基于K-means聚类分析处理遥感影像的河湖“四乱”因素识别

Identification of"four illegal actions"on rivers and lakes based on K-means cluster analysis for processing remote sensing images

2018年水利部部署开展河湖“清四乱”专项行动,为提高河湖乱占、乱采、乱堆、乱建(“四乱”)问题的识别效率,应用K-means聚类分析方法对卫星遥感影像进行检测,分析了2020~2022年上海市26条(总长649 km)河道与淀山湖、元荡湖两个湖泊的“四乱”问题。结果表明:2020~2022年,28个河湖河口线外延6 m范围内有274处变化,其中疑似“四乱”问题82处。相比同期人工巡查,基于K-means聚类分析的遥感影像因素识别技术在识别河湖“四乱”中具有巡查效率高、巡查范围广、巡查成本低等优势。研究成果可以为下阶段河湖治理和保护监管以及信息化建设提供参考。

In 2018,the Ministry of Water Resources proposed the special action to clear the"four illegal actions"in rivers and lakes.In order to improve the efficiency of identifying factors related to illegal occupation,mining,stacking,and construction on rivers and lakes,K-means clustering analysis method was applied to detect satellite remote sensing images.The"four illegal actions"problem of 26 rivers with a total length of 649 km and two rivers of Dianshan Lake and Yuandang Lake in Shanghai from 2020 to 2022 were analyzed.The results showed that from 2020 to 2022,there were 274 changes within 6 m of 28 rivers and lakes,including 82 suspected"four illegal actions"problems.Compared with manual inspections during the same period,remote sensing image factor recognition technology based on K-means clustering analysis had advantages such as high inspection efficiency,wide inspection scope,and low inspection cost in identifying the"four illegal actions"problems in rivers and lakes.The research results can provide a reference for the next stage of rivers and lakes management,protection and supervision,as well as information construction.