地理空间信息2025，Vol.23Issue(9)：22-26,84,6.DOI:10.3969/j.issn.1672-4623.2025.09.006
耦合机器学习与连通性修正的河流选取模型
River Selection Model Coupled with Machine Learning and Connectivity Correction
摘要
Abstract
The traditional river selection methods are prone to problems such as low efficiency when handling large datasets.Therefore,taking river data from Henan Province and Ningxia Hui Autonomous Region for example,we used support vector machine(SVM)model to initially select river elements,and introduced improved Dijkstra algorithm to construct the river network paths.Then,we proceeded connectivity correction on the initially selected results to achieve the automatic selection of river elements.The results show that the SVM model effectively simplifies river elements,outperforming the random forest algorithm in terms of efficiency and accuracy in identifying high-priority rivers.The improved Dijkstra algorithm addresses issues such as river discontinuities in the initial selection and significantly reduces the time required for constructing river network paths in large datasets.The integration of machine learning and connectivity correction maintains the continuity and integrity of rivers to a certain extent,meeting the cartographic needs at different scales.关键词
SVM/Dijkstra算法/河流选取/线要素Key words
SVM/Dijkstra algorithm/river selection/line element分类
天文与地球科学引用本文复制引用
车一鸣,邓维熙,冯飞军,张静静,卢文渊,任东宇..耦合机器学习与连通性修正的河流选取模型[J].地理空间信息,2025,23(9):22-26,84,6.基金项目
宁夏自然科学基金资助项目(2023AAC03572) （2023AAC03572）
宁夏优秀人才支持计划资助项目. （）
