Submitted to: Poultry Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/18/2010
Publication Date: 6/15/2010
Publication URL: http://hdl.handle.net/10113/45796
Citation: Yang, C., Chao, K., Kim, M.S., Chan, D.E., Early, H., Bell, M. 2010. Machine vision system for online wholesomeness inspection of poultry carcasses. Poultry Science. 89(6):1252-1264. Interpretive Summary: USDA/FSIS inspectors currently inspect and remove birds that exhibit signs of bacterial contamination or disease from poultry processing lines. Inspectors must visually examine 35 bird-per-minute (bpm). Subject to human variability, this inspection process makes inspectors prone to fatigue and repetitive injuries, and also limits the maximum possible output speed for the processing plants. The need to increase production throughput to satisfy increasing chicken consumption and demand places additional pressure on both chicken producers/processors and the U.S. food safety inspection program. Automated line-scan imaging inspection systems can help alleviate this pressure as well as provide further improvements to food safety and quality inspection. This paper presents the development of a line-scan spectral imaging system for high-speed wholesomeness inspection of chickens. In-plant testing of the system on a 140 bpm commercial processing line demonstrated over 99 percent accuracy in sorting contaminated or diseased birds from wholesome birds. Automated online pre-sorting of young broiler chickens on commercial processing lines is an ideal application for this spectral line-scan imaging system. For this purpose, the spectral line-scan imaging technology has been recently reviewed and approved by the USDA-FSIS Risk and Innovations Management Division. Commercialization of this system for industry use will be the first application of spectral line-scan imaging technology for a food safety inspection task.
Technical Abstract: A line-scan machine vision system and multispectral inspection algorithm were developed and evaluated for differentiation of wholesome and systemically diseased chickens on a high-speed processing line. The inspection system acquires line-scan images of chicken carcasses on a 140 bird-per-minute processing line and is able to automatically detect individual birds entering and exiting the field of view of the camera, locate a specified Region of Interest (ROI) for spectral image analysis, and produce a decision output for each bird. The same spectral line-scan imaging system was used for hyperspectral data acquisition/analysis to develop the multispectral detection and differentiation algorithm and for multispectral implementation of the algorithm for real-time online inspection on the processing line. Results showed that effective multispectral inspection could be achieved by analysis of a selected ROI across the breast area from images at the 580 nm and 620 nm wavebands. Overall system performance was evaluated during two 8-hour shifts in which the system inspected over 100,000 chickens, with system results compared to FSIS inspector tallies of wholesome and systemically diseased birds for that same time period. During system verification, the system accurately classified wholesome and systemically diseased chickens that were observed by a veterinarian posted beside the system to perform real-time identifications of the same birds. The high accuracy of the results demonstrated that the spectra line-scan imaging system and multispectral detection and differentiation algorithm can be effectively used for online high-speed pre-sorting applications for young broiler chickens.