图像变换--灰度调整、滤波增强---二值化（阈值分割）--形态学处理--特征提取

一：色彩变换，从颜色上控制
思考怎样应用灰度的变换、滤波变换减少地物

直方图均衡化histeq
自适应均衡化adapthisteq
灰度调节（灰度映射）imadjust【直接变换、非线性变换、灰度映射】
图像相减imsubtract（一张求减可以调节灰度值、两张相减可以消除一些地物）
组合型：直方图均衡自适应调整
图像相减

背景色剔除：
缨帽滤波变换：调整背景色，增大对象间隙

图像补色：照亮谷底

 

滤波变换：消除物体filter2（）、imfilter、medfilt2
不常用：抖动

二：二值化，中介

然后二值化：im2bw
组合型：自动灰度阈值分割二值化：J = im2bw(I,graythresh(I));
im2bw(I,100/255)
二值化修饰imshow(bw,[1 0 0;0 1 0])
求反色
求补色

三、形态学处理，消除不必要的地物
然后形态学处理，再减去些东西，{消除物体的方法}（膨胀、腐蚀、开运算、闭运算、与运算、加减运算、空洞填充、移除边界连通的目标、缨帽组合变换）

四特征提取：进行最终的提取
最后通过特征提取进行提取（骨架化、边缘像素测定、绘制轮廓线、边缘提取、等值线提取）

 

下面是一些处理的函数:

显示：
hist(I);

whos

imfinfo('')

imwrite(I,'XX.tif');

纹理映射方式观察
[x,y,z]= cylinder;warp(x,y,z,I);

指定输出图像像素的大小：J = imresize(I,1.25);【tif】

图像的旋转：J = imrotate(I,90,'bilinear');【tif】

图像的裁剪：J = imcrop(I,[60 40 100 90]);【tif】

区域属性度量 stats = regionprops(bwlabel(I),'all');stats(23)【png】【也不用imshow】

 

图像转换

索引图像转换为RGB显示：imshow(ind2rgb(X,map))
灰度图像转换为RGB图像：imshow(cat(3,I,I,I))
将数据矩阵转换成灰度图像imshow(mat2gray(I))
mat2gray(fitsread('solarspectra.fts'))【fts格式的处理】
imshow(fitsread('solarspectra.fts'),[]);
【解释：通过算子对图像滤波后的图像即是数据矩阵，当然数据矩阵也能直接显示
当然如果数据矩阵不想直接转化成灰度图像时，另一种简单的方法是调节灰度值
imshow(J,[])（其中J为滤波后的数据矩阵）】
将灰度图像转化为索引图像grayslice(I,16);
将RGB图像转换为灰度图像rgb2gray(RGB)
将真彩png图像转化为索引图像[X,map] = rgb2ind(I,128);
将真彩、索引色、灰度图转化为二值图：imshow(im2bw(X,map,0.4));imshow(im2bw(I,0.4))
将索引图转化为灰度图load trees imshow(ind2gray(X,map))
[X,map] = imread('forest.tif');
I = ind2gray(X,map);

 

灰度调整：增强（只改变灰度和颜色，不消除物体，不改变图像的原有物质）
imshow(I,[]);
histeq(I)
adapthisteq(I)

imsubtract(I,50)（图像相减）

灰度调节imadjust(I,stretchlim(I),[]);{提高对比度}
imadjust(I,stretchlim(I),[0 1]);

直方图均衡自适应调整bw1 = imadjust(adapthisteq(I,'NumTiles',[10 10]));

图像的二值化level = graythresh(I),J = im2bw(I,level);
J = im2bw(I,graythresh(I));【自动灰度阈值分割】

自定义阈值分割figure,imhist(I);J=im2bw(I,100/255);【通过灰度来进行二值化】

二值图像的调节色彩imshow(bw,[1 0 0;0 1 0])

求反色imshow(~I)
求补色imshow(imcomplement(I));

通过乘法来改变亮度和对比度（乘以数值改变亮度，乘以本身图像改变对比度）
J= immultiply(I,0.2);
I16 = uint16(bw8);bw9 = immultiply(I16,I16);

直接灰度变换J = imadjust(I,[0 1],[1 0],1.5);[PNG]【灰度倒置线性变换】
灰度的非线性变换J = 45*log(double(I)+1);figure,imshow(J,[])

直方图灰度变换J = imadjust(I,[0.15 0.9],[0 1]);figure;imshow(J);figure,imhist(J,64)【区域映射】【增强减弱图像的对比度】
灰度范围的映射J = imadjust(I,[0 0.2],[0.5 1]);【此方法有利于消除背景区域，然后方便边界的提取】
索引图的灰度变换[X,map] = imread('forest.tif');I = ind2gray(X,map);J = imadjust(I,[],[],0.5);

改变图像以调节对比度
抖动 imshow(dither(I));

滤波和灰度调整是为了消除图像中的一些东西
目的相同，原理不同
J= filter2([1 2;-1 -2],I)
J = filter2([1 2 1;0 0 0;-1 -2 -1],I);
J = filter2(fspecial('sobel'),I);

I2 = imfilter(I,ones(5,5)/25);

中值滤波K = medfilt2(j);【把杂色过滤掉】

 

