Research On Algorithm Of PCB Defect Detection Based On Image Registration

In the modern electronic electronic and equipment products,printed circuit board PCB is the connected wires among the various electronic components in various fields obtained large scale application. Detection is one of the important links in printed circuit board in printed circuit board manufacturing process, it is the necessary means in quality assurance and the production cost controlled. As the PCB's output is more and more big, the wiring density and accuracy are increasingly high,traditional detecting way as artificial visual and needle type detection can't satisfy the modern PCB production needs. Many enterprises need to achieve on-line detection based on machine vision of PCB defects detection system.The paper researches defect detection algorithm of PCB based on the source of PCB's defect and its common features.With the application of digital image processing and pattern recognition technology, the paper completes the defect detection work of PCB. The work apply to the reference algorithm in PCB defect detection.First, the paper introduces the basic principle of image acquisition. In the process of image acquisition, the noise and blur of image can't be avoided and it will influence the quality of image, so we must process the image in advance. The pretreatment of image mainly include the smooth filtering, sharpen and contrast enhancement of image. The paper choose the Canny algorithm to extract the edge of image.The registration of the template image and tested image is the most important point in this paper. The paper completes the registration of images based on the FAST algorithm.The FAST algorithm detect feature points of image and the MSAC algorithm optimize the feature points matching. Then the paper correct the deviation of size and angle of the template image to complete the registration work. The advantage of FAST algorithm is its detection rate and it can improve the efficiency of PCB defect detection.After the image registration,we need to detect and distinguish image defect. The paper put forward a defect detection algorithm based on the similarity between the template image and tested image. The paper experiment the similarity and substraction of images to complete the image defect detection. The paper use MATLAB R2013a to program and research the PCB defect detection and algorithm.