High Reliability Flame Detector

Corvid Technologies proposes to produce a low false positive optical flame detector (OFD) to address the Navy’s need to reduce costs and loss of confidence associated with OFD false alarms. Corvid will utilize the basic classification scheme of current spectral based OFDs and expand upon the possible spectral bands used during classification. To effectively extend this design, Corvid will collect spectral data from various flame and not flame sources via burn tests and off the shelf spectrometers. This data will be used to run feature selection methods to determine the most valuable spectral bands for classifying flames with an optical flame detector. Corvid will then produce a bread board prototype of the new band combinations utilizing our existing argon plasma detector. We will reobtain spectral data using our bread board prototype and train a classification tree to classify flames and common false alarms sources as either a flame or not a flame.