Submitted to: ASABE Annual International Meeting
Publication Type: Proceedings
Publication Acceptance Date: 3/17/2006
Publication Date: 8/8/2006
Citation: Yoon, S.C., Lawrence, K.C., Smith, D.P., Park, B., Windham, W.R. 2006. Bone fragment detection in chicken breast fillets using back-illuminated structured light. ASABE Annual International Meeting. Portland, OR, July 9-12,2006. Paper No. 063017. Interpretive Summary: Bone fragments in poultry meat are significant physical hazards to the consumer, may cause injury, and thus, need to be detected. Commercial processors lose customers and spend considerable resources each year for insurance claims and legal fees from broken bones in de-boned meat. The predominant technology for bone detection is X-ray imaging which is based on projecting ionizing radiation through a meat sample onto an image detector. Optical imaging which is based on non-ionizing radiation at the visible and near-infrared spectral range of light has not been explored to measure the internal and functional information of uncooked or cooked poultry products. The objective of this study was to develop a viable optical imaging technique for detection of broken bones in uncooked skinless chicken breast meat. The study found that bones embedded in compressed 1-cm thick samples were always detectable by an optical imaging system. The study also found that shadows back-illuminated on the surface of the meat sample were unique when a bone was embedded. These findings can contribute to a food safety effort with developing a system to detect broken bones on-line and in real-time.
Technical Abstract: This paper is concerned with the detection of bone fragments embedded in de-boned skinless chicken breast fillets using transmittance optical imaging at the visible and near-infrared (VIS/NIR) wavelength range. Naturally occurring materials on the surface of breast meat, such as fats and connective tissues, often hinder the detection of bone because these materials’ transmittance values at the VIS/NIR wavelength range are very similar to embedded bones. Transmittance images of chicken breast fillets also suffer from blurring and noise, and thus a low contrast, due to strong multiple-scattering properties of the fillets. For a feasibility test, sample thickness was controlled to be 1-cm thickness by a compressed sample holder. Multi-imaging modes (transmittance/reflectance/both sides of sample) were studied to find the best possible images showing bones. A structured line light source with a transmittance mode was utilized to exploit unique scattering patterns characteristic of embedded bones. The study found that bones in 1-cm thick compressed samples were detectable when the multi-modes of imaging were fused in the final decision making. Experimental results with chicken breast samples and bone fragments are provided.