Computer-aided classification of the Citadel Trial sidescan sonar images


  1. Fawcett, J.A.
  2. Crawford, A.
  3. Hopkin, D.
  4. Couillard, M.
  5. Myers, V.
  6. Zerr, B.
Corporate Authors
NATO Undersea Research Centre, La Spezia (ITALY);University of Western Ontario, London ONT (CAN);Defence R&D Canada - Atlantic, Dartmouth NS (CAN);GESMA, Brest (FRANCE)
In October 2005, Defence Research and Development Canada (DRDC) Atlantic participated with the Nato Undersea Research Centre (NURC) and the Groupe D’études sous-marines de l'Atlantique (GESMA, France) in a joint trial (CITADEL) with the 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 this report, an automated detection method is used to detect these objects as well as clutter images (false alarms). The extracted small images are then considered for classification. The detection and classification results will be considered in terms of minelike/clutter (Receiver Operating Characteristic curve) and in terms of individual target type classification (confusion matrices). The features will be computed on the basis of traditional shadow/highlight segmentation and also using the image pixels themselves in a manner similar to some of the techniques used in facial recognition. The classification performance will be considered as a function of the type of features used: shadow features only, shadow and highlight features and these features in combination with the image-based features. In addition, the data collected at the second site provides a verification of the results obtained from the first site.
Target identification;Classification algorithms;Computer aided classification;Automatic target detection
Report Number
DRDC-ATLANTIC-TM-2007-162 — Technical Memorandum
Date of publication
01 Jul 2008
Number of Pages
Electronic Document(PDF)

Permanent link

Document 1 of 1

Date modified: