IR Signature Measurement Using Neural Spectroscopy (PR01-480)

"This proposal addresses the need for fast, rugged, portable, and highly automated coating assessment for military and aerospace vehicles. Spectroscopic identification and quantification of emissivity variations induced by coating degradation is perhapsthe most versatile and unambiguous method to achieve this goal, but these systems are often plagued by complex data interpretation requirements. One solution which has not been implemented for this type of work in the mid-IR is neural spectroscopy, wherecomponent features in aggregate spectra can be isolated and identified using properly trained neural nets. The proposed development will result in a very flexible automated analysis tool for chemical processes which exhibit spectral features in themid-IR, which encompasses an extraordinarily broad class of applications. We envision applications of this product development in the materials, pharmaceutical, semiconductor, and petrochemical industries and currently do business with several potentialcustomers. The proposed product will complement our existing line of temperature measurement equipment quite well."