Fusion Technique to Extract Detail Information from Moderate Resolution Data for Global Scale Image Map Production

摘要

The Global Irrigated Area mapping project at IWMI addresses the need to develop improved regional or global image products that incorporate spatial and temporal heterogeneity of agriculture. Fusion of freely available high temporal resolution MODIS composite image and high spatial resolution GeoCover Landsat TM mosaic is an appropriate approach to monitor temporal and spatial variability in agriculture at global scale. Methodologies were developed using the techniques of IHS color transformation and Principal Component Analysis to produce global image fusion products. In these processes, the spectral characteristics of the original MODIS multi-spectral image are maintained. In effect one can clearly identify the textural and structural features of land cover as well as the vegetation cover at the period of MODIS data acquisition. Moreover it improves classification accuracy and spatial disaggregation.