<title>Abstract</title> Unmanned Aerial Vehicles (UAVs) are critical for applications like aerial surveillance, disaster response, and urban air mobility. Operating safely in these dynamic environments requires robust autonomous navigation and obstacle avoidance capabilities. While traditional methods depend on predefined rules and sensor-based heuristics, they often lack the real-time adaptability needed for complex scenarios. To address this limitation, we propose BS-CYOLOv5, a novel framework that integrates a …