混合因子模型在雷达目标识别中的应用

高分辨一维距离像是雷达自动目标识别的重要特征之一,它对目标姿态变化很敏感,只有通过进一步处理才能够实现有效的目标识别。针对距离像的这种姿态敏感性,本文提出了一种基于混合因子建模的雷达目标识别框架,它通过对从各个姿态角下获得的目标一维距离像出发构建目标的距离像概率生成模型,然后利用该模型通过比较条件概率大小的方法判别目标类属。对5类飞机数据的实验结果表明该框架对任意姿态角距离像的目标识别有很好性能。

Application of MFA model in automatic target recognition

In automatic target recognition of radar,high-resolution range profile(HRRP) is an important feature which is very sensitive to attitude change of the target.Only through further processing can it be used for effective recognition.In this paper,a framework for automatic target recognition is put forward based on mixtures of factor analyzer(MFA).By modeling the range profile of target,the shortcoming of the HRRP's aspect sensitivity is overcome.In the framework,we first establish the probability-generation model using the obtained HRRP under every pose,then determine its class attribute by comparison of the conditional probability of a test HRRP.The results of recognition experiment with five class of airplanes show that the framework has excellent performance of target recognition using HRRP under arbitrary pose.