DK157579B - PROCEDURES AND PLACES FOR DETERMINING QUALITY CHARACTERISTICS OF INDIVIDUAL CREATURES - Google Patents

Description

-1-

DK 157579 B

The present invention relates to a method and plant for determining the quality characteristics of individual cattle carcasses.

5 At slaughterhouses, the individual carcass slaughter is classified, so that a price can be obtained between the producer and the slaughterhouse in accordance with the quality of the body. An educated classifier assesses the shape and fat cover of the individual carcasses and determines its category.

10 The classifier then assigns the carcass a grade in accordance with a classification system.

The system used within the EEC comprises five mold classes, designated E, U, R, 0 and P. In order to obtain the grade E (excellent), the body must have an excellent shape and there must be no shortcomings 15 in its essential parts. All profiles must be convex to super-convex, and the muscles must be exceptionally well developed. The thigh should be very full and the back wide and very thick forward to the shoulder, should be very full.

Grades U, R, 0 and P are given to less valuable bodies.

20 The P (poor) grade is thus given to bodies where the profiles are concave to very concave and the muscles are poorly developed. In these bodies, the thigh is poorly developed, the back is narrow with visible legs, and the shoulder is flat with visible legs.

To achieve a greater differentiation in pricing, each 25 class is divided into three subclasses. The class E- thus denotes a carcass located in the lower third of the E-class, while the class E + denotes a body in the upper third.

The system includes five obesity classes, 1-5. The grade 1 is given to bodies without or with low fat cover and without fat in the chest cavity. The grade 5 30 means that the entire carcass is covered with fat and that there are heavy fat deposits in the chest cavity. The thigh is almost completely covered with fat, so the fat joints are no longer clear. In the chest cavity the muscles between the ribs are penetrated by fat.

DK 157579B

-2-

Cattle carcasses representing the extremes of the scales can be relatively easily distinguished from one another. On the other hand, it takes a great deal of practice to safely classify bodies that differ only at one stage on the scale. This is especially true for bodies that occur in the middle classes, where good and bad properties occur in the mixture, and these must be weighed against each other. Unfortunately, the carcasses are predominantly in the 10 middle classes. This means that the classification can become uncertain with the risk of being settled on the wrong price group.

Another factor is that the classification only expresses the value of the carcass to a certain extent. The classification can only be regarded as a rough division of the bodies into different groups, which, in experience, can each represent a price category. In the same class, carcasses of different composition with regard to meat, fat and bone may appear very well, and thus with different kilogram prices.

However, the exact value of the individual carcass can only be determined by cutting the body and assessing the individual 20 cuts. It requires a great deal of work and assessment and is not possible for practical reasons. Among the qualities that influence the value of the carcass include the yield of salable meat, the distribution of meat in the body, the fat content in the body, the muscle fullness, the meat avenue, the fat color, the marbling, the darkness and the taste.

There is a need for a semi-or fully automated method that can replace the aforementioned subjective classification method with a safe objective classification of bovine moths. If possible, the method should also be able to provide information on other characteristics of the individual carcasses that affect the body's value.

30 In one publication, an attempt is made to provide an objective classification of cattle carcasses by video recording.

(S.E. Sørensen: "Possibilities for Application of Video Image Analysis in Beef Carcass Classification" in In Vivo Measurement of Body Composition in Meat Animals, Elsevier (1984), pp. 113-122).

-3-

DK 157579 B

The carcass is positioned in front of a camcorder on a dark background, and a video image is recorded and the gray values are stored in a 5 memory. When processing the values in a computer system, the carcass outline is determined. Various parameters, such as the width of the chest, middle, thigh, as well as body area, can be calculated on the basis of the outline, and it is suggested that the shape and meat-bone ratio can be assessed if the parameters are used in a formula obtained by regression analysis.

The gray value distribution within the perimeter of the carcass can also be calculated and it is proposed to evaluate the obesity class using the result in a formula obtained by regression analysis.

The publication mentions that the use of video image analysis for carcass classification is not at all simple and that several potential problems must be taken into account. It is concluded that a thorough optimization of the analytical procedures is necessary and that the method may need to include the third dimension in order to work optimally.

It is known to classify pig carcasses using the body weight of 20 and some meat and fat thickness measurements. They are converted to a meat percentage used in pricing. Thicknesses are measured using an instrument with a probe that an operator places at certain locations in the carcass.

