基于方位特征序列的地雷鉴别算法

在使用低频超宽带合成孔径雷达(UWB-SAR)对地雷进行探测的过程中，根据目标电磁散射随方位角和入射角的变化特性，提出一种利用双峰间距和频率凹点特征沿方位向变化的隐马尔科夫模型(HMM)鉴别算法。该算法首先针对目标感兴趣区域(ROI)图像估计其各方位回波响应，然后利用时频原子提取时域双峰间距和频率凹点，进而得到随方位角变化的特征序列，再通过SAR工作时方位角和入射角的变化特点以及训练样本确定HMM参数，并在此基础上计算疑似目标新的特征矢量，采用马氏距离进行判别。实验结果表明了本文所提方法在目标鉴别方面的有效性。

A new landmine discrimination approach based on sequential aspect features

Low-frequency ultra-wideband synthetic aperture radar (UWB-SAR) is a promising technology for landmine detection. According to the scattering characteristics of body-of-revolution (BOR) targets along with azimuth angles and incident angles, a Hidden Markov model (HMM) discrimination algorithm is proposed which uses sequential features such as double-hump distance and notch frequency. First, the algorithm estimates the target scatterings in all azimuths based on regions of interest (ROI). Second, Sequential aspect features are extracted by sparse time-frequency representation. Then the HMM parameters are trained with labeled samples and the probability of occurrence is computed to discriminate suspicions targets. The experimental results indicate that the proposed algorithm is effective in BOR target discrimination.