车牌识别系统MATLAB源代码完整

clc; clear all; close all;

[filename, pathname, filterindex] = uigetfile({'*.jpg;*.tif;*.png;*.gif','All Image Files';...

'*.*','All Files' }, '选择待处理图像', ...

'images\01.jpg');

file = fullfile(pathname, filename);%文件路径和文件名创建合成完整文件名

id = Get_Id(file);%得到file中的所有对象

Img = imread(file);%根据路径和文件名读取图片到Img

[Plate, bw, Loc] = Pre_Process(Img); % 车牌区域预处理

result = Plate_Process(Plate, id); % 车牌区域二值化处理

% 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bw = Segmation(result);

words = Main_Process(bw); % 主流程处理

Write_Mask(words, id); % 写出到模板库

str = Pattern_Recognition(words); % 识别

function id = Get_Id(file)

% 获取图像id信息

% 输入参数：

% file——图像路径

% 输出参数：

% id——图像id信息

info = imfinfo(file);

FS = [422227 354169 293184 235413 214202 ...

130938 490061 120297 98686 137193 ...

80558 46208 69947 58110 62115 ...

59072 52168 60457 53979 50223];

id = find(FS == info.FileSize);

if isempty(id)

warndlg('未建立该图像模板库，可能运行出错！', '警告');

id = 1;

end

function R = Cubic_Spline(P)

% 三次样条插值

% 输入参数：

% P——节点矩阵

% 输出参数：

% R——样条节点矩阵

% 计算相邻插值点之间的弦长

chordlen = sqrt(sum(diff(P, [], 1).^2,2));

% 将弦长参数归一化到[0,1]上

chordlen = chordlen/sum(chordlen);

% 计算每个插值节点处的累加弦长，作为给点处的参数

cumarc = [0; cumsum(chordlen)];

x = cumarc;

N = size(P, 1);

R = [];

% 以下部分为一元三次样条插值的程序，对于空间三维数据，以同样的累加

% 弦长作为参数，对x，y，z分量做三次一元样条插值得到的结果便是对三维数据% 做三次样条插值

for k = 1 : size(P, 2)

y = P(:, k);

m = zeros(1, N);

M = zeros(1, N);

n = m;

d = m;

A = eye(N);

A = 2*A;

m(1) = 1;

n(N) = 1;

m(N) = 1;

n(1) = 1;

for i = 2:N-1

m(i) = (x(i+1)-x(i))/(x(i+1)-x(i-1));

n(i) = 1 - m(i);

d(i) = 6*((y(i+1) - y(i))/(x(i+1) - x(i)) - (y(i) - y(i-1))/(x(i) - x(i-1)))/(x(i+1)-x(i-1));

end

for j = 1 : N-1

A(j, j+1) = m(j);

A(j+1, j) = n(j+1);

end

p = A(2:N-1, 2:N-1);

q = d(2:N-1);

Q = inv(p)*q';

M = zeros(1,N);

M(1,1) = 0;

M(1,N) = 0;

M(1,2:N-1)= Q;

S = [];

temp = [];

for i = 1:N-1

% 对每一个分量计算出来的插值曲线进行采样，以便将其画出。

s = 50;% 采样点个数

z = linspace(x(i),x(i+1),s);

h = x(i+1)-x(i);

for j = 1:length(z)

S(j) = M(i)*((x(i+1)-z(j))^3)/(6*h)+M(i+1)*((z(j)-x(i))^3)/(6*h)+(y(i)-M(i)*(h^2)/6)*((x(i+1)-z(j))/h)+( y(i+1)-M(i+1)*h^2/6)*((z(j)-x(i))/h);

end

temp = [temp S];

end

R(:, k) = temp;

end

function mask = Get_PointSplineMask(Img, Ptn)

% 获取封闭有序节点的蒙板图像

% Img——图像矩阵

% Ptn——封闭有序节点

% 输出参数：

% mask——蒙板图像

if ndims(Img) == 3

I = rgb2gray(Img);

else

I = Img;

end

mask = zeros(size(I));

