Unmanned aerial vehicles precision landing on a moving platform using image matrix segmentation method

Authors : Muhammad Farris Khyasudeen, Muhammad Badri Mhd Noor, Saaidal Razalli Azzuhri, Norlida Buniyamin

Pages : 1-12

View : 92 | Download : 114

Publication Date : 2024-06-30

Article Type : Research

Abstract :This paper presented the theory, planning, control and method for Unmanned Aerial Vehicles performing precision landings on a moving platform. Unmanned Aerial Vehicles performing flight mission often having issues of its retrieving due to the landing sequence inaccuracy which may lead to crash. Thus, in this paper, by using adaptive method and image processing, the H480 hexacopter, equipped with a gimbaled camera detecting the moving platform attached to a ground rover using a pattern recognition algorithm. Using AprilTag as the unique pattern, the H480 follows the moving platform and moves via pitch and roll instructions while constantly descending towards the ground. In this paper, the system proposed the degree of pitch and roll changes with regards to the position of the AprilTag i.e., the further the tag location detected from the camera center, the higher the degree of movement, such that the tag will be forced to be in the center of the camera frame. The system divides a camera frame into an 11x11 matrix in which each cell within the matrix suggests different pitch and roll degrees for the H480 movement. As a result, the system manages to assist the landing process for the H480 to reach the moving platform successfully with less than 0.5m offset from the center of the target.
Keywords : Computer vision, Image processing, Object detection, Robot navigation, Unmanned aerial vehicle