Tripwire Detection for Landmines

Budd, Chris and Stockie, John (1998) Tripwire Detection for Landmines. Canadian Industrial Problem Solving Workshops > 2nd IPSW [Calgary 1/6/1998 - 5/6/1998].

Full text available as:

PDF - Requires Adobe Acrobat Reader or other PDF viewer.

Abstract/Summary

Pre-processed Radon transform-based detection methods were identified as the most viable technique. There are some issues that arose when testing the various methods. A) The 'log' edge detection in Matlab seemed better than the default 'sobel' method, but is still worth considering other edge detection algorithms that might deal more effectively with noisy or grainy image. B) The algorithm works best when the wire is horizontal and spans the length of the image. If constrained to thin-strip type images then a simple solution is to run the algorithm twice, once with the image and a second time with the same image rotated through 45 degrees.

Item Type:	Study Group Report
Study Group:	Canadian Industrial Problem Solving Workshops > 2nd IPSW [Calgary 1/6/1998 - 5/6/1998]
Company Name:	ITRES Research Limited
Industrial Sector:	Aerospace and defence
Additional Contributors:	Jessop, Chris and Kavousian, Shabnam and Lamoureux, Michael and Madden, Joshua and Nyland, Edo and Rout, Bruce and Salauskas, Arunas and Saunders, David and Vetter, Bill and Tomoda, Satoshi
ID Code:	124
Deposited By:	Richard Booth
Deposited On:	24 January 2008

Problem Statement

Canada is a global leader in the decommissioning of land mines. ITRES is working with the Defense Research Establishment Suffield to develop techniques for visually recognizing trip-wires from images. The problem is to do this in real-time.

Archive Staff Only: edit this record