Improved Pattern Recognition Techniques for Real Time Operation
Abstract
Passive Fourier transform infrared remote sensors have shown tremendous potential for use in the automated detection of atmospheric pollutants. To be maximally effective as an early warning alarm device, however, passive infrared sensors must be able to operate in a variety of infrared backgrounds and under conditions in which the background is rapidly changing. For economic reasons, the sensor should be the simplest device that can meet the aformentioned operational requirements. In the work described here, new signal processing strategies have been developed that allow target chemical species to be detected directly from time-domain infrared data. This methodology eliminates the Fourier transform computation through the development of digital filters keyed to the characteristic infrared frequencies of the target molecules. Through this approach, both the computational and instrumental requirements for the sensor are significantly reduced. The performance of this methodology will be illustrated with a variety of test data. Keywords: Dimethyl methyl phosphonate, Sulfur hexafluoride; Fourier spectroscopy; Standoff chemical vapor detection algorithm.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1989
- Accession Number
- ADA211395
Entities
People
- Gary W. Small
Organizations
- Iowa State University