Research On Workpieces Sorting Technology Based On Machine Vision

Machine vision is a hot and important issue of robotics.It is a combination of artificial intelligence and robotics.Workpieces sorting system is one of the most common application of machine vision in industry,but also an important index to measure the level of industrial automation.So the research on workpieces sorting technology based on machine vision has important industrial value and profound realistic significance.In this paper,workpieces sorting system is studied based on workpieces and industrial robot.The main work and research results are as follows:(1)According to the principle of calibration and experimental environment,this paper proposes a calibration method based on circular characteristics.Through the camera calibration,experiment can calculate the spatial relationship between workpiece coordinate system and the robot base coordinate system.(2)According to original images from camera,design image preprocessing flow of workpiece sorting system.Through four steps: the image enhancement,filtering,binarization and edge detection,separate workpieces from the background image,finally extract accurate workpiece edge information.(3)Recognition methods are based on the geometrical characteristics of the workpieces.Based on the principles and formulas,we research two recognition methods to extract the contour information of workpieces,including Hu moment and Zernike moment.This paper proposes a modified Zernike moment fast algorithm.At the same time,add SVM asclassifier of off-line training and on-line identification.(4)Set up an industrial robots workpieces sorting experimental platform based on machine vision.The system is divided into two modules including machine vision and control execution.Simulate the condition of industrial assembly line and complete the workpieces sorting experiment.The experiment result shows that the proposed method is feasible.