Mechanism and algorithm for addressing the impact of multiple scattering on surface elevation extraction in photon-counting LiDAR data

The Ice, Cloud and Land Elevation Satellite-2 (ICESat-2), equipped with the advanced topographic laser altimeter system (ATLAS), utilizes an innovative photon-counting LiDAR technique to conduct precise global elevation measurements. While it offers significant advantages in surface elevation retrieval, its performance can be compromised by substantial multiple scattering in highly scattering turbid media, leading to surface elevation deviations. Despite this issue, the mechanisms and effective algorithms to address the impact of multiple scattering in photon-counting LiDAR data have remained largely unexplored. To fill this gap, this study employs the DART-Lux model to simulate the photon-counting LiDAR signals, incorporating laser multiple scattering, and conducts a comprehensive analysis of its effect on photon spatial distribution and surface elevation retrieval. Additionally, a novel method based on the photon distribution characteristics is proposed to remove multiple scattering photons for surface elevation retrieval in highly scattering turbid media, which combines 2D Cloth Simulation Filtering (CSF) and Empirical Mode Decomposition (EMD) with adaptive classification threshold. Experimental results reveal that multiple scattering photons predominantly accumulate below the surface, with photon density decreasing as elevation declines. This causes a downward shift in the elevation density peak, resulting in surface elevation underestimation. The proposed method in this study effectively mitigates the impact of multiple scattering, a challenge that conventional surface extraction algorithms struggle to address. Through analyzing the surface types in different scenarios including day/night and strong/weak beams, the results indicate that our proposed method outperforms other methods, with an average bias of 0.009 m, MAE of 0.032 m, RMSE of 0.059 m, and R² of 0.990. Our method demonstrates high robustness and precision, particularly over land ice. In summary, this study is the first to not only analyze the impact of multiple scattering on the spatial distribution of photons and surface elevation retrieval, but also provide an effective method for separating multiple scattering photons in photon-counting LiDAR data.