View
507
Download
3
Category
Preview:
DESCRIPTION
This paper describe how license plate number can be extracted in Assamese language. This was implemented in mat lab.
Citation preview
CHARACTER RECOGNITION FROM NUMBER PLATE WRITTEN
IN ASSAMESE LANGUAGE
Presented ByPresented By
SUBHASH BASISHTHA
DEPARTMENT OF INFORMATION TECHNOLOGY
ASSAM CENTRAL UNIVERSITY
Presentation Outline
Basic Introduction to License Plate Types of License Plate Objectives Of The System Aim Basic Module Of The System Block Diagram of The System Results And Discussion Conclusion References
2
Introduction To License Plate
A vehicle registration plate is a metal or plastic plate attached to a motor vehicle or trailer for official identification purposes[1] .
The registration identifier is a numeric or alphanumeric code that uniquely identifies the vehicle within the issuing region’s database.
3
Plate format of License Plate
Plates for private car and two-wheeler owners have black lettering on a white background [2].
e.g., (TN 81 NZ 0025).
Commercial vehicles such as taxis and trucks have a yellow background and black text
e.g.,
4
AP 32 VA 2223
Contd…
Vehicles belonging to foreign consulates have white lettering on a light blue background.
e.g.,
The President of India and state governors travel in official cars without licence plates. Instead they have the Emblem of India in gold embossed on a red plate[2].
5
22 UN 14
Current format The current format of the registration index consists of 3 parts,
They are
The first two letters indicate the state to which the vehicle is registered.
The next two digit numbers are the sequential number of a district. Due to heavy volume of vehicle registration, the numbers were given to the RTO(Regional Transport officers) offices of registration as well.
The third part is a 4 digit number unique to each plate. A letter(s) is prefixed when the 4 digit number runs out and then two letters and so on[3].
6
AS 32 VA 2223
Objectives
With an everyday increase in the number of cars on our
roads and highways, we are facing numerous problems,
for example[4]:
Identification of stolen cars
Smuggling of Cars
Invalid license plates
Usage of cars in terrorist attacks/illegal activities
7
Aim
To address these issues, we intend to develop a
prototype system in MATLAB which can
perform license plate recognition written in
Assamese Language.
8
Basic Modules of the System
License Plate BinarizationConverting the RGB License plate image to Binary image
Character SegmentationSegment the alpha numeric characters on a license plate written in Assamese Language.
Optical Character Recognition (OCR)Compares the segmented characters with our database set[5].
9
Block Diagram
10
11
Capture
The image of the vehicle is captured using a high resolution mobile camera and after that following steps are performed one by one on that image[6].
12
Binarization
It is the process of converting the captured image into binary image where a fixed value is choose as a standard threshold value and classify all pixels with value above this threshold as a white and all other pixels as black.
13
Noise Removal
Dilation is process of improvising the quality of the captured images. Dilation is used for filling up the holes present in an image, can be used for joining the broken lines and also for noise removal[7].
14
Segmentation
Character segmentation can be done by using the blob analysis.
Matlab function “bwlabel” is used for labeling the connected pixels together in a sequence to form groups of connected objects.
Function “regionprpos” is used to measure the set of properties such as Area, Centroid, and Bounding Box for each connected components.
15
Horizontal & Vertical Segmentation
Detect the horizontal lines in the image with a pixel value of zero.
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[8].
16
Character Recognition
Character recognition is the process of Template matching approach[9].
Matching technique can be performed in three classes.
Direct Matching Deformable Template and Elastic Matching Relaxation Matching
17
Contd…
For resizing, all the input characters images must be equal-sized with the database characters[10].
In our approach, we have resize each segmented images into size 32x15.
18
Contd…19
Template Matching
Template matching is one of the most common and easy
classification method for recognizing the characters.
Template matching is one of the most common and easy
classification method for recognizing the characters.
Data Set
For recognition purpose we need a standard database to compare the segmented characters.
In our case we used 58 characters(out of which10 are numerals) in Assamese language[6].
20
Load The Image From File21
Img=imread(‘filename’)
Binarization22
f1=rgb2gray(img);
Morphological Operations
23
se=strel('rectangle',[15 17]);
img=imdilate(img,se);
img=imfill(img,'holes');
se=strel(‘disk',20);d1-imopen(img,se);
Characters Segmentation24
Normalized characters25
After normalization
Data Set
The below figure shows some samples of our dataset-
26
Experimental Results27
Total License Plate Captured 100
Segmented Characters 895
Correctly Recognized Characters 845
Incorrectly Recognized Characters
50
Accuracy 94.4%
Conclusion & Future Work
License Plate Recognition process requires a very high degree of accuracy when we have to capture image from different angle, different distance, low light etc. These types of anomalies are needed to consider for getting better accuracy. In this paper we have discussed License plate image taken straightly and from 40 meter distance. So in our approach some license image may not detect properly. In future we will work on it to test different images from far distance and various angles. We will also try to include more character samples of various shape and size into our database so that to achieve a higher level accuracy in recognition.
28
REFERENCES
[1] Nafiz Arica and Fatos T. Yarman-Vural. An Overview of Character Recognition Focused on Off-Line Handwriting” IEEE transactions on systems, man, and cybernetics—part c: applications and reviews, vol. 31, no. 2, may 2001.
[2] M.Horowitz, “Efficient use of a picture correlator” J.Opt. Soc. Am vol.47, pp.327,1957
[3] http://www.matlabprojecthelp.com/char-segmantation.
[4]Serkan Ozbay ,Ergun Ercelebi Automatic Vehicle Identification by Plate Recognition World Academy Science Engineering and technology 9 2005
29
Contd…
[5]Jorge Martinez-Carballido, Ruben Alfonso-Lopez, Juan M. Ramirez-Cortes, “License Plate Digit Recognition using 7x5 Binary Templates at an outdoor Parking Lot entrance”IEEE Trans,pp-18-21,2011.
[6]H.K.Chethan,Hemantha Kumar G. Raghavendra.R ,”A Novel Edge Based Method to Extract Text in Camera captured images”IEEE Trans,International Conference on Advances in Computing, and Telecommunication Technologies,pp.853-855,2009
[7]Serkan Ozbay, and Ergun Ercelebi, “Automatic Vehicle Identification by Plate Recognition” proc of World Academy of Science,Engineering and Technology,vol-9,pp.222-225,2005
30
Contd…
[8]S. Banerjee, K. Mullick and U. Bhattacharya, A robust approach to extraction of texts from camera captured images, Proc. of the 5th International Workshop on Camera-Based Document Analysis and Recognition (CBDAR 2013), Washington DC, USA, pp. 53-58, 2013
[9]Prakriti Banik, Ujjwal Bhattacharya and Swapan K. Parui, Segmentation of Bangla Words in Scene Images, Proc. of Indian Conf. on Comp. Vision, Graphics and Image Processing, ACM Conf. Proc. Series, 2012.
31
Contd…
[10]T. Chattopadhyay, U. Bhattacharya, B. B. Chaudhuri, On the enhancement and binarization of mobile captured vehicle identification number for an embedded solution, Proc. of Document Analysis Systems, pp. 235-239, IEEE Comp. Soc. Press, 2012..
32
Recommended