An Image-Based Path Planning Algorithm Using a UAV Equipped with Stereo Vision

On November 12, 2025, a novel image-based path planning algorithm was introduced, leveraging stereo vision technology in UAVs to enhance navigation safety by accounting for terrain depth. Traditional two-dimensional images often fail to reveal critical features like craters and hills, which can jeopardize path safety. By generating a disparity map from captured images, the algorithm identifies candidate way-points using advanced computer vision techniques, including edge and corner detection. The initial and desired points are automatically detected through ArUco marker pose estimation. The algorithm's performance was rigorously compared with well-known methods such as A* and the Probabilistic Road Map algorithm (PRM) in both virtual simulations and a physical laboratory setup. Results indicate that the proposed algorithm significantly improves path planning in complex terrains, showcasing its effectiveness and potential applications in real-world UAV operations.