Research And Implementation Of Key Technologies In Image Mosaic

Image mosaic (Imgae Mosaic) technology is to match a group of image, which have overlapping part, at space by re-fused to form a sample that contains the image sequence information of the wide viewing angle scenes, complete, high-definition images of the new technology. The technology is widely used in virtual reality technology, micro-image analysis, digital video, motion analysis, medical image processing and remote sensing image processing and other fields. Image mosaic is the hot spots of the computer vision.This thesis elaborates the supporting theory of image mosaic and its key technology, and study several major methods of image mosaic in-depth.First of all, this thesis studied interest points extraction methods and the advantages and disadvantages of algorithm for the demonstration, and studied the effects on splicing and solutions of images distortion. These lay the theoretical foundation for the Image mosaic method study.Image mosaic depends largely on the success of image registration.This thesis, starting with the existing methods of image mosaic, focused on the key technology -Image Registration. In this thesis, the registration algorithm used is to point the direction of interest in the characteristics and criteria for each combination of information, first of all in the interest of the image points and their point of view of information through wavelet multi-scale edge detection, re-use algorithm harris corner points of interest, guidelines for the use of mutual information of the matching point has been the need to have the support of image mosaic.The thesis applies simulation to discuss the advantages and disadvantages of different algorithms by Matlab. The analysis of experiment results shows that the method is used to improve the registration error small, high accuracy. Finally, the algorithm achieves seamless image stitching using delphi7 programming.