TiMo: Spatiotemporal Foundation Model for Satellite Image Time Series

• URL: http://arxiv.org/abs/2505.08723v1
• Date: Tue, 13 May 2025 16:35:11 GMT
• Title: TiMo: Spatiotemporal Foundation Model for Satellite Image Time Series
• Authors: Xiaolei Qin, Di Wang, Jing Zhang, Fengxiang Wang, Xin Su, Bo Du, Liangpei Zhang,
• Abstract summary: TiMo is a novel hierarchical vision transformer foundation model tailored for SITS analysis.<n>At its core, we introduce atemporal attention mechanism that dynamically captures multiscale patterns across both time and space.<n>Extensive experiments across multipletemporal tasks-including deforestation monitoring-demonstrate TiMo's superiority over state-of-theart methods.
• Score: 39.22426645737932
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Satellite image time series (SITS) provide continuous observations of the Earth's surface, making them essential for applications such as environmental management and disaster assessment. However, existing spatiotemporal foundation models rely on plain vision transformers, which encode entire temporal sequences without explicitly capturing multiscale spatiotemporal relationships between land objects. This limitation hinders their effectiveness in downstream tasks. To overcome this challenge, we propose TiMo, a novel hierarchical vision transformer foundation model tailored for SITS analysis. At its core, we introduce a spatiotemporal gyroscope attention mechanism that dynamically captures evolving multiscale patterns across both time and space. For pre-training, we curate MillionST, a large-scale dataset of one million images from 100,000 geographic locations, each captured across 10 temporal phases over five years, encompassing diverse geospatial changes and seasonal variations. Leveraging this dataset, we adapt masked image modeling to pre-train TiMo, enabling it to effectively learn and encode generalizable spatiotemporal representations.Extensive experiments across multiple spatiotemporal tasks-including deforestation monitoring, land cover segmentation, crop type classification, and flood detection-demonstrate TiMo's superiority over state-of-the-art methods. Code, model, and dataset will be released at https://github.com/MiliLab/TiMo.

Related papers

• TARDIS STRIDE: A Spatio-Temporal Road Image Dataset and World Model for Autonomy [44.85881816317044]
  We show how to permuting 360-degree panoramic imagery into rich interconnected observation, state and action nodes.<n>We benchmark this dataset via TARDIS, a transformer-based generative world model.<n>We demonstrate robust performance across a range of agentic tasks such as controllable image synthesis, instruction following, autonomous self-control, and state-of-the-art georeferencing.
  arXiv  Detail & Related papers  (2025-06-12T21:08:11Z)
• UniSTD: Towards Unified Spatio-Temporal Learning across Diverse Disciplines [64.84631333071728]
  We introduce bfUnistage, a unified Transformer-based framework fortemporal modeling.<n>Our work demonstrates that a task-specific vision-text can build a generalizable model fortemporal learning.<n>We also introduce a temporal module to incorporate temporal dynamics explicitly.
  arXiv  Detail & Related papers  (2025-03-26T17:33:23Z)
• SatMamba: Development of Foundation Models for Remote Sensing Imagery Using State Space Models [0.0]
  Foundation models refer to deep learning models pretrained on large unlabeled datasets through self-supervised algorithms.<n>Various foundation models have been developed for remote sensing, such as those for multispectral, high-resolution, and hyperspectral images.<n>This research proposes SatMamba, a new pretraining framework that combines masked autoencoders with State Space Model.
  arXiv  Detail & Related papers  (2025-02-01T14:07:21Z)
• EarthView: A Large Scale Remote Sensing Dataset for Self-Supervision [72.84868704100595]
  This paper presents a dataset specifically designed for self-supervision on remote sensing data, intended to enhance deep learning applications on Earth monitoring tasks.<n>The dataset spans 15 tera pixels of global remote-sensing data, combining imagery from a diverse range of sources, including NEON, Sentinel, and a novel release of 1m spatial resolution data from Satellogic.<n>Accompanying the dataset is EarthMAE, a tailored Masked Autoencoder developed to tackle the distinct challenges of remote sensing data.
  arXiv  Detail & Related papers  (2025-01-14T13:42:22Z)
• Deep Multimodal Fusion for Semantic Segmentation of Remote Sensing Earth Observation Data [0.08192907805418582]
  This paper proposes a late fusion deep learning model (LF-DLM) for semantic segmentation. One branch integrates detailed textures from aerial imagery captured by UNetFormer with a Multi-Axis Vision Transformer (ViT) backbone. The other branch captures complex-temporal dynamics from the Sentinel-2 satellite imageMax time series using a U-ViNet with Temporal Attention (U-TAE)
  arXiv  Detail & Related papers  (2024-10-01T07:50:37Z)
• SpectralEarth: Training Hyperspectral Foundation Models at Scale [47.93167977587301]
  We introduce SpectralEarth, a large-scale multi-temporal dataset designed to pretrain hyperspectral foundation models. We pretrain a series of foundation models on SpectralEarth using state-of-the-art self-supervised learning (SSL) algorithms. We construct four downstream datasets for land-cover and crop-type mapping, providing benchmarks for model evaluation.
  arXiv  Detail & Related papers  (2024-08-15T22:55:59Z)
• SatSwinMAE: Efficient Autoencoding for Multiscale Time-series Satellite Imagery [1.6180992915701702]
  We extend the SwinE model to integrate temporal information for satellite time-series data. The architecture employs a hierarchical 3D Masked Autoencoder (MAE) with Video Swin Transformer blocks. Our approach shows significant performance improvements over existing state-of-the-art foundation models for all the evaluated downstream tasks.
  arXiv  Detail & Related papers  (2024-05-03T22:55:56Z)
• Timer: Generative Pre-trained Transformers Are Large Time Series Models [83.03091523806668]
  This paper aims at the early development of large time series models (LTSM) During pre-training, we curate large-scale datasets with up to 1 billion time points. To meet diverse application needs, we convert forecasting, imputation, and anomaly detection of time series into a unified generative task.
  arXiv  Detail & Related papers  (2024-02-04T06:55:55Z)
• Rethinking Urban Mobility Prediction: A Super-Multivariate Time Series Forecasting Approach [71.67506068703314]
  Long-term urban mobility predictions play a crucial role in the effective management of urban facilities and services. Traditionally, urban mobility data has been structured as videos, treating longitude and latitude as fundamental pixels. In our research, we introduce a fresh perspective on urban mobility prediction. Instead of oversimplifying urban mobility data as traditional video data, we regard it as a complex time series.
  arXiv  Detail & Related papers  (2023-12-04T07:39:05Z)
• Unsupervised 4D LiDAR Moving Object Segmentation in Stationary Settings with Multivariate Occupancy Time Series [62.997667081978825]
  We address the problem of unsupervised moving object segmentation (MOS) in 4D LiDAR data recorded from a stationary sensor, where no ground truth annotations are involved. We propose a novel 4D LiDAR representation based on a time series that relaxes the problem of unsupervised MOS. Experiments on stationary scenes from the Raw KITTI dataset show that our fully unsupervised approach achieves performance that is comparable to that of supervised state-of-the-art approaches.
  arXiv  Detail & Related papers  (2022-12-30T14:48:14Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.