Submarine Mapping Using Multibeam Bathymetry
Submarine Mapping Using Multibeam Bathymetry
Conclusions
[3] Jain, A., Hong, L., Pankanti, S., and Bolle, R.
A software-hardware partitioning proposal for «An identity authentification system using
designing an embedded digital system for fingerprint fingerprints». In Proceedings of the IEEE, September
identification has been presented. The paper descri- 1997, vol. 85, pp. 1365-1388.
bes the most important stages involved in the
minutiae extraction of a fingerprint image. The [4] Jain, A.K., Hong, L., and Bolle, R.M. «On line
algorithm has been executed obtaining a profiler that fingerprint verification». IEEE Transactions on Pattern
allows the most critical parts with higher Analysis and Machine Intelligence 19, 4 (1997), pp.
computational cost to be detected. Hardware is 302-314
recommended to implement these stages and soft-
ware to implement the others. [5] Zhang, T.Y., and Suen, C.Y., «A Fast Parallel
Algorithm for Thinning Digital Patterns». CACM, 27
5. References (1984), pp. 236-239.
[1] Anil K. Jain, Ruud Bolle and Sharath Pankanti, [6] Simon-Zorita, D., Ortega-Garcia, J., Cruz-
«Biometrics, Personal identification in Networked LLanas, S., and González-Rodríguez, J. «Minutiae
Society». Kluwer Academic Publishers, 1999 extraction écheme for fingerprint recognition
systems. In Proceedings of the International
[2] Hong, L., Wan, Y., and Jain, A.K. «Fingerprint Conference on Image Processing (October
image enhancement: Algorithm and performance
evaluation». IEEE Transactions on Pattern Analysis 2001), vol. 3, pp. 254-247.
and Machine Intelligence, 20, 8 (1998), pp. 777-789.
Swath-Bathymetry processing
The first step is to import the raw data into CARAIBES
software in order to get navigation (date, time, and
geographic coordinates) and depth data (date, time, Fig. 2. Snapshot windows showing different stages of swath
bathymetric processing with Caraibes. Left: Ping and beam
beam number, two-way travel time). After navigation
editing. Right: Soundings editing across a defined path.
post-processing, which mainly consists of an
automatic filtering of anomalous values, smoothing
and interpolation, the two types of data are merged
and a georeferenced sounding file (x, y, z) is generated.
A Digital Terrain Model (DTM), a binary grid, is first
interpolated from raw soundings to search for errors,
such as the effect of tide (for shallow waters),
variations in water column sound velocity, and
motions of the vessel (roll, pitch and heave) (Bourillet
et al., 1996). Once these corrections are introduced,
data are filtered and cleaned up using different
methods, such as raw data automatic cleaning by
comparison with a reference DTM (using a band-pass
filter), and raw data manually cleaning with the ping
graphical editor to get more accuracy on the depth
data control (Fig. 2). After filtering, bathymetric data
is interpolated at nodes of a regular-spacing grid in
order to get a final DTM. The choice of the spacing of
the grid nodes takes into account the density of the
soundings, quality of the data and finally, the scale
of the final document. The software allows many
operations on the DTMs such as spline or numeric
filter smoothing, interpolation or mosaic of several
DTMs. The final DTM can be displayed under
different cartographic views such as slope gradient
representation, bathymetric cross sections along a
specific path, shaded relief maps and, 2D and 3D
visualizations (Fig. 3).