灰度调整与消除背景后的图像相加有异曲同工之妙，都能够进行亮更亮，暗更暗的效果。

 

处理方法，先进行背影的去除，然后进行（图像相加）增强（第一次增加对比度），然后进行第二次的效果增强。对比度调节，到了这个时候明暗已经很突出了，不过还有些微小的灰度变化，然进行一下彻底地图像二值化，进行完全的黑白分明，这样就能进行边缘提取，要是感觉还不行的话，那就进行形态学的膨胀或腐蚀的操作，消除微小物，然后在进行分割操作

 

图像的提取

1 背景色的处理

图像消除背景色，先提取背影图，然后再进行图像相减(背景色不一致的可以用这个方法)

背景图的提取（开运算）
se = strel('disk',25);
background = imopen(I,se);
figure,imshow(background);

blocks = blkproc(I,[32 32],'min(x(:))');
background = imresize(blocks,[256 256],'bilinear');
�ckground = imresize(blkproc(I,[32 32],'min(x(:))'),[256 256],'bilinear');

减去背景色bw3 = imsubtract(I,background);

图像相减imsubtract(I,J);

图像的开运算imopen(I,se);se = strel('disk',25);

查看背景色是否正常的方法
figure,surf(double(I(1:8:end,1:8:end))),zlim([0 255]);
set(gca,'ydir','reverse');

对一副图像的各种处理中的图像可进行相减来消除一些东西
imsubtract(I,J);

两幅同样的图像可以进行背景去除后的相加操作是特征更加明显
imadd(I,J);
这点和直接加一个数值的区别是后者是整体相加，而前者是亮亮相加

经过灰度变换后使图像的灰度对比很大时就可以进行边缘提取了

边缘提取edge(I);
BW2 = edge(I,'canny');
BW = edge(I,'sobel')
BW = edge(I,'prewitt')
BW = edge(I,'roberts')
BW = edge(I,'log')
bws = edge(I,'sobel',(graythresh(topI)));

先进行灰度划分然后进行边缘检测bws = edge(I,'sobel',(graythresh(I)*.1));

图像块边缘检测J = nlfilter(I,[3 3],inline('max(x(:))'));【tif】【此方法进行边缘的扩充】【邻域检测】

等高线提取imcontour(I),imcontour(I,3);【过程提取比较慢】

形态学的膨胀imdilate(I,se);
【可通过膨胀填补空隙】
bw = imdilate(K,[strel('line',3,90) strel('line',3,0)]);
bwsdil = imdilate(bws,[strel('line',3,90)]);
可参照帮助文档strel 和imdilate

空洞的填充J = imfill(I,'holes');J = imfill(I);

移除与边界连通的目标：J= imclearborder(I,4);|J= imclearborder(I);

腐蚀操作：se = strel('diamond',1);J = imerode(I,se);J = imerode(J,se);figure,imshow(J);【此方法可以对边缘进行平滑】

绘制轮廓线J = bwperim(I);

两幅图像叠加的显示segout = I;segout(bwoutline)= 255;figure,imshow(segout);【一般与绘制轮廓线同时使用，是图像分割的方法】

开运算：bwco = imopen(bwc,strel('disk',6));
se = strel('rectangle',[40 30]);
bw2 = imopen(bw1,se);【此方法用来删除图像中较小的物体】

闭运算：bwc = imclose(bw,strel('disk',6));【抽取原始图像较大的物体】

图像“与”操作imshow（bw & bwco）【包含小对象和包含大对象的图像进行“与”操作，从而得到原始图像中所有的小对象】

骨架化b2 = bwmorph(b1,'skel',Inf);【骨架提取适用于道路的提取】
b3 = bwmorph(b1,'remove');【骨骼提取、边缘检测最好先将图像转化为二进制图像】

边缘像素测定bw2 = bwperim(bw1);|BW2 = bwperim(BW1,8);

缨帽变换（高帽滤波变换）：J = imtophat(gI,strel('disk',10));【消除背景中那些亮度不一致的背景】

低帽滤波变换Ibot = imbothat(afm,se);figure,imshow(Ibot,[]);

增大对象间的间隙【原图像加上高帽滤波变换的图像，然后再减去低帽滤波的图像】

Ienhance = imsubtract(imadd(Itop,afm),Ibot);

谷点图像的增强，照亮谷点图像
Iec = imcomplement(Ienhance);

 

下面试一下实验的小代码可以作为参考

在每幅图像处理前都可以用下面的方法进行检测
《《《《《《《《《《《《《《
clear,close all
I = imread('f:\s.png');
%figure,imshow(I);
%I = imread('rice.png');
imshow(I);
figure,imhist(I);
bw = im2bw(I,230/255);
figure;imshow(bw);
阈值分割
********************
边缘检测
检测灰度梯度相差很大的区域

clear,close all
%I = imread('f:\s.png');
%figure,imshow(I);
I = imread('rice.png');

figure;
imshow(I);
bw1 = edge(I,'Roberts');
figure,imshow(bw1);
bw2 = edge(I,'sobel');
figure,imshow(bw2);
bw3 = edge(I,'prewitt');
figure,imshow(bw3);
bw4 =edge(I,'canny');
figure,imshow(bw4);
******************************
通过图像分割检测细胞 P217
clear,close all
%I = imread('f:\s.png');
%figure,imshow(I);
I = imread('cell.tif');
figure,imshow(I),title('yuanshitu');

