DINFNN: Data Inpainting Fourier Neural Network for Cloud-Induced Extensive Missing Area in Sea Surface Temperature
- 28 October 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) , 1-14
Abstract
Sea surface temperature (SST) serves as a critical indicator of oceanic environmental changes. However, the uninterrupted observation of vast oceanic areas via remote sensing is frequently impeded by cloud cover, resulting in persistent data gaps. Consequently, the completion of SST data emerges as an essential technique. Recently, the utilization of deep neural networks, particularly generative models, has shown promising results by effectively leveraging historical data for training. However, these approaches often concentrate on spatial domain features, neglecting the inherent chaotic nature of the ocean as a complex system, thus failing to accurately reflect the complexity of oceanic processes. Therefore, this study introduces a novel approach to SST completion focusing on frequency domain features, utilizing Fourier neural operators. We propose an innovative Data INpainting Fourier Neural Network, namely DINFNN, for data completion to facilitate feature learning in complex oceanic systems. Our method employs a triple-stream neural network to capture periodic steady-state features, adjacent temporal features, and current context. By integrating high-pass and low-pass filters and dynamically combining them, we adaptively extract essential frequency domain features for completion. Particularly, a Two-Cascaded Fusion module is utilized dynamically to amalgamate these distinct attributes to form a composite feature set, encompassing both steady and dynamic frequency domain elements, aimed at achieving a comprehensive SST field reconstruction. In our experimental evaluations, our method demonstrates significant improvements under various conditions. Specifically, when the cover ratio is 68% and the signal-to-noise ratio (N/S) is set to 0.1, 0.2, and 0.3, we observe enhancements of 11.2%, 10.8%, and 16.2% in R 2 , respectively. Furthermore, corresponding reductions of 7.2%, 13.1%, and 7.3% in RMSE are achieved, respectively. Moreover, comprehensive ablation experiments confirm the effectiveness of each component within our method and emphasize the superiority of DINFNN over conventional operators.Keywords
Funding Information
- National Key Research and Development Program of China (2023YFC3108700)
- National Natural Science Foundation of China (U23A20320)
- Postdoctoral Fellowship Program of China Postdoctoral Science Foundation (GZB20240694)
This publication has 39 references indexed in Scilit:
- Machine learning for phase-resolved reconstruction of nonlinear ocean wave surface elevations from sparse remote sensing dataOcean Engineering, 2023
- Grid search hyperparameter tuning in additive manufacturing processesManufacturing Letters, 2023
- Long-term predictions of turbulence by implicit U-Net enhanced Fourier neural operatorPhysics of Fluids, 2023
- Fourier Imager Network (FIN): A deep neural network for hologram reconstruction with superior external generalizationLight: Science & Applications, 2022
- Research on Sea Surface temperature Reconstruction from long-term MODIS dataIOP Conference Series: Earth and Environmental Science, 2021
- Using Sea Surface Temperature Observations to Constrain Upper Ocean Properties in an Arctic Sea Ice‐Ocean Data Assimilation SystemJournal of Geophysical Research: Oceans, 2019
- Impact of ocean warming on sustainable fisheries management informs the Ecosystem Approach to FisheriesScientific Reports, 2017
- Retrieving ocean surface current by 4-D variational assimilation of sea surface temperature imagesRemote Sensing of Environment, 2008
- Fisheries oceanography using satellite and airborne remote sensing methods: a reviewFisheries Research, 2000
- Gradient-based learning applied to document recognitionProceedings of the IEEE, 1998