Vision-Based Collision Avoidance and Path Planning for UAVs Using Bearing and Pixel Area

This paper presents an innovative collision avoid-ance and path planning framework for unmanned aerial vehicles (UAVs) using minimal camera-based inputs. The system leverages visual data to predict the future trajectories of nearby flying objects and compute low collision risk paths while maintaining progress toward designated targets. This solution extracts only two essential parameters from the vi-sual feed-bearing and pixel area-enabling practical obstacle detection and avoidance. Furthermore, our approach avoids the target observability problem without relying on extensive ownship maneuvers, allowing collision avoidance with minimal movement. Designed for UAVs operating in shared airspace with manned aircraft, the proposed framework emphasizes autonomous decision-making to improve operational safety. Simulation results demonstrate the system's capability to effec-tively plan avoidance maneuvers and generate feasible routes in complex and dynamic environments.