HK1097702A - Method for determining the quality of a slaughtered animal carcass and quantities therein - Google Patents

Description

Method for evaluating the quality and quantity of carcasses of slaughtered animals

The invention relates to a method for non-invasively determining the trade classification, trade value, market value and quality of slaughter animal carcasses by means of optical image processing, which method can be used preferably in slaughter houses and meat processing plants. Larger slaughter animals, such as pigs, are usually split along the spine and suspended from hooks for transport between different locations by means of special transport systems. Each split of slaughtered pigs will be enrolled, weighed and evaluated at a determined location.

In terms of evaluation, slaughtered pig carcasses will be graded according to legal trade classification based on lean meat percentage. There are several methods available for determining the lean meat percentage, in which the fat thickness (S) and the meat mass (F) are both related to each other and measured in millimeters, the calculation of the lean meat percentage being by means of an estimation formula established by the authorities.

One method is to measure the values of the terms S and F, i.e., measure the rib rows 7 cm lateral to the cut line for the height of the rib 2/3 at the fourth position. Another conventional measurement method, known as the two-point (ZP) method, determines the amount of fat (S) at the thinnest point of the gluteus medius (MGM) fat and the amount of meat (F) as the thickness of the lumbar muscles, measured as the minimum distance from the anterior (head) end of the gluteus medius to the superior (back) edge of the spinal canal, on the split pig halves of slaughtered carcasses cut along the spinal column.

Substitution of (S) and (F) into the official formula specific for Germany for the calculation of lean meat percentage (MF%)

The value is ranked according to the associated rule.

The measured values can be obtained by both manual and automatic calculations according to a two-point method. A series of documents known from the prior art describe automated solutions for this purpose with optical image processing.

Documents DD 298310 a5/DE 4131556C 2 and DE 4109345C 2 describe methods for determining and analyzing the split of slaughtered animal carcasses by image processing, in which the ratio of the outer contour, fat layer, lean meat and backfat is determined, and the split of the animal carcass, including the spine and all intermediate vertebral layers, are recorded in the image. As a fixed point for determining parameters for splitting and classification of the carcass, it is determined by means of object analysis, as with the other vertebrae, starting from the sacrum of the spinal column. The disadvantages of this method are, on the one hand, the object analysis with predefined contours and object parameters, which leads to high computer-technical costs, and, on the other hand, the selection of the sacrum as a fixation point is not always sufficiently reliable in the case of splitting errors in the actual operation.

Document DE 19733216C 1 describes a method for evaluating the split of slaughtered animals using optical image processing, which can be classified according to a standard two-point method using optical images to evaluate the extended psoas muscle area while excluding the source of subjective errors. The accuracy of the estimation in the evaluation and thus in the classification of the carcass is not improved with this evaluation method compared to the evaluation methods known hitherto.

A method for evaluating the split halves of slaughtered animals by means of optical image processing is known from the document DE 19847232C 2, in which a photogrammetry is used to simulate a conventional two-point evaluation method. Two defined points were marked in the back waist, gluteal thigh area (Schinkenregion), where the first point was the terminal end of the side clavicle (Schlo β knottochen) and the second point was the terminal end of the side MGM (gluteus medius), and a straight line was photographically drawn towards the midline of the backfat. In order to achieve the result of the actual evaluation, the length of the segment of each block is used, which is given by the backfat thickness at a second defined point height on a perpendicular line to the line running parallel to the clavicle. Although the subjective measurement errors of the manually operated ZP method are eliminated in this method, the accuracy of the evaluation is not substantially improved.

Another method is known from DE 19936032C 1, which automatically evaluates the quality of the splits of slaughtered animals, in particular slaughtered pigs, by means of optical image processing, which, compared to known methods, achieves a higher repeatable evaluation accuracy, which is only insignificantly affected during splitting of the slaughtered animal and is not affected by optical imaging in which the imaging plane and the splitting plane are not exactly perpendicular, wherein the optical images of the splitting planes of the carcass of the slaughtered animal are evaluated in the area of the hip thigh and the back waist and are photographically plotted on the basis of defined reference points.

The spine, clavicle, gluteus medius thinnest fat layer, and backfat contour in the acquired region are all used as determined reference points.

The lean meat percentage, which plays a decisive role in the quality assessment, is calculated by adding the partial line segments, which are dependent on one another and perpendicular to the straight course of the spinal cord central canal in the region of the meat and fat layers, using the constants obtained by the regression calculation for each term and a basic constant.

Although in the context of this method the measured value of the total amount of fat (S) is determined at the correct point in a manner complying with the legal rules, the meat quantity (F) is not determined and therefore the lean meat percentage (MF%) cannot be calculated using the official formula and thus the work of classification into trade classification cannot be carried out.

