The Research And Application Of Algorithm For Multi-point Pattern Matching

Point pattern matching is a hot problemand key technology in the field of computer vision,in face recognition,moving object detection,object recognition,automatic navigation and other application fields has widely application value.Point pattern matching is an NP difficult problem in computer science,researchers have proposed a variety of ways to solve the problem.Compared with the point pattern matching,the multipoint pattern matching problem has higher computational complexity.At present,the study of multi pattern matching problem is a little,and lack of effective solving method.Proposing a effective multipoint pattern matching algorithm will provide a new solution for image registration,structure comparison,target detection,etc.Therefore,in this paper,we study a multipoint pattern matching algorithm,according to 3d spatial point pattern matching problem,based on the positive shape context algorithm and maximum group algorithm,puts forward a new three dimensional multi pattern matching algorithm,the effectiveness of the algorithm is verified by experiment.This paper proposes an approximation algorithm for solving a new three dimensional multi pattern matching problem-cliques of shape context for multiple point pattern matching(CSMPPM).The model algorithm is put forward an improved positive shape context algorithm as a foundation for point pattern matching algorithm,and based on the point pattern matching results build undirected graph,solving all the biggest group in undirected graph,thus solving multipoint pattern matching problem.Innovation algorithm mainly reflected in three aspects:Firstly,the PSC arithmetic is perfecting that pro-processing the point set before the match and ruled out the interface points that may have a significant impact on the results.Next,in the evaluation criteria,the improved Q-value method is added,which considers the robustness and flexibility of the similarity evaluation.In terms of computational accuracy,this paper redefines the standard for matching that only correct matching called match,rather than find a match point for success.In order to solve the problem of matching multiple points,this paper introduces the maximum group algorithm,in order to achieve the matching of multiple points.In this paper,simulation data and real data are used to compare experiments,and compared with other classical algorithms.The experimental results show that the PSC algorithm proposed in this paper has better effect in dealing with the three-dimensional point pattern matching problem,and the result of the algorithm is better than other algorithms.The multi-point pattern matching algorithm proposed in this paper has good robustness.It can be used as a basic algorithm in multi-point pattern matching.If the algorithm and other algorithms combined,there may be a better match.If the algorithm used in the production of various areas of life will have better results.