RGB 颜色空间转 HSI 颜色空间的matlab程序实现

程序实现的时候注意把RGB像素值归一化就是了，然后千千万万要根据公式来，不能大意，我之前就有theta为复数的情况，是sqrt对负数进行运算进而得到了复数的结果(BUG)

程序是由matlab实现的，利用了矩阵运算的便捷性，这里如果直接操作单个元素的话。。。会很慢很慢，慢到能把我电脑卡死。。。如果改写成C可能程序的性能还是会有问题

好吧，介绍一下HSI和RGB的转化方法

%******************************************

% Code writer : EOF

% Code file : RGB2HSI.m

% Code date : 2014.10.15

%

% Code description:

% This code is a demo for how to transfrom pixel'value in

% RGB-space into HSI-space.

%

%******************************************

Img_Original = imread('/home/jasonleaster/Software/opencv-2.4.9/Project_in_XTU/PDF/mountain.png');

Height_Original = size(Img_Original,1);

Width_Original = size(Img_Original,2);

Channel_Orignal = size(Img_Original,3);

Theta = zeros(1,Height_Original * Width_Original);

S = zeros(1,Height_Original * Width_Original);

I = zeros(1,Height_Original * Width_Original);

R_temp = double(Img_Original(:,:,1));

G_temp = double(Img_Original(:,:,2));

B_temp = double(Img_Original(:,:,3));

R = R_temp./(R_temp + G_temp + B_temp);

G = G_temp./(R_temp + G_temp + B_temp);

B = B_temp./(R_temp + G_temp + B_temp);

for row = 1:Height_Original

for col = 1: Width_Original

% initialize min with 255(the max value in RGB space)

min = 255;

for channel = 1:Channel_Orignal

if min > Img_Original(row,col,channel)

min = Img_Original(row,col,channel);

end

end

Dark_channel(row,col) = double(min)/255;

end

end

Theta = acos(0.5*((R - G) + (R - B))./(sqrt((R -G).^2 + (R-B).*(G-B))));

S = 1 - (3./(R + G + B)).* double(Dark_channel);

I = 255*(R + G + B)./3;

S = 100*S;

for row = 1: Height_Original

for col = 1:Width_Original

if B(row,col) <= G(row,col)

H(row,col) = Theta(row,col);

else

H(row,col) = 2*pi - Theta(row,col);

end

end

end

H = H*180/pi;