I am doing a lane detect program for use in conjunction with ROS Kinetic for an IGVC robot. I have a python program in OpenCV3 that produces an IPM with detected lanes as white and black everything else. My thought process is transforming this "overhead view" into a laser scan (or PCL), then layer it with my Lidar's /scan to treat the lines as walls.
- Is this possible / Am I going about this in an overly difficult way?
- Is there a way to transform my IPM (black with white "obstacles") to a laser scan inside OpenCV that I can then grab with CV bridge?
I have a Realsense D415 so I could use RGBD data if that helpful.