Research on Online Inspection System for Non-Standard Parts Based on Contour Matching Algorithm
With the application of machine vision in the inspection of standard parts,the research on machine vision-based inspection of non-standard parts has received more and more attention.The vision inspection method that meets the requirements of on line inspection of non-standard parts is investigated.The vision system is used to collect images of non-standard parts.The contour of the workpiece and the contour of the corresponding computer aided design(CAD)model are extracted by the contour extraction algorithm.The graphical invariant moments of the contour of the workpiece and the contour of the corresponding CAD model are computed.The degree of approximation of the graphic invariant moments is compared and it is judged whether the workpiece complies with the styling requirements.In the validation test,19 pieces of stamping mould inserts are randomly selected for the inspection of the inspection method and cross-validated among the workpieces.The text results show that the inspection rate is 100%correct,and the average inspection time is 1.4 seconds per piece.Inspection based on CAD drawings,without collecting images of workpieces beforehand.Similarity matching based on graphical invariant moments has low requirements for the inspection environment and high configurability of hardware.Since only the contour information is matched,the detection efficiency is high,which can meet the diversified application requirements of online inspection of non-standard parts.
Non-standard parts inspectionMachine visionGraphical invariant momentsContour matchingShooting system
NETL
NSTL
万方数据
