Page 704 - NGTU_paper_withoutVideo

Modern Geomatics Technologies and Applications

set for the scale parameter, shape and compactness, respectively. In the image segmentation of Landsat-8 OLI image of 2019,
the scale parameter was set to 1 and the values of both the shape and compactness parameters were considered 0.6. After
performing segmentation, each of the image segments should be assigned to one of the pre-defined object classes. The
executive results of knowledge based classification process are shown in Fig. 4.
For performing the proposed post-classification change detection algorithm, the obtained object based classification
maps are compared with each other. The general change map is illustrated in Fig. 5. Table 2 depicts some statistical
information containing initial and last areas of each object class and the difference percentage determined from multi-temporal
changes.

Table 2. Statistical change information
Object Initial Area Last Area Differencing Area Difference Percentage
Classes (Pixels) (Pixels) (Pixels) (%)
Soil 493628 409791 -83837 -16.98
Vegetation 17164 34970 17806 103.74
Built up 35481 101015 65534 184.702

As it is depicted in the produced change map and statistical analysis of the post-classification change detection results,
soil object class is decreased for about 17% and built up areas are increased for about 184% in the 17 time interval. This is
concluded due to reduce the 83837 pixels rom soil area. Moreover, 65534 pixels are added to the built- up object class between
the years 2002 and 2019. Agricultural fields in this study area are mostly destructed due to the developments in constructing
built up areas. The increasing amount of about 104% in vegetation covers relates to the trees and grasslands in new constructed
built up areas.

Fig. 5. Change map produced from post-classification change detection algorithm

5. Conclusion
This article aims to investigate the changes caused by Mehr housing constructions in Pardis area located in the northeast
of Tehran. The new city of Pardis with about 120,000 housing units is known as the housing capital of Mehr in Iran and will
have a population of about 500 to 700 thousand people. The proposed post-classification land use/ cover change detection
method is applied on the Landsat images those were taken from the Pardis area in the 17 time interval. The post-classification
change detection is performed based on the capabilities of the object based image analysis procedures based on the spectral and
structural features.
As it is illustrated in the produced land use/ cover change map between 2002 and 2019, the constructions are rapidly
increased in Pardis area and this causes the serious impacts on the environment. According to the obtained post-classification
change detection results, the soil object class is decreased for about 17% and built up object class is increased for about 184%.
Construction in the area is associated with ecosystem change. For instance, the phase 11 of this project is completely
located in a mountainous area and new constructions have changed the natural structure of this area. Also, Phase 8 of the
Pardis project is associated with many environmental problems that have always had many concerns. The uncontrolled
increase of constructions without considering the infrastructure has caused problems such as water pollution in the villages
adjacent to the Pardis city and its sewage and treatment plant. Also, due to irregular land use changes, the agricultural lands of
this region have almost disappeared and only the names of them remain.
5