迭代自组织数据法在数据关联中的研究与改进  

Research and Improvement of Iterative Self Organizing Data in Data Association

作　　者：杨林[1] 鄢茂林[1] 朱靖[1] 

机构地区：[1]海军装备研究院,北京102249

出　　处：《传感技术学报》2006年第2期541-544,共4页Chinese Journal of Sensors and Actuators

摘　　要：针对目标密集环境下易出现的误关联问题,提出了ISODATA算法的解决方法。该算法将目标点迹的特征参量构成待分类的样本空间,构建了基于ISODATA算法的模型,利用航迹的历史状态数据进行关联运算,有效地解决了多目标密集环境下的误关联问题,分析了记忆步长与计算量和关联结果的关系。通过仿真研究说明了算法在目标密集环境下的有效性。A method of data association based on ISODATA algorithm is put forward, in allusion to the problem of error association in multi-sensors multi-targets data association. Construct the Sample space with the parameters of track features, and construct limited memory mode based on ISODATA algorithm with historical track data. Solve the problem in the situation of multi-targets, and analyze the relationship between the calculation, accuracy and memory step. Testify the validity of the algorithm in such a situation though simulation,

关 键 词：数据关联 样本空间 多传感器 聚类分析 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]