Vision-Based System for Satellite On-Orbit Self-Servicing

The paper puts forward a new concept of the Satellite On-Orbit Self-Servicing using vision. In order to verify this concept, a prototype vision system has been presented by integrating vision techniques. Cooperative object recognition has been designed and implemented based on polygonal approximation. To realize real-time visual servoing control, part of recognition module has been optimized to increase process speed. Instead of time-consuming 3D pose estimation, a simple redundant feature based visual servoing method is presented. A new quantitive sensitive measure is used to select the feature points so as to minimize the joint error and increase the response speed. Finally, in a representative laboratory environment, a typical satellite servicing experiment is carried out based on shared control concept. The results of the experiment demonstrate the effectiveness of the system.