张茂省,贾俊,王毅,牛千,毛伊敏,董英.基于人工智能(AI)的地质灾害防控体系建设[J].西北地质,2019,52(2):103-116
基于人工智能(AI)的地质灾害防控体系建设
Construction of Geological Disaster Prevention and Control System Based on AI
投稿时间:2019-02-04  修订日期:2019-03-26
DOI:10.19751/j.cnki.61-1149/p.2019.02.011
中文关键词:  大数据;人工智能;Python;智能防灾减灾;地质灾害;防控体系
英文关键词:big data;artificial intelligence;Python;intelligent disaster prevention and mitigation;geological disasters;prevention and control system
基金项目:国家重点研发项目(2018YFC1504700),国家自然科学基金重点项目(41641011、41530640),中国地质调查局项目(DD20160261、DD20189270、DDT20190463)联合资助
作者单位
张茂省 中国地质调查局西安地质调查中心/西北地质科技创新中心, 自然资源部黄土地质灾害重点实验室, 陕西 西安 710054 
贾俊 中国地质调查局西安地质调查中心/西北地质科技创新中心, 自然资源部黄土地质灾害重点实验室, 陕西 西安 710054 
王毅 西北大学, 陕西 西安 710069 
牛千 中国地质调查局西安地质调查中心/西北地质科技创新中心, 自然资源部黄土地质灾害重点实验室, 陕西 西安 710054 
毛伊敏 江西理工大学, 江西 赣州 341000 
董英 中国地质调查局西安地质调查中心/西北地质科技创新中心, 自然资源部黄土地质灾害重点实验室, 陕西 西安 710054 
摘要点击次数: 266
全文下载次数: 288
中文摘要:
      近年来,人工智能(AI)技术得到快速发展,并日益融入到经济社会各个领域,成为当代创新发展的新标志,智能防灾减灾将成为未来发展的趋势和研究的热点。在回顾AI发展现状与趋势的基础上,系统梳理出以往地质灾害风险防控的数据依据和传统技术方法,分析了可能采用的潜在AI方法,初步搭建了基于AI的地质灾害风险防控体系建设方案。研究表明,AI技术为地质灾害风险防控提供了新的技术途径,但目前尚无可照搬或可移植的成熟技术或解决方案。智能防灾减灾体系包括早期识别、风险评估、风险防控等3个主要环节,其中最重要的环节是早期识别,传统方法与AI技术融合的关键参数为斜坡失稳概率或泥石流发生概率;根据所依据的数据资料将早期识别方法归纳为图像识别、形变识别、位移识别、内因识别、诱因识别和综合识别等6种方法;提出了从数据层、方法层和应用层3个层次构建基于大数据智能混合优化的地质灾害风险防控平台。认为数据驱动的智能模型与理论驱动的物理模型融合是地质灾害风险防控发展的趋势。
英文摘要:
      In recent years, artificial intelligence (AI) technology has developed rapidly and is increasingly integrated into various fields of economy and society. It has become a new symbol of contemporary innovation and development. Intelligent disaster prevention and mitigation will become the future development trend and research hotspot. On the basis of reviewing the status and trend of (AI) development, this paper systematically sorts out the data basis and traditional technical methods of previous geological disaster risk prevention and control, analyzes the potential AI methods that may be used, and initially builds a an AI-based geological disaster risk prevention and control system. Research shows that AI technology provides a new technical approach for the prevention and control of geological disaster risks, but there are no mature technologies or solutions that can be copied or portable. The proposed intelligent disaster prevention and mitigation system includes three main parts:early identification, risk assessment, and risk prevention and control. The key point is early identification, while the key parameter for the fusion between traditional methods and AI technology is the probability of slope instability or the probability of derrs flow occurrence; According to the data, the early identification methods are summarized into six methods:image recognition, deformation recognition, displacement recognition, internal factor recognition, incentive identification and comprehensive recognition. This paper proposes a geological disaster risk prevention and control platform based on big data intelligent hybrid optimization from three levels:data layer, method layer and application layer. It is believed that the fusion between data-driven intelligent models and theoretically driven physical models is the trend of geological disaster risk prevention and control development.
查看全文  查看/发表评论  下载PDF阅读器
关闭