Ptn = Cubic_Spline(Ptn); % 样条插值

for i = 1 : size(Ptn, 1)-1

pt1 = Ptn(i, :); % 线段起点

pt2 = Ptn(i+1, :); % 线段终点

x1 = pt1(1); y1 = pt1(2);

x2 = pt2(1); y2 = pt2(2);

% 直线段参数

A = (y1 - y2)/(x1*y2 - x2*y1);

B = (-x1 + x2)/(x1*y2 - x2*y1);

% 直线段取点

xk = linspace(min(x1, x2), max(x1, x2));

if B == 0

yk = linspace(min(y1, y2), max(y1, y2));

else

yk = (-1 - A*xk)/B;

end

% 赋值操作

for j = 1 : length(xk)

if ~isnan(round(yk(j))) && ~isnan(round(xk(j))) && ...

~isinf(round(yk(j))) && ~isinf(round(xk(j))) && ...

round(yk(j)) > 0 && round(xk(j)) > 0

mask(round(yk(j)), round(xk(j))) = 1;

end

end

end

mask = logical(mask); % 类型转换

mask = bwmorph(mask, 'bridge'); % 桥接操作

mask = imfill(mask, 'hole'); % 补洞操作

function Im = Image_Rotate(Img, num, flag)

% 旋转校正函数

% 输入函数：

% Img——图像矩阵

% num——图像序号

% flag——显示图像窗口

% 输出函数：

% Im——结果图像

if nargin < 3

flag = 0;

end

role = [6 0 0 -13 5 10 0 10 0 -5 2 -1 2 2 0 -5 -2 0 6 2];

Im = imrotate(Img, role(num), 'bilinear');

if flag

figure(2);

subplot(1, 2, 1); imshow(Img); title('原图像');

subplot(1, 2, 2); imshow(Im); title('旋转图像');

end

function words = Main_Process(bw, flag_display)

% 主流程处理，分割字符并获取

% 输入参数：

% bw——车牌区域图像

% flag_display——显示图像标记

% 输出参数：

% words——车牌字符数据

if nargin < 2

flag_display = 1;

end

[m, n] = size(bw);

k1 = 1;

k2 = 1;

s = sum(bw); % 列积分投影

j = 1; % 列游标

while j ~= n

% 寻找车牌图像左侧边界

while s(j) == 0 && j <= n-1

j = j + 1;

end

k1 = j-1; % 车牌图像左侧边界

% 寻找车牌图像右侧边界

while s(j) ~= 0 && j <= n-1

j = j + 1;

end

k2 = j-1; % 车牌图像右侧边界

Tol = round(n/6.5); % 字符区域宽度约束

if k2-k1 > Tol

[val, num] = min(sum(bw(:, [k1+5:k2-5])));

bw(:, k1+num+5)=0; % 抹去该字符end

end

% 再切割

bw = Segmation(bw);

% 切割出7 个字符

[m, n] = size(bw);

wideTol = round(n/20); % 区域宽度最小约束rateTol = 0.25; % 中心区域比值约束

flag = 0;

word1 = [];

while flag == 0

[m, n] = size(bw);

left = 1;

wide = 0;

% 找到空隙位置

while sum(bw(:,wide+1)) ~= 0

wide = wide+1;

end

if wide < wideTol

% 认为是左侧干扰

bw(:, 1:wide) = 0; % 抹去干扰区域

bw = Segmation(bw);

else

% 提取字符区域

temp = Segmation(imcrop(bw, [1 1 wide m]));

[m, n] = size(temp);

tall = sum(temp(:)); % 该字符所有像素之和

% 该字符图像2/3图像区域像素和

two_thirds = sum(sum( temp(round(m/3):2*round(m/3), :) ));

rate = two_thirds/tall; % 中间区域像素/全局区域像素

if rate > rateTol

flag = 1;

word1 = temp; %提取WORD 1

end

bw(:, 1:wide) = 0; % 抹去已处理的区域

bw = Segmation(bw);

end

end

% 分割出第二个字符

[word2, bw] = Word_Segmation(bw);

