Advancing ALS Applications with Large-Scale Pre-training: Dataset Development and Downstream Assessment

2025-01-09

AI Summary

To address the performance limitations of airborne laser scanning (ALS) point cloud downstream tasks—stemming from scarce annotations and dataset bias—this work introduces the first large-scale, contiguous U.S.-wide ALS point cloud dataset. We propose a GIS-driven geospatially oriented sampling strategy that stratifies sampling by land cover and topographic features, substantially outperforming random sampling. Furthermore, we pioneer the adaptation of the BEV-MAE masked autoencoder to ALS point clouds, establishing an end-to-end self-supervised pretraining paradigm. Experiments demonstrate that the pretrained representations consistently surpass from-scratch baselines across diverse downstream tasks—including tree species classification, terrain identification, and semantic segmentation. Critically, performance scales monotonically with dataset size under oriented sampling, whereas random sampling yields no such gains. The dataset, code, and pretrained models will be publicly released.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Machine Learning: Unsupervised & Self-Supervised Learning
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

The pre-training and fine-tuning paradigm has revolutionized satellite remote sensing applications. However, this approach remains largely underexplored for airborne laser scanning (ALS), an important technology for applications such as forest management and urban planning. In this study, we address this gap by constructing a large-scale ALS point cloud dataset and evaluating its impact on downstream applications. Our dataset comprises ALS point clouds collected across the contiguous United States, provided by the United States Geological Survey's 3D Elevation Program. To ensure efficient data collection while capturing diverse land cover and terrain types, we introduce a geospatial sampling method that selects point cloud tiles based on land cover maps and digital elevation models. As a baseline self-supervised learning model, we adopt BEV-MAE, a state-of-the-art masked autoencoder for 3D outdoor point clouds, and pre-train it on the constructed dataset. The pre-trained models are subsequently fine-tuned for downstream tasks, including tree species classification, terrain scene recognition, and point cloud semantic segmentation. Our results show that the pre-trained models significantly outperform their scratch counterparts across all downstream tasks, demonstrating the transferability of the representations learned from the proposed dataset. Furthermore, we observe that scaling the dataset using our geospatial sampling method consistently enhances performance, whereas pre-training on datasets constructed with random sampling fails to achieve similar improvements. These findings highlight the utility of the constructed dataset and the effectiveness of our sampling strategy in the pre-training and fine-tuning paradigm. The source code and pre-trained models will be made publicly available at url{https://github.com/martianxiu/ALS_pretraining}.

Problem

• Laser Scanning (ALS)
• Forest Management
• Urban Planning

Innovation

• BEV-MAE Model
• Geospatial Sampling
• Aircraft Laser Scanning (ALS)