物探化探计算技术2025，Vol.47Issue(3)：466-474,9.DOI:10.12474/wthtjs.20240314-0007
基于高分辨率遥感影像的城市水体语义分割和制图
Semantic segmentation and mapping of urban water bodies based on high-resolution remote sensing images:a case study in Chengdu,China
摘要
Abstract
Urban water bodies are critically crucial for urban planning and development,underscoring the need for rapid and precise acquisition of urban water body information.Typically,the large-scale water body extraction is achieved using low-to medium-resolution remote sensing images.However,detecting edges and extracting smaller water bodies necessitate high-resolution remote sensing images.For this purpose,a high-resolution remote sensing image dataset has been developed for urban water bodies in Chengdu.The dataset employs a semantic segmentation approach using the SegFormer model,which integrates an attention mechanism.Experimental results validate the efficacy of combining this dataset with the model for large-scale extraction of urban water body information,achieving extraction accuracy rates exceeding 96%.This study demonstrates potential applications in extracting and mapping water bodies across other extensive regions,serving as a benchmark for different methods that utilize high-resolution remote sensing images to extract large-scale water body information.关键词
城市水体/高分辨率遥感/SegFormer/大规模Key words
urban water body/high-resolution remote sensing/SegFormer/large-scale分类
天文与地球科学引用本文复制引用
王婷婷,程熙,李斌,王潆晨,宋泽毅,孟丽康..基于高分辨率遥感影像的城市水体语义分割和制图[J].物探化探计算技术,2025,47(3):466-474,9.基金项目
火灾感知软件集成及预警模型功能测试项目(80303-AH20230088) （80303-AH20230088）
