Predicting synthetic aperture sonar performance with augmented sonar imagery


  1. Fawcett, J.A.
Corporate Authors
Defence Research and Development Canada, Atlantic Research Centre, Halifax NS (CAN)
The prediction of synthetic aperture sonar/sidescan sonar detection performance for a specific sonar survey is a topic of much interest. There are many ways to define performance. In this report, the detection and false alarm rates of specified automatic target recognition (ATR) methods for minelike objects will be the performance statistics of interest. Previous researchers have investigated the expected sonar performance as a function of image and sonar features. The approach taken in this report follows more closely the work of other authors on augmented reality. Synthetic target images are inserted into real sonar images obtained from a specific survey. Thus, the background sonar imagery and the clutter represents a survey exactly. The issue is making the inserted target realistic, in terms of its distortions and the noise characteristics of its shadow and highlight regions. These features need not be exactly correct but they need to be sufficiently correct to yield the correct ATR performance and behaviour. This report will describe the approach taken to insert targets with noise statistics which vary with respect to range and the surrounding image statistics. The approach will be applied to sonar data from a NATO Centre for Maritime Research and Experimentation (CMRE) trial (three different sites) and the predicted ATR performances will be compared to the actual performances. In addition, the performance model can be used to predict performances as a function of other p

Il y a un résumé en français ici.

sonar;performance;automatic target recognition
Report Number
DRDC-RDDC-2017-R020 — Scientific Report
Date of publication
01 Mar 2017
Number of Pages
Electronic Document(PDF)

Permanent link

Document 1 of 1

Date modified: