Environment Mapping Using Stereo Vision

This is an ongoing project. Presently we are aiming at generating a map of 3-D obstacles placed in front of the bot. The bot would be using two cameras laterally displaced in one particular axis and obtain two images one from each camera. Working on these two images the Z-depth map needs to be generated. Feature based stereo correspondence is being used. Once this is achieved efforts will be made to make this real time and the code would then run with the two cameras continuously streaming video. The Z-depth map would be generated and the bot would be moving from a certain location to a destination avoiding the obstacles.