Matlab中开发一个名为regiongrow的M函数来完成基本的区域生长。
该函数为 [g,NR,SI,TI]=regiongrow(f,S,T) 输入中:f为输入图像,S为种子,T为阈值(标量时为全 局阈值) 输出中:g为分割后的图像,NR为连通区域的数目,SI为一幅 包含有种子点的图像。SI也为一幅图像,包含在连通性处理前, 通过阈值检测的像素。
Matlab程序举例如下:(程序使用时候,regiongrow一定要先定义,这个我不用交吧?)
i=imread('eight.tif');
figure(1);imshow(i);
% i=doulbe(i);
[m,n]=size(i);
[y1,x1]=getpts;
x1=round(x1);y1=round(y1);
seed=[x1,y1];
th_mean=40;
yout=regiongrow(i,seed,th_mean);
figure(2);imshow(yout);title('区域增长');
%原图:
%增长之后的:
%%%%%%%%%%%%%%regiongrow区域增长
function yout=regiongrow(I,seed,th_mean)
[m,n]=size(I);
[I h]=size(seed);
yout=zeros(m,n);
for i=1:I
yout(seed(i,1),seed(i,2))=1;
end
for i=1:I
sum(seed(i,1),seed(i,2));
end
seed_mean=mean(sum);
ok=true;
s_star=1;
s_end=1;
while ok
ok=false;
for i=s_star:s_end
x=seed(i,1);
y=seed(i,2);
if x>2&&(x+1)<m&&y>2&&(y+1)<n
for u=-1:1
for v=-1:1
if yout(x+u,y+v)==0
avs(I(x+u,y+v)-seed_mean)<=th_mean
yout(x+u,y+v)=I
ok=true;
seed=[seed;[x+u,y+v]];
end
end
end
end
end
s_star=s_end+1;
[I h]=size(seed);
s_end=1
end
上述代码不太正确,重新弄个简单的