Research On The Defect Detection Method Of Metal Mobile Phone Backplane Based On Machine Vision

The detection of the surface defects of the metal mobile phone shell is an important part of the appearance detection in the mobile phone finished product quality detection.Because of the complex process in the production process of the metal mobile phone shell,it is inevitable to produce defects on the surface.In order to ensure the product quality,effective quality detection and screening must be carried out.The traditional appearance quality inspection is manual inspection,which has many problems such as low efficiency,high labor cost and different inspection standards.It is more and more difficult to meet the high standard requirements of modern industrial production for product quality.In recent years,with the application of machine vision to all aspects of automatic production,the technology continues to mature,and has the advantages of stable and reliable,high detection accuracy,fast detection speed,etc.,machine vision detection gradually replaces manual detection.In this paper,the metal mobile phone shell on the electronic manufacturing automation production line is taken as the research object.Aiming at the defects such as scratches,scratches,pits and sand spots that may appear on the surface of the shell,a set of effective detection and recognition methods are designed.The main research contents are as follows:First of all,the overall design of the metal mobile phone shell surface defect detection system is introduced.The design and implementation of mechanical transmission module,visual imaging system and electrical control system are described in detail.Then,the algorithm of ROI segmentation on the surface of metal mobile phone shell is studied.In the process of extracting ROI region of mobile phone shell,firstly,background segmentation of mobile phone shell is carried out,and the segmentation methods based on threshold,edge and region are studied respectively.Combined with the effect of current background segmentation methods,a background segmentation method based on threshold is realized.After the actual comparison,the segmentation effect of this method is better than that of other methods.Then,aiming at the segmentation of non edge reflective non hole logo area of mobile phone shell,an image rotation correction method based on offset angle and affine transformation and an area of interest segmentation method based on pixel gray projection are studied and implemented.Finally,the detection and recognition algorithm of metal mobile phone shell surface defects is studied.In the aspect of defect detection,the traditional defect texture feature extraction algorithm is analyzed and compared with experiments.Aiming at the phenomenon of false detection caused by the irregular texture interference on the surface of the metal mobile phone shell and the uneven illumination in the local area,a defect detection method combining homomorphic filtering and significance detection LC is proposed,which can effectively suppress the metal mobile phone shell table In addition,it can keep the edge information of the defect and improve the detection accuracy.In the aspect of defect recognition,a method of defect recognition based on SVM classifier is adopted,and the decision-making is made according to the shape and geometry features of different defects,so as to realize the recognition and classification of the surface defects of mobile phone shell.The detection and recognition method designed in this paper can effectively detect the possible defects in the main area of the mobile phone shell surface and identify the types of defects.The experimental results show that the average detection accuracy can reach more than 96%,which is higher than the traditional detection method of mobile phone shell surface defects,so it has a good practical value.