Spherical and Nonspherical Aerosol and Particulate Characterization Using Optical Pattern Recognition Techniques

New methods that exploit the capabilities of the optical correlator and optical pattern recognition techniques are investigated for characterizing spherical and nonspherical aerosols and particles. Optical correlators utilizing the classical matched filter and the phase-only filter are introduced, and the special case of the opaque spherical particle is investigated. The increasing influence of the cutoff frequency on the optical correlator as the particles become smaller is examined. A holographic modification to the commercially available wedge-ring detector is described to replace the classical matched and phase-only filters, and some of the possible benefits are enumerated. The general field of optical neural networks is introduced and the specific area of optical heteroassociative processors is presented along with possible advantages of coupling the output from the optical correlator into an optical processor. The overall objective of size distribution determination in nearly real-time under conditions of high -particulate loading is specified. Experimental data are compared to analytical results utilizing the classical matched filter in a standard 4f optical correlator for sizing large opaque spherical particles with diameters between 150 μm and 350 mum. A modified 4f optical correlator utilizing the classical matched filter for small and medium opaque spherical particles with diameters between 5 μm and 100 μm was also investigated. Computer simulations and experimental data are presented for various ternary phase-amplitude filters used to size opaque spherical particles in the size range between 5 μm and 25 mum. Additional work includes developing and demonstrating a holographic ring detector capable of distinguishing spherical particles with diameters in the range from 2.5 μm to 100 μm and demonstrating the ability of the holographic wedge-ring detector to distinguish simultaneously and characterize the three standard classes of unclassified military smokes and obscurants. Digital and optical implementations of the holographic ring detector directly coupled into an optical neural network are studied as a means of increasing the speed of the decision process. This consists of testing various optical setups utilizing single and combination holographic optical elements, including demonstrating the ability to characterize these particles at very high speed.