|
" Population Estimation in a Desert City Using Parameters from Satellite Imagery "
Lin, Weiying
Deng, Chengbin
| Document Type
|
:
|
Latin Dissertation
|
| Language of Document
|
:
|
English
|
| Record Number
|
:
|
1104732
|
| Doc. No
|
:
|
TLpq2247838998
|
| Main Entry
|
:
|
Deng, Chengbin
|
|
|
:
|
Lin, Weiying
|
| Title & Author
|
:
|
Population Estimation in a Desert City Using Parameters from Satellite Imagery\ Lin, WeiyingDeng, Chengbin
|
| College
|
:
|
State University of New York at Binghamton
|
| Date
|
:
|
2019
|
| student score
|
:
|
2019
|
| Degree
|
:
|
M.A.
|
| Page No
|
:
|
72
|
| Abstract
|
:
|
Increasing population living in urban areas produces huge challenges in urban sustainable development due to its impacts on natural environment and limited public resources. Timely and reliable population information is useful for both public practices and private sectors. With the development of remote sensing techniques, increasing number of studies have attempted to estimate population distribution by using geospatial techniques. However, these studies mainly focused on temperate areas (e.g. America, Australia, and China) dominated by vegetated covers while placed less emphasis on regions with complicated desert landforms. This thesis, therefore, explored the effectiveness of two popular population indicators for mapping population distribution in Dubai, United Arab Emirates (UAE). One is remote-sensing-derived impervious surface fractions from three spectral analytical approaches (i.e. simple spectral mixture analysis, normalized spectral mixture analysis, and random forest) and the other is nighttime light (NTL) radiance imagery. With the four different indicators, population models were built, respectively, at the community level (the smallest unit in UAE) to estimate spatial population distribution. Their performances were then compared. There are three major findings in the thesis. First, random forest outperforms the other two spectral mixture analysis in terms of mapping impervious surface area in the desert environment. Second, the robustness of population estimation model is related to the precision of imperviousness distribution mapping. Finally, when compared to NTL, percent imperviousness is a more suitable indicator to estimate population density at the finer scale in a desert environment.
|
| Subject
|
:
|
Geography
|
|
|
:
|
Remote sensing
|
| |