Section: New Results

Computer Vision in Bad Weather

Participants : Raoul de Charette, Aitor Gomez, Sule Kahraman.

Common assumption of any perception system is to consider the atmosphere transparent so that the light rays travel directly from a point in the scene to the camera. While this assumption is true in clear weather, in fog/rain/hail or snow conditions this assumption isn't valid and all perception system will struggle. This can have a dramatic impact in autonomous driving. Following some of his previous works in former labs, Raoul de Charette lead several works to investigate and quantify the influence of rain and fog on computer vision for autonomous driving. Two internships were conducted in that axis (Aitor Gomez, Sule Kahraman) and there are on going results and research to be output. More detail can be found in [41].