matlab图像处理算法源码.docx

matlab图像处理算法源码252020年4月19日文档仅供参考matlab图像分割算法源码1.图像反转MATLAB程序实现如下：J=double(I);J=-J+(256-1); %图像反转线性变换H=uint8(J);subplot(1,2,1),imshow(I);subplot(1,2,2),imshow(H);2.灰度线性变换MATLAB程序实现如下：I=imread(xian.bmp);subplot(2,2,1),imshow(I);title(原始图像);axis(0,250,0,200);axis on; %显示坐标系I1=rgb2gray(I);subplot(2,2,2),imshow(I1);title(灰度图像);axis(0,250,0,200);axis on; %显示坐标系J=imadjust(I1,0.1 0.5,); %局部拉伸，把0.1 0.5内的灰度拉伸为0 1subplot(2,2,3),imshow(J);title(线性变换图像0.1 0.5);axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系K=imadjust(I1,0.3 0.7,); %局部拉伸，把0.3 0.7内的灰度拉伸为0 1subplot(2,2,4),imshow(K);title(线性变换图像0.3 0.7);axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系3.非线性变换（对数变换图像）%图像的对数变换原理就是将源图像的像素点的像素值经过对数函数映射到目标空间，然后%用目标值来代替原来像素点的像素值。那么它的作用就是对于一些视频采集设备，它们采%集的图像会有一个指数的失真，用对数变换后就抵消掉了原来的指数失真得到原来的图%像，还有对于灰度值偏低的图像来说它是有一个锐化的效果的。MATLAB程序实现如下：I=imread(xian.bmp);I1=rgb2gray(I);subplot(1,2,1),imshow(I1);title(灰度图像);axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系J=double(I1);J=40*(log(J+1);H=uint8(J);subplot(1,2,2),imshow(H);title(对数变换图像);axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系4.直方图均衡化% 直方图均衡化 解释网址：MATLAB程序实现如下：I=imread(xian.bmp);I=rgb2gray(I);figure;subplot(2,2,1);imshow(I);subplot(2,2,2);imhist(I); %获取图像数据直方图I1=histeq(I); %直方图均衡化figure;subplot(2,2,1);imshow(I1);subplot(2,2,2);imhist(I1);5.线性平滑滤波器%空间域的平滑滤波一般采用简单平均法进行，就是求邻近像元点的平均亮度值。邻域的大小与平滑的效果直接相关，邻域越大平滑的效果越好，但邻域过大，平滑会使边缘信息损失的越大，从而使输出的图像变得模糊，因此需合理选择邻域的大小。用MATLAB实现领域平均法抑制噪声程序：I=imread(xian.bmp);subplot(2,3,1)imshow(I)title(原始图像)I=rgb2gray(I);I1=imnoise(I,salt & pepper,0.02); % subplot(2,3,2)imshow(I1)title(添加椒盐噪声的图像)k1=filter2(fspecial(average,3),I1)/255; %进行3*3模板平滑滤波k2=filter2(fspecial(average,5),I1)/255; %进行5*5模板平滑滤波k3=filter2(fspecial(average,7),I1)/255; %进行7*7模板平滑滤波k4=filter2(fspecial(average,9),I1)/255; %进行9*9模板平滑滤波subplot(2,3,3),imshow(k1),title(3*3模板平滑滤波);subplot(2,3,4),imshow(k2),title(5*5模板平滑滤波);subplot(2,3,5),imshow(k3),title(7*7模板平滑滤波);subplot(2,3,6),imshow(k4),title(9*9模板平滑滤波);6.中值滤波器%中值也称中位数，即数据按升序或者降序排列，假如有n个数据，当n为偶数时，中位数为第n/2位数和第(n+2)/2位数的平均数；如果n为奇数，那么中位数为第（n+1）/2位数的值。用MATLAB实现中值滤波程序如下：I=imread(xian.bmp);I=rgb2gray(I);J=imnoise(I,salt & pepper,0.02);subplot(231),imshow(I);title(原图像);subplot(232),imshow(J);title(添加椒盐噪声图像);k1=medfilt2(J); %进行3*3模板中值滤波k2=medfilt2(J,5,5); %进行5*5模板中值滤波k3=medfilt2(J,7,7); %进行7*7模板中值滤波k4=medfilt2(J,9,9); %进行9*9模板中值滤波subplot(233),imshow(k1);title(3*3模板中值滤波);subplot(234),imshow(k2);title(5*5模板中值滤波);subplot(235),imshow(k3);title(7*7模板中值滤波);subplot(236),imshow(k4);title(9*9模板中值滤波);7.用Sobel算子和拉普拉斯对图像锐化：%索贝尔算子（Sobel operator）是图像处理中的算子之一，主要用作边缘检测。在技术上，它是一离散性差分算子，用来运算图像亮度函数的梯度之近似值。在图像的任何一点使用此算子，将会产生对应的梯度矢量或是其法矢量 I=imread(xian.bmp);subplot(2,2,1),imshow(I);title(原始图像);axis(50,250,50,200);grid on; %显示网格线axis on; %显示坐标系I1=im2bw(I);subplot(2,2,2),imshow(I1);title(二值图像);axis(50,250,50,200);grid on; %显示网格线axis on; %显示坐标系H=fspecial(sobel); %选择sobel算子 J=filter2(H,I1); %卷积运算subplot(2,2,3),imshow(J); title(sobel算子锐化图像);axis(50,250,50,200);grid on; %显示网格线axis on; %显示坐标系h=0 1 0,1 -4 1,0 1 0; %拉普拉斯算子J1=conv2(I1,h,same); %卷积运算subplot(2,2,4),imshow(J1); title(拉普拉斯算子锐化图像);axis(50,250,50,200);grid on; %显示网格线axis on; %显示坐标系8.