Accuracy Assessment of User-Derived RFCs for Ortho-Rectification of High-Resolution Satellite Imagery

  • Afify H. Department of Public Works, Faculty of Engineering, Tanta University, Tanta, El-Gharbiya, Egypt
  • Zhang K. Department of Geodesy and Geomatics Engineering, University of New Brunswick, 15 Dineen Dr., P.O.Box 4400, Fredericton, New Brunswick, Canada E3B 5A3


High-resolution satellite images acquired by IKONOS and QuickBird are currently considered as the most ideal data source for digital mapping applications. Three geometric correction methods can be used to ortho-rectify these satellite images: (1) 3D rigorous physical model, (2) 3D Rational Function Model (RFM) with vendor-supplied Rational Function Coefficients (RFCs), and (3) 3D RFM with user-derived RFCs from Ground Control Points (GCPs). Several investigations have been reported concerning the accuracy attainable by the first two methods. However, reports on geometric accuracy achievable using the third method can rarely be found in the literature. Therefore, an evaluation was performed in this research to identify the attainable geometric accuracies of RFMs when user-derived RFCs are utilized. Two panchromatic images from IKONOS and QuickBird, for Fredericton, New Brunswick, Canada were used. Third-, second-, and first-order RFMs were applied using different numbers of GCPs. This paper presents accuracy test results of user-derived RFCs and RFM using an Independent Check Points ICPs-based planimetric accuracy assessment, as well as an accuracy assessment of ortho-rectification using user-derived RFCs. The results demonstrate that competitive accuracies, compared to those obtained using the first and second methods, can be achieved using user-derived RFMs with the recommended numbers of GCPs.