Deep Spatial Feedback Refined Network With Multilevel Feature Fusion for Hyperspectral Image Subpixel Mapping
- 26 June 2024
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 62 (01962892) , 1-20
Abstract
The issue of mixed pixels is a prevalent challenge in hyperspectral images, largely due to imaging modalities and hardware limitations. Sub-pixel mapping (SPM) can address this issue by distinguishing between land cover classes at a sub-pixel scale. In recent years, deep convolutional neural networks have demonstrated their potential and effectiveness for sub-pixel mapping. However, during the SPM process, the unique strengths of features at different network levels are often neglected, resulting in insufficient interaction between features at varying levels. Consequently, the network cannot fully extract and utilize the information from higher and lower-level features, nor can it thoroughly understand and recognize the input images. This ultimately hampers the model’s SPM performance. In this study, we propose a deep spatial feedback refined network with multi-level feature fusion (SFRNet-MLFF) for sub-pixel mapping of hyperspectral remote sensing images. This innovative network incorporates a feedback mechanism and a multi-level feature fusion technique, both of which significantly enhance the interaction and aggregation of information across diverse features. Furthermore, throughout the entire network process, we apply solution space constraints to various spatial resolution contexts. These strategic implementations are expected to substantiate the network’s performance in sub-pixel mapping. Experimental results on three hyperspectral datasets demonstrate that SFRNet-MLFF produces results with more distinct feature outlines. These results are better equipped to indicate the precise location and distribution patterns of land cover classes in low-resolution images at higher spatial resolutions. Furthermore, SFRNet-MLFF significantly improves the overall accuracy compared to the state-of-the-art deep learning SPM networks (DLSPMNets).Keywords
Funding Information
- National Natural Science Foundation of China (U21A2013, 62071438)
- Open Fund of Hebei Key Laboratory of Ocean Dynamics, Resources and Environments, Qinhuangdao (HBHY2302)
- Open Fund of State Key Laboratory of Remote Sensing Science (OFSLRSS202312)
- Fundamental Research Funds for the Central Universities, China University of Geosciences (2642022009)
- global change and air-sea interaction ii (GASI-01-DLYG-WIND0)
- Foundation of State Key Laboratory of Public Big Data (PBD2023-28)
- Open Fund of Key Laboratory of Regional Development and Environmental Response (2023(A)003)
- Open Fund of Wenzhou Future City Research Institute (WL2023007)
- Open Fund of Key Laboratory of Space Ocean Remote Sensing and Application, Ministry of Natural Resources (202401001)
This publication has 59 references indexed in Scilit:
- Very fine spatial resolution urban land cover mapping using an explicable sub-pixel mapping network based on learnable spatial correlationRemote Sensing of Environment, 2023
- High-quality super-resolution mapping using spatial deep learningiScience, 2023
- Coupled retrieval of heavy metal nickel concentration in agricultural soil from spaceborne hyperspectral imageryJournal of Hazardous Materials, 2023
- Multi-level features fusion via cross-layer guided attention for hyperspectral pansharpeningNeurocomputing, 2022
- Generating 2m fine-scale urban tree cover product over 34 metropolises in China based on deep context-aware sub-pixel mapping networkInternational Journal of Applied Earth Observation and Geoinformation, 2022
- Estimating the distribution trend of soil heavy metals in mining area from HyMap airborne hyperspectral imagery based on ensemble learningJournal of Hazardous Materials, 2021
- The mixed pixel effect in land surface phenology: A simulation studyRemote Sensing of Environment, 2018
- Generating a series of fine spatial and temporal resolution land cover maps by fusing coarse spatial resolution remotely sensed images and fine spatial resolution land cover mapsRemote Sensing of Environment, 2017
- Adaptive pixel unmixing based on a fuzzy ARTMAP neural network with selective endmembersSoft Computing, 2015
- A new sub-pixel mapping algorithm based on a BP neural network with an observation modelNeurocomputing, 2008