Estimation and discrimination of aerosols using multiple wavelength LWIR lidar

This paper presents an overview of recent work by the Edgewood Chemical Biological Center (ECBC) in algorithm development for parameter estimation and classification of localized atmospheric aerosols using data from rapidly tuned multiple-wavelength range-resolved LWIR lidar. The motivation for this work is the need to detect, locate, and discriminate biological threat aerosols in the atmosphere from interferent materials such as dust and smoke at safe standoff ranges using time-series data collected at a discrete set of CO2 laser wavelengths. The goals of the processing are to provide real-time aerosol detection, localization, and discrimination. Earlier work by the authors has produced an efficient Kalman filter-based algorithm for estimating the range-dependent aerosol concentration and wavelength-dependent backscatter signatures. The latter estimates are used as feature vectors for training support vector machines classifiers for performing the discrimination. Several years of field testing under the Joint Biological Standoff Detection System program at Dugway Proving Ground, UT, Eglin Air Force Base, FL, and other locations have produced data and backscatter estimates from a broad range of biological and interferent aerosol materials for the classifier development. The results of this work are summarized in our presentation.