application of support vector machine method in hyperspectral mapping of ophiolite mélanges-a case study from eastern iran

the lithologies of regions, which located near the collision zone, are very different from other geology setting. mapping in these areas needs extensive and exact studies and tools because of the variety of rocks, intensive tectonic uplift and complicated units. hyperspectral sensor is one of the most advanced tools with hundreds of bands that each measures a very narrow range of wavelengths and continuous bands in visible and infrared spectrums, so it can identify various terrains despites with spectral similarities and complications. in present study, as the first survey of hyspectral data efficiency for separating ophiolite melange units in iran, we applied spectral - based method of support vector machine classification method on hyperion image, in east of iran. based on various laboratory- field studies, the lithology of studied area can be separated into five general groups (ophiolite series, metamorphic units, oligocene - miocene to quaternary volcanic units, limestone and flysh units). in this region for calculation of processing results accuracy rate, some scattered locations and points were sampled according to field surveys. these samples were analyzed in microscopic section and by electron microprobe. points of grand though selected based on these field-laboratory studies for compute results accuracy rate. according to results, the average overall accuracy for all lithology has reached 52% in total colored- mélanges of the studied area at the east of iran.  the user accuracy factor of svm method is highest for the lithology with more spectral separability. these coefficients are acceptable ratios in separation of ophiolites as actual complicated units.