Research On Positioning And Grasping System Of Power Meter Based On Binocular Stereo Vision

To establish an intelligent power meter verification fully automatic flowing working platform,different types of power meter need to be transported from the area waiting for verification to the conveyor belt of the meter verification line.For this practical application background requirement,this paper studies the positioning and grasping system of power meter based on binocular stereo vision.The binocular stereo vision is used to recognition the three-dimensional coordinates and attitude of the meter in space,and control the industrial robot to grasp and place meter of arbitrary attitude.The specific research work in this paper is as follows:The rotary platform is designed and constructed as the place area of the meter waiting for verification.For the purpose of power meter rotating with rotary platform,this paper studies the method of vision-based moving target detection.Firstly,the Gaussian Mixture Model is used to detect the rotating meter,and the detected result is used to judge the position of the meter in the image coordinate system of vision system,when it reaches the specified position range,control turntable stop motion,finally the contour of the meter is acquired by the background difference method.For the type of meter waiting for verification,a vision-based meter model identification method is studied.By analyzing the image of the meter waiting for verification,it is determined that the meter model is identified by detecting its character mode.In view of this,the method of recognizing the meter model based on the combination of the Maximum Stable Extreme Region(MSER)algorithm and the multi-classification of image based on the Bag Of Word(BOW)model is studied.First,the image of the meter model text area is obtained by the MSER algorithm,and this image is used as a test image and input into a successfully trained image classifier.This classifier is based on the BOW model training constructed by image visual features,and finally obtains the model of the meter.The method of obtaining the three-dimensional coordinates of the center point of the meter by binocular stereo vision system is studied.Firstly,the binocular camera calibration and hand-eye calibration method in the hand-eye system of Eyein-Hand is studied.Then the left and right images are corrected by stereo calibration,and the camera coordinate system and the robot end coordinate system are converted.Through systematic analysis and comparison of the advantages and disadvantages of the stereo matching algorithm based on Semi-global Block Matching(SGBM)and Scale-invariant Feature Transform(SIFT)feature operators,a more stable stereo matching algorithm based on SIFT feature operator is determined to complete the stereo matching of the meter image,and the obtained disparity value is used to calculate the three-dimensional coordinates of the center point of the meter.A power meter positioning and grabbing system is be designed and implemented,it includes a binocular stereo vision system,an industrial robot,a rotary platform,and an electric gripper.A three-dimensional coordinate of the center point of the meter and its tilt angle are used to describe the attitude of the meter,and grasping experiment of the meter is completed.The experimental results show that the stability and accuracy of the positioning and grasping of this system can meet the demand of power meter grabbing,which has important application value.