Microdrill Vision Detection Method And Device Based On Inner Cone Mirror Scanning Panoramic Imaging

With the continuous development of digital industry,PCB is becoming more and more intensive and complex.As the main processing tool of printed circuit board micro drill bit diameter(common diameter of 0.2-2.5mm)also with the printed circuit board constantly refined.The difficulty of pre-delivery inspection is gradually improved,and the efficiency of traditional optical machinery micro drill bit detection is out of step with the increasing demand.Due to the characteristics of micro drill bit shape,the complete side edge image acquisition is more difficult and the steps are complex.Only a few researchers use multicamera,multi-step method to extract the image information of micro drill bit side edge,and most researchers only discuss the method of detecting defects in the edge of micro drill bit.Aiming at the background of this kind of research,this thesis designs a set of inner cone mirror device,solves the problem of multi-camera,and also uses its ingenious design to overcome the micro depth limit of high magnification lens,and obtains the image of large area clear side edge,Greatly reduce the number of image Mosaic.The image obtained by the inner cone mirror is a ring image,and the image is gradually distorted as the diameter of the ring increases due to processing and other reasons.In this thesis,the clear region of the image is extracted by the definition evaluation method,the mathematical model of ring image mapping to rectangular image is established,and the annular image is expanded into rectangular image,and the rectangular image is stitched in the way of position mapping based on the method of side edge image location information and local feature fusion in the control system.Finally,the complete micro-drill panoramic side edge image is obtained.Two different methods for detecting micro-drilling defects are proposed for the characteristics of micro-drill end edge and side edge image.In detecting the end edge of the micro-drill,the end edge contour is detected by extracting the edge,and the corner detection determines the important characteristic information points.Then the angle point of the image to be inspected is matched with the standard sample,and the conformity is judged according to the similarity degree.When detecting the end edge,the helical incision feature is extracted by transverse differential,image equalization and expansion are enhanced,and the side edge is qualified by calculating the number of effective connected domains.In the Visual Studio 2015 development environment,based on OPENCV,the micro-drilling detection system is designed and developed,which can basically complete the detection of micro-drilling surface defects.