Abstract
Abstract. The high-resolution satellite imageries (HRSI) are as primary dataset for different applications such as DEM generation, 3D city mapping, change detection, monitoring, and deformation detection. The geo-location information of HRSI are stored in metadata called Rational Polynomial Coefficients (RPCs). There are many methods to improve and modify the RPCs in order to have a precise mapping. In this paper, an automatic approach is presented for the RPC modification using global Digital Elevation Models. The main steps of this approach are: relative digital elevation model generation, shift parameters calculation, sparse point cloud generation and shift correction, and rational polynomial fitting. Using some ground control points, the accuracy of the proposed method is evaluated based on statistical descriptors in which the results show that the geo-location accuracy of HRSI can be improved without using Ground Control Points (GCPs).
Highlights
Nowadays, high-resolution satellite imageries have been widely used in vast range of remote sensing and photogrammetric applications
A different approach is adopted for the Rational Polynomial Coefficients (RPCs) modification by incorporating the global Digital Elevation Models (DEMs) into the intersection solution without using any Ground Control Points (GCPs) based on the considered idea by (Jeong, et al, 2012)
The average values of the calculated shift parameters for SRTM and ASTER GDEM are presented in Table 2, separately
Summary
High-resolution satellite imageries have been widely used in vast range of remote sensing and photogrammetric applications. The final 3D coordinates of the generated object points in the terrain-independent method, suffer from the shift and drift errors due to the inherent bias in the measured physical sensor parameters This necessitates the use of a minimum number of GCPs to decrease the shift and drift errors and modify the RPCs. This necessitates the use of a minimum number of GCPs to decrease the shift and drift errors and modify the RPCs In this case, the number and accuracy of GCPs should be sufficient to have a precise mapping using RPCs. Since the gathering of GCPs is a costly, time consuming, and limited process in many application, significant efforts are dedicated for bias compensation of high-resolution satellite images using existing dataset such as reference maps, LiDAR points, ortho photos, and Digital Elevation Models (DEMs). The template-based edge matching technique can be used to determine the local shift between the image and large scale topographic map features and model the RPCs bias compensation parameters precisely In this technique, the georeferencing accuracy is about one pixel (Oh, et al, 2015). Using 3D LiDAR point cloud, the accuracy of relative and absolute orientation of highresolution satellite images can be improved to sub pixel values
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
