Application of Lincar Distinetion Analysis for Wheat Discrimination from Other Crops on Satellite Images

In this paper, has been tried to discriminate wheat_cultivated fields from those crops that have similar characteristics to this class, spectrally and phonologically, in bahar_hamedan region. So based on cultivation calendar of study region, decided to apply bitemporal or dual_season multispectral spot5 images for this aim. Both images divided to discrete objects(polygpns) using multiresoluion segmentation algorithm. The zonal attributes of segments entered to discriminant analysis as variables. After the stepwise discrimination analysis, all the segments were classified using outputted discrimination functions. ‘’discrimination accuracy’’ and ‘’ classification accuracy’’ were defined as discrete concepts. Also, with assuming the DNs of each band pixels as variables “pixel based” discriminant analysis was applied. Comparison of both methods results indicated that generally, accuracy of pixel_based classification is upper than other method.