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

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.