Implicit spatial-frequency fusion of hyperspectral and lidar data via kolmogorov-arnold networks

Zekun Long, Judy X. Yang, Jing Wang, Ali Zia, Guanyiman Fu, Jun Zhou

arXiv:2605.14239 · 2026-05-15 공개 · arXiv · PDF

lidar kan hyperspectral frequency-geometry-fusion hsi-classification implicit-aggregation spline-based houston-2013

Abstract

Hyperspectral image (HSI) classification is challenging in complex scenes due to spectral ambiguity, spatial heterogeneity, and the strong coupling between material properties and geometric structures. Although LiDAR provides complementary elevation information, most HSI-LiDAR fusion methods rely on CNNs or MLPs with fixed activation functions and linear weights. These methods struggle to model structural discontinuities in LiDAR data, intricate spectral features of HSI, and their interactions. In addition, fusion of the two modalities in both spatial and frequency domains with LiDAR guidance remains underexplored. To address these issues, we propose the Implicit Frequency-Geometry Fusion Network (IFGNet), which leverages Kolmogorov-Arnold Networks (KANs) with learnable spline-based functions to adaptively capture highly nonlinear relationships between hyperspectral and LiDAR features. Furthermore, IFGNet introduces a LiDAR-guided implicit aggregation module in both spatial and frequency domains, enhancing geometry-aware spatial representations while capturing global structural patterns. Experiments on the Houston 2013 and MUUFL benchmarks demonstrate that IFGNet consistently outperforms existing fusion methods in overall accuracy, average accuracy, and Cohen's Kappa, while maintaining an efficient architecture.

한국어 요약

📋 한 줄 요약

**[원격 탐사 / 멀티모달 융합]** Kolmogorov-Arnold Network(KAN)와 LiDAR 가이드의 공간·주파수 도메인 융합을 결합한 초분광-LiDAR 분류 네트워크 IFGNet 제안.

🎯 핵심 기여도

💡 핵심 아이디어

고정 활성화의 단순 가산 융합 대신, KAN의 학습 가능한 비선형 변환을 통해 LiDAR가 HSI의 공간·주파수 표현을 가이드하도록 하면, 재료 특성과 기하 구조의 강한 결합을 더 자연스럽게 표현할 수 있다.

🔬 기술적 접근법

📊 주요 결과

💭 의의 및 한계

**의의**: KAN을 원격 탐사 멀티모달 융합에 성공적으로 적용한 사례로, 분광·기하 결합 모델링의 새 경로를 제시. **한계**: 두 개의 벤치마크 데이터에 국한, 더 큰 도시·자연 시나리오나 시계열 멀티모달 데이터로의 확장은 미검증.

🚀 실용적 활용