摘要
针对全波形三维激光测绘雷达(LiDAR)在数字地形测量中如何降低背景噪声问题,提出了一种基于经验模态分解(EMD)和小波阈值的自适应降噪方法。在扫测的地形信号经EMD分解后,计算内蕴模式函数(IMF)与经过2/3阶重构的扫测信号之间的互相关函数,从而改善小波阈值自适应地对IMF中的高频噪声成分进行滤除。实验结果表明,与EMD重构降噪法、小波阈值降噪法和传统的EMD-小波联合降噪法比较,这种方法在对全波形LiDAR回波信号的噪声剔除和地物信号保留方面具有明显的优势,降噪后信号的误差能缩小10%~20%,波形相关性能提升5%~20%,信噪比能提升20%~40%。
Aiming at the problem of reducing the background noise derived from full-waveform LiDAR in digital terrain survey, this paper has put forward a self-adapting denoising method based on Empirical Mode Decomposition(EMD) and wavelet threshold method(WTM).The cross-relation functions, between the signal reconstructed by 2/3 degree and Intrinsic Mode Functions(IMFs) obtained from EMD,is utilized to improve WTM.As thus, the improved WTM has the ability to filter the noise with high frequency in the IMFs.The experiment result shows that, compared with EMD reconstruction method, wavelet threshold method and the traditional EMD-wavelet directly denoising method, the proposed method has better performance in reducing the noise and reserving the original topographic signal.Meanwhile, after being denoised, the signal error is reduced from 10% to 20%,the waveform relativity and the signal-noise ratio are improved from 5% to 20%,from 120% to 140%,respectively.
作者
石志远
徐卫明
周波
孟浩
SHI Zhiyuan;XU Weiming;ZHOU Bo;MENG Hao(Department of Military Oceanography and Hydrography,Dalian Naval Academy,Dalian 116018,China;Department of Basic Sciences,Dalian Naval Academy,Dalian 116018,China)
出处
《海洋测绘》
CSCD
北大核心
2021年第6期54-57,72,共5页
Hydrographic Surveying and Charting
基金
国家自然科学基金(61071006)。
关键词
信号降噪
全波形激光雷达
经验模态分解
小波阈值
噪声自适应
signal denoising
full-waveform LiDAR
empirical mode decomposition
wavelet threshold
noise self-adapting
