PCB Inspection System Based On Visual Processing

Printed circuit board is the main core of modern electronics products, PCB vision inspection is an important content of electronic product testing. But with the improvement of density, complexity and number of the printed circuit board,the traditional methods of PCB defect inspection such as got-up inspection can't meet the requirement of modern industrial productions.Firstly the thesis introduces several methods of PCB inspection. After the comparison, the method that based on visual processing for PCB inspection is selected for practical use. And introduces a system of PCB inspection that based on visual processing, the system includes the hardware and software structure.According to the detailed descriptions of Automatic Optical Inspection (Automatic Optic I-AOI), this thesis analyzes the function of each module of the complete PCB visual inspection systems in detail.PCB image processing and analysis methods are required for visual processing mode PCB inspection. This paper analyzes the relevant algorithms of the image processing, such as image denoising and enhancement, edge recognition, morphological filtering, pattern matching, threshold setting and etc..In the defect inspection system of bared PCB, the image is segmented by binaryzation method and mixed method is used to identify the defects. In the defect inspection of PCBA, threshold setting and morphological filtering are selected to identify solder issue. Pattern matching is selected for the inspection of the components missing and polarity reverse.On the platform of Lab VIEW and IMAQ Vision, some programs are developed.In analysis of the defects of bared PCB and PCBA, some image processing methods are put forward. The experimental results show that the system is capable of inspecting some defects of bared PCB and PCBA, which provide the methods used in the system is succeed and could be used in the working.