Automatic Extraction of Landforms in a Complex Terrain of Western Ghats Region of India Using Earth Observation Datasets and CIS

摘要

In this study, the SRTM digital elevation model (DEM) of 30m resolution and SENTINEL-2B satellite data (10m) were used to delineate distinct landforms in complex terrain of Pune district, Maharashtra, India using multi-resolution segmentation algorithm. The study area covered the Pune district of Western Ghats, Maharashtra, India. The SRTM DEM (30m resolution) was used to generate slope, aspect, hillshade, flow accumulation, contours, curvatures (plan, profile, total) and drainage network and Iso-cluster unsupervised classification was used for generating land use / land cover. Methods employed included of multi-resolution segmentation algorithm for generating landforms. In particular, the terrain variables employed for generating landforms in the study area are: elevation (m, above mean sea level), slope (percentage), contour (50m interval), plan curvature (m m~(-1)), profile curvature (m m~(-1)), Total curvature (m m~(-1)). hillshade, flow accumulation and drainage (200 stream cell count) and generated land use / land cover. The knowledge-based classification of the primitive objects based on their elevation and shape parameters, resulted in the extraction of the landforms. The analysis shows that alluvial plain occupies about 22.0 per cent of the TGA of the district, mostly spread in Daund. Baramati and Indapur thesils. Whereas, pediments were noticed mainly in Sirur and Purender thesils with an area of 16.4 per cent of the TGA of the district. The resulted boundaries in comparison to those by previous studies were found satisfactory. The study demonstrates that automatic extraction of landforms can be carried out using high resolution satellite data and DEM through multi-resolution segmentation algorithm.