A neural network approach for improved detector performance of spectral matched filters in hyperspectral imagery

Robert S. Rand

doi:10.1117/12.825748

17 August 2009 A neural network approach for improved detector performance of spectral matched filters in hyperspectral imagery

Robert S. Rand

Author Affiliations +

Proceedings Volume 7457, Imaging Spectrometry XIV; 74570T (2009) https://doi.org/10.1117/12.825748
Event: SPIE Optical Engineering + Applications, 2009, San Diego, California, United States

Abstract

The detection of sub-pixel materials in a hyperspectral scene is often accomplished using spectral matched filters or subspace projection. These methods rely on estimates of background second order statistics or subspaces in a scene that are usually based on either on global statistics of the entire scene or on adaptive local statistics. Global statistics have the disadvantage of including materials of interest in the background estimate and this implies the method assumes these materials occupy an insignificant portion of the scene. Adaptive methods that use a small number of samples surrounding the pixel of interest to estimate a background covariance eliminate much of this disadvantage, but this comes at the cost of significantly increasing computation time and potentially unstable estimates for some backgrounds. A number of spectral matched filter methods have been developed with increasing sophistication, but experience indicates that the method used to compute the background statistics may have a greater impact on overall detector performance. This research investigates the use of a neural network approach to estimate the background statistics needed for certain spectral matched filters requiring global statistics. The context of the effort is terrain, urban, and shallow-water mapping using hyperspectral imagery, where the materials of interest inherently occupy a significant portion of a scene or where certain background classes have problematic second-order statistics. Results of experiments within this context are shown.

Citation Download Citation

Robert S. Rand "A neural network approach for improved detector performance of spectral matched filters in hyperspectral imagery", Proc. SPIE 7457, Imaging Spectrometry XIV, 74570T (17 August 2009); https://doi.org/10.1117/12.825748

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available