Submitted to: Proceedings International Seminar on Nondestructive Evaluation Technology
Publication Type: Proceedings
Publication Acceptance Date: June 18, 2001
Publication Date: June 20, 2001
Citation: Park, B., Windham, W.R., Lawrence, K.C., Buhr, R.J. 2001. Hyperspectral imaging for identifying fecal and ingesta contaminants. Proceedings International Seminar on Nondestructive Evaluation Technology. Interpretive Summary: The Food Safety and Inspection Service personnel visually inspect poultry carcasses for the presence of feces. There is a zero tolerance for the presence of fecal material on carcasses being processed. Any carcass detected with fecal material must be removed from the line and treated. The presence of fecal material on the skin indicates a problem during processing (broken digestive system and spilling of contents). Feces varies in color, consistency, and composition depending on location in the digestive system. To aid in the inspection of fecal contamination, we developed a hyperspectral imaging system for identification of fecal and ingesta that contaminate the surface of carcasses. Algorithms were developed for image processing and identification of fecal contamination. The results of this research show that the imaging system can identify uncontaminated carcasses from the contaminated carcasses. The system has potential for real-time identification of fecal contamination on poultry carcasses in the poultry processing plant.
Technical Abstract: A hyperspectral imaging system with line-scan mode including camera with visible and far red prism-grating-prism spectrograph, fiber optic line lighting, motorized lens control, and hyperspectral image processing software was developed, to identifiy of fecal and ingesta contamination on poultry carcasses. Both spectral and spatial image data between 400 and 900 nm with 512 spectral bands, which were created by 4x2 binning image acquisition, were acquired from various fecal and ingesta contaminated carcasses. Four dominant wavelengths (434, 517, 565, 628 nm) were selected by principal component analysis from visible/near- infrared spectroscopy to apply for wavelength selection of hyperspectral images. A calibration model for the imaging system was developed from calibration lighting sources (HgAr, Kr, Lasers) for accurate band selection from hyperspectral images to identify spatial and spectral characterization of fecal and ingesta contaminants. Hyperspectral image processing algorithms, specifically band ratio of dual-wavelength (565/517) images and histogram stretching, were effective on the identification of fecal and ingesta contamination of carcasses under the various sample conditions. This algorithm can be further applied for real-time identification of fecal contamination on carcasses with further development of real-time multispectral image processing algorithms in the processing line. This paper presents the research results that hyperspectral imaging can be used effectively for detecting feces (from duodenum, ceca, and colon) and ingesta on poultry carcasses. Test results indicated that the detection accuracy was 97.3 percent for linear and 100 percent for non-linear histogram stretching algorithm.