From document DE 19952628 a1 a method is known for determining the trade value of individual parts of live pig carcasses, wherein the body weight, the weight ratio and the meat ratio of the individual parts, such as hip thigh, rib row, cut rib row, fillet, shoulder, abdomen and/or other parts which can be traded or further processed separately, are evaluated by means of an online evaluation of the splitting of live pigs. In order to carry out this method, an estimated value describing the carcass structure is determined from the outer contour lines from the carcass splits and the areas derived therefrom, the position and course of the spinal column and the lengths and areas of the partial regions of the carcass of the slaughtered animal derived therefrom and the fat information obtained for the splitting of live pigs represented by the relative thickness and course of nearly the entire subcutaneous fat layer of the back region. The estimates are correlated with each other by taking into account the statistical correlation that exists between the estimates, whereby the weight of the target part, its weight ratio based on the carcass after slaughter, and the meat ratio can be determined on-line in the slaughter line. In the practice of this method, video recordings must be made of whole live pig splits and the pictorial images are subjected to expensive processing and evaluation, just to assess trade value. Since the picture area over the entire splitting plane is rather large, the evaluation speed is negatively affected and the weight of each block cannot be determined very accurately.

The object of the present invention is to develop a multi-step method for non-invasively determining the trade classification, trade value, market value and quality of slaughtered animal carcasses by means of optical image processing, which complies with the conditions of relevant official standards and regulations and which operates accurately, quickly and inexpensively.

The features disclosed in claims 1 and 2 enable the invention to achieve the above object. Preferred refinements are provided in the dependent claims.

The principle of the multistep method for non-invasively assessing trade classification, trade value, market value and quality of slaughter animal carcasses is that first of all basic data of the slaughter animal carcasses are acquired as data volumes in the actual operation of the slaughterhouse, and then the ratio data (Verhllsniden) is used to perform a simulation calculation on the output assessment of each block. The ratio data are obtained by means of the correlations of the weight ratios produced by the individual blocks in the results of the dissection experiments and the calculated characteristic measurement values and parameters processed in the hip thigh-waist region in parallel with an automated grading method using optical images for evaluating the dissected slaughter animal carcass imaging records.

The complete method for determining the quality and quantity of slaughtered animal carcasses essentially consists of three illustrative incremental steps, but the results of each step can be calculated and indicated independently of one another from the optical image evaluation data of the image recording area.

The dissection experiments of a certain number of slaughter carcasses (here pig carcasses) or carcass splits are carried out in the regulations of european and national legislation for the approval of methods for classification of trade. Within the scope of the dissection experiments, the lean meat percentage was calculated from the inner back weight, the weight of the muscles of the shoulders (including connective tissue), the rib cage, the hip thigh and the abdomen, the weight of the dissected part and the weight of the remaining part according to standard methods. These dissection experiments, including all details, were recorded.

All relevant data are included with the accurate record and used as a data volume, wherein the weight ratio of the fluctuating individual block outputs of non-syngeneic slaughter animal carcasses is contained with a statistically high degree of accuracy.

The accuracy of a licensed automated grading method for estimating lean meat percentage of a slaughter animal carcass must here at least comply with the accuracy achieved by means of simple regression calculations in the case of a 120-rack slaughter carcass dissection experiment, for example.

A known method can be used as an automated grading method which, by means of optical image evaluation of the images recorded with optical sensors from the cut side of the split of the slaughtered carcass, calculates characteristic measurement values and parameters only in the waist and hip thigh region by selection of the determined points. The characteristic measurement values and parameters, such as length, angle and area, and also brightness information or color information presented with the image, are corrected using the resulting data from the weight ratio produced by the blocks of the dissection experiment and ratio data is obtained therefrom, which are stored together with the starting data for later recursive calculations.

In particular, exact measured values of the fat mass (S) and the meat mass (F) are calculated here, wherein the lean meat percentage (MF%) of slaughtered pigs is directly calculated in germany according to the two-point method using the official formula, whereby the classification of trade classes can be carried out immediately as a first method step.

The grading of slaughtered pig carcasses is carried out in a similar manner using national specific formulas.

As important basic data for slaughtering carcasses in slaughterhouses and processing plants, the determination of their weight is made from the total weight of the split halves, obtained by cutting along the spinal column, suspended from hooks and by selecting a determination point characteristic measurement values and parameters in the waist and hip thigh area are calculated by means of optical image evaluation of the images recorded with optical sensors from the cut side of the split halves of the slaughtering carcasses. The determination of characteristic values, lengths, angles and areas in the imaging region is carried out in dependence on the determined points.

In a second step of the method, the rib rows are evaluated on the basis of the calculated lengths of the vertical segments in the straight section of the spine in the imaging region, which relates to the outer contour line, the fat line course and the ratio between the two.

The yield of each block is estimated by a simulation calculation of recursive calculations with ratio data from the splitting experiments using other existing characteristic values. The sum of the individual block evaluations obtained here generally gives the trade value as a third method step.

