Research On Perforated Workpiece Recognition And Detection Methods Based On Machine Vision

With the rapid development of the industrial era from informationization to intelligence,The quality requirements of various workpieces are increased.The recognition and detection of workpieces is an essential part of the production process.In the workpiece detection process,traditional manual operations have many problems such as slow speed,large error,no direct data storage,and security risks.The workpiece detection technology based on machine vision can solve the problems effectively.The thesis studies perforated workpiece recognition and specification detection.The main contents of the research include the following aspects:(1)A Perforated Domain BRIEF(PDB)algorithm is proposed in this thesis,which includes: targeted preprocessing,using the connected region to determine the feature points of the perforated workpiece,and the feature points are described by a binary coding method which is based on the pixel grayscale difference.The design of the PDB algorithm combines the structural features of the perforated workpiece to reduce the storage space and improve the matching efficiency of the workpiece feature points.In this thesis,the comparison experiment between the PDB algorithm and several classical feature extraction algorithms is designed.The experimental results show that the PDB algorithm takes less time to identify the perforated workpiece and the matching rate is higher.(2)A Gaussian point Hough transform detection method for holes detection in the workpiece is proposed in this thesis.In this algorithm,a Gaussian transform temporary space is established,which is used to perform Gaussian transformation on each hole sub-image and use the Gaussian kernel function to extend the coordinates of each operation to its neighborhood pixel interval to determine the center coordinates.Then the minimum mean square error estimation is used to obtain the hole radius to determine the optimal hole parameters.This thesis completed the comparison experiment between Gaussian Hough transform algorithm and traditional Hough transform algorithm.The results show that the Gauss Hough transform algorithm has better average recall,precision and time-consuming.It demonstrates the accuracy and efficiency of the Gaussian transform algorithm for holes detection.(3)A system is designed for the identification and detection of perforated workpieces.The functions of the system include: creating a template library with the characteristics of the perforated workpiece and a template library of the specification,identifying the type of the workpiece,counting the number of holes in the workpiece,detecting the position and size of the holes,and completing the hole integrity and the conformity detection.The combination of creating a database and detecting is implemented in the system.It can analyze and process the performance and data of the workpiece in real time.