| With the development of electronic industry technology, the carrier of electronic information-PCB(Printed Circuit Board) develops toward the high-density, high-precision, miniaturization direction. Therefore, improving quality testing of the PCB production, is the basis and key of the improvement of electronic industry. Traditional manual testing methods existing poor reliability and low level of automation cannot satisfy the production requirements of the modern electronic products. The PCB detection achieves a great progress with the development of non-contact detection based on the machine vision.Firstly, this paper introduces the detective technology and the current situation of circuit boards in present. According to the working principle of the detection system, the hardware solutions for machine vision inspection system are designed. Secondly, The collected PCB image is enhanced selecting a gradation transformation, smoothed eliminating noise through median filter and segmented separating the foreground and background by Otsu method. Furthermore, by analyzing the basic characteristics of common defects, the detection and identification method achieving common defect detection and automatic classification based on regional characteristics is designed. Finally, the development of man-machine interface is implemented by using MATLAB software GUI design, and the prepared program can achieve not only image acquisition and image pre-processing, but also the detection and result analysis of the common pitfalls: PCB board short circuit, open circuit, burrs, defects and voids.A lot of experimental results suggest that the system above can meet the needs of the experimental platform mount technology accompanying with the advantages of precision in testing recognition, small error and high reliability.And the software interface designed in this system is user-friendly and easy in operation. |
PDF Full Text Request