霍夫变换实现直线检测——MATLAB

网上好多不能用，就重新写了一个：

% 入口图像为 BW，出口图像为fclc,closeBW=imread('D:\picture\9dafa605d53eea243812bb29.jpg');BW=rgb2gray(BW);thresh=[0.01,0.17];sigma=2;%定义高斯参数f = edge(double(BW),'canny',thresh,sigma);figure(1),imshow(f,[]);title('canny 边缘检测');[H, theta, rho]= hough(f, 0.5);%imshow(theta,rho,H,[],'notruesize'),axis on,axis normal%xlabel('\theta'),ylabel('rho');[r,c]=houghpeaks(H,5);hold onlines=houghlines(f,theta,rho,r,c);figure,imshow(f,[]),title('Hough Transform Detect Result'),hold onfor k=1:length(lines)     xy=[lines(k).point1;lines(k).point2];plot(xy(:,1),xy(:,2),'LineWidth',4,'Color',[.6 .6 .6]);end

Matlab2012版本的内核函数有所改变，houghpeaks, houghlines等函数有所改变，感谢cuicd2011的提醒，R2012a版本的直线检测稍有变化：

clc,close  BW=imread('D:\picture\9dafa605d53eea243812bb29.jpg');  BW=rgb2gray(BW);  thresh=[0.01,0.17];  sigma=2;%定义高斯参数  f = edge(double(BW),'canny',thresh,sigma);  figure(1),imshow(f,[]);  title('canny 边缘检测');    [H, theta, rho]= hough(f,'RhoResolution', 0.5);  %imshow(theta,rho,H,[],'notruesize'),axis on,axis normal  %xlabel('\theta'),ylabel('rho');    peak=houghpeaks(H,5);  hold on    lines=houghlines(f,theta,rho,peak);  figure,imshow(f,[]),title('Hough Transform Detect Result'),hold on  for k=1:length(lines)      xy=[lines(k).point1;lines(k).point2];      plot(xy(:,1),xy(:,2),'LineWidth',4,'Color',[.6 .6 .6]);  end  

实验结果：