Accuracy assessment of the CARTOSAT DEM using robust statistical measures

Abstract

CARTOSAT-1-derived 10 m spatial resolution digital elevation model (DEM) for entire India was compiled for several applications. Overall assessment of the accuracy of this product requires additional regional studies involving ground truth control points and accuracy verification methods with a higher level of precision, such as the global positioning system (GPS).The study presented in this paper compares the accuracy of CARTOSAT-1 datasets with respect to eight sites over different terrains in India with the same GPS system. Robust statistical analysis including mean errors, standard deviation error, root mean square error (RMSE), skewness, kurtosis measures and Shapiro–Wilk’s normality test were used for evaluating error. The results of this study show a linear trend between the DEM and the ground control points (GCP). The mean error is very high in highlands ranging up to − 14.06 m, whereas in moderate terrain it ranges around 2.65 m and in the lowland about 1.20 m. RMSE ranges up to 28.82 m in rugged high-altitude topographies, 6.24 m in moderate and 1.98 m in low-altitude regions. However, this study shows that CARTODEM is one of the finest DEM that can be used for the Indian subcontinent as it works more accurately over plain and moderately undulating lands.