A new technique for seasonal land-cover change analysis using directional brightness differencing

Due to seasonal spectral variability in land-cover of cool temperate climatic conditions, the method that suits best for seasonal land-cover change identification remains uncertain. The study tested 11 different binary change detection methods and compared their capability in detecting land-cover change/no-change information in different seasons. Multi-date Thematic Mapper (TM) data pertaining to different seasons were used for a wide set of change image generation. A relatively new approach was applied for optimal threshold value determination for separation of change/no-change areas. Research indicated that irrespective of the method used, the results using vegetation index change images, particularly Normalized Difference Vegetation Index (NDVI)-based change images outperformed all other tested techniques in change detection process (overall accuracy> 90% and Kappa value> 0.85 for all six change periods).