基于高斯与小波的机载测深LiDAR波形分解算法

An algorithm for Airborne LiDAR Bathymetric Waveform Decomposition Based on the Gaussian Model and Wavelet

波形分解是机载测深LiDAR数据处理的关键环节,为水深计算、底质类型反演和水体浑浊度分析等提供基础信息。针对传统测深LiDAR波形分解算法受噪声干扰严重、对微弱及叠加信号分解不准确的问题,提出一种新的波形分解算法。对原始波形经小波滤波后,计算滤波前后尾段波形的差异,估计回波信号的噪声;利用高斯模型,从原始波形数据中不断分解出经LM算法优化参数后的波形分量,直到剩余波形中最大峰值与优化后的参数小于一定阈值。通过南海实测数据进行验证,实验结果表明:该算法分解弱回波能力强,不论在浅水(回波发生叠加)还是深水,其分解精度均优于传统算法。

Decomposition of waveform is a key link in airborne LiDAR bathymetry data processing, which can provide the basic information for deriving water depth, inverting seafloor sediment tyoe, and analyzing water turbidity. Traditional waveform fitting algorithms are less robust to noise, and cannot detect weak signals and superposed signals in an accurate way. To overcome these problems, a method was proposed in this paper. Firstly,the proposed algorithm is used to calculate the difference between the last part of the original waveform and the waveform after wavelet threshold denoising. Then, the Gaussian model is adopted to continuously extract Gaussian components which are optimized by the LM algorithm. Then this algorithm is tested by measured data obtained from the South China Sea, demonstrating that it can detect weak signals. The experiment results show that,whether in shallow or deep water, the fitting precision of the proposed algorithm outperforms the traditional algorithms.