摘要
城市的扩张改变了地表类型,越来越多的农田、水域等土地利用类型被水泥、沥青等不透水地面取代,易造成生态系统的破坏。估算不透水面对建设生态海绵城市有重要意义,该研究利用多源遥感数据,通过随机森林方法构建城市不透水面提取方法,并对广州典型试验区的不透水面进行提取与验证,结果表明其总体分类精度高于90%,在抑制错分和漏分时有良好效果,研究成果可为不透水面提取提供技术借鉴。
The expansion of cities has changed the surface type.Land types such as farmland and water areas have been replaced by impermeable surfaces such as cement asphalt,which can easily cause damage to the ecosystem.Estimating impermeable surfaces is of great significance in building ecological sponge cities.This study uses multi-source remote sensing data to construct an urban impervious surface extraction method through the random forest method,and extracts and verifies the impervious surface in a typical experimental area in Guangzhou.The results show that its overall classification accuracy is higher than 90%,and it has good effect in suppressing misclassification and missing classification.The research can provide technical reference for impermeable surface extraction.
作者
严传勇
YAN Chuanyong(Huizhou City Huayu Water Resources and Hydropower Engineering Survey and Design Co.,Ltd.,Huizhou 516000,China)
出处
《广东水利水电》
2025年第4期92-96,共5页
Guangdong Water Resources and Hydropower
关键词
不透水面
随机森林
遥感
impermeable surface
random forest
remote sensing
