Research On Recognition And Picking Methods Based On Machine Vision Of Mechanical Parts

It is an important issue related to production efficiency and safety to determine whether the machine can work safely in industrial production.Whether the mechanical parts are clean is an important factor that affects the stable operation of mechanical equipment,so the cleaning of mechanical parts is one of the important manufacturing links before assembly manufacturing.After the cleaning,mechanical parts need to be sorted and stored.When a variety of mechanical parts need to be sorted,the robot cannot complete the sorting task well.It will increase the industrial cost because workers will take part in sorting work.In this paper,the automatic sorting of mechanical parts after cleaning in industrial production is studied,as are the identification and positioning of a variety of scattered mechanical parts.The main contents include:(1)Research on classification methods for mechanical parts based on deep learning.According to the requirements of industrial,the dataset of mechanical parts is created.According to the manufacturer requirement,the SPP structure in the traditional YOLOv5 network is improved,and a new SPC structure is proposed.The characteristic of the YOLOv5 network equipped with an SPC structure is that it does not contain a pooling layer.Compared with the traditional YOLOv5 network,the precision and recall rate are enhanced.(2)Research on the control method of a robot grasping mechanical parts An industrial camera was used to capture images for the computer,and the computer was used to calculate the picking coordinates of mechanical parts in the camera coordinate system.The transformation relationship between the picking coordinates of the mechanical parts in the camera coordinate system and the picking coordinates of the mechanical parts in the robot base coordinate system was obtained by camera calibration and hand-eye calibration.The picking coordinates of mechanical parts in the robot’s base coordinate system and the part type information are transmitted from the computer to the robot by the Modbus/TCP protocol.(3)Research on Communication between Computers and Industrial Robots According to the demand for industrial production,the program based on the Modbus/TCP protocol is designed to ensure that the computer can guide the robot to carry out industrial operations.(4)Designed the overall system,developed the software with good operability,and design the user interface.Let the camera,robot and computer combine to become the system.