Enhancing Object Detection Capabilities of Autonomous Vehicles in Adverse Weather Conditions

Abstract

In this study, we detail a streamlined object detection framework intended for autonomous vehicles, primarily designed to initiate emergency braking. Our focus is predominantly on utilizing solely a digital camera as the sensory component under circumstances commonly affected by atmospheric conditions such as haze. The performance of optical sensors, digital cameras included, typically deteriorates due to such environmental factors. We evaluate various“ dehazing ”techniques pertinent to vehicular applications within this document and suggest an architecture that incorporates an effective dehazing algorithm to enhance safety measures. Experimental findings demonstrate the practicality of this architecture and its suitability for real-time operational needs.