Source: UNIV OF MARYLAND submitted to
DETECTION OF BONE FRAGMENT IN POULTRY MEAT USING COMBINED X-RAY AND LASER IMAGING
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
TERMINATED
Funding Source
NRI COMPETITIVE GRANT
Reporting Frequency
Annual
Accession No.
0186041
Grant No.
98-35503-7022
Project No.
MD-ENBE-7190
Proposal No.
2000-05602
Multistate No.
(N/A)
Program Code
(N/A)
Project Start Date
Sep 15, 2000
Project End Date
Mar 31, 2004
Grant Year
2000
Project Director
Tao, Y.
Recipient Organization
UNIV OF MARYLAND
(N/A)
COLLEGE PARK,MD 20742
Performing Department
BIOLOGICAL AND RESOURCES ENGINEERING
Non Technical Summary
With the United States producing over 13.2 billion pounds of boneless chicken meat annually and consumers increasingly demanding safe, high quality boneless meat at a low cost, the accurate and efficient detection of bone fragments and other hazards in poultry meat has great urgency. Current x-ray technology addressing these problems has very limited success (over 30% detection errors) mainly due to its inability to consistently recognize bone fragments in meat of uneven thickness. This project will develop a technology that integrates x-ray imaging with laser 3-D imaging to compensate for the uneven thickness of poultry meat and to enhance the x-ray accuracy in bone fragment detection and augment the inspection system's capability to rapidly and accurately detect bone fragments or hazards. This research will be conducted in three stages. First, we will determine x-ray absorption coefficients and specify optimal settings for bone fragment detection. Next, we will develop a high resolution laser 3-D imaging system, determining variations in thickness of poultry fillets at real-time. Finally, we will combine x-ray and laser 3-D imaging techniques, producing integrated images, developing image processing algorithms for bone fragment recognition. The results from this study will advance the x-ray technology in food inspection and promise to improve the quality, safety, and cost-effectiveness of US poultry products.
Animal Health Component
(N/A)
Research Effort Categories
Basic
30%
Applied
70%
Developmental
(N/A)
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
50132602020100%
Knowledge Area
501 - New and Improved Food Processing Technologies;

Subject Of Investigation
3260 - Poultry meat;

Field Of Science
2020 - Engineering;
Goals / Objectives
The goals will be evaluated by the following aims: to develop an effective imaging inspection technique and an integrated dynamic vision system to detect bone fragments in de-boned poultry meat; to overcome the obstacles of x-ray absorption variations and false detection due to uneven meat thickness and false patterns; to recognize major types of bone fragments from pulley bones, rib bones, thin fan bones, solid structure bone pieces, and sharp fragments embedded in meat fillets; to make correct decisions to remove meat pieces containing bones from processing lines to ensure the quality and safety of boneless poultry products.
Project Methods
Comprehensive research on both hardware in x-ray physics and sensing electronics and software image processing algorithms will be conducted for differentiating bones from non-bone poultry materials. X-ray absorption of different poultry materials including various types of bones, mussels, fat, blood vessels and clots, and skins will be studied to obtain sensing parameters and contrast imaging signals. A special high-contrast x-ray imaging system will be developed to scan meat objects on a conveyor belt. To obtain instantaneous meat curvature profile, a high-resolution 3-D laser range imaging system will be developed to produce thickness image which will be then converted to x-ray intensity images using pre-determined x-ray absorption coefficients and corresponding equations. Both x-ray and laser intensity images will be combined and transformed into thickness-invariant x-ray image. Furthermore, image pattern recognition algorithms will be developed to identify bones, filter noise, and reach detection decisions. A control system will be developed to remove bone-inclusive meat pieces from lines and achieve automated vision inspection.

Progress 09/15/00 to 03/31/04

Outputs
Through previous studies, we have developed a new technology that uses the synergism of x-ray and laser imaging for sensitive detection of bone fragments in de-boned poultry meat. The new technique overcomes the weakness of conventional x-ray technology and eliminates false image patterns by x-ray absorption compensation. Through intensive experiments, our lab results have proved the theory and demonstrated that the method is able to detect various frequent and hard-to-detect bone fragments, particularly fan bones and small bone fragments, regardless of their locations in chicken fillets As a result, with this synergetic sensing approach, the concept sought to eliminate false patterns and enable thickness-invariant pattern recognition. The experimental result demonstrated that this concept eliminated the effect of uneven thickness of chicken meat, and was able to produce thickness-invariant x-ray images and detect bone fragments embedded in chicken fillets. The result has opened the way to the next in-depth research leading to a new automated detection technology.