% 分割出第三个字符

[word3, bw] = Word_Segmation(bw);

% 分割出第四个字符

[word4, bw] = Word_Segmation(bw);

% 分割出第五个字符

[word5, bw] = Word_Segmation(bw);

% 分割出第六个字符

[word6, bw] = Word_Segmation(bw);

% 分割出第七个字符

[word7, bw] = Word_Segmation(bw);

wid = [size(word1, 2) size(word2, 2) size(word3, 2) ...

size(word4, 2) size(word5, 2) size(word6, 2) size(word7, 2)];

[maxwid, indmax] = max(wid);

maxwid = maxwid + 10;

wordi = word1;

wordi = [zeros(size(wordi, 1), round((maxwid-size(word1, 2))/2)) wordi zeros(size(wordi, 1), round((maxwid-size(word1, 2))/2))];

word1 = wordi;

wordi = word2;

wordi = [zeros(size(wordi, 1), round((maxwid-size(word2, 2))/2)) wordi zeros(size(wordi, 1), round((maxwid-size(word2, 2))/2))];

word2 = wordi;

wordi = word3;

wordi = [zeros(size(wordi, 1), round((maxwid-size(word3, 2))/2)) wordi zeros(size(wordi, 1), round((maxwid-size(word3, 2))/2))];

word3 = wordi;

wordi = word4;

wordi = [zeros(size(wordi, 1), round((maxwid-size(word4, 2))/2)) wordi zeros(size(wordi, 1), round((maxwid-size(word4, 2))/2))];

word4 = wordi;

wordi = word5;

wordi = [zeros(size(wordi, 1), round((maxwid-size(word5, 2))/2)) wordi zeros(size(wordi, 1), round((maxwid-size(word5, 2))/2))];

word5 = wordi;

wordi = word6;

wordi = [zeros(size(wordi, 1), round((maxwid-size(word6, 2))/2)) wordi zeros(size(wordi, 1), round((maxwid-size(word6, 2))/2))];

word6 = wordi;

wordi = word7;

wordi = [zeros(size(wordi, 1), round((maxwid-size(word7, 2))/2)) wordi zeros(size(wordi, 1), round((maxwid-size(word7, 2))/2))];

word7 = wordi;

% figure(1);

% subplot(2, 7, 1); imshow(word1); title('字符1');

% subplot(2, 7, 2); imshow(word2); title('字符2');

% subplot(2, 7, 3); imshow(word3); title('字符3');

% subplot(2, 7, 4); imshow(word4); title('字符4');

% subplot(2, 7, 5); imshow(word5); title('字符5');

% subplot(2, 7, 6); imshow(word6); title('字符6');

% subplot(2, 7, 7); imshow(word7); title('字符7');

% 切割出的字符归一化大小为40*20,此处演示

word11 = imresize(word1, [40 20]);

word21 = imresize(word2, [40 20]);

word31 = imresize(word3, [40 20]);

word41 = imresize(word4, [40 20]);

word51 = imresize(word5, [40 20]);

word61 = imresize(word6, [40 20]);

word71 = imresize(word7, [40 20]);

% subplot(2, 7, 8); imshow(word11); title('字符1');

% subplot(2, 7, 9); imshow(word21); title('字符2');

% subplot(2, 7, 10); imshow(word31); title('字符3');

% subplot(2, 7, 11); imshow(word41); title('字符4');

% subplot(2, 7, 12); imshow(word51); title('字符5');

% subplot(2, 7, 13); imshow(word61); title('字符6');

% subplot(2, 7, 14); imshow(word71); title('字符7');

