A modified exterior orientation approach for IRS – 1C stereo images

A modified exterior orientation approach for IRS – 1C stereo images

SHARE

R. S. Tiwari1, P. K. Garg1, S. K. Govil2 and Sonia Sarkis1
1Remote Sensing and Photogrammetric Engineering Section, Department of Civil
Engineering, University of Roorkee, Roorkee – 247 667
2Indian Institute of Remote Sensing
4, Kalidas Road, Dehradun – 248 001

The stereo-photogrammetric principles of derivation of 3-D ground coordinates from aerial and terrestrial photographs are well know, and are abundantly available in the literature. The mathematical formulations for such determinations, in aerial and terrestrial photogrammetry, have also been well established and are being utilized for various purposes. The availability of high resolution linear array images such as SPOT, IRS-1C/1D has, however, changed the way in which satellite images may be used for extraction of data. These images can now be used for many such applications in which only aerial photo were used previously. Linear array images are also known as dynamic images since the position and orientation of the sensor in space are changing continuously during the scanning period of image. The purpose of this paper is to investigate as to how some of the available mathematical formulations, such as space resection/intersection used in aerial photogrammetry, may be applied in case of satellite stereo images and what modifications, if any, are required to be made when the existing formulations are used for stereo-satellite images.

In this paper a modified approach for exterior orientation (space resection) for linear array of images such as IRS-1C PAN stereo images, has been developed using four orbital parameters, considering the image as a frame photo. A simple method for orientation of image was developed, based on rotating the orbital system of each image to Geocentric coordinate system to get X, Y and Z of each point in geocentric coordinate system. The ground coordinates of any conjugate point (M, Y and 7 Z) are then determined by Space Intersection method for stereo images. A minimum of 3 GCPs are required for least square solution of the method, though more are preferable. A computer program has been developed in Fortran 77 for the extraction of height information from IRS-1C PAN stereo data. The accuracy in height has been found to be less than 2 pixels. The height data determined from this method can be used to generate DEMs and orthoimages from linear array imageries which then may be utilized for deriving data for planning and designing of engineering projects.