Research On Image Mosaic For Quality Inspection Of PCB

Printed Circuit Board (PCB), a kind of information carrier which integrates varieties of electronic devices, has been applied widely in electronic fields nowadays. These years, a kind of vision inspection technology which based on computer and image processing is becoming very popular because of the PCB turning to have more layers and higher density.To make sure the precision of vision inspection system is good, the camera's visual range will be limited because of the high resolution. In fact, the order of magnitude about the AOI's visual range is cm at present, so it hard to get a whole image at one time. In order to get a large acreage and high definition image with low coat, an image need to be segmented to several parts and each of them need to be acquired, then these small images are mosaicked into a large image. After that the image could be use in size detection,pattern recognition and defect detection. Now most image mosaic methods are based on hardware that is to acquire and mosaic images by controlling the displacement and speed of the precision worktable. This way is very rapid, but the hardware device is precise, the control system is complex, and the cost is high. So it's significant to research a image mosaic method based on software to meet the high speed and precision PCB vision Inspection system.The main contributions of this paper are the followings:⑴Have analyzed the characteristics of the PCB images. PCB images are varied, but its texture have common characteristics; Apply Several typical mosaic methods to the PCB images, analyze their strengths and weaknesses; Decide which strategy will be used in this paper initially.⑵Propose an image mosaic method base on corner by analyzing the PCB's feature. First, we deduce that the minimal overlapping area in one image can act as the template; the maximum overlapping area in the other image can act as the searching area based on several hypotheses about image's overlapping area. Second, use image processing technique such as image preprocessing,image segmentation and edge detection to each area to get two object contour images. Third, extract corners on boundary base on curvature, and then match these two point-groups by hausdorff distance and genetic algorithms. Finally, evaluate the performance of mosaic; fuse images if the performance is high.⑶Build the image mosaic hardware system for PCB inspection. Make an account about basic compositions and work principle of the system. Expound how to design the important compositions, for examples, Lighting System and Image Acquisition System. Explain how to control table and camera's movements. ⑷Finish the PCB image mosaic program. This program is developed by LabVIEW which base on Graphics Language and IMAQ Vision module of National Instruments.⑸Finally, the method which proposed to this article has carried on theexperimental confirmation. The experimental results show that: the method is feasible and effective.