It is also known to measure meat and fat percentages of livestock carcasses 25 using said instrument. However, an objective classification of cattle in accordance with the current systems cannot be carried out in this way, as there is no applicable correlation between such thickness measurements and classes.

Various methods have been proposed in the patent literature for obtaining measurement results that can express quality characteristics of individual pig or bovine carcasses and possibly use an objective classification.

-4-

DK 157579 B

French Patent No. 2,546,423 discloses an apparatus for automatically measuring classification parameters in carcass carcasses.

The apparatus comprises an arrangement consisting of a frame with a support surface for a spotted cattle carcass. A complicated mechanical system with mechanical sensors measures the carcass's essential dimensions. This apparatus requires a high degree of maintenance of the mechanical parts and the veterinary authorities may require 10 sterilization of each sensor between each measurement. The apparatus is unable to determine the obesity class or composition of cattle carcasses.

West German Patent No. 27 28 913 discloses an apparatus for grading pig carcasses in accordance with the system used in West Germany 15.

The apparatus includes a dark chamber equipped with a color video camera and some daylight sources. A spotted pig carcass is placed in the chamber in such a way that the cutting surface faces the camera and light sources, and parameters on the cutting surface such as meat and fat thicknesses are determined by the camera and processed in a microprocessor system. The camera detects the chamber wall as black, greasy as white and the flesh as red. In this way, the camera distinguishes between the different materials.

On the basis of the registration and the weight of the carcass, a quality class can be calculated in the computer of the apparatus and a mark with the class is automatically placed on the carcass by means of a stamping machine.

French Patent No. 2,545,010 also discloses an apparatus for grading pig carcasses. In this apparatus, the height of an ultrasound probe is adjusted by video recording of the cut surface of the mid-spotted 30 carcass. The camera's recording is also used to determine the thickness of the cut surface. These thicknesses, together with the thicknesses determined by the ultrasound probe, are included in the calculation that determines the class of the pig carcass.

-5-

DK 157579 B

The literature thus proposes various methods in which a video recording is used in semi or fully automatic classification of carcasses, but no one has provided a method for classification of cattle carcasses sufficiently precise in practice to be able to determine composition and shape class and obesity class in an objective manner. . Such a method would be very difficult to provide in comparison with the present objective classification of pig carcasses in which it is sufficient to determine a few meat and fat thicknesses. Even within the same category, carcass carcasses can be very disparate.

Furthermore, meat and fat thicknesses cannot always be determined using a video recording of the carcass's cut surface, because some slaughterhouses do not stain the bodies until after the classification.

It is the object of the present invention to provide a method and plant by means of which quality characteristics which are of importance to the value of the individual cattle carcasses can be accurately determined by means of a video system. Preferably, an accurate, objective determination of the body composition and the shape and obesity classes approved by the authorities must be provided.

This object is achieved by the method according to the invention, characterized in that a carcass or half-spotted body is placed in a light-shielding chamber in front of a light-emitting contrast surface, an image of a whole side of the carcass is aligned with an electronic camera facing the contrast surface whereas the carcass appears as a dark silhouette against the contrast surface, the carcass contour is determined by processing the recorded image using a data processing system, a calculation of parameters is performed in the data processing system on the basis of the particular contour for determining the quality characteristics of the individual body, and the particular quality characteristics of the carcass are communicated.

-6-

DK 157579 B

The method of the invention provides a very accurate picture of the contour of the individual carcasses as there is no illumination of the side facing the camera and the background appears as a luminous surface. A very high contrast is provided in the image · between body and background, avoiding reflections resulting from reflections of the front illumination on the body surface.

By the method of the invention, it has become possible to determine the composition and shape and shape class of the individual cattle carcasses in an accurate and objective manner under the production conditions on a slaughter line. By means of the method, the shape and obesity class have been determined with sufficient accuracy to replace the objective classification used today.

Other quality characteristics that are of significance to the value of the 15 individual bodies such as the meat distribution can also be determined on the basis of the individual body's contour. A more accurate pricing than hitherto can be provided on the basis of this data and the bodies can be used in more accordance with their actual properties.

The contour of the body is determined by processing the recorded image in a processing system. When performing calculations, information about the size and dimensions of the body can then be provided. The parameters can be related to the form classes. For example, the parameters can be inserted into a formula embedded in the computer system that reflects the shape class on the basis of some variables in the contour image.

In one embodiment, one or more dimensions of the carcass are determined on the basis of the recorded contour and recognizable points thereof. This reduces the amount of speech processing needed to determine the shape class. The dimensions obtained can be part of an algorithm that determines the shape class of the body.

