Research On The Image Matching Method Based On Point Pattern Matching

Vision automatic alignment technique is one of the most important techniques in IC encapsulation. The vision automatic alignment system comprises optical lighting system, optical imaging system, CCD device, and image processing software etc, in which the high precision image processing algorithms are the determinant part for the final accuracy of alignment. This paper contains the intensive research on the subject of image matching methods, related image preprocessing methods, and methods to enhance the accuracy of image matching. Image matching is a fundamental problem in image processing, and it has lots of applications, such as industry inspection, traffic management, and remote sense survey. Simply put, its goal is to find out the spatial relationship between two different images. This paper first intensively studies the known methods of image matching. Then studies the point pattern matching to match images. Finally, the paper presents several ways to enhance the accuracy of image matching.The main content of the paper comprise: (1) classification, summary, and concise analysis of known image matching methods. (2) Study on image preprocessing methods. This section discusses in detail the advantages and defect of linear filter and median filter, and presents the detail of SUSAN filter. Moreover, this section also briefly introduces image segmentation. (3) Study on methods to extract feature points from images. This section briefly introduces simply edge detecting methods, the detail of wavelet based strong edge detecting method, and SUSAN corner point extracting method. (4) Study on point pattern matching. This section presents the detail of definition of the point pattern matching problem and its process. (5) Study on methods to enhance the accuracy of image matching. This section introduces subpixel edge detecting method, energy minimization method, and optical flow method. (6) Study on image matching software. This section includes the develop platform, structure, and user guide. At the end, a summary of the paper and the view of future image matching methods are given .