Land cover change detection by integrating object-based data blending model of Landsat and MODIS

Accurate information on land cover changes is critical for global change studies, land cover mapping and ecosystem management. Although there are numerous change detection methods, pseudo changes can occur if data are acquired from different seasons, which presents a significant challenge for land cover change detection. In this study, land cover change detection by integrating object-based data blending model of Landsat and MODIS is proposed to solve this issue. The Estimation of Scale Parameter (ESP) tool under Minimum Mapping Unit (MMU) restriction is employed to identify the optimal scale for Landsat image segmentation. The Object Based Spatial and Temporal Vegetation Index Unmixing Model (OB-STVIUM) disaggregates MODIS NDVIs to Landsat objects using the spatial analysis and the linear mixing theory. Then, the change detection method of NDVI Gradient Difference (NDVI-GD) is developed to detect change and no-change objects considering the NDVI shape and value differences simultaneously. The results of the study indicate that the approach proposed in this study can effectively detect change areas when Landsat images are acquired from different seasons. OB-STVIUM is more suitable for change detection application compared with the Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) and NDVI Linear Mixing Growth Model (NDVI-LMGM), because it is less sensitive to the number and acquisition time of Landsat images.