梯度算子检测边缘用MATLAB实现如下：I=imread(xian.bmp);subplot(2,3,1);imshow(I);title(原始图像);axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系I1=im2bw(I);subplot(2,3,2);imshow(I1);title(二值图像);axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系I2=edge(uint8(I1),roberts);%figure;subplot(2,3,3);imshow(I2);title(roberts算子分割结果);axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系I3=edge(uint8(I1),sobel);subplot(2,3,4);imshow(I3);title(sobel算子分割结果);axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系I4=edge(uint8(I1),Prewitt);subplot(2,3,5);imshow(I4);title(Prewitt算子分割结果);axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系9.LOG算子检测边缘用MATLAB程序实现如下：I=imread(xian.bmp);subplot(2,2,1);imshow(I);title(原始图像);I1=rgb2gray(I);subplot(2,2,2);imshow(I1);title(灰度图像);I2=edge(I1,log);subplot(2,2,3);imshow(I2);title(log算子分割结果);10.Canny算子检测边缘用MATLAB程序实现如下：I=imread(xian.bmp);subplot(2,2,1);imshow(I);title(原始图像)I1=rgb2gray(I);subplot(2,2,2);imshow(I1);title(灰度图像);I2=edge(I1,canny);subplot(2,2,3);imshow(I2);title(canny算子分割结果);10.边界跟踪（bwtraceboundary函数）clcclear allI=imread(xian.bmp);figureimshow(I);title(原始图像);I1=rgb2gray(I); %将彩色图像转化灰度图像 threshold=graythresh(I1); %计算将灰度图像转化为二值图像所需的门限BW=im2bw(I1, threshold); %将灰度图像转化为二值图像figureimshow(BW);title(二值图像);dim=size(BW);col=round(dim(2)/2)-1; %计算起始点列坐标 52和图像大小有关row=find(BW(:,col),1); %计算起始点行坐标connectivity=8;num_points=180;contour=bwtraceboundary(BW,row,col,N,connectivity,num_points); %提取边界figureimshow(I1);hold on;plot(contour(:,2),contour(:,1), g,LineWidth ,2);title(边界跟踪图像);12.Hough变换clcclearI= imread(xian.bmp);rotI=rgb2gray(I);subplot(2,2,1);imshow(rotI);title(灰度图像);axis(0,250,0,200);grid on; axis on; BW=edge(rotI,prewitt);subplot(2,2,2);imshow(BW);title(prewitt算子边缘检测后图像);axis(0,250,0,200);grid on; axis on; H,T,R=hough(BW);subplot(2,2,3);imshow(H,XData,T,YData,R,InitialMagnification,fit);title(霍夫变换图);xlabel(theta),ylabel(rho);axis on , axis normal, hold on;P=houghpeaks(H,5,threshold,ceil(0.3*max(H(:);x=T(P(:,2);y=R(P(:,1);plot(x,y,s,color,white);lines=houghlines(BW,T,R,P,FillGap,5,MinLength,7);subplot(2,2,4);,imshow(rotI);title(霍夫变换图像检测);axis(0,250,0,200);grid on; axis on; hold on;max_len=0;for k=1:length(lines)xy=lines(k).point1;lines(k).point2;plot(xy(:,1),xy(:,2),LineWidth,2,Color,green);plot(xy(1,1),xy(1,2),x,LineWidth,2,Color,yellow);plot(xy(2,1),xy(2,2),x,LineWidth,2,Color,red);len=norm(lines(k).point1-lines(k).point2);if(lenmax_len)max_len=len;xy_long=xy;endendplot(xy_long(:,1),xy_long(:,2),LineWidth,2,Color,cyan);13.直方图阈值法用MATLAB实现直方图阈值法：I=imread(xian.bmp);I1=rgb2gray(I);figure;subplot(2,2,1);imshow(I1);title(灰度图像)axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系m,n=size(I1); %测量图像尺寸参数GP=zeros(1,256); %预创立存放灰度出现概率的向量for k=0:255 GP(k+1)=length(find(I1=k)/(m*n); %计算每级灰度出现的概率，将其存入GP中相应位置endsubplot(2,2,2),bar(0:255,GP,g) %绘制直方图title(灰度直方图)xlabel(灰度值)ylabel(出现概率) I2=im2bw(I,150/255); subplot(2,2,3),imshow(I2);title(阈值150的分割图像)axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系I3=im2bw(I,200/255); %subplot(2,2,4),imshow(I3);title(阈值200的分割图像)axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系14. 