低信噪比下机载多脉冲激光雷达姿态不敏感性特征提取研究（英文）

Research on attitude insensitive feature extraction algorithm of airborne pulsed laser radar target at low SNR

针对纳秒量级脉宽机载多脉冲激光照射目标的回波波形,提出几何分割比这一目标姿态不敏感特征,给出了特征提取算法,可用于与激光目标运动状态结合开展目标检测与跟踪。首先,基于激光目标波形数学模型分析目标波形与目标特征的关系,指出激光目标的几何分割比特征具备姿态不敏感特性。接着,结合目标特征提取对目标波形和噪声不失真分离的要求,提出在小波域利用对称小波基进行改进Donoho阈值降噪处理,在时域利用小波重构信号阈值获取目标波形序列,再由一阶二阶差分计算检测目标波形峰值点,进而提取激光目标散射中心及几何分割比等特征。最后,通过仿真实验验证算法在低信噪比下提取激光目标特征信息的有效性。在与机载多脉冲激光目标模拟器联试实验中,利用文中算法提取激光目标特征结合目标的运动状态信息,开展多帧相关匹配检测,多脉冲激光目标检测的可靠性明显提高。

Aiming at airborne multi-pulse laser radar target echo waveform irradiated by the nanosecond pulse width laser, the geometric section ratio feature is proposed, which is insensitive to target attitude and can be used for target detection and tracking combined with target motion state information. And then, the feature extraction algorithm is given. Firstly, by analyzing the relationship between target waveform and target features based on the laser target waveform model, it is pointed out that geometric section ratio of laser target is an attitude insensitive feature. Next, for the separation of target waveform from noises without distortion, the improved Donoho threshold de-noising algorithm is given in wavelet domain using discrete wavelet transform, the target waveform sequence length is obtained with wavelet reconstruction signal, the peaks of target waveform are detected through first order and second order difference equations, and then laser target scattering center and geometric section ratio can be extracted. Finally, the performance of the proposed algorithm is verified through simulation experiment. In the test with airborne multi-pulse laser target simulator, the laser target multi-frame matching detection experiment is developed, using the features which are extracted by the algorithms proposed in this paper and combining with the target motion state information .The multi-pulsed laser target detection become more reliable.