High-Resolution Pavement Rutting Dataset (HPRD): Transverse Profiles, Lane Edges, and Multi-Algorithm Rut Depths for Asphalt Roads

This dataset contains the processed high-resolution Line Laser Imaging System (LLIS) data underlying the study "From Legacy to 3D Rutting Measurement Systems: A Quantitative Framework for Continuity, Accuracy, and Efficiency." It comprises over 2.2 million transverse pavement profiles collected with a vehicle-mounted AMES Transverse Profiling System on asphalt roads in Hong Kong, organised into ten road sections — H01–H04 and U01–U06 — totalling 275 segments of 100 m. For each 100 m segment the dataset provides: Intensity grids (.npz; reflected-light intensity; shape 8333×4096; float32), Range grids (.npz; surface elevation in millimetres; shape 8333×4096; float32), Lane edges (.npy; detected left/right lane-edge column indices; shape 2×8333; int16; with a QC plot .png and a processing log .txt), and Rut depths (.csv; per-profile left/right rut depth from three algorithms: AASHTO PP 69-14 five-point, wire model, and a 1.83 m straight-edge). Each road folder contains Intensity/, Range/, Lane_Edge/, and RutDepth_Results/ sub-folders; files are named <Road>Section<NNN><Type>. Grids load with NumPy under the key 'generated_grid'; rut depths load with pandas. Longitudinal spacing 12 mm. Full data dictionary in README.md, software versions in requirements.txt, and a complete file inventory in dataset_manifest.csv. The processing pipeline (three Python/Jupyter notebooks) and all acquisition/analysis parameters are documented in the README.