Visual servo technique of autonomous mobile robot in bended pipe

摘要

Presents a visual servoing control method considering the problem of movement of an autonomous mobile robot in a bended pipe. Considering the turning of the robot at bends, a vision based intelligent control system is presented for the coordinated control of three wheels of the robot. The key techniques of the pipeline mobile robot are how to recognize the existence of bended pipe online, how to recognize the orientation of the bend and the curvature of the bend. The paper discusses the image recognition problems based on a CCD vision sensor. For the recognition of existence of bended pipe, the paper studies the image processing problem and provides an image recognition algorithm based on the movement invariant quadrature. For the recognition of orientation of the bended pipe, the paper provides a method based on the characteristics rotating vector and a neural network is utilized to recognize the orientation of the bended pipe. Finally the paper discusses the problem of the curvature recognition of the bended pipe and shows that a genetic algorithm can be used to achieve the optimization.