数字图象处理项目报告 final

'《数字图象处理》课程项目实施报告题目：车牌自动定位与字符分割组号：29任课教师：严壮志组长：13122925鲁珅20%成员：13121104王国平20%成员：13123032尹悦20%成员：13120824胡子泉20%成员：13124177杨亨敏20%联系方式：15921611856二零一六年二月二十六日 《数字图像处理》课程项目报告2016年2月27日一、引言图像分割是图像处理的主要目标，本项目综合利用图像采集、图像增强、彩色图像处理、图像分割等知识，根据牌照的色彩和形状、牌照字符的排列规则及形状等特点，设计实现车牌区域的自动分割及字符的自动分割。二、课程项目目的通过该项目，我们旨在了解图像分割的基本原理，能够对图像处理系统及应用有一个整体认识，对图像识别中涉及的关键技术如直方图分析、阈值分割、特征匹配等有更进一步的理解，提高数字图像处理课程的实际应用能力。三、基本原理随着中国车辆的逐渐增多车辆，车牌识别也应用越来越广泛。虽然车牌识别的方法多种多样，但真正能实用的却是不多。我们小组总结下来对于车牌识别最重要的环节——车牌定位，不外乎两种途径。其一中国的车牌无外乎黄、蓝、白等几种颜色的车牌，因此我们可以根据彩色图像的车牌和字体的色彩，对车牌进行定位。其二我们可以通过对车牌的预处理，将图像转换为灰度图像，通过基本的图像处理方法，如膨胀腐蚀来对车牌进行定位。两种或给方法各有优缺点，方法一明显比第二种方法定位的准确率高，但是识别颜色范围有限，而且需要进行较多的数据运算。第二种方法虽然准确率低，但是运算数据量小，比较快速。车牌能准确定位后，接下来就比较好做了。仅统计二值图像的数据即可对车牌进行分割，从而提取出每个字体的特征，与已知的码本进行匹配比较就可以达到识别车牌的目的。本课题先根据色彩对车牌进行了定位，而后进行了一些相应的处理，达到了较好的效果。 四、具体操作步棸如下：4.1彩色车牌定位以蓝色车牌为例，将彩色车牌读入后，采用水平垂直双向投影法，将图像分别投影到X、Y坐标轴，然后分别沿X、Y轴扫描图像。当沿Y轴扫描时，一边扫描一边统计图像中蓝色像素点的个数，第一次扫描到蓝色像素点最多的行时停止扫描，并记录下蓝色像素点最多的行，然后以此行为基点，分别向上、向下扫描，直到统计的像素点小于像素点阈值时，停止扫描，记录上下行的Y轴坐标PY1、PY2，即PY1、PY2之间区域就是Y轴方向的车牌边缘位置。同理进行X轴方向的像素扫描，确定X轴方向的车牌区域。在得到车牌的大概位置I之后，再对I区域扫描，确定该区域内白色像素点个数，然后求出该区域内白色像素点数所占比例，以判断所得的I区域是否为正确的车牌位置。代码如下（以下为Y方向）：[y,x,z]=size(I);%I为彩色图像矩阵是三维矩阵myI=double(I);Blue_y=zeros(y,1);%建立矩阵用于储存蓝色像素点数目fori=1:yforj=1:xif((myI(i,j,1)<=48)&&((myI(i,j,2)<=100)&&(myI(i,j,2)>=40))&&((myI(i,j,3)<=200)&&(myI(i,j,3)>=80)))%蓝色像素点判断依据Blue_y(i,1)=Blue_y(i,1)+1;endendend[tempMaxY]=max(Blue_y);PY1=MaxY;%确定上边界PY2=MaxY;while((Blue_y(PY2,1)>=5)&&(PY2=40)&&(myI(i,j,2)<=90)&&(myI(i,j,3)<=220)&&(myI(i,j,3)>=80)if((myI(i,j,1)<=48)&&((myI(i,j,2)<=100)&&(myI(i,j,2)>=40))&&((myI(i,j,3)<=200)&&(myI(i,j,3)>=80)))%if((myI(i,j,1)<=0.7)&&(myI(i,j,1)>=0.6)&&((myI(i,j,2)<=1.8)&&(myI(i,j,2)>=0.6))&&((myI(i,j,3)<=1.7)&&(myI(i,j,3)>=0.45)))Blue_y(i,1)=Blue_y(i,1)+1;%蓝色象素点统计endend endbaisebili=0;changkuanbi=0;k=0;l=0;while(~((baisebili>=0.06)&&(baisebili<=0.5)&&(changkuanbi>=0.20)&&(changkuanbi<=0.6)))%gaishanqingxieif(k==0)%第一次进来[tempMaxY]=max(Blue_y);%Y方向车牌区域确定temp(最多点数):所有行中，最多的累积像素点MaxY（最多点所在行）:该行中蓝点最多iftemp<=20l=1;break;endPY1=MaxY;%有最多蓝点的行付给PY1while((Blue_y(PY1,1)>=Y_threshlow)&&(PY1>1))%找到图片上边界PY1=PY1-1;end%PY1：存储车牌上边界值PY2=MaxY;while((Blue_y(PY2,1)>=Y_threshlow)&&(PY2=30))&&((myI(i,j,3)<=160)&&(myI(i,j,3)>=80)))if((myI(i,j,1)<=65)&&((myI(i,j,2)<=100)&&(myI(i,j,2)>=40))&&((myI(i,j,3)<=160)&&(myI(i,j,3)>=90)))%这里由82修改成90.