| The measurement of workpiece size is very important in the production process.Effective measurement of workpiece size can improve the quality of workpiece.In the production process of workpiece,workpiece defect is an important factor that restricts the quality of workpiece.In recent years,with the development of computer technology,using computer vision to measure workpiece size and detect workpiece defects has become one of the research hotspots.At present,with the development of computer hardware technology,in the field of deep learning this year,machine vision ushered in a period of rapid development,and more and more application research based on machine vision has made significant breakthroughs.Based on machine vision technology,this paper achieves accurate measurement of workpiece size and target detection and recognition of workpiece defects,and achieves the desired results.The research focus of this topic is on the measurement of workpiece size and the detection of workpiece defects,and some image preprocessing operations are carried out for workpiece images.The following is the main research content.(1)It explain the machine vision theory in detail,introduce and analyze the online accurate measurement method using machine vision technology,and build a hardware experiment platform;(2)he image acquisition system is introduced.The camera calibration technology is studied.The calibration method of standard parts is adopted.(3)In this paper,the preprocessing method of workpiece image is studied,which mainly analyses the image denoising and enhancement,and finally achieves the effective segmentation of workpiece image.(4)After image preprocessing,the dimensions of polygonal and circular workpieces were measured by Harris corner detection and least square method.(5)On the basis of workpiece dimension measurement,workpiece defects are effectively detected and identified by traditional methods and depth learning methods respectively.In the research of deep learning algorithm,it mainly involves R-CNN algorithm and its derivative Faster-RCNN algorithm,and has achieved good results.6.Based on the above research,the demo prototype system of workpiece dimension measurement and defect detection and recognition is built. |
