TEA: Temporal Adaptive Satellite Image Semantic Segmentation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917190612025344 |
|---|---|
| author | Kang, Juyuan Zhu, Hao Zhu, Yan Zhang, Wei Chen, Jianing Xiao, Tianxiang Ma, Yike Jiang, Hao Dai, Feng |
| author_facet | Kang, Juyuan Zhu, Hao Zhu, Yan Zhang, Wei Chen, Jianing Xiao, Tianxiang Ma, Yike Jiang, Hao Dai, Feng |
| contents | Crop mapping based on satellite images time-series (SITS) holds substantial economic value in agricultural production settings, in which parcel segmentation is an essential step. Existing approaches have achieved notable advancements in SITS segmentation with predetermined sequence lengths. However, we found that these approaches overlooked the generalization capability of models across scenarios with varying temporal length, leading to markedly poor segmentation results in such cases. To address this issue, we propose TEA, a TEmporal Adaptive SITS semantic segmentation method to enhance the model's resilience under varying sequence lengths. We introduce a teacher model that encapsulates the global sequence knowledge to guide a student model with adaptive temporal input lengths. Specifically, teacher shapes the student's feature space via intermediate embedding, prototypes and soft label perspectives to realize knowledge transfer, while dynamically aggregating student model to mitigate knowledge forgetting. Finally, we introduce full-sequence reconstruction as an auxiliary task to further enhance the quality of representations across inputs of varying temporal lengths. Through extensive experiments, we demonstrate that our method brings remarkable improvements across inputs of different temporal lengths on common benchmarks. Our code will be publicly available. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_04956 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | TEA: Temporal Adaptive Satellite Image Semantic Segmentation Kang, Juyuan Zhu, Hao Zhu, Yan Zhang, Wei Chen, Jianing Xiao, Tianxiang Ma, Yike Jiang, Hao Dai, Feng Computer Vision and Pattern Recognition Crop mapping based on satellite images time-series (SITS) holds substantial economic value in agricultural production settings, in which parcel segmentation is an essential step. Existing approaches have achieved notable advancements in SITS segmentation with predetermined sequence lengths. However, we found that these approaches overlooked the generalization capability of models across scenarios with varying temporal length, leading to markedly poor segmentation results in such cases. To address this issue, we propose TEA, a TEmporal Adaptive SITS semantic segmentation method to enhance the model's resilience under varying sequence lengths. We introduce a teacher model that encapsulates the global sequence knowledge to guide a student model with adaptive temporal input lengths. Specifically, teacher shapes the student's feature space via intermediate embedding, prototypes and soft label perspectives to realize knowledge transfer, while dynamically aggregating student model to mitigate knowledge forgetting. Finally, we introduce full-sequence reconstruction as an auxiliary task to further enhance the quality of representations across inputs of varying temporal lengths. Through extensive experiments, we demonstrate that our method brings remarkable improvements across inputs of different temporal lengths on common benchmarks. Our code will be publicly available. |
| title | TEA: Temporal Adaptive Satellite Image Semantic Segmentation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2601.04956 |