ALGORITHM DEVELOPMENT WITH VISIBLE/NEAR-INFRARED SPECTRA FOR DETECTION OF POULTRY FECES AND INGESTA

Abstract

The USDA Agricultural Research Service has developed a method and a hyperspectral imaging system to detectfeces (from duodenum, ceca, and colon) and ingesta on poultry carcasses. The method first involves the use of multivariatedata analysis on visible/near-infrared (Vis/NIR) reflectance spectra of fecal and uncontaminated skin samples forclassification of contaminates and selection of key wavelengths. Four dominant wavelengths (434, 517, 565, and 628 nm)were identified by intensity of principal component (PC) loading weights. Key wavelengths were validated on hyperspectralimages of contaminated broiler carcasses. Specifically, with a quotient of 565 nm/517 nm, 100% of the fecal contaminateswere detected in a limited population of broilers fed a corn/soybean meal diet. The objectives of this research was to validatethe 565 nm/517 nm quotient to classify uncontaminated skin from feces/ingesta with broilers fed corn, milo, or wheat dietsand to investigate the use of single-term linear regression (STLR) to select key wavelengths for classification. Feces (N =369) and uncontaminated broiler breast skin (N = 96) were analyzed from 440 to 880 nm. The overall accuracy of detectingcontamination for any type of feed with the 565 nm/517 nm quotient was 99% with 16 uncontaminated skin samples classifiedas contaminates (false positive). STLR optimized a new quotient of 574 nm/588 nm, which classified 100% of contaminatescorrectly with no false positives. The shift in the denominator from 517 to 588 nm is possibly due to greater fecal colorvariation from broilers fed wheat or milo. In addition, dividing by 588 nm minimized the effect of lightness (L*) onclassification. The use of the STLR to scan the spectral data to find wavelengths correlated with the dependent variable isan alternative to selecting key wavelengths based on the intensity of PC loading weights. Although models from Vis/NIRspectroscopy and STLR performed well, they need to be validated on hyperspectral images of uncontaminated andcontaminated carcasses.