DINCoDE: A Data Interpolation Network With a Collaborative Dual Encoder for Reconstructing Missing Sea Surface Temperature Data

Xin Yang

doi:10.1109/TGRS.2025.3638341

Return

DINCoDE: A Data Interpolation Network With a Collaborative Dual Encoder for Reconstructing Missing Sea Surface Temperature Data

2025-01-01

PRE

DOI：10.1109/TGRS.2025.3638341

Original

Original request for help

Save

Abstract

En 中文

Sea surface temperature (SST) data derived from infrared satellites are frequently missing due to cloud cover. To address this challenge, we propose a novel interpolation model, the data interpolation network with a collaborative dual encoder (DINCoDE). The model adopts a hybrid dual-encoder architecture that integrates a convolutional neural network (CNN) and a Transformer encoder, enabling the joint extraction of local and global spatiotemporal features. A composite loss function is employed to ensure both pixel-level accuracy and structural fidelity in the reconstructed data. Experimental results on the OSTIA SST dataset ( $N=6575$ ) demonstrate that DINCoDE outperforms the data-interpolating convolutional autoencoder (DINCAE) and data-interpolating empirical orthogonal function (DINEOF) models across all evaluation metrics [RMSE $= 0.25~^{\circ }$ C, mean absolute error (MAE) $= 0.16~^{\circ }$ C, $R^{2} =0.99$ , and structural similarity (SSIM) = 0.99] and generates more physically realistic spatial gradient fields. Notably, the pretrained model demonstrated robust performance when applied to the Advanced Himawari Imager (AHI) SST dataset (RMSE $= 0.38~^{\circ }$ C, MAE $= 0.27~^{\circ }$ C, and $R^{2} =0.99$ ), highlighting its strong generalization ability and potential for operational deployment. Moreover, our analysis reveals that the intrinsic spatial complexity and variability of the SST field are key factors influencing reconstruction performance. DINCoDE offers a high-accuracy, robust, and generalizable method for reconstructing SST data gaps caused by cloud occlusion, providing a reliable solution for SST data reconstruction tasks in oceanographic applications.

Keywords:

Convolutional neural network (CNN)

data interpolation

dual-encoder network

missing data reconstruction

sea surface temperature (SST)

Transformer