| With the rapid development of electronic industry recently, as it's fundamental products, Print Wire Boards (PWB) also development very fast. The traditional inspecting methods for PWB depend on manual work a lot, which are very underdeveloped and face austere challenge. It is apparent that the test process is time consuming, of high cost, and the bad test result. Therefore, the study of a automatic inspection system instead of manual work is meaningful in today's PWB production.Aiming at the primary defects of PWB, an analysis on the requirement of PWB defect inspection was implemented and in this paper an image inspection method was studied. The principle of the image inspection method is: under illuming system, use the electrical coupling property of plane matrix CCD, make the surface status of PWB via optical lens an image on the surface of matrix CCD chip, video signal is outputted from CCD, transmit it to PC through image acquiring card.The resolution is a very important parameter of CCD. It is composed by optical lens resolution, CCD chip resolution and circuit frequency width. In image inspection, the main error is caused by CCD optical lens aberration. Using B -y graph, thiserror can be modified. The PCI and DMA technology was implemented in image acquiring card.Because of noise influence, the derived image should have median filtering. When getting the object in the image, image segmentation is required. In the thresholding methods, the moment-preserving thresholding is the better method.In the orientation of PWB and the circle inspection, the Hough transform was used. As the result, the center and the diameter of circle are obtained. Tree-shape class and fuzzy class method were used to class the PWB defects. Mathematical morphology dilation arithmetic is used to estimate the degree of nick defect.This paper has finished the theoretical investigation and part experiment investigation of using image PWB inspection, and has carried out system debugging. |
PDF Full Text Request