YOLOV7-BASED MOVING OBJECT DETECTION IN DENSE FOG CONDITIONS

Detecting moving objects in dense foggy conditions is a challenging problem in computer vision, with critical applications in smart transportation systems, surveillance, and autonomous driving. Fog particles scatter and absorb light, significantly reducing visibility and making it difficult for traditional computer vision algorithms to accurately detect moving objects. To address this challenge, researchers have proposed learning-based approaches that leverage deep neural networks to recognize moving objects and adapt to the unique characteristics of foggy environments. In this study, we present a learning-based method utilizing the YOLOv7 framework to effectively detect moving objects in dense fog conditions. The proposed approach involves four key stages: feature extraction, feature fusion, object detection, and non-maximum suppression. The results achieved are highly promising when compared to state-of-the-art techniques.