Impacts
Bone fragments and physical contaminants in chicken fillets are the number one consumer complaint in the poultry industry. With the U. S. producing 13.2 billion pounds of boneless chicken meat annually, accurate detection of bone fragments and other hazardous materials in poultry meat becomes essential. These hazardous materials pose lawful threats to the food quality and safety to consumers, especially to small children and the elderly. The research will advance x-ray technology in food inspection and will meet the strong demand from the poultry industry for increased food quality and safety of boneless chicken meat, reduced customer complains, and improved cost-effective productions. It will benefit both the poultry industry and the consumers in general. This research is a step forward to the potential new technology with positive impact to food quality and safety and cost-effective productions.

Publications

  • Tao, Y. 2002. Machine Vision and Productivity: Techniques and collaborations in fruit/food quality sorting, X-ray detection of hazardous materials in foods, and other high-speed machine vision on-line inspections. Forum on Ag & Bio System Engineering Development Strategy. Yangling, China, June 26. (Invited). Tao, Y. 2002. Machine vision and imaging for inspection of agricultural materials. Zhejiang University, Hangzhou, China, June 28. (Invited).
  • Chen Z. and Y. Tao. 2001. Food safety inspection using from presence to classification object-detection model. Pattern Recognition. Vol. 34(2001) 2331-2338.
  • Chen Z. and Y. Tao. 2001. Multi-resolution local multi-scale contrast enhancement of X-ray images for poultry meat inspection. Applied Optics. Vol. 40(8):1195-2000.
  • Chen Z. and Y. Tao. 2001. Subband correlation characteristics of Daubechies wavelet representation. Optical Engineering. Vol. 40(3):362-371.
  • Chen Z., Y. Tao, and X. Chen. 2001. Wavelet-based adaptive thresholding method for image segmentation. Optical Engineering. Vol.40(5):868-874.
  • Tao Y., Chen Z., H. Jing, and J. Walker. 2001. Internal inspection of deboned poultry using x-ray imaging and adaptive thresholding. Trans. of ASAE. Vol.44(4):1005-1009.
  • Tao, Y., X. Chen, and H. Jing. 2001. Integrated techniques for accurate detection of bone fragments. 36th National Meeting on Poultry Health and Processing. Oct.17-19, 2001. Ocean City, MD. (Invited).
  • Cheng, X. and Y. Tao. 2001. Infrared Imaging and Wavelet-Based Segmentation Method for Apple Defect Inspection. ASAE paper 01-3109.
  • Chen X. and Y. Tao. 2001 Minimum-Error Thickness Compensation for X-ray Inspection of Deboned Poultry Meat. ASAE paper 01-3062.
  • Jing, H. and Y. Tao. 2001. Scattering Laser Line Image Recovery for 3D Poultry Meat Processing. ASAE paper 01-3161.
  • Tao, Y. and J. Ibarra. 2000. Thickness invariant X-ray imaging detection of bone fragments in de-boned poultry meat - model analysis. Transaction of ASAE. Vol.43(2):453-459.
  • Tao, Y. 2000. Combined x-ray and laser imaging for sensitive detection of bone fragments for poultry de-boning lines. July 9-12, 2000. ASAE paper 003144. St Joseph, MI.
  • Chen, X. and Y. Tao. 2000. High resolution imaging detection of physical contamination in de-boned poultry meat on processing lines. July 9-12, 2000. ASAE paper 003123. St Joseph, MI.
  • Chen, Z. and Y. Tao. 2000. Contrast enhancement of x-ray imaging and postprocessing for food internal inspection. July 9-12, 2000. ASAE paper 003124. St Joseph, MI.
  • Jing H. and Y. Tao. 2000. Multiple camera real-time depth detector for poultry meat. July 9-12, 2000. ASAE paper 003126. St Joseph, MI.


Progress 01/01/02 to 12/31/02

Outputs
Comprehensive experiments have proved the concept and demonstrated that the combined x-ray and laser imaging technique has been able to detect frequent hard-to-detect bone fragments including fan bones and fractured rib and pulley bone pieces (but not cartilage yet) regardless of their locations and uneven meat thickness without being affected by skin, fat, and blood clots or blood vines. At the current stage, the research has resolved some engineering issues and has produced the synchronized x-ray and laser operating system, simultaneous image processing algorithms, and related software engineering in previously proposed tasks. Further research is needed including optical distortion corrections, algorithm optimization, non-bias image transformation, automated hazardous item rejection, and other closely related technical issues for a successful new technology in detection of hazardous items in foods and biomaterials.

Impacts
Bone fragments and physical contaminants in chicken fillets are the number one consumer complaint in the poultry industry. With the U. S. producing 13.2 billion pounds of boneless chicken meat annually, accurate detection of bone fragments and other hazardous materials in poultry meat becomes essential. These hazardous materials pose lawful threats to the food quality and safety to consumers, especially to small children and the elderly. The research will advance x-ray technology in food inspection and will meet the strong demand from the poultry industry for increased food quality and safety of boneless chicken meat, reduced customer complains, and improved cost-effective productions. It will benefit both the poultry industry and the consumers in general.