因为图片20090504809%if((myI(i,j,1)<=0.7)&&(myI(i,j,1)>=0.6)&&((myI(i,j,2)<=1.8)&&(myI(i,j,2)>=0.6))&&((myI(i,j,3)<=1.7)&&(myI(i,j,3)>=0.45)))Blue_x(1,j)=Blue_x(1,j)+1;endendend [tempMaxX]=max(Blue_x);PX1=MaxX-6*(PY2-PY1);ifPX1<=1PX1=1;endwhile((Blue_x(1,PX1)<=X_threshhigh)&&(PX1=xPX2=x;endwhile((Blue_x(1,PX2)<=X_threshhigh)&&(PX2>PX1))%阈值PX2=PX2-1;end%确定出X方向车牌终点PX1,PX2%=========================================================a=PY2-PY1+1;b=PX2-PX1+1;White=0;fori=PY1:PY2forj=PX1:PX2if(std([myI(i,j,1)myI(i,j,2)myI(i,j,3)],1,2)<=22)&&(myI(i,j,1)>=90)&&(myI(i,j,1)<=255)White=White+1;%白色象素点统计endendendbaisebili=White/(a*b)changkuanbi=a/bk=k+1%===========================蓝色区域不是车牌区域=================elseif(k~=0)Blue_y(PY1:PY2,1)=0;[tempMaxY]=max(Blue_y);iftemp<=20l=1;break;endPY1=MaxY; while((Blue_y(PY1,1)>=Y_threshlow)&&(PY1>1))%找到图片上边界%阈值为5PY1=PY1-1;end%PY1：存储车牌上边界值PY2=MaxY;while((Blue_y(PY2,1)>=Y_threshlow)&&(PY2=20))&&((myI(i,j,3)<=160)&&(myI(i,j,3)>=80)))Blue_x(1,j)=Blue_x(1,j)+1;endendend%这里修改成了从中间向两边扩展,这个方法不好。因车牌中间某些位置可能出现断层。%采用增强型两边往中间收缩。[tempMaxX]=max(Blue_x);PX1=MaxX-6*(PY2-PY1);ifPX1<=1PX1=1;endwhile((Blue_x(1,PX1)<=X_threshhigh)&&(PX1=xPX2=x;endwhile((Blue_x(1,PX2)<=X_threshhigh)&&(PX2>PX1))%阈值PX2=PX2-1;end%确定出X方向车牌终点%PX1=1;%while((Blue_x(1,PX1)<=X_threshhigh)&&(PX1PX1))%阈值%PX2=PX2-1;%endPX1,PX2%=========================================================a=PY2-PY1+1;b=PX2-PX1+1;White=0;fori=PY1:PY2forj=PX1:PX2if(std([myI(i,j,1)myI(i,j,2)myI(i,j,3)],1,2)<=16)&&(myI(i,j,1)>=90)&&(myI(i,j,1)<=255)White=White+1;%白色象素点统计endendendbaisebili=White/(a*b)changkuanbi=a/bk=k+1endendif(l==1)myI=double(I);Y_threshlow=5;%这个数值很重要。决定了提取的彩图的质量X_firrectify=5;%ganraotransaction%==============黄色车牌识别========Y方向============================Blue_y=zeros(y,1);fori=1:yforj=1:x%if(myI(i,j,1)<=48)&&(myI(i,j,2)>=40)&&(myI(i,j,2)<=90)&&(myI(i,j,3)<=220)&&(myI(i,j,3)>=80)if((myI(i,j,1)>=140)&&(myI(i,j,1)<=250)&&((myI(i,j,2)<=200)&&(myI(i,j,2)>=100))&&((myI(i,j,3)<=50)&&(myI(i,j,3)>=0)))%if((myI(i,j,1)<=0.7)&&(myI(i,j,1)>=0.6)&&((myI(i,j,2)<=1.8)&&(myI(i,j,2)>=0.6))&&((myI(i,j,3)<=1.7)&&(myI(i,j,3)>=0.45)))Blue_y(i,1)=Blue_y(i,1)+1;%蓝色象素点统计endend endk=0changkuanbi=0;while(~((changkuanbi>=0.20)&&(changkuanbi<=0.6)))%if(k==0)[tempMaxY]=max(Blue_y);%Y方向车牌区域确定temp(最多点数):所有行中，最多的累积像素点MaxY（最多点所在行）:该行中蓝点最多iftemp<=20%2048*1536照相msgbox("车牌定位出错","warning");endPY1=MaxY;%有最多蓝点的行付给PY1while((Blue_y(PY1,1)>=Y_threshlow)&&(PY1>1))%找到图片上边界PY1=PY1-1;end%PY1：存储车牌上边界值PY2=MaxY;while((Blue_y(PY2,1)>=Y_threshlow)&&(PY2=30))&&((myI(i,j,3)<=160)&&(myI(i,j,3)>=80)))if((myI(i,j,1)>=140)&&(myI(i,j,1)<=250)&&((myI(i,j,2)<=200)&&(myI(i,j,2)>=100))&&((myI(i,j,3)<=50)&&(myI(i,j,3)>=0)))%这里由82修改成90.