%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% 字符分割模块算法
% 定位剪切后的彩色车牌图像－－灰度－－二值化－－统一到黑底白字－－去除上下边框
% －－切割出最小范围－－滤波－－形态学处理－－分割出7个字符
% 去除上下边框算法：
% 1.黑白跳变小于阈值则被视为背景；2.连续白线大于某阈值则该白线被认为是背景
% 3.单行白色大于阈值则被认为是背景，考虑FLAG的值；
% 4.做完以上处理后，上边1/2 中搜索连续两条黑线，认为该黑线以上为背景；在下边1/2 中搜索连续两条黑线，认为该黑线以下为背景
% 归一化为 40*20 ，商用系统程序中归一化为 32*16 ，此处仅演示作用
function [d]=lpcseg(jpg)
I=imread('car1.jpg');
I1=rgb2gray(I);
I2=edge(I1,'robert',0.15,'both');
se=[1;1;1];
I3=imerode(I2,se);
se=strel('rectangle',[25,25]);
I4=imclose(I3,se);
I5=bwareaopen(I4,2000);
[y,x,z]=size(I5);
myI=double(I5);
tic
white_y=zeros(y,1);
for i=1:y
for j=1:x
if(myI(i,j,1)==1)
white_y(i,1)= white_y(i,1)+1;
end
end
end
[temp MaxY]=max(white_y);
PY1=MaxY;
while ((white_y(PY1,1)>=5)&&(PY1>1))
PY1=PY1-1;
end
PY2=MaxY;
while ((white_y(PY2,1)>=5)&&(PY2<y))
PY2=PY2+1;
end
IY=I(PY1:PY2,:,:);
white_x=zeros(1,x);
for j=1:x
for i=PY1:PY2
if(myI(i,j,1)==1)
white_x(1,j)= white_x(1,j)+1;
end
end
end
PX1=1;
while ((white_x(1,PX1)<3)&&(PX1<x))
PX1=PX1+1;
end
PX2=x;
while ((white_x(1,PX2)<3)&&(PX2>PX1))
PX2=PX2-1;
end
PX1=PX1-1;
PX2=PX2+1;
dw=I(PY1:PY2-8,PX1:PX2,:);
t=toc;
figure(1),subplot(3,2,1),imshow(dw),title('定位剪切后的彩色车牌图像')
imwrite(dw,'dw.jpg');
[filename,filepath]=uigetfile('dw.jpg','输入一个定位裁剪后的车牌图像');
jpg=strcat(filepath,filename);
a=imread(jpg);
%figure(1);subplot(3,2,1),imshow(a),title('1.定位剪切后的彩色车牌图像')
b=rgb2gray(a);
imwrite(b,'2.车牌灰度图像.jpg');
figure(1);subplot(3,2,2),imshow(b),title('2.车牌灰度图像')
g_max=double(max(max(b)));
g_min=double(min(min(b)));
T=round(g_max-(g_max-g_min)/3); % T 为二值化的阈值
[m,n]=size(b);
d=(double(b)>=T); % d:二值图像
imwrite(d,'3.车牌二值图像.jpg');
figure(1);subplot(3,2,3),imshow(d),title('3.车牌二值图像')
% 旋转
rotate=0;
d=imread('3.车牌二值图像.jpg');
bw=edge(d);
[m,n]=size(d);
theta=1:179;
% bw 表示需要变换的图像，theta 表示变换的角度
% 返回值 r 表示的列中包含了对应于 theta中每一个角度的 Radon 变换结果
% 向量 xp 包含相应的沿 x轴的坐标
[r,xp]=radon(bw,theta);
i=find(r>0);
[foo,ind]=sort(-r(i));
k=i(ind(1:size(i)));
[y,x]=ind2sub(size(r),k);
[mm,nn]=size(x);
if mm~=0 && nn~=0
j=1;
while mm~=1 && j<180 && nn~=0
i=find(r>j);
[foo,ind]=sort(-r(i));
k=i(ind(1:size(i)));
[y,x]=ind2sub(size(r),k);
[mm,nn]=size(x);
j=j+1;
end
if nn~=0
if x % Enpty matrix: 0-by-1 when x is an enpty array.
