|
Object-based and pixel-based methods in land cover changes detection using Landsat and Worldview imagery (case study: West of Tehran)
|
A. Habibi Razi1   , Z. Azizi *  1 |
| 1- Department of Remote Sensing and GIS, Faculty of Natural Resource and Environment, Science and Research Branch, Islamic Azad University, Tehran, Iran |
|
|
Abstract: (329 Views) |
Aims: Humans can affect the regional natural and biological environment by changing land use. In order to detect such changes in the earth's surface features, remote sensing is a practical methods. Surface features are detected and then analyzed in the studied area using the chronological charactirstics of the satellite data.
Methodology Various digital methods have been developed to show changes in land cover and the earth surface transformations using remote sensing. Landsat 7 satellite images taken in December 2000 and those of Worldview2 taken in December 2020 were used over a 20-year period to compare object-based (Support Vector Machine) and pixel-based (Maximum Likelihood) methods. The mentioned satellite images were used to investigate land use changes trend in the western region of Tehran.
Findings: The spatial resolution of the images and the algorithm used to extract features of the earth's surface were more efficient in world View 2 images. The overall accuracy and kappa coefficient for Landsat 7 images were 56.9767% and 0.3962, respectively, while for Worldview2, they were 86.0465% and 0.8069, respectively. Over the 20-year period, bare lands area decreased by 9.61 square kilometers, vegetation-covered area decreased by 5.13 square kilometers, constructed areas increased by 13.13 square kilometers, and water areas increased by approximately one square kilometer due to the construction of Chitgar Lake.
Conclusion: The results of the evaluations indicate that the support vector machine method is more compatible than the maximum likelihood method to detect land use types and land cover characteristics, as it incorporates additional factors beyond spectral reflection, such as shape, size, and texture factors between bands. |
|
| Keywords: Variation detection, maximum similarity, classification, support vector machine |
|
|
Full-Text [PDF 2868 kb]
(132 Downloads)
|
Article Type: Original Research |
Subject:
Urban Planning
Received: 2022/03/6 | Accepted: 2023/05/27 | Published: 2023/06/5
|
| * Corresponding Author Address: Science and Research Branch, Daneshgah Boulevard, Simon Bulivar Blvd, Islamic Azad University, Tehran, Iran. Postal Code: 1477893855 (zazizi@srbiau.ac.ir) |
|
|
|
|
|
|
| Add your comments about this article |
|
|
|