% 赋值操作

words.word1 = word11;

words.word2 = word21;

words.word3 = word31;

words.word4 = word41;

words.word5 = word51;

words.word6 = word61;

words.word7 = word71;

if flag_display

figure;

subplot(2, 7, 1); imshow(word1); title('字符1', 'FontWeight', 'Bold');

subplot(2, 7, 2); imshow(word2); title('字符2', 'FontWeight', 'Bold');

subplot(2, 7, 3); imshow(word3); title('字符3', 'FontWeight', 'Bold');

subplot(2, 7, 4); imshow(word4); title('字符4', 'FontWeight', 'Bold');

subplot(2, 7, 5); imshow(word5); title('字符5', 'FontWeight', 'Bold');

subplot(2, 7, 6); imshow(word6); title('字符6', 'FontWeight', 'Bold');

subplot(2, 7, 7); imshow(word7); title('字符7', 'FontWeight', 'Bold');

subplot(2, 7, 8); imshow(word11); title('字符1', 'FontWeight', 'Bold');

subplot(2, 7, 9); imshow(word21); title('字符2', 'FontWeight', 'Bold');

subplot(2, 7, 10); imshow(word31); title('字符3', 'FontWeight', 'Bold');

subplot(2, 7, 11); imshow(word41); title('字符4', 'FontWeight', 'Bold');

subplot(2, 7, 12); imshow(word51); title('字符5', 'FontWeight', 'Bold');

subplot(2, 7, 13); imshow(word61); title('字符6', 'FontWeight', 'Bold');

subplot(2, 7, 14); imshow(word71); title('字符7', 'FontWeight', 'Bold'); end

function mask = Mask_Process(Img, id)

% 图像蒙版处理函数

% 输入参数：

% Img——图像矩阵

% id——图像序号

% 输出参数：

% mask——模板图像

% 如果已经存在模板图像则不再取模板

filename = sprintf('mask\\mask%d.jpg', id);

if exist(filename, 'file')

mask = imread(filename);

if ~isa(mask, 'logical')

mask = im2bw(mask);

end

sz = size(Img);

if ~isequal(sz(1:2), size(mask))

mask = imresize(mask, sz(1:2));

mask = logical(mask);

end

return;

end

I = Img;

Ptn = [];

figure;

subplot(1, 3, 1); imshow(I);

title('取点_{左键取点，右键退出}', 'Color', 'r', ...

'FontWeight', 'Bold');

hold on;

set(gcf, 'units', 'normalized', 'position', [0 0 1 1]);

[x, y, button] = ginput(1); % 点击

while button == 1

plot(x, y, 'r+', 'LineWidth', 2); % 绘制节点

Ptn = [Ptn; x y]; % 存储节点

[x, y, button] = ginput(1); % 点击

end

if size(Ptn, 1) < 2

return;

end

Ptn = [Ptn; Ptn(1, :)];

plot(Ptn(:, 1), Ptn(:, 2), 'ro-', 'LineWidth', 2, 'MarkerFaceColor', 'k'); title('原图像', 'Color', 'k', ...

'FontWeight', 'Bold');

mask = Get_PointSplineMask(I, Ptn); % 获取蒙板图像

subplot(1, 3, 2); imshow(mask);

title('蒙板图像', 'Color', 'k', ...

'FontWeight', 'Bold');

if ndims(I) == 3

I1 = I .* uint8(cat(3, mask, mask, mask));

else

if ~isequal(size(I), size(mask))

mask = imresize(mask, size(I));

mask = logical(mask);

end

I1 = I .* mask;

end

subplot(1, 3, 3); imshow(I1);

title('蒙板分割结果', 'Color', 'k', ...

'FontWeight', 'Bold');

imwrite(mask, filename);

function str = Pattern_Recognition(words)

% 车牌字符识别

% 输入参数：

% words——车牌字符图像

% 输出参数：

% str——识别结果

pattern = [];

dirpath = './标准库/';

%读取所有样本文件

files = ls([dirpath '*.jpg']);

for t = 1 : length(files)

imagedata = imread([dirpath, files(t,:)]); %读取样本文件imagedata = im2bw(imagedata, 0.5);

pattern(t).feature = imagedata;

end

% 字符库列表

chepaiword=[...

'沪' 'C' 'E' 'N' '2' '6' '7' '浙' 'B' 'T' '2' '0' '0' '8' ...

