论文部分内容阅读
在使用经验模式分解(Empirical Mode Decomposition,EMD)对激光雷达回波信号进行去噪处理时,由于信号含有脉冲及间歇等间断事件而产生模态混叠,导致不能很好地分解出有用信号成分,影响去噪效果。针对这一问题,提出了一种形态滤波与EMD相结合的组合算法。首先,使用自适应多尺度形态滤波器作为前置单元,对信号进行初步处理,剔除信号中的间断事件干扰。之后,应用EMD对处理过的信号去噪。采用仿真数据及真实激光雷达回波数据进行了去噪实验。实验结果表明,文中算法相比于直接EMD去噪,在仿真试验中信噪比提高了8.89 d B,均方根误差降低了0.0514;在真实回波数据去噪实验中,6 km以后平均信噪比提高了3.356 4 d B。该组合算法有效地抑制了模态混叠现象,具有良好的去噪效果及应用前景。
When using the empirical mode decomposition (EMD) to denoise the lidar echo signal, the mode aliasing occurs due to the signal containing pulse and intermittent intermittent events, resulting in the inability to decompose the useful signal component , Affect the denoising effect. Aiming at this problem, a combination algorithm of morphological filtering and EMD is proposed. First of all, using adaptive multi-scale morphological filter as the pre-unit, the signal is processed preliminarily, and the interference in the signal is eliminated. After that, the processed signal is denoised using EMD. De-noising experiment was carried out by using simulation data and real laser radar echo data. The experimental results show that compared with the direct EMD denoising algorithm, the SNR increases by 8.89 d B in the simulation and the root mean square error decreases by 0.0514. In the real echo data denoising experiment, the average signal after 6 km The noise ratio has been increased by 3.356 4 days. The combined algorithm effectively suppresses the phenomenon of mode aliasing and has good denoising effect and application prospect.