STATISTICAL INVESTIGATION OF ACCURACY OF SATELLITE ELEVATION DATA: A CASE STUDY

In this work we have used SRTM 90 and 30m, ASTER 30m and 60 cm Google Earth (GE) based elevation data generated using online tool Terrain Zonum, validation was performed with respect to Survey of India (SOI) toposheet. These results are important for those areas where limited field data are available. It also provides a benchmark for validation of future DEM with respect to SOI toposheet. Statistical tests were applied to assess vertical accuracy. The results highlight a minimum and maximum RMSE of 1.54m and 3.75m for temples and settlement, SRTM-30m, ASTER-30m and temples (1.84m) and 4.54m (for settlement) respectively with respect to SOI toposheet. Similarly, SRTM-90m indicate min and max RMSE of 1.88m and 6.57m for temples and other type of trees respectively. While Google Earth (GE) model show minimum 16.08m (for settlement) and maximum is 29.35m (for temples). We observed that SRTM-30m has good vertical accuracy for all classes (forest, towers, settlement, boundary pillars, temples and other types of tree). SRTM-30m score first rank for study area (Tehri District of Uttarakand, India) on the basis of statistical analysis.