Relative Pose measurement of a non-cooperative spacecraft based on feature information fusion

For a space robot to capture a non-cooperative spacecraft, a novel method for measurement of the relative pose of the spacecraft based on feature information fusion is proposed. The method chooses the outlines of the spacecraft itself and the engine nozzle as the recognition features. It uses the least square-elipse fitting algorithm and the polygon approximation method to extract and identify the features. Based on the disparity theory of stereo vision, the actual positions of feature points are computed. The geometry relationship between the plane of the engine nozzle and the surface of the spacecraft is utilized to calculate the position and attitude of the spacecraft to be captured. The experimental results indicate that the method possesses the characters of simple programming, efficienct computation, and high measurement accuracy, which satisfies the requirements of pose measurements in capturing non-cooperative targets.