Making a pseudo Lidar with cameras and deep learning

Vision vs LiDARs. These times are interesting.

The self drivig would be solved via either LiDARs getting cheap, or Machine learning using vision to tackle tradeoffs of LiDARs.