'皖' 'N' '6' '9' '3' '3' '5' '皖' 'A' 'M' '3' '7' '6' '7' ...

'沪' 'B' '9' '1' '2' '1' '3' '沪' 'F' 'S' '2' '6' '9' '5' ...

'吉' 'A' 'E' '3' '1' '2' '7' '沪' 'B' 'B' '3' '3' '9' '3' ...

'沪' 'J' 'B' '7' '7' '7' '8' '京' 'H' 'Z' '9' '1' '3' '9' ...

'晋' 'K' 'C' '3' '8' '4' '1' '苏' 'D' 'V' '0' '0' '0' '1' ...

'苏' 'A' 'V' '8' '8' '8' '8' '浙' 'B' '8' '8' '8' '8' '8' ...

'沪' 'F' 'Q' '1' '2' '7' '7' '苏' 'E' 'D' '0' '0' '0' '9' ...

'沪' 'E' 'C' '1' '2' '8' '8' '苏' 'G' 'B' '0' '5' '7' '8' ...

'沪' 'B' 'B' '2' '0' '2' '9' '沪' 'D' '5' '4' '2' '2' '6'];

for m = 1 : 7;

for n = 1 : length(files);

switch m

case 1

distance(n)=sum(sum(abs(words.word1-pattern(n).feature)));

case 2

distance(n)=sum(sum(abs(words.word2-pattern(n).feature)));

case 3

distance(n)=sum(sum(abs(words.word3-pattern(n).feature)));

case 4

distance(n)=sum(sum(abs(words.word4-pattern(n).feature)));

case 5

distance(n)=sum(sum(abs(words.word5-pattern(n).feature)));

case 6

distance(n)=sum(sum(abs(words.word6-pattern(n).feature)));

case 7

distance(n)=sum(sum(abs(words.word7-pattern(n).feature)));

end

end

[yvalue,xnumber]=min(distance);

filename = files(xnumber, :);

[pathstr, name, ext] = fileparts(filename);

result(m) = chepaiword(str2num(name));

end

str = ['识别结果为：' result];

msgbox(str, '车牌识别', 'modal');

str = result;

function result = Plate_Process(plate, id, flag)

% 车牌区域处理

% 输入参数：

% plate——车牌区域

% id——车牌序号

% flag——显示图像标记

% 输出参数：

% result——结果图像

if nargin < 3

flag = 1;

end

if ndims(plate) == 3

plate1 = rgb2gray(plate);

% figure(1)

% imshow(plate1)

else

plate1 = plate;

end

Im = Image_Rotate(plate1, id);

th = [0.85 0.50 0.85 0.70 0.30 ...

0.40 0.65 0.45 0.45 0.65 ...

0.51 0.80 0.40 0.75 0.60 ...

0.55 0.40 0.40 0.76 0.72];

bw = im2bw(Im, th(id)); % 车牌区域二值化

h = fspecial('average', 2); % 均值滤波模板

bw1 = imfilter(bw, h, 'replicate'); % 均值滤波

% figure(1)

% imshow(bw1)

mask = Mask_Process(bw1, id); % 模板处理

bw2 = bw1 .* mask; % 模板滤波

result = bw2;

if flag

figure;

subplot(2, 2, 1); imshow(plate); title('车牌区域图像', 'FontWeight', 'Bold');

subplot(2, 2, 2); imshow(Im); title('车牌区域校正图像', 'FontWeight', 'Bold');

subplot(2, 2, 3); imshow(bw1); title('车牌区域二值图像', 'FontWeight', 'Bold');

subplot(2, 2, 4); imshow(bw2); title('滤波二值图像', 'FontWeight', 'Bold'); end

function [Plate, bw, Loc] = Pre_Process(Img, parm, flag)

