The Design And Research Of Automatic Optical Inspection System For Printed Circuit Board

This paper offers a PCB(Printed Circuit Board) automatic optical inspection system according to the basic theory of AGI(Automatic Optical Inspection). The system will automatically find the PCB defects by collecting images and comparing images with standard images.The buildup and the theory of this PCB optical inspection system is discussed in the following several parts. First, with the beginning of PCB inspection system development and mechanical vision used status, the principle and scheme of PCB optical inspection are proposed entirely. Secondly, the hardware composition and structure are expatiated. Thirdly, the type choice of CCD(Charge Coupled Device), the choice of lamp-house, the work platform are discussed. After that, the principles of image processing used to detect the deflects of PCB are expounded. Finally, the experiment result is given.In order to enhance the capability and inspection speed, a lot of research work was done in this paper in the field of image capture, pre-processing, feature extracting, image matching and defects detecting. In the procedure of image capture, a project method is used to extract the PCB image. The method using region growing in PCB printed circuit board image segmentation is presented. The experiments indicate the method extracts the characteristics well and removes the background noise effectively. In the feature matching, the traditional method using Hough transform to match the orientation circles is discussed and a simple matching method using characteristic image selected by operator is adopted in this thesis. At last, morphologic arithmetic is used to remove the background noise and highlight the defects.According to the result under the laboratory environment, this system is capable of detecting the defects with a relative high-speed.