Evaluation of the Accuracy of Supervised and Unsupervised Classification Methods for Landsat 8 Image Study Area: Tartous City

Authors

  • Fadi Chaaban
  • Ahmad Ali
  • Ahmad Mohammed

Keywords:

land cover changes, Evaluate the accuracy of classification methods remote sensing, Landsat 8 image, geographic information systems (GIS), Tartus city

Abstract

Conservation of vegetation is an important issue for urban planners. and this requires constant information, on a wide geographical scale, and through this accurate information plans, resulting from satellite work after several operations and analysis of the optimal utilization of visualizations to show as much as possible accurate information.

The objective of this research is to evaluate the accuracy of the supervised and unsupervised classification methods for the Landsat 8 image taken in 2017 for the city of Tartous and its surrounding. The accuracy of the classification methods will be evaluated through Accuracy Assessment. This method is based on an assessment of classification of a number of points, The user determines the number of points required to verify their classification. Here we chose almost 100 points, after making corrections of the radiometric errors and noise in the image data, in addition to several improvements to improve and increase image clarity ( contrast enhancement and image transformation). The supervised and unsupervised classification methods were used on a subset image of Landsat 8 taken in 2017 in natural colors for Tartous city (Syria). The study showed that the Maximum likelihood method (supervised classification) gave the best results (86.21%, Kappa 0.80) compared to other supervised and unsupervised classification methods.

 

Published

2020-10-01

How to Cite

1.
شعبان ف, علي أ, محمد أ. Evaluation of the Accuracy of Supervised and Unsupervised Classification Methods for Landsat 8 Image Study Area: Tartous City. Tuj-eng [Internet]. 2020Oct.1 [cited 2024Mar.28];42(4). Available from: https://journal.tishreen.edu.sy/index.php/engscnc/article/view/9930

Most read articles by the same author(s)

1 2 > >>