% 车牌图像预处理，提取车牌区域

% 输入参数：

% Img——图像矩阵

% parm——参数向量

% flag——是否显示处理结果

% 输出参数：

% Plate——分割结果

if nargin < 3

flag = 1;

end

if nargin < 2 || isempty(parm)

if size(Img, 2) > 900

parm = [0.35 0.9 90 0.35 0.7 90 2];

end

if size(Img, 2) > 700 && size(Img, 2) < 900

parm = [0.6 0.9 90 0.6 0.8 90 0.5];

end

if size(Img, 2) > 500 && size(Img, 2) < 700

parm = [0.5 0.54 50 0.6 0.7 50 3];

end

if size(Img, 2) < 500

parm = [0.8 0.9 150 0.8 0.9 150 3];

end

end

I = Img;

[y, x, z] = size(I); % y方向对应行、x方向对应列、z方向对应深度

% 图像尺寸过大会影响处理效率，所以进行放缩处理

if y > 800

rate = 800/y;

I = imresize(I, rate);

end

[y, x, z] = size(I); % y方向对应行、x方向对应列、z方向对应深度

myI = double(I); % 数据类型转换

bw1 = zeros(y, x);

bw2 = zeros(y, x);

Blue_y = zeros(y, 1);

% 对每一个像素进行分析，统计满足条件的像素所在的行对应的个数for i = 1 : y

for j = 1 : x

rij = myI(i, j, 1)/(myI(i, j, 3)+eps);

gij = myI(i, j, 2)/(myI(i, j, 3)+eps);

bij = myI(i, j, 3);

% 蓝色RGB的灰度范围

if (rij < parm(1) && gij < parm(2) && bij > parm(3)) ...

|| (gij < parm(1) && rij < parm(2) && bij > parm(3)) Blue_y(i, 1) = Blue_y(i, 1) + 1; % 蓝色象素点统计

bw1(i, j) = 1;

end

end

end

% Y方向车牌区域确定

[temp, MaxY] = max(Blue_y);

Th = parm(7);

% 向上追溯，直到车牌区域上边界

PY1 = MaxY;

while ((Blue_y(PY1,1)>Th) && (PY1>1))

PY1 = PY1 - 1;

end

% 向下追溯，直到车牌区域下边界

PY2 = MaxY;

while ((Blue_y(PY2,1)>Th) && (PY2

PY2 = PY2 + 1;

% 对车牌区域的修正

PY1 = PY1 - 2;

PY2 = PY2 + 2;

if PY1 < 1

PY1 = 1;

end

if PY2 > y

PY2 = y;

end

% 得到车牌区域

IY = I(PY1:PY2, :, :);

%%%%%%%% X 方向%%%%%%%%%%

% 进一步确定X方向的车牌区域

Blue_x = zeros(1,x);

for j = 1:x

for i = PY1:PY2

rij = myI(i, j, 1)/(myI(i, j, 3)+eps);

gij = myI(i, j, 2)/(myI(i, j, 3)+eps);

bij = myI(i, j, 3);

% 蓝色RGB的灰度范围

if (rij < parm(4) && gij < parm(5) && bij > parm(6)) ...

|| (gij < parm(4) && rij < parm(5) && bij > parm(6)) Blue_x(1,j) = Blue_x(1,j) + 1; % 蓝色象素点统计

bw2(i, j) = 1;

end

end

end

% 想右追溯，直到找到车牌区域左边界

PX1 = 1;

while (Blue_x(1,PX1)

PX1 = PX1 + 1;

end

% 想左追溯，直到找到车牌区域右边界

PX2 = x;

while (Blue_x(1,PX2)PX1)

PX2 = PX2 - 1;

end

% 对车牌区域的修正

PX1 = PX1 - 2;

PX2 = PX2 + 2;

if PX1 < 1

PX1 = 1;

if PX2 > x

PX2 = x;

end

% 得到车牌区域

IX = I(:, PX1:PX2, :);

% 分割车牌区域

Plate = I(PY1:PY2, PX1:PX2, :);

Loc.row = [PY1 PY2];

