Author(s): Rifaat M. Abdalla; Jonathan Li; Zhi Chen; Yasir A. El Awad; Imran Maqsood
Linked Author(s):
Keywords: Change detection; Stream channel; Landsat TM; Multitemporal; Multispectral; Algorithm comparison
Abstract: This study provides an application of multitemporal and multispectral Landsat TM images to detect stream channel changes in the Qu’Appelle River. Two different digital change detection techniques were applied and compared. These techniques are: image differencing and principal components analysis (PCA). IKONOS panchromatic imagery with 1-m spatial resolution acquired in December 2000 along with 1:50,000-scale map of the study area was used as a topographic reference. The image difference technique has shown higher accuracy on detecting change, comparing to other techniques. The overall accuracy and Kappa coefficients for image differencing change detection method were 70% using the standard error matrix.
Year: 2002