Geography And Urben Planning
Mehri Roozbahani; Gholamreza Janbaz ghobadi; Sadroddin Motevalli; Jalal Azimi amoli
Abstract
The main aim of this paper is to detect the ten-year changes in urban green spaces of Tehran metropolis, from 2010 to 2019, using the time series of Landsat 5, 7 and 8 images. The change detection was done in both annual and ten-year scale and the results are analyzed in two spatial scales; city level ...
Read More
The main aim of this paper is to detect the ten-year changes in urban green spaces of Tehran metropolis, from 2010 to 2019, using the time series of Landsat 5, 7 and 8 images. The change detection was done in both annual and ten-year scale and the results are analyzed in two spatial scales; city level and municipal district-level. Detection of changes was done by a post-classification approach. The innovation of the study is efforts to reach the best results in the image classification step, for which in addition to optical and thermal bands' various features including some vegetation indices, water and built-up index, image texture components, and principal components were also used. Three classification methods including maximum likelihood, artificial neural network and support vector machine were implemented. The results indicated that the support vector machine has had the best result with 91.06% mean overall accuracy. The change detection showed a 10.58% decrease in the Tehran green spaces in the period under review. The greatest decrease, about 7.46 Km2, occurred in the period 1390-91 and the largest increase was 7.61 Km2 in the period 1394-95. Among the 22 municipal districts, regions 1 and 22 with 5.2 and 2.37 Km2, respectively, have had the highest decrease in urban green space, and regions 2 and 19 with 0.5 and 0.47 Km2, respectively, have had the highest increase.
