Integration of Vision and LIDAR for Navigation of Micro Aerial Vehicle

Abstract

In this work, Vision-based obstacle size estimation algorithm and distance estimation based on the LIDAR (Light Detection and Ranging) sensor for autonomous navigation of MAV (Micro Aerial Vehicle) were proposed. First, the LIDAR sensor installed on the MAV was used to measure the obstacle distance. When the threshold distance between the MAV and the obstacle is equal to 1.5m, then the obstacle size (width and height) can be measured using the object images acquired using the camera sensor based on the proposed vision-based object size measurement algorithm. The collision can be avoided with the obstacle using the LIDAR sensor which works on time on flight principle, in addition to that based on obstacle’s width and height with the tolerance of 0.01m, the MAV can change the flight route by either increase the altitude or roll/yaw. In addition, the proposed obstacle detection and collision avoidance algorithm implemented using the Raspberry Pi 3 flight controller can be used for real-time collision avoidance with obstacles.