Matching Algorithm Research Based On Feature Point And Its Application In Panorama Image Mosaic

Image matching technology is a fundamental and critical technology in many computer vision applications.It is designed to spatially align two or more images taken at different times,at different viewpoints,or by different sensors.Image matching has important applications in 3D reconstruction,change monitoring and video tracking,image fusion,medical image analysis and remote sensing image processing.It is the focus of computer vision research.Therefore,it is very meaningful to design a fast and robust matching algorithm that achieves the speed and accuracy of industrial requirements.First of all,aiming at the problem that the fault-tolerant rate of error feature matching is not high for traditional feature-based image matching algorithm.In this paper,a Gaussian Field Criterion(GFC)based on Gaussian field criterion is proposed,which can obtain high-precision feature matching results when the scene has complex non-rigid transformation.The algorithm builds the initial feature matching based on the existing scale-invariant feature transform algorithm,and then focuses on eliminating the unknown error match.Specifically,this paper transforms mismatched culling into image transformation regression fitting problem,and designs a Gaussian field criterion to achieve robust estimation under erroneous samples,and then achieves mismatched precision culling.The robust criterion can handle both linear and nonlinear image transformations.In the linear case,we use a general homography to model the transformation,while in the nonlinear case,the non-rigid functions located in a reproducing kernel Hilbert space are considered,and a regularization term is added to the objective function to ensure its well-posedness.Moreover,we apply a sparse approximation to the non-rigid transformation and reduce the computational complexity from cubic to linear.Extensive experiments on various natural and remote sensing images show the effectiveness of our approach,which is able to yield superior results compared to other state-of-the-art methods.Then,the spatial neighborhood relationship(representing the topological structures of an image scene)is generally well preserved between two feature points of an image pair.Several mismatch-removing methods that maintain the local neighborhood structures of potential true matches have been proposed.Defining local neighborhood structures is a crucial issue in the feature matching problem.In this study,we propose a robust and efficient topological structure measurement method called top K rank preservation(TopKRP)for mismatch removal from given putative point set matching correspondences.We transform data from the feature space to the ranking list space.Thus,the topological structure similarity of two feature points can be simply calculated by comparing their ranking lists,which are measured by the top K ranking distance.TopKRP is validated on 10 public image pairs with typical scenes.Experimental results demonstrate that the proposed approach outperforms several state-of-theart feature matching methods,especially when the number of mismatches is large.Finally,we applied the GFC algorithm to the panoramic image mosaic of the actual scene,and designed four sets of experiments in different scenes,including infrared images and visible images.The experimental results show a good image mosaic effect and prove the practicability of the algorithm.