Submitted to: Proceedings of SPIE
Publication Type: Proceedings
Publication Acceptance Date: 10/4/2006
Publication Date: 11/12/2006
Citation: Liu, Y., Chao, K., Chen, Y.R., Nou, X., Chan, D.E., Yang, C. 2006. Detection of fecal / ingesta contaminants at slaughter plants from a number of characteristic visible and near infrared bands. Proceedings of SPIE. 6381:6381OU-1 - 6381OU-9.
Interpretive Summary: The Food Safety Inspection Service (FSIS) of U.S. Department of Agriculture (USDA) has adopted the Pathogen Reduction, Hazard Analysis and Critical Control Points (HACCP) systems, which require all meat and poultry plants to develop written sanitation standard operating procedures to show how they will meet daily sanitation requirements. Currently, FSIS inspectors use the established guidelines to identify fecal remains, the most likely source of pathogenic contamination, at surfaces of equipment, utensils, and walls at slaughter plants. Certainly, the verification is both labor intensive and prone to both human error and inspector-to-inspector variation. Therefore, the Agricultural Research Service of USDA has been developing hyper- and multi- spectral reflectance and fluorescence imaging techniques for the use in real-time on-line detection of fecal spots on chicken carcasses. One key factor in the successful applications is to have a few essential spectral bands, which not only reflect the chemical / physical information in the samples, but also maintain the successive discrimination and classification efficiency. In this study, a few selected visible and NIR bands were sought to explore the potential for the classification of fecal / ingesta (“F/I”) objectives from rubber belt and stainless steel (“RB/SS”) backgrounds. Spectral features of “F/I” objectives and “RB/SS” backgrounds showed large differences in both visible and NIR regions, due to the diversity of their chemical compositions. Such spectral distinctions formed the basis on which to develop simple three-band ratio algorithms for the classification analysis. Meanwhile, score-score plots from principal component analysis (PCA) indicated the obvious cluster separation between “F/I” objectives and “RB/SS” backgrounds, but the corresponding loadings did not show any specific wavelengths for developing effective algorithms. Furthermore, 2-class soft independent modeling of class analogy (SIMCA) models were developed to compare the correct classifications with those from the ratio algorithms. Results indicated that using ratio algorithms in the visible or NIR region could separate “F/I” objectives from “RB/SS” backgrounds with a success rate of over 97%. The visible/NIR spectroscopic technique shows promise for sanitation verification and poultry safety at processing plant.
Technical Abstract: The study presented a direct analysis of visible and NIR reflectance spectra for the classification of “F/I” objectives from “RB/SS” backgrounds. Spectral differences in both visible and NIR regions revealed a number of significant bands, which were used to develop simple 3-band ratio algorithms for discriminant analysis. The results showed that the three-band based algorithms could classify “F/I” objectives from “RB/SS” backgrounds with a success of over 97%, which was at least the same accuracy as those from the 2-class SIMCA models. Meanwhile, PCA was performed on both spectral sets, and the score-score plot showed a clear separation between “F/I” objectives and “RB/SS” backgrounds. However, the optimal loadings did not provide any specific characteristic bands that could further improve the classification rate. The finding of three visible or NIR bands is most promising in the development of simple goggle and binocular sensing system for in-situ inspection of fecal and ingesta contaminants at slaughter plants.