Research On Digital Image Mosaic Technology In Gis

At present, image mosaic has wide applications in the variety of fields. This paper presents a comprehensive analysis based on the current situation of the domestic and international research about the technology of image mosaic, and the problem about image mosaic in GIS needs further research. The major contents can be illustrated as follows.Firstly, an image sequence auto-sorting algorithm is proposed. The algorithm takes the range value of impulse function as the magnitude of the correlation between images, so as to make a group of partly overlapped images ordered. The algorithm avoids the artificial threshold setting, and achieves automatic sorting of the sequence of images truly. By the analysis of the corresponding theory, an image sequence auto-sorting algorithm based on phase correlation and its optimized algorithm are given.Secondly, the image enhancement methods are researched. Template operation and its decomposition algorithm are given in the course of image smoothing, and the examples were given to verify the feasibility of the decomposition algorithm. An image sharpening algorithm with Sobel operator mask method is given. The algorithm firstly detects the image gradient information by the traditional derivative method, and sets the threshold, then takes the largest part of the gradient magnitude image as the location of edges and contours. The original image is enhanced by high-frequency according to binary gradient information, so that the process of enhancing the image edges and contours do not generate noise distortion of the image.Thirdly, an image matching algorithm based on wavelet transform is proposed. The algorithm controls the noise through first wavelet decomposition, in order to enhance immunity and reduce the calculation and improve matching speed. And then in the image 1-T matching and image 2-T matching calculation process, the algorithm timely updates the threshold to improve the matching accuracy. An improved gradually fading out algorithm is proposed. The algorithm can eliminate splicing gap better, and then gain seamless image.Finally, experiments show the feasibility and effectiveness of these algorithms. And extensive experiments are conducted to testify these proposed algorithms and the experimental results are given and analyzed carefully.