# Extract all red pixels from image to binary image

33 views (last 30 days)
Philip on 7 Dec 2011
Does anybody know of a way to extract all 'red' values [255 0 0] from an image to make a binary image where 1's indicate the locations of red pixels?
I have done this with a loop, but I'm sure there is a much quicker way to do it...
Here's my loop, in case it clears up the objective:
for m = 1:size(img,1)
for n = 1:size(img,2)
if img(m,n,1) == 255 && img(m,n,2) == 0 && img(m,n,3) == 0
HaloRegion_Map(m,n) = 1;
end
end
end

Walter Roberson on 7 Dec 2011
HaloRegion_Map = img(:,:,1) == 255 & img(:,:,2) == 0 && img(:,:,3) == 0;

Sean de Wolski on 7 Dec 2011
X = uint8(rand(100,100,3)>.5)*255; %sample image with some red
red = all(bsxfun(@eq,X,reshape([255 0 0],[1 1 3])),3); %logical image of red parts
And, for fun, a speed comparison:
X = uint8(rand(500,500,3)>.5)*255;
img = X;
t1 = 0;
t2 = t1;
for ii = 1:100
tic
red = all(bsxfun(@eq,X,reshape([255 0 0],[1 1 3])),3);
t1 = t1+toc;
tic
HaloRegion_Map = (img(:,:,1) == 255) & ~img(:,:,2) & ~img(:,:,3);
t2 = t2+toc;
end
[t1 t2]
ans = 0.3921 0.6233