LU504930B1 - A MATLAB BASED LICENSE PLATE RECOGNITION SYSTEM - Google Patents

Abstract

The present invention relates to the technical field of license plate recognition, particularly to a MATLAB-based license plate recognition system comprising a preprocessing module, a feature extraction module and a license plate recognition module; a preprocessing module for preprocessing an input license plate image; a feature extraction module for extracting character features of the license plate from the preprocessed image; a license plate recognition module uses a support vector machine classifier to identify the extracted character features; The advantageous effect is that the MATLAB-based license plate recognition system proposed in the present invention can be quickly implemented by using MATLAB as a development platform, and can be customized and optimized according to specific needs.

Description

A MATLAB-based license plate recognition system LU504930

Technical part

The present invention relates to the technical field of license plate recognition, in particular to a MATL AB-based license plate recognition system.

Technology in the background

As the number of vehicles increases, so does the demand for them

License plate recognition systems in areas such as traffic management and

Security monitoring. A license plate recognition system can

Automatically identify a vehicle's license plate information and improve efficiency and

Improve traffic management accuracy.

In the prior art, many license plate recognition systems are used with the help of

Image processing and pattern recognition techniques implemented.

However, the existing systems have some problems, such as: B. complex

Algorithms, low accuracy rates and high demands on computing power.

Content of the invention

The aim of the present invention is to create a MATLAB-based

To provide license plate recognition system to solve the problems raised in the above prior art.

In order to achieve the above objective, the present invention provides the following technical solution: a MATLAB-based license plate recognition system, the system being a

Preprocessing module, a feature extraction module and a license plate recognition module;

A pre-processing module for pre-processing the input license plate image; a feature extraction module for extracting character features of the

license plate from the preprocessed image; a license plate recognition module that uses a support vector machine classifier to identify the extracted character features.

Preferably, the preprocessing module includes graying, binarization and

Denoising the image.

Preferably, gray scaling converts the color image to a grayscale image to simplify subsequent processing;

Binarization converts the grayscale image into a binary image for better

to enable edge detection and morphological processing;

The denoising process removes the noise from the binary image by applying a

Filter to improve the accuracy of subsequent processing.

Preferably, the feature extraction module includes edge detection, morphological processing and character segmentation of the image.

Preferably, edge detection is used to detect edge information

License plate image used to achieve better segmentation of characters;

Morphological processing improves the accuracy of character segmentation by applying morphological operations to further process the

edge information;

Character segmentation to segment the processed image into individual characters as

Preparation for subsequent detection.

Preferably, the license plate recognition module trains an SVM model zir}/504930

Classification and identification of the extracted characters to obtain recognition results

to obtain a license plate.

Preferably, the license plate recognition system further comprises an image input module and a result output module;

An image input module for receiving an image of the license plate to be recognized and inputting the same via a camera or an image file;

A result output module for outputting recognition results that are stored on the

screen or saved to a file.

The advantageous effects of the present invention compared to the prior art are:

The one proposed in the present invention = MATLAB based

License plate recognition system can be done by using MATLAB as

Development platform can be implemented quickly and depending on the specific

Adapted and optimized to your needs.

Description of the attached drawings

Figure 1 shows a schematic diagram of the present invention loading a

license plate image;

Figure 2 shows a schematic diagram of a histogram of the present invention plotting a grayscale map;

Figure 3 shows a schematic diagram of the present invention according to the

image hollowing;

Figure 4 shows a schematic diagram of the present invention for smoothing a

image;

Figure 5 shows a schematic diagram of the present invention after removing small objects from a binary image;

Figure 6 shows a schematic diagram of the color image of the license plate of the present invention after position cropping;

Figure 7 shows a schematic diagram of the present invention according to the

License plate character segmentation and processing;

Figure 8 shows a schematic diagram of the present invention according to the

Normalization of the license plate.

Detailed description

In order to better understand the purpose of the present invention, the technical solutions proposed for a clear and complete description and the advantages, the following embodiments of the present invention are attached in connection with the

Drawings described in more detail. It should be understood that the specific

Embodiments described herein are part of the embodiments herein

Invention, not all of the embodiments, only for explaining the embodiments of the present invention, and is not used to describe the embodiments of the present

Invention to limit all other embodiments by those of ordinary skill in the art

Territory preserved without creative work under the premise of protecting the

Scope of the present invention.

Embodiment 1

The present invention offers a technical solution: a MATLAB based one

License plate recognition system, the system having a pre-processing module, ek}504980

Feature extraction module and a license plate recognition module;

The preprocessing module is used to preprocess the input

License plate image, including gray scaling, binarization and denoising. At the

Gray scaling converts the color image into a grayscale image for further purposes

to simplify processing. Binarization converts the grayscale image into a binary image to enable better edge detection and morphological processing. The

Denoising removes noise from binary images by applying filters

Improve accuracy of subsequent processing.

