چکیده
عنوان مقاله [English]
In order to capability investigation on Landsat-7 Satellite ETM+ Data in Separability Forest Type and mapping in the Zagros Region, a small window digital data dating July 2001 from Ghalajeh Forests in the Kermanshah province were analyzed. Ground data were provided the cluster sampling method and with 0.36 ha. No radiometric error was found then the quality investigations. Orthorectification was implemented using 55 Ground Control Points and RMS Error equal 0.39 the X axis aspect and 0.46 the Y axis aspect. Atmospheric correction was implemented using of Cost model. Vegetation indexes to produce appropriate for soil reflectance effect decrease. Synthetic bands were generated using some suitable processing methods to produce the spectral arithmetic bands like principal component analysis and Tasseled Cap transformation. Multispectral and panchromatic data were fused using HIS and statistics-based pansharp fusion techniques. Selection of the appropriate bands for classification was performed using divergence separability index. Classification was done by supervised method using maximum likelihood (ML), minimum distance to mean (MD) and parallel epiped (PPD) classifiers. Accuracy assessment of classified maps was performed by ground truth map. The results showed that the classified map using maximum likelihood classifier could be better than other maps with an overall accuracy 68% and Kappa coefficient 0.29 for Oak and mixed forest types. Results generally showed a moderately capability ETM+ data to provide forest types map in the Zagros forest. These study would suggest the utility of best spatial resolution imagery and in other regions.
)