期刊
FRONTIERS IN WATER
卷 4, 期 -, 页码 -出版社
FRONTIERS MEDIA SA
DOI: 10.3389/frwa.2022.786040
关键词
artificial intelligence of things; flood analytics information system; at-risk locations; flood image processing; convolutional neural networks
Floods can cause severe damage to lives, properties, infrastructure, and the environment. The combination of artificial intelligence, big data, and the Internet of Things has the potential to predict floods more accurately and advance flood analytics research. This research discusses the workflow of a Flood Analytics Information System (FAIS) as an example of an AIoT prototype to drive the next generation of flood informatics systems.
Floods are among the most destructive natural hazards that affect millions of people across the world leading to severe loss of life and damage to properties, critical infrastructure, and the environment. The combination of artificial intelligence (AI), big data, and the Internet of Things (IoTs), has the potential to more accurately predict these extreme events and accelerate the convergence of advanced techniques for flood analytics research. This convergence-so called the Artificial Intelligence of Things (AIoT)-is transformational for both technologies and science-based decision making since AI adds value to IoT through interpretable machine learning (ML) while IoT leverages the power of AI via connectivity and data intelligence. The aim of this research is to discuss the workflow of a Flood Analytics Information System (FAIS; version 4.00) as an example of AIoT prototype to advance and drive the next generation of flood informatics systems. FAIS integrates crowd intelligence, ML, and natural language processing (NLP) to provide flood warning with the aim of improving flood situational awareness and risk assessments. Various image processing algorithms, i.e., Convolutional Neural Networks (CNNs), were also integrated with the FAIS prototype for image label detection, and floodwater level and inundation areas calculation. The prototype successfully identifies a dynamic set of at-risk locations/communities using the USGS river gauge height readings and geotagged tweets intersected with watershed boundary. The list of prioritized locations can be updated, as the river monitoring system and condition change over time (typically every 15 min). The prototype also performs flood frequency analysis (FFA) by fitting multiple probability distributions to the annual flood peak rates and calculates the uncertainty associated with the model. FAIS was operationally tested (beta-tested) during multiple hurricane driven floods in the US and was recently released as a national-scale flood data analytics pipeline.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据