Self-Supervised Pre-Training Boosts Semantic Scene Segmentation on LiDAR Data
This work addresses the time-consuming labeling issue in LiDAR data for researchers and practitioners in remote sensing, though it is incremental as it applies an existing self-supervised technique to a specific domain.
The paper tackled the problem of labeling airborne LiDAR point clouds for semantic segmentation by proposing a self-supervised pre-training method using Barlow Twins, which boosted performance when fine-tuned, particularly for under-represented categories.
Airborne LiDAR systems have the capability to capture the Earth's surface by generating extensive point cloud data comprised of points mainly defined by 3D coordinates. However, labeling such points for supervised learning tasks is time-consuming. As a result, there is a need to investigate techniques that can learn from unlabeled data to significantly reduce the number of annotated samples. In this work, we propose to train a self-supervised encoder with Barlow Twins and use it as a pre-trained network in the task of semantic scene segmentation. The experimental results demonstrate that our unsupervised pre-training boosts performance once fine-tuned on the supervised task, especially for under-represented categories.