부산대학교 도서관

Land use-land cover (LULC) mapping has immerged as a useful and important Remote Sensing (RS) and Geographic Information System (GIS) technique for improving the management of natural resources for the progress of a country like Pakistan. Therefore, a research study was conducted to develop LULC maps for the District of Multan, Pakistan. For this purpose, economically available multi temporal (time series) images with acceptable resolution from satellite (LANDSAT 7 ETM+) were obtained for Rabi and Kharif season of 2011-12 to perform supervised classification and identification of crops in the study area. Image Processing was performed in ERDAS Imagine 2011 version for obtaining the high-class time series normalized difference vegetation indices (NDVI) for each LANDSAT 7 imagery. Four classes were targeted for crops out of the seven clusters created using target crop signatures with 95% maximum likely-hood. The resulting crop types were validated by 86 ground truthing points. Over all 74 % efficiency was found using error matrix technique. The regional irrigation water requirements of specific crops were estimated using the generated LULC maps of exerted crop area. The calculated cropped areas through ArcGIS 9.3 version were of 0.226 Mha for cotton, 0.207 Mha for wheat, 0.014 Mha for rice and 0.007 Mha for sugarcane. The total regional crop water requirement of the study area was of 1653.62 Mm3 for cotton, 911.25 Mm3 for wheat, 97.93 Mm3 for rice and 112.25 Mm3 for sugarcane. The LULC mapping technique should be used to develop a decision support system for water, land and other natural resources management at regional scale for efficient resource utilization and sustainable development. [ABSTRACT FROM AUTHOR]