Research on Robot Vision Servo Based on image big data

Abstract

With the integration and development of information technology and control theory, industrial robot has become an indispensable automation equipment in modern industrial production, improving product quality and work efficiency. However, there is still a need to improve robot technology in terms of intelligence and flexibility. Robot vision servo control technology uses image sensors to obtain the information of the external field of vision, which can improve its environmental perception and response ability to cope with different production work. The important control method of robot system is visual servo control. Because of its good performance, wide application and convenience, visual servo control technology has become one of the hot spots in the field of mechatronics. In this paper, the problem of visual feedback and noise processing in the vision system is described. The six degree of freedom pose correction of the three focus tensor is made use of the big image data, and some tensor elements are selected to define the visual feedback, which effectively avoids the display estimation of the camera pose. Based on the extracted tensor features, an adaptive controller is designed to drive the camera to the desired attitude. The visual servo control system improves the real-time performance of big data image processing and reduces the calibration error, achieving the expected purpose. Finally, the future research direction is prospected, which is expected to become an important support technology for robot application in dynamic and variable environment.