The tail and neck portion of the body can be relatively easily determined by analyzing the contour, and by means of these two points of fixation the width of the body can be determined midway between the points and other desired places.

-7-

DK 157579 B

The contour obtained by the method according to the invention can also be used to provide a boundary line, a video recording of the body with front lighting. In this way, a very accurate boundary line is drawn between the body and the background, and irrelevant image content and error sources can be removed before analyzing the image content, for example before determining the fat areas. There is a correlation between the obesity class and the percentage fat cover.

10 By introducing a probe into the carcass, the video measurement can be supplemented and further information can be obtained on the current carcass, for example more accurate information on the composition, expected yield and muscle fullness. The probe can measure fat and meat thicknesses, and these measurements, together with the image recording information 15, can give a relatively accurate indication of the carcass's quality characteristics.

Accordingly, one embodiment is peculiar in that meat and fat thicknesses are recorded at insertion of a probe in the body at selected locations and the measurements are used as parameters in 20 data processing systems to determine quality characteristics. The probe may have a probe near the tip having a light emitter and a light receiver. Such a light reflection based meter is relatively accurate in determining the thicknesses.

To improve accuracy, thickness measurements can be made at 25 anatomically identical locations on the right and left half of the body, so that measurement accuracy can be increased by calculating the average thicknesses and / or checking for unacceptable deviations.

As mentioned above, the contour can be used to provide a boundary line in a front lighting recording.

In a particular embodiment, it is possible to provide a recording of an image of the individual body, at which the individual points of the image can be separated from the background with good accuracy and at which the value of the points reflects with good certainty whether the surface of the carcass constitutes the places in question. of meat or 35 fat. Thus, these data may form the basis for an objective determination of the obesity class.

-8-

DK 157579 B

The embodiment is characterized in that a second image of the carcass is recorded by means of the electronic camera under illumination of the side of the body facing the camera, that the recorded image and the particular contour are processed in the data processing system for masking the background in said second image and separation of meat and fat areas on the body surface and that the areas obtained are used as parameters in the data processing system in determining the body's quality properties.

An optical distinction between meat and fat areas on the surface of bovine carcasses can be difficult to perform for natural reasons, such as fat discoloration, obstruction on the meat, light reflections on the surface 15, and illumination in lighting and camera sensing affinity under slaughter conditions.

It has been found possible to reduce these disturbances to a level where they affect the determination of the class of the individual bodies to a lesser extent. The reduction is achieved by defining a sub-image of the recorded image by means of the particular contour of the individual body, and the light reflection values of the image elements in said sub-image are statistically processed for separation of the meat and fat areas.

In the present embodiment, it is important for accuracy to define the sub-image and treat the light reflection values herein statistically to provide a meat-fat limit value. The light reflection values above this limit value may represent grease ranges. The number of such values expresses the area of the fat body of the actual body. Along with other measured parameters, the 30 fat tire area can be included in an algorithm that assigns the body to one of the five fat classes used.

The body may conveniently be illuminated with several light sources, thereby reducing shadow effects and providing a more uniform illumination on the body surface.

-9-

DK 157579 B

A preferred embodiment is characterized in that the side of the body facing the camera is illuminated with diffuse light during the registration of said second image. This gives a very uniform illumination of the body surface without shadow effects caused by the surface structure. In addition, mirrors of lamps and the like are avoided. in the surface. The individual measured light reflection values from the body in this embodiment provide an almost true picture of the body surface's reflectivity.

10 For the present subject, a color camera may be used as this may separate the red meat areas from the white to yellow fat areas.

However, due to the statistical speech processing of the light reflection values, a black and white camera is preferably used for the recording.

15 The light reflection value of each picture element in the recording is then represented by only one number. In order to accurately record the reflection values, a grayscale scale with a certain number of steps is used. Conveniently, a scale with, for example, 256 steps is used.

Even at a high number of grayscale levels, the meat and fat areas of the 20 black and white recorded image can be difficult to separate. Therefore, an illumination is preferably used when recording, which gives a greater difference in the light reflectivity of meat and fat than in white light illumination. For example, UV light can be used which gives different fluorescence in meat and fat surfaces.

Preferably, green light illumination is used as it is extinguished on red surfaces such as meat while reflected well on white or whitish surfaces such as grease.

The light emitting contrast surface used in the method of the invention may be a frosted glass surface illuminated from the rear.

