Page Banner

United States Department of Agriculture

Agricultural Research Service

Title: Performance of Supervised Classification Algorithms of Hyperspectral Imagery for Identifying Fecal and Ingesta Contaminants

Authors
item PARK, BOSOON
item LAWRENCE, KURT
item Windham, William
item Smith, Douglas

Submitted to: Transactions of the ASABE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: September 14, 2006
Publication Date: December 16, 2006
Citation: Park, B., Lawrence, K.C., Windham, W.R., Smith, D.P. 2006. Performance of supervised classification algorithms of hyperspectral imagery for identifying fecal and ingesta contaminants. Transactions of the ASABE. 49(6): 2017-2024.

Interpretive Summary: Reduction in the potential health risks to consumers from human pathogens in food is an important food safety issue. When food pathogens cause human illness and death, there is a loss of productivity in those affected. Human foodborne illnesses also result in economic loss and damage to producers and the food industry. For the poultry industry, while a number of factors can influence bacterial contamination of chicken carcasses, fecal contamination at the poultry processing plants is one of the issues be considered. Scientific inspection methods play an important role in the detection and prevention of various types of contamination and in the production of safe foods. Since hyperspectral imaging may be used to detect fecal contamination of poultry, these experiments were conducted to determine the optimum image classification methods that may be used to identify fecal contaminants on poultry carcasses. This new imaging technology can improve the FSIS poultry safety inspection program by incorporating scientific testing and detection of fecal contamination in poultry processing plants.

Technical Abstract: In order to select the optimum classifier, the performance of six different supervised classification algorithms including parallelepiped, minimum distance, Mahalanobis distance, maximum likelihood, spectral angle mapper, and binary coding were investigated and compared for identifying fecal and ingesta contaminants of hyperspectral poultry imagery. A pushbroom line-scan hyperspectral imager was used for hyperspectral image acquisition with 512 narrow bands between 400 and 900-nm wavelengths. Three different feces from digestive tracts (duodenum, ceca, colon), and ingesta were used as contaminants. These contaminants were collected from the broiler carcasses fed by corn, milo, and wheat with soybean mixture. The classification accuracies varied from 62.9% to 92.3% depending on the classification methods. The highest classification accuracy for identifying contaminants from corn fed carcasses was 92.3% with a spectral angle mapper classifier. While, the accuracy was 82.1% with maximum likelihood method for milo fed carcasses and 91.2% accuracy was obtained for wheat fed carcasses when same classification method was employed. The mean classification accuracy for classifying fecal and ingesta contaminants obtained in this study was 90.3%.

Last Modified: 9/10/2014
Footer Content Back to Top of Page