Study On Image Matching Technology

Image matching technology is an extremely important modern information processing technology, especially in the field of image processing. Image matching is to study how to choose some features and similar standards based on the reference image and real-time image to search for strategies for correlation computing and to determine the best space responding point for matching. Its main issues focus on the feature space, similarity measurement and searching strategies. The fatal step is to ascertain effective matching methods, meanwhile, the effect of matching needs high probability, small error, fast speed and good real-time.Image matching methods cover two categories: one is based on the gray, the other is feature. Based on the gray method, this paper mainly focuses on the study of several fast image matching computing on the basis of Wavelet Transform matching which helps to improve reliability and positioning accuracy and fulfilled the following steps in this paper.Firstly the thesis introduces the application and current development of image matching technology, analysizes some traditional image matching computing and find out the advantages and disadvantages of them. It has much more work to study image matching on the gray-based method than on the feature-based methods, however the correlation computing on the feature-based method are more accurate. But this method cannot adapt better to picture type than the gray-based method. For the rotating image matching, it would cause errors when the angle of rotation enlarged. The traditional matching computing cannot meet the need of the requirements of real-time.Then it explores fast and reliable image matching method according to the gray space, similarity measurement of the normalized cross-correlation coefficient which have got de-mean and search strategy based on the wavelet pyramid. In order to improve the speed of image matching this paper studied how to decompose the image by wavelet transform, match them first by coarse resolution then by careful resolution. This multi-resolution pyramid matching has greatly accelerated the speed of image matching. Discrete Hartley Transform (DHT) is used to compute cross-correlation and it is a real computation and reduces the amount of computation and memory compared with...