The weights of the individual parts were estimated from the weight of the carcass, and the market value was given again from the sum.

It is also conceivable in this connection that the weight of the individual blocks is estimated solely using the characteristic values, lengths, angles and areas calculated in the imaging area, instead of first determining the total weight of the carcass of the slaughtered animal and calculating it.

With the aid of the brightness and color information, the slaughtered carcasses are subjected to quality classification.

The advantage of the invention lies in particular in the use of known, non-invasive, automated methods for calculating the measurement values for determining the lean meat percentage (MF%) of slaughtered pigs in accordance with official regulations. It is possible to use both imaging methods for evaluating the dissected plane imaging and methods for slaughtering carcasses by means of nuclear spin tomography or computer tomography or ultrasound measurements along the dorsal spine.

The allowable tolerance, the permissible assessment tolerance according to the official regulations for measuring lean meat percentage, is complied with and even lower than.

The limitation is imposed on the imaging area of the hip thigh and waist area used for the evaluation, so that exact measurement values can be determined, thereby obtaining more accurate evaluation at a higher rate.

The trade value of a slaughtered carcass can be calculated from the data of the valuable parts. Taking the entire quality into account, the market value can be calculated.

The method can be used to replace various methods known hitherto for determining lean meat percentage and optionally trade value, so that all parameters for processing, finishing and pricing can be determined accurately, quickly and inexpensively.

The invention is further explained by way of example in accordance with fig. 1 as an imaging area for determining characteristic measured values and parameters for the splitting of a slaughtered carcass.

In the case of a dissection experiment on a sufficient number of live pig carcasses, in order to obtain the basic data, the weight of the slaughtered carcasses is first determined after slaughter and cooling, where the carcasses may have been split along the spine, and then digital images of the waist and hip thigh are made using imaging methods, and the images are analysed to obtain a profile course of meat tissue, adipose tissue and bone. The average length and area of the individual lengths passing through the contour region is measured and brightness and/or color values are obtained, depending on the contour course. However, actual dissection experiments were performed and accurately recorded, where the weight ratio of all blocks was measured separately and stored.

The measured values and parameters obtained with the automatic image analysis are correlated with the weight of the slaughtered carcass and the weight of the output of the individual blocks in each case, wherein specific ratio data are calculated. These ratio data are statistically reliable due to the large range of data volumes derived from a large number of dissection experiments.

In the case of the dissection experiments and in the case of slaughter line operations, it is preferred to take characteristic measurements and parameters at the waist and hip thigh in each case in the same way under the guidance of the method described in the document DE 19936032C 1.

All details of the image region 1 in the hip thigh and waist region of the carcass splits are recorded and evaluated in an image recording manner in accordance with fig. 1.

The image area 1 is recorded against a dark background over the entire hip thigh and waist area with its outer contours 2.1 and 2.2.

According to the histogram analysis, the critical parameter (schwellian) of the luminance of the pig carcass in each case was first renormalized (renormalizing) and then different tissue groups were selected in computer technology on the basis of the color and/or luminance differences of the image region 1. Impurities such as blood are filtered out of the image by self-detection in a conventional manner.

In the next step, the light-colored fat is separated from the dark-colored muscle and the determination of the fat area 3 and the meat area 4 is carried out in this way. In the meat area 4, a contour tracking algorithm is applied to determine the contour of the gluteus medius muscle (MGM)5, which is then geometrically positioned. In addition, the vertebrae 6 at the end of the spinal column and the clavicle 7 are visible in the image region 1. In this case, the vertebrae 6 with the spinal canal 8 in the straight part of the spinal column can be calculated here according to a periodicity criterion.

A straight line 9 of the spinal column in the direction of the straight section is taken as the starting line of measurement by the upper edge (dorsal) of the spinal canal 8. On this line 9, a vertical line 10 is provided at the level of the front (head) end 11 of the gluteus medius muscle (MGM)5, whose line length corresponds to the amount of muscle (F) as the thickness of the psoas muscle, as the shortest distance between the front end 11 of the gluteus medius muscle (MGM)5 and the upper (back) edge of the spinal canal 8. The extension of the vertical line 10 up to the outer contour line 2.2 defines the course of the fat through which the head of the gluteus medius MGM 5 passes.

A line 12 from the contour of the gluteus medius MGM 5 to the outer contour line 2.2 is determined at the height of the thinnest fat layer on the gluteus medius MGM 5 side, and its line length represents the fat amount (S).

The calculation of the lean meat percentage (MF%) is carried out on-line by the specific official formula according to the two-point method from the numerical terms (F) and (S) in millimeters as the unit of detection, and then the classification of trade is carried out on the basis of the calculated lean meat percentage.

