針對非合作目標的自適應(yīng)網(wǎng)格聚類算法
發(fā)布時間:2018-09-08 10:04
【摘要】:武器系統(tǒng)的探測設(shè)備通常面對的是非合作目標,觀測樣本在特征空間中的分布形式難以預(yù)期,噪聲、不規(guī)則的類簇形狀以及差異化的類簇密度給聚類分析帶來極大挑戰(zhàn)。提出了一種自適應(yīng)的網(wǎng)格聚類算法,該算法包括基于k-近鄰方法的空間分辨率自適應(yīng)網(wǎng)格化處理方法,以及基于自適應(yīng)分水嶺變換的類簇結(jié)構(gòu)檢測與劃分方法。實現(xiàn)了對噪聲以及密度差異極大類簇的自適應(yīng)處理,同時保留了網(wǎng)格聚類方法對類簇形狀不敏感、不需要類個數(shù)作為先驗參數(shù)等優(yōu)點。通過雷達、電子偵察以及復(fù)雜人造數(shù)據(jù)集的仿真,證明了該算法的有效性。
[Abstract]:The detection equipment of weapon systems usually faces non-cooperative targets, and the distribution of observation samples in the feature space is difficult to predict. Noise, irregular cluster shape and differentiated cluster density pose great challenges to cluster analysis. An adaptive mesh clustering algorithm is proposed, which includes spatial resolution adaptive mesh processing method based on k- nearest neighbor method and cluster structure detection and partition method based on adaptive watershed transformation. The adaptive processing of noise and density difference cluster is realized, while the grid clustering method is not sensitive to the shape of cluster and does not need the number of classes as a priori parameter. The effectiveness of the algorithm is proved by radar, electronic reconnaissance and simulation of complex artificial data sets.
【作者單位】: 北京理工大學(xué)機電學(xué)院;北京遙感設(shè)備研究所;
【基金】:國防“973”計劃項目(613196)
【分類號】:TJ03;TP311.13
本文編號:2230202
[Abstract]:The detection equipment of weapon systems usually faces non-cooperative targets, and the distribution of observation samples in the feature space is difficult to predict. Noise, irregular cluster shape and differentiated cluster density pose great challenges to cluster analysis. An adaptive mesh clustering algorithm is proposed, which includes spatial resolution adaptive mesh processing method based on k- nearest neighbor method and cluster structure detection and partition method based on adaptive watershed transformation. The adaptive processing of noise and density difference cluster is realized, while the grid clustering method is not sensitive to the shape of cluster and does not need the number of classes as a priori parameter. The effectiveness of the algorithm is proved by radar, electronic reconnaissance and simulation of complex artificial data sets.
【作者單位】: 北京理工大學(xué)機電學(xué)院;北京遙感設(shè)備研究所;
【基金】:國防“973”計劃項目(613196)
【分類號】:TJ03;TP311.13
【相似文獻】
相關(guān)期刊論文 前2條
1 楊邦翔;自適應(yīng)網(wǎng)格法應(yīng)用于Emery問題的數(shù)值模擬[J];彈道學(xué)報;1995年04期
2 尤國釗;膛口流場數(shù)值計算的發(fā)展概況和展望[J];彈道學(xué)報;1993年02期
相關(guān)會議論文 前1條
1 張樹道;董師舜;;爆轟驅(qū)動的自適應(yīng)網(wǎng)格數(shù)值模擬[A];中國工程物理研究院科技年報(2002)[C];2002年
相關(guān)博士學(xué)位論文 前1條
1 王燕;非合作目標精確定位技術(shù)研究[D];哈爾濱工程大學(xué);2006年
相關(guān)碩士學(xué)位論文 前1條
1 李永振;一種在軌衛(wèi)星抓捕技術(shù)研究[D];中北大學(xué);2015年
,本文編號:2230202
本文鏈接:http://sikaile.net/kejilunwen/ruanjiangongchenglunwen/2230202.html
最近更新
教材專著