Preserving Vertical Structure in 3D-to-2D Projection for Permafrost Thaw Mapping

TL;DR

This paper introduces a novel projection decoder with learned height embeddings and stratified sampling to preserve vertical structure in 3D-to-2D lidar data, improving permafrost thaw prediction in forests.

Contribution

It presents a new height-dependent feature transformation method combined with stratified sampling, enhancing vertical structure preservation in 3D-to-2D projections for thaw mapping.

Findings

Z-stratified projection outperforms standard methods in complex forests.

The approach enables high-resolution, scalable permafrost monitoring from UAV lidar data.

Method effectively differentiates ground, understory, and canopy signals.

Abstract

Forecasting permafrost thaw from aerial lidar requires projecting 3D point cloud features onto 2D prediction grids, yet naive aggregation methods destroy the vertical structure critical in forest environments where ground, understory, and canopy carry distinct information about subsurface conditions. We propose a projection decoder with learned height embeddings that enable height-dependent feature transformations, allowing the network to differentiate ground-level signals from canopy returns. Combined with stratified sampling that ensures all forest strata remain represented, our approach preserves the vertical information critical for predicting subsurface conditions. Our approach pairs this decoder with a Point Transformer V3 encoder to predict dense thaw depth maps from drone-collected lidar over boreal forest in interior Alaska. Experiments demonstrate that z-stratified projection outperforms standard averaging-based methods, particularly in areas with complex vertical vegetation structure. Our method enables scalable, high-resolution monitoring of permafrost degradation from readily deployable UAV platforms.