Research On The Detection And Positioning Algorithm For Loaded Circuit Board

With the advance of the semiconductors design and manufacturing technology, printed circuit boards become increasingly high density, multi-layers and high capability. PCB becomes so complex that manual inspection or In Circuit Testing is not any more feasible. As a consequence, the sampling inspection is getting more and more difficult to apply in the electronics mass-production manufacturing facilities, where achieving 100% quality is crucial, thus magnifying the importance of implementing a reliable, and cost effective automatic inspection process.The recent achievements in the computer technology, combined with the latest algorithms in image processing, pattern recognition, and artificial intelligence have resulted in better and cheaper equipment for industrial image analysis. This development allowed the electronic industry to use machine visual inspection to manufacturing processes that include printed circuit boards, IC chips, photo-masks, etc. A lot of applications for vision inspections in the electronic manufacturing have been developed in USA and Japan. But the research for such kind of applications is still in a primary stage in China. That's why most of the high capability inspection systems are being imported. However, their high price brings a lot of economical pressure to the manufacturers in China. In addition, quite often the imported systems are not completely adapted for use by the Chinese operators, so not allowing using these expensive systems in an efficient way.In order to meet the requirements of manufacturers, and improve the vision inspection applications in China, a variety of approaches for automated visual inspection of printed circuits boards have been studied in this thesis. The research concentrates mainly on the following three points:1. After a thorough study of the linear and nonlinear least square theories, it proposes an inspection algorithm based on least square theory for round solder joints. This algorithm is suitable for the shape inspection and classification of the round solders joints on the printed circuit board with DIP (Dual In-line Package) components.2. Based on the theory of morphology, and according to the concepts of the Hough-transform and connected space search, it proposes an algorithm to inspect the short and break lead. This algorithm is used for all kind of PCBs with fault leads.3. According to the theories of image matching and correlative functions, it presents an algorithm to inspect the assembly faults. This algorithm is used to diagnose the assembly faults, for example, missing, inverted and misplaced components.The feasibility and the robustness of the proposed algorithms have been proven in series of experiments and demonstrations described in the thesis.