Statistical Methods for Analysis of Hyperspectral Anomaly Detectors
Abstract
Most hyperspectral (HS) anomaly detectors in the literature have been evaluated using a few HS imagery sets to estimate the well-known ROC curve. Although this evaluation approach can be helpful in assessing detectors rates of correct detection and false alarm on a limited dataset, it does not shed lights on reasons for these detectors strengths and weaknesses using a significantly larger sample size. This paper discusses a more rigorous approach to testing and comparing HS anomaly detectors, and it is intended to serve as a guide for such a task. Using randomly generated samples, the approach introduces hypothesis tests for two different kinds of data: (i) idealized homogeneous samples and (ii) idealized heterogeneous samples, where model parameters can vary the difficulty level of these tests. In (i), a simulation experiment is devised to address a more generalized concern the expected degradation of correct detection as a function of increasing noise on a given alternative hypothesis. In (ii), fundamental features of a spectral sample (magnitude and shape) are modeled separately so that strengths and weaknesses of competing detectors can be independently assessed for each feature. Additionally, detectors ability to suppress transition of regions in the imagery is assessed in (ii).
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 2007
- Accession Number
- ADA480123
Entities
People
- Dalton Rosario
Organizations
- United States Army Research Laboratory