A Scale-Attention Optimized Hybrid Deep Learning Model for High-Accuracy Soil Salinity Mapping in Arid Oases

Accurate monitoring of soil salinization in arid oasis regions is crucial for agricultural sustainability and ecological security. However, existing deep learning-based approaches often suffer from insufficient utilization of multi-scale information and inadequate modelling of feature interactions, limiting their accuracy in retrieving complex salinity patterns. To address these limitations, this study proposes a scale-attention optimized hybrid deep learning model that integrates multi-scale 1D convolutional neural networks (1D-CNN), bidirectional gated recurrent units (Bi-GRU), and Transformer mechanisms. The model employs a multi-scale feature extraction module to capture remote sensing signals across different scales, a scale attention mechanism to adaptively weight the most informative features, and a Bi-GRU-Transformer module to explore complex sequential and global feature relationships. The proposed framework is applied to the oasis irrigation zone in Weili County, Xinjiang, using hyperspectral data from the ZY-1E satellite, topographic indices, and spectral-derived variables. Experimental results demonstrate that our method achieves a coefficient of determination (R²) of 0.952 and a root mean square error (RMSE) of 0.867 g·kg⁻¹ on the test set, outperforming conventional 1D-CNN, GRU-Transformer, and other benchmark models with improvements of 2.8% in R² and 18.9% in RMSE.