Research On Image-based Visual Servo System Of Robot Arm

Machine vision technology has always been of vital importance in the use and development of robotic arms.It enables the mechanical arm to have the ability to perceive the external environment,and can autonomously perform work movements without manual control.The traditional mechanical arm visual servo system often needs to rely on accurate camera calibration,but the working environment of the mechanical arm is complex and changeable,and accurate camera calibration is often difficult to achieve in the actual working process.Therefore,the robotic vision servo method without calibration technology has become a research hotspot.Traditional calibration technology takes a lot of time,and the accuracy of the system is affected by the accuracy of the calibration.Under external interference,it is often necessary to re-calibrate due to large system errors.The visual servo system without calibration technology can perfectly solve the above problems.According to the classification of visual servoing,this paper chooses to use the image-based visual servoing method as the foundation,deeply researches the uncalibrated visual servoing technology,and completes the corresponding software development.This article first introduces the research background and significance of visual servoing of mechanical arm,summarizes the research results of visual servoing methods,and analyzes the problems that appear in the practical application of visual servoing system.Afterwards,according to the classification of visual servoing,the two visual servoing methods based on position and image are introduced in detail,and the visual servoing method based on image is the research focus of this article.Then the kinematics basis of the robotic arm and the visual foundation of the robotic arm are explained,and the definition and calculation of the image Jacobian matrix are derived.In the image-based visual servo system,the real-time update process of the image Jacobian matrix has a crucial influence on the quality of the servo results.Therefore,this article combines the dynamic update process of the Kalman filter algorithm and designs a real-time update method of the image Jacobian matrix based on the Kalman filter according to the characteristics of the robotic arm servo system.The whole system is simplified from nonlinear to a linear time-varying system,and the image Jacobian matrix is used as the state of the system for real-time estimation.Subsequently,the image features and extraction methods are further introduced to ensure the accuracy and applicability of the image features.Then completed the design of the entire system,and completed the simulation analysis through the Matlab platform.Finally,the software development of each functional module of the system was completed on the Visual Studio 2010 platform.Combining with the existing robotic arm in the laboratory,an experiment was carried out to track the target by the robotic arm under actual conditions.Fully verify the feasibility of the image-based mechanical arm visual servo system studied in this paper.