基于多特征融合对海导航雷达运动目标PDA算法  

Possibility Data Association Algorithm Based on Multi-feature Fusion of Sea Navigation Radar Target

作　　者：熊伟[1] 谭洁[1] 王聪[1] 

机构地区：[1]海军航空工程学院信息融合研究所,山东烟台264001

出　　处：《指挥信息系统与技术》2014年第6期11-14,51,共5页Command Information System and Technology

基　　金：山东省自然科学基金青年基金(ZR2012FQ004)资助项目

摘　　要：对海雷达视频图像背景复杂,综合利用运动目标的位置、面积和灰度均值等特征在空间和时间上的关联性,提出了一种基于多特征融合的对海雷达运动目标概率数据关联(PDA)算法。首先,描述了对海导航雷达视频图像空间运动目标模型;然后,分析了用运动目标位置、面积和灰度均值特征实现数据关联的可行性,计算特征关联度;最后,运用概率融合方法计算综合关联概率,完成目标状态更新。试验结果表明,该算法跟踪精度和稳定度高,跟踪误差小。Aimed at the complex background of sea navigation radar video image,apossibility data association（PDA）algorithm is proposed based on information fusion of multi-feature in infrared sea target tracking by using the association of location,the acreage and the grey scale mean value in space and time.Firstly,the algorithm can describe the space motion model for the sea navigation radar video image.Then,the feasibility of the data association of the location,the acreage and the grey scale mean value are analyzed.The association probabilities are calculated.Finally,the possibility of synthetic data association of all targets is calculated by using the probability fusion method for updating the target state.Experimental results show that the algorithm has advantages of high tracking accuracy and stability with few tracking errors.

关 键 词：雷达视频图像 目标跟踪 多特征融合 概率数据关联 对海雷达 

分 类 号：TN959[电子电信—信号与信息处理]