Obstacle avoidance methods for rotor UAVs using RealSense camera

This work proposes an obstacle avoidance method based on depth camera for UAV navigation. Firstly, we use a RealSense depth-camera to obtain depth images of the UAV flight environment. Then the image processing board is applied to stratify and segment the depth image and to obtain the orientation and distance information of the obstacle, thus generating the corresponding obstacle avoidance instructions. Pixhawk on the UAV uses the serial port to communicate with the image processing board to receive instructions and control the UAV to avoid any obstacle. Finally, the UAV obstacle avoidance experiment in different environments is completed on the robot operating system (ROS). The results show that the method is high in efficiency, precision and reliability, and beneficial to airborne realization.