Visual Perception Challenges in Adverse Weather for Autonomous Vehicles: A Review of Rain and Fog Impacts
Yongsheng Qiu, Yuanyao Lu, Yuantao Wang, Chaochao Yang
Abstract
Autonomous driving systems (ADS) offer efficient solutions for transportation, but their visual perception capabilities are significantly affected under adverse weather conditions such as rain and fog, posing a threat to vehicle safety. Rain droplets and fog can interfere with sensors, leading to inaccurate information acquisition and increased driving risks. Therefore, enhancing the environmental perception of ADS in adverse weather conditions is a critical research focus. This paper systematically evaluates the impacts and challenges of adverse weather conditions, specifically rain and fog, on ADS environmental perception sensors. It explores the imaging characteristics of rain and fog and their effects on visual perception, summarizing dehazing and deraining algorithms. Ad-ditionally’ this paper reviews the development status of advanced visual tasks such as object detection in autonomous driving under adverse weather conditions. The aim of this paper is to provide a comprehensive understanding of the reliability of ADS in adverse weather conditions, laying a foundation for future de-velopments and practical applications in autonomous driving technology.