用时空变压器预测长期植被变化，精准捕捉复杂生态系统的动态。
Spatio-Temporal Transformers for Long-Term NDVI Forecasting
- 构建统一架构，同时建模空间局部关系与区域气候影响。
- 在40年遥感数据上训练，下一年预测MAE仅0.0328，R²达0.8412。
- 适合处理不规则采样和快速生态变化的异质景观分析。
异质景观中长期卫星图像时序（SITS）分析面临巨大挑战，尤其在地中海地区，复杂的空间模式、季节变化及跨尺度的数十年环境变迁相互作用。本文提出时空变压器长时预测框架（STT-LTF），突破纯时间分析局限，将空间上下文建模与时间序列预测融合。该框架通过统一的Transformer架构处理多尺度空间块与长达20年的时序数据，捕捉局部邻域关系与区域气候影响。采用全面的自监督学习策略，结合空间掩码、时间掩码与未来时点采样，利用40年未标注的Landsat影像进行稳健训练。不同于自回归方法，STT-LTF可直接预测任意未来时间点，避免误差累积，引入空间块嵌入、周期性时间编码与地理坐标，学习异质地中海生态系统中的复杂依赖关系。在Landsat数据（1984–2024）上的实验表明，其下一年预测的平均绝对误差（MAE）为0.0328，决定系数R²达0.8412，优于传统统计方法、基于CNN的方法、LSTM网络与标准Transformer。该框架对不规则采样与可变预测时长的适应能力，使其特别适用于经历快速生态转变的异质景观分析。
原文摘要 · Abstract (English)
Long-term satellite image time series (SITS) analysis in heterogeneous landscapes faces significant challenges, particularly in Mediterranean regions where complex spatial patterns, seasonal variations, and multi-decade environmental changes interact across different scales. This paper presents the Spatio-Temporal Transformer for Long Term Forecasting (STT-LTF ), an extended framework that advances beyond purely temporal analysis to integrate spatial context modeling with temporal sequence prediction. STT-LTF processes multi-scale spatial patches alongside temporal sequences (up to 20 years) through a unified transformer architecture, capturing both local neighborhood relationships and regional climate influences. The framework employs comprehensive self-supervised learning with spatial masking, temporal masking, and horizon sampling strategies, enabling robust model training from 40 years of unlabeled Landsat imagery. Unlike autoregressive approaches, STT-LTF directly predicts arbitrary future time points without error accumulation, incorporating spatial patch embeddings, cyclical temporal encoding, and geographic coordinates to learn complex dependencies across heterogeneous Mediterranean ecosystems. Experimental evaluation on Landsat data (1984-2024) demonstrates that STT-LTF achieves a Mean Absolute Error (MAE) of 0.0328 and R^2 of 0.8412 for next-year predictions, outperforming traditional statistical methods, CNN-based approaches, LSTM networks, and standard transformers. The framework's ability to handle irregular temporal sampling and variable prediction horizons makes it particularly suitable for analysis of heterogeneous landscapes experiencing rapid ecological transitions.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。