热带地理 ›› 2020, Vol. 40 ›› Issue (1): 110-118.doi: 10.13284/j.cnki.rddl.003168

• 论文 • 上一篇    下一篇

基于多源数据的主体功能区划分方法——以广州市为例

罗伟玲, 王艳阳(), 张恒   

  1. 广东国地规划科技股份有限公司,广州 510650
  • 收稿日期:2019-05-01 修回日期:2019-07-02 出版日期:2020-01-10 发布日期:2020-02-24
  • 通讯作者: 王艳阳 E-mail:yanyangwang@163.com
  • 作者简介:罗伟玲(1981—),女,湖南省长沙人,高级工程师,硕士,主要从事空间规划及大数据等方面的研究工作,(E-mail)289061973@qq.com 。

A Novel Method of Division Major Function Oriented Zoning Using Multi-Source Data in Guangzhou, China

Luo Weiling, Wang Yanyang(), Zhang Heng   

  1. Guangdong Guodi Planning Technology Co., Ltd., Guangzhou 510650, China
  • Received:2019-05-01 Revised:2019-07-02 Online:2020-01-10 Published:2020-02-24
  • Contact: Wang Yanyang E-mail:yanyangwang@163.com

摘要:

以广州市为例,基于城市感知数据、遥感影像等多源数据,采用卷积神经网络对遥感影像进行了语义提取,将提取结果与兴趣点(Points of Interest)样方密度的功能用地识别结果进行补充校验,根据政策和规划文件建立功能用地与主体功能区之间的关联,利用信息熵分析广州市功能用地混合程度,以辅助判别主体功能区,最终得到广州市主体功能区划分结果。将划分结果与《广州市主体功能区规划(2008—2020年)》和《广州市城市总体规划(2017—2035年)》草案对比验证,结果表明文章所提出的方法精准度较高,并能体现广州市空间格局形态,反映主体功能区实际分布情况。

关键词: 多源数据, 主体功能区, POI, 广州

Abstract:

Major Function Oriented Zoning (MFOZ) is a part of the national land space development and protection system and is the main basis for other special space development and layout plans; further, it plays a role in strategic planning and target transmission in the spatial planning system. The study area is the city of Guangzhou, China, and based on a 96.6% matching rate for the interest point (POI) data, the village-level administrative unit is selected as the basic calculation unit of the study. Functional land type identification is performed using a deep network convolutional neural network, and semantic extraction of the remote sensing images is conducted. The extraction results and the POI data function land obtained by the sample density method undergo data reclassification and are assigned to an assignment model. The result is supplemented and verified. In accordance with the relevant policies and planning documents, an association is established between the functional land and the MFOZ, and the rules of interpretation for the functional area and MFOZ are developed. As the POI data is peculiar, an information entropy method with quantitative metric elements is used to analyze the degree of mixing of functional land use in Guangzhou, and the road network data is used to determine the regional traffic development to assist in judging the MFOZ. A cluster analysis of the functional land of the village-level administrative unit from the geographical space is performed by dividing Guangzhou City into five functional areas: core promotion zone, adjustment optimization zone, key expansion zone, moderate development zone, and prohibited development zone. The resulting zone map is compared with the draft of the Guangzhou Major Function Oriented Zoning Planning (2008-2020) document and the Guangzhou City Master Plan (2017-2035). We observe that the proposed method has a high accuracy and appropriately reflects the Major Function Oriented Zoning. The actual distribution of the functional area reflects the spatial pattern of Guangzhou: a core upgrading area dominated by the main urban area, an adjustment and optimization area mainly in the Panyu and Baiyun Districts, a large-scale expansion area dominated by the Huadu District and Huangpu, a moderate development zone dominated by the Conghua and Zengcheng Districts, and a prohibited construction zone.

Key words: multi-source data, Major Function Oriented Zoning, POI, Guangzhou

中图分类号: 

  • F301.2