简单车辆检测源码

​​简单车牌检测​​

main：

clc;
clear;
close all;
I=imread('1.png');
figure(1),imshow(I);title('原图');
I1=rgb2gray(I);
figure(2),subplot(1,2,1),imshow(I1);title('灰度图');
figure(2),subplot(1,2,2),imhist(I1);title('灰度图直方图');
I2=edge(I1,'canny',[0.2,0.55]);
figure(3),imshow(I2);title('canny算子边缘检测');
se=[1;1;1];
I3=imerode(I2,se);
figure(4),imshow(I3);title('腐蚀后图像');
se=strel('rectangle',[40,40]);
I4=imclose(I3,se);
figure(5),imshow(I4);title('平滑图像的轮廓');
I5=bwareaopen(I4,2000);
figure(6),imshow(I5);title('从对象中移除小对象');
[y,x,z]=size(I5);
myI=double(I5);
    %begin横向扫描
     Blue_y=zeros(y,1);
      for i=1:y
         for j=1:x
             if(myI(i,j,1)==1) 
          %如果myI(i,j,1)即myI图像中坐标为(i,j)的点为蓝色
          %则Blue_y的相应行的元素white_y(i,1)值加1
           Blue_y(i,1)= Blue_y(i,1)+1;%蓝色像素点统计 
             end  
         end    
      end
[temp MaxY]=max(Blue_y);%temp为向量white_y的元素中的最大值，MaxY为该值的索引（ 在向量中的位置）
PY1=MaxY;
     while ((Blue_y(PY1,1)>=50)&&(PY1>1))
          PY1=PY1-1;
     end    
 PY2=MaxY;
     while ((Blue_y(PY2,1)>=10)&&(PY2<y))
        PY2=PY2+1;
     end
     IY=I(PY1:PY2,:,:);
%IY为原始图像I中截取的纵坐标在PY1：PY2之间的部分
 %end横向扫描
 %begin纵向扫描
      Blue_x=zeros(1,x);%进一步确定x方向的车牌区域
 for j=1:x
         for i=PY1:PY2
             if(myI(i,j,1)==1)
                  Blue_x(1,j)= Blue_x(1,j)+1;               
             end  
         end       
 end
 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 
 PX1=PX1-1;
 PX2=PX2+15;
 PY1=PY1-2.5;
 %end纵向扫描
 dw=I(PY1:PY2,PX1:PX2,:);
 imwrite(dw,'dw.jpg');
figure(7),subplot(1,2,1),imshow(IY),title('行方向合理区域');
figure(7),subplot(1,2,2),imshow(dw),title('定位剪切后的彩色车牌图像')
%利用radon变换做水平矫正
bw=rgb2gray(dw);
bw1=edge(bw,'sobel','horizontal');
theta=0:179;
r=radon(bw1,theta);
[m,n]=size(r);
c=1;
for i=1:m
    for j=1:n
        if r(1,1)<r(i,j)
            r(1,1)=r(i,j);
            c=j;
        end
    end
end
rot=90-c+2;
pic=imrotate(bw,rot,'crop');
figure(8),subplot(3,1,1),imshow(bw),title('1.定位后的这牌灰度图像');
figure(8),subplot(3,1,2),imshow(pic),title('2.利用radon算子做水平方向矫正');
%利用radon变换做垂直方向的矫正
binaryImage = edge(pic,'canny'); 
binaryImage = bwmorph(binaryImage,'thicken'); 
theta = -90:89;
[R,xp] = radon(binaryImage,theta);
%%%%%
[R1,r_max] = max(R);
theta_max = 90;
while (theta_max > 50 || theta_max<-50)
    [R2,theta_max] = max(R1);                      
    R1(theta_max) = 0; 
    theta_max = theta_max - 91;
end
%角度计算完毕
H=[1,0,0; tan(-theta_max),1,0;0,0,1];
T=maketform('affine',H);
pic=imtransform(pic,T);
figure(8),subplot(3,1,3), imshow(pic);title('3.利用radon算子做垂直方向矫正');
%字符分割前的预处理
g_max=double(max(max(pic)));
g_min=double(min(min(pic)));
T=round(g_max-(g_max-g_min)/3); % T 为二值化的阈值
[m,n]=size(pic);
d=im2bw(pic,T/256);  % d:二值图像
figure(9);subplot(2,2,1),imshow(d),title('1.车牌二值图像')
% 滤波
h=fspecial('average',3);
d=im2bw(round(filter2(h,d)));
figure(9),subplot(2,2,2),imshow(d),title('2.均值滤波后')
% 某些图像进行操作
% 膨胀或腐蚀
se=strel('square',3); % 使用一个3X3的正方形结果元素对象对创建的图像膨胀
