Multi-aspect computer-aided classification of the Citadel Trial sidescan sonar images

PDF

Authors
  1. Fawcett, J.A.
  2. Crawford, A.
  3. Hopkin, D.
  4. Myers, V.
  5. Couillard, M.
  6. Zerr, B.
Corporate Authors
Defence R&D Canada - Atlantic, Dartmouth NS (CAN);Defence R&D Canada - Centre for Operational Research and Analysis, Ottawa ON (CAN);GESMA, Brest (FRANCE)
Abstract
In October 2005, Defence Research and Development Canada (DRDC) Atlantic participated with the Nato Undersea Research Centre (NURC) and Groupe D'Études Sous-marine de l'Atlantique (GESMA, France) in a joint trial (CITADEL) with the DRDC remote, semi-submersible vehicle DORADO. A large number of sidescan sonar images of minelike and non minelike objects on the seabed were obtained for a variety of ranges and aspects and at two different sites. In earlier reports, the computer-automated detection and classification of small sonar images from this trial was discussed. Many of these images are of the same object from different sonar passes. Thus there are multiple looks of the same objects at different ranges and aspects of the sonar. In this paper, the strategies for classification from 2 sets of features, corresponding to 2 sonar images of the same object, will be considered. A classifier based upon combining the feature sets from each of the 2 images will be considered. As well, the fusion of the individual image classification results using Dempster-Shafer theory will be discussed. The multi-aspect classification performance will be considered when the training and testing data sets are both from the same site and when the training data is from one site and the testing data from a different site.

Il y a un résumé en français ici.

Report Number
DRDC-ATLANTIC-TM-2008-029 — Technical Memorandum
Date of publication
01 May 2008
Number of Pages
42
DSTKIM No
CA033240
CANDIS No
532550
Format(s):
Electronic Document(PDF)

Permanent link

Document 1 of 1

Date modified: