HYPERSPECTRAL IMAGING APPLICATION FOR FOOD SAFETY ENGINEERING

Authors: Park, Bosoon; Lawrence, Kurt; Windham, William; Smith, Douglas

Submitted to: Meeting Abstract
Publication Type: Proceedings
Publication Acceptance Date: 9/14/2004
Publication Date: 10/11/2004
Citation: Park, B., Lawrence, K.C., Windham, W.R., Smith, D.P. 2004. Hyperspectral imaging application for food safety engineering [abstract]. Proceedings Korean Society for Food Engineering. October 11, 2004, Seoul, Korea. p.21-37.

Interpretive Summary: Food safety is an important issue for public health and reductions in the potential health risks to consumers from human pathogens in food are crucial for food safety. While a number of factors can influence bacterial contamination of chicken carcasses, one of the leading causes is fecal contamination at the poultry processing plants. Improper handling of carcasses and equipment that is not good working condition can contribute to the problem. The goal of food safety engineering is to develop new knowledge, technologies and systems for the detection and prevention of various types of contamination in foods. Since hyperspectral imaging technique has been demonstrated to be a potential tool for poultry safety inspection, particularly fecal contamination, a hyperspectral image processing method was developed for rapid and accurate detection of fecal contaminants on broiler carcasses. This new imaging technology can improve the FSIS poultry safety inspection program, especially HACCP, by incorporating scientific testing and systematic detection of fecal contamination in poultry processing plants.

Technical Abstract: A pushbroom hyperspectral imaging system including area scan camera, prism-grating-prism spectrograph, quartz halogen lighting, motorized lens control, and a hyperspectral image processing software was developed for detection of fecal and ingesta contamination of poultry carcasses. A calibration model was developed using various lighting sources (HgAr, Kr, and Lasers) for accurate wavelength selection from hyperspectral images to identify spatial and spectral characteristics of fecal and ingesta contaminants. Three different type of feces from different portions of the digestive tracts (duodenum, ceca, colon), and ingesta collected from the carcasses of broiler fed corn, milo, or wheat with soybean meal were used as contaminants. Hyperspectral image processing algorithms, specifically simple image ratio of two wavelengths, 565-nm image divided by 517-nm image, and thresholding effectively identified fecal and ingesta contaminants on broiler carcasses. The highest accuracy for classifying contaminants was 92.3% for corn fed carcasses when a spectral angle mapper (SAM) classifier was used. The overall mean accuracy and corresponding mean kappa coefficient to classify fecal and ingesta contaminants were 90.1% and 0.89. These algorithms can be further applied for real-time identification of fecal contamination on poultry carcasses for on-line poultry safety inspection.