因为图片20090504809%if((myI(i,j,1)<=0.7)&&(myI(i,j,1)>=0.6)&&((myI(i,j,2)<=1.8)&&(myI(i,j,2)>=0.6))&&((myI(i,j,3)<=1.7)&&(myI(i,j,3)>=0.45)))Blue_x(1,j)=Blue_x(1,j)+1;endendend[tempMaxX]=max(Blue_x);PX1=MaxX-6*(PY2-PY1);ifPX1<=1 PX1=1;endwhile((Blue_x(1,PX1)<=X_threshhigh)&&(PX1=xPX2=x;endwhile((Blue_x(1,PX2)<=X_threshhigh)&&(PX2>PX1))%阈值PX2=PX2-1;end%确定出X方向车牌终点%PX1=1;%while((Blue_x(1,PX1)<=X_threshhigh)&&(PX1PX1))%阈值%PX2=PX2-1;%end%确定出X方向车牌终点PX1,PX2%=========================================================a=PY2-PY1+1;b=PX2-PX1+1;changkuanbi=a/b;k=k+1;elseif(k~=0)Blue_y(PY1:PY2,1)=0;[tempMaxY]=max(Blue_y);%Y方向车牌区域确定temp(最多点数):所有行中，最多的累积像素点MaxY（最多点所在行）:该行中蓝点最多iftemp<=20%2048*1536照相msgbox("车牌定位出错","warning");endPY1=MaxY;%有最多蓝点的行付给PY1while((Blue_y(PY1,1)>=Y_threshlow)&&(PY1>1))%找到图片上边界PY1=PY1-1;end%PY1：存储车牌上边界值PY2=MaxY;while((Blue_y(PY2,1)>=Y_threshlow)&&(PY2=30))&&((myI(i,j,3)<=160)&&(myI(i,j,3)>=80)))if((myI(i,j,1)>=140)&&(myI(i,j,1)<=250)&&((myI(i,j,2)<=200)&&(myI(i,j,2)>=100))&&((myI(i,j,3)<=50)&&(myI(i,j,3)>=0)))%这里由82修改成90.因为图片20090504809%if((myI(i,j,1)<=0.7)&&(myI(i,j,1)>=0.6)&&((myI(i,j,2)<=1.8)&&(myI(i,j,2)>=0.6))&&((myI(i,j,3)<=1.7)&&(myI(i,j,3)>=0.45)))Blue_x(1,j)=Blue_x(1,j)+1;endendend[tempMaxX]=max(Blue_x);PX1=MaxX-6*(PY2-PY1);ifPX1<=1PX1=1;endwhile((Blue_x(1,PX1)<=X_threshhigh)&&(PX1=xPX2=x;endwhile((Blue_x(1,PX2)<=X_threshhigh)&&(PX2>PX1))%阈值PX2=PX2-1;end%确定出X方向车牌终点PX1,PX2%=========================================================a=PY2-PY1+1;b=PX2-PX1+1;changkuanbi=a/b;end endend%初次修正Y_firrectify=fix((PY2-PY1)/5);%适当扩大这个值可以正确旋转PY1=PY1-Y_firrectify;%对车牌区域的修正,向上PY2=PY2+Y_firrectify;%对车牌区域的修正,向下PX1=PX1-X_firrectify;%对车牌区域的修正PX2=PX2+X_firrectify;%对车牌区域的修正，%二次修正子函数function[PY2,PY1,PX2,PX1,threshold]=xiuzheng(PY2,PY1,PX2,PX1)S=(PY2-PY1)*(PX2-PX1)ifS<=25000threshold=50;Y_secrectify=3;X_secrectify=3;elseifS>25000&&S<=45000threshold=100;Y_secrectify=-3;X_secrectify=3;elseifS>45000&&S<=80000threshold=200;Y_secrectify=-3;X_secrectify=3;elseifS>80000&&S<=150000threshold=300;Y_secrectify=-10;X_secrectify=-10;elseifS>150000&&S<=400000threshold=600;Y_secrectify=-10;X_secrectify=-10;elsethreshold=1800;Y_secrectify=-10;X_secrectify=-10;end%对车牌区域的进一步修正PY1=PY1-Y_secrectify;PY2=PY2+Y_secrectify; %---Executesonbuttonpressinhuiduhua.functionhuiduhua_Callback(hObject,eventdata,handles)%hObjecthandletohuiduhua(seeGCBO)%eventdatareserved-tobedefinedinafutureversionofMATLAB%handlesstructurewithhandlesanduserdata(seeGUIDATA)%---Executesonbuttonpressinhuiduhua.globalbw;bw=rgb2gray(bw);axes(handles.axes1);imshow(bw);set(handles.edit1,"String","车牌灰度处理后图像");%---Executesonbuttonpressinqingxiejiaozheng.functionqingxiejiaozheng_Callback(hObject,eventdata,handles)%hObjecthandletoqingxiejiaozheng(seeGCBO)%eventdatareserved-tobedefinedinafutureversionofMATLAB%handlesstructurewithhandlesanduserdata(seeGUIDATA)globalbw;globalqingxiejiao;qingxiejiao=rando_bianhuan(bw);bw=imrotate(bw,qingxiejiao,"bilinear","crop");axes(handles.axes1);imshow(bw);set(handles.edit1,"String","倾斜校正后的车牌图像");%倾斜变换子函数用到Rando变换functionqingxiejiao=rando_bianhuan(I)I=edge(I);theta=1:180;[R,xp]=radon(I,theta);[I,J]=find(R>=max(max(R)));%J为倾斜角qingxiejiao=90-J;%-----车牌图像处理部分，与识别无关，纯粹图像处理%---Executesonbuttonpressinhuidubianhuan.functionhuidubianhuan_Callback(hObject,eventdata,handles)%hObjecthandletohuidubianhuan(seeGCBO)%eventdatareserved-tobedefinedinafutureversionofMATLAB%handlesstructurewithhandlesanduserdata(seeGUIDATA)globalbw; b1=bw;b1=imadjust(b1,[0.3,0.7],[]);%J=imadjust(I,[low_in;high_in],[low_out;high_out])%将图像I中的亮度值映射到J中的新值，即将low_in至high_in之间的值映射到low_out至high_out之间的值。%low_in以下与high_in以上的值被剪切掉了，也就是说，low_in以下的值映射到low_out，high_in%以上的值映射到high_out。它们都可以使用空的矩阵[]，默认值是[01]。axes(handles.axes1);imshow(b1);set(handles.edit1,"String","灰度变换后的车牌图像");%---Executesonbuttonpressinaverage.functionaverage_Callback(hObject,eventdata,handles)%hObjecthandletoaverage(seeGCBO)%eventdatareserved-tobedefinedinafutureversionofMATLAB%handlesstructurewithhandlesanduserdata(seeGUIDATA)globalbw;b3=bw;b3=filter2(fspecial("average",2),b3)/255;axes(handles.axes1);imshow(b3);set(handles.edit1,"String","均值滤波后的车牌图像");%---Executesonbuttonpressinzhongzhi.functionzhongzhi_Callback(hObject,eventdata,handles)%hObjecthandletozhongzhi(seeGCBO)%eventdatareserved-tobedefinedinafutureversionofMATLAB%handlesstructurewithhandlesanduserdata(seeGUIDATA)globalbw;b2=bw;b2=medfilt2(b2);axes(handles.axes1);imshow(b2);set(handles.edit1,"String","中值滤波后的车牌图像");%-----以上为车牌图像处理部分，与识别无关，纯粹图像处理%---Executesonbuttonpressincachu.functioncachu_Callback(hObject,eventdata,handles)%hObjecthandletocachu(seeGCBO)%eventdatareserved-tobedefinedinafutureversionofMATLAB%handlesstructurewithhandlesanduserdata(seeGUIDATA) globalbw;globalthreshold;bw=im2bw(bw,graythresh(bw));%figure,imshow(bw);bw=bwmorph(bw,"hbreak",inf);%figure,imshow(bw);bw=bwmorph(bw,"spur",inf);%figure,imshow(bw);title("擦除之前");bw=bwmorph(bw,"open",5);%figure,imshow(bw);title("开运算");bw=bwareaopen(bw,50);axes(handles.axes1);imshow(bw);set(handles.edit1,"String","擦除黑点后的车牌图像");%---Executesonbuttonpressinfanse.functionfanse_Callback(hObject,eventdata,handles)%hObjecthandletofanse(seeGCBO)%eventdatareserved-tobedefinedinafutureversionofMATLAB%handlesstructurewithhandlesanduserdata(seeGUIDATA)globalbw;bw=~bw;axes(handles.axes1);imshow(bw);set(handles.edit1,"String","反色后的车牌图像");%---Executesonbuttonpressinjiebian.functionjiebian_Callback(hObject,eventdata,handles)%hObjecthandletojiebian(seeGCBO)%eventdatareserved-tobedefinedinafutureversionofMATLAB%handlesstructurewithhandlesanduserdata(seeGUIDATA)globalbw;bw=touying(bw);axes(handles.axes1);imshow(bw);set(handles.edit1,"String","上下边界截取后的图片");%截取时用到投影计算黑点的子函数functionbw_fir=touying(image_bw)X_yuzhi=1;[y,x]=size(image_bw);assignin("base","y",y);assignin("base","x",x);Y_touying=(sum((~image_bw)"))";%往左边投影统计黑点assignin("base","Y_touying",Y_touying);X_touying=sum((~image_bw));%往下面投影 assignin("base","X_touying",X_touying);%找黑体边缘Y_up=fix(y/2);Y_yuzhi=mean(Y_touying((fix(y/2)-10):(fix(y/2)+10),1))/2;%由1.6调整为2assignin("base","Y_yuzhi",Y_yuzhi);while((Y_touying(Y_up,1)>=Y_yuzhi)&&(Y_up>1))%找到图片上边界Y_up=Y_up-1;endY_down=fix(y/2);while((Y_touying(Y_down,1)>=Y_yuzhi)&&(Y_down20000&&SS<=30000shedingyuzhi=4;elseifSS>30000&&SS<=50000shedingyuzhi=4;elseifSS>50000&&SS<=80000shedingyuzhi=4;elseshedingyuzhi=4;endganrao=SS/100;%=========================================================%定义数组histogram存储垂直方向的黑点数histogram=sum(~imfenge);%=================文字分割=============================k=1;forh=1:x-1if((histogram(1,h)<=shedingyuzhi)&&(histogram(1,h+1)>shedingyuzhi))||((h==1)&&histogram(1,h)>shedingyuzhi)fenge(1,k)=h;k=k+1;elseif((histogram(1,h)>shedingyuzhi)&&(histogram(1,h+1)<=shedingyuzhi))||((h==x-1)&&histogram(1,h)>shedingyuzhi)fenge(1,k)=h+1;k=k+1;endend %k记录边界值的个数%fenge记录边界的位置k=k-1;%去掉多产生的一个K值ifk<10msgbox("提取出错","警告");end%==============================================if(sum(histogram(1,fenge(1,1):fenge(1,2)))=max(baifenbi));shibiehanzi=chepai;%在数字中，从0开始所以要减一。这里不用ifshibiehanzi==1shibiehanzi="京";elseifshibiehanzi==2shibiehanzi="津";elseifshibiehanzi==3shibiehanzi="沪";elseifshibiehanzi==4shibiehanzi="渝";elseifshibiehanzi==5shibiehanzi="冀";elseifshibiehanzi==6shibiehanzi="晋";elseifshibiehanzi==7shibiehanzi="辽";elseifshibiehanzi==8shibiehanzi="吉";elseifshibiehanzi==9shibiehanzi="黑";elseifshibiehanzi==10shibiehanzi="苏";elseifshibiehanzi==11shibiehanzi="浙";elseifshibiehanzi==12shibiehanzi="皖";elseifshibiehanzi==13 