基于Hough变换和聚类的舰艇编队队形识别算法  

Warship Formation Recognition Algorithm Based on Hough Transform and Clustering

作　　者：张翼飞[1] 董受全[1] 毕开波[1] 

机构地区：[1]海军大连舰艇学院导弹系,辽宁大连116018

出　　处：《兵工学报》2016年第4期648-655,共8页Acta Armamentarii

基　　金：海军装备部军内科研项目(2012年)

摘　　要：编队队形识别技术是反舰导弹武器系统目标识别领域中的一项重要研究内容,具有队形识别能力的反舰导弹可以有效增强对密集型舰艇编队当中重要目标的选择能力,进而直接提升导弹的命中概率和作战效能。基于Hough变换技术研究了一种舰艇编队队形识别算法,在无探测噪声影响时具有很好的识别率。当目标信息受污染较严重时,进一步采用了改进的K均值聚类算法对Hough变换后得到的积累矩阵局部峰值进行聚类处理,根据峰值聚类的结果准确提取出待识别队形的参数,从而有效抑制了探测噪声带来的不利影响。仿真结果表明,采用该算法可以正确识别出舰艇编队队形,在目标信息受污染较严重时也具有较好的识别效果,具有较好的鲁棒性。对该算法复杂度及目标指示误差对算法精度的影响进行了分析。Formation recognition is an important research task in the area of target recognition for antiship missile weapon systems. Perfect formation recognition capability can improve the target selection of anti-ship missiles for compact warship formation,thus enhancing the hit probability and operational effectiveness of anti-ship missiles. The formation recognition algorithm is researched base on Hough transform,which has higher recognition rate without the influence of detection noise. If the target information is polluted badly,the improved K-means clustering algorithm is used to cluster the local peaks in an accumulation matrix. The shape parameters of formation to be recognized can be extracted from the clustering results so that the adverse influence due to detection noise is restrained effectively. Even though the target information is polluted badly,the algorithm has better recognition accuracy and robustness. The complexity of the algorithm and the effect of target designation error on the accuracy of the algorithm are analyzed.The simulation results show that the proposed algorithm has the perfect capability of formation recognition.

关 键 词：飞行器控制、导航技术 队形识别 HOUGH变换 改进K均值聚类 峰值聚类 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]