30 During the first recording, the carcass is unlit on the side facing the camera. The body therefore appears as a black silhouette on a light background that is made up of the matte glass plate. This silhouette forms the contour that can be used to determine the shape class.

DK 157579B

-10-

In the two-shot embodiment, it also defines the area where fat is to be sought in another recording, and thus can serve as a mask for the second recording. This recording can be unmasked in a computer program by the first recording, so that it will only include reflection values from the carcass itself.

In order to obtain figures representative of the individual body's obesity class, a sub-picture can be defined within the recording of the carcass side. The boundary of the sub-image can be determined by the dimensions of the 10 individual body. Easily recognizable points can be used to determine the boundary line of the image.

In one embodiment, a partial image whose lower boundary line is the horizontal line passing through the apex of the contour curve along the neck and forelegs of the body is used.

The recording can also be cut across the horizontal line through the back leg of the body, which has a predetermined length in the recording, for example 10 pixels.

The light reflectance values in each image element in the defined sub-image represent a meat or fat range, depending on the criterion used for the statistical processing of the data. As a criterion, a threshold value obtained by speech processing of the light reflection values is preferably used. This divides the data into two groups. Registered light reflection values below the threshold fall in the group that gives a measure of carcass area 25. Reflection values above the threshold belong to the second group that gives a measure of the fat area.

The threshold value is preferably determined on the basis of the frequency of the pixels at the different reflection values.

At these frequencies, a discrimination analysis can be made to determine the threshold. The threshold value is preferably calculated as the light reflection value at which the variance of the frequency of the pixels is maximum. This provides the best possible separation between two overlapping groups.

-11-

DK 157579 B

On the basis of the relative or absolute sizes thus calculated of the fat cover of the individual bodies, and preferably also on the basis of the fat thickness measured by the above-mentioned probe gauge, the obesity class of the individual carcass can be determined.

Within the found fat ranges, the body color class (fat color) - if desired - can be determined by the average measured color.

The method according to the invention may consist of: a carcass or a half, stained carcass is placed in a light-shielding chamber in front of a light-emitting contrast surface, a first image of an entire side of the body is recorded with an electronic camera facing the contrast surface, while the body appears as a dark silhouette against the contrast surface, the side of the carcass facing the camera is illuminated, 15 - another image of the body side is recorded with an electronic camera, while said side is illuminated, the contour of the carcass is determined by processing the registered first image in a data processing system, a partial image of it registered second image is defined by means of the particular contour and the light reflection values of the elements of said sub-image are processed to separate meat and fat areas, meat and fat thicknesses - if desired - measured at selected locations of the body by insertion of a probe into the body, 25 - a calculation of the parameters is performed in data the processing system on the basis of the contour, areas and thicknesses that have been devised to determine the particular quality properties of the individual body and the specific properties of the body are communicated.

In calculating the composition, shape and obesity class, and other quality characteristics such as yield, available data on the individual body, such as weight, are also preferably taken into account.

The invention also relates to a plant for use in determining the quality characteristics of individual cattle carcasses. The plant is peculiar in that it comprises -12-

DK 157579 B

a light shielding chamber with a light-emitting contrast surface and means intended to place a carcass or half, speckled carcass in front of the contrast surface, 5 - an electronic camera facing the contrast surface to record an image of a whole side of the body while appearing as a a dark silhouette against the contrast surface, a data processing system adapted to determine the contour of the body in processing the recorded image, which system 10 is based on the particular contour adapted to perform a calculation of parameters for determining quality characteristics of the individual body, and a data channel for communicating the particular quality properties of the body * 15 The light emitting contrast surface may have substantially uniform light density. The plant may comprise an instrument with a probe for introduction into the carcass for recording meat and fat thicknesses. The plant may further comprise an arrangement for illuminating the side of the carcass facing the camera and a program in the data processing system to unmask the background of the second image and to separate meat and fat areas in a recorded image.

The system may be designed to provide diffuse light. It may comprise a frosted translucent plate, behind which are arranged several light sources.

The information obtained by the method according to the invention, in addition to the said quality characteristics, can also be used for automatic control of a transport and treatment system, for example for automatic sorting of carcasses and parts thereof.

The method and plant of the invention can be used to classify various categories of cattle. In the present description, cattle shall mean calves, cattle, young bulls, bulls, heifers and cows and sheep (including lamb).