The feature extraction module is used to extract the features of the license plate from the preprocessed image. The module includes edge detection, morphological processing and character segmentation. Edge detection is used for

Detecting edge information in the license plate image to better segment the

characters to reach. Morphological processing improves the accuracy of the

Character segmentation by applying morphological operations for further

Processing of edge information. Character segmentation involves dividing the processed image into individual characters for subsequent recognition.

The license plate recognition module uses Support Vector Machine (SVM)

Classifiers to identify the extracted character features. The Support Vector

Machine (SVM) is a widely used pattern recognition method with high accuracy and generalization ability. By training an SVM model, the extracted

Character features are classified and recognized in order to obtain a recognition result

Obtain license plate.

The license plate recognition system of the present invention further includes

Image input module and a result output module. The image input module is used to receive an image of the license plate to be identified via a camera or a

Image file can be entered. The result output module is used to output the

Recognition result that can be displayed on the screen or saved to a file.

The license plate recognition system of the present invention has the following advantages:

Simplicity, efficiency and accuracy. By using MATLAB as

Development platform, the license plate recognition system can be implemented quickly and can be adapted and optimized depending on specific needs.

Embodiment 2

Referring to images 1 to 7 of the accompanying drawings, the specific includes

Realization based on the first embodiment:

I. Preprocessing the license plate image 1. Loading the license plate image: function[d]=main(jpg) [filename,pathname]=uigetfile({‘** jpg’, JPEG file(* jpg)’ }); if(filename==0),return,end global FILENAME?% define global variables.

FILENAME=[pathnamefilename];

I=imread(FILENAME); figure(1),,mshow(D);title(,original image*); %The original license plate image is displayed as shown in Image 1. LU504930 2. Convert the color image to grayscale and draw a histogram: 11=rgb2gray(1); %Convert the color image to a grayscale image; figure(2),subplot(1,2,1),imshow(I1);title(‘Grayscale Image’);. figure(2),subplot(1,2,2),imhist(I1);title(,Grayscale Image Histogram’);% The histogram of the grayscale image is drawn as shown in Figure 2; 3. Edge detection using Roberts operator:

I2=edge(I1,'roberts',0.18,'both');%Select the threshold 0.18, edge detection with the roberts operator figure(3),imshow(I12);title(,edge detection image with the roberts operator Operator*); 4. The image performs an erosion operation: se=[1;1;1];

I3=imerode(12,se);%Implementation of the corrosion operations on the image, i.e. the

Reversing the expansion of the operation figure(4),imshow(I3);title(‘Image after corrosion’); 5. Smooth image se=strel(,Rectangle',[25,25]);%Construct structural elements to construct a se square 14=imclose(I3,se);%Cluster image, fill image figure(5),imshow (I14);title(,Smooth Image');

The result is shown in Figure 3; 6. Remove the small objects from the binary image 15=bwareaopen(14,2000);%remove the part of the cluster whose gray value is less than 2000 figure(6),imshow(I5);title(, Remove small objects from objects' );

The result is shown in Figure 4; [y,x,z]=size(I5);%returns the size of each dimension of IS, stored in x,y,z myl=double(I5);%Conversion of IS to double precision tic%tic means start of time measurement , toc means end of time measurement

Blue y=zeros(y,1);%creates a zero array of y*1 fori=1:y forj=1:x if(myl(i,j,1)==1) %If myl(i, j,1), i.e. the value of the point with coordinates (i,j) in the image of myI is 1, i.e. the point is the license plate background color Blue % then the value of Blue y(i,1) plus 1

Blue y(i,1)=Blue y(i,1)+1;%Statistics of the blue pixel point end end end [tempMaxY]=max(Blue_y);%Y direction license plate size determination %temp is the maximum of the elements of the white y -Vector, MaxY is the index of this

value

PY1=MaxY; while((Blue_y(PY1,1)>=5)&&(PY1>1))

PY1=PY1-1: LU504930 end

PY2=MaxY. while((Blue_y(PY2,1)>=5)&&(PY2<y)) 5 PY2=PY2+1: End end

IY=KPY1:PY2,;,;); %x direction license plate area determination %%%%%%X direction %%%%%%%%%

Blue x=zeros(1,x);%Further determination of the license plate area in the x direction forj=1:x fori=PY1:PY2 if(myl(i,j,1)==1)

