Submitted to: International Journal of Poultry Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: September 4, 2005
Publication Date: September 21, 2005
Citation: Windham, W.R., Smith, D.P., Berrang, M.E., Lawrence, K.C., Feldner, P.W. 2005. Effectiveness of hyperspectral imaging system for detecting cecal contaminated broiler carcasses. International Journal of Poultry Science. 4(9):657-662. Interpretive Summary: On-line visual and manual inspection of fecal contaminated broiler carcasses is conducted by the Food Safety and Inspection Service (FSIS) to ensure a safe product to consumers. The inspection processes is both labor intensive and prone to human error. The USDA Agricultural Research Service has developed a method and a hyperspectral imaging system to detect feces and ingesta on poultry carcasses. To further improve the method it is necessary to test the size or weight of the feces that the system can detect. The system was tested to identify cecal contaminants on broiler carcasses from small to large spots. The imaging system correctly identified 100% cecal mass applied. Within a cecal spot, some of the pixels in the image or not identified as a contaminant. These pixels are a mixture of the contaminant and the carcass skin. Detection of these mixed pixels is essential for contaminant identification of small spots (less than 10 mg). Detection of mixed pixels in large contaminants is not significant to overall contaminant identification.
Technical Abstract: Broiler processing may result in fecal contamination of the surfaces of carcasses. Fecal contaminants on broiler carcasses are prohibited due to the potential presence of bacterial pathogens. The objective of this study was to determine the effectiveness of the hyperspectral imaging system to detect cecal contamination of known mass. On each of three replicate sample days, twenty-four eviscerated, pre-chilled broiler carcasses were collected from a commercial processing plant. Broiler carcasses were cut longitudinally into contra-lateral halves using a sanitized saw. Cecal contents from the same flock were also collected and used to contaminate carcass. Contents of multiple ceca were combined, homogenized and used to contaminate carcass. Carcass halves were imaged uncontaminated and cecal contents (10, 50, or 100 mg) were applied to the carcass half, and then re-imaged. Cecal detection results varied due to contaminate detection threshold. The imaging system correctly identified 100% cecal mass applied at a threshold of 1.00 and 1.05 but also incorrectly identified 252 and 65 carcass features, respectively that were not contaminates (false positives). False negative were only associated with the 10mg mass and a detection threshold of 1.10. The percentage of the cecal ground truth detected also varied due to the detection threshold. Averaged across cecal mass, the percentage of the cecal ground truth detected was 74, 55 and 35% for the 1.00, 1.05 and 1.10 threshold, respectively. The percentage of contaminated pixels not detected were a spectral mixture of ceca and uncontaminated skin. Detection of mixed pixels in small contaminants (ie. 10mg and less) or an aggregate of several single-pixels is essential for contaminant identification. Detection of mixed pixels in large contaminants is not significant to overall contaminant identification.