3
 Simulation using proteus
 Hardware
Week 4
Hardware Interfacing 5
BASIC PROJECT
 Input image ( from real environment)
 Algorithm using (matlab)
 output
Microcontroller serial interfacing with hardware.
6
WORKFLOW
 Image was taken from real environment .
 Process Digital Images of License Plates using
existing/modified algorithms.
 Algorithms will perform alpha numeric conversions on
the captured license plate images into text entries.
 System would check the extracted entries against a
database in real time.
 The entire system is implemented in MATLAB is used for
detection and recognition . 7
BASIC MODULES OF THE SYSTEM
 Detection is done by Character Segmentation
Locates the alpha numeric characters on a license
plate.
 Optical Character Recognition (OCR)
Translates the segmented characters into text entries.
8
Block Diagram
9
Start
Localization
End
Characters And Numbers Segmentation
Feature Extraction Of Segmented
Image
Recognize The Extracted Features
Show The License Plate
10
LICENSE PLATE LOCALIZATION
11
Edge Detection
Morphological
Operations
Extracting The
Plate Region
Flow Chart of extraction in
Matlab
12
13
Start
Load Image From File
Morphological Operations Are Applied On The Image
Convert Image Into Grayscale
Median Filter To remove noise in The Image
Edge enhancement In The Image
Convolution for brightening image
Intensity scaling
Show The License Plate
Filling all the regions of Image
Thinning to isolate characters
End
LOAD THE IMAGE FROM FILE
14
a=imread(‘car 10.JPG’)
PREPROCESSING
Preprocessing is very important for the good
performance of character segmentation.
Preprocessing consists of :
 Resizing image
 Rgb to gray
 Noise removal ( we used median filter) .
CHANGING THE TYPE
16
c=rgb2gray(b);
EXTRACTING PLATE REGION
 It is result of dilation after noise removal .
17
18
EDGE ENHANCEMENT
19
gdiff=imsubtract(d,e);
where ‘d’ is dilated and
‘e’ is eroded image
MORPHOLOGICAL OPERATIONS
 Filling (holes )
 Thinning ( for character isolation)
 Finding connected components of area more than
200 pixels
20
CHARACTERS SEGMENTATION
21
Preprocessing Horizontal And
Vertical Segmentation
HORIZONTAL & VERTICAL
SEGMENTATION
 Detect the horizontal lines in the image with a pixel value
of zero.
 Converting the image into binary.
 Use simple “for loops” to detect the portions of the image
that had connected objects with a pixel value of ‘0’ and
hence accordingly, the image was read.
22
CHARACTERS RECOGNITION
23
Template Matching
Template matching is one of the most
common and easy classification method
for recognizing the characters.
We used code for OCR
TEMPLATE MATCHING
The used templates are given in the figure below:
24
OUTPUT
Correlation is used to match the image from the license
plate and the template’s image. The following figure
shows the numbers in a text file.
26
27
EXPERIMENTAL RESULTS
28
WHY CHOSE MATLAB FOR PROJECT
 To move to a Real Time Environment.
 For fast computation.
29
PROBLEMS WITH THE MATLAB
SYSTEM
The problems that we faced during Localization were:
 Algorithm did not work perfectly for more than one
image.
 Manual Changes were required in the code every time ,
manually we had to change parameters in code that was
kind of hit and trial method.
30

Number plate recognition system using matlab.

