Towards Onboard Continuous Change Detection for Floods
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908899084337152 |
|---|---|
| author | Kyselica, Daniel Herec, Jonáš Kutis, Oliver Pitoňák, Rado |
| author_facet | Kyselica, Daniel Herec, Jonáš Kutis, Oliver Pitoňák, Rado |
| contents | Natural disaster monitoring through continuous satellite observation requires processing multi-temporal data under strict operational constraints. This paper addresses flood detection, a critical application for hazard management, by developing an onboard change detection system that operates within the memory and computational limits of small satellites. We propose History Injection mechanism for Transformer models (HiT), that maintains historical context from previous observations while reducing data storage by over 99\% of original image size. Moreover, testing on the STTORM-CD flood dataset confirms that the HiT mechanism within the Prithvi-tiny foundation model maintains detection accuracy compared to the bi-temporal baseline. The proposed HiT-Prithvi model achieved 43 FPS on Jetson Orin Nano, a representative onboard hardware used in nanosats. This work establishes a practical framework for satellite-based continuous monitoring of natural disasters, supporting real-time hazard assessment without dependency on ground-based processing infrastructure. Architecture as well as model checkpoints is available at https://github.com/zaitra/HiT-change-detection . |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_13751 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Towards Onboard Continuous Change Detection for Floods Kyselica, Daniel Herec, Jonáš Kutis, Oliver Pitoňák, Rado Computer Vision and Pattern Recognition Machine Learning 68T07 I.2.10; I.4.6; I.4.10 Natural disaster monitoring through continuous satellite observation requires processing multi-temporal data under strict operational constraints. This paper addresses flood detection, a critical application for hazard management, by developing an onboard change detection system that operates within the memory and computational limits of small satellites. We propose History Injection mechanism for Transformer models (HiT), that maintains historical context from previous observations while reducing data storage by over 99\% of original image size. Moreover, testing on the STTORM-CD flood dataset confirms that the HiT mechanism within the Prithvi-tiny foundation model maintains detection accuracy compared to the bi-temporal baseline. The proposed HiT-Prithvi model achieved 43 FPS on Jetson Orin Nano, a representative onboard hardware used in nanosats. This work establishes a practical framework for satellite-based continuous monitoring of natural disasters, supporting real-time hazard assessment without dependency on ground-based processing infrastructure. Architecture as well as model checkpoints is available at https://github.com/zaitra/HiT-change-detection . |
| title | Towards Onboard Continuous Change Detection for Floods |
| topic | Computer Vision and Pattern Recognition Machine Learning 68T07 I.2.10; I.4.6; I.4.10 |
| url | https://arxiv.org/abs/2601.13751 |