How to combine thresholded images

Hi,

I was following one of the tutorials on thresholding colour images and I am having trouble with one step. It is a colour jpg image and I am trying to remove background noise, large & small outlying spots and then count the number of spots.

Basically I separate the image into R , G and B components and then generate a threshold value and use the im2bw function. This step works fine, but when I try to combine them together to analyse the final binary image I am not getting the right thing.

The function I am using to combine the three is rgb=(r&g&b) . When I combine just r and g everything works fine and I see the red spots and the green spots come together. But when I include b in the function, it subtracts spots from the final image.

Is there something I am doing wrong, does this function not serve my purpose or is there something better I can do?

Thank you.

Answer 1

[Image Analyst]

Please post your image and code. rgb is a very bad name for an image that is not RGB - it's deceptive. Your rgb image is actually a binary (logical) image, NOT an RGB image. What you did was to create a binary image that is true/white/1 ONLY where it's true/white/1 in the r binary image AND the g binary image AND the b binary image. Perhaps you want OR, I don't know. I can't really say more until you post your image(s) and code.

Answer 2

[Meshooo]

Use cat

RGB = cat(3, R, G, B)

[msij]

When I use cat, it won't let me look at the image with imshow or use the bwlabel functions.

[Adam]

Use 'image' with the result of cat

[Image Analyst]

He has that image - that's just his original! If you did the same thing with lower case r,g,b you'd get an image with values around 0 and 1 and it would be so dark you could not see anything. So it's not a problem of whether he used imshow() or image(). And of course using bwlabel in that is just nonsense. So it's just a matter of doing the wrong thing(s) instead of the right thing(s). And you don't want to use cat() on the lower case r, g, b binary images.