bws = edge(I,'sobel',(graythresh(I)*.1));
figure,imshow(bws),title('tiqu');

se90 = strel('line',3,90);
se0 = strel('line',3,0);

bwsdil = imdilate(bws,[se90 se0]);
figure,imshow(bwsdil),title('dilated gradient mask');

bwdfill = imfill(bwsdil,'holes');
figure,imshow(bwdfill);
title('binary image with filled holes');

bwnobord = imclearborder(bwdfill,4);
figure,imshow(bwnobord),title('cleared border image ')

seD = strel('diamond',1);
bwfinal = imerode(bwnobord,seD);
bwfinal = imerode(bwfinal,seD);
figure,imshow(bwfinal),title('segmented image');

bwoutline = bwperim(bwfinal);
Seout = I;
Segout(bwoutline) = 255;
figure,imshow(Seout),title('outlined original image');

****************************************
利用图像分割结合形态学操作测试图像中的微小结构 P219
clear,close all
%I = imread('f:\s.png');
%figure,imshow(I);
I = imread('f:\sy.tif');
figure,imshow(I),title('original image');

ic = imcomplement(I);%求反色
bw = im2bw(ic,graythresh(ic));%阈值分割
figure,imshow(ic),title('complement of image');
figure,imshow(bw);
title('thresholding the image to show small structures');

se = strel('disk',6);
bwc = imclose(bw,se);
bwco = imopen(bwc,se);
figure,imshow(bwc),title('closing the thresholded image');
figure,imshow(bwco),
title('open the image to show large objects');

mask = bw & bwco;
figure,imshow(mask),title('the "and" of these two images');

*******************************
图像粒度测定 P221
clear,close all
%I = imread('f:\s.png');
%figure,imshow(I);
I = imread('f:\s.png');
figure,imshow(I),title('original image');

ic = imcomplement(I);%求反色
figure,imshow(ic);

%I = imread('snowflakes.png');
%figure,imshow(I);
claheI = adapthisteq(I,'NumTiles',[10 10]);
claheI = imadjust(claheI);
figure;
imshow(claheI);
gI = imadjust(im2double(I),[],[0 1]);
figure,imshow(gI),title('adjusted grayscale iamge');

se = strel('disk',10);
topI = imtophat(gI,se);
figure,imshow(topI),title('top-hat iamge');

bws = edge(I,'sobel',(graythresh(topI)));
figure,imshow(bws);
*************************************
%腐蚀膨胀联合操作 P236
clear,close all
%I = imread('f:\s.png');
%figure,imshow(I);
I = imread('f:\s.png');
figure,imshow(I),title('original image');

bw1 = imread('circbw.tif');
figure,imshow(bw1);
se = strel('rectangle',[40 30]);

bw2 = imopen(bw1,se);
figure,imshow(bw2);

bw3 = imerode(I,se);
figure,imshow(bw3);

bw4 = imdilate(I,se);
figure,
imshow(bw4);
*************************************
%骨架化 P237
clear,close all
%I = imread('f:\s.png');
%figure,imshow(I);

b1 = imread('circbw.tif');
b2 = bwmorph(b1,'skel',Inf);
figure,
imshow(b1);
figure,imshow(b2);

b3 = bwmorph(b1,'remove');
imview(b3)

BW = imread('circles.png');
figure,
imshow(BW);

*****************************
边界像素检测 P238
bw1 = imread('circbw.tif');
bw2 = bwperim(bw1);
imshow(bw1);
figure,imshow(bw2);

BW1 = imread('circbw.tif');
BW2 = bwperim(BW1,8);
imview(BW1)
imview(BW2)

BW2 = bwmorph(BW,'remove');
imview(BW2)

BW3 = bwmorph(BW,'skel',Inf);
imview(BW3)

********************************
基于分水岭的图像

clear,close all
%I = imread('f:\s.png');
%figure,imshow(I);

afm = imread('spine.tif');
figure;
imshow(afm);

se = strel('disk',15);
Itop = imtophat(afm,se);
Ibot = imbothat(afm,se);
figure,imshow(Itop,[]);
figure,imshow(Ibot,[]);

Ienhance = imsubtract(imadd(Itop,afm),Ibot);
figure,imshow(Ienhance),title('original of enhanced image');

Iec = imcomplement(Ienhance);
figure,imshow(Iec),title('complement of enhanced image');

Iemin = imextendedmin(Iec,22);
%Iompose = imimposemin(Iec,Iemin);
figure,imshow(Iemin);
%figure,imshow(Iimpose);

I = imread('rice.png');
figure, imshow(I), title('original image')
BW = im2bw(I, graythresh(I));
L = bwlabel(BW);
RGB = label2rgb(L);
RGB2 = label2rgb(L, 'spring', 'c', 'shuffle');
imview(RGB), imshow(RGB2)

rgb = label2rgb(Iemin);
figure,imshow(rgb);

