The Defect Detection Method For Solder Joint Of The Integrated Circuit Board Based On Machine Vision

With the rapid development of electronic manufacturing industry, the application of integrated circuit board has been increasingly spread. How to detect the products’ quality efficiently has been a major issue for enterprise. It is an inevitable trend that automatic detection technology will instead of artificial detection. The inspection methods vary with different detection objects, therefore it is necessary to design corresponding system to different objects.The targeted board is the main component of signal receiver which is used to receive signal from automobile key. Since the chip pin joints will be got light blocked after the antenna and antenna bracket placed on the PCB, the current existing inspection systems are unable to be used for this kind of integrated circuit board. What’s more, the enterprise still use manual inspection. Aiming at the problem of low efficiency and lack of accuracy, this paper proposes a vision detection method for signal receiver of automobile.According to the special spatial structure of the integrated circuit board, an inspection system is designed to meet the original needs in this paper. And it has been verified through the experiments.The research includes four aspects.1. Summarized solder joint defects detection methods. In combination with special structure of the targeted board, it introduces a practical detection method based on machine vision.2. An automatic inspection platform with three freedom degrees based on GT-400 is designed that can inspect solder joints in different positions through the movement of the motion platform.3. The vision camera has to be placed at a certain angle to get images of chip pin solder joints, so some part of pins will get out of the depth of filed. As a result of this, the image is blurred. This paper proposes a reasonable method to control the capture angles to get clear images.4. According to the actual demands, the detection software is developed to achieve the goal that the motion control and image process, which based on VC++6.0 and MFC. It contains track learn module and on-line process module, etc.