Based On Image Processing PCB Solder Joint Defects Detection

PCB solder joint defects detection in modern production line is usually through the automatic optical inspection (AOI) system. But the development and production of military circuit board usually have characteristics such as small batch, kinds of varieties or even single production. These features limit the use of the AOI system in the military circuit defects detection. At the same time, the AOI equipments are usually with high cost and not suitable for large-scale fielding our military.According to the small production scale characteristics, this article used digital image processing and pattern recognition in the solder defects detection of military circuit board to realize the extraction and recognition of the solder defects. The main completed work are as follows:(1) The image preprocessing. Because the images of circuit board from CCD digital camera are always fuzzy, with low contrast and noise interference, so this article choosed the proper image preprocessing method to improve the image contrast, restrain the noise, and sharpen the image;(2) The image alignment. According to the results of the hough transform, this project completed the translation and rotation of the test image, for the matching of the test image and the standard image;(3) The image processing. This article completed the threshold segmentation and feature extraction of the solder joints images after pretreatment to be ready for the defects pattern recognition;(4) The defects pattern recognition. According to different types of solder defects this article choosed different kinds of defects pattern recognition to realize the extraction and model recognition of circuit solder defects automaticly.