Publications

  • Tao, Y. 2002. Machine Vision and Productivity: Techniques and collaborations in fruit/food quality sorting, X-ray detection of hazardous materials in foods, and other high-speed machine vision on-line inspections. Forum on Ag & Bio System Engineering Development Strategy. Yangling, China, June 26. (Invited).
  • Tao, Y. 2002. Machine vision and imaging for inspection of agricultural materials. Zhejiang University, Hangzhou, China, June 28. (Invited). Chen Z. and Y. Tao. 2001. Food safety inspection using "from presence to classification" object-detection model. Pattern Recognition. Vol. 34(2001) 2331-2338.


Progress 01/01/01 to 12/31/01

Outputs
The research continues and is approaching the goals of the project. The x-ray lab including hardware and software has been reestablished and set up running, although some delays occurred due to the lab building construction. Experiments are accelerated for the detection and image processing software development. The breakthrough includes concurrent x-ray and laser parallel processing and seamless synthesis of images toward the hazard detection. The test data shows the expected results as we proposed. During this period with extensive research efforts, we have published six (6) refereed journal articles and presented 10 papers at national/international conferences about our research results.

Impacts
Bone fragments and physical contaminants in chicken fillets are the number one consumer complaint in the poultry industry. With the U. S. producing 13.2 billion pounds of boneless chicken meat annually, accurate detection of bone fragments and other hazardous materials in poultry meat becomes essential. These hazardous materials pose lawful threats to the food quality and safety to consumers, especially to small children and the elderly. The research will advance x-ray technology in food inspection and will meet the strong demand from the poultry industry for increased food quality and safety of boneless chicken meat, reduced customer complains, and improved cost-effective productions. It will benefit both the poultry industry and the consumers in general.

Publications

  • Tao Y., Chen Z., H. Jing, and J. Walker. 2001. Internal inspection of deboned poultry using x-ray imaging and adaptive thresholding. Trans. of ASAE. Vol.44(4):1005-1009.
  • Chen Z. and Y. Tao. 2001. Food safety inspection using from presence to classification object-detection model. Pattern Recognition. Vol. 34(2001) 2331-2338.
  • Chen Z. and Y. Tao. 2001. Multi-resolution local multi-scale contrast enhancement of X-ray images for poultry meat inspection. Applied Optics. Vol. 40(8):1195-2000.
  • Chen Z. and Y. Tao. 2001. Subband correlation characteristics of Daubechies wavelet representation. Optical Engineering. Vol. 40(3):362-371.
  • Chen Z., Y. Tao, and X. Chen. 2001. Wavelet-based adaptive thresholding method for image segmentation. Optical Engineering. Vol.40(5):868-874.
  • Tao, Y. and J. Ibarra. 2000. Thickness invariant X-ray imaging detection of bone fragments in de-boned poultry meat - model analysis. Transaction of ASAE. Vol.43(2):453-459.
  • Tao, Y., X. Chen, and H. Jing. 2001. Integrated techniques for accurate detection of bone fragments. 36th National Meeting on Poultry Health and Processing. Oct.17-19, 2001. Ocean City, MD. (Invited).
  • Chen X. and Y. Tao. 2001 Minimum-Error Thickness Compensation for X-ray Inspection of Deboned Poultry Meat. ASAE paper 01-3062.
  • Jing, H. and Y. Tao. 2001. Scattering Laser Line Image Recovery for 3D Poultry Meat Processing. ASAE paper 01-3161.
  • Tao, Y. 2000. Imaging analyses of biological materials for food safety. Invited by American Institute for Medical and Biological Engineering (AIMBE). Washington DC. March 3, 2000. (Invited)
  • Tao, Y. 2000. Increase values of agricultural products. Invited by US-China Science, Technology, and Business. McLean, VA. Nov. 18, 2000. (Invited).
  • Tao, Y. 2000. Combined x-ray and laser imaging for sensitive detection of bone fragments for poultry de-boning lines. July 9-12, 2000. ASAE paper 003144. St Joseph, MI.
  • Chen, X. and Y. Tao. 2000. High resolution imaging detection of physical contamination in de-boned poultry meat on processing lines. July 9-12, 2000. ASAE paper 003123. St Joseph, MI.
  • Chen, Z. and Y. Tao. 2000. Contrast enhancement of x-ray imaging and postprocessing for food internal inspection. July 9-12, 2000. ASAE paper 003124. St Joseph, MI.
  • Jing H. and Y. Tao. 2000. Multiple camera real-time depth detector for poultry meat. July 9-12, 2000. ASAE paper 003126. St Joseph, MI.
  • Tao, Y., X. Chen, H. Jing, Z. Chen, and J. Walker. 2000. Highly-sensitive imaging detection of physical contaminant on poultry de-boning lines. In Proceeding of USDA Food Safety Consortium Annual Meeting. Pp.63-68. Sept.17-19, 2000. Fayetteville, AR.