Research On AOI Image Getting Path Planning And Solder Joint Locating

Long life and high reliability is the key to space equipment. At present, AOI(Automated Optical Inspection) technology has been popularly utilized in circuit-board quality inspection. AOI systems integrated performance of camera/lighting system design and vision algorithms to meet the quality control requirements. It is necessary to plan the path for AOI to shorten the time of image-getting and improve the speed of inspection.Path-planning is regarded as a combinatorial optimization problem which is the combination of clustering and sequencing problem. By researching the images-getting principle of AOI system, the sequencing problem can be regarded as TSP (Traveling Salesman Problem), and the optimization method for the TSP can be directly applied to this problem. But it is very difficult to directly apply the typical clustering algorithms to the problem because it is different from the typical clustering problem in the viewpoint of decision variables and optimization criteria. Because of the disadvantage of existing clustering algorithms, a heuristic algorithm is suggested to overcome the dependence on the initial distribution of cluster center.Methods which solve the two sub-problems hierarchically is prone to trapping in local optimum, a path-planning method based on HGA is suggested. HGA can allow the solution to get out of the local problems, solve the two sub-problems simultaneously and approach the global problem. This algorithm improves the method of coding, fitness function, operations of HGA, auto-adaptive parameter and a genetic simulated annealing algorithm is proposed. Simulated test results show the feasibility and efficiency of the proposed algorithms.When solder joint quality is inspected on a PCB image by image processing, it is necessary to design a method to locate the solder joint. A quick Hough transform is used to detect circular PCB mark based on discussing advantages and disadvantages of several Hough transforms. After position calibration the range of solder joint location is confirmed. Then, template matching is used to finish locating the solder joint accurately. In this paper, the author respectively presents two locating methods based on image mosaic and path-planning. On the base of path planning, the author finally implemented solder joint location and achieves the anticipated goal.