x=x;
else % 可能 x 为空值
x=90; % 其实就是不旋转
end
d=imrotate(d,abs(90-x)); % 旋转图像
rotate=1;
end
end
imwrite(d,'4.Radon 变换旋转后的二值图像.jpg');
figure(1),subplot(3,2,4),imshow(d),title('4.Radon 变换旋转后的二值图像')
% 统一到白底黑字
[m,n]=size(d);
% flag=0 表示原来就是白底黑字，否则表示原来是黑底白字
flag=0;
c=d([round(m/3):m-round(m/3)],[round(n/3):n-round(n/3)]);
if sum(sum(c))/m/n*9>0.5
d=~d;flag=1;
end
% 对反色后的图像预处理，整列几乎为白的认为是背景
if flag==1
for j=1:n
if sum(sum(d(:,j)))/m>=0.95
d(:,j)=0;
end
end
% 对以上处理后的图像再处理
% 在左边 1/2 处找连续两条黑线，认为该黑线左边为背景；在右边 1/2 处找连续两条黑线，认为该黑线右边是背景
% 左边 1/2
jj=0;
for j=1:round(n/2)
if sum(sum(d(:,[j:j+0])))==0
jj=j;
end
end
d(:,[1:jj])=0;
% 右边 1/2
for j=n:-1:round(n/2)
if sum(sum(d(:,[j-0:j])))==0
jj=j;
end
end
d(:,[jj:n])=0;
end
imwrite(d,'5.统一成黑底白字.jpg');
figure(1),subplot(3,2,5),imshow(d),title('5.背景色统一成黑底白字')
figure(2),subplot(5,1,1),imshow(d),title('5.黑底白字的二值车牌图像')
% 去除上下边框
% STEP 1 黑白跳变小于阈值则被视为背景
% 上面 2/5
y1=10; % y1: 跳变阈值
for i=1:round(m/5*2)
count=0;jump=0;temp=0;
for j=1:n
if d(i,j)==1
temp=1;
else
temp=0;
end
if temp==jump
count=count;
else
count=count+1;
end
jump=temp;
end
if count<y1
d(i,:)=0;
end
end
% 下面 2/5
for i=3*round(m/5):m
count=0;jump=0;temp=0;
for j=1:n
if d(i,j)==1
temp=1;
else
temp=0;
end
if temp==jump
count=count;
else
count=count+1;
end
jump=temp;
end
if count<y1
d(i,:)=0;
end
end
imwrite(d,'6.黑白跳变小于某阈值的行则被视为背景.jpg');
figure(2),subplot(5,1,2),imshow(d),title('6.黑白跳变小于某阈值的行则被视为背景')
% STEP 2 单行白色大于阈值则被认为是背景，考虑 FLAG 的值
% 上面 2/5
y2=round(n/2); % y2: 阈值
for i=1:round(m/5*2)
if flag==0
temp=sum(d(i,:));y2=round(n/2);
if temp>y2
d(i,:)=0;
end
else
temp=m-sum(d(i,:));y2=m-round(n/2);
if temp<y2
d(i,:)=0;
end
end
end
% 下面 2/5
for i=round(3*m/5):m
if flag==0
temp=sum(d(i,:));y2=round(n/2);
if temp>y2
d(i,:)=0;
end
else
temp=m-sum(d(i,:));y2=m-round(n/2);
if temp<y2
d(i,:)=0;
end
end
end
imwrite(d,'7.单行白色点总数大于某阈值则该行被认为是背景.jpg');
figure(2),subplot(5,1,3),imshow(d),title('7.单行白色点总数大于某阈值则该行被认为是背景')
% STEP 3 单行白色大于阈值则被认为是背景，考虑 FLAG 的值
% 上面 2/5
y2=round(n/2); % y2: 阈值
for i=1:round(m/5*2)
if flag==0
temp=sum(d(i,:));y2=round(n/2);
if temp>y2
d(i,:)=0;
end
else
temp=m-sum(d(i,:));y2=m-round(n/2);
if temp<y2
d(i,:)=0;
end
end
end
% 下面 2/5
for i=round(3*m/5):m
if flag==0
temp=sum(d(i,:));y2=round(n/2);
if temp>y2
d(i,:)=0;
end
else
temp=m-sum(d(i,:));y2=m-round(n/2);
if temp<y2
d(i,:)=0;
end
end
end
imwrite(d,'8.单行白色点总数大于某阈值则该行被认为是背景.jpg');
figure(2),subplot(5,1,4),imshow(d),title('8.单行白色点总数大于某阈值则该行被认为是背景')
% STEP 4 做完以上处理后，上边 1/2 中搜索连续两条黑线，认为该黑线以上为背景；
% 在下边 1/2 中搜索连续两条黑线，认为该黑线以下为背景
% 上边 1/2
for i=1:round(m/2)
if sum(sum(d([i,i+0],:)))==0
ii=i;
end
end
d([1:ii],:)=0;
% 下边 1/2
for i=m:-1:round(m/2)
if sum(sum(d([i-0:i],:)))==0
ii=i;
end
end
d([ii:m],:)=0;
imwrite(d,'9.搜索上下两条黑线后的结果.jpg');
figure(2),subplot(5,1,5),imshow(d),title('9.搜索上下两条黑线后的结果')
% 反旋转
if rotate==1
d=imrotate(d,-abs(x-90));
end
imwrite(d,'10.反旋转去毛刺后.jpg');
figure(3),subplot(3,2,1),imshow(d),title('10.反旋转去毛刺后')
% 切割处最小范围
d=qiege(d);e=d;
imwrite(d,'11.切割处最小范围.jpg');
figure(3),subplot(3,2,2),imshow(d),title('11.切割处最小范围')
figure(3),subplot(3,2,3),imshow(d),title('11.均值滤波前')
% 滤波
h=fspecial('average',3);
d=im2bw(round(filter2(h,d)));
imwrite(d,'12.均值滤波后.jpg');
figure(3),subplot(3,2,4),imshow(d),title('12.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进行膨胀
% 'line'/'diamond'/'ball'...
