空中飞机目标噪声稳健分类方法

Robust classification scheme for airplane targets

针对低信噪比条件下微多普勒调制易被噪声污染的问题,提出了一种基于复数域概率主成分分析(Complex Probabilistic Principal Component Analysis,CPPCA)模型的噪声稳健分类算法来实现低分辨雷达体制下三类飞机目标(喷气式飞机、螺旋桨飞机和直升机)的分类.算法依据三类飞机多普勒谱调制的差异,提出两维反映这种差异的微动特征.为了提高微动特征在低信噪比条件下的分类性能,利用CPPCA模型对雷达复回波信号建模并结合Akaike信息量准则(Akaike’s Information Criterion,AIC)来自适应地确定回波中主成分的个数从而实现对数据的噪声抑制.基于实测数据的实验结果表明,该算法在较低信噪比条件下能够获得较好的噪声抑制和分类性能.

A robust classification scheme to categorize airplane targets into three kinds, i. e. , turbojet aircraft, prop aircraft and helicopter, by using the conventional low-resolution radar system is proposed aiming to solve the problem that the micro-Doppler modulation is con- taminated easily by the noise component under the low singal to noise ratio（SNR） cases. Based on the different characteristics of the micro-Doppler modulation of the three kinds of airplane, this classification scheme firstly extracts two dimensional feature vectors to depict these differ- ences. In order to elevate the classification performance under the low SNR cases, we utilize the Complex Probabilistic Principal Component Analysis （CPPCA） to model the complex-val- ued echo from the airplane targets. The Akaike^s Information Criterion （AIC） is applied to the CPPCA model to determine the number of principal components adaptively for denoising the returned echo. The experimental results on measured data indicate that the proposed method can achieve the good noise reduction and classification performances under the test condition of relatively low SNR.