基于海底回波增强的机载LiDAR测深水体旅行时提取算法

An algorithm for extracting airborne LiDAR bathymetric travel time in water column based on seabed echo enhancement

机载LiDAR测深(Airborne LiDAR Bathymetry, ALB)技术具有高精度、高效率、强机动性、水陆两用等优势,特别适合海岸带、海岛礁等浅水海域复杂地形的快速探测。激光穿透水体时能量将迅速衰减,导致部分海底回波难以有效提取,海底真实位置判别困难。为此,本文提出一种基于回波增强的机载LiDAR测深水体旅行时提取算法。通过Gold去卷积算法来恢复目标横截面形状,确定海底初始回波范围;随后采用双指数函数拟合水体后向散射有效范围,进而求取波形漫衰减系数Kd值;最后结合海底激光雷达方程,利用K_(d)值对海底初始回波范围内波形进行增强,并利用高斯函数分解增强后回波,确定海底位置参数,从而实现ALB波形的水体旅行时提取。利用青岛胶州湾RIEGL VQ-840-G ALB实验数据对本文算法的可行性进行验证,将本文算法与理查德森-露西(Richardson-Lucy,RL)去卷积模型、峰值探测模型进行了比对,结果表明本文算法与单波束同名点之间高程误差的均方根误差(Root Mean Square Error, RMSE)为18.5 cm,较上述两种算法分别降低了29.9%、41.4%。因此,本文算法具有可行性,能够满足ALB波形的水体旅行时高精度提取,可为机载LiDAR测深数据精细化处理提供一定技术支撑。

The airborne LiDAR bathymetry(ALB)technology has the advantages of high precision,high efficiency,strong mobility and dual use of water and land.It is especially suitable for the rapid detection of complex terraforming in shallow waters such as coastal zones,islands and reefs.When the laser penetrates the water,the energy will attenuate rapidly,which makes it difficult to extract part of the seabed echo effectively and distinguish the true position of the sea bottom.Therefore,an airborne LiDAR bathymetric travel time in the water column extraction algorithm based on echo enhancement is proposed in this paper.The Gold deconvolution algorithm was used to restore the cross section shape of the target and determine the initial range of the seabed.Then,the effective range of backscattering was fitted by double exponential function,and the diffuse attenuation coefficient Kd was obtained.Finally,combined with the seabed LiDAR equation,the waveform in the initial range of the seabed is enhanced by Kd value,and the enhanced echo is decomposed by Gaussian function to determine the seabed position parameters,so as to realize the travel time in the water column extraction of ALB waveform.The feasibility of the proposed algorithm was verified by using the experimental data of RIEGL VQ-840-G ALB in Qingdao Jiaozhou Bay,and the proposed algorithm was compared with the Richardson-Lucy deconvolution model and the peak detection model.The results show that the root mean square error(RMSE)between the proposed algorithm and the single-beam point with the same name is 18.5 cm,which is 29.9%and 41.4%lower than the above two algorithms,respectively.Therefore,the proposed algorithm is feasible and can satisfy the high precision extraction of ALB waveform during water column traveling,which can provide certain technical support for the fine processing of airborne LiDAR bathymetry data.