A spectral gradient difference based approach for land cover change detection

Change detection with remotely sensed imagery plays an important role in land cover mapping, process analysis and dynamic information services. Euclidean distance, correlation and other mathematic metrics between spectral curves have been used to calculate change magnitude in most change detection methods. However, many pseudo changes would also be detected because of inter-class spectral variance, which remains a significant challenge for operational remote sensing applications. In general, different land cover types have their own spectral curves characterized by typical spectral values and shapes. These spectral values are widely used for designing change detection algorithms. However, the shape of spectral curves has not yet been fully considered. This paper proposes to use spectral gradient difference (SGD) to quantitatively describe the spectral shapes and the differences in shape between two spectra. Change magnitude calculated in the new spectral gradient space is used to detect the change/no-change areas. Then, a chain model is employed to represent the SGD pattern both qualitatively and quantitatively. Finally, the land cover change types are determined by pattern matching with the knowledgebase of reference SGD patterns. The effectiveness of this SGD-based change detection approach was verified by a simulation experiment and a case study of Landsat data. The results indicated that the SGD-based approach was superior to the traditional methods.