Poli, Daniela
The resulting y-parallaxes are a function of the height, the along-track wind component (only for clouds) and of the nadir
and forward zenith angle. The nadir and forward zenith angle are known at 11 equally distributed points of the first and
last scan line from the GBT header and can be linearly interpolated for all pixels (Bailey, 1995). If no cloud motion
information is available, the cloud height 4 is calculated from the uncorrected y-parallax y, after (Prata and Turner,
1997):
= Yp
tan((x,) — tan(x,)) N
where x; x, forward/nadir zenith angle
yp: parallax in y-direction
If either the cloud wind vector or the total cloud wind velocity is available from another source (e.g. Meteosat cloud
winds), the y-parallax can be corrected for the wind-induced along-track amount. The x-parallax is a function of the
across-track wind component and of the resampling error (for gridded products). For the wind corrections, the exact time
difference between the same pixel in the forward and the nadir scan has to be calculated from the along-track distance
and the satellite velocity.
4 CONCLUSIONS
The potential of deriving cloud-top heights from stereo satellite images with 288 m and 1000 m resolution has been
demonstrated. The same algorithms could be used with both datasets. Preprocessing and post-processing blunder
detection algorithms led to improvements. The accuracy potential of our matching approach, as shown with the MOMS
data, is well into the subpixel range, being able to fulfil the accuracy requirements for applications in weather and
climate. Similar matching problems were encountered, with slightly more for MOMS due to the larger parallax range and
higher resolution. Large blunders remain undetected in the results, caused mainly by the surface discontinuities, mixing
of image neighboring areas of large vertical separation, reflectance differences, and poor approximate values.
Approaches to reduce these problems have been already identified and will be implemented in future investigations.
ACKNOWLEDGEMENTS
We would like to thank Dirk Stallmann, Institute for Photogrammetry, University of Stuttgart for providing valuable help
in processing of the MOMS-02 imagery and Chris Mutlow, Rutherfold Appleton Laboratory, for valuable comments
about the ATSR2 data. The ATSR2 data was kindly provided by the ESA NRT service.
This work is funded by the Bundesamt für Bildung und Wissenschaft (BBW) within the EU-project CLOUDMAP (BBW
Nr. 97.0370).
REFERENCES
Bailey, P., 1995. SADIST-2 v100 products. ER-TN-RAL-AT-2164, Rutherford Appleton Laboratory.
Baltsavias, E.P., 1991. Multiphoto geometrically constrained matching. Ph. D. dissertation, Institute of Geodesy and
Photogrammetry, ETH Ziirich, Mitteilungen No. 49, 221 p.
Baltsavias, E. P, Stallmann, D., 1996. Geometrical potential of MOMS-02/P data for point positioning, DTM and
orthoimage generation. IAPRS, Vol. 31, Part B4, pp. 110 - 116.
Buongiorno, A., 1999, WWW ATSR Near Real Time service, quick guide, version 1.0. ESA/ ESRIN.
Danson, F.M., Higgins, N.A., Trodd, N.M., 1999. Measuring land-surface directional reflectance with the along-track
scanning radiometer. Photogrammetric Engineering and Remote Sensing, 65 (12), pp. 1411-1417.
1168 International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000.