se=eye(2); % eye(n) returns the n-by-n identity matrix 单位矩阵
[m,n]=size(d);
if bwarea(d)/m/n>=0.365
       d=imerode(d,se);
elseif bwarea(d)/m/n<=0.235
       d=imdilate(d,se);
end
figure(9),subplot(2,2,3),imshow(d),title('3.膨胀或腐蚀处理后')
%对图像的边框进行裁剪，只留下有效字符部分
[y1,x1,z1]=size(d);
I3=double(d);
TT=1;
Y1=zeros(y1,1);
 for i=1:y1
    for j=1:x1
             if(I3(i,j,1)==1) 
                Y1(i,1)= Y1(i,1)+1 ;
            end  
     end       
 end
Py1=1;
Py0=1;
while ((Y1(Py0,1)<30)&&(Py0<y1))
      Py0=Py0+1;
end
Py1=Py0;
 while((Y1(Py1,1)>=30)&&(Py1<y1))
         Py1=Py1+1;
 end
d=d(Py0:Py1,:,:);
figure(9),subplot(2,2,4);
imshow(d),title('4.目标车牌区域');
%在列方向上进行灰度值累加
X1=zeros(1,x1);
for j=1:x1
    for i=1:y1
             if(I3(i,j,1)==1) 
                X1(1,j)= X1(1,j)+1;
            end  
     end       
end
figure(10);
plot(0:x1-1,X1),title('列方向像素点灰度值累计和'),xlabel('列值'),ylabel('累计像素量');
Px0=1;
Px1=1;
%%%%%%%%%%%%分割字符%%%%%%%%%%%%%%%%%%
for i=1:7
  while ((X1(1,Px0)<2)&&(Px0<x1))
      Px0=Px0+1;
  end
  Px1=Px0;
  while (((X1(1,Px1)>=3)&&(Px1<x1))||((Px1-Px0)<10))
      Px1=Px1+1;
  end
  Z=d(:,Px0:Px1,:);
  switch strcat('Z',num2str(i))
      case 'Z1'
          PIN0=Z;
      case 'Z2'
          PIN1=Z;
      case 'Z3'
          PIN2=Z;
      case 'Z4'
          PIN3=Z;
      case 'Z5'
          PIN4=Z;
      case 'Z6'
          PIN5=Z;
      otherwise 
          PIN6=Z;
  end
  figure(11);
  subplot(1,7,i);
  imshow(Z);
  Px0=Px1;
end
%载入训练好的神经网络%
load('-mat','ym');
%将分割后的后的车牌字符归一化处理%
PIN0=pretreatment(PIN0);
PIN1=pretreatment(PIN1);
PIN2=pretreatment(PIN2);
PIN3=pretreatment(PIN3);
PIN4=pretreatment(PIN4);
PIN5=pretreatment(PIN5);
PIN6=pretreatment(PIN6);
P0=[PIN0',PIN1',PIN2',PIN3',PIN4',PIN5',PIN6'];
%开始识别字符
for i=2:7
  T0= sim(net ,P0(:,i));
  T1 = compet (T0) ;
  d =find(T1 == 1)-1
 if (d==0)
    str='A';
 elseif (d==1)
     str='B';
 elseif (d==2)
     str='C';
 elseif (d==3)
     str='D';
 elseif (d==4)
     str='E';
 elseif (d==5)
     str='F';
 elseif (d==6)
     str='G';
 elseif (d==7)
     str='H';
 elseif (d==8)
     str='I';
 elseif (d==9)
     str='J';
 elseif (d==10)
     str='K';
 elseif (d==11)
     str='L';
 elseif (d==12)
     str='M';
 elseif (d==13)
     str='N';
 elseif (d==14)
     str='O';
 elseif (d==15)
     str='P';
 elseif (d==16)
     str='Q';
 elseif (d==17)
     str='R';
 elseif (d==18)
     str='S';
 elseif (d==19)
     str='T';
 elseif (d==20)
     str='U';
 elseif (d==21)
     str='V';
 elseif (d==22)
     str='W';
 elseif (d==23)
     str='X';
 elseif (d==24)
     str='Y';
 elseif (d==25)
     str='Z';
 else
    str=num2str(d);
 end
 switch i
     case 2
         str1=str;
     case 3
         str2=str;
     case 4
         str3=str;
     case 5
         str4=str;
     case 6
         str5=str;
     otherwise
         str6=str;
  end
end 
s=strcat('京',str1,'.',str2,str3,str4,str5,str6); 
figure(12);
imshow('dw.jpg'),title(s);