The invention will now be described in more detail with reference to the accompanying drawing, in which: FIG. 1 shows a system for classifying cattle carcasses from above with a positioned body, -13-

DK 157579 B

FIG. 2 the system of FIG. 1, and FIG. 3 shows the system of FIG. 2 and 3 with a half, speckled carcass.

The system comprises a stainless steel cabin 1. Two doors 2 and 5 of the cabin can be opened and closed by means of pneumatic cylinder units 4. Through the cabin extends a sliding bar 5 · Along this rod a carcass suspended in a hook device can be advanced 19. Parallel to the guide rod there is in the cabin a pneumatic cylinder unit 6 with a carrier device 7 (Fig. 2) intended to pull the hook device and the suspended carcass to the registration area in the middle of the cabin. Here, there is a swing arm 9 operated by a cylinder unit 8 which is at the height of the rear leg of the body.

By activating the cylinder unit 8, the arm 9 pushes the carcass along the sliding bar until the arm reaches the final position, in which the carcass has taken a correct position for registration in the middle of the cabin. The arm 9 positions the entire carcass side to the camera as the two rear legs of the carcass abut the arm 9 in the final position (Fig. 1).

At a subsequent stroke of the cylinder unit 6, the carrier device 20 7 is intended to pull the hook device and the body out of the cabin and along the rod 5. until they are again driven by the chain conveyor system or other transport systems of the percussion line.

A frame 10 is pivotally suspended in the roof of the cabin so that it can pivot to an active position when actuated by the piston rods by two cylinder units 11 located at the bottom of the cabin. Upon impact, the lower part of the frame is moved away from the cabin wall 1a. Approximately in the center of the frame is a thin, horizontal bar 12 attached to the vertical sides of the frame.

30 A frosted glass plate 13 is placed in front of the cabin wall 1a which faces the frosted side away from the wall to reduce light reflections. The glass plate extends substantially throughout the height and width of the cabin. Immediately behind the glass plate 13 is a diffuser in the form of a milky white acrylic plate 13a. Between this plate and the cabin wall 1a are arranged a plurality of fluorescent tubes 14 with a density such that a uniform light density is achieved on the glass plate 13. When the tubes are switched on, the glass plate in the cabin appears as a uniform milky white light surface.

-14-

DK 157579 B

In front of the wall 1b of the cabin is a corresponding frosted glass plate 15 and an acrylic plate 15a which extends substantially throughout the height and width of the cabin. Between the glass plate and the wall lb there are a number of green fluorescent tubes 16 arranged at such a density that the glass plate appears as a substantially uniform light surface when the tubes are switched on.

At each of doors 2 and 3 is mounted a lamp 17 consisting of a reflector, two green vertical fluorescent lamps and a diffuser. The lamps are directed to the area of the cabin where a carcass is shown.

In a cutout in the middle of wall 1b is a CCD camera 18 of the type CCD. The camera is directed against the opposite wall 1a of the cabin, so that the carcass 19 can be accommodated (Fig. 1). The field of view of the camera is limited by a mask.

15 The camera is connected to a computer system containing an image recording unit and a data processing unit. The system additionally includes a control unit for the mechanical components of the system, including the control cylinder units and a program which ensures that the functions are performed at the correct time and in the correct order.

20 The computer system includes external devices for recording additional data for the individual carcasses, for example a keyboard for manually entering the carcass numbers or codes, a weight arrangement for recording the carcass weight, and a probe for recording meat and fat thicknesses one or more places on the carcass.

25 The quality characteristics of a carcass can be recorded as follows in the plant described:

A carcass is fed to the door 2 of the conveyor line conveyor system, then pulled by means of the carrier unit 7 of the cylinder unit 6 along the rod 5 to approx. center of cabin 1. Doors 2 and 3 are closed by means of cylinder units 4.

-15-

DK 157579 B

The cylinder unit 8 is actuated, whereby the arm 9 guides the hook assembly and carcass 19 to the exact position shown in the middle of the cabin. The cylinder units 11 are then actuated so that the lower 5 of the frame 10 pivots a distance away from the wall 1a. The carcass thereby comes into contact with the rod 12. In the obtained resting position, one side of the body faces the camera 18.

By means of the camera 18, a recording of the side of the carcass is made while the backlight from the fluorescent lamps 14 is turned on. Next, the front lighting consisting of the fluorescent lamps 16 and the lamps 17 is turned on and a second recording of the carcass side is made by means of the camera. The backlight is also turned on for this second recording. The recordings are sent electronically to the computer system, in which they are immediately processed as described below.

