Study Of Image Mosaic Method Based On Ratio Template Matching

Image Mosaic technology is one important branch of digital image processing research and an significant research subject in virtual reality, computer vision, computer graphics and imagery processing research. With the rapid development of science and technology, image splicing technology has been widely applied fields such as digital video frequency, movement analysis, seabed survey, virtual reality technology, medicine image analysis, remote sensing imagery processing and the military and so on, playing a very important role in those fields. Therefore, the image splicing technology research has a booming application prospect and application value.Because imperfect devices and the different interference when the sensor incept the source image, in the process of image forming and transmission, the image we get frequently can't match our expectation, such as grayscale off-lying, geometric distortion , muzzy image and so on. So before the image fusion, we must pre-process the source image, to inhibit interference signals, and enhance useful signals. Image matching is a kind of technology to make sure the overlap between the images for matching. Which are got by different sensors, different time, and different angles. So that technology is the core for image matching, and it is directly related to the success rate and the speed of the image mosaic algorithm, which is also the focus of years of study. Image fusion technology is used to eliminate the dynamic scene either the illumination change causes, because of the geometry adjusts intensity or the color's discontinuousness question between the neighboring images. After the matching of the overlap region, two images are determined to one seamless image. But as a result of the camera condition influenced by the environment and hardware, the Image Mosaic usually has differences, such as translation, revolving, chromatic aberration and combination deformation, which bring more difficulty to Image Mosaic. The Image Mosaic technical requires the analysis to many images, so,it has the characteristic of multiplicity, focal property and complexity.The algorithm based on ratio template matching mainly relay on the color ratio of diplostichous pixel on the overlap of two contextual images, which is a relative relationship. It will constantly make mistake if the image mosaic only relay on such a relative relationship. In this article, the gradient factor is adopted, according to the row gradient maximum point, then we establish the curves to the two templates, and definite most similar part about the two curves, so we can definite the position for matches. This method may enhance the accuracy of match and reduce the amount of calculation greatly.Firstly, I studied image pre-processing technology, image matching technology and image fusion technology, carried on the classification of the existing technology and comparing, has analyzed good and bad points of different technologies; Secondly, thorough studying deeply the image algorithm based on the ratio template matches, proposed a new the Image Mosaic improvement algorithm on it.From the experimental results, through the image mosaic system studied in this article, basing on the ratio of template matching, we reality the seamless image,it proves the image mosaic system have favorable effect.