A Deep Learning Architecture for Land Cover Mapping Using Spatio-Temporal Sentinel-1 Features

2025-03-10

AI Summary

Sparse SAR time-series data (e.g., in Siberia) severely degrade land cover classification accuracy. Method: We propose a hybrid architecture integrating seasonal spatiotemporal representation with Swin-Unet. Specifically, we introduce a lightweight, seasonality-aware spatiotemporal representation via seasonal clustering—replacing redundant dense time series—and design a Swin-Unet model that jointly leverages Transformer-based long-range dependencies and CNN-based local texture modeling. The model takes as input multi-season Sentinel-1 SAR backscatter time-series composites. Contribution/Results: Our approach significantly enhances generalization and robustness under data sparsity and class imbalance. It achieves high overall accuracy (O.A.) across ecologically heterogeneous regions—including the Amazon, Africa, and Siberia—with particularly pronounced gains in low-sample regimes, outperforming conventional CNNs and pure Transformer baselines.

Technology Category

• Machine Learning: Semi-Supervised Learning
• Computer Vision: Remote Sensing / Geospatial AI
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Sustainability and carbon-aware systems for Web, mobile, and WoT

Abstract

Land Cover (LC) mapping using satellite imagery is critical for environmental monitoring and management. Deep Learning (DL), particularly Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs), have revolutionized this field by enhancing the accuracy of classification tasks. In this work, a novel approach combining a transformer-based Swin-Unet architecture with seasonal synthesized spatio-temporal images has been employed to classify LC types using spatio-temporal features extracted from Sentinel-1 (S1) Synthetic Aperture Radar (SAR) data, organized into seasonal clusters. The study focuses on three distinct regions - Amazonia, Africa, and Siberia - and evaluates the model performance across diverse ecoregions within these areas. By utilizing seasonal feature sequences instead of dense temporal sequences, notable performance improvements have been achieved, especially in regions with temporal data gaps like Siberia, where S1 data distribution is uneven and non-uniform. The results demonstrate the effectiveness and the generalization capabilities of the proposed methodology in achieving high overall accuracy (O.A.) values, even in regions with limited training data.

Problem

• Develops a deep learning model for land cover mapping using Sentinel-1 SAR data.
• Focuses on improving classification accuracy with seasonal spatio-temporal features.
• Evaluates model performance across diverse ecoregions like Amazonia, Africa, and Siberia.

Innovation

• Swin-Unet architecture for LC mapping
• Seasonal spatio-temporal SAR feature utilization
• Improved accuracy in data-scarce regions