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
imwrite(d,'13.膨胀或腐蚀处理后.jpg');
figure(3),subplot(3,2,5),imshow(d),title('13.膨胀或腐蚀处理后')
% 寻找连续有文字的块，若长度大于某阈值，则认为该块有两个字符组成，需要分割
d=qiege(d);
[m,n]=size(d);
figure,subplot(2,1,1),imshow(d),title(n)
k1=1;k2=1;s=sum(d);j=1;
while j~=n
while s(j)==0
j=j+1;
end
k1=j;
while s(j)~=0 && j<=n-1
j=j+1;
end
k2=j-1;
if k2-k1>=round(n/6.5)
[val,num]=min(sum(d(:,[k1+5:k2-5])));
d(:,k1+num+5)=0; % 分割
end
end
% 再切割
d=qiege(d);
% 切割出 7 个字符
y1=10;y2=0.25;flag=0;word1=[];
while flag==0
[m,n]=size(d);
left=1;wide=0;
while sum(d(:,wide+1))~=0
wide=wide+1;
end
if wide<y1 % 认为是左侧干扰
d(:,[1:wide])=0;
d=qiege(d);
else
temp=qiege(imcrop(d,[1 1 wide m]));
[m,n]=size(temp);
all=sum(sum(temp));
two_thirds=sum(sum(temp([round(m/3):2*round(m/3)],:)));
if two_thirds/all>y2
flag=1;word1=temp; % WORD 1
end
d(:,[1:wide])=0;d=qiege(d);
end
end
% 分割出第二个字符
[word2,d]=getword(d);
% 分割出第三个字符
[word3,d]=getword(d);
% 分割出第四个字符
[word4,d]=getword(d);
% 分割出第五个字符
[word5,d]=getword(d);
% 分割出第六个字符
[word6,d]=getword(d);
% 分割出第七个字符
[word7,d]=getword(d);
subplot(5,7,1),imshow(word1),title('1');
subplot(5,7,2),imshow(word2),title('2');
subplot(5,7,3),imshow(word3),title('3');
subplot(5,7,4),imshow(word4),title('4');
subplot(5,7,5),imshow(word5),title('5');
subplot(5,7,6),imshow(word6),title('6');
subplot(5,7,7),imshow(word7),title('7');
[m,n]=size(word1);
% 商用系统程序中归一化大小为 32*16,此处演示
word1=imresize(word1,[40 20]);
word2=wordprocess(word2);
word3=wordprocess(word3);
word4=wordprocess(word4);
word5=wordprocess(word5);
word6=wordprocess(word6);
word7=wordprocess(word7);
subplot(5,7,15),imshow(word1),title('1');
subplot(5,7,16),imshow(word2),title('2');
subplot(5,7,17),imshow(word3),title('3');
subplot(5,7,18),imshow(word4),title('4');
subplot(5,7,19),imshow(word5),title('5');
subplot(5,7,20),imshow(word6),title('6');
subplot(5,7,21),imshow(word7),title('7');
imwrite(word1,'14.字符分割归一化后 1.jpg');
imwrite(word2,'14.字符分割归一化后 2.jpg');
imwrite(word3,'14.字符分割归一化后 3.jpg');
imwrite(word4,'14.字符分割归一化后 4.jpg');
imwrite(word5,'14.字符分割归一化后 5.jpg');
imwrite(word6,'14.字符分割归一化后 6.jpg');
imwrite(word7,'14.字符分割归一化后 7.jpg');
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clc
word='';
word(1)=wordrec(word1);
word(2)=wordrec(word2);
word(3)=wordrec(word3);
word(4)=wordrec(word4);
word(5)=wordrec(word5);
word(6)=wordrec(word6);
word(7)=wordrec(word7);
clc
save I 'word1' 'word2' 'word3' 'word4' 'word5' 'word6' 'word7'
clear
load I;
load bp net;
word='';
word(1)=wordrec(word1);
word(2)=wordrec(word2);
word(3)=wordrec(word3);
word(4)=wordrec(word4);
