VIS/NIR SPECTROSCOPIC SYSTEM FOR HIGH-SPEED POULTRY CARCASSES INSPECTION

Authors: Chao, Kuanglin - Kevin Chao; Chen, Yud

Submitted to: Proceedings of the XXVIII International Congress of CIOSTA-CIGRV
Publication Type: Proceedings
Publication Acceptance Date: 10/11/2004
Publication Date: 10/11/2004
Citation: Chao, K., Chen, Y.R. 2004. VIS/NIR spectroscopic system for high-speed poultry carcasses inspection. In: Proceedings of the CIGR International Conference, Beijing, China. October 11-14, 2004 CDROM.

Interpretive Summary: Automatic processing systems are needed in the poultry industry to improve product safety, quality, consistency, and increase processing efficiency by increasing line throughputs and reducing wastewater output. One of the most important aspects of automation in poultry processing systems is the inspection of poultry carcasses. Currently, each chicken intended for sale to U.S. consumers is required by law to be inspected post-mortem by a USDA/FSIS (United States Department of Agriculture/Food Safety and Inspection Service) inspector for its wholesomeness. These inspectors visually examine the exterior, the inner surfaces of the body cavity, and the organs of each carcass for indications of diseases or defects. For effective inspection and occupational considerations, each inspector is limited to a maximum of 35 birds per minute. This current inspection system limits the production efficiency of processing plants that are seeking to satisfy increasing consumer demand for poultry products. One possible solution to this problem is for poultry processing plants to install on-line instrumental inspection systems that can accurately screen for wholesome carcasses. Inspectors then would only have to "re-inspect" questionable carcasses to insure that wholesome carcasses are not discarded. The Instrumentation and Sensing Laboratory, ARS, USDA, has developed a Vis/NIR spectroscopic system for on-line poultry carcass inspection. On-line trials of the visible/near-infrared chicken inspection system were conducted during a 5-day period in a poultry processing plant in Athens, Georgia. Spectra of 450 wholesome and 426 unwholesome chicken carcasses were measured. The instrument measured the spectra of veterinarian-selected carcasses on a processing line running at speeds of 140 and 180 birds per minute (bpm). The automatic inspection system was able to identify wholesome broiler carcasses at a 140 bpm speed with a 95 percent accuracy rate and unwholesome birds with a 92 percent accuracy rate. At 180 bmp, the accuracy rates for wholesome and unwholesome birds were 94 and 92 percent respectively. This information is useful to the FSIS, and poultry equipment manufacturers and processing plants.

Technical Abstract: The Instrumentation and Sensing Laboratory, Agricultural Research Service, U.S. Department of Agriculture, has developed a visible/near-infrared spectroscopic system for high-speed on-line poultry carcasses inspection. The inspection system, which was an area scanning system designed to measure the reflectance spectra of poultry carcasses in the visible to near-infrared regions, consisted of a fiber optic probe, a spectrograph, a spectroscopic charge coupled device detector, a quartz tungsten halogen light source, an industrial computer, and in-house developed software modules. On-line trials of the visible/near-infrared chicken inspection system were conducted during a 5-day period in a poultry processing plant in Athens, Georgia. Spectra (431-943 nm) of 450 wholesome and 426 unwholesome chicken carcasses were measured. The instrument measured the spectra of veterinarian-selected carcasses on a processing line running at speeds of 140 and 180 birds per minute. Results showed that this visible/near-infrared system can be used to differentiate between wholesome and unwholesome poultry carcasses at high speeds. For the 140 bird per minute line speed, the best model achieved classification accuracies of 95% for wholesome and 92% for unwholesome birds. For the 180 bird per minute line speed, the best model achieved classification accuracies of 94% and 92% for wholesome and unwholesome birds, respectively.