  • 1.
    CAR RECOGNITION SYSTEM USING MATLAB ProjectSupervisor Sir Umer Javed Group Members Sania Arif (1547) Namra Afzal (1528) Laraib Mumtaz (1522) Batch F11 BSEE Faculty of Engineering and Technology IIUI
  • 2.
    WHY DID WECHOOSE THIS PROJECT?  Identification of stolen cars  Smuggling of Cars  Invalid license plates  Usage of cars in terrorist attacks/illegal activities  Applications in traffic systems (highway electronic toll collection, red light violation enforcement, border and customs checkpoints, etc.). 2
  • 3.
    AIM  We intendedto develop a system in MATLAB which can perform detection as well as recognition of Car Number plate  The objective of this project is to recognize car number plate using serial communication. 3
  • 4.
    WORK DIVISION  Matlabcode ( Namra Afzal 1528 /BSEE /FET F11)  Microcontroller interfacing (Sania Arif / 1547 /BsEE FET f11)  Hardware ( Laraib Mumtaz/ 1522/ BSEE /FET F11) 4
  • 5.
    TIME DIVISION Week 1 Studied project  Project Planning Week 2 matlab coding Week 3  Simulation using proteus  Hardware Week 4 Hardware Interfacing 5
  • 6.
    BASIC PROJECT  Inputimage ( from real environment)  Algorithm using (matlab)  output Microcontroller serial interfacing with hardware. 6
  • 7.
    WORKFLOW  Image wastaken from real environment .  Process Digital Images of License Plates using existing/modified algorithms.  Algorithms will perform alpha numeric conversions on the captured license plate images into text entries.  System would check the extracted entries against a database in real time.  The entire system is implemented in MATLAB is used for detection and recognition . 7
  • 8.
    BASIC MODULES OFTHE SYSTEM  Detection is done by Character Segmentation Locates the alpha numeric characters on a license plate.  Optical Character Recognition (OCR) Translates the segmented characters into text entries. 8
  • 9.
  • 10.
    Start Localization End Characters And NumbersSegmentation Feature Extraction Of Segmented Image Recognize The Extracted Features Show The License Plate 10
  • 11.
    LICENSE PLATE LOCALIZATION 11 EdgeDetection Morphological Operations Extracting The Plate Region
  • 12.
    Flow Chart ofextraction in Matlab 12
  • 13.
    13 Start Load Image FromFile Morphological Operations Are Applied On The Image Convert Image Into Grayscale Median Filter To remove noise in The Image Edge enhancement In The Image Convolution for brightening image Intensity scaling Show The License Plate Filling all the regions of Image Thinning to isolate characters End
  • 14.
    LOAD THE IMAGEFROM FILE 14 a=imread(‘car 10.JPG’)
  • 15.
    PREPROCESSING Preprocessing is veryimportant for the good performance of character segmentation. Preprocessing consists of :  Resizing image  Rgb to gray  Noise removal ( we used median filter) .
  • 16.
  • 17.
    EXTRACTING PLATE REGION It is result of dilation after noise removal . 17
  • 18.
  • 19.
  • 20.
    MORPHOLOGICAL OPERATIONS  Filling(holes )  Thinning ( for character isolation)  Finding connected components of area more than 200 pixels 20
  • 21.
  • 22.
    HORIZONTAL & VERTICAL SEGMENTATION Detect the horizontal lines in the image with a pixel value of zero.  Converting the image into binary.  Use simple “for loops” to detect the portions of the image that had connected objects with a pixel value of ‘0’ and hence accordingly, the image was read. 22
  • 23.
    CHARACTERS RECOGNITION 23 Template Matching Templatematching is one of the most common and easy classification method for recognizing the characters. We used code for OCR
  • 24.
    TEMPLATE MATCHING The usedtemplates are given in the figure below: 24
  • 25.
    OUTPUT Correlation is usedto match the image from the license plate and the template’s image. The following figure shows the numbers in a text file.
  • 26.
  • 27.
  • 28.
  • 29.
    WHY CHOSE MATLABFOR PROJECT  To move to a Real Time Environment.  For fast computation. 29
  • 30.
    PROBLEMS WITH THEMATLAB SYSTEM The problems that we faced during Localization were:  Algorithm did not work perfectly for more than one image.  Manual Changes were required in the code every time , manually we had to change parameters in code that was kind of hit and trial method. 30

Editor's Notes

  • #11 Localization is basically highlighting some parts and ignoring other stuff in background .
  • #12 Egde (enhancement + brightening by (diff + convolve) Morphological ( e, d , fill , thin ) connected components
  • #20 Then edge brightening was done by convolve command