Visual Servoing Based on Geometric Features by Surface Structured Light

摘要

Traditional visual servoing based on passive vision cannot deal with the non-textured object. Therefore, we propose a method that combines image-based visual servoing with surface structured light. The goal of our method is to realize positioning task of 3 DOFs robot in translational motion. In our approach, the phase map of a non-textured target object is got at first through structured light. Then, the edge is extracted from phase map and is utilized for the robot servo control. An error definition is adopted and the fuzzy PID controller is used for the positioning task. A non-textured white plane has then experimented and the result is satisfying.