A vision-based environment analysis algorithm for optical threat detection

PDF

Authors
  1. Paquin, V.
  2. Morin, A.
  3. Fortin, J.
Corporate Authors
Defence R&D Canada - Valcartier, Valcartier QUE (CAN)
Abstract
This paper describes an image analysis algorithm that is part of an advanced optical threat detection system installed on parol vehicles. This algorithm processes wide field of view (WFOV) image sequences captured by a color camera to select regions of the environment that are most likely to contain threats and that will be later inspected by a slower narrow field of view (NFOV) active sensor. A set of features on which the region selection will rely is extracted. Given the limited availability of operation context sequences, a heuristic approach was preferred over learning methods for feature selection and weighting. The algorithm has primary application in scenes where little a priori knowledge is available. The proposed approach was validated on real-world images. Preliminary tests results showed that the suggested method can effectively reduce the number of scan areas and that using several redundant cues enables the system to gracefully degrade.
Report Number
DRDC-VALCARTIER-SL-2010-074 — Scientific Literature
Date of publication
01 Dec 2009
Number of Pages
7
DSTKIM No
CA033834
CANDIS No
533264
Format(s):
Document Image stored on Optical Disk

Permanent link

Document 1 of 1

Date modified: