Submitted to: New Food
Publication Type: Trade journal
Publication Acceptance Date: 7/1/2002
Publication Date: 9/1/2002
Citation: Lawrence, K.C., Windham, W.R., Park, B., Smith, D.P. 2002. Contaminant detection on poultry carcass surfaces. New Food. 5(3):21-24. Interpretive Summary: Although the United States has one of the safest food supplies in the world, an estimated 76 million persons contract food-borne illnesses in the U. S. each year. Citing a potential correlation between fecal contaminants and bacterial pathogens, FSIS has established a standard that requires the surfaces of meat and poultry carcasses during slaughter to be free of fecal contaminants. ARS is developing an on-line imaging system to inspect all carcasses for surface fecal contamination. Scientists are using a series of imaging technologies to identify fecal contaminated carcasses. The process starts with basic spectroscopic measurements of pure fecal material and poultry skin, meat, and fat. Advanced statistical techniques are then used to select several key spectral wavelengths which are highly correlated with the contaminants. Next, another imaging technology know as hyperspectral imaging was used to evaluate those key wavelengths on contaminated whole poultry carcasses. Hyperspectral analysis indicated that single wavelengths were inadequate for detecting fecal contaminants, but the ratio of two-wavelength images resulted in contaminant detection. Unfortunately, hyperspectral imaging is too slow for real-time contaminant detection. Therefore, a multispectral imaging system was used to collect images at the two key wavelengths which were verified by the hyperspectral system. Preliminary results of the multispectral system are very encouraging and early indications are the system should be able to operate at speeds up to 180 birds per minute with accuracies above 95%.
Technical Abstract: A real-time imaging system has been developed to detect fecal contaminants on poultry carcasses. The system, which will require two to three multispectral cameras to fully image the outside of a poultry carcass, has shown promise in early trials. The system was developed from spectroscopic first principles and utilizes synchronized multispectral images at several discrete wavelengths to identify the contaminants at poultry processing line speeds. Implementations of such a system should provide continuous monitoring of processed poultry and result in a safer food product by preventing carcasses with fecal contaminants from entering the chill tank and thus, minimize cross contamination. A systems approach to detecting contaminants on food surfaces in general has been used and U. S. and international patents on the method and technology have been submitted and are now allowed. The method and system: (1) identifies the light characteristics (spectral properties) of contaminants on the surface of food products; (2) develops image processing routines that accentuate the difference between the contaminant spectral features and the food spectral features; and (3) establishes a detection system for identifying surface contaminants in real-time based on those spectral properties and image-processing routines (Windham et al., 1999). Preliminary results of this system are very encouraging and early indications are the system should be able to operate at speeds up to 180 birds per minute with accuracies above 95%.