A photon-counting LiDAR bathymetric method based on adaptive variable ellipse filtering

Abstract

The Ice, Cloud, and Land Elevation Satellite-2 (ICESat-2), equipped with the advanced topographic laser altimeter system (ATLAS), is a new measurement strategy that includes a photon-counting technique. Employing such an advanced system facilitates highly accurate measurement of changes in the land surface of the Earth, as well as water depths, thereby providing a novel opportunity for nearshore bathymetry. The main problem in photon-counting bathymetry is the precise identification and separation of photons with different density distributions on the water surface and water-column. Further, automatic detection of the signal photon of the water surface and bottom from the noise photon presents a problem. To overcome such problems, we propose a novel, validated, adaptive variable ellipse filtering bathymetric method (AVEBM). This method can precisely identify and separate the photons in the above-water, water surface, and water-column regions. Filter parameters can be determined automatically by the density distribution of photons in different water environments and water depths, and the signal photons of the water surface and bottom can be detected precisely to conduct bathymetry. To ensure and improve the accuracy and reliability of the AVEBM, datasets of different photon densities were selected and employed in conducting experiments. Further, we used in situ data and remote sensing images to validate and estimate the detection accuracy of the signal photon and bathymetric accuracy. The experimental results indicated that bathymetric accuracy based on the AVEBM method reached 0.48 m of root mean square error (RMSE) when we employed three ATL03 datasets in combination for Qilianyu Islands. Moreover, the bathymetric accuracy reached 0.79 m of RMSE when comparing with the SDB result in Shanhu Island. Results showed that the proposed AVEBM method facilitated highly accurate signal photon detection and bathymetry.