Submitted to: Transactions of the ASAE
Publication Type: Peer reviewed journal
Publication Acceptance Date: 11/21/2005
Publication Date: 8/8/2006
Citation: Lawrence, K.C., Windham, W.R., Smith, D.P., Park, B., Feldner, P.W. 2006. Effect of broiler carcass washing on fecal contaminant imaging. Transactions of the ASABE. 49(1):133-140. Interpretive Summary: The Food Safety & Inspection Service has a mandate that requires processed poultry carcasses to be free of fecal contaminants before they go into the chiller tanks. We are developing an imaging system that will be able to detect the contaminants on carcasses as they move down the processing line. Earlier research has shown that we can detect contaminants applied on the carcasses with an imaging system known as hyperspectral imaging. Since the industry uses mechanical bird washing cabinets to clean and remove feces, this research examined the effect of bird washing on the ability to detect the contaminants. Results indicate that the imaging system successfully detected 98 % of pre-washed contaminants, but incorrectly detected 36 % of the post-washed stains from the washed-off feces. These stains are not considered contaminants, so we must make modifications to the imaging system to not pick up the stains as contaminants.
Technical Abstract: The Food Safety and Inspection Service has mandated that there be no fecal contaminants on poultry carcasses when the carcasses enter the chiller tank because of a risk of cross-contamination of pathogens. Since the inception of the hazard analysis, critical control point mandate, the poultry industry has increased the amount of water use per bird to ensure compliance with this mandate. This paper reports on research to develop a method to identify fecal contaminants on poultry carcasses with a hyperspectral imaging system and to evaluate the effectiveness of this system for detecting contaminant residuals and stains on mechanically washed carcasses. The imaging system easily identified fecal contaminants (98%) prior to mechanical washing but also incorrectly identified 196 carcass features that were not contaminants (false positives). However, almost half of the false positives came from only five carcasses. Results confirm the feasibility of using such a system for detecting fecal contaminants. For washed carcasses, the hyperspectral imaging system significantly detected about 45% of the cecal stains and 34% of the duodenum stains. Contaminant wash times of 8 or 12 s. did not significantly affect either the observation of visible stains or the hyperspectral detection of those stains. However, the hyperspectral imaging system detected significantly more cecal stains at the longer contaminant exposure time of 12 min. than the shorter exposure time of 2 min. For the hyperspectral contaminant detection, no other contaminant exposure-time effects were observed. Based on the interpretation of the FSIS regulation of fecal contaminants, fecal stains are not normally considered contaminants. Therefore, to comply with the FSIS regulation while not adversely affecting the processing plants production, the hyperspectral imaging system should be modified to prevent detection of fecal stains.