DISPOSITIF POUR EVALUER LA QUALITE DE MOITIES DE CARCASSES ANIMALES

METHOD FOR ASSESSING THE QUALITY OF ANIMAL HALF-CARCASSES

Abrégé français

L'invention concerne un procédé permettant de déterminer de manière automatique la qualité de carcasses animales, par traitement optique d'images, notamment de porcs venant d'être abattus. Comparativement aux procédés connus, ce procédé permet de parvenir à une meilleure précision d'estimation reproductible, sur laquelle des erreurs intervenant dans le processus de séparation de l'animal abattu et qui sont dues au fait que la détection d'images n'intervient pas absolument perpendiculairement au plan de séparation, ne peuvent influer que de manière peu significative. A cet effet, il est prévu d'évaluer de manière photogrammétrique une prise optique de la moitié de carcasse animale dans le plan de séparation, dans la zone du jambon et de la longe. Les points de référence marquants retenus dans la zone sélectionnée sont : la colonne vertébrale (3), l'os de jointoiement du bassin (2), l'épaisseur de couenne la plus mince sur le muscle grand fessier (16) et les contours de la couenne arrière (6; 12). La proportion de viande maigre prépondérante pour estimer la qualité est calculée par addition de sections partielles, perpendiculaires au cours linéaire du canal médullaire arrière, et dont le rapport mutuel est établi, tout en tenant compte de constantes résultant de calculs de régression pour chaque terme, ainsi que d'une constante de base.

Abrégé anglais

The invention relates to a method which, using optical image processing,
ensures an automatic quality assessment of animal half-carcasses, in
particular, of slaughtered pigs, whereby, compared to prior art methods, a
higher reproducible estimation accuracy, which can not be significantly
influenced by errors made during the cleaving process of the slaughtered
animal, is attained by carrying out image acquisition in a manner that is not
absolutely perpendicular in relation to the cleaving plane. To this end, the
invention provides that an optical recorded image of the animal half-carcass
is photogrammetrically evaluated in the cleaving plane, in the area of the ham-
loin region. The vertebral column (3), the hinge bone (2), the thinnest fat
thickness on the musculus gluteus medium (MGM) (16), and the contours of the
fat back (6; 12) in the selected region are used as distinctive reference
points. The portion of lean meat which leads to the assessment of quality is
calculated by adding the partial lengths in the region of the meat and of the
fat layer which are set in proportion to one another and which are
perpendicular to the straight progression of the back marrow channel while
taking into account a base constant as well as constants for each term, which
are established from regression calculations.

Revendications

7
Claims
1. Method of evaluating the quality of slaughtered animal halves by means of
optical image
processing, wherein by means of photogrammetric evaluation of a recorded
optical image in
the splitting plane of said slaughtered animal halves in the region of the ham-
loin area,
different tissue types are selected using computer technology and specific
parameters for
evaluation purposes are ascertained,
characterised in that
the significant reference points used are the spinal column (3) comprising the
spinal cord
channel, the carcass-side lower point of the pin bone (2), the thinnest layer
of fat on the
MGM (16) and the contours of the back fat (6; 12) in the selected region,
a straight line (4) in the direction of the last, straight portion of the
spinal column (3) is placed
in the middle of the spinal cord channel, as a first starting line for the
measurements, and
placed on this straight line (4) as a second starting line (5) is a
perpendicular section
extending as far as the outer fat contour (6), wherein an inner extension (7)
of the second
starting line (5) tangentially contacts the lower point of the pin bone (2),
in parallel with the
second starting line (5) further sections, whose lengths are defined by the
first starting line
and the outer fat contour (6), are calculated at the level of the tail-side
end of the last vertebra
of the spinal column (3), the thinnest layer of fat on the MGM (16), between
the ultimate and
penultimate vertebra of the spinal column (3) as well as between the
penultimate and third
from last vertebra of the spinal column (3), wherein the length of partial
sections of the
second starting line (5) and the further sections in parallel therewith is
determined by the
point of intersection thereof with the fat inner contour line (12), and
the lean flesh proportion, which is crucial for the purpose of the evaluation,
is calculated by
the addition of summands of partial sections, which are set in ratio with
respect to each other
and are perpendicular to the straight progression of the spinal cord channel,
in the region of
the flesh and the layer of fat, thus incorporating constants for each term,
which are
ascertained by means of regression calculations, and a basic constant.
2. Method according to claim 1, characterised in that the market value is
determined by
calculating the area from partial sections and associated sections on the
straight line (4) as a

8
first starting line by reference to characteristic ratio values of the area of
fat layer parts and of
partial areas which are defined by the fat inner contour line (12), the pin
bone (2) and the
penultimate vertebra of the spinal column (3).
3. Method according to claim 2, characterised in that ratio values of the
lengths of the
diagonals of partial areas in the flesh and fat region with respect to each
other are used for the
purpose of determining market value.
4. Method according to claim 2, characterised in that ratio values of the
lengths of sections,
which define the partial areas, in the flesh and fat region with respect to
each other are used
for the purpose of determining market value.

