PE&RS December 2003

Direct Georeferencing

One of the objectives of the Direct Georeferencing column is to briefly present the new developments in the Direct Georeferencing technology and its integration with modern technologies for mapping applications around the world. Recently, rapid development in digital image acquisition systems and digital image processing capabilities resulted in the development of a number of commercial digital camera systems. These systems represent the present and the future of data acquisition in the mapping industry. Examples of these systems are Leica ADS 40, Z/I Imaging DMC, and Emerge DSS. In the next few months, the Direct Georeferencing column will be inviting the developers of all known commercial digital camera systems to present an article or a series of articles introducing a brief technical background about the georeferencing aspects of each system. This is intended to help mapping professionals, educators, students, engineers, and system developers with an overview of the latest developmentsin this field.

Since the author has been directly involved with the development of the DSS, a number of articles will be published to briefly present the performance of the DSS based on real data from different flight missions. This article is the second in such a series, where a brief overview of the preliminary analysis done to evaluate the performance of the DSS using the flight data flown in Japan in February, 2003.

In the Japanese test field, over 60 ground control points were available. The ground control points are well distributed and their positions are accurately surveyed. The availability of such accurate and well distributed GCPs allows for successful calibration and verification of the different system components of the DSS and for testing the performance of the system as a whole. Note that no ground control point is needed for the DSS data processing in general, except for quality control purposes.

The data flow was presented in the August issue of PE&RS (Volume 69, Number 8), and is repeated here for convenience. The POS AV data are run through POSPac software, while the image data are run through ImageView Software. The radiometrically calibrated images together with their exterior orientation parameters are then imported into a commercial third-party Automatic Tiepoint Generation (ATG) software to generate a number of pass/tie points per each image. Then, the image coordinates of tie/pass/GCP, the exterior orientation parameters, and any other available information are imported in POSCal software to calibrate the camera/IMU boresight and to do quality control of the camera calibration parameters and datum definition. Then, the data are looped once or twice between the ATG (using ISAT software for the data at hand) and POSCal software to obtain the ideal solution for the calibration parameters. Then, as a final quality control step, the QC’d exterior orientation parameters, camera calibration parameters, and datum shifts (if any) are imported into a digital photogrammetric workstation to check the DSS-computed checkpoint ground coordinates against the land-surveyed ones, and to check the remaining y-parallax.

Table 1 lists the results for one of the flights flown in Japan where the flight altitude was about 2000 m and the lens in use was of 55 mm focal length. The ground sample distance is 0.3 m. The first column lists the statistics of the remaining y-parallax in micrometers (µm), while the remainder of Table 1 lists the difference between the DSS-computed checkpoint coordinates and the land-surveyed ones. Note that the results shown in Table 1 reveal that the calibration and quality control procedure designed for the DSS is accurate enough for the desired applications. For example, although the DSS is primarily used for ortho-mapping, the maximum reported y-parallax is 7.8 µm for this project, which allows for comfortable stereo vision. More importantly, the ‘Mean’ values of the checkpoint coordinate differences are small enough to indicate that the system calibrated parameters have minimal or no remaining biases. The last row shows the RMS values for checkpoint X, Y, and Z components individually (one sigma).

Summary and Outlook
A brief overview of the DSS in-flight performance is introduced in this article. Among the next few article in the Direct Georeferencing Column, further analysis will be presented in some detail. Also, a number of articles will be published to present the performance analysis of the currently operational commercial digital camera systems

Further Reading
Mostafa, M.M.R., 2003. Performance Analysis of the DSS. Direct Georeferencing Column, PE&RS, 69(8):835.
Mostafa, M.M.R., 2003. Design and Performance of the DSS. CD-ROM Proceedings of the 49th Photogrammetric Week, Stuttgart, Germany, September 1-5, 2003.

Edited by Dr. Mohamed M.R. Mostafa, Applanix Corporation.