首页|基于GIS服务的遥感影像智能识别系统研究与应用

基于GIS服务的遥感影像智能识别系统研究与应用

扫码查看
为解决水利行业遥感影像智能识别系统样本制作效率低、影像检测成果使用和管理不便、用户体验差、缺乏可靠水安全地形要素 AI 智能提取系统的问题,开展基于 GIS 服务的遥感影像智能识别系统研究.基于 WebGIS、数据库、大数据、互联网、人工智能等技术,研究构建遥感影像大数据管理与发布、样本数据集自动化生产与管理系统.在 U-Net 模型的基础上引入影像预处理和栅格转矢量模块,添加 2 个残差块以加深网络,从而得到 D-Unet模型,并用于地形要素提取.结果表明:基于 GIS 服务的遥感影像智能识别系统能够实现对多源卫星、无人机遥感影像的高效管理,以及样本数据集生产的自动化,在地形要素识别方面具有较高精度,且具有较好的用户体验.
Research and application of remote sensing image intelligent identification system based on GIS services
Aiming to address the problems of low efficiency in sample production,inconvenient usage and management of image detection results,poor user experience,and the lack of a reliable water safety topographic feature AI intelligent extraction system in the water conservancy industry,this study conducts research on a remote sensing image intelligent identification system based on GIS services.Drawing upon technologies such as WebGIS,databases,big data,the internet,and artificial intelligence,this research constructs a system for the management and release of large remote sensing image data,and an automated production and management system for sample datasets.By integrating image preprocessing and raster-to-vector conversion modules into the U-Net model,and adding two residual blocks to deepen the network,a D-Unet model is developed for topographic feature extraction.The results indicate that the remote sensing image intelligent identification system based on GIS services can efficiently manage multi-source satellite and drone remote sensing images and automate the production of sample datasets.It achieves high accuracy in topographic feature identification and offers a good user experience.

GIS serviceremote sensing imageintelligent identificationdeep learningsample productionD-Unet

唐宏、罗丹、朱长富

展开 >

中水珠江规划勘测设计有限公司,广东 广州 510610

广西右江水利开发有限责任公司,广西 南宁 530029

GIS服务 遥感影像 智能识别 深度学习 样本生产 D-Unet

2024

水利信息化
水利部南京水利水文自动化研究所

水利信息化

影响因子:0.571
ISSN:1674-9405
年,卷(期):2024.(5)