自动阈值法：Otsu法用MATLAB实现Otsu算法：clcclear allI=imread(xian.bmp);subplot(1,2,1),imshow(I);title(原始图像)axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系level=graythresh(I); %确定灰度阈值BW=im2bw(I,level);subplot(1,2,2),imshow(BW);title(Otsu法阈值分割图像)axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系15.膨胀操作I=imread(xian.bmp); %载入图像I1=rgb2gray(I);subplot(1,2,1);imshow(I1);title(灰度图像) axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系se=strel(disk,1); %生成圆形结构元素I2=imdilate(I1,se); %用生成的结构元素对图像进行膨胀subplot(1,2,2);imshow(I2);title(膨胀后图像);axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系16.腐蚀操作MATLAB实现腐蚀操作I=imread(xian.bmp); %载入图像I1=rgb2gray(I);subplot(1,2,1);imshow(I1);title(灰度图像) axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系se=strel(disk,1); %生成圆形结构元素I2=imerode(I1,se); %用生成的结构元素对图像进行腐蚀subplot(1,2,2);imshow(I2);title(腐蚀后图像);axis(0,250,0,200);grid on; %显示网格线axis on; %显示坐标系17.开启和闭合操作用MATLAB实现开启和闭合操作I=imread(xian.bmp); %载入图像subplot(2,2,1),imshow(I);title(原始图像);axis(50,250,50,200);axis on; %显示坐标系 I1=rgb2gray(I);subplot(2,2,2),imshow(I1);title(灰度图像);axis(50,250,50,200);axis on; %显示坐标系 se=strel(disk,1); %采用半径为1的圆作为结构元素I2=imopen(I1,se); %开启操作I3=imclose(I1,se); %闭合操作subplot(2,2,3),imshow(I2);title(开启运算后图像);axis(50,250,50,200);axis on; %显示坐标系subplot(2,2,4),imshow(I3);title(闭合运算后图像);axis(50,250,50,200); axis on; %显示坐标系18.开启和闭合组合操作I=imread(xian.bmp); %载入图像subplot(3,2,1),imshow(I);title(原始图像);axis(50,250,50,200);axis on; %显示坐标系 I1=rgb2gray(I);subplot(3,2,2),imshow(I1);title(灰度图像);axis(50,250,50,200);axis on; %显示坐标系 se=strel(disk,1); I2=imopen(I1,se); %开启操作I3=imclose(I1,se); %闭合操作subplot(3,2,3),imshow(I2);title(开启运算后图像);axis(50,250,50,200);axis on; %显示坐标系subplot(3,2,4),imshow(I3);title(闭合运算后图像);axis(50,250,50,200);axis on; %显示坐标系se=strel(disk,1); I4=imopen(I1,se);I5=imclose(I4,se);subplot(3,2,5),imshow(I5); %开闭运算图像title(开闭运算图像);axis(50,250,50,200);axis on; %显示坐标系 I6=imclose(I1,se);I7=imopen(I6,se);subplot(3,2,6),imshow(I7); %闭开运算图像 title(闭开运算图像);axis(50,250,50,200);axis on; %显示坐标系 19.形态学边界提取利用MATLAB实现如下：I=imread(xian.bmp); %载入图像subplot(1,3,1),imshow(I);title(原始图像); axis(50,250,50,200);grid on; %显示网格线axis on; %显示坐标系I1=im2bw(I);subplot(1,3,2),imshow(I1);title(二值化图像);axis(50,250,50,200);grid on; %显示网格线axis on; %显示坐标系I2=bwperim(I1); %获取区域的周长subplot(1,3,3),imshow(I2); title(边界周长的二值图像);axis(50,250,50,200);grid on;axis on; 20.形态学骨架提取利用MATLAB实现如下：I=imread(xian.bmp);subplot(2,2,1),imshow(I);title(原始图像);axis(50,250,50,200);axis on; I1=im2bw(I);subplot(2,2,2),imshow(I1);title(二值图像);axis(50,250,50,200);axis on; I2=bwmorph(I1,skel,1);subplot(2,2,3),imshow(I2);title(1次骨架提取);axis(50,250,50,200);axis on; I3=bwmorph(I1,skel,2);subplot(2,2,4),imshow(I3);title(2次骨架提取);axis(50,250,50,200);axis on; 21.直接提取四个顶点坐标I = imread(xian.bmp);I = I(:,:,1);BW=im2bw(I); figureimshow(BW)x,y=getpts