图像处理第一次作业.docx

1.绘制出一幅2M×2N的灰度图像直方图：对该图像进行均衡化处理，绘制出均衡后图像的直方图：对该图像进行灰度变换，使变换后的图像大致如下图所示的归一化直方图，并绘制出变换后实际的直方图。 (1)程序 clear all; %直方图均衡化 A = imread(lena.jpg,jpg); B = uint8(A(:,:,1)*0.3+A(:,:,2)*0.59+A(:,:,3)*0.11);%B=rgb2gray(A); [m,n] = size(B);%测量图像尺寸参数,m*n=490*490 B1=zeros(1,256); for k=0:255 B1(k+1)=length(find(B==k)); end B2=zeros(1,256); for i=1:256 for j=1:i B2(i)=B2(i)+B1(j); end end B3=round(double(255*B2)/(m*n)); B4=B; for i=0:255 B4(find(B==i))=B3(i+1); end for k=0:255 B5(k+1)=length(find(B4==k)); end figure; subplot(2,2,1); bar(0:255,B1,b); title(原图像直方图); subplot(2,2,2); bar(0:255,B5,b); title(均衡后的直方图); subplot(2,2,3); B6=histeq(B); dif=B6-B4; imshow(dif); title(matlab函数均衡化减去自己??写函数均衡化); %直方图规定化 C1=zeros(1,256); x=0:255; for i=1:128 C1(i)=1/(127*128)*(128-i); C1(257-i)=C1(i); end figure; stem(x,C1,.); axis([0 255 0 max(C1(:))]); C2=zeros(1,256); for i=1:256 for j=1:i C2(i)=C2(i)+C1(j); end end C2=round(C2*255); D1=zeros(1,256); for i=1:256 for j=1:255 if B3(j)=C2(i)B3(j+1)=C2(i) if abs(B3(j)-C2(i))=abs(B3(j+1)-C2(i)) D1(i)=j; else D1(i)=j+1; end end end end D2=zeros(1,256); D2(1)=C1(1); for i=2:256 if D1(i-1)0 for k=(D1(i-1)+1):D1(i) D2(i)=D2(i)+B1(k); end else D2(i)=0; end end figure; bar(0:255,D2,b); title(规定化后的直方图) (2)实验结果 2.分别在两幅灰度图像中加入一定量的高斯噪声和椒盐噪声，噪声强度自定，然后采用3×3的均值滤波器和3×3的中值滤波器分别对噪声图像进行处理，给出两种处理方法的峰值信噪比(PSNR)。 (1)程序 clear all A=imread(lena.jpg,jpg); B=uint8(A(:,:,1)*0.3+A(:,:,2)*0.59+A(:,:,3)*0.11);%B=rgb2gray(A); B_gau=imnoise(B,gaussian,0,0.002); %添加高斯噪声 B_salt=imnoise(B,salt pepper,0.02); %添加椒盐噪声 fli_med_gau=medfilt2(B_gau);%中值滤波函数 fli_med_salt=medfilt2(B_salt); h1=fspecial(average,3);%均值滤波函数 fli_avg_gau=uint8(round(filter2(h1,B_gau))); h2=fspecial(average,3);%均值滤波函数 fli_avg_salt=uint8(round(filter2(h2,B_salt))); [m,n] = size(B); %测量图