***********************************************************************************************************

I = imread('pout.tif');
imshow(I);
figure,
imhist(I);
I2 = histeq(I);
figure,imshow(I2);
figure,imhist(I2);
imwrite(I2,'pout2.png');
imfinfo('pout2.png');
***************************************

clear,close all
I = imread('rice.png');
figure;
imshow(I)

se = strel('disk',15)
background = imopen(I,se);
figure,
imshow(background);
figure,surf(double(background(1:8:end,1:8:end))),zlim([0 255]);
set(gca,'ydir','reverse');

I2 = imsubtract(I,background);
figure,
imshow(I2)

I3 = imadjust(I2,stretchlim(I2),[0 1]);
figure,
imshow(I3)

level = graythresh(I3);
bw = im2bw(I3,level);
figure,
imshow(bw)
whos
**********************

clear,close all
I = imread('rice.png')
figure,imshow(I),colormap(I)
J= filter2([1 2;-1 -2],I)
imshow(I)
figure,
imshow(J,[])
*******************

clear,close all

moonfig = imread('moon.tif');
imshow(moonfig)
imview(moonfig)
imview close all
****************

clear,close all
I= imread('testpat1.png');
figure,
imshow(I);
J= filter2([1 2;-1 -2],I);
figure,
imshow(J,[]);

clear ,close all
%显示灰度图像
load clown
clims = [10 60];
imagesc(X,clims)
colormap(bone)
%colormap('default')
%colormap(gray)
%colormap(winter)
% See also hsv, gray, hot, bone, copper, pink, flag, colormap,

rgbplot.

clear ,close all

mri = uint8(zeros(128,128,1,27));
for frame= 1: 27
[mri(:,:,:,frame),map]= imread('mri.tif',frame);
end
imshow(mri(:,:,:,3),map);

*****************************************

clear ,close all

mri = uint8(zeros(128,128,1,27));
for frame= 1: 27
[mri(:,:,:,frame),map]= imread('mri.tif',frame);
end
figure;
imshow(mri(:,:,:,1),map);
figure;
montage(mri,map);
figure
mov = immovie(mri,map);
movie(mov);
******************************

clear ,close all

[X1,map1] = imread('forest.tif');
[X2,map2] = imread('trees.tif');
Imshow(X1,map1),
figure,
imshow(X2,map2);
figure;
subplot(1,2,1),imshow(X1,map1);
subplot(1,2,2),imshow(X2,map2)

figure;
subplot(1,2,1),subimage(X1,map1);
subplot(1,2,2),subimage(X2,map2);
figure;
[x,y,z]= cylinder;
I = imread('testpat1.png');
subplot(1,2,1),imshow(I);
subplot(1,2,2),
warp(x,y,z,I);
******************************************
clear ,close all
I = imread('rice.png');
J = filter2(fspecial('sobel'),I);
K = mat2gray(J);
imshow(I)
figure;
imshow(K);
figure,
imshow(J)
******************************
clear ,close all
imview close all
I = imread('snowflakes.png');
X = grayslice(I,16);
imview(I)
imview(X,jet(16))
**********************

clear ,close all
imview close all
RGB = imread('peppers.png');
[X,map] = rgb2ind(RGB,128);
figure;
imshow(X,map);
*************
clear ,close all
imview close all
load trees
bw = im2bw(X,map,0.4);
imshow(X,map);
figure,imshow(bw);

*************
clear ,close all
imview close all
load trees
bw = ind2gray(X,map);
imshow(X,map);
figure,imshow(bw);

*****************
clear ,close all
imview close all
I = imread('cameraman.tif');
BW = dither(I);
imview(BW)
***************
clear ,close all
imview close all

I = imread('rice.png');
J = imread('cameraman.tif');
K = imadd(I,J,'uint16');
subplot(1,3,1),imshow('rice.png');
subplot(1,3,2),imshow('cameraman.tif');
subplot(1,3,3),imshow(K,[])
*************
clear ,close all
imview close all

I = imread('rice.png');
J = imread('cameraman.tif');
K = imadd(I,J,'uint16');
subplot(1,3,1),imshow('rice.png');
subplot(1,3,2),imshow('cameraman.tif');
subplot(1,3,3),imshow(K,[])
figure;

I = imread('rice.png');
J = imadd(I,50);
subplot(1,2,1),imshow(I);
subplot(1,2,2),imshow(J)
******************

clear ,close all
I = imread('rice.png');
%blocks = blkproc(I,[32 32],'min(x(:))');
background = imresize(blkproc(I,[32 32],'min(x(:))'),[256

256],'bilinear');
bw = imsubtract(I,background);
figure,imshow(I);
figure;
imshow(bw,[]);

*******************

I = fitsread('solarspectra.fts');
I1 = mat2gray(I);%数据矩阵转化为灰度图像
bw = edge(I1);
figure,
imshow(I);
figure;
imshow(I1),figure,imshow(bw);

*************

clear ,close all
I = imread('rice.png');
figure;
imshow(I);
figure;
imcontour(I);
figure;
I = imread('circuit.tif');
imcontour(I,3)
*********
clear ,close all

