disparity map creation opencv

I have 2 image one is left the other is right image. I want to implement disparity mapping. There are some questions in my mind.

Firstlty dispartiy image is 1 channel gray image but the left and right images are 3 channels bgr images. If we convert these 3 channels image to gray image, do i loss information? If yes, how can i get disparity image from 3 channels i开发者_Go百科mg to 1 channel disparity image. I mean how can demote to 3 channels image to 1 channels image to get disparity image?

Thanks for answers..

Yes, you are losing information by combining all three channels into a single channel grayscale image.

There are certainly algorithms that use 3 channels to compute a disparity map, but opencv does not contain functions that implement these algorithms.

Have you considered taking out each channel R, G and B and computing the correspondence for each seperately? Of course, you would need to figure out a clever way to combine the results into one final disparity map, but it could be fun!

Try first to find the corresponding pixels, then apply SAD difference on the two matrices, then find the minimum value and display it at the disparity map.