Hypergraph Convolution Network Classification for Hyperspectral and LiDAR Data
Conventional remote sensing classification approaches based on single-source data exhibit inherent limitations, driving significant research interest in improved multimodal data fusion techniques. Although deep learning methods based on convolutional neural networks (CNNs), transformers, and graph c...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-05-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/25/10/3092 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850126898192973824 |
|---|---|
| author | Lei Wang Shiwen Deng |
| author_facet | Lei Wang Shiwen Deng |
| author_sort | Lei Wang |
| collection | DOAJ |
| description | Conventional remote sensing classification approaches based on single-source data exhibit inherent limitations, driving significant research interest in improved multimodal data fusion techniques. Although deep learning methods based on convolutional neural networks (CNNs), transformers, and graph convolutional networks (GCNs) have demonstrated promising results in fusing complementary multi-source data, existing methodologies demonstrate limited efficacy in capturing the intricate higher-order spatial–spectral dependencies among pixels. To overcome these limitations, we propose HGCN-HL, a novel multimodal deep learning framework that integrates hypergraph convolutional networks (HGCNs) with lightweight CNNs. Specifically, an adaptive weight mechanism is first designed to preliminarily fuse the spectral features of hyperspectral imaging (HSI) and Light Detection and Ranging (LiDAR), enhancing the feature representation ability. Then, superpixel-based dynamic hyperedge construction enables the joint characterization of homogeneous regions across both modalities, significantly boosting large-scale object recognition accuracy. Finally, local detail features are captured through a parallel CNN branch, complementing the global relationship modeling of the HGCN. Comprehensive experiments conducted on three benchmark datasets demonstrate the superior performance of our method compared to existing state-of-the-art approaches. Notably, the proposed framework achieves significant improvements in both training efficiency and inference speed while maintaining competitive accuracy. |
| format | Article |
| id | doaj-art-ffe9271f51a14f389bc1d35d40b441f7 |
| institution | OA Journals |
| issn | 1424-8220 |
| language | English |
| publishDate | 2025-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj-art-ffe9271f51a14f389bc1d35d40b441f72025-08-20T02:33:48ZengMDPI AGSensors1424-82202025-05-012510309210.3390/s25103092Hypergraph Convolution Network Classification for Hyperspectral and LiDAR DataLei Wang0Shiwen Deng1College of Geographical Sciences, Harbin Normal University, Harbin 150025, ChinaSchool of Mathematical Sciences, Harbin Normal University, Harbin 150025, ChinaConventional remote sensing classification approaches based on single-source data exhibit inherent limitations, driving significant research interest in improved multimodal data fusion techniques. Although deep learning methods based on convolutional neural networks (CNNs), transformers, and graph convolutional networks (GCNs) have demonstrated promising results in fusing complementary multi-source data, existing methodologies demonstrate limited efficacy in capturing the intricate higher-order spatial–spectral dependencies among pixels. To overcome these limitations, we propose HGCN-HL, a novel multimodal deep learning framework that integrates hypergraph convolutional networks (HGCNs) with lightweight CNNs. Specifically, an adaptive weight mechanism is first designed to preliminarily fuse the spectral features of hyperspectral imaging (HSI) and Light Detection and Ranging (LiDAR), enhancing the feature representation ability. Then, superpixel-based dynamic hyperedge construction enables the joint characterization of homogeneous regions across both modalities, significantly boosting large-scale object recognition accuracy. Finally, local detail features are captured through a parallel CNN branch, complementing the global relationship modeling of the HGCN. Comprehensive experiments conducted on three benchmark datasets demonstrate the superior performance of our method compared to existing state-of-the-art approaches. Notably, the proposed framework achieves significant improvements in both training efficiency and inference speed while maintaining competitive accuracy.https://www.mdpi.com/1424-8220/25/10/3092superpixelshypergraph convolutional networkshyperedgehyperspectral image |
| spellingShingle | Lei Wang Shiwen Deng Hypergraph Convolution Network Classification for Hyperspectral and LiDAR Data Sensors superpixels hypergraph convolutional networks hyperedge hyperspectral image |
| title | Hypergraph Convolution Network Classification for Hyperspectral and LiDAR Data |
| title_full | Hypergraph Convolution Network Classification for Hyperspectral and LiDAR Data |
| title_fullStr | Hypergraph Convolution Network Classification for Hyperspectral and LiDAR Data |
| title_full_unstemmed | Hypergraph Convolution Network Classification for Hyperspectral and LiDAR Data |
| title_short | Hypergraph Convolution Network Classification for Hyperspectral and LiDAR Data |
| title_sort | hypergraph convolution network classification for hyperspectral and lidar data |
| topic | superpixels hypergraph convolutional networks hyperedge hyperspectral image |
| url | https://www.mdpi.com/1424-8220/25/10/3092 |
| work_keys_str_mv | AT leiwang hypergraphconvolutionnetworkclassificationforhyperspectralandlidardata AT shiwendeng hypergraphconvolutionnetworkclassificationforhyperspectralandlidardata |