matlab车牌自动定位,超级实用的车牌定位matlab程序

%程序功能：实现车牌自动定位

%作者：重庆大学 田建国 QQ:363966533

%编写时间:2007.04.10

%修改时间:2007.07.06;修改内容:增加合并区域操作，第88行开始

clear;clc;close all

filename='241.jpg';

I=im2gray(filename);%调用自编函数读取图像，并转化为灰度图象；

tic   %计时开始

[height,width]=size(I);

%预处理

I_edge=zeros(height,width);% 创建height*width的矩阵

for i=1:width-1            % 对每一列开始遍历

I_edge(:,i)=abs(I(:,i+1)-I(:,i));% 每列的值赋为原图像中左右两列相减的绝对值(即梯度)

end

% 归一化处理(0～255)

I_edge=(255/(max(max(I_edge))-min(min(I_edge))))*(I_edge-min(min(I_edge)));

figure

imshow(I_edge);title('归一化处理')

[I_edge,y1]=select(I_edge,height,width);   %%%%%%调用select函数

figure

imshow(I_edge);title('选择')

BW2 = I_edge;%

%%%%%%%%%%%%%%%%一些形态学处理

SE=strel('rectangle',[10,10]);

IM2=imerode(BW2,SE);%腐蚀

figure

imshow(IM2);title('腐蚀');

IM2=bwareaopen(IM2,20);%开运算，消除细小物体

figure

imshow(IM2);title('开运算');

IM3=imdilate(IM2,SE);%膨胀

figure

imshow(IM2);title('膨胀');

%%%%%%%%%%%%%%%%%%%%%%%%%%

%%%%%%%%%%%%%%%%%%投影以粗略估计车牌位置

p_h=projection(double(IM3),'h');                %调用projection函数

if(p_h(1)>0)

p_h=[0,p_h];

end

p_v=projection(double(IM3),'v');                %调用projection函数

if(p_v(1)>0)

p_v=[0,p_v];

end

%%%%%%

p_h=double((p_h>5));

p_h=find(((p_h(1:end-1)-p_h(2:end))~=0));

len_h=length(p_h)/2;

%%%%%

p_v=double((p_v>5));

p_v=find(((p_v(1:end-1)-p_v(2:end))~=0));

len_v=length(p_v)/2;

%%%%%%%%%%%

%%%%%%%%%%%%%%%%%粗略计算车牌候选区

k=1;

for i=1:len_h

for j=1:len_v

s=IM3(p_h(2*i-1):p_h(2*i),p_v(2*j-1):p_v(2*j));

if(mean(mean(s))>0.1)

p{k}=[p_h(2*i-1),p_h(2*i)+1,p_v(2*j-1),p_v(2*j)+1];

k=k+1;

end

end

end

k=k-1;

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

%%%%%%%%%%%%%%进一步缩小车牌候选区

for i=1:k

edge_IM3=double(edge(double(IM3(p{i}(1):p{i}(2),p{i}(3):p{i}(4))),'canny'));

[x,y]=find(edge_IM3==1);

p{i}=[p{i}(1)+min(x),p{i}(2)-(p{i}(2)-p{i}(1)+1-max(x)),...

p{i}(3)+min(y),p{i}(4)-(p{i}(4)-p{i}(3)+1-max(y))];

p_center{i}=[fix((p{i}(1)+p{i}(2))/2),fix((p{i}(3)+p{i}(4))/2)];

p_ratio(i)=(p{i}(4)-p{i}(3))/(p{i}(2)-p{i}(1));

end

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

%对上面参数和变量的说明：p为一胞元，用于存放每个图像块的左上和右下两个点的坐标；

%存放格式为：p{k}=[x1,x2,y1,y2]；x1,x2分别为行坐标，y1,y2为列坐标

%p_center为一胞元,用于存放每个图像块的中心坐标,p_center{k}=[x,y];x,y分别为行,列坐标

%p_ratio为一矩阵，用来存放图像块的长宽比例

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

%%%%%%%%%%合并临近区域%%%%%%%

%如果有多个区域则执行合并

if k>1

n=0;

ncount=zeros(1,k);

for i=1:k-1

%%%需要调整if条件中的比例

%%%需要调整

%检查是否满足合并条件

if(abs(p{i}(1)+p{i}(2)-p{i+1}(1)-p{i+1}(2))<=height/30&&abs(p{i+1}(3)-p{i}(4))<=width/15)

p{i+1}(1)=min(p{i}(1),p{i+1}(1));

p{i+1}(2)=max(p{i}(2),p{i+1}(2));

p{i+1}(3)=min(p{i}(3),p{i+1}(3));

p{i+1}(4)=max(p{i}(4),p{i+1}(4));  %向后合并

n=n+1;

ncount(n)=i+1;

end

end

%如果有合并，求出合并后最终区域

if(n>0)

d_ncount=ncount(2:n+1)-ncount(1:n);%避免重复记录临近的多个区域。

index=find(d_ncount~=1);

m=length(index);

for i=1:m

pp{i}=p{ncount(index(i))};

%pp_center{i}=p_center{ncount(i)};

%重新记录合并区域的比例

pp_ratio(i)=(pp{i}(4)-pp{i}(3))/(pp{i}(2)-pp{i}(1));

end

p=pp;%更新区域记录

p_ratio=pp_ratio; %更新区域比例记录

clear pp;clear pp_ratio; %清除部分变量

end

end

k=length(p); %更新区域个数

%%%%%%%%%%%%%%合并结束%%%%%%%%%%%%%%%%%%%%%

%%%%%%%%根据区域比例判断是否为车牌区域%%%%%%%%%%%%

m=1;T=0.6*max(p_ratio);%0.8参数需要调整

for i=1:k

if(p_ratio(i)>=T&p_ratio(i)<20)

p1{m}=p{i};

m=m+1;

end

end

p=p1;clear p1;

k=m-1;   %更新区域数

%%%%%%%%%%%判定结束%%%%%%%%%%%%%%%%%%%%%

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

%d=zeros(k+1,k+1);

%for i=1:k+1

%   for j=i+1:k+1

%d(i,j)=sqrt((p_center{i}(1)-p_center{j}(1))^2+(p_center{i}(2)-p_center{j}(2))^2);

%  end

%end

%说明:d用于存放第i,j个图像块中心点的距离;

%T=sqrt(height^2+width^2)/10;%阈值

%[x,y]=find(d>0&d

%for i=1:length(x)

%    p{x(i)}(1)=min(p{x(i)}(1),p{y(i)}(1));

%    p{x(i)}(2)=max(p{x(i)}(2),p{y(i)}(2));

%    p{x(i)}(3)=min(p{x(i)}(3),p{y(i)}(3));

%    p{x(i)}(4)=max(p{x(i)}(4),p{y(i)}(4));

%end

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

toc                                  %计时结束

clear edge_IM3;clear x; clear y;     % 清空部分变量

%%%%%%%%%%%%%%%%显示

figure;

subplot(221);imshow(I);

subplot(222);imshow(BW2);

subplot(223);imshow(IM2);

subplot(224);imshow(IM3);

%%%%%%%%%%%%%%%%%显示

figure;

for i=1:k

subplot(1,k,i);

index=p{i};

imshow(I(index(1)-2:index(2),index(3):index(4)));

end

if(k==1)

imwrite(I(index(1)-2:index(2),index(3):index(4)),'cp.jpg');

end

%存储车牌图像

%%%%%%%%%%%%%%%%

%figure;

%I1=I.*uint8(IM3);imshow(I1)

[本帖最后由 vincent92 于 2009-3-18 01:08 编辑]