数字图像处理(直方图)剖析课件.pptVIP

数字图像处理的基本概念第三课1

2.4像灰度直方2.4.1概念定灰度直方反映的是一幅像中各灰度像素出的率。以灰度坐，坐灰度的率，制率同灰度的系就是灰度直方。它是像的一个重要特征，反映了像灰度分布的情况。下是一幅像的灰度直方。率的算式2

3

2.4灰度直方(histogram)imhist(InMATLAB)直方图=imhist/prod(size(I0));4

彩色像直方A=imread(‘lena.tiff’);%入像数据N=size(A,1)*size(A,2);%得像素数h1=imhist(A(:,:,1))/N;%色分量直方h2=imhist(A(:,:,2))/N;%色分量直方h3=imhist(A(:,:,3))/N;%色分量直方holdon;xlim([0,255]);stem(0:255,h1,Marker,None,Color,r);stem(0:255,h2,Marker,None,Color,g);stem(0:255,h3,Marker,None,Color,b);text(40,0.016,直方,FontSize,22)5

彩色像直方A=imread(‘lena.tiff’);%入像数据N=size(A,1)*size(A,2);%得像素数h1=imhist(A(:,:,1))/N;%色分量直方h2=imhist(A(:,:,2))/N;%色分量直方h3=imhist(A(:,:,3))/N;%色分量直方holdon;xlim([0,255]);plot(0:255,h1,Marker,None,Color,r);plot(0:255,h2,Marker,None,Color,g);plot(0:255,h3,Marker,None,Color,b);legend(Red,Green,Blue,2);text(40,0.016,直方,FontSize,22)6

彩色像直方axes(‘Position’,[0.1,0.1,0.8,0.2]);%生成坐stem(0:255,h1,Marker,None,Color,r);set(gca,YColor,r,Xlim,[0,255]);axes(Position,[0.1,0.3,0.8,0.2]);stem(0:255,h2,Marker,None,Color,[0,0.6,0]);set(gca,YColor,[0,0.6,0],Ytick,[0.005,0.01],Xlim,[0,255]);axes(Position,[0.1,0.5,0.8,0.2]);stem(0:255,h3,Marker,None,Color,b);set(gca,YColor,b,Ytick,[0.01,0.02],Xlim,[0,255]);7

灰度直方的性1、直方图只反映像素值分布信息，不包括像素位置信息2、不同图像可能具有相同的直方图直方图相同8

2.4灰度直方3.一幅图像分成多个区域，多个区域的直方图之和即为原图像的直方图。Directplus9

①判断像量化是否恰当10

②确定像二化的11

具有二峰性的灰度图像12

③算直方中物体的面当物体部分的灰度值比其他部分的灰度值大时，可利用直方图统计图像中物体面积：其中v是灰度级为I的像素出现的频数。ii13

④算像信量H(Entropy,)熵：熵反映了图像信息丰富的程度，在图像编码中意义重大。14

直方Example-------I=imread(pout.tif);imhist(I);%算并制直方Seealsohisteq,hist.15

按列的直方histc(pascal(3),1:6)producesthearray[311;010;011;000;pascal(3)000;001]ans=每列目标数据的个数111123136统计16

其他型的?英文学名SectorGraph,有名PieGraph。常用于学模。2D形，手画，常用作。pie,pie3inMATLABpie([0.1,0.2,0.4,0.3])pie([0.1,0.2,0.4,0.3],[1,0,0,0])%explodeexplode17

ph=pie([2435],{North,South,East,West});holdon;z=0.6*exp(i*linspace(0,pi*2,100));fill(real(z),imag(z),0.8*[1,1,1]);18

其他型的?3D?Example：?pie3([0.1,0.2,0.4,0.3],[1,0,0,