基于遥感影像的土地利用特征提取与城乡梯度差异分析——以河北省涿州市为例

Diversity analysis of urban and rural land use based on Sentinel-2 remote sensing image: A case study of Zhuozhou City in Hebei Province

关键词:  土地利用,面向对象分类,隶属度函数,决策树分类,遥感影像

Key words:  land use  object-oriented classification  subjection function  decision tree classifier  Sentinel-2 remote sensing image

摘要:

为快速获取区域土地利用特征和精细刻画城乡土地利用差异，以河北省涿州市为研究对象，基于Sentinel-2 影像数据，采取面向对象方法进行影像分割，利用隶属度函数与决策树方法相结合的非监督分类算法对涿州市土地利用进行分类，并选取了不同方向的城乡梯度样带进行了土地利用特征分析。结果表明，应用模糊决策树方法的涿州市土地利用分类结果总体精度为93.7%，Kappa系数0.892，分类精度较高。分析上述结果发现:涿州市土地利用类型以耕地与城乡居民点用地为主，林地、草地、水体等自然生态空间比例较低，土地利用的城乡梯度特征明显；耕地集中分布在距离城市中心4~7 km的东南、南、西方向；城乡居民点整体分布分散，在距离城市中心3 km以内、5 km、8~9 km呈现明显的集聚特征。建议涿州市依据预期人口规模和集聚特征优化建设用地布局，提高建设用地集约利用强度，同时提高林地、草地、水体等生态空间比例。

Abstract:

In order to quickly get the regional land use characteristics and describe the urban and rural land use differences, this study adopts object-oriented method for image segmentation, and USES unsupervised classification algorithm combining membership function and decision tree methods to classify the land use of Zhuozhou City based on Sentinel-2 image data, and selects the different directions of urban-rural gradient belt to analysis the characteristics of land use. The results show that the overall accuracy of the classification method is 93. 7%, the Kappa coefficient is 0. 892, and the classification accuracy is relatively high. Based on the classification results, it is found that Zhuozhou's land use types are mainly farmland and urban and rural residential land, while the proportions of natural ecological space such as forest land, grassland and water body are relatively low. The spatial distribution shows an obvious urban-rural gradient characteristics. The cultivated land is concentrated in the southeast, south and west directions 4-7 km away from the city center. The overall distribution of urban and rural residential areas is scattered, with certain clustering characteristics within 3, 5, and 8-9 km from the city center. It is suggested Zhuozhou should optimize the layout of construction land according to the expected population size and land use agglomeration characteristics, improve the intensity of intensive utilization of construction land, and increase the proportion of ecological space such as forest land, grassland and water body.