REPORT OF THE SSG 4.187

WAVELETS IN GEODESY AND GEODYNAMICS

 

Chair: W. Keller

 

The main fields of the SSG activities have been

  1. publications
  2. software development
  3. education
  4. comparison of different wavelet based algorithms

1 Publications

The members of the SSG have published so far 22 papers in scientific journals and on conferences. These contributions can be divided into four groups

  1. pattern recognition,
  2. spherical wavelets,
  3. filtering and prediction,
  4. enhancement of numerical processes.

1.1 Pattern recognition

Most of the publications about wavelet applications belong to this field. The goal is to find certain signatures which, on different scales, are hidden in the signal.

One group of pattern recognition techniques deals with the interpretation of signal registration from different geodetic measuring systems : Superconducting gravimeters, airborne gravimeters and so on.The publications [3],[9],[19] and [21] can be counted to this group.

A more exotic application of pattern recognition using wavelets is the study of atmospheric turbulence and seafloor topography as reported in [2] and [4].

Last but not least, wavelets are frequently used for gravity field modeling. A topic which is discussed in the contributions [5] and [10].

 

1.2 Spherical wavelets

This topic is the main focus of the Kaiserslautern University Group with W. Freeden as its head. The group aims at a construction of a wavelet analysis on the sphere. Among the many publications of this group the contributions [1], [6]and [7] should be mentioned.

 

1.3 Filtering and prediction

The ability of wavelets to decompose a given signal into a sequence of non-overlapping (or slightly overlapping) frequency bands can be used for an optimization of prediction and filtering algorithms. The publications [13], [14], [15], [16] and [17] deal with this problem.

The time-frequency resolution property of wavelets is the key for the extension of the known filtering algorithms from the stationary to the non-stationary case. These questions are discussed in the publications [11] and [12].

 

 1.4 Enhancement of numerical processes

Numerical algorithms do not take effect uniformly on all scales. Usually their main effect is concentrated on a certain scale-range. Wavelets can be used to decompose the numerical process in a number of sub-processes, each of them operating on a certain scale-range. Neglecting the minor effective sub-processes can improve the numerical performance of the whole process considerably. This idea is studied in the contributions [8] and [20].

 

2 Software development

In order to provide the geodetic community with ready-to-use tools for wavelet analysis the most important wavelet-related algorithms as

windowed Fourier transformation
continuous wavelet transformation
discrete wavelet transformation
2D-discrete wavelet transformation

were coded in ANSI-C and tested with the GNU compiler collection gcc. These programs are command-line driven. In order to facilitate the use of these algorithms Tck/Tk graphical user interfaces and utilities for the visualization of the results using the GMT tools were supplied. Since gcc, Tcl/Tk and GMT are not available on every platform a transformation to the platform independent JAVA system is currently going on. Both the C and the JAVA versions of the algorithms can be downloaded from the SSG 4.187 home-page

http://www.uni-stuttgart.de/iag

 

3 Education

In order to provide young scientists a basic knowledge in wavelet theory SSG members were involved as lecturers in two summer courses on wavelets:

School of Wavelets in the Geosciences , Delft October 4-9, 1998
graduate course Wavelets in Geodesy and Geodynamics , University of Calgary, August 4-21, 2000

 

4 Comparison of algorithms

In order to provide a scale for comparison of different wavelet algorithms a synthetic polar motion test data-set was created by Dr. Schmidt and was made available to the SSG via the SSG home-page. The members of the SSG were requested to analyze the data with their own algorithms and to submit the results. From the results conclusions about the strength and the weakness of different algorithms should be drawn. Unfortunately, the resonance to this project was disappointingly low.

 

References

[1]Bayer M. , Freeden W., Maier T.: WAVELET ANALYSIS OF THE GEOMAGNETIC FIELD WITHIN SOURCE REGIONS OF IONOSPHERIC AND MAGNETOSPHERIC CURRENTS Berichte der Arbeitsgruppe Technomathematik, N0.: 217 University of Kaiserslautern, September 1999

[2]Beth, S. ;Boos, T.; Freeden, W.; Casott, N.; Deussen, D.; Witte, B: On the wavelet determination of scale exponents in energy spectra and structure functions and their application to CCD camera data In: K.-P. Schwarz (Ed.) Geodesy Beyond 2000: The Challenges of the First Decade Birmingham, 1999 IAG General Assembly Birmingham , July 19 - 30, 1999 pp. 221 - 226 Springer-Verlag Berlin Heidelberg 2000

[3]Bruton, A.M.; Schwarz, K.-P.,Skaloud,J.: The use of wavelets for the analysis and de-noising of kinematic geodetic measurements In: K.-P. Schwarz (Ed.) Geodesy Beyond 2000: The Challenges of the First Decade Birmingham, 1999 IAG General Assembly Birmingham , July 19 - 30, 1999 pp. 227 - 232 Springer-Verlag Berlin Heidelberg 2000