shibiehanzi="闽";elseifshibiehanzi==14shibiehanzi="赣";elseifshibiehanzi==15shibiehanzi="鲁";elseifshibiehanzi==16shibiehanzi="豫";elseifshibiehanzi==17shibiehanzi="鄂";elseifshibiehanzi==18shibiehanzi="湘";elseifshibiehanzi==19shibiehanzi="粤";elseifshibiehanzi==20shibiehanzi="琼";elseifshibiehanzi==21shibiehanzi="川";elseifshibiehanzi==22shibiehanzi="贵";elseifshibiehanzi==23shibiehanzi="云";elseifshibiehanzi==24shibiehanzi="陕";elseifshibiehanzi==25shibiehanzi="甘";elseifshibiehanzi==26shibiehanzi="青";elseifshibiehanzi==27shibiehanzi="藏";elseifshibiehanzi==28shibiehanzi="桂";elseifshibiehanzi==29shibiehanzi="蒙";elseifshibiehanzi==30shibiehanzi="新";elseifshibiehanzi==31shibiehanzi="宁";elseifshibiehanzi==32shibiehanzi="港";end functionshibiezimu=shibiezimu(zimu,xiuzhengzimu)[y,x,z]=size(xiuzhengzimu);fork=1:24sum=0;fori=1:yforj=1:xifzimu(i,j,k)==xiuzhengzimu(i,j)%统计黑白sum=sum+1;endendendbaifenbi(1,k)=sum/(x*y);endchepai=find(baifenbi>=max(baifenbi));shibiezimu=chepai;%在数字中，从0开始所以要减一。这里不用ifshibiezimu==1shibiezimu="A";elseifshibiezimu==2shibiezimu="B";elseifshibiezimu==3shibiezimu="C";elseifshibiezimu==4shibiezimu="D";elseifshibiezimu==5shibiezimu="E";elseifshibiezimu==6shibiezimu="F";elseifshibiezimu==7shibiezimu="G";elseifshibiezimu==8shibiezimu="H";elseifshibiezimu==9shibiezimu="J";elseifshibiezimu==10shibiezimu="K";elseifshibiezimu==11shibiezimu="L";elseifshibiezimu==12shibiezimu="M";elseifshibiezimu==13shibiezimu="N";elseifshibiezimu==14shibiezimu="P"; elseifshibiezimu==15shibiezimu="Q";elseifshibiezimu==16shibiezimu="R";elseifshibiezimu==17shibiezimu="S";elseifshibiezimu==18shibiezimu="T";elseifshibiezimu==19shibiezimu="U";elseifshibiezimu==20shibiezimu="V";elseifshibiezimu==21shibiezimu="W";elseifshibiezimu==22shibiezimu="X";elseifshibiezimu==23shibiezimu="Y";elseifshibiezimu==24shibiezimu="Z";endfunctionshibiezm_sz=shibiezm_sz(shuzizimu,xiuzhengzm_sz)[y,x,z]=size(xiuzhengzm_sz);fork=1:34sum=0;fori=1:yforj=1:xifshuzizimu(i,j,k)==xiuzhengzm_sz(i,j)%统计黑白sum=sum+1;endendendbaifenbi(1,k)=double(sum/(x*y));endchepai=find(baifenbi>=max(baifenbi));chepai=chepai(1,1);%===================数字直接返回=========================if(chepai>=1)&&(chepai<=10)zm_sz=chepai-1;zm_sz=num2str(zm_sz);elseif(chepai>=11)&&(chepai<=34) %==================字母对应序号转字母====================%=========11-15ABCDE16-20FGHIJ21-25KLMNP26-30QRSTU31-35VWXYZifchepai==11zm_sz="A";elseifchepai==12zm_sz="B";elseifchepai==13zm_sz="C";elseifchepai==14zm_sz="D";elseifchepai==15zm_sz="E";elseifchepai==16zm_sz="F";elseifchepai==17zm_sz="G";elseifchepai==18zm_sz="H";elseifchepai==19zm_sz="J";elseifchepai==20zm_sz="K";elseifchepai==21zm_sz="L";elseifchepai==22zm_sz="M";elseifchepai==23zm_sz="N";elseifchepai==24zm_sz="P";elseifchepai==25zm_sz="Q";elseifchepai==26zm_sz="R";elseifchepai==27zm_sz="S";elseifchepai==28zm_sz="T";elseifchepai==29zm_sz="U";elseifchepai==30zm_sz="V";elseifchepai==31 