I = imread('peppers.png');colormap
imshow(I);
j = imadjust(I,[0 1],[1 0],1.5);
figure;
subimage(j)

clear ,close all

[X,map] = imread('forest.tif');
I = ind2gray(X,map);
J = imadjust(I,[],[],0.5);
imshow(I)
figure,imshow(J);

clear ,close all

[X,map] = imread('forest.tif');
I = ind2gray(X,map);
J = imadjust(I,[0.15 0.5],[0 1],1.5);
%J = imadjust(I,[],[],0.5);
imshow(I)
figure,imshow(J,[]);

J1 = edge(J);
figure;imshow(J1);
************************
I = imread('cameraman.tif');
imshow(I)
figure,imhist(I)

clear ,close all

I = imread('cameraman.tif');
j = imadjust(I,[0 0.2],[0.5 1]);
imshow(j);
figure,imhist(64)
***************
clear ,close all

I = imread('tire.tif');
j = histeq(I);
subplot(1,2,1),imshow(I);
subplot(1,2,2),imshow(j)
figure
subplot(1,2,1);
imhist(I,64);
subplot(1,2,2);
imhist(j,64)
************************
clear ,close all

I = imread('tire.tif');
j = histeq(I);
subplot(1,2,1),imshow(I);
subplot(1,2,2),imshow(j)
figure
subplot(1,2,1);
imhist(I,64);
subplot(1,2,2);
imhist(j,64)
*******************
clear ,close all
%对比度自适应直方图均衡化
I = imread('pout.tif');
j= adapthisteq(I);
imshow(I)
figure,
imshow(j)
figure
imhist(j,64)
******************
clear ,close all
%去相关拉伸
[I col] = imread('forest.tif');
S = decorrstretch(ind2rgb(I,col));
subplot(2,1,1), imshow(I,col)
subplot(2,1,2), imshow(S)
*****************
clear ,close all
%中值滤波
I = imread('eight.tif');
j = imnoise(I,'salt & pepper',0.02);
figure,imshow(I);
figure;
K = medfilt2(j);
subplot(1,2,1),imshow(j)
subplot(1,2,2),imshow(K)
************************
clear,close all
%边缘提取
I = imread('circuit.tif');
BW1 = edge(I,'prewitt');
BW2 = edge(I,'canny');
figure;
imshow(I)
figure;
imshow(BW1);
figure, imshow(BW2)
***************************
clear,close all
%边缘提取
I = imread('circuit.tif');
BW1 = edge(I,'prewitt');
BW2 = edge(I,'canny');
figure;
imshow(I)
figure;
imshow(BW1);
figure, imshow(BW2)
*********************
clear ,close all
%Otsu自动阈值分割方法

I=imread('coins.png');
figure,imshow(I);
level=graythresh(I); %确定灰度阈值
J=im2bw(I,level);
figure,imshow(J)

************************
clear,close all
bw = imread('text.png');
se = strel('line',11,90);%生成线形的结构元素
bw2 = imdilate(bw,se);%对图像进行膨胀
subplot(121), imshow(bw), title('原始图像')
subplot(122), imshow(bw2), title('膨胀后的图像')
**********************
clear,close all
I = imread('cameraman.tif');
se = strel('ball',5,5);%生成球形的结构元素
I2 = imdilate(I,se);%对图像进行膨胀
subplot(121), imshow(I), title('原始图像')
subplot(122), imshow(I2), title('膨胀后的图像')
*****************
clear,close all
I = imread('cameraman.tif');
se = strel('ball',5,5);%生成球形的结构元素
I2 = imdilate(I,se);%对图像进行膨胀
subplot(121), imshow(I), title('原始图像')
subplot(122), imshow(I2), title('膨胀后的图像')

clear,close all
I = imread('cameraman.tif');
I1=256-I;%对图像进行取反操作
se = strel('ball',5,5);%生成球形的结构元素
I2 = imdilate(I1,se);%对图像进行膨胀
I3=256-I2;%对膨胀后的图像取反
subplot(121), imshow(I), title('原始图像')
subplot(122), imshow(I3), title('膨胀后的图像')

clear,close all
originalBW = imread('circles.png');
se = strel('disk',11);%生成圆盘形的结构元素
erodedBW = imerode(originalBW,se);%对图像进行腐蚀
subplot(121), imshow(originalBW), title('原始图像');
subplot(122), imshow(erodedBW); title('腐蚀后的图像');

clear,close all
I = imread('cameraman.tif');
se = strel('ball',5,5);%生成球形的结构元素
I2 = imerode(I,se);%对图像进行腐蚀
subplot(121), imshow(I), title('原始图像')
subplot(122), imshow(I2), title('腐蚀后的图像')

clear,close all
BW1 = imread('circbw.tif');%读取图像
subplot(131), imshow(BW1); title('原始图像');
SE = strel('rectangle',[40 30]);%生成矩形结构元素
BW2 = imerode(BW1,SE);%对图像进行腐蚀
subplot(132), imshow(BW2); title('腐蚀后的图像');
BW3 = imdilate(BW2,SE);%对图像进行膨胀
subplot(133), imshow(BW3); title('先腐蚀后膨胀的图像');

clear,close all

BW1 = imread('circbw.tif'); %读取图像
subplot(121), imshow(BW1); title('????????');
SE = strel('rectangle',[40 30]); %生成矩形结构元素
BW2 = imopen(BW1,SE);%对图像直接进行开运算
subplot(122), imshow(BW2); title('开运算后的图像');

clear,close all
originalBW = imread('circles.png');
subplot(121), imshow(originalBW); title('原始的图像');
se = strel('disk',10);
closeBW = imclose(originalBW,se);
subplot(122), imshow(closeBW); title('关运算后的图像')

clear,close all
BW1 = imread('circbw.tif');
BW2 = bwmorph(BW1,'skel',Inf);
subplot(121), imshow(BW1) ; title('原始图像');
subplot(122), imshow(BW2) ; title('骨架提取后的图像')

