No abstract

Anomaly detection for hyperspectral imaging is typically based on the Mahalanobis distance. The sample statistics for Mahalanobis distance are not resistant to the anomalies that are present in the sample pixels. Consequently, the sample statistics do not estimate the corresponding population parameters accurately. In this paper, we will present an algorithm for hyperspectral anomaly detection based on the Mahalanobis distance computed using robust statistics which are estimated based on the minimum generalized variance of the sample pixels. Numerical results based on actual hyperspectral images will be presented.

To facilitate in-depth hazard prediction models, we must understand the spectral properties of expulsion plumes from conventional weapon attacks. Precise data on the spectral absorption of three chemical weapon agent simulants, in the infrared regime, are required to properly determine the mass of simulant in expulsion plumes from field demonstrations and small scale tests. Data for triethyl phosphate (a Soman simulant), triethyl phosphite (a Sarin simulant), and tributyl phosphate (a VX simulant) are presented. A thermal evaporation cell was designed and built that incorporated features that are not commercially available.

This paper describes a statistically based efficient algorithm for the laboratory spectral calibration of a low-resolution terrestrial hyperspectral imaging camera operating in the visible range. Didymium oxide is used as a wavelength standard. The observed periodic spatial distortion in the wavelength abscissa of the hypercube is removed automatically through statistical modeling and the calibration of the wavelength abscissa is determined using a numerical method. The performance and cost of this algorithm in removing the spatial distortion are assessed quantitatively.

No abstract