Remote sensing technology is a useful tool in obtaining information with respect to soil characteristics and land degradation aspects. This work considers the possibilities of using band ratios (plant cover and soil indices) for classification of saline lands in a part of Neyshabor plain region in Khorasan Razavi State in which ETM land satellite data of July 10 2002 was also used. Satellite images was reviewed and due to primary and radiometric corrections, the study area was selected from complete frame image. A Selection of sampling points with uniform distributions was performed and soil samples from 0 -10 cm of the studied point were collected and their geographical positions were recorded by GPS. 39 soil samples were analyzed for EC, pH, soluble Ca, Mg, Na, Cl, CO3, HCO3, SO4 ions. The position of samples was determined on each image, and Digital Number (DN) in each point was detected. Results indicated that the regression coefficients obtained for main and composite bands were very low, but revealing that the use of composite bands (plant cover indices) compared to main bands resulted to an increasing in the regression coefficient in the equations. In addition, when statistical analysis is limited to Data of higher EC, regression coefficients increase compared to when total sample parameters was statistically analyzed. All in all, it was realized that, due to very low regression coefficients among numerical vales obtained from main bands and plant indices and EC of soils, the classification of satellite images on the bases of a suitable regression model is not possible and also these coefficients for classification of the images are not sufficient.

Keywords