BibTeX:

@article{IJIRSTV2I8025,
     title={Improving the Elevation Accuracy of Cartosat-1 Dem},
     author={Visakh S, Dr. S Muralikrishnan and M Sreedhar},
     journal={International Journal for Innovative Research in Science & Technology},
     volume={2},
     number={8},
     pages={117--128},
     year={},
     url={http://www.ijirst.org/articles/IJIRSTV2I8025.pdf},
     publisher={IJIRST (International Journal for Innovative Research in Science & Technology)},
}

Automatic image matching technique when applied to high resolution space images generate point cloud data describing the surface. A DSM is an elevation model of surface reflectance features and includes the height of cultural features such as buildings, roads and vegetation as well as bare earth. The surface model thus defined needs to be converted to 'bare- earth model' DEM before it can be used for any natural resource applications. Here in this study area an attempt has been made to improve the elevation accuracy of Cartosat-1 DSM which is generated using Augmented Stereo Strip Triangulation developed by ISRO, using high precision ground control points. The Cartosat-1 DSM is having a posting of 10m. The improvement in elevation accuracy is achieved by applying bias correction using ICESat/GLAS points and slope based filtering. For bias corrections the points falling on the ground are taken and other points are excluded in computation as they are treated as outliers. For the study area of Gandhak area covering 19,250 Sq. Km, the RMSE and Absolute LE90 for input DSM is 5.96m and 6.25m respectively. For the output DEM the RMSE and Absolute LE90 is 1.98 m and 2.08 m respectively. Statistical validation of input DSM and output DEM with ground control points shows an improvement of 4.17m in Absolute LE90 for output DEM with respect to input DSM. The values for output DEM shows that the 90% of elevations are within 2.08 m when compared with reference points. To access the output quality of DEM derivatives of DEM, drainage and contours are generated. The output DEM derived from filtering methods depicts the continuity in drainage network and contours along the river. Many isolated contours formed in DSM in vegetation areas are removed in the output DEM. Drainage continuity is very much useful for runoff and discharge flow estimation which are critical parameters for hydrological modelling. This study aptly demonstrates the improvement in elevation accuracy of Cartosat-1 DSM using ICESat/GLAS data and slope based filtering methods.

Digital Elevation Model, Remote Sensing, Slope Based Filtering, RMSE