For our research project, we would like to perform collision detection and obstacle avoidance with Machine Vision. I would like to use deep learning or maybe even better extreme online learning techniques. I found out that extreme online learning is fast and might be better when not so much data is available to label.

Also I have done some research and find out that labeling data would be a time-consuming process for our application as we might need to do it manually. So please can someone suggest some platforms and tools where to start with obstacle avoidance /detection and deep learning/extreme online learning w to hit the ground? I would like to use monocular camera and point clouds and ROS for this project

Would appreciate your help Thanks