Submitted to: Transactions of the ASAE
Publication Type: Peer reviewed journal
Publication Acceptance Date: 6/18/2004
Publication Date: 10/1/2004
Citation: Kim, I., Kim, M.S., Chen, Y.R., Kong, S.G., 2004. Detection of skin tumors on chicken carcasses using hyperspectral fluorescence imaging. Transactions of the ASAE. Volume 47(5): 1785-1792. Interpretive Summary: U.S. Consumption of poultry and poultry products has steadily increased. Currently, individual poultry carcasses are subjected visual to inspection by Food Safety Inspection Service (FSIS) inspectors. Improvement of the existing inspection methods is needed. Development of rapid and reliable inspection systems to ensure safe production of poultry during post-harvest processing has become an important issue. The Instrumentation and Sensing laboratory has been involved in development of noninvasive optical sensing methods, as a rapid means to inspect chicken carcasses in the processing plants. This paper reports the use of hyperspectral fluorescence imaging techniques for detection of skin tumors on the chicken carcass. Algorithms using spectral and spatial information were devised to automatically detect chicken carcasses with skin tumors. This information is useful to the FSIS and poultry processing plants.
Technical Abstract: This paper presents a method for detecting skin tumors on chicken carcasses using hyperspectral fluorescence imaging data, which provide both spectral and spatial information. Since these two different kinds of information are complementary to one another, it is necessary to exploit them in a synergistic manner. Chicken carcasses are examined first using spectral information and results are used to determine candidate regions for skin tumors. Next, a spatial classifier selects the real tumor spots from the candidate regions. It was shown that the method detects chicken carcasses with tumors, but failed to detect some of the tumors that were small. This study uncovered meaningful spectral bands for detecting tumors using hyperspectral fluorescence images. A detection system that is built on this concept can increase detection rate, and the processing time can be reduced, because the procedure for detection is simplified by using a limited number of features to maintain low computational complexity. The method and findings can be employed in implementing customized chicken tumor detection systems.