Research On Dimension Measurement And Typical Surface Defect Detection Of Metal Mobile Phone Shell Based On Machine Vision

As an important part of mobile phone,the metal mobile phone shell will inevitably be affected by dimension defect and surface appearance defect in the process of production.Therefore,effective detection means must be used in the production process to ensure the quality of the products.Based on the current huge production volume of mobile phones,it is difficult to meet actual production requirements by adopting manual visual inspection and online sampling.Due to its high-efficiency,full-automatic,and intelligent features,Machine Vision technology is gradually used in the field of automated inspection.Therefore,the main task of this paper is to study the size measurement and typical surface defect detection of metal mobile phone shell based on machine vision technology,and providing an automatic vision detection scheme for metal mobile phone shell.Firstly,the imaging system of metal mobile phone shell visual inspection is studied.And the optical characteristics of the surface of metal mobile phone,the color characteristics of the light source and the influence of different lighting modes on the detection effect are analyzed.Then,the imaging model of CCD camera is studied,and on this basis,the principle and method of camera calibration are discussed.After that,the geometric transformation relationship between the visual inspection system and the actual physical coordinate system is established.Secondly,the key algorithm of mobile phone shell size measurement is studied,including the implementation of algorithms for edge extraction and size calculation of metal mobile phone shell.In particular,a subpixel edge detection method based on gradient direction and Sigmoid function fitting is proposed in this paper,which can realize sub-pixel edge detection with accuracy within 0.06 pixels.Thirdly,the key algorithms of surface defect detection for metal cell phone shell are studied,a method based on morphological filtering is applied to effectively suppress the surface texture and noise of the metal mobile phone shell.And then,an improved method of image enhancement of gamma gray scale transformation and a defect segmentation method based on double threshold of area and gray level are proposed,which can achieve effective segmentation of the defect area in matte texture background.By calculating and statistically validating the relevant defect features,a rule-based algorithm for classifying typical defects on the surface of a metal phone case is proposed.Finally,based on the research content and requirements of this paper,the hardware and software design of the metal mobile phone shell visual inspection system is completed,then the system is built and the relevant experimental research is carried out on this basis.Experimental results show that the size measurement accuracy of the metal mobile phone shell reaches 0.03 mm,and the average measurement time is 336ms/piece;the detection accuracy of the typical surface defects of the metal mobile phone shell reaches 95%,and the average detection time is 1.164s/piece,both of which can meet the actual production requirements.