The most common way of determining classification accuracy is through computation of error matrix (also called as confusion matrix). Detail procedure for computing error matrix is reported elsewhere [29,30]. Result of accuracy assessment of the present study is presented in Table 1. Desirable minimum limit of overall accuracy for land use land cover classification has been reported as 85% [29,31,32]. Further, accuracy for each individual class of a classified image should exceed 70% [32].
Finally the map is converted into vector format and the crop area polygons are exported as vector layer for further analysis using GIS software (ArcGIS).