基于动态时间规整算法的车辆目标分类研究

将动态时间规整(Dynamic Time Warping)算法应用于地面车辆目标的分类识别中。基于微多普勒效应原理,建立了轮式车辆和履带式车辆雷达回波模型,对两种车辆目标微多普勒信号的差异性进行了分析,并结合实测数据,验证了理论分析的正确性。在杂波抑制及速度归一化处理的基础上,利用动态时间规整算法,将提取出的车辆目标的累积失真距离作为目标分类识别的依据,实现了轮式车辆和履带式车辆的自动分类。基于实测数据的实验结果表明,该方法在不同信噪比条件下都具有较好的分类性能。

Study on Classification of Ground Vehicles Based on Dynamic Time Warping

In this paper, dynamic time warping (DTW) is utilized to the classification and recognition of ground vehicles. Radar returned echo model of wheeled vehicles and tracked vehicles is established based on micro-Doppler effect. The distinctions between the micro-Doppler signals of these two kinds of vehicles are analyzed. In addition,the correctness of the theoretical analysis is verified by the measured data. On the basis of clutter suppression and velocity normalization, taking the parameters of cumulative distances as a characteristic, the classification of wheeled vehicles and tracked vehicles is achieved. Experiment results based on the measured data show the proposed methods simultaneously achieves good classification performance under different SNR conditions.