Land use change mapping is one of the basic requirements for effective monitoring and management of the environment. Although several change detection approaches have been proposed and used, past experience has shown that each method may have its own advantages. Selecting the best approach is an important challenge. The integration of spectral/spatial criteria seems to lead to better results. In this study, for change detection analysis, we have used the Landsat Thematic Mapper (TM®) imagery of Zanjan city,acquired in 1985 and 2010. To map the changes, the spectral/spatial criteria were integrated with different weights by using the Fuzzy Analytical Hierarchy Process (FAHP). Results indicate that the FAHP methods generally resulted in a higher accuracy than spectral and spatial criteria. The proposed semivariogram features have a better, early and automatic change detection performance. When using spatial information, both the omission and commission errors decreased significantly. In addition, the uncertainty derived from choosing only one separate method decreases when using the FAHP method. This method will likely be widely used for the change detection of two images that have, on the one hand, the same spectral characteristics and, on the other, different texture characteristics.
Cite (Atıf) : Eisavi, V., Homayouni, S., Karami, J., 2016. Integration of remotely sensed spatial and spectral information for change detection using FAHP. Journal of the Faculty of Forestry Istanbul University 66(2): 524-538. DOI: 10.17099/jffiu.90466