Two Fast Image Matching Algorithms Based On Gray Value

Image matching is an important technology of image information processing area, it plays an important role in many aspects such as remote technology, medical processing, computer vision, industrial area, and it is also the basis of some other image analysis technique such as image segmentation. Recently, a large number of image information need to be handled with image matching algorithms which have strong real-time performance. But traditional image matching algorithm has certain restrictions. Therefore, how to improve the real-time performance of image matching algorithm is a very significant study.This paper mainly studies the image matching algorithm based on the gray value. At first, it summaries and experiments several image matching algorithms which used more frequently. At the sencond time, it summaries some metric formulas on this basis, a rapid image matching method based projection is provided, reduceing the computation of the matching process with the recursive thinking. The method could improve the real-time performance of the matching process. Finally, the combination of pixel sampled idea could further improve the matching time and ensure the accuracy of matching process.For the rapid image matching method based projection, combinating the method of multiresolution, this paper presents a fast multi-resolution matching algorithm based on wavelet transform. At the bottom level, the algorithm uses the fast matching algorithm based projection; About the selection of searching strategy, it uses the multi-resolution, and it determine the scope of searching according to the corresponding relationships between image pixels after geometric transformation; At the top level, SSDA matching algorithm is used, it could ensure the accuracy of matching process. Experiments show that the algorithm has strong real-time performance.Finally, it summarizes the key work done in the paper and gives some outlook.