實驗目的
1、掌握p參數分割的工作原理和算法實現
2、掌握均勻性度量法分割的工作原理和算法實現
實驗圖片
鏈接:https://pan.baidu.com/s/1gSpYLw9Xz5OK_hSqSGeUwQ
提取碼:o4au
實驗內容
實現P-參數法的圖像分割的代碼
測試代碼如下:
Im=imread('yw2_g.jpg');
[Im2]=pParam0(Im,0.7974);
imshow(Im2);
實驗代碼:
function Im2 = pParam0(im,perct)
bestDelta = inf;
BestThrd = 0;
[m,n] = size(im);
for Thrd = 0:255
ind1 = find(im<=Thrd);
ind2 = find(im>Thrd);
if(~isempty(ind1) && ~isempty(ind2))
p1 = length(ind1)/(m*n);
p2 = length(ind2)/(m*n);
Delta = abs(p2-perct);
if(Delta < bestDelta)
BestThrd = Thrd;
bestDelta = Delta;
end
end
end
Im2 = zeros(m,n);
Im2( find(im > BestThrd) ) =1;
Im2 = logical(Im2);
實驗結果:
實現均勻性度量法的圖像分割的代碼
測試代碼:
Im=imread('cameraman.tif');
[Im2,BestClThrd]=jyxdl(Im);
imshow(Im2);
實驗代碼:
function [Im2,BestClThrd] = jyxdl(Im)
BestCost = inf;
BestClThrd = 0;
[m,n] = size(Im);
for ClThrd = 0:255
ind1 = find(Im<=ClThrd);
ind2 = find(Im>ClThrd);
if(~isempty(ind1) && ~isempty(ind2))
mu1 = mean(Im(ind1));
mu2 = mean(Im(ind2));
sigma1_sq = sum((Im(ind1)-mu1).^2);
sigma2_sq = sum((Im(ind2)-mu2).^2);
p1 = length(ind1)/(m*n);
p2 = length(ind2)/(m*n);
Cost = p1*sigma1_sq + p2*sigma2_sq;
if( Cost < BestCost )
BestClThrd = ClThrd;
BestCost = Cost;
end
end
end
Im2 = zeros(m,n);
Im2 ( find(Im > BestClThrd )) =1;
Im2 = logical(Im2);
end
實驗結果:
學如逆水行舟,不進則退