Background noise reduction for airborne bathymetric full waveforms by creating trend models using Optech CZMIL in the Yellow Sea of China

Abstract

Raw full waveforms of green lasers used in airborne LiDAR bathymetry (ALB) are contaminated by background and random noise related to the environment and ALB devices. Traditional thresholding methods have been widely used to reduce background noise in raw full waveforms on the basis of the assumption of constant background noise. However, background noise that is mainly related to background solar radiation and detector dark current changes over time. Thresholding methods perform poorly on the full waveforms with a wide variation range of background noise. A background noise reduction method considering its wide variation is proposed to decrease the background noise by creating trend models. First, each green full waveform is divided into two parts: pulse- and non-pulse-return waveforms. Second, a linear interpolation is conducted on the non-pulse-return waveform to impute the missing noise. Third, a low-pass filter is used to filter the random noise with high frequency in the imputed non-pulse-return waveform and obtain the trend model of background noise of the full waveform. Finally, the derived background noise model is used to decrease the background noise in the pulse-return waveform. The proposed method is applied to decrease the background noise in raw green full waveforms collected by the Optech coastal zone mapping and imaging LiDAR (CZMIL). The mean and standard deviation of residual noise in the CZMIL waveform reduced by the trend model of background noise are − 0.03 and 3.5 digitizer counts, respectively. The proposed background noise reduction method is easy to apply and can reduce the background noise to a significantly low level. This method is recommended for preprocessing the raw full waveforms of green lasers collected by Optech CZMIL for ALB.