An Efficient Hybrid Classification Approach for Land Use/Land Cover Analysis in a Semi-Desert Area Using ${\rm ETM}{+}$ and LISS-III Sensor

Land-use and land-cover (LU/LC) studies help in assessing and monitoring the status of the natural resources, detecting the changes in spatial and temporal scale and predict them for the future. Due to changing environments and increasing anthropogenic pressures, the demand for a LU/LC database at the global level is increasing. Therefore, a comprehensive understanding of LU/LC at both local and regional scales is important since it plays a pivotal role in socioeconomic development and global environmental changes. There are many approaches for LU/LC analysis such as supervised classification, unsupervised classification and onscreen digitization but simplest and most popular approach on IRS LISS-III and ${\rm Landsat\hbox{-}ETM}{+}$ satellite data revealed a serious problem in some semidesert areas caused by spectral confusion because of the similar radiometric response like scrub land with harvested land, built-up with bare hills and many other. Present study suggests hybrid classification approach for LU/LC classification, which is found highly useful in achieving high accuracy for areas where spectral classes of images are inseparable.