Image analysis – Applications – Animal – plant – or food inspection
Patent
1994-10-13
1998-08-11
Boudreau, Leo H.
Image analysis
Applications
Animal, plant, or food inspection
382164, 382165, 382203, 382282, 382257, 348 89, G06K 900, G06K 934
Patent
active
057938792
DESCRIPTION:
BRIEF SUMMARY
FIELD OF INVENTION
The present invention relates image analysis to locate a target section in an image. The invention has been developed for use in evaluation of meat quality parameters, particularly utilising contact-free determination, but the invention has wider application than this field alone.
BACKGROUND
In the meat industry it is becoming increasingly important for meat products to be accurately described to ensure that the correct products are directed to specific markets, maximum returns are gained for the different products and products can be changed over time to suit the requirements of the market. This is particularly so in the meat processing section of the industry which operates between the producers (farmers) and wholesale or retail consumers who purchase either whole carcases or boned-out cuts of meat in cartons. The requirements of the consumers (in various markets such as Australia, Japan, USA, Korea, EEC etc) are varied, with the potential for both large premiums and also heavy discounts for providing either the correct or wrong products respectively.
In the USA and Japan, for example, centrally controlled grading schemes have evolved to cater for this requirement by inspecting the carcases in the abattoir and assigning an overall grade to the meat products in the carcase which is based on several individual characteristics. The grading is based on an expert assessment of a cross section cut through the carcase at a specified site, usually between two particular ribs. This cut exposes several muscles, in particular the longissimus dorsi or ribeye muscle which is one of the more valuable portions of meat in the carcase, as well as intermuscular fat deposits, a cross section of the subcutaneous fat layer and any bones/ribs that may be present. Both US and Japanese grading schemes attach great significance in the grading schemes to the level of intramuscular fat contained within the ribeye (marbling), the colour of the ribeye muscle, the area of the ribeye and the colour of the intermuscular fat surrounding the muscles. These are all important quality measures although the ribeye area, intermuscular fat and subcutaneous fat thickness can also be used as a quantitative measure of the carcase by predicting the relative amount of meat in the whole carcase (i.e., lean meat yield %). In the US system the carcase is assigned a marbling category (slight, abundant, etc.) based on comparison to a set of photographs published by the U.S. Department of Agriculture (USDA), who are also responsible for training all graders within the industry. Meat colour is judged by reference to a few descriptive terms (bright, cherry red, etc.) and similarly for fat. Ribeye area is estimated in square inches but is mainly used to assign a yield grade rather than quality grade. Along with other characteristics (age, etc.) these descriptions are used to assign the carcase/meat into a few overall grades i.e., premium, choice, select, etc.
The Japanese grading system, for example, goes a stage further by having a reference set of artificially made "chips" which are used for comparative scoring of meat colour, fat colour and marbling. Thus, an individual carcase can be given a score of marbling (1-12, indicating no fat to abundant fat) meat colour (1-9, indicating light to dark) and fat colour (0-9, indicating white to dark/yellow) where the marbling score in particular has a great bearing on the value assigned to the meat.
In Australia, AUSMEAT have developed a Chiller Assessment Method which is similar to U.S. and Japanese methods. Usually, the assessor is also trained to make objective measurements of subcutaneous fat depth at a particular site with a rule, and ribeye area using either a grid and manual counting technique or a digitising tablet which computes the area from an outline traced around the ribeye boundary with a stylus.
A typical abattoir processing>500 carcases per day would need to train several skilled, senior personnel to be qualified chiller assessors to cope with assessing every carcase. Thes
REFERENCES:
patent: 4226540 (1980-10-01), Barten et al.
patent: 4413279 (1983-11-01), Gorl
patent: 4665554 (1987-05-01), Sternberg
patent: 5136906 (1992-08-01), Antonissen et al.
patent: 5170440 (1992-12-01), Cox
patent: 5172423 (1992-12-01), France
patent: 5194036 (1993-03-01), Chevalier et al.
patent: 5339815 (1994-08-01), Lin et al.
patent: 5457754 (1995-10-01), Han et al.
Choo et al., "Computation of Features of 2-D Polycurve-Encoded Boundaries," Systems, Man, and Cybernetics, 1989 International Conference, pp. 1041-1047.
Chen, Y.R., McDonald, T.P., and Crouse, J.D., "Determining Percent Intra-Muscular Fat On Ribeye Surface By Image Processing", ASAE Paper No. 89-3009, The American Society of Agricultural Engineers, St. Joseph, MI, Jun. 1989.
McDonald, Tim, et al. "Location and isolation of beef carcass ribeyes in cluttered scenes." (Paper No. 89-3005 presented at ASAE/CSAE meeting, Jun. 25-28, 1989, Quebec, Canada).
Barrett-Lennard David
Benn Alan
Hay Peter J.
Boudreau Leo H.
Meat Research Corporation
Mehta Bhavesh
LandOfFree
Image analysis for meat does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Image analysis for meat, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image analysis for meat will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-398146