一种基于改进Mean Shift的空间点目标聚合显示方法研究

Aggregation Display of Spatial Points Based on Improved Mean Shift Algorithm

目前,战场态势的目标显示仍然存在上下粒度一般粗,数据量大时目标相互压盖等问题。现有较为成熟的算法通常采用基于网格和距离的方法,并未考虑目标的种类或属性重要程度,这对战场态势中基于任务编组的目标聚合是不准确的。论文提出了一种基于权重的改进Mean Shift聚类算法,通过目标的某些关键属性构建目标权重序列,考虑了目标内和目标之间的多种因素。实验表明该算法较好地实现了空间点目标的聚合。

At present,there are still some problems in object display of battlefield situation,such as the same coarseness of high and low granularity and the coverage of objects when data amount is huge.Existing methods usually adopt grid and dis⁃tance-based methods without considering kinds and properties of objects,which is not accurate for task force-based object aggrega⁃tion in battlefield situation.By building object weight sequence based on some key attributes,a modified MeanShift clustering algo⁃rithm which considers the relations between objects is proposed to improve the effect of object aggregation.Experiment results dem⁃onstrate that the proposed algorithm achieves excellent result of aggregation display of spatial points.