Loc.col = [PX1 PX2];

bw = bw1 + bw2;

bw = logical(bw);

bw(1:PY1, :) = 0;

bw(PY2:end, :) = 0;

bw(:, 1:PX1) = 0;

bw(:, PX2:end) = 0;

if flag

figure;

subplot(2, 2, 3); imshow(IY); title('行过滤结果', 'FontWeight', 'Bold');

subplot(2, 2, 1); imshow(IX); title('列过滤结果', 'FontWeight', 'Bold');

subplot(2, 2, 2); imshow(I); title('原图像', 'FontWeight', 'Bold');

subplot(2, 2, 4); imshow(Plate); title('分割结果', 'FontWeight', 'Bold'); end

function e = Segmation(d)

% 车牌图像切割

% 输入参数：

% d——车牌图像

% 输出参数：

% e——切割后的车牌图像

[m, n] = size(d);

% 初始化参数

top = 1; % 矩形框上参数

bottom = m; % 矩形框下参数

left = 1; % 矩形框左参数

right = n; % 矩形框右参数

% 寻找矩形框上参数

while sum(d(top,:))==0 && top<=m

top = top + 1;

% 寻找矩形框下参数

while sum(d(bottom,:))==0 && bottom>=1

bottom = bottom - 1;

end

% 寻找矩形框左参数

while sum(d(:,left))==0 && left<=n

left = left + 1;

end

% 寻找矩形框右参数

while sum(d(:,right))==0 && right>=1

right = right - 1;

end

dd = right - left; % 矩形框长

hh = bottom - top; % 矩形框宽

e = imcrop(d, [left top dd hh]); % 车牌图像矩形框裁剪

function [word, result] = Word_Segmation(d)

% 提取字符

% 输入参数：

% d——车牌图像

% 输出参数：

% word——字符图像

% result——处理结果

word = [];

flag = 0;

[m, n] = size(d);

wideTol = round(n/20); % 区域宽度最小约束

rateTol = 0.25; % 中心区域比值约束

while flag == 0

[m, n] = size(d);

wide = 0;

while sum(d(:,wide+1)) ~= 0 && wide <= n-2

wide = wide + 1;

end

temp = Segmation(imcrop(d, [1 1 wide m]));

[m1,n1] = size(temp);

if widerateTol

d(:, 1:wide) = 0;

if sum(sum(d)) ~= 0

d = Segmation(d); % 切割出最小范围

else

word = [];

flag = 1;

end

else

word = Segmation(imcrop(d, [1 1 wide m]));

d(:, 1:wide) = 0;

if sum(sum(d)) ~= 0;

d = Segmation(d);

flag = 1;

else

d = [];

end

end

end

result = d;

function Write_Mask(words, id)

% 写出到模板库

% 输入参数：

% words——字符图像集合

% id——序号

imwrite(words.word1, sprintf('模板库\\%d_字符%d.jpg', id, 1)); imwrite(words.word2, sprintf('模板库\\%d_字符%d.jpg', id, 2)); imwrite(words.word3, sprintf('模板库\\%d_字符%d.jpg', id, 3)); imwrite(words.word4, sprintf('模板库\\%d_字符%d.jpg', id, 4)); imwrite(words.word5, sprintf('模板库\\%d_字符%d.jpg', id, 5)); imwrite(words.word6, sprintf('模板库\\%d_字符%d.jpg', id, 6)); imwrite(words.word7, sprintf('模板库\\%d_字符%d.jpg', id, 7));

% imwrite(words.word1, sprintf('模板库\\%d.jpg', (id-1)*7+1)); % imwrite(words.word2, sprintf('模板库\\%d.jpg', (id-1)*7+2)); % imwrite(words.word3, sprintf('模板库\\%d.jpg', (id-1)*7+3)); % imwrite(words.word4, sprintf('模板库\\%d.jpg', (id-1)*7+4)); % imwrite(words.word5, sprintf('模板库\\%d.jpg', (id-1)*7+5)); % imwrite(words.word6, sprintf('模板库\\%d.jpg', (id-1)*7+6)); % imwrite(words.word7, sprintf('模板库\\%d.jpg', (id-1)*7+7));