Object-oriented classification of a high-spatial resolution SPOT5 image for mapping geology and landforms of active volcanoes: Semeru case study, Indonesia
The present work explores the object-oriented classification (OOC) of high-spatial resolution (HSR) satellite panchromatic imagery for mapping the geology of the persistently active Semeru volcano and its ring plain, east Java, Indonesia. A panchromatic SPOT5 image and a digital elevation model (DEM) have been used to identify geologic units, structures, landforms and deposits. The panchromatic image was georeferenced and enhanced using histogram equalization. The enhanced image was segmented into polygons using the EnviFx (Exelis) Software. The polygons were delineated and classified on the basis of spectral (panchromatic hues and textures), topographic (slope, elevation) information and geologic/geomorphic processes. The validity of classification was evaluated by interpreting Google Earth images, aerial photographs and limited field observations. The classification consists of three hierarchical levels across the volcanic area of about 745 km2. The first operational spatial level includes seven large volcano domains based on the spectral content of the volcanic, tectonic and lithological structures, and principal catchments. The second operational level, based on contextual relationships (topography, drainage network, vegetation cover type, stratigraphy and slope dynamics), encompasses 20 geological units that range between 30 and 80 km2 in area. Among the units, the third operational level distinguishes as many as 47 geomorphological sub-units (0.25-25 km2) according to slope gradient, deposit type, mass-wasting process and weathering. The resulting map provides a detailed pattern of geologic and geomorphic features unlike previous geologic maps that identified only 11 stratigraphic units. We show that the high-spatial resolution panchromatic SPOT5 scene can help to safely map the geology and landforms of persistently active volcanoes such as Semeru. We have applied the OOC technique on one HSR GeoEye panchromatic image to map another active volcano, Merapi in Central Java, after considerable geomorphic changes due to the large eruption in 2010.