Research On Contral Method Of Loading Robot Based On Automatic Vision Guidance

With the vigorous development of many disciplines such as optics,electronics,control science,robotics,etc.,in the field of robot control,visual control is one of the current research hotspots.In the field of industrial robots,the main uses of vision are:product appearance inspection,such as bottle neck inspection in pharmaceutical production;parts identification,such as the collection and tracing of automotive parts information data in the automotive industry;guidance and positioning,such as The machine tool of the stamping workshop in the automotive industry uses machine vision to position when loading and unloading,which can guide the robot arm to accurately grasp;high-precision detection,in the occasion of high precision requirements,when the naked eye can not meet the requirements of the operation,you need to use visual equipment Perform work instead of the human eye.In this paper,the control process of the three-degree-of-freedom transfer manipulator based on visual feedback is taken as the research object,and the kinematics and dynamics of the manipulator are analyzed.Image acquisition is carried out by visual sensors,and image processing and image feature extraction and positioning are completed after image information is obtained.The main tasks are as follows:First,it discusses the research status of robotic visual control,introduces position-based visual servo control and image-based visual control;it introduces the basic theory of robotic arm kinematics and the basic structural characteristics of the reprinted robotic arm.Basically,MATLAB software was used to carry out forward kinematics and inverse kinematics simulations,and the kinematics model was verified.Secondly,the Lagrange equation was used to establish the dynamic equation of the manipulator,and the manipulator dynamic model was obtained.The PD control of the manipulator based on model compensation is designed,and the stability of the trajectory tracking control system of the manipulator is proved by the Lyapunov stability principle.The experimental results show that the PD controller based on model compensation can track the expected trajectory well.Then,the visual inspection system is studied.in the process of collecting and transmitting images through the camera,the image will contain noise points in the imagedue to the interference of various noise sources.In industrial applications,these noises will cause the image recognition results Error,therefore,eliminating these noises and highlighting the useful information in the image is helpful to obtain more accurate results of the recognition and positioning of cooperative target points,and the image information collected by the camera is processed for image enhancement,smoothing,edge detection and so on.Finally,the recognition of cooperative targets is studied,and the recognition method based on SIFT local features is used for recognition.In order to improve the positioning accuracy,after performing template matching,the image moment algorithm is used to calculate the exact position of the cooperative target.