Obstacle detection algorithm for unmanned aerial vehicle inspection in substations based on fusion of vision and radar

When unmanned aerial vehicles are subjected to external interference, their ability to detect obstacles in substations can decrease. Therefore, this paper proposes an obstacle detection algorithm for unmanned aerial vehicle inspection in substations based on fusion of vision and radar. Using radar to scan cloud data of objects around drones, calculate the distance between drones and non-drone objects, and obtain complete drone operation data. Integrating visual technology with LiDAR for calibration, solving the rotation and translation matrices between the visual sensor and LiDAR, enhancing the accuracy of the detection algorithm and comparing the radial optical flow value with the detection threshold. Those greater than the optical flow value are inspection obstacles in the substation, thus achieving obstacle detection in substation drone inspection. The experimental results show that the proposed algorithm has good obstacle detection ability and can effectively reduce the distance error of obstacle detection.