Pose estimation measured by the camera is increasingly used in helicopter landing. We present a pose estimation algorithm for autonomous landing. You Only Look Once (YOLO) algorithm is extended to estimate the 6-demensional relative position and attitude. We modify the structure by reducing the layer of network and adding a full connect layer to output the relative position and attitude. The traditional method such as Oriented FAST and Rotated BRIEF (ORB) and Perspective-n-Point (PNP) is also studied. Compared with them, modified YOLO method preforms better on the accuracy of pose estimation.
Pose Estimation Algorithm for Helicopter Landing Based on YOLO and PNP
Lect. Notes Electrical Eng.
2021-10-30
10 pages
Article/Chapter (Book)
Electronic Resource
English