摘要
水利行业对于整合"天-空-地"一体化遥感监测技术的应用需求越来越高,然而传统的遥感影像分析模型往往难以适应日益多样化的行业应用需求.因此,通过集成分布式存储、虚拟化管理、自动化运维和云安全等技术,以PaaS和SaaS方式提供云服务,构建了一个融合先进人工智能技术、面向行业应用的智能化遥感影像数据处理云平台HydrSAI.该平台集成了影像预处理、样本采集与建库、模型构建与优化、高性能矢量切片以及云计算服务技术,并配备可扩展的影像库、样本库和模型库.该平台在深圳市人为扰动图斑识别以及大藤峡库区开展了试验应用,其识别效率相较于目视识别得到显著提升.HydrSAI可为水利部门提供高效、便捷、智能化的遥感解译支持.
Abstract
The water conservancy industry has an increasing demand for"air-space-ground"integrated remote sensing moni-toring technology.However,traditional image analysis models often struggle to adapt to the growing diversity of applications.To ad-dress this,our goal is to integrate technologies such as distributed storage,virtual management,automated operation and mainte-nance,and cloud security,to offer cloud services in the form of PaaS and SaaS.The objective is to establish an intelligent image-processing cloud platform called HydrSAI that incorporates advanced artificial intelligence technology and is oriented to industrial applications.This platform integrates image preprocessing,sample collection,specific database construction,model building and optimization,high-performance vector slicing,and cloud computing services,and is equipped with scalable image,sample,and model libraries.This platform was applied to identify artificially disturbed patches in Shenzhen City,and Datengxia reservoir area,showing much better interpretation effect compared to visual interpretation.The HydrSAI serves as an intelligent cloud service platform,providing efficient and convenient image interpretation toolsets for the water conservancy applications.
基金项目
国家自然科学基金项目(32371966)
数字孪生大藤峡建设(一期)项目(DXH2022012)
粤港澳大湾区水安全要素"空-天-地"立体观测关键技术项目(SKR-2022037)
&&(2022YF009)
国家科技期刊平台
NSTL
万方数据