Abstract
In this paper, we consider the advantages that segmentation can give towards detecting sub-pixel point targets in hyperspectral imagery. We will show that by segmenting the image, we can find a better estimate for the covariance matrix; the resulting target estimation is vastly improved. A method to deal with edge points is discussed.
Original language | English |
---|---|
Article number | 25 |
Pages (from-to) | 103-114 |
Number of pages | 12 |
Journal | Proceedings of SPIE - The International Society for Optical Engineering |
Volume | 5617 |
DOIs | |
State | Published - 1 Dec 2004 |
Event | Optically Based Biological and Chemical Sensing for Defence - London, United Kingdom Duration: 25 Oct 2004 → 28 Oct 2004 |
Keywords
- Hyperspectral
- Segmentation
- Target detection
ASJC Scopus subject areas
- Electronic, Optical and Magnetic Materials
- Condensed Matter Physics
- Computer Science Applications
- Applied Mathematics
- Electrical and Electronic Engineering