Residences information extraction from Landsat imagery using the multi-parameter decision tree method

Abstract

The rapid and accurate grasp of changes in residences is crucial for urban planning and urbanisation. However, the traditional methods for extracting residences exists several problems, which lead to inaccurate extraction results. In this study, the Landsat image is used to establish a new method for extracting the residences quickly and accurately. The specific steps are as follows: (1) We calculate surface albedo to exclude the interference of waters and shadows; (2) Using single-band threshold method, we eliminate the interference of shadows; (3) Normalized Difference Vegetation Index is calculated to exclude the effects of vegetation; (4) Roads are removed by calculating the shape index. Verification shows that the accuracy of this extraction method is 92.81%, which is more accurate than the traditional methods and solves the problems existed in the traditional methods. This novel method is a new reference for other land cover research on the technical aspect.

Keywords:

• Landsat
• Geographic Information System
• decision tree
• residences
• extraction

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This research work was supported jointly by National Key Research Program of China (No. 2016YFC0502300 and 2016YFC0502102), Chinese Academy of Science and Technology Services Network Program (No. KFJ-STS-ZDTP-036) and International Cooperation Agency International Partnership Program (No.132852KYSB20170029 and No. 2014-3), Guizhou High-Level Innovative Talent Training Program ‘Ten’ Level Talents Program (No. 2016-5648), United Fund of Karst Science Research Center (No. U1612441), International Cooperation Research Projects of the National Natural Science Fund Committee (No. 41571130074 and 41571130042), Science and Technology Plan of Guizhou Province of China (No. 2017–2966).