Registration Methods Based On SIFT And Triangular Mesh

Image processing,including image mosaic,splicing,registration,etc,and image registration in the field of image processing occupies a very important position,image registration refers to the shooting time interval is larger,multiple sensor access,in different conditions obtaining two or more images after the process to registration or mosaic process,it has been widely used in motion estimation,angiography,remote sensing data analysis and other fields.With the continuous improvement'real-time and registration accuracy of image registration,researchers have proposed many registration algorithms,such as image registration based on image region grayscale,image registration based on feature and image registration based on hybrid model.In all image registration algorithms,SIFT algorithm is the most widely used image registration,the SIFT algorithm is the best robust algorithm in corner feature extraction,it extracts corner features belonging to local features that has a good scale,translation,illumination and affine invariance.Because of its good performance,image registration technology based on SIFT algorithm has become more and more people's research hotspot,this paper also focuses on improving SIFT algorithm,so that it can be more accurately image registrating.The article first introduces the basic technology and theory of image registration,analyzing systematically related approaches and pointing out shortcomings,then the article introduces the principle and implementation of the SIFT algorithm with the best image registration robustness and put forward the corresponding improvement measures,including:1)add”centering factor”in the extract edge response phase,so that the SIFT algorithm can extract the feature points of the relatively smooth region;2)the article also describes the delaunay triangulation algorithm removing the misalignment point of SIFT algorithm,and the Bowyer-Watson algorithm of the most suitable extracted image triangulation in incremental insertion algorithm was selected by analyzing time-consuming phenomenon,so under these details it can be combined with the SIFT algorithm perfectly.;3)this paper analyzes the causes of the triangular grid cross phenomenon,and uses the Top K algorithm to solve this problem well;4)in the Bowyer-Watson algorithm,the runway translation,the shortest path of the graph,and the optimization of the registration triangle description set are used to remove the misregistration points in the SIFT algorithm,and the detail of the hybrid model algorithm is described in detail to obtain the best registration effect.Whether it is experiment results or data analytics that the algorithm can extract a certain amount of the target object center and the smooth feature points,and the Bowyer-Watson algorithm occuping the entire algorithm time not to surpass 50% that has certain rationality,at the same time,the experimental results confirm the accuracy,fastness and stability of the registration algorithm of this article.