zm_sz="W";elseifchepai==32zm_sz="X";elseifchepai==33zm_sz="Y";elseifchepai==34zm_sz="Z";endendshibiezm_sz=zm_sz;functionedit2_Callback(hObject,eventdata,handles)%hObjecthandletoedit2(seeGCBO)%eventdatareserved-tobedefinedinafutureversionofMATLAB%handlesstructurewithhandlesanduserdata(seeGUIDATA)%Hints:get(hObject,"String")returnscontentsofedit2astext%str2double(get(hObject,"String"))returnscontentsofedit2asadouble%---Executesduringobjectcreation,aftersettingallproperties.functionedit2_CreateFcn(hObject,eventdata,handles)%hObjecthandletoedit2(seeGCBO)%eventdatareserved-tobedefinedinafutureversionofMATLAB%handlesempty-handlesnotcreateduntilafterallCreateFcnscalled%Hint:editcontrolsusuallyhaveawhitebackgroundonWindows.%SeeISPCandCOMPUTER.ifispc&&isequal(get(hObject,"BackgroundColor"),get(0,"defaultUicontrolBackgroundColor"))set(hObject,"BackgroundColor","white");endfunctionedit3_Callback(hObject,eventdata,handles)%hObjecthandletoedit3(seeGCBO)%eventdatareserved-tobedefinedinafutureversionofMATLAB%handlesstructurewithhandlesanduserdata(seeGUIDATA)%Hints:get(hObject,"String")returnscontentsofedit3astext%str2double(get(hObject,"String"))returnscontentsofedit3asadouble %---Executesduringobjectcreation,aftersettingallproperties.functionedit3_CreateFcn(hObject,eventdata,handles)%hObjecthandletoedit3(seeGCBO)%eventdatareserved-tobedefinedinafutureversionofMATLAB%handlesempty-handlesnotcreateduntilafterallCreateFcnscalled%Hint:editcontrolsusuallyhaveawhitebackgroundonWindows.%SeeISPCandCOMPUTER.ifispc&&isequal(get(hObject,"BackgroundColor"),get(0,"defaultUicontrolBackgroundColor"))set(hObject,"BackgroundColor","white");end%---ExecutesonbuttonpressinDCT.functionDCT_Callback(hObject,eventdata,handles)%hObjecthandletoDCT(seeGCBO)%eventdatareserved-tobedefinedinafutureversionofMATLAB%handlesstructurewithhandlesanduserdata(seeGUIDATA)globalbw;globalI;I1=I;%setFI=abs((fft2(I1)));NFI=255*mat2gray(FI);%归一化SFI=fftshift(NFI);imgray=rgb2gray(SFI);%灰阶grayImage=rgb2gray(I1);dctgrayImage=dct2(grayImage);colormap(gray(4));colorbar;dctgrayImage(abs(dctgrayImage)<0.1)=0;I2=idct2(dctgrayImage)/255;axes(handles.axes1);subplot(2,2,1),imshow(I1);title("原始图像");subplot(2,2,2),imshow(imgray),title("FFTof原始图像");subplot(2,2,3),imshow(log(abs(dctgrayImage)),[]);title("DCT变换灰度图像");subplot(2,2,4),imshow(I2),title("DCT逆变换的灰度图像");%---Executesonbuttonpressinreset.functionreset_Callback(hObject,eventdata,handles) %hObjecthandletoreset(seeGCBO)%eventdatareserved-tobedefinedinafutureversionofMATLAB%handlesstructurewithhandlesanduserdata(seeGUIDATA)clear;clc;close'