***********************************

BW1 = imread('circbw.tif');
BW2 = bwperim(BW1,8);
subplot(121), imshow(BW1) ; title('原始图像');
subplot(122), imshow(BW2) ; title('边缘检测后的图像');

************************
I = imread('rice.png');
subplot(121), imshow(I) ; title('原始图像');
se = strel('disk',12); %生成圆盘形状的结构元素
J = imtophat(I,se); %使用top-hat对图像进行滤波
subplot(122), imshow(J) ; title('top-hat滤波后的图像');

**********************
clear,close all
I = imread('pout.tif');
subplot(121), imshow(I) ; title('原图像');
se = strel('disk',3);%生成圆形的结构元素
I1=imtophat(I,se);%top-hat滤波
I2=imadd(I,I1);%原图像加上top-hat滤波后的图像
I3=imbothat(I,se);%bottom-hat滤波
J = imsubtract(I2, I3);%再减去bottom-hat滤波后的图像
subplot(122), imshow(J) ; title('增强后的图像');
************************
BW = im2bw(imread('coins.png'));%读取图像
BW1 = imfill(BW,'holes');%对图像进行填充
subplot(121), imshow(BW), title('原始图像')
subplot(122), imshow(BW1); title('填充后的图像')
********************************
I = imread('glass.png');
BW = imextendedmin(I,50);%求取极小值的二值图像，阈值为50
subplot(121), imshow(I), subplot(122), imshow(BW)

*********************************
bw = zeros(5,5);
bw(2,2) = 1; bw(4,4) = 1; %设定像素值为1
[D,L] = bwdist(bw)%计算欧氏距离

*****************
bw = zeros(200,200);
bw(50,50) = 1; bw(50,150) = 1; bw(150,100) = 1; %设定像素值为1
D1 = bwdist(bw,'euclidean');%使用欧氏距离
D2 = bwdist(bw,'cityblock');%使用曼哈顿距离
D3 = bwdist(bw,'chessboard');%使用棋盘距离
D4 = bwdist(bw,'quasi-euclidean');%使用准欧氏距离
figure%依次显示四种距离矩阵并显示等值线
subplot(2,2,1), subimage(mat2gray(D1)), %欧氏距离
title('Euclidean')
hold on, imcontour(D1)%绘制等值线
subplot(2,2,2), subimage(mat2gray(D2)),%曼哈顿距离
title('City block')
hold on, imcontour(D2)%绘制等值线
subplot(2,2,3), subimage(mat2gray(D3)),%棋盘距离
title('Chessboard')
hold on, imcontour(D3)%绘制等值线
subplot(2,2,4), subimage(mat2gray(D4)),%准欧氏距离
title('Quasi-Euclidean')
hold on, imcontour(D4)%绘制等值线

****************

bw = zeros(50,50,50);
bw(25,25,25) = 1;%设定一个像素值为1
D1 = bwdist(bw,'euclidean');%欧氏距离
D2 = bwdist(bw,'cityblock');%曼哈顿距离
D3 = bwdist(bw,'chessboard');%棋盘距离
D4 = bwdist(bw,'quasi-euclidean');%准欧氏距离
figure; subplot(2,2,1), isosurface(D1,15), %显示欧氏距离
axis equal, view(3); camlight, lighting gouraud, %光照处理
title('Euclidean')
subplot(2,2,2), isosurface(D2,15), %显示曼哈顿距离
axis equal, view(3); camlight, lighting gouraud, %光照处理
title('City block')
subplot(2,2,3), isosurface(D3,15), %显示棋盘距离
axis equal, view(3); camlight, lighting gouraud, %光照处理
title('Chessboard')
subplot(2,2,4), isosurface(D4,15), %显示准欧氏距离
axis equal, view(3); camlight, lighting gouraud, %光照处理
title('Quasi-Euclidean')

**********
BW1 = imread('text.png');
c = [43 185 212];%规定了对象的列数
r = [38 68 181];%规定了对象的行数
BW2 = bwselect(BW1,c,r,4);%选择对象，连通性为4
subplot(121), imshow(BW1), subplot(122), imshow(BW2),%显示选择的对象

****************
clear,close all
BW1 = imread('text.png'); BW = imread('circbw.tif');
SE = ones(5);
BW2 = imdilate(BW,SE);%对图像膨胀
increase = (bwarea(BW2) - bwarea(BW))/bwarea(BW)%计算面积增加的比例
******
BW = imread('circles.png'); imshow(BW);
eul=bweuler(BW)%计算欧拉数
***********************************

f = @(x) sum(x(:)) >= 3;%创建查询表使用的函数
lut = makelut(f,3);%创建查询表
BW1 = imread('text.png');
BW2 = applylut(BW1,lut);%使用查询表操作
subplot(121), imshow(BW1); subplot(122), imshow(BW2);
**************************

