Detecting Waterborne Debris with Sim2Real and Randomization

From palpable marine debris to microplastics, marine debris pollution has been a perennial problem. In recent years, there is an emergence of largescale clean-up efforts making its way around the world. Complimentary to large-scale clean-up efforts, there is a nascent area in the use of unmanned and remote vehicles for detecting and removing debris. In this project, our focus is on marine debris detection where we propose to train a waterborne debris detector based on a mixture of real and synthetic training samples. We leverage on the capabilities of game engines to generate a variety of distributions and types of marine debris with variations in camera angles and distance from objects. The use of the game engine allows us to get segmentation masks at no expense as the ground truth is known. We further augment the images via domain randomization and mix with the real dataset. All our datasets and model implementations will be publicly available.