본문 바로가기
  • Home

Comparison of Built-up Area Extraction Methods: Classification Method vs. Vegetation Index Method

  • Journal of the Korean Cartographic Association
  • Abbr : JKCA
  • 2012, 12(3), pp.35-44
  • Publisher : The Korean Cartographic Association
  • Research Area : Social Science > Geography > Geography in general > Cartography

Jeong, Jae Joon 1

1성신여자대학교

Accredited

ABSTRACT

Satellite image classification is one of the useful methodologies for land-cover (land-use) map generation. But prevailing supervised classification results vary according to training areas and classifiers. Meanwhile, vegetation indices originally developed for testing vital degree of vegetations are modified up to 20 kinds and NDBI (Normalized Difference Built-up Index), NDVI (Normalized Difference Vegetation Index) and MNDWI (Modified Difference Water Index) are suitable for detecting built-up area, vegetation area, and water area respectively. In this research, Landsat TM satellite images are transformed into 3 bands generated using NDBI, NDVI and MNDWI equation and then unsupervised classification method is used in classification. The proposed method can conveniently classify the image because the method doesn't require training area selection and statistical assumptions. Accuracy differences using reference data are about 2%. Because the proposed method does not need statistical requirements such as sample distribution and class divergence and is quite simple, it will be useful in built-up area extraction.

Citation status

* References for papers published after 2022 are currently being built.