Blue _x(1,j)=Blue_x(1,))+1; end end end

PX1=1; while((Blue_x(1,PX1)=1)) while((Blue_x(1,PX1)3)&&(PX1<x))

PX1=PX1+1: end

PX2=x; while((Blue_x(1,PX2)<3)&&(PX2>PX1))

PX2=PX2-1; end end

PX1=PX1-1;% correction of the license plate area

PX2=PX2+1; dw=I(PY1:PY2-8,PX1:PX2,;); t=toc; figure(7),subplot(1,2,1),1mshow(IY),title(, Appropriate area of travel direction’); %Row direction license plate area determination figure(7),subplot(1,2,2),imshow(dw),title(,Positioning the cropped color image of the

License Plate®); The license plate area is shown in Figure 5;

IL. character segmentation and processing 1. further processing of the license plate

Grayscale conversion, binarization, average filtering, corrosion expansion and

Character segmentation of the segmented color license plate image to separate the individual character images from the license plate image to form the license plate, pre-processing the segmented characters (binarization, normalization), and then analyzing the extraction of the segmented character image to identify the textual form of the license plate. The code is as follows: imwrite(dw, 'colored-license.jpg');%Write the colored license plate to the colored license plate file a=imread(,Coloured-label.jpg');%Read the data in the license plate file b =rgb2gray(a);%Convert the license plate image to grayscale imwrite(b, License plate gray image jpg');%Write the gray image to the file LU504930 figure(8);subplot(3,2,1),imshow(b),title (, License plate grayscale image') g_max=double(max(max(b))); g_min=double(min(min(b)));

T=round(g_max-(g max-g min)/3);%T is the threshold for binarization [m,n]=SizeBe(b); d=(double(b)>=T);%d:binarized image imwrite(d, ‘Before average filtering jpg’); subplot(3,2,2),imshow(d),title(' Before mean filtering") % Before mean filtering %Filter h=fspecial('average',3); %Create a predefined filter operator, average is the mean Filters, the

Size of template is 3*3 d=im2bw(round(filter2(h,d)));% Use the specified filter h to filter h by d, i.e. H. the average value imwrite(d, 'After average filter. jpg"); subplot(3,2,3),imshow(d),title(,By Average Filter®) % Some images are manipulated

Yexpand or erode

Yse=strel('square',3);%expand the created image with a 3x3 square

Result item object

Y'Line'"Rhombus'"Sphere'…. se=eye(2);%eye(n)returnsthen-by-nidentitymatrix unity matrix [m,n]=size(d);%return size information for matrix b and store in m,n ifbwarea(d)/m/n> =0.365%Calculate whether the ratio of the total area of the object in

Binary image to total area is greater than 0.365 d=imerode(d,se);% If greater than 0.365, then the image is eroded elseifbwarea(d)/m/n<=0.235%Calculate whether the ratio of the total area of the object to the total area in the binary image less than 0.235, d=imdilate(d,se);% If less than, then the expansion process is carried out end imwrite(d, 'after expansion or corrosion treatment jpg'); subplot(3,2,4),imshow(d),title(,after expansion or corrosion treatment");

The result of the run is shown in Figure 6; 2. Character segmentation

In automatic license plate recognition, character segmentation has the

Task of going forward and going back. It performs character segmentation on the

Based on the preliminary positioning of the license plate and then uses the

Segmentation results for character recognition. There are many algorithms for that

Character recognition because the space between the characters of the license plate is large and no

If a character gets stuck, it looks for a continuous block of text, and if the length is greater than a certain threshold, it is assumed that the block consists of two characters that need to be segmented. In general, segmented characters are further processed to meet the needs of the next step of character recognition. However, license plate recognition does not require much processing 504930 to achieve the purpose of correct recognition. Here only the normalization process is done and then the post-processing is done.

When searching for a continuous block of text whose length is a certain

exceeds the threshold, the block is assumed to consist of two characters and needs to be split %First create the qiege and getword subfunctions, and then call the

Subroutine to split and normalize the license plate characters. d=qiege(d); [m,n]=size(d); %subplot(3,2,5),imshow(d),title(n) k1=1;k2=1;s=sum(d);j=1; whilej~=n whiles(j)==0 j=+1; end kl=j; whiles(j)-=0&&]<=n-1

JH end k2=j-1; itk2-k1>=round(n/6.5) [val,numJ=min(sum(d(;,[k1+5:k2-5]))); d(;,k1+num+5)=0;% cut end end % cut around d=qiege(d); % cut 7 characters y1=10;y2=0.25;flag=0;word1=[]; whileflag==0 [m,n]=size(d); left=1;wide=0; whilesum(d(:,wide+1))~=0 wide=wide+1; end ifwide<y1% think it is a left-handed disorder. d(:,[1-wide])=0; d=giege(d); else temp=qiege(imcrop(d,[11widem])); [m,n]=size(temp); all=sum(sum(temp)); two_thirds=sum(sum(temp([round(m/3):2*round(m/3)],:))); iftwo_thirds/all>y2 flag=1;word1=temp; LU504930 end d(;,[1:wide])=0;d=qiege(d); end end % Split the second character [word2,d]=getword(d); % Split the third character [word3,d]=word(d). % Split the fourth character [word4,d]=word(d). % Split the fifth character [wordS,d]=word(d). % Split the sixth character [word6,d[=word(d). % Split the seventh character [word7,d[=word(d).

Image (9). subplot(2,7,1),imshow(word1),title(‘1’); subplot(2,7,2),imshow(word2),title(‘2’); subplot(2,7,3),imshow(word3),title(‘3’); subplot(2,7,4),imshow(word4),title(‘4’); subplot(2,7,5),imshow(word5),title(‘5’); subplot(2,7,6),imshow(word6),title(‘6’); subplot(2,7,7),imshow(word7),title(‘ 7’); [m,n]=size(word1);

In % of commercial system programs, the normalized size is 40*20, as shown here. word 1=imresize(word1,[4020]); word2=imresize(word2,[4020]); word3=imresize(word3,[4020]); word4=imresize(word4,[4020]); wordS=imresize(word5,[4020]); wordé=imresize(word6,[4020]); word7=imresize(word7,[4020]); subplot(2,7,8),,mshow(word1),title(‘ 1°); subplot(2,7,9),imshow(word2),title(‘2’); subplot(2,7,10),imshow(word3),title(‘3”); subplot(2,7,11),imshow(word4),title(‘4’); subplot(2,7,12),imshow(wordS),title(‘ 5”); subplot(2,7,13),imshow(word6),title(‘ 6°); subplot(2,7,14),imshow(word7),title(‘ 7’); imwrite(word1,’1 jpg’); imwrite(word2,’2 jpg’); imwrite(word3,’3 jpg’);

imwrite(word4,’4.jpg’); LU504930 imwrite(wordS,’5 jpg’); imwrite(word6,’6.jpg’); imwrite(word7,7 jpg); The results of the run are shown in Figure 7; 6. Conclusion

The MATL AB based simulation of the license plate recognition method was carried out by taking 500 license plate images captured by the camera using the methods described in this

Draft proposed procedural method were identified, and ultimately only 20 were used

License plate not recognized correctly, which corresponds to a recognition rate of 96%.

Although embodiments of the present invention have been shown and described, those skilled in the art will recognize that various changes, modifications, substitutions and variations may be made to these embodiments without departing from them

The principles and spirit of the present invention, the scope of which is limited by the appended claims and their equivalents.

Claims (7)

Claims LU504930 1. A MATLAB-based license plate recognition system, characterized in that: the system comprises a preprocessing module, a feature extraction module and a license plate recognition module; A pre-processing module for pre-processing the input license plate image; a feature extraction module for extracting character features of the license plate from the preprocessed image; a license plate recognition module for recognizing the extracted character features using a support vector machine classifier. 2. A MATLAB-based license plate recognition system according to claim 1, characterized in that the pre-processing module includes graying, binarizing and denoising the image. 3. A MATLAB-based license plate recognition system according to claim 2, characterized in that: the gray scaling converts a color image into a grayscale image to simplify subsequent processing; Binarization converts the grayscale image into a binary image to enable better edge detection and morphological processing; The denoising process removes noise from the binary image by applying a filter to improve the accuracy of subsequent processing. 4. A MATLAB-based license plate recognition system according to claim 1, wherein: the feature extraction module includes performing edge detection, morphological processing, and character segmentation on the image. 5. A MATLAB-based license plate recognition system according to claim 4, characterized in that: edge detection is used to detect edge information in the license plate image for better segmentation of the characters; Morphological processing improves the accuracy of character segmentation by applying morphological operations to further process the edge information; Character segmentation segments the processed image into individual characters in preparation for subsequent recognition. 6. A MATLAB-based license plate recognition system according to claim 1, characterized in that the license plate recognition module trains an SVM model to classify and recognize the extracted character and obtains a recognition result of the license plate. 7. A MATLAB-based license plate recognition system according to claim 1, characterized in that the license plate recognition system further comprises an image input module and a result output module; An image input module for receiving an image of a license plate to be recognized and inputting the same via a camera or an image file; a results output module for outputting recognition results to be displayed on a screen or saved to a file.