Research And Implementation Of Panoramic Image Mosaic Technology

Image mosaic technology,as one of the important foundations of computer vision and digital image processing technology,has become an important research content in the field of computer image.Image mosaic technology is to match and fuse several overlapping images according to the overlapping region,and finally to form a seamless image.This technology has been widely used in image analysis,scene virtualization and object survey.In this thesis,image mosaic technology is taken as the research background.Aiming at the problems of slow extraction speed of feature points,low matching accuracy and image duplication in image fusion process,image matching and image fusion algorithms with better performance are studied and designed to improve the efficiency and accuracy of image mosaic.The achievements of this thesis mainly include the following parts;(1)Aiming at the problem of inefficient feature extraction and description process and large matching error of feature points in image with more details,an improved speeded up robust features(ISURF)algorithm is designed.The algorithm improves the efficiency by optimizing the image and reducing the dimension of the image descriptor before feature point extraction and reduces the error of feature point matching by setting threshold and bidirectional matching.(2)In order to solve the problem of overlapping image in the case of large overlap range,this thesis proposes a regional weighted average method(RWAM)based on the weighted average algorithm.The algorithm solves the problem of duplication by adjusting the range of overlapping regions.(3)On the basis of the above research results,this thesis designs and implements the ISURF algorithm and the existing image registration algorithm on the VS2015 platform with the help of Opencv image library and C++language.The results show that the ISURF algorithm is better than the original one in terms of the proportion of feature points used and the running time.Some image registration algorithms run faster and have higher accuracy.At the same time,the proposed RWAM algorithm and weighted average algorithm are designed and implemented,and the fusion results are compared.The results show that the RWAM algorithm can better solve the problem of ghosting after fusion.Finally,according to the process of image mosaic,an ideal image mosaic system is designed with the method proposed in this thesis.