After the second recording, the tubes 16 and the lamps 17 are extinguished. The doors 2 and 3 are opened by the cylinder units b. 5 · 20 When the body is outside the cabin, an operator enters the stroke number and color class into the computer system. The operator also measures meat and fat thicknesses using a probe that he inserts in several places on the body. The body is then handed over to the slaughter line's usual transport system for further treatment.

25 The first recording is processed in the computer system. A program defines the contour curve of the carcass on the basis of a significant increase in the recorded reflection values occurring on the boundary between the carcass and the background.

The contour curve is used when calculating the composition and shape class of the carcass 30 in the computer. The convexity of the thigh is calculated, for example, by the length of the three horizontal line pieces connecting points on the contour of the back of the thigh and a vertical reference line. When the midline is relatively longer than the other two, the shape of the thigh is convex.

-16-

DK 157579 B

The contour curve also defines the area in which the carcass is and is used for unmasking the second recording in the computer system.

In a speech processing, this area is cut off at the top and bottom. Above 5 is the boundary line at the Y coordinate in which the sum of the number of pixels within the curve in the direction of the X axis becomes 10 or more for the first time when moving from high Y values to decreasing Y values.

Below is the boundary line at the Y coordinate, in which a horizontal 10 line intersects the curve at two points only when moving from low Y values to increasing Y values at the neck region.

The closed curve provided in this way in processing the first recording defines the area within which meat and fat must be recorded in the second recording with front lighting.

In processing the image elements in the second recording, a masking of the image elements is made which falls outside the range determined by the first recording and the above processing. The masking is done by a program in the computer system.

Thus, in the system memory is stored a number of numbers, in which every 20 numbers represents a reflection value for a point on the carcass outer surface that can be detected by the camera. A high value corresponds to a fat range, while a low value represents meat. An intermediate value may, in some recordings, mean meat and in other cases fat, depending on natural variations such as aging or contamination of the light sources, fat color, uniform reflection conditions, etc.

The stored quantity is subjected to a treatment which defines more securely than at a fixed threshold, which numbers represent meat and which numbers represent fat.

DK 157579 B

-17-

The processing of the number is preferably done by expressing it in a histogram with grayscale values as the basis and the frequency as a function thereof. The histogram can be normalized. The variance of a number of grayscale-5 values in the grayscale range is calculated. The maximum separation between meat and fat is achieved by the gray tone threshold value, where the variance is greatest. The ratio of the number of numbers that stands for fat to the number of numbers in the whole quantity indicates the percentage of the surface covered by fat.

10 The fat percentage obtained and the fat thickness determined by the probe gauge are used in the calculation of the carcass's obesity class. The calculations used include information on the carcase weight and possibly also shape.

On the basis of the measurements and calculations, an additional 15 quality characteristics of the carcass can be determined, such as the muscular fullness of certain muscles, which are calculated by registered weight, contour and meat thickness.

Thereafter, the computer system can print a label which the operator can adhere to the body before it is again transported by the slaughter line 20 chain transport system. The banknote may contain information on the stroke number, date, category, weight, obesity class, shape class, body composition and codes of transport and treatment instructions with regard to the individual characteristics and quality characteristics of the body or parts thereof.

In FIG. 3, the figures refer to the same parts as in FIG. 1 and 2. 19a is a half speckled carcass that has been pulled into the cabin and rests on the rod 12 with the outside facing the camera. The other half of the splintered body is outside the cabin during video recording.

The procedure is the same as described above for an entire carcass (Figs. 1 and 2).

-18-

DK 157579 B

Example In this example, 2948 mixed-category carcasses are used (whole and spotted young bulls, bulls, bulls, whole calves, heifers and cows). Practiced 5 classifiers perform a subjective classification of the shape and obesity of the individual bodies according to the EEC classification system.

Furthermore, objective measurements are performed on the individual bodies by means of the system described above. Two video recordings are made to determine the contour of the body and the nature of the surface. Probe 10 is inserted into the body four locations to determine the meat and fat thicknesses.

389 bodies are cut into salable meat, fat trimmings and bones according to a standard procedure, and the composition of each carcass is determined.

The data obtained is processed in a computer unit by means of a program system called "Statistical Analysis System". In this way, information on constants and other criteria is obtained in the formulas to be used to determine the quality characteristics of other carcasses.

