The Research Of Target Highlight Modeling Based On The Planar Elements Clustering

The research on characteristics of underwater submarine target is an important issue in the development of mordern self-guided torpedo system.The accuracy of the underwater scale target model has a greater impact on the accuracy of the simulation results of the performance of the self-guided system.The component-based bright spot model based on geometric partition is a commonly used underwater scale target model.This modeling method mainly uses the equally spaced arrangement of the bright spots of the target's size and shape simple manifestation,but it fails to fully describe the target scattering characteristics and bright spot distribution in different situations.According to the experimental study on the target echo characteristics of the existing underwater scale,it can be seen that the acoustic scattering of the submarine target varies greatly under different situations and the spatial displacement of the bright spot caused by the submarine target is alse obvious.Based on the theories of scarttered sound field and planar element method,this paper analyzed the design ideas of many kinds of clustering analysis algorithms,proposeed a bright spot modeling method of submarine target,and builds a more accurate submarine target highlight model.Based on the acoustic scattering characteristics of the submarine target,the modeling method designed in this paper is to build a submarine target clustering bright spot model through the reference and improvement of a variety of clustering analysis algorithms.By extracting the bright spot of the field test data,the validation of the model has been completed.The main contents of the thesis include the following aspects:1)The basic thory of planar element method is studied.The planar element method is used to simulate the acoustic scattering characteristics of cylinder and Benchmark submarine model.The secondary splitting of sub-surface,in order to build a submarine target mesh model fo the modeling of bright spots,as the basis for the study of subject.2)To carry out analysis and research on many kinds of clustering analysis algorithms,extract the idea of clustering algorithm which is suitable for the construction of bright spot model,take the calculation result of plate meta-method as clustering research samle set,put forward the primary positioning of bright spot based on k-means clustering algorithm and local search algorithm based on FLP bright spot area selection algorithm to carry out the submarine target highlight model simulation modeling to detemine the number of bright spots in different thunderbolt situation the number of space position and relative intensity,highlights of the correlation between the bright spot of the combined cluster study to complete submarine target clustering highlights the construction of the model.3)Using the two-points detection and extraction method and the spatial location estimation method of the self-guided system in the present stage to process the field test data,the spatial distribution and intensity fluctuation of the bright spots are obtained,which is used as the verfication standard to carry out the verfication of the cluster bright spots model.At the same time,the geometric division model is added to the comparative study to analyze the comparision results.The results of comparison show that the clustering model is more accurate in describing the completeness of the scale target,the accuracy of the bright spot position and the fluctuation and estimation in the expermental data,which is more accurate and resonable than the geometric partition model.