I want to perform stereo processing to estimiate the depth of the scene.
I have two different scenarios where I want to apply Sum of Squared Differences / Birchfield Tomasi for disparity computation.
- Two RGB images from the same camera with predefined baseline
- RGB image combined with gray level image from different camera.
So far, I have applied the following steps for preprocessing:
- histogram equalization
- bilateral filtering
I'm not sure, since the gain dynamic of the gray image camera is much stronger then the one of the rgb camera.
What can I do to adapt bias and gain of the two images for case 2.?
Are there any other filters that I should apply (for 1. and 2.) to improve disparity coputation?