Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2505.21357 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908765941399552 |
|---|---|
| author | Li, Wenyuan Liang, Shunlin Chen, Keyan Chen, Yongzhe Ma, Han Xu, Jianglei Ma, Yichuan Guan, Shikang Fang, Husheng Shi, Zhenwei |
| author_facet | Li, Wenyuan Liang, Shunlin Chen, Keyan Chen, Yongzhe Ma, Han Xu, Jianglei Ma, Yichuan Guan, Shikang Fang, Husheng Shi, Zhenwei |
| contents | Accurate crop mapping fundamentally relies on modeling multi-scale spatiotemporal patterns, where spatial scales range from individual field textures to landscape-level context, and temporal scales capture both short-term phenological transitions and full growing-season dynamics. Transformer-based remote sensing foundation models (RSFMs) offer promising potential for crop mapping due to their innate ability for unified spatiotemporal processing. However, current RSFMs remain suboptimal for crop mapping: they either employ fixed spatiotemporal windows that ignore the multi-scale nature of crop systems or completely disregard temporal information by focusing solely on spatial patterns. To bridge these gaps, we present AgriFM, a multi-source remote sensing foundation model specifically designed for agricultural crop mapping. Our approach begins by establishing the necessity of simultaneous hierarchical spatiotemporal feature extraction, leading to the development of a modified Video Swin Transformer architecture where temporal down-sampling is synchronized with spatial scaling operations. This modified backbone enables efficient unified processing of long time-series satellite inputs. AgriFM leverages temporally rich data streams from three satellite sources including MODIS, Landsat-8/9 and Sentinel-2, and is pre-trained on a global representative dataset comprising over 25 million image samples supervised by land cover products. The resulting framework incorporates a versatile decoder architecture that dynamically fuses these learned spatiotemporal representations, supporting diverse downstream tasks. Comprehensive evaluations demonstrate AgriFM's superior performance over conventional deep learning approaches and state-of-the-art general-purpose RSFMs across all downstream tasks. Codes will be available at https://github.com/flyakon/AgriFM. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_21357 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | AgriFM: A Multi-source Temporal Remote Sensing Foundation Model for Agriculture Mapping Li, Wenyuan Liang, Shunlin Chen, Keyan Chen, Yongzhe Ma, Han Xu, Jianglei Ma, Yichuan Guan, Shikang Fang, Husheng Shi, Zhenwei Computer Vision and Pattern Recognition Machine Learning Accurate crop mapping fundamentally relies on modeling multi-scale spatiotemporal patterns, where spatial scales range from individual field textures to landscape-level context, and temporal scales capture both short-term phenological transitions and full growing-season dynamics. Transformer-based remote sensing foundation models (RSFMs) offer promising potential for crop mapping due to their innate ability for unified spatiotemporal processing. However, current RSFMs remain suboptimal for crop mapping: they either employ fixed spatiotemporal windows that ignore the multi-scale nature of crop systems or completely disregard temporal information by focusing solely on spatial patterns. To bridge these gaps, we present AgriFM, a multi-source remote sensing foundation model specifically designed for agricultural crop mapping. Our approach begins by establishing the necessity of simultaneous hierarchical spatiotemporal feature extraction, leading to the development of a modified Video Swin Transformer architecture where temporal down-sampling is synchronized with spatial scaling operations. This modified backbone enables efficient unified processing of long time-series satellite inputs. AgriFM leverages temporally rich data streams from three satellite sources including MODIS, Landsat-8/9 and Sentinel-2, and is pre-trained on a global representative dataset comprising over 25 million image samples supervised by land cover products. The resulting framework incorporates a versatile decoder architecture that dynamically fuses these learned spatiotemporal representations, supporting diverse downstream tasks. Comprehensive evaluations demonstrate AgriFM's superior performance over conventional deep learning approaches and state-of-the-art general-purpose RSFMs across all downstream tasks. Codes will be available at https://github.com/flyakon/AgriFM. |
| title | AgriFM: A Multi-source Temporal Remote Sensing Foundation Model for Agriculture Mapping |
| topic | Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2505.21357 |