matlab实现分水岭算法处理图像分割

此程序为优化后的分水岭算法，避免了图像过分割

1. I= imread('D:\Images\pic_loc\1870405130305041503.jpg');
2. imshow(I);
3. h=fspecial('sobel'); %h = fspecial(type) creates a two-dimensional filter h of the specified type. fspecial returns h as
4. %a correlation kernel, which is the appropriate form to use with imfilter. type is a string having one of these values.
5. fd=double(I);%double使数据变成双精度
6. g=sqrt(imfilter(fd,h,'replicate').^2+imfilter(fd,h','replicate').^2);
7. figure;
8. imshow(g);
9. g2=imclose(imopen(g,ones(3,3)),ones(3,3));
10. figure;
11. imshow(g2);
12. im=imextendedmin(g2,10);   %
13. Lim=watershed(bwdist(im)); %watershed分水岭算法 Lim的值greater than or equal to 0，等于0是分水岭脊像素
14. em=Lim==0;
15. g3=imimposemin(g2,im|em);
16. g4=watershed(g3);
17. figure;
18. imshow(g4);
19. g5=I;
20. g5(g4==0)=255;
21. figure;
22. imshow(g5);

实验结果：

from: http://blog.csdn.net/abcjennifer/article/details/6669074