Description

CA 02371911 2001-10-22
Method for evaluating the quality of slaughtered animal halves
The invention relates to a method for evaluating the quality of slaughtered
animal halves by
optical image processing which can be used in particular for the purpose of
classifying and
determining the market value of halves of slaughtered pigs and also in
principle of other large
and small slaughtered animals.
Generally in slaughter houses and meat processing works the halves of
slaughtered pigs are
suspended on hooks, the halves are registered, weighed and evaluated according
to an official
trade classification by measuring the thickness of fat and flesh at locations
which are
specifically prescribed by law in each State. Furthermore, the market value is
determined
using a plurality of other specific parameters of the slaughtered animal
carcass and these
parameters are mainly not standardized.
In addition to manual methods for determining the thickness of fat and flesh
on halves of
slaughtered animals, different prior art classification methods are known
which follow the
principle of optical image processing.
The documents DD 298 310 AS / DE 41 31 556 C2 and DE 41 09 345 C2 describe a
method
for determining or analysing halves of slaughtered animal carcasses by means
of image
processing, wherein the outer contour, layer of fat, flesh and back fat ratio
are determined,
whereby images are recorded of the halves of slaughtered animals including the
backbone and
all the intermediate vertebral layers. The sacrum of the spinal column is used
as the fixing
point for determining the parameters for splitting and classifying the carcass
and the sacrum
is also determined in the same way as the other vertebrae by means of object
analysis. On the
one hand, a disadvantage of this method is the high cost of computer
technology required to
analyse the object using pre-defined contour and object parameters, on the
other hand owing
to possible splitting errors during the actual processing, it is not always
possible to select the
sacrum in a sufficiently reliable manner as a fixing point.

CA 02371911 2001-10-22
2
The document DE 197 33 216 C 1 describes a method for evaluating halves of
slaughtered
animals using optical image processing, which renders it possible based on the
standard two-
point method to classify a carcass using an optical image evaluation of the
extended loin
region whilst excluding subjective error sources. The accuracy of the
estimation for
evaluation purposes and thus for classification of the carcass has not thereby
been improved
in comparison to hitherto known evaluation methods.
A method of evaluating slaughtered animal halves by optical image processing
is also known,
wherein a photogrammetric method is used as a simulation of the conventional
two-point
evaluation method. In the loin and ham region, two significant points, of
which the first point
is the carcass-side end of the pin bone, the second point is the carcass-side
end of the MGM
(Musculus Gluteus Medium), and a straight line in the direction of the average
progression of
the back fat are recorded photogrammetrically. For actual evaluation purposes,
the lengths of
partial sections are used which are provided on a perpendicular on the
straight line, which is
displaced in parallel with the pin bone, at the level of the second
significant point by means of
the thickness of the back fat. Although in the case of this method the
subjective measuring
errors of the manually performed two-point method are eradicated, the accuracy
of the
estimation for evaluation purposes is not, however, increased.
It is the object of the invention to develop a method which by means of
optical image
processing ensures the automatic evaluation of quality of slaughtered animal
halves, in
particular of slaughtered pigs, wherein in contrast to the known methods it is
possible to
achieve a higher reproducible accuracy of estimation which can only be
influenced
insignificantly by errors in the processing of splitting the slaughtered
animal and cannot be
influenced by not absolutely perpendicular image recording with respect to the
splitting plane.
The object is achieved by the features stated in claim 1. Preferred
developments are stated in
the subordinate claims.
The essence of the invention resides in the photogrammetric evaluation of a
recorded optical
image of the slaughtered animal half in the splitting plane, in the region of
the ham-loin area.

CA 02371911 2001-10-22
3
In this region, the pin bone and the vertebrae can be identified optically,
the Musculus
Gluteus Medium and the back fat can always be selected optically by the colour
and
brightness differences with respect to other tissue parts, which means that
they can be
selected in a reliable manner with respect to computer technology. The
significant reference
points used for the purpose of the photogrammetric evaluation are the spinal
column, the pin
bone, the thinnest layer of fat on the MGM and the contours of the back fat in
the selected
region.
The lean flesh proportion which is crucial for evaluating quality is
calculated by the sum of
partial sections, which are set in ratio with respect to each other and are
perpendicular to the
straight progression of the spinal cord channel, in the region of the flesh
and layer of fat, thus
incorporating constants for each term, which are determined from regression
calculations, and
a basic constant.
The partial sections which are required for the calculation are determined in
that a straight
line in the direction of the last, straight portion of the spinal column is
placed in the middle of
the spinal cord channel, as the first starting line for the measurements.
Placed on this straight
line as the second starting line is a perpendicular section extending as far
as the outer contour
of fat, of which the inner extension tangentially contacts the lower point of
the pin bone.
In parallel with the second starting line, further sections, whose lengths are
defined by the
first starting line and the outer contour of fat, calculated at the level of
the tail-side end of the
last vertebra, the thinnest layer of fat on the MGM, between the ultimate and
penultimate
vertebra and also between the penultimate and third from last vertebra. The
length of the
partial sections of the second starting line and the further sections in
parallel therewith are
determined by the point of intersection thereof with the fat inner contour
line between fat and
flesh.
It is also feasible to determine the market class by means of the cross-
sectional area
calculation, which is possible in the following step, from the partial
sections and the
associated sections on the first starting line by reference to characteristic
proportional
variables of fat-layer parts and of partial areas which are defined by means
of the fat inner
contour, the pin bone, the penultimate vertebra.
Alternatively, the ratios of the section lengths of the diagonals of partial
areas with respect to
each other can also be used for determination purposes.

