Self-Supervised Monocular Depth Estimation in Autonomous Driving

It is easy for humans to estimate depth in a scene, but what about machines? Typically, robots and self-driving cars use LiDAR sensors to gauge the depth of a scene. However, LiDAR is an expensive sensor that is beyond the reach of many personal vehicles. Robo-Taxis may be reasonable in business models that provide service across a city, but not for personal vehicles. As a result, some companies are using camera-only approaches to infer depth information from monocular images. I will discuss some of the state-of-the-art approaches for monocular…