Cone Sleeve Center Measurement Algorithm based on Multi-information Integration

Abstract

In order to meet the high-precision docking requirements of autonomous aerial refueling, this paper proposes a high-precision and fast positioning algorithm for the center of the cone sleeve based on two-dimensional and three-dimensional multi-information fusion. The algorithm performs target detection and tracking on the cone sleeve in the image after calibration and calibration, and adaptively selects the region of interest in the positioning area; the region of interest is calibrated based on the gray-scale constrained two-dimensional model and the three-dimensional model based on the depth gradient. The two-dimensional and three-dimensional information fusion result is back-projected and compensated to the region of interest to realize the high-precision center point position measurement of the center point. The algorithm is used in the semi-physical simulation experiment on the ABB mechanical arm to verify its feasibility, robustness, fast and high precision and other characteristics.