Research On Pick-and-Place Robotics Svstem Based On Machine Vision

In recent years, NC cutting machines are widely applied to the cutting of soft materials in cutting industry. However, the automatization of the entire cutting system is greatly limited by the manual sorting of the materials next to the cutting. Responding to the great sorting demand for the diversified garment design in cutting industry, this paper presents the research and modeling process of a pick-and-place robotics system based on machine vision, focusing on the Staubli RX90 robot. The system consists of machine vision subsystem and control subsystem, involving the robot, the manipulator, the computer, the CCD camera, the image acquisition card, the cutting platform, and the software. The system works in this way:the garment image is captured through the CCD camera from the cutting platform first, then the pretreatment, classification, feature extraction and target matching based on discharging system of the image data is done in turn, finally the motion of pick-and-place can be finished by the robot as the location information of the pattern has been transferred into the robot controller.In this paper, the positioning technology through machine version is studied in the first place. Next, according to the analysis of grasp demands for the garment of different types, four structure schemes of manipulators are developed. Then, a kinematic model built by D-H method is used to describe the motion relation of the RX90 robot, and a simulation based on MATLAB is done to obtain the motion trajectory of the robot and to validate the model Finally, the overall design scheme of the pick-and-place robotics system based on machine version is given, the working principles, working process, and hardware structure of the system are analyzed, platform is built to carry out several experiments on different types of garments to verify the pick-and-place function of the system.