Combined rigorous-generic direct orthorectification procedure for IRS-p6 sensors Meisam Yousefzadeh a,b,⇑ , Barat Mojaradi c a Islamic Azad University, Zanjan branch, Iran b University of Twente, Faculty of Geo-information Science and Earth Observation (ITC), Enschede, The Netherlands c Dept. of Geomatics Engineering, Iran University of Science and Technology, Tehran, Iran article info Article history: Received 25 June 2011 Received in revised form 13 September 2012 Accepted 13 September 2012 Available online 26 October 2012 Keywords: Direct georeferencing Orthoimage Pushbroom Ancillary data DEM Google Earth abstract The accessibility of global data, such as the digital elevation model (DEM), and the development of global visualisations allow promising new methods for minimising the distortion in the online generation of rough orthorectified products to be developed. Direct georeferencing (DG) has attracted a considerable amount of attention in the applications of pushbroom raw images in orthorectification or mono-plotting using ancillary satellite data. This study builds on recent DG studies to achieve an ‘‘orthoimage’’ from raw data and to determine potential mapping errors due to the DG procedure. Thus, this paper focuses on establishing a simple method for mitigating the misalignments of space-borne imageries to be used in direct orthorectification. Towards this goal, instead of image resectioning, affine transformation in dif- ferent coordinate systems is employed in the orthorectification algorithm to compensate for the systematic DG errors. For a given point, the elevation corresponding to the obtained planimetric coordinate is extracted using available topographic maps and global DEMs, such as SRTM’s and ASTER’s DEMs. As a result, parameters no longer need to be updated, as in the conventional orthophoto generation methods. To eval- uate the proposed procedures, experiments were conducted over three different IRS-p6 sensors in five datasets with different swath widths and tilt angles. The obtained results also demonstrate that the geo- graphic coordinate system and a simple 2D affine transformation can efficiently correct misalignments. Ó 2012 International Society for Photogrammetry and Remote Sensing, Inc. (ISPRS) Published by Elsevier B.V. All rights reserved. 1. Introduction In a typical imaging system, an actual 3D object is projected onto a 2D image plane after passing through a camera/sensor. The mathematical model linking the 3D object coordinates to the 2D image coordinates is known as the physical camera model. The reconstruction of the camera model or viewing geometry includes the exterior and interior orientations of the sensor. The exterior orientation (EO) describes the location of the projection centre and attitude of the bundle of rays, whereas the geometry of the bundle of rays is reconstructed by the measured image posi- tions and interior orientation (IO). Linear array satellite images, re- ferred to as ‘‘pushbroom images’’, have a set of the EO parameters for each image line, and the IO is restricted to this line. Thus, for a given satellite image sensor, the reconstruction of the imaging geometry involves a combination of the IO and EO (Radhadevi and Solanki, 2008; Tighe et al., 2009). The DG methods utilise ancillary data to retrieve the physical geometry of the imagery. The sensor’s EO parameters can be obtained from the physical parameters of the orbit through the Keplerian elements (Gugan and Dowman, 1988; Valadan Zoej and Foomani, 1999) or state vectors, where the position, velocity and attitude values for a particular time are contained in the ancil- lary data. This method was proposed by Toutin (2004), imple- mented in the PCI Geomatica software and has also been recently evaluated in the literature (Poli, 2001, 2005). Theoretically, in this approach, which is referred to as the ‘‘rig- orous camera model’’, the DG procedure does not require any ground control points (GCPs) or additional processes. Thus, the effectiveness and reliability of this method depends on the accu- racy of the internal and external orientation data. In other words, the results may not be satisfactory if the interior orientation parameters, which are given by the pre-flight laboratory calibra- tion and the exterior orientation measurements are inaccurate. If those data are available at a very high accuracy, the values of the misalignments and the shifts between the positioning and imaging instruments must be considered. Thus, multiple GCPs are needed for these bias and shift corrections (Poli, 2005). In contrast to the camera model, a popular alternative is the rational polynomial coefficients (RPC) model, which is used by 0924-2716/$ - see front matter Ó 2012 International Society for Photogrammetry and Remote Sensing, Inc. (ISPRS) Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.isprsjprs.2012.09.005 ⇑ Corresponding author at: University of Twente, Faculty of Geo-information Science and Earth Observation (ITC), Enschede, The Netherlands. E-mail addresses: yousefzade29109@itc.nl (M. Yousefzadeh), mojaradi@iust.ac.ir (B. Mojaradi). ISPRS Journal of Photogrammetry and Remote Sensing 74 (2012) 122–132 Contents lists available at SciVerse ScienceDirect ISPRS Journal of Photogrammetry and Remote Sensing journal homepage: www.elsevier.com/locate/isprsjprs