基于Matlab的遥感影像BP神经网络分类算法

clear all; clc; %<----------------------------贝叶斯算法进行遥感图像分类------------------------ %读入样本图像1 Y=imread('date\yangqu.tif'); %读入待分类遥感图像 I=imread('date\wuhan.bmp'); %将彩色图像降维 R=I(:,:,1); G=I(:,:,2); B=I(:,:,3); %将灰度值归一化处理 R=im2double(R); G=im2double(G); B=im2double(B); [M,N]=size(R); P=[R;G;B]; a1=1;a2=1;a3=1; for i = 1 : M for j = 1 : N if (Y(i,j)==1) T1(:,a1) = I(i,j,:); a1=a1+1; elseif (Y(i,j)==2) T2(:,a2) = I(i,j,:); a2=a2+1; elseif (Y(i,j)==3) T3(:,a3) = I(i,j,:); a3=a3+1; end end end T1 = im2double(T1); T2 = im2double(T2); T3 = im2double(T3); P=[T1,T2,T3]; %样区数据 T=[]; %标记结果 TT=[0;0;0]; T=concur(TT,a1-1); TT=[0.4;0.7;0.7]; TT=concur(TT,a2-1); T=[T,TT]; TT=[0.6;0.7;0.4]; TT=concur(TT,a3-1); T=[T,TT]; %建立BP网络，中间层8，输出层3，tansig、purelin分别为中间层、输出层的转换函数， echo on net=newff(minmax(P),[8,3],{'tansig','purelin'},'trainlm'); clc %设置训练参数 net.trainparam.show=50; net.trainParam.epochs=1000; net.trainParam.goal=1e-4; net=init(net);%重新初始化 [net,tr]=train(net,P,T); A4=imread('date\wuhan.bmp'); A4=im2double(A4);%将各波段的灰度值类型转换为人工神经网络可用的类型 RB=A4(:,:,1);%提取出图像红色波段的灰度值 GB=A4(:,:,2);%提取出图像绿色波段的灰度值 BB=A4(:,:,3);%提取出图像蓝色波段的灰度值 %初始化三个矩阵 [m,n]=size(RB);%确定图像的行数和列数 RB1=reshape(RB,1,n*m); GB1=reshape(GB,1,m*n); BB1=reshape(BB,1,m*n); X=[RB1;GB1;BB1]; OUT=sim(net,X); IN=sim(net,X); clc [m,n]=size(RB); r1=OUT(1,:);%将数据还原成原始图像的rgb行列格式 r1=reshape(r1,m,n); g1=OUT(2,:); g1=reshape(g1,m,n); b1=OUT(3,:); b1=reshape(b1,m,n); r2=IN(1,:);%将数据还原成原始图像的rgb行列格式 r2=reshape(r2,m,n); g2=IN(2,:); g2=reshape(g2,m,n); b2=IN(3,:); b2=reshape(b2,m,n); R=[]; G=[]; B=[]; for i=1:m for j=1:n if r1(i,j)<0.9 R(i,j)=r1(i,j); G(i,j)=g1(i,j); B(i,j)=b1(i,j); else R(i,j)=r2(i,j); G(i,j)=g2(i,j); B(i,j)=b2(i,j); end end end Rr=abs(R)*255;%还原成图像格式 Gg=abs(G)*255; Bb=abs(B)*255; Rr=uint8(Rr); Gg=uint8(Gg); Bb=uint8(Bb); RGB=cat(3,Rr,Gg,Bb); clc imshow(RGB); title('05年分类图像');