<OL> <LI> Evaluate the feasibility of imaging & other technologies for identification of surface contamination on poultry carcasses & processed meat. <LI>Determine optimal spectral regions & data preprocessing treatments for imaging contaminated meat. <LI>Develop algorithms to identify the site & type of contamination. <LI>Develop a real time on-line system for contaminate identification. <LI>Implement system to assist regulatory & industry personnel in attaining compliance, optimizing efficiency & providing a safe product.
A real time on-line system will be developed to detect contamination of ingesta, feces, and other contaminants on poultry carcasses & processed meats. Visible &/or NIR spectroscopy will be evaluated with diode array technology using multispectral imaging techniques & with two-dimensional images using hyperspectral imaging techniques to determine optimum operating wavelengths and conditions. To aid in the evaluation of hyperspectral instrumentation & image processing techniques, collaboration with Institute of Tech. Development, Stennis Space Center, MS., & ARS Instrumental & Sensing Lab, BARC will be established. Once the optimum technique & spectral region are determined, algorithms will be developed to classify the type & location of the contaminate in real time. The system will then be calibrated & validated for carcasses contaminated in a processing plant environment at commercial line speeds. Cooperation will be established with FSIS personnel & processing plant management to ensure the technology meets their needs & standards.