# Adaptive graph contrastive learning with hard negative mining for multimodal hyperspectral and LiDAR classification

## TL;DR

This paper introduces a self-supervised method for classifying hyperspectral and LiDAR data using adaptive graph learning and contrastive techniques.

## Contribution

AGCL introduces adaptive graph construction and hard negative mining for self-supervised multimodal remote sensing classification.

## Key findings

- AGCL improves classification accuracy through adaptive graph refinement and cross-modal alignment.
- Hard negative mining enhances contrastive learning by reducing false negatives.
- The method outperforms existing approaches on benchmark datasets.

## Abstract

Joint classification of hyperspectral imagery (HSI) and light detection and ranging (LiDAR) data has attracted increasing attention in remote sensing. However, effective multimodal fusion and robust feature modeling remain challenging due to data heterogeneity. Graph neural networks (GNNs) are well suited for modeling non-Euclidean structures and cross-modal relations, but most existing GNN-based methods rely on supervised learning, limiting their applicability in label-scarce scenarios. We propose adaptive graph contrastive learning (AGCL), a self-supervised graph framework for HSI and LiDAR classification. AGCL performs adaptive graph construction through input-conditioned neighborhood selection and learns dynamic affinity matrices for flexible message passing. A hard negative mining strategy constructs informative negative samples for contrastive learning. During self-supervised pretraining, AGCL jointly optimizes intra-modal consistency, cross-modal alignment, and graph topology reconstruction without labeled data. The learned representations are then transferred to downstream classification via supervised fine-tuning. Experiments on three benchmark datasets demonstrate the effectiveness of the proposed framework.

•Adaptive graph refinement learns task-aware and evolving graph topologies•Multi-order aggregation captures both local and long-range dependencies•Cross-modal contrastive learning aligns HSI and LiDAR representations•Hard negative mining reduces false negatives caused by spatial and modal correlations

Adaptive graph refinement learns task-aware and evolving graph topologies

Multi-order aggregation captures both local and long-range dependencies

Cross-modal contrastive learning aligns HSI and LiDAR representations

Hard negative mining reduces false negatives caused by spatial and modal correlations

Remote sensing; computer science; computing methodology; artificial intelligence; computational intelligence