HAVANA: Hard negAtiVe sAmples aware self-supervised coNtrastive leArning for Airborne laser scanning point clouds semantic segmentation

TL;DR

This paper introduces HAVANA, a self-supervised contrastive learning method for airborne laser scanning point cloud segmentation that effectively utilizes unlabeled data and improves performance, especially with limited labels.

Contribution

It proposes a novel hard-negative sample aware contrastive learning approach with an AbsPAN strategy to enhance pre-training for point cloud segmentation.

Findings

Outperforms supervised training without pre-training on benchmark datasets.

Achieves over 94% of full-label performance with only 10% labels.

Effective in scenarios with severe label scarcity.

Abstract

Deep Neural Network (DNN) based point cloud semantic segmentation has presented significant achievements on large-scale labeled aerial laser point cloud datasets. However, annotating such large-scaled point clouds is time-consuming. Due to density variations and spatial heterogeneity of the Airborne Laser Scanning (ALS) point clouds, DNNs lack generalization capability and thus lead to unpromising semantic segmentation, as the DNN trained in one region underperform when directly utilized in other regions. However, Self-Supervised Learning (SSL) is a promising way to solve this problem by pre-training a DNN model utilizing unlabeled samples followed by a fine-tuned downstream task involving very limited labels. Hence, this work proposes a hard-negative sample aware self-supervised contrastive learning method to pre-train the model for semantic segmentation. The traditional contrastive learning for point clouds selects the hardest negative samples by solely relying on the distance between the embedded features derived from the learning process, potentially evolving some negative samples from the same classes to reduce the contrastive learning effectiveness. Therefore, we design an AbsPAN (Absolute Positive And Negative samples) strategy based on k-means clustering to filter the possible false-negative samples. Experiments on two typical ALS benchmark datasets demonstrate that the proposed method is more appealing than supervised training schemes without pre-training. Especially when the labels are severely inadequate (10% of the ISPRS training set), the results obtained by the proposed HAVANA method still exceed 94% of the supervised paradigm performance with full training set.