Change Detection through Four Techniques Using Multi-Temporal Landsat Thematic Mapper Data: A Case Study on Falavarjan Area, Isfahan, Iran

Abstract

The present study was aimed to assess the applicability of four techniques for detecting changed and unchanged areas in terms of land cover in Falavarjan area, (Isfahan, Iran). The images of the multi-temporal Landsat Thematic Mapper (TM) data acquired on 17 September, 1990 and 13 August, 2010 were used to apply land cover change analysis. The Images were respectively radiometrically and geometrically corrected. The root mean square errors were less than 0.5 pixels for each image. Finally, the image differencing method was used to produce the change image. To separate out the changed and unchanged areas in the difference image, four techniques including Metternicht's method, statistical method, Liu's method and Kapur's method were employed. Among them, the Metternicht's method followed by statistical thresholding technique yielded more accurate binary images.

Keywords: change detection, fuzzy logic, image differencing, Landsat images, thresholding