CANONICAL RELATIONS OF SUBSPACES IN MULTI-SENSOR DATA ANALYSIS


Polat O. M. , Ozkazanc Y.

22nd IEEE Signal Processing and Communications Applications Conference (SIU), Trabzon, Turkey, 23 - 25 April 2014, pp.1219-1222 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Volume:
  • Doi Number: 10.1109/siu.2014.6830455
  • City: Trabzon
  • Country: Turkey
  • Page Numbers: pp.1219-1222

Abstract

In multisensor data analysis, scene details can be extracted via subspace methods without any prior information on the scene In these decomposition techniques, data is projected into a new space so that the information in the data is highlighted In this study, Principal Component Analysis, Independent Component Analysis and Minumum Noise Fractions method are applied to a multi-sensor data composed of radar, visible, and infrared images. Canonical correlations between these subspaces are investigated via Canonical Correlation Analysis. This equalization subspace offers a new point of view in the realnt of multi-sensor data analysis.