测绘通报 ›› 2020, Vol. 0 ›› Issue (8): 1-4.doi: 10.13474/j.cnki.11-2246.2020.0238

• 自然资源监测 •    下一篇

南昌地区不透水面信息自动提取

何佳君1, 刘向铜1, 涂梨平2, 徐看1, 何紫玲1, 宋昊航1, 刘欢1   

  1. 1. 东华理工大学测绘工程学院, 江西 南昌 330013;
    2. 江西核工业测绘院, 江西 南昌 330038
  • 收稿日期:2019-11-04 修回日期:2020-04-21 出版日期:2020-08-25 发布日期:2020-09-01
  • 作者简介:何佳君(1998-),男,硕士生,研究方向为INSAR图像处理。E-mail:1005402179@qq.com
  • 基金资助:
    江西省教育厅科学技术研究项目(GJJ170439);江西省教育科学“十三五”规划课题(17YB070;17YB079);刘宝珺地学青年科学基金(DMSM2017010)

Automatic extraction of impervious surface information in Nanchang

HE Jiajun1, LIU Xiangtong1, TU Liping2, XU Kan1, HE Ziling1, SONG Haohang1, LIU Huan1   

  1. 1. Faculty of Geomatics, East China University of Technology, Nanchang 330013, China;
    2. Jiangxi Faculty of Nuclear Industry Surveying and Mapping, Nanchang 330038, China
  • Received:2019-11-04 Revised:2020-04-21 Online:2020-08-25 Published:2020-09-01

摘要: 随着城市化进程的不断加速,不透水面成为一种十分关键的地表覆盖类型之一,如何快速获取高精度的不透水面信息成为目前遥感研究的重点话题。考虑不透水面构成复杂、光谱曲线复杂的特征,本文基于决策树思想,通过建成区指数(BUAI)、改进的归一化水体指数(MNDWI),制作研究区植被、水体掩膜,并根据研究区土壤反射光谱特点制作研究区土壤掩膜。通过基于决策树思想的提取模型,结合掩膜文件,逐步剔除透水面,得到不透水面信息。结果表明:该模型提取的不透水面信息总体精度达到了98.27%,Kappa系数为0.86。通过和实地信息对比分析,认为该模型提取结果较为可靠,适用于大范围地区的不透水面信息提取。

关键词: 不透水面, Landsat 8, 指数, 掩膜, 遥感

Abstract: With the accelerating urbanization process, impervious surface has become one of the most important types of surface coverage. How to obtain impervious surface information with high-precision quickly has become an important question in remote sensing. Considering the complex structure and spectral curve of the impervious surface, based on the idea of decision tree, using built-up area index (BUAI) and modified normalized difference water index (MNDWI), this paper makes the mask of vegetation and water in study area. Then, the characteristics of soil reflectance spectrum are utilized to make soil mask. Through the extraction model based on the idea of decision tree, combined with the mask file, the permeable surface objects are removed step by step, and the impermeable surface are obtained. The results show that the overall accuracy of the impervious surface extracted by the model is 98.27%, and the Kappa coefficient is 0.86. Compared with the investigation of the study area, the consequences are believed to be more reliable. It is suitable for the extraction of impervious water surface information in large areas.

Key words: impervious surface, Landsat 8, index, mask, remote sensing

中图分类号: