Decision surfaces of binary tests in hyperspectral detection

PDF

Authors
  1. Bouffard, F.
Corporate Authors
Defence R&D Canada - Valcartier, Valcartier QUE (CAN)
Abstract
This report is first and foremost an analysis of the decision surfaces associated with common detection statistics used for hyperspectral target detection. The intention is to clarify the process leading to detection using those statistics and show where their differences and similarities lie. The large number of available detection statistics, along with their numerous parameters, makes formal comparisons complicated and sometimes produce ambiguous or uncertain conclusions. This document tries to establish a solid framework on which to base this analysis and then demonstrates how detection statistics can be grouped in distinct classes based on their decision surface's geometry. In doing so, possibly new detection statistics (or methods that can be used to build them) are exposed.

Il y a un résumé en français ici.

Keywords
Decision test;decision surface;hyperspectral target detection
Report Number
DRDC-VALCARTIER-TR-2013-098 — Technical Report
Date of publication
01 Apr 2013
Number of Pages
48
DSTKIM No
CA046618
CANDIS No
806995
Format(s):
Electronic Document(PDF)

Permanent link

Document 1 of 1

Date modified: