Computer vision in the poultry industry

Authors: CHAO, KUANGLIN; PARK, BOSOON; KIM, MOON

Submitted to: Book Chapter
Publication Type: Book / Chapter
Publication Acceptance Date: 9/28/2011
Publication Date: 8/17/2012
Citation: Chao, K., Park, B., Kim, M.S. 2012. Computer vision in the poultry industry. Book Chapter. p. 330-351.

Interpretive Summary: To satisfy increasing consumer demand for poultry products, the chicken industry depends on automation to effectively process high product volumes for a variety of markets. More recently, research has sought to develop automated computer vision technologies that can effectively address food safety concerns, to facilitate or complement the work performed by human inspectors. Inspection of birds for wholesomeness and fecal contamination are two important areas currently included in the growing list of responsibilities performed at processing plants by food safety inspectors. To help address these concerns, researchers at the U.S. Department of Agriculture have been developing spectroscopy and spectral imaging technologies suitable for high-speed chicken processing lines. The work has led to rapid line-scan imaging methods for hyperspectral/multispectral inspection, now a feasible technology for automated online inspection of chickens. This chapter presents a brief discussion of automation currently used in poultry processing operations, early research to develop spectroscopic inspection methods, and the development of imaging inspection methods from whole-target filter-based and common-aperture systems to rapid spectral line-scan imaging systems. Two case studies are presented regarding online wholesomeness inspection and fecal contamination detection for freshly slaughter chickens, along with discussion of the benefits to poultry processors and consumers that can be realized with the use of these automated vision systems. Future trends for spectral imaging and automation for food safety inspection of poultry are also discussed.

Technical Abstract: Computer vision is becoming increasingly important in the poultry industry due to increasing use and speed of automation in processing operations. Growing awareness of food safety concerns has helped add food safety inspection to the list of tasks that automated computer vision can assist. Research to develop the technology has included multiple approaches including spectroscopy and spectral imaging. Case studies presented in this book chapter focus on recent research targeting line-scan imaging for two food safety applications in poultry processing operations: automated wholesomeness inspection and fecal contamination detection for freshly slaughtered birds. Current research has demonstrated the feasibility of spectral line-scan imaging to detect unwholesome conditions and fecal contamination on high speed bird processing lines. Successful commercialization of these technologies has great potential that benefits multiple aspects of poultry processing.