经高斯滤波后用riberts算子边缘检测

clc;close all;clear all

%---------------------------------------------生成高斯平滑滤波模板-------------------------------------------------------------%

hg=zeros(3,3); %设定高斯平滑滤波模板的大小为3*3

delta=0.5;

for x=1:1:3

for y=1:1:3

u=x-2;

v=y-2;

hg(x,y)=exp(-(u^2+v^2)/(2*pi*delta^2));

end

end

h=hg/sum(hg(:));

%------------------------------------------------读入图像------------------------------------------------------------------------%

f = imread('F:\ZPB\3.tif'); % 读入图像文件

f=rgb2gray(im2double(f));

imshow(f)

title('原始图像');

[m,n]=size(f);

ftemp=zeros(m,n);

rowhigh=m-1;

colhigh=n-1;

%%%---------------------------------------------高斯滤波----------------------------------------%%%

for x=2:1:rowhigh-1

for y=2:1:colhigh-1

mod=[f(x-1,y-1) f(x-1,y) f(x-1,y+1); f(x,y-1) f(x,y) f(x,y+1);f(x+1,y-1) f(x+1,y) f(x+1,y+1)];

A=h.*mod;

ftemp(x,y)=sum(A(:));

end

end

f=ftemp

figure,imshow(f)

title('通过高斯滤波器后的图像');

 

%----------------------------------------利用roberts算子进行边缘检测--------------------------------------------------%

sx=[-1 -2 -1;0 0 0;1 2 1];

sy=[-1 0 1;-2 0 2;-1 0 1];%%%%%你可以替换成其他算子，这里是罗伯特算子

 

% sx=[-1 -2 -1;0 0 0;1 2 1];

% sy=[-1 0 1;-2 0 2;-1 0 1];这个是Sobel算子，类似的，你可以替换成canny算子等等

for x=2:1:rowhigh-1

for y=2:1:colhigh-1

mod=[f(x-1,y-1) f(x-1,y) f(x-1,y+1); f(x,y-1) f(x,y) f(x,y+1);f(x+1,y-1) f(x+1,y) f(x+1,y+1)];

fsx=sx.*mod;

fsy=sy.*mod;

ftemp(x,y)=sqrt((sum(fsx(:)))^2+(sum(fsy(:)))^2);

end

end

fr=im2uint8(ftemp);

figure,imshow(fr)

title('用roberts算子边缘检测的原始图像');

 

%--------------------------------------域值分割，这部分可不参---------------------------------------------------------------%

TH1=60; %设定阈值

for x=2:1:rowhigh-1

for y=2:1:colhigh-1

if (fr(x,y)>=TH1)&((fr(x,y-1) <= fr(x,y)) & (fr(x,y) > fr(x,y+1)) )

fr(x,y)=200;

elseif(fr(x,y)>=TH1)&( (fr(x-1,y) <=fr(x,y)) & (fr(x,y) >fr(x+1,y)))

fr(x,y)=200;

else fr(x,y)=50;

end

end

end

figure,imshow(fr)

title('用roberts算子边缘检测并细化后的图像');