上面代码有些稍微重复，有些是前面的代码进一步加工处理的结果，看时请自行鉴别其意义！

 

下面是提取遥感图像上我的家乡的几个村庄

具体图片见相册！

代码如下：（可以借鉴下，这些代码只对原图有效，换张图就完全不一样了）

clear,close all
imview close all
I = imread('c:\家乡.jpg');
bw1 = rgb2gray(I);
surf(double(bw1(1:8:end,1:8:end))),zlim([0 255]);title('背景色观察');
set(gca,'ydir','reverse');
se = strel('disk',25);
background = imopen(bw1,se);
bw2 = imsubtract(bw1,background);
se1=strel('disk',10)
bw3 = imtophat(bw2,se1);
bw4 = imadjust(bw3,stretchlim(bw3),[0 1]);
bw5 =medfilt2(bw4);
bw6 = imadjust(bw5,stretchlim(bw5),[0 1]);
bw7=im2bw(bw4,100/255);
%图像变换以及图像增强（灰度变换、滤波变换）
figure;
subplot(3,3,1),imshow(I),title('源图像');
subplot(3,3,2),imshow(bw1),title('灰度图像');
subplot(3,3,3),imshow(background),title('背景图像');
subplot(3,3,4),imshow(bw2),title('背景色消除后图像');
subplot(3,3,5),imshow(bw3),title('缨帽变换');
subplot(3,3,6),imshow(bw4),title('灰度变换1');
subplot(3,3,7),imshow(bw5),title('中值滤波');
subplot(3,3,8),imshow(bw6),title('灰度变换2');
subplot(3,3,9),imshow(bw7),title('灰度直方图二值化');
%图形学处理、特征提取
se2 = strel('disk',5);
e1 = imclose(bw7,se2);
se3 = strel('rectangle',[4 3]);
bw8 = imerode(e1,se3);
e2 = imopen(bw8,se2);
se4 = strel('rectangle',[3 4]);
bw9 = imerode(bw8,se4);
bw10 = imfill(bw9,'holes');
se5 = strel('diamond',10);
bw11 = imerode(bw10,se5);
se6 = strel('rectangle',[5 4]);
bw12 = imerode(bw11,se6);
se7 = strel('rectangle',[25 20]);
bw13 = imdilate(bw12,se7);
bw14 = bwperim(bw13);
bw15 = I;
bw15(bw14)= 255;
figure,
subplot(3,3,1),imshow(e1),title('闭运算');
subplot(3,3,2),imshow(bw8),title('腐蚀1');
subplot(3,3,3),imshow(e2),title('开运算');
subplot(3,3,4),imshow(bw9),title('腐蚀2');
subplot(3,3,5),imshow(bw10),title('填补空洞');
subplot(3,3,6),imshow(bw11),title('腐蚀3');
subplot(3,3,7),imshow(bw12),title('腐蚀4');
subplot(3,3,8),imshow(bw13),title('膨胀');
subplot(3,3,9),imshow(bw14),title('边缘提取');
figure,imshow(bw15);title('叠加合成');
imview(bw15);

 

 

 

 

 

 

 

 

 

 

最终结果略......

