| In this thesis, digital image processing methods are studied. And a intelligent detection system, which is based on the template reference comparison method and machine vision inspection technology, is designed for PCB ( Printed Circuit Board ).In the production process of PCB, there are many defects such as surface smudginess, solder mask unevenness, text excursion, pad damage, track damage etc. The same specs of PCBs should be inspected one by one in order to find and fix their defects. The development and application of PCB detection equipments is the key to solve this problem. The intelligent detection system based on the machine vision inspection technology is designed for this. This system is capable of finding out and identifying the defects of PCBs with capturing and analyzing their images. Meanwhile, in the interest of detectability and speed, great effort is made for image capture, image preprocesses, defect detection and identification. In the procedure of image capture, a equipment of self-adjusting luminous intensity is desired and a method for the image data format fast conversion is achieved so that the speed of image capture is improved, which is under the premise of guaranteeing the image quality. In the procedure of image preprocesses, some image preprocesses algorithms such as Gray Balance, Image Smooth, Median Filter are used for enhancing the SNR (Signal Noise Ratio) of the PCB image. In the procedure of defect detection and identification, a series of digital image processing algorithms are adopted by the way of combining the SSDA (Sequential Similarity Detection Algorithm ) Template Matching Algorithm with Image Subtraction Algorithm. This process achieve the fast alignment and reference comparison between the PCB image with the template image.The feasibility and validity of this system have been tested under laboratory environment. The experimental results show that this system is capable of detecting the micro defects with a relative high speed. |
PDF Full Text Request