Researches On AOI-Based PCB Defects Detection

The production process of printed circuit board (PCB) is complex which could produce defects easily. Compared to traditional manual visual inspection and common electrical measurement method, automatic optic inspection (AOI) has some advantages that it can increase the accuracy of inspection and avoid components mechanical damage and so on which has very important significance for PCB defects detection.Firstly, the analysis of AOI technology status at home and abroad is presented in the introduction. Secondly, the principle of PCB defects detection system is described and each part of the system’s hardware is analyzed. Finally, PCB image positioning algorithm and position defects algorithm are researched and the system software is also developed.According to PCB location, component’s coordinates can be obtained by PCB design data. So PCB location is the basis of position defects detection. An Optimized Generalized Hough Transform (OGHT) algorithm based on gradient direction vector and scale segmentation is applied which is used for PCB location. Compared to traditional Generalized Hough Transform (GHT) algorithm, OGHT is not only robust to noise and shape defects but also can reduce needed space, improve computation speed and algorithm’s performance by applying multi resolution. The experiment results also demonstrate the above improvements.In the detection of PCB position defects, the normalized cross correlation algorithm and phase matching algorithm have been presented and analyzed. Template matching approach is applied to get component’s position. The character recognition can be realized according to character gray feature extracting and the minimum distance classification meanwhile the component’s position defects can be detected by comparing the results of character recognition and PCB design data.In the end, software of the system is developed in the VC++6.0 environment and the pictures of experiment results are also presented.