word(5)=wordrec(word5);
word(6)=wordrec(word6);
word(7)=wordrec(word7);
word=strcat('识别结果:',word);
subplot(5,3,14),imshow([]),title(word,'fontsize',24)
% 该子程序用于切割出最小范围
function e=qiege(d)
[m,n]=size(d);
top=1;bottom=m;left=1;right=n; % init
while sum(d(top,:))==0 && top<=m
top=top+1;
end
while sum(d(bottom,:))==0 && bottom>=1
bottom=bottom-1;
end
while sum(d(:,left))==0 && left<=n
left=left+1;
end
while sum(d(:,right))==0 && right>=1
right=right-1;
end
dd=right-left;
hh=bottom-top;
e=imcrop(d,[left top dd hh]);
% 分割字符
function [word,result]=getword(d)
word=[];flag=0;y1=8;y2=0.5;
% if d==[]
% word=[];
% else
while flag==0
[m,n]=size(d);
wide=0;
while sum(d(:,wide+1))~=0 && wide<=n-2
wide=wide+1;
end
temp=qiege(imcrop(d,[1 1 wide m]));
[m1,n1]=size(temp);
if wide<y1 && n1/m1>y2
d(:,[1:wide])=0;
if sum(sum(d))~=0
d=qiege(d); % 切割出最小范围
else word=[];flag=1;
end
else
word=qiege(imcrop(d,[1 1 wide m]));
d(:,[1:wide])=0;
if sum(sum(d))~=0;
d=qiege(d);flag=1;
else d=[];
end
end
end
%end
result=d;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% 该子程序用于字符归一化处理
function d=wordprocess(d)
[m,n]=size(d);
%top 1/3, bottom 1/3
for i=1:round(m/3)
if sum(sum(d([i:i+0],:)))==0
ii=i;d([1:ii],:)=0;
end
end
for i=m:-1:2*round(m/3)
if sum(sum(d([i-0:i],:)))==0
ii=i;d([ii:m],:)=0;
end
end
if n~=1
d=qiege(d);
end
% d=..这个可以通过训练过程设置大小
% d=imresize(d,[32 16]); % 商用系统程序中归一划大小为：32*16
d=imresize(d,[40 20]);
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% 编号：A-Z 分别为 1-26; 0-9 分别为 27-36;
% 京 津 沪 渝 港 澳 吉 辽 鲁 豫 冀 鄂 湘 晋 青 皖 苏
% 赣 浙 闽 粤 琼 台 陕 甘 云 川 贵 黑 藏 蒙 桂 新 宁
% 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59
% 60 61 62 63 64 65 66 67 68 69 70
% 使用 BP 网络
function word=wordrec(xx)
% clear
% clc
load bp net;
xx=im2bw(xx);xx=double(xx(:)); % 使用阈值将图像转换为二进制图像
a=sim(net,xx); % 归一划为： 32*16,则 xx=512*1;
[val,num]=max(a);
if num<=26
word=char(double('A')+num-1);
elseif num<=36
word=char(double('0')+num-1-26);
else
switch num
case 37
word='京';
case 38
word='津';
case 39
word='沪';
case 40
word='渝';
case 41
word='港';
case 42
word='澳';
case 43
word='吉';
case 44
word='辽';
case 45
word='鲁';
case 46
word='豫';
case 47
word='冀';
case 48
word='鄂';
case 49
word='湘';
case 50
word='晋';
case 51
word='青';
case 52
word='皖';
case 53
word='苏';
case 54
word='赣';
case 55
word='浙';
case 56
word='闽';
case 57
word='粤';
case 58
word='琼';
case 59
word='台';
case 60
word='陕';
case 61
word='甘';
case 62
word='云';
case 63
word='川';
case 64
word='贵';
case 65
word='黑';
case 66
word='藏';
case 67
word='蒙';
case 68
word='桂';
case 69
word='新';
case 70
word='宁';
end
end