SCC-PG: Intelligent Flood Detection and Warning System to Assist Homeless Communities and Emergency Management Entities

无家可归的人以惊人的速度增长，特别是在圣地亚哥等西海岸城市。没有避难所的人比一般人面临更高的洪水风险，因为他们更有可能生活在最容易受到洪水影响的地方，并且与现有的洪水预警系统断开连接。这种困境造成了不平等的灾害和环境影响，使最脆弱的人在关键时刻获得的信息最少。从技术方面来看，缺乏智能洪水探测系统会导致预测不准确和响应延迟。该项目旨在通过将图像数据集成到当前最先进的洪水数据采集和城市基础设施建模中来克服这些挑战。该研究的更广泛影响包括提高社区对使用新智能技术的认识，并通过让居民，企业和从业人员参与研究计划的开发来支持积极的洪水和应急管理。参与的学生将有机会通过参与团队合作学习多学科的研究课题。新的智能系统和连接的决策框架将转移到美国西部和东部沿海的其他洪水易发社区，这些社区由于气候变化而面临更频繁的洪水，同时也面临前所未有的住房负担能力危机。该项目将创建一个由研究人员，从业人员和相关公民组成的多元化，多学科的社区，以开发新技术和综合理论和方法，以提高洪水数据采集，检测和监测的时间和准确性。将开发一套基于深度学习的图像处理系统，并使用大型完全分割的洪水图像数据集进行训练，以检测洪水形成和监测洪水事件。将开发一个洪水模型，以模拟和预报实时洪水及其影响。该项目将提高洪水预警和监测系统的响应时间和准确性，以支持智能和互联的沿海社区的复原力和应急管理。在规划阶段，将组建并召集一个社区咨询小组，就智能防洪系统的开发提供咨询意见，并指导人类主体的数据收集活动，包括与无庇护者进行结构化访谈。该项目增强了支持智能和互联社区所需的知识，包括1）来自各种信息来源的新颖数据收集技术，如地面摄像机，2）基于人工智能的洪水视觉传感和分析各种数据，和3）该奖项反映了NSF的法定使命，并被认为值得支持通过使用基金会的知识价值和更广泛的影响审查标准进行评估。

Unsheltered homelessness has grown at staggering rates, particularly across West Coast cities such as San Diego. Unsheltered people are at higher risk than the general population of experiencing flooding risk, as they are more likely both to be living in the most flood-vulnerable locations as well as disconnected from existing flood warning systems. This predicament results in unequal disasters and environmental impact, burdening the most vulnerable people with the least information in critical moments. From a technical aspect, the absence of an intelligent system for flood detection causes inaccuracy in prediction and delays in responses. This project aims to overcome these challenges by integrating imagery data into the current state-of-the-art flood data acquisition and urban infrastructure modeling. The broader impact of the study involves raising community awareness about the use of new smart technologies and supporting proactive flood and emergency management by engaging residents, businesses, and practitioners in the development of the research program. Involved students will have the opportunity to learn multi-disciplinary research topics through engaging in collaborative teamwork. The new smart system and connected decision-making framework will be transferable to other flood-prone communities across the U.S. West and East Coasts that are confronting more frequent floods as a result of climate change coinciding with an unprecedented housing affordability crisis. This project will create a diverse, multidisciplinary community of researchers, practitioners, and concerned citizens to develop novel technologies and integrated theories and methods to improve the time and accuracy of flood data acquisition, detection, and monitoring. A set of deep learning-based image processing systems will be developed and trained using large fully segmented flood image datasets to detect formation and monitor flood events. A flood model will be developed to simulate and forecast real-time floods and impacts. This project will enhance the response time and accuracy of flood early warning and monitoring systems to support the resilience and emergency management of smart and connected coastal communities. During the planning phase, a Community Advisory Group will be assembled and convened to advise on the development of the intelligent flood system and guide the human subject’s data collection activities, including structured interviews with unsheltered people. This project enhances the knowledge needed to support smart and connected communities by including 1) novel data collection techniques from various sources of information, such as ground-based cameras, 2) artificial intelligence-based flood visual sensing and analyzing diverse data, and 3) substantive community engagement that centers the needs of unsheltered people.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.