面向洪水的 onboard 持续变化检测
Towards Onboard Continuous Change Detection for Floods
AI总结：
本文提出HiT机制，用于小卫星上的持续洪水检测，通过减少数据存储并保持检测精度，实现高效的 onboard 处理。
摘要：
通过连续卫星观测进行自然灾害监测需要在严格操作约束下处理多时相数据。本文通过开发一个在小卫星内存和计算限制内运行的变化检测系统，解决洪水检测问题，这是灾害管理的关键应用。我们提出了适用于Transformer模型的History Injection机制（HiT），该机制在保持历史上下文的同时，将数据存储减少超过原始图像大小的99%。此外，在STTORM-CD洪水数据集上的测试表明，HiT机制在Prithvi-tiny基础模型中与双时相基线保持相同的检测精度。所提出的HiT-Prithvi模型在Jetson Orin Nano上实现了43 FPS。本文建立了一个实用的卫星连续自然灾害监测框架，支持实时灾害评估，而无需依赖地面处理基础设施。架构及模型检查点可在https://github.com/zaitra/HiT-change-detection获取。
英文摘要：
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 .
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
