植物生态学报 ›› 2020, Vol. 44 ›› Issue (3): 205-213.DOI: 10.17521/cjpe.2019.0236

• 研究论文 • 上一篇    下一篇

基于百度街景图像的行人视角城市街道植被绿化格局分析

冯思远1,魏亚楠1,王振娟1,于新洋1,2,3,*()   

  1. 1山东农业大学资源与环境学院, 山东泰安 271018
    2中国科学院地理科学与资源研究所, 北京 100101
    3Tropical Research and Education Center, University of Florida, Homestead Florida, 33031, USA
  • 收稿日期:2019-09-05 接受日期:2020-01-27 出版日期:2020-03-20 发布日期:2020-04-30
  • 通讯作者: 于新洋
  • 基金资助:
    国家自然科学基金(41877003);“十二五”国家科技支撑计划项目(2015BAD23B0202)

Pedestrian-view urban street vegetation monitoring using Baidu Street View images

FENG Si-Yuan1,WEI Ya-Nan1,WANG Zhen-Juan1,YU Xin-Yang1,2,3,*()   

  1. 1Department of Resources and Environment, Shandong Agricultural University, Tai’an, Shandong 271018, China
    2Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing 100101, China
    3Tropical Research and Education Center, University of Florida, Homestead Florida, 33031, USA
  • Received:2019-09-05 Accepted:2020-01-27 Online:2020-03-20 Published:2020-04-30
  • Contact: Xin-Yang YU
  • Supported by:
    National Natural Science Foundation of China(41877003);“Twelfth Five-Year” National Science and Technology Support Plan Project(2015BAD23B0202)

摘要:

城市街道绿化植被作为城市景观的重要组成部分, 其分布格局对城市景观美学发展及行人身心健康有显著影响, 立足行人视角准确监测街道绿植分布信息对城市规划与管理有明确的辅助作用。该文针对已有研究多采用沿天底方向垂直向下观测的遥感影像监测地表植被而对行人视角的绿色植被分布格局研究涉及不多的现状, 基于免费获取的百度街景图像, 选取绿植覆被典型的泰安市区为案例区, 结合网络信息抓取与空间地理信息处理技术, 分析百度街景图像提取侧视绿植信息的可行性, 统计并对比其计算结果与遥感影像提取结果的关系, 以期为城市规划与管理提供辅助参考信息。网络抓取案例区273个样点共3 276幅百度街景图像, 利用计算机监督分类提取图像中的绿植区域; 基于空间分析模型分析街道绿色植被的分布格局; 利用SPSS软件趋势拟合模块分析百度街景图像与遥感影像提取的植被信息的相关性。主要结果为: 百度街景图像可作为主数据源提取城市街道的侧视绿植分布情况; 案例区不同区域植被分布指数区别较大, 空间格局差异明显; 百度街道植被分布指数与基于遥感图像提取的10、20、50 m缓冲距离范围内植被覆盖面积呈显著正相关关系, 但两者的变化趋势并非完全一致。百度街道植被分布结果可作为遥感监测结果的辅助信息更好地指导城市绿色景观规划与精准管理。

关键词: 街景图像, 街道植被分布指数, 侧视视角, 植被格局, 监督分类

Abstract:

Aims The distribution pattern of green vegetation in urban streets has significant impacts on urban ecological environment and physical/mental health of local residents. Accurate detecting and monitoring of street green information is of great significance for precise urban planning and management, while there are few studies focusing on urban greenery estimation using profile image system.
Methods In this study, combining network information capturing and geospatial information analysis technology, Taiʼan city was selected as the case study area. Based on the Baidu application programming interface (API), a total of 3 276 Baidu Street View (BSV) images of 273 research samples were obtained and processed, and the green vegetation pixels in the image were extracted by computer supervised classification and compared with the artificial extraction results. Based on the proposed Baidu Street Vegetation Distribution Index (BSVDI), we monitored the street vegetation’s distribution characteristics from the pedestrian perspective, and analyzed the street- scale vegetation distribution pattern.
Important findings The BSV image could be used as the main data source to monitor the distribution of green trees and lawns in pedestrian’s perspective on the street scale. BSVDI was higher in the center, northeast and southeast of the study area than the other regions. BSVDI and remote sensing extracted vegetation covered area are significantly and positively correlated, with correlation coefficient of 0.76, 0.63 and 0.49 in the buffered distance of 10, 20 and 50 m, respectively. However, the change trends of the BSVDI and remote sensing results were not completely consistent in each study site. This study implies that the combination of BSVDI and remote sensing monitoring results can better guide urban green landscape planning and precise management.

Key words: street view image, street greenery index, profile view, vegetation pattern, supervised classification