基于AlexNet的自适应杂波智能抑制方法

An Adaptive Clutter Intelligent Suppression Method Based on AlexNet

由于受观测环境以及雷达参数等因素的影响,构造协方差矩阵的杂波样本数据并不满足独立同分布,导致传统的自适应杂波抑制方法性能降低。针对静止雷达平台,本文提出一种基于AlexNet的自适应杂波智能抑制方法,首先,通过分析海杂波幅度分布特性,建立样本数据库;然后,通过迁移AlexNet在ImageNet数据集上的分类模型,并使用杂波数据集微调网络参数,实现对海杂波数据的准确分类,从而获取服从独立同分布的杂波样本数据,提高自适应杂波抑制方法性能。相比现有杂波抑制方法,所提方法具有人工参与度低、杂波分类准确率高、以及杂波抑制效果更好等优点;最后,通过CSIR实测数据验证了本文所提方法的有效性。

Due to the influence of observation environment and radar platform parameters,the clutter sample data of constructing covariance matrix does not satisfy independent identical distribution,which results in the performance deterioration of the traditional adaptive clutter suppression method.In this paper,an adaptive clutter intelligent suppression method based on AlexNet is developed for stationary radar platforms.Firstly,the sample datasets are established by analyzing the amplitude characteristics of sea clutter.Secondly,by transferring the AlexNet classification model on the ImageNet dataset,and then using the clutter datasets to fine-tune the network parameters.In doing so,sufficient clutter sample data with the independent identical distribution are obtained based on the accurate classification,which improves the performance of the adaptive clutter suppression method.Compared with the existing clutter suppression methods,the proposed method has the advantages in artificial participation,clutter classification,and clutter suppression performances.Finally,the effectiveness of the proposed method is verified by the measured data of CSIR datasets.