The formulas are obtained by multiple regression equations. The weight of the individual bodies is used in the formulas. In some of the formulas, a level correction may occur according to the category of the object.

Table 1 shows factors of importance in the formulas.

DK 157579 B

TABLE 1

Yield Fat Pro - Remember - Form - Obesity of meat (1) cents (2) fullness (3) class (4) class (5)

Video recording Weight / body length + + Weight / area ++

Forpartareal /

Body length + + + + Thigh tilt + + +

Thigh / body length convexity + + +

Body width / length = 0.25 +

Body width / length = 0.45 (+)

Probe measurement Meat thickness, fillet (+) (+) (+) (+) (+)

Fat Thickness, Filet (+) (+) (+)

Fat thickness, culotte + + + +

Level correction at:

Jersey +++++

Speckle + + + Gender / Age + + + +

Slaughter weight + + + + + (1) - Percentage of salable meat (including some fat) (2) - Percentage of fat removed if. standard cutting method (3) - Index determined by the fillet's cross-sectional area and the weight of the hindquarters (average 100) (4) - Quality marking as close as possible to the EUROP shape class (5) - Quality marking as close as possible to the EUROP obesity class -20-

DK 157579 B

After the formulas have been determined, the relevant data for each carcass is used in the formulas. The meat yield, fat percentage, muscle fullness, shape class and obesity class are calculated for each body. The values obtained are compared with the corresponding quality properties that have been determined by a classifier or by cutting.

Good agreement was found between the plant's determination and the subjective regulations.

Accuracy

Form class 1.01 subclass stage *

Obesity class 0.57 grade Meat yield 1, ¾%

Fat percentage 1.35%

Muscle density 2.92 units * three subclasses = one class

Tables 2 and 3 show the number of carcasses in each form and obesity class. The plant in question is very reliable for classifying bodies with mixed mold properties. Between 86 and 91% of these bodies are classified in the same form subclass found by the subjective assay, or differ from a subclass step at most. Only 2% of the body is classified into classes that are more than two subclass stages from the class defined by the classifier. The plant is also reliable for making an objective determination of the body's obesity class, with only 1.5% of the provisions deviating more than one class step from the subjective determination.

-21-

DK 157579 B

TABLE 2 Form class

Classification with the plant according to the invention +4/3 +3/3 +2/3 +1/3 Over- -1/3 -2/3 -3/3 -4/3 Number of% before

The classmate’s 6 p.m. pm. pm. pm. etc. pm. pm. pm. for +/- rating or stamp. 1/3 at

more meaning more p- 0 136 47 ---- 57 (93) p 0 o 7 18 29 16 - - - 70 (90) p + 1 1 16 58 61 34 6 177 86 0- 2 8 34 l40 174 80 14 1 - 453 87 0 5 10 71 229 290 130 17 0 0 752 86 0+ 1 4 30 142 272 166 23 0 0 638 91 R- 0 2 10 77 189 144 35 3 1 461 89 R 0 0 2 26 56 71 26 4 0 185 83 R + 0 2 3 5 22 23 15 5 0 75 (67) U- 0 0 1 3 6 14 10 2 0 36 (64) uo 00138 11 4o 27 (44) u + - 00032230 10 (50) E- - -0010400 5 (20) E - - -000101 2 (0) E + ---- 00000 0 alt 9 28 177 705 1153 688 164 22 2 2948% 0 1 6 24 39 23 6 1 0 100 86.3

• DK 157579 B

-22- TABLE 3 Obesity class

Classification with the plant according to the invention +3 +2 +1 Over- -1 -2 -3 Number of% within

The classmate’s 6 p.m. pm. pm. single vote pm. pm. +/- 1 class rating

Class 1 1 5 128 48 - 182 97

Class 2 4 256 1132 97 1489 100

Class 3 10 129 486 401 9 - 1035 98

Class 4 - - 26 86 80 17 - 209 92

Class 5 21 10 1 1 33 94 Total 1 19 539 1773 588 27 1 2948% 0 1 18 60 20 1 0 100 98.5

-23- · DK 157579B

reproducibility

The ability of the plant and classifier to produce the same results in dual measurements is examined by repeating the measurements on 4l bodies. The accuracy of the plant and classifier in repeating the measurements is shown in Table 4.

TABLE 'i

Yield Fat Muscle- Form- Obesity- of meat percent fill class class

Plant if. invent.

Residue spread 0.56 # 0.67 # 2.00 units. 0.36 units 0.15 o'clock.

