Alaska Arctic Highways LiDAR dataset for advanced infrastructure modeling.

The Alaska Arctic Highways LiDAR Dataset (AAHLD) is a curated collection of high-resolution 3D point clouds acquired along approximately 95 km of Arctic roadway corridors in Alaska, including the Dalton, Steese, and Nome-Council highways. Collected between 2022 and 2024 using a combination of ground-vehicle and uncrewed-aircraft LiDAR platforms, the dataset comprises 826 spatial tiles per processing stage, organized into raw, segmentation-reference, and filter-reference products, for a total of 2478 LAZ files and approximately 16.14 GiB. The data were captured with a LiDAR USA Surveyor 32 system, which the manufacturer specifies to support vertical accuracy better than 2 cm and point densities exceeding 300 pts/m2 under appropriate acquisition conditions. The release also includes Nome-only CSV damage labels documenting 3906 unique damage instances across potholes, cracks, edge cracks, and washed-out road sections. The dataset is openly available through the University of North Dakota (UND) Commons repository at https://commons.und.edu/data/39/ under a Creative Commons Attribution-NonCommercial 4.0 International license. AAHLD is intended to support research in Arctic infrastructure monitoring, road segmentation, road damage detection, and machine learning model development in cold-region transportation environments.