By Topic

Automatic interpretation of sonar imagery using qualitative feature matching

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$33 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
D. M. Lane ; Dept. of Comput. & Electr. Eng., Heriot-Watt Univ., Edinburgh, UK ; J. P. Stoner

This paper reports on the automatic interpretation of sector scan sonar imagery. Previous work has resulted in a blackboard-based system, employing a mixture of image-, signal-, and rule-based processing to extract appropriate feature information from sonar scans. We here describe a system capable of carrying out classifications of observed objects based on available feature measures, such as size, shape, and gray level characteristics. The problem of determining feature measures which are invariant to changes in sonar setting, object position/orientation, and noise characteristics is addressed by using qualitative measures to describe object features during matching for recognition. Invariance comes from dynamically selecting the threshold values used to map the numerical feature values derived from the image data to these qualitative measures. Descriptions of the qualitative appearance of known objects are maintained as “exemplars.” Recognition therefore takes place by matching observed object descriptions to exemplars in either a constrained or unconstrained fashion. Descriptions are presented for the feature measures used, the quantitative-to-qualitative mapping, and the matching procedures, with results showing the discrimination provided by the feature measures, the changing numerical boundary values between qualitative attributes, and the overall success of the recognition processing for single sonar scans. The overall interpretation is shown to be 86% successful for objects viewed on different sonars in different conditions, provided features measures giving good discrimination between objects are employed

Published in:

IEEE Journal of Oceanic Engineering  (Volume:19 ,  Issue: 3 )