摘要
针对我国大城市居住用地的地块级精细结构和宏观空间分布格局,该文提出了一种区域尺度的城市居住用地遥感提取方法。基于Swin Transformer构建UPerNet神经网络,并结合后处理方法构建城市居住用地提取模型。利用资源三号遥感影像,对全国36个重点城市开展居住用地提取,同时进行景观格局分析以及与人口数量、房价等社会经济因子的相关性分析。结果表明,提出的方法能够准确地提取城市居住用地,F1值为84.43%,交并比为73.05%;居住用地面积在选取城市中的平均占比为22.76%,经济发达城市居住用地面积占比较低;部分城市存在过度发展房地产的现象。该方法能够应用于城市居住用地的微观结构监测与宏观差异比较,对优化城市居住用地空间布局、调整住房供给结构具有支撑作用。
The paper proposed a regional-scale remote sensing extraction method for urban residential area in China's large cities.The UPerNet neural network is constructed based on Swin Transformer and combined with post-processing methods to build an urban residential area extraction model.Using the remote sensing images of ZY-3 to extract residential area in 36 key cities across the country,and the landscape pattern analysis and the correlation analysis between residential area and socio-economic factors such as the number of population and housing price are carried out at the same time.The results show that the proposed method can accurately extract urban residential area with an F1 value of 84.43%and an intersection over union of 73.05%;the average proportion of residential area in the selected cities is 22.76%,and the area of residential land in economically developed cities is relatively low;some cities have the phenomenon of over-development of real estate.The method can be applied to the monitoring of the microstructure of urban residential area and the comparison of macroscopic differences,which is supportive to the optimization of the spatial layout of urban residential area and the adjustment of the housing supply structure.
作者
白佳玮
王光辉
陆尘
闫志刚
杜虓宇
BAI Jiawei;WANG Guanghui;LU Chen;YAN Zhigang;DU Xiaoyu(Key Laboratory of Land Environment and Disaster Monitoring Ministry of Natural Resources,China University of Mining and Technology,Xuzhou,Jiangsu 221l16,China;School of Environment and Spatial Informatics,China University of Mining and Technology,Xuzhou.Jiangsu 221116,China;Land Satellite Remoting Sensing Application Center,MNR,Beijing 100048,China)
出处
《测绘科学》
CSCD
北大核心
2023年第5期92-103,共12页
Science of Surveying and Mapping
基金
自然资源部国土环境与灾害监测重点实验室开放基金项目(LEDM2021B08)
国家自然科学基金项目(41971370)。
关键词
城市居住用地
高分遥感影像
语义分割
景观格局
社会经济因子
Urban residential area
high-resolution remote sensing imagery
semantic segmentation
landscape pattern
socioeconomic factors