Correlation coeff. 0.91 0.71 0.98 0.98 0.91

classifies

Residue spread 0.51 and excl. 0.32 at.

Correlation coefficient of 0.88 0.79

The plant according to the invention exhibits significantly better determinations of shape and obesity classes than those of the classifier.

The plant according to the invention can not only replace a subjective classification, but the results are more reliable. It will also be considerably easier to compare the results of the various slaughterhouses.

Claims (14)

A method for determining the quality characteristics of 5 individual cattle carcasses, characterized in that a carcass (19) or a half-spotted body (19a) is placed in a light-shielding compartment (1) in front of a light-emitting contrast surface (13), an image of a the entire side of the carcass is recorded with an electronic camera (18) facing the contrast surface (13). while the carcass appears as a dark silhouette against the contrast surface, the carcass contour is determined by processing the recorded image using a data processing system, a calculation of parameters is performed in the data processing system on the basis of the particular contour for determining the quality characteristics of the individual body, and the certain quality characteristics of the carcass are announced. Method according to claim 1, characterized in that one or more dimensions of the carcass are determined on the basis of the contour of the registered body and recognizable points therein. Method according to claim 1 or 2, characterized in that meat and fat thicknesses are measured at selected locations on the carcass by the introduction of a probe into the body and the measurements are used as parameters in the data processing system to determine the quality characteristics. Method according to one of Claims 1 to 3, characterized in that a second image of the carcass is recorded by means of the electronic camera (18) illuminating the side of the body facing the camera, the recorded image and the particular contour 30 is processed in the data processing system for masking the background of said second image and separation of meat and fat areas on the body surface and that the areas obtained are used as parameters in the data processing system in determining the quality characteristics of the body. • DK 157579 B -25- Method according to claim 4, characterized in that a sub-image of the recorded image is defined by the particular contour of the individual body and the light reflection values 5 of the image elements in said sub-image are treated statistically for separation of the meat and fat areas. Method according to one of Claims 4-5, characterized in that the side of the carcass facing the camera is illuminated with diffuse light during recording of said second image. Method according to one of Claims 5 to 6, characterized in that a partial image is used, the lower bound of which is a horizontal line which intersects the apex of the curve along the neck and forelegs of the body. Method according to claim 1, characterized in that 15. a carcass (19) or a half, carcass (19a) is placed in a light-shielding space (1) in front of a light-emitting contrast surface (13), a first view of a whole side of the body is recorded with the electronic camera (18) facing the contrast surface, while the body 20 appears as a dark silhouette against the contrast surface, the side of the carcass facing the camera is illuminated, another image of the side of the body is recorded with an electronic camera, while said side is illustrated, the contour of the carcass is determined by processing the recorded first image in a data processing system, a sub-image of the registered second image is defined by the particular contour, and the light reflection values of the elements in said sub-image are processed to separate meat and fat areas, 30. meat and fat thicknesses - if desired - are measured at selected locations of the body by introducing a probe into the body, a calculated The parameters are performed in the data processing system on the basis of the contour, areas and thicknesses found to determine the particular quality properties of the individual body and the specific characteristics of the body are communicated. -26- DK 157579 B 9. A system for determining the quality characteristics of individual bovine carcasses, characterized in that it comprises: 5. a light-shielding compartment (1) with a light-emitting contrast surface (13) and means (7 * 8,9) intended for placing a carcass or a carcass; half, speckled carcass in front of the contrast surface, an electronic camera (18) facing the contrast surface to record an image of an entire side of the body while appearing 10 as a dark silhouette against the contrast surface, a data processing system adapted to determine the contour of the body during processing of the recorded image, which system on the basis of the particular contour is arranged to perform a calculation of parameters for determining the quality characteristics of the individual body, and a data channel for communicating the specific quality characteristics of the body. Installation according to claim 9 *, characterized in that the light emitting contrast surface (13) has substantially uniform light density. Plant according to claims 9 ~ 10, characterized in that it comprises an instrument with a probe for insertion into the carcass for recording meat and fat thicknesses. System according to one of claims 9-H *, characterized in that it comprises an arrangement (15,16,17) for illuminating the side of the carcass facing the camera (18) and a program in the data processing system for separate meat and fat areas in a registered image. System according to claim 12, characterized in that the arrangement 30 is arranged to give diffuse light. Installation according to claim 13, characterized in that the arrangement comprises a frosted translucent plate (15a), behind which are arranged several light sources. 35