matlab终结 赞赞赞

灰度处理和图像增强 图像提取、形态学处理、图像分割 图像格式的转化 图像的输出   显示： hist(I); whos imfinfo('') imwrite(I,'XX.tif'); 纹理映射方式观察 [x,y,z]= cylinder;warp(x,y,z,I); 指定输出图像像素的大小：J = imresize(I,1.25);【tif】 图像的旋转：J = imrotate(I,90,'bilinear');【tif】 图像的裁剪：J = imcrop(I,[60 40 100 90]);【tif】 区域属性度量 stats = regionprops(bwlabel(I),'all');stats(23)【png】【也不用imshow】   灰度调整：增强（只改变灰度和颜色，不消除物体，不改变图像的原有物质） imshow(I,[]); histeq(I) adapthisteq(I) imsubtract(I,50)（图像相减） 灰度调节imadjust(I,stretchlim(I),[]);{提高对比度} imadjust(I,stretchlim(I),[0 1]); 直方图均衡自适应调整bw1 = imadjust(adapthisteq(I,'NumTiles',[10 10])); 图像的二值化level = graythresh(I),J = im2bw(I,level); J = im2bw(I,graythresh(I));【自动灰度阈值分割】 自定义阈值分割figure,imhist(I);J=im2bw(I,100/255);【通过灰度来进行二值化】 二值图像的调节色彩imshow(bw,[1 0 0;0 1 0]) 求反色imshow(~I) 求补色imshow(imcomplement(I)); 通过乘法来改变亮度和对比度（乘以数值改变亮度，乘以本身图像改变对比度） J= immultiply(I,0.2); I16 = uint16(bw8);bw9 = immultiply(I16,I16); 直接灰度变换J = imadjust(I,[0 1],[1 0],1.5);[PNG]【灰度倒置线性变换】 灰度的非线性变换J = 45*log(double(I)+1);figure,imshow(J,[]) 直方图灰度变换J = imadjust(I,[0.15 0.9],[0 1]);figure;imshow(J);figure,imhist(J,64)【区域映射】【增强减弱图像的对比度】 灰度范围的映射J = imadjust(I,[0 0.2],[0.5 1]);【此方法有利于消除背景区域，然后方便边界的提取】 索引图的灰度变换[X,map] = imread('forest.tif');I = ind2gray(X,map);J = imadjust(I,[],[],0.5); 改变图像以调节对比度 抖动 imshow(dither(I)); 滤波和灰度调整是为了消除图像中的一些东西 目的相同，原理不同 J= filter2([1 2;-1 -2],I) J = filter2([1 2 1;0 0 0;-1 -2 -1],I); J = filter2(fspecial('sobel'),I); I2 = imfilter(I,ones(5,5)/25); 中值滤波K = medfilt2(j);【把杂色过滤掉】   灰度调整与消除背景后的图像相加有异曲同工之妙，都能够进行亮更亮，暗更暗的效果。   处理方法，先进行背影的去除，然后进行（图像相加）增强（第一次增加对比度），然后进行第二次的效果增强。对比度调节，到了这个时候明暗已经很突出了，不过还有些微小的灰度变化，然进行一下彻底地图像二值化，进行完全的黑白分明，这样就能进行边缘提取，要是感觉还不行的话，那就进行形态学的膨胀或腐蚀的操作，消除微小物，然后在进行分割操作   图像的提取 1 背景色的处理 图像消除背景色，先提取背影图，然后再进行图像相减(背景色不一致的可以用这个方法) 背景图的提取（开运算） se = strel('disk',25); background = imopen(I,se); figure,imshow(background); blocks = blkproc(I,[32 32],'min(x(:))'); background = imresize(blocks,[256 256],'bilinear'); �ckground = imresize(blkproc(I,[32 32],'min(x(:))'),[256 256],'bilinear'); 减去背景色bw3 = imsubtract(I,background); 图像相减imsubtract(I,J); 图像的开运算imopen(I,se);se = strel('disk',25); 查看背景色是否正常的方法 figure,surf(double(I(1:8:end,1:8:end))),zlim([0 255]); set(gca,'ydir','reverse'); 对一副图像的各种处理中的图像可进行相减来消除一些东西 imsubtract(I,J); 两幅同样的图像可以进行背景去除后的相加操作是特征更加明显 imadd(I,J); 这点和直接加一个数值的区别是后者是整体相加，而前者是亮亮相加 经过灰度变换后使图像的灰度对比很大时就可以进行边缘提取了 边缘提取edge(I); BW2 = edge(I,'canny'); BW = edge(I,'sobel') BW = edge(I,'prewitt') BW = edge(I,'roberts') BW = edge(I,'log') bws = edge(I,'sobel',(graythresh(topI))); 先进行灰度划分然后进行边缘检测bws = edge(I,'sobel',(graythresh(I)*.1)); 图像块边缘检测J = nlfilter(I,[3 3],inline('max(x(:))'));【tif】【此方法进行边缘的扩充】【邻域检测】 等高线提取imcontour(I),imcontour(I,3);【过程提取比较慢】 形态学的膨胀imdilate(I,se); 【可通过膨胀填补空隙】 bw = imdilate(K,[strel('line',3,90) strel('line',3,0)]); bwsdil = imdilate(bws,[strel('line',3,90)]); 可参照帮助文档strel 和imdilate 空洞的填充J = imfill(I,'holes');J = imfill(I); 移除与边界连通的目标：J= imclearborder(I,4);|J= imclearborder(I); 腐蚀操作：se = strel('diamond',1);J = imerode(I,se);J = imerode(J,se);figure,imshow(J);【此方法可以对边缘进行平滑】 绘制轮廓线J = bwperim(I); 两幅图像叠加的显示segout = I;segout(bwoutline)= 255;figure,imshow(segout);【一般与绘制轮廓线同时使用，是图像分割的方法】 开运算：bwco = imopen(bwc,strel('disk',6)); se = strel('rectangle',[40 30]); bw2 = imopen(bw1,se);【此方法用来删除图像中较小的物体】 闭运算：bwc = imclose(bw,strel('disk',6));【抽取原始图像较大的物体】 图像“与”操作imshow（bw & bwco）【包含小对象和包含大对象的图像进行“与”操作，从而得到原始图像中所有的小对象】 骨架化b2 = bwmorph(b1,'skel',Inf);【骨架提取适用于道路的提取】 b3 = bwmorph(b1,'remove');【骨骼提取、边缘检测最好先将图像转化为二进制图像】 边缘像素测定bw2 = bwperim(bw1);|BW2 = bwperim(BW1,8); 缨帽变换（高帽滤波变换）：J = imtophat(gI,strel('disk',10));【消除背景中那些亮度不一致的背景】 低帽滤波变换Ibot = imbothat(afm,se);figure,imshow(Ibot,[]); 增大对象间的间隙【原图像加上高帽滤波变换的图像，然后再减去低帽滤波的图像】 Ienhance = imsubtract(imadd(Itop,afm),Ibot); 谷点图像的增强，照亮谷点图像 Iec = imcomplement(Ienhance);