Issues and Challenges of Remote Sensing–Based Local Climate Zone Mapping for High-density Cities
Refereed conference paper presented and published in conference proceedings

Other information
AbstractThe local climate zone (LCZ) classification system provides a standard organizing principle for urban climate studies, such as urban heat island analysis, in which urban structures are classified into 17 standard LCZ classes according to urban land surface properties, such as land use and land coverage. Based on the standard definitions of the LCZ, freely available Landsat satellite data have been used to generate LCZ maps for cities worldwide. This research aims to evaluate the performance of the existing methods for high-density cities and to discuss the main issues, challenges, and possible solutions for further applications. Our experimental results indicate that three main factors, including training samples, input features, and the classifiers used, had a decisive effect on the final mapping result. In particular, low-quality training samples coupled with complex urban scenarios tended to yield low-quality LCZ mapping results. To improve the LCZ mapping results, some practical solutions and suggestions were given for future applications.
Acceptance Date10/02/2017
All Author(s) ListYong Xu, Chao Ren, Meng Cai, Ran Wang
Name of ConferenceJoint Urban Remote Sensing Event (JURSE 2017)
Start Date of Conference06/03/2017
End Date of Conference08/03/2017
Place of ConferenceDubai
Country/Region of ConferenceUnited Arab Emirates
LanguagesEnglish-United States

Last updated on 2018-22-01 at 05:55