Evaluation of full-waveform bottom return extraction in whitewater rapids using bathymetric LiDAR dataset

The application of LiDAR and remote sensing methods has long been considered challenging or unfeasible in the turbulent waters of whitewater rapids. Therefore, this study presents a method to survey whitewater rapids using signal processing of full-waveform LiDAR recordings. The subtraction of an idealized water column backscattering response from the recorded waveform and further analysis of the remaining signal allows the extraction of bottom echoes on a single-waveform basis.

To evaluate the extracted points, we surveyed a block ramp with both total station measurements and helicopter-based bathymetric LiDAR. We could demonstrate an increase in the water bottom coverage in whitewater areas through the application of the presented method, which is able to extract bottom echoes for 70% of the waveforms in the surveyed area. This improves LiDAR full-waveform processing by extracting bottom echoes where current processing algorithms were not able to and thus reducing the mean absolute vertical distance to the reference data from 43.9 cm to 13.3 cm.

Context and methodology Full-waveform bathymetric LiDAR dataset with reference data of a surveyed (whitewater) block ramp. The dataset is split into three parts: The full-waveform data as a pickled Pandas DataFrame. The point cloud data (pielach.txt) of the general region where the data was acquired (approximately 45 m x 45 m), the whitewater area of interest (pielach_whitewater.txt) where each point corresponds to a full-waveform data recording, and the new points generated with the method (new_points.txt).

The reference dataset (all_reference.txt), which was acquired by a two-person team using a total station and a reflector prism pole. The data was acquired to serve as a secondary validation of the original whitewater bathymetric LiDAR study by Rhomberg-Kauert et al. (2025b). Important to note: The data is given in a local project coordinate system and thus not georeferenced.

Technical details Full-waveform data (waveform_data.json): A JSON file of a Pandas DataFrame containing the waveform of each point in the pielach_whitewater.txt (in the same order). Each row of the DataFrame contains the "Id" matching the waveforms to the point cloud data and the waveforms. There, each waveform is given as a list of measurements: The sample interval index (approximately 0.5 ns) and the amplitude in analog-to-digital converted (ADC) units ([[index_0, amplitude_0], [index_1, amplitude_1], ...., [index_n, amplitude_n]]).

E.g.: id wfm 0 562282.0 [[535, 69], [536, 316], [537, 753], [538, 1260... 1 562363.0 [[538, 238], [539, 1402], [540, 2794], [541, 2... Point cloud data (pielach_whitewater.txt): The whitewater area of interest with all attributes needed to run the method described in Rhomberg-Kauert et al. (2025b). E.g.: x y z wfm_sbl_id reflectance amplitude ... wfm_ampl_offset 0 -3023.950680 -6780.087890 260.26425 3206350.0 -17.820000 18.250000 ... -8.128573e-08 1 -3021.864260 -6781.127930 260.35101 3210972.0 -14.090000 21.980000 ... -8.128468e-08 The reference data (all_reference.txt): The reference data of each pole measurement is given in the same local coordinate system as the point clouds.

E.g.: x y z 0 -3029.306114 -6776.488996 259.104533 1 -3029.149941 -6775.951857 259.128578 Code setup To run the notebooks and the provided code of the method, the following folder structure has to be implemented after downloading the files: _whitewater_LiDAR data all_new_points.txt all_reference.txt filtered_new_points.txt pielach_whitewater.txt pielach.txt waveform_data.json method __init__.py bathy_convolution.py clean_data.py metrics_and_plotting.py waveform_plotting.py wfm_averaging.py Pielach_Waverforms.ipynb Pielach_Whitewater.ipynb Additionally, to run the method, the following packages are required (requirements.txt): ----- matplotlib==3.10.8 numpy==2.2.6 pandas==2.3.3 scipy==1.15.3 seaborn==0.13.2 scikit-learn==1.7.2 ----- IPython 8.38.0 jupyter_client 8.8.0 jupyter_core 5.9.1 jupyterlab 4.5.2 notebook 7.5.2 ----- References Jan Rhomberg-Kauert, Lucas Dammert, Theresa Himmelsbach, et al. "Evaluation of full-waveform bottom return extraction in whitewater rapids using bathymetric LiDAR", Proc.

SPIE 13666, Remote Sensing for Agriculture, Ecosystems, and Hydrology XXVII, 1366612 (30 Oct 2025); doi.org/10.1117/12.3068055%20. Rhomberg-Kauert, J. , T. Himmelsbach , F. Pöppl , L. Dammert , M. Pfennigbauer , and G. Mandlburger . 2026 . “ Mapping River Bed Topography in Whitewater Rapids Using Bathymetric LiDAR .” River Research and Applications 42 , no. 4 : 886 – 900 . doi.org/10.1002/rra.70109%20.