[4]Dreher, S.; Lalancette, M.F.: Estimating Lithosphere Characteristics With Wavelet Admittance and Coherence of Gravity and Seafloor Topography. paper given at the IAG International Symposium on Gravity, Geodesy and Geodynamics held in Banff, July 31 - August 4, 2000

[5]Doufexopopolou, M.G. ; Bartha, G.C.: A Strategy to Approach a Multi-scale Space/Time Problem by Only data Information.paper given at the IAG International Symposium on Gravity, Geodesy and Geodynamics held in Banff, July 31 - August 4, 2000

[6]Freeden W. , Michel V. : LEAST SQUARES GEOPOTENTIAL APPROXIMATIONBY WINDOWED FOURIER TRANSFORM AND WAVELET TRANSFORM Berichte der Arbeitsgruppe Technomathematik, N0.: 216 University of Kaiserslautern, September 1999

[7]Freeden W.: Multiscale Modelling of Spaceborne Geodata B.G. Teubner Verlag, Stuttgart, 1999

[8]Gilbert A., Keller W.: Deconvolution with Wavelets Journal of Geodesy. Journal of Geodesy 2000(74) pp. 306-320

[9]Haagmans,R.H.N.; Dorst, L.L.; Francis, O.: Wavelets and Statistical Testing for Improved Analysis of Super Conducting Gravimeter Data at the main Tidal Frequencies.paper given at the IAG International Symposium on Gravity, Geodesy and Geodynamics held in Banff, July 31 - August 4, 2000

[10]Horowitz, F.G., Strykowski, P.; Hornby, P.; Boschetti, F.; Cox, S.J.D.: A Multiskale Skeletonization of The EGM96 Gravity Disturbances, Locally Compared with Australian Geophysical data.paper given at the IAG International Symposium on Gravity, Geodesy and Geodynamics held in Banff, July 31 - August 4, 2000

[11]Keller W.: A Wavelet Approach to Non-Stationary Collocation In: K.-P. Schwarz (Ed.) Geodesy Beyond 2000: The Challenges of the First Decade Birmingham, 1999 IAG General Assembly Birmingham , July 19 - 30, 1999 pp. 208 - 213 Springer-Verlag Berlin Heidelberg 2000

[12]Keller, W.: A Wavelet Solution to Non-stationary 2D Collocation. data.paper given at the IAG International Symposium on Gravity, Geodesy and Geodynamics held in Banff, July 31 - August 4, 2000

[13]Keller, W.: Wavelets in Geodesy and Geodynamics Lecture Notes, Universtity of Calgary 2000

[13]Kotsakis C.: The Multiresolution Character of Collocation Journal of Geodesy in print

[14]Kotsakis C.: The Role of Statistical Collocation as a Tool for Constructing Generalized Multiresolution Analyses in the L2(R) Hilbert Space Bolletina di Geodesia e Scienze Affini in press

 [15]Kotsakis C.: Wavelets and Collocation: An Interesting Similarity In: K.-P. Schwarz (Ed.) Geodesy Beyond 2000: The Challenges of the First Decade Birmingham, 1999 IAG General Assembly Birmingham , July 19 - 30, 1999 pp. 227 - 232 Springer-Verlag Berlin Heidelberg 2000

[16]Kotsakis, C.: Multiresolution Aspects of Linear Approximation Methods in Hilbert Spaces Using Gridded Data Department of Geomatics Engineering UCGE Reports Number 20138

 [17]Kotsakis, C.; Sideris M.G.: Aliasing Error Modelling in Single-input Single-output Linear Estimation Systems. paper given at the IAG International Symposium on Gravity, Geodesy and Geodynamics held in Banff, July 31 - August 4, 2000

[19]Liu, L.T.; Hsu, H.T. ; Sun H.P.: Wavelet-characterized Approach to Harmonic Analysis of Tide Gravity Observations.paper given at the IAG International Symposium on Gravity, Geodesy and Geodynamics held in Banff, July 31 - August 4, 2000

[20]Salamonowitcz, P.H.: A Wavelet Based Gravity Model with Application to the Evaluation of Stoke's Integral.paper given at the IAG International Symposium on Gravity, Geodesy and Geodynamics held in Banff, July 31 - August 4, 2000

[21]Liu,L.T.;Hsu,H.T.: Wavelet application in Airborne Gravimetry..paper given at the IAG International Symposium on Gravity, Geodesy and Geodynamics held in Banff, July 31 - August 4, 2000

[22]Klees, R.; Haagmanns,R.: Wavelets in the Geosciences. Lecture Notes in Earth Sciences, Springer-Verlag, Berlin, Heidelberg, New York, 2000

 

 

TABLE OF CONTENTS                                                          TOP