pretreatment.m
function  inpt  = pretreatment(I)
%图像归一化处理以及特征提取
%if isrgb(I)
%   I1 = rgb2gray(I);
%else
%   I1=I;
%end
%I1 = rgb2gray(I);
%将图片统一划为50*20大小
I=imresize(I,[50 20]);
I=im2bw(I,0.9);
[m,n]=size(I);
inpt=zeros(1,m*n);
%将图像按列转换成一个行向量
for j=1:n
    for i=1:m
        inpt(1,m*(j-1)+i)=I(i,j);
    end
end

bp.m
function bp()
%use for bp training
I0=pretreatment(imread('A.jpg'));
I1=pretreatment(imread('B.jpg'));
I2=pretreatment(imread('C.jpg'));
I3=pretreatment(imread('D.jpg'));
I4=pretreatment(imread('E.jpg'));
I5=pretreatment(imread('F.jpg'));
I6=pretreatment(imread('G.jpg'));
I7=pretreatment(imread('H.jpg'));
I8=pretreatment(imread('I.jpg'));
I9=pretreatment(imread('J.jpg'));
I10=pretreatment(imread('K.jpg'));
I11=pretreatment(imread('L.jpg'));
I12=pretreatment(imread('M.jpg'));
I13=pretreatment(imread('N.jpg'));
I14=pretreatment(imread('O.jpg'));
I15=pretreatment(imread('P.jpg'));
I16=pretreatment(imread('Q.jpg'));
I17=pretreatment(imread('R.jpg'));
I18=pretreatment(imread('S.jpg'));
I19=pretreatment(imread('T.jpg'));
I20=pretreatment(imread('U.jpg'));
I21=pretreatment(imread('V.jpg'));
I22=pretreatment(imread('W.jpg'));
I23=pretreatment(imread('X.jpg'));
I24=pretreatment(imread('Y.jpg'));
I25=pretreatment(imread('Z.jpg'));
P=[I0',I1',I2',I3',I4',I5',I6',I7',I8',I9',I10',I11',I12',I13',I14',I15',I16',I17',I18',I19',I20',I21',I22',I23',I24',I25'];
%输出样本%
T=eye(26,26);
%%bp神经网络参数设置
net=newff(minmax(P),[800,23,26],{'logsig','logsig','logsig'},'trainrp');
net.inputWeights{1,1}.initFcn ='randnr';
net.layerWeights{2,1}.initFcn ='randnr';
net.trainparam.epochs=5000;
net.trainparam.show=50;
net.trainparam.goal=0.0000000001;
net=init(net);
%%%训练样本%%%%
[net,tr]=train(net,P,T);
save('ym','net');