| Autor: |
Sun, Ying, Zhang, Xinchang, Zhao, Yuan, Xin, Qinchuan |
| Předmět: |
|
| Zdroj: |
International Journal of Remote Sensing; Mar2017, Vol. 38 Issue 5, p1258-1276, 19p |
| Abstrakt: |
Rapid land-use/land-cover (LULC) changes such as urbanization have tremendous impacts on regional climate and environment. Satellite images acquired by fast-developing remote-sensing techniques provide frequent observations of the land surface, thereby allowing for continuous mapping of urbanization activities. In this study, we investigated the annual urbanization activities over the past three decades in Guangzhou, one of the largest metropolises in China. To enhance the efficiency of training sample extraction in long-term land-cover mapping, we developed a three-step method: 1) three spectral indices were derived to extract the candidates of training samples based on decision trees; 2) a spatial filter was used to extract homogenous samples for each land-cover type; and 3) temporal consistency checking was performed for the samples of urban areas. We applied the developed method to time-series Landsat images and produced annual land-cover maps of Guangzhou from 1987 to 2015. We evaluated the produced land-cover maps and found an average overall accuracy of 89.80% for all studied years. Our results show that dramatic urbanization has occurred in the region of the Guangzhou city, where built-up areas have mostly expanded to the northwest, east, and south of the central regions of Guangzhou. The average growth rate of urban areas in Guangzhou from 1987 to 2015 was at 38.72 km2 per year, which was generally consistent with the government survey data. Future studies are required to understand how rapid urbanization in Guangzhou influences social economy and environmental sustainability. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
