MORE | Spring 2019

Detection and Tracking of Objects Outside the Line of Sight of a Camera

Non-line-of-sight (NLOS) imaging is an emerging field of research with applications for autonomous vehicle collision avoidance, search and rescue operations, endoscopy, and others. This problem is challenging because NLOS objects are outside of the line-of-sight (LOS) of both the camera and the illumination source(s). Traditionally NLOS reconstruction requires expensive, bulky and specialized detectors. This work overcomes such shortcomings by using a data-driven approach to handle the problem using only a traditional camera and a projector as the light source. Additionally, this work also involves identifying the optimal patches to shine light on for recovering NLOS information better.