The Algorithm Of Panoramic Image Stitching Line Fusion And Elimination Based On Improved IGG Model

Image mosaic is a hot research topic in the field of digital image processing and computer vision,which mainly combine the images of the same scene with overlapping regions so as to obtain a high quality panoramic image.The thesis discusseed theory and method of panoramic image mosaic based on SIFT feature and almost focused on improving the quality of stitching.In the process of stitching,in order to eliminate the influence of mismatching points on the transformation model,the parameters of the transformation model in the process of image mosaic were optimized,and the precision of image registration was improved.At the same time,this thesis used multi-resolution fusion method and adaptive optimal seam combination to eliminate ghosting and uneven illumination caused by the discontinuous transition phenomenon so as to obtain high quality seamless panoramic image stitching.The main research work includes:1.The accuracy of the transform model determines the quality of image mosaic.In order to eliminate the influence of mismatching points on the transformation model,this thesis used the IGG function with the ability of robust to optimize the parameters of the transformation model so as to eliminate the influence of parameter calculation of the matching matrix,improve the accuracy of the transformation matrix.2.This thesis studied the joint gap and discontinuity of the uneven exposure which could be effectively eliminated under the use of Multiresolution image fusion technology.The traditional multiresolution fusion algorithm is used to fuse the decomposed images directly.But it also has large computation and low real-time performance.In this thesis,the adaptive region is used for image fusion,which improved the fusion efficiency and eliminated the discontinuity of the transition,and achieved seamless splicing.3.The traditional image fusion algorithm can only be used to fuse the overlapping regions,which can ensured the smooth transition of the image.However,due to the large variation of the intensity of the image sequence,only the overlap region is fused,which would lead to a greater difference between the overlap region and other regions,Therefore,this thesis proposed a method to adjust the brightness of the whole image,which made the brightness of the whole image converge.4.This thesis designed and implemented a panoramic image mosaic system,which could extracted SIFT feature points of image,used KD-tree structure based on BBF algorithm for registration and used the improved LM algorithm to optimize the parameter model so as to get the results of a joint with a graph.In this thesis,the image fusion is performed by the combination of multi-resolution image fusion algorithm and stitching stitching method.Finally,we achieved the high quality panoramic image mosaic and eliminated the seam.At the end of this thesis,the content of the research was summarized,and the future research direction was pointed out.