Integrating Phenological Priors with Deep Spatio-Temporal Features for tree species mapping
Mapping large-scale tree species distributions is essential for accurately estimating forest carbon storage. Previous studies have shown that Satellite Image Time Series (SITS) can be effective for classifying tree species. However, many of these studies rely heavily on manual feature engineering or...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2025-07-01
|
| Series: | ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
| Online Access: | https://isprs-annals.copernicus.org/articles/X-G-2025/559/2025/isprs-annals-X-G-2025-559-2025.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849704569733382144 |
|---|---|
| author | Z. Ma N. Zhu Z. Dong R. Chen B. Yang Z. Chen C. Long R. Ding |
| author_facet | Z. Ma N. Zhu Z. Dong R. Chen B. Yang Z. Chen C. Long R. Ding |
| author_sort | Z. Ma |
| collection | DOAJ |
| description | Mapping large-scale tree species distributions is essential for accurately estimating forest carbon storage. Previous studies have shown that Satellite Image Time Series (SITS) can be effective for classifying tree species. However, many of these studies rely heavily on manual feature engineering or overlook critical geoscientific and forestry knowledge. Such domain-specific insights are particularly important in Earth observation because the same species can exhibit diverse spatio-temporal behaviors across different regions, leading to lower accuracy and limited model robustness. In this work, we propose a novel model, PTSViT, which integrates phenological information with deep spatio-temporal features to address these limitations. Our model’s loss function incorporates phenological priors, utilizing ground-based phenological data and tree species labels as supervisory signals to guide the learning of spatio-temporal encoders. We evaluate PTSViT on a newly created dataset, GXData, which includes 11 major tree species in GuangXi. Our model surpasses previous approaches across all evaluation metrics, demonstrating the value of integrating prior knowledge for automated, accurate tree species mapping. |
| format | Article |
| id | doaj-art-861ea4dab61a45cfa7347db5b8595944 |
| institution | DOAJ |
| issn | 2194-9042 2194-9050 |
| language | English |
| publishDate | 2025-07-01 |
| publisher | Copernicus Publications |
| record_format | Article |
| series | ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
| spelling | doaj-art-861ea4dab61a45cfa7347db5b85959442025-08-20T03:16:43ZengCopernicus PublicationsISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences2194-90422194-90502025-07-01X-G-202555956610.5194/isprs-annals-X-G-2025-559-2025Integrating Phenological Priors with Deep Spatio-Temporal Features for tree species mappingZ. Ma0N. Zhu1Z. Dong2R. Chen3B. Yang4Z. Chen5C. Long6R. Ding7LIESMARS, Wuhan University, Wuhan, ChinaLIESMARS, Wuhan University, Wuhan, ChinaLIESMARS, Wuhan University, Wuhan, ChinaGuangxi Zhuang Autonomous Region Institute of Natural Resources Remote Sensing, Nanning, ChinaLIESMARS, Wuhan University, Wuhan, ChinaLIESMARS, Wuhan University, Wuhan, ChinaLIESMARS, Wuhan University, Wuhan, ChinaLIESMARS, Wuhan University, Wuhan, ChinaMapping large-scale tree species distributions is essential for accurately estimating forest carbon storage. Previous studies have shown that Satellite Image Time Series (SITS) can be effective for classifying tree species. However, many of these studies rely heavily on manual feature engineering or overlook critical geoscientific and forestry knowledge. Such domain-specific insights are particularly important in Earth observation because the same species can exhibit diverse spatio-temporal behaviors across different regions, leading to lower accuracy and limited model robustness. In this work, we propose a novel model, PTSViT, which integrates phenological information with deep spatio-temporal features to address these limitations. Our model’s loss function incorporates phenological priors, utilizing ground-based phenological data and tree species labels as supervisory signals to guide the learning of spatio-temporal encoders. We evaluate PTSViT on a newly created dataset, GXData, which includes 11 major tree species in GuangXi. Our model surpasses previous approaches across all evaluation metrics, demonstrating the value of integrating prior knowledge for automated, accurate tree species mapping.https://isprs-annals.copernicus.org/articles/X-G-2025/559/2025/isprs-annals-X-G-2025-559-2025.pdf |
| spellingShingle | Z. Ma N. Zhu Z. Dong R. Chen B. Yang Z. Chen C. Long R. Ding Integrating Phenological Priors with Deep Spatio-Temporal Features for tree species mapping ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
| title | Integrating Phenological Priors with Deep Spatio-Temporal Features for tree species mapping |
| title_full | Integrating Phenological Priors with Deep Spatio-Temporal Features for tree species mapping |
| title_fullStr | Integrating Phenological Priors with Deep Spatio-Temporal Features for tree species mapping |
| title_full_unstemmed | Integrating Phenological Priors with Deep Spatio-Temporal Features for tree species mapping |
| title_short | Integrating Phenological Priors with Deep Spatio-Temporal Features for tree species mapping |
| title_sort | integrating phenological priors with deep spatio temporal features for tree species mapping |
| url | https://isprs-annals.copernicus.org/articles/X-G-2025/559/2025/isprs-annals-X-G-2025-559-2025.pdf |
| work_keys_str_mv | AT zma integratingphenologicalpriorswithdeepspatiotemporalfeaturesfortreespeciesmapping AT nzhu integratingphenologicalpriorswithdeepspatiotemporalfeaturesfortreespeciesmapping AT zdong integratingphenologicalpriorswithdeepspatiotemporalfeaturesfortreespeciesmapping AT rchen integratingphenologicalpriorswithdeepspatiotemporalfeaturesfortreespeciesmapping AT byang integratingphenologicalpriorswithdeepspatiotemporalfeaturesfortreespeciesmapping AT zchen integratingphenologicalpriorswithdeepspatiotemporalfeaturesfortreespeciesmapping AT clong integratingphenologicalpriorswithdeepspatiotemporalfeaturesfortreespeciesmapping AT rding integratingphenologicalpriorswithdeepspatiotemporalfeaturesfortreespeciesmapping |