Suspicious Activity Detection in Adverse Weather Conditions using YOLOv7

This research focuses on developing a robust and efficient system for suspicious activity detection in adverse weather conditions using the YOLOv7 model. The proposed system intends to annotate potentially hazardous actions within difficult environmental variables such as rain, fog, and crowded settings, including instances of gun use, fighting, and holding weapons like swords. We have designed the model to perform well in challenging weather circumstances, boosting its applicability in real-world scenarios, by utilizing the cutting-edge YOLOv7 architecture. We demonstrate the system's effectiveness in reliably detecting and localizing suspicious actions through thorough testing and evaluation, even in challenging weather and busy locations. The research advances public safety measures by developing a tool that can improve security monitoring and reaction under adverse weather situations