Adaptive Histogram Equalization with Fine-Tuned DNN-YOLOv9 Technique for Vehicle Detection in Adverse Weather

Autonomous vehicles struggle in adverse weather due to reduced visibility and impaired sensor performance. Cameras lose detail, lidars face backscatter, and radars experience interference, impacting perception and decision-making. This research proposes solution which combines image enhancement with advanced object detection to address these issues. Enhanced Super-Resolution GANs (ESRGAN) and Histogram Equalization improve image quality in low-visibility conditions. YOLOv9 ensures accurate and real-time detection. Using the DAWN dataset, the approach boosts detection reliability in adverse weather. Results indicate improved object detection accuracy, and resilience in AV navigation. This paper advances autonomous vehicle technology by mitigating weather-related limitations, supporting consistent and dependable operation across adverse and unpredictable conditions.