CA 02371911 2001-10-22
4
By determining section ratios for the purpose of evaluating the quality of
slaughtered animal
halves, the advantages of the invention reside particularly in the elimination
of sources of
error by not absolutely perpendicular recorded images in the splitting plane
which otherwise
cause incorrect evaluations and incorrect classifications. Splitting errors in
the slaughtering
process scarcely influence the evaluation. The accuracy of the estimation is
substantially
increased, whereby it is possible to achieve a more precise, higher quality
evaluation.
The invention is explained in detail as an exemplified embodiment with
reference to Figure 1
as a recording region for the evaluation.
As shown in Figure 1, an image region 1 of the ham and loin region of a
slaughtered animal
half is evaluated photogrammetrically and said image region includes its
entire width, the
carcass-side end of the pin bone 2 and the lower end of the spinal column 3
including
vertebrae.
A straight line 4 in the direction of the straight portion of the spinal
column 3 is placed in the
middle of the spinal cord channel as the first starting line for the
measurements. Placed on
this straight line 4 as the second starting line 5 is a perpendicular section
extending as far as
the outer fat contour 6, wherein an inner extension 7 of the second starting
line 5 tangentially
contacts the lower point of the pin bone 2.
In parallel with the second starting line 5, four further sections are
calculated, whose lengths
are likewise defined by the straight line 4 as the first starting line and the
outer fat contour 6.
A first parallel 8 is calculated at the level of the thinnest layer of fat on
the MGM, a second
parallel 9 is calculated on the tail-side end of the last vertebra of the
spinal column 3, a third
parallel 10 is calculated between the last and the penultimate vertebra of the
spinal column 3
and the a fourth parallel 11 is calculated between the penultimate and third
from last vertebra
of the spinal column 3. The second starting line 5 and the parallels 8; 9; 10;
11 are bisected
by the fat inner contour line 12, whereby partial sections are produced. As a
consequence,
partial sections stl 13 and st2 14 are determined on the second starting line
5. Furthermore,
the points of intersection determine the length of the outer partial section
stzp 15 of the first
parallel 8, which corresponds to the thinnest layer of fat on the Musculus
Gluteus Medium
16, the length of an inner partial section stk 17 on the second parallel 9,
the lengths of the

CA 02371911 2001-10-22
partial sections st3 18 and st4 19 on the third parallel 10 and on the fourth
parallel 11 the
lengths of the partial sections sty 19 and st6 20.
The lean flesh proportion (mfa) is calculated e.g. according to the following
formula:
st6 sty stl sty st6 st2
mfa = p0 + pl stl + p2 stl + p3 stk + p5 stzp + pg stzp + p9 stzp
which is used directly for evaluation purposes and serves to classify market
value
classification.
The constants p0, pl, p2, p3, p8, p9 are calculated as reference variables by
regression
calculations by means of a standardised statistics program, such as e.g.
Microsoft Excel, from
the actual lean flesh proportion of slaughtered animal halves.
It is likewise possible to determine the market value by means of the ratio of
areas or section
lengths of the delimitation thereof. For this purpose, the variables of the
area, which is
defined by the straight line 4, the fat inner contour line 12, the partial
sections stl 13 and st3
18, are calculated with respect to the area of the fat between the partial
sections st2 14 and st4
19 and a coefficient is determined which corresponds to the market value.
Alternatively, the
market value can be determined by means of coefficients from the values of the
diagonals of
these areas.

CA 02371911 2001-10-22
6
List of reference numerals used
1 image region
2 pin bone
3 spinal column
4 straight line
second starting line
6 outer fat contour
7 inner extension of the second starting line 5
8 first parallel
9 second parallel
third parallel
11 fourth parallel
12 fat inner contour line
13 stl
14 st2
outer partial section (stzp)
16 Musculus Gluteus Medium - MGM
17 stk
18 st3
19 st4
sty
21 st6

Dessin représentatif