Juxtaposition to the vertical line 10 makes it possible to calculate the distance of the straight line of a further vertical line 13 on the line 9 from the outer contour line 2.2, the starting points of which on the straight line 9 fall in each case in an imaginary vertical extension of the layer between the vertebrae 6.

The vertical line 13 is interrupted by the inner contour 14 of the fat region 3, so that part of the line segments lie in the muscle and part of the line segments lie in the fat, their length being used as the length of the fat segment and the length of the segment muscle and their length ratio to one another for the evaluation of the rib rows.

The average fat thickness through the gluteus medius MGM 5 side in the region between the extension of the vertical line 10 to the outer contour line 2.2 and another vertical line 15 on the straight line 9 at the height of the posterior terminal (tail) end 16 of the gluteus medius MGM 5 was used for evaluation of the gluteal thigh, which was also used for determination of trade value.

Proceeding from the described exemplary embodiment, a large number of other lengths, angles and areas in the image region 1 are determined, which are used to differentiate the ratio data more finely.

Therefore, the abdomen is represented in terms of the average subcutaneous fat layer thickness 17 in the rib cage region, and is represented by other measurement values in terms of the gluteal thigh, the rib cage, and the abdomen from above the head end 11 of the gluteus medius (MGM)5 and the shoulder in the image region 1.

The use of data previously obtained from image analysis, including the calculation of the total weight of the slaughtered carcass obtained from two originally integral splits hung on hooks, the calculation of the yield of each block based on the ratio data present in the data volume, the calculation of the trade value from the sum of the evaluations of each block and the calculation of the market value from the sum of the weights of each block.

It is conceivable to determine the weight of the individual parts directly from the image-analyzed measured values, as is the hip thigh or rib row.

Furthermore, the slaughtered carcasses and/or parts are quality-classified according to the brightness and/or colour values present.

A further development of the method makes it possible, in particular for use in a decommissioning enterprise, to have a self-learning effect using a self-consistency check of the data volume, wherein the variability of the evaluation results produced by the individual blocks is limited, in particular, further, by correcting the weighing results of the individual blocks during the processing with the values present in the data volume, optionally supplemented with further data.

The expanded data volume obtained in this way will be used for upgrading small-scale slaughter operations, so that smaller slaughterhouses also achieve more accurate estimation results.

All the steps of the method are carried out with electronic data processing equipment, which has suitable interfaces for imaging equipment and weighing equipment, among others.

Reference numerals used

1 image area

2 outer contour line (2.1; 2.2)

3 fat noodles

4 muscle noodles

5 gluteus medius (MGM)

6 vertebrae

7 clavicle

8 spinal canal

9 straight line

10 vertical line

11 front (head) end

12 connecting wire

13 remaining vertical lines

14 inner contour line

15 additional vertical lines

16 rear (tail) end

17 middle subcutaneous fat layer

Claims (9)

1. Method for non-invasively determining trade classifications, trade values, market values and quality of slaughter animal carcasses based on optical image processing, which determines all details, lengths, angles, areas, luminance information and/or color information in the hip thigh and waist region in an image region (1), records and uses the total weight of the slaughter carcasses and the result data of the dissection experiments produced for individual blocks of the fluctuations of non-syngeneic slaughter animal carcasses, characterized in that the result data of the weight ratio of the individual block outputs obtained in the dissection experiments of a sufficient number of slaughter carcasses are corrected to one another with characteristic measured values and parameters calculated from the two halves of the slaughter carcasses in the hip thigh and waist region with the total weights calculated in and ratio data are obtained therefrom, and the obtained ratio data are used in a slaughter line operation taking into account the total weight of the two proto carcass splits and the total weight of the two proto carcasses and the pair of carcasses The carcass-specific characteristic measurements and parameters calculated at the hip thigh and waist regions were simulated to estimate the block yield.

2. Method for non-invasively determining trade classifications, trade values, market values and quality of slaughter animal carcasses based on optical image processing, which determines all details, lengths, angles, areas, luminance information and/or color information in the hip thigh and waist region in an image region (1), records and uses the total weight of the slaughter carcasses and the result data of a dissection experiment for each block of fluctuations of non-syngeneic slaughter animal carcasses, characterized in that the result data of the weight ratio of the block outputs obtained in the dissection experiment of a sufficient number of slaughter carcasses are corrected to one another and ratio data are obtained therefrom, and the obtained ratio data of the slaughter carcasses and the characteristic measurement data and parameters calculated for the carcass-specific in the hip thigh and waist region are used in a slaughter line operation to model the obtained ratio data of the slaughter carcasses and the characteristic measurement data and parameters calculated for the carcass-specific in the hip thigh and waist region The yield of each block is estimated by calculation.

3. The method as claimed in claim 1 or 2, characterized in that, in part of the steps of the image evaluation for the online calculation of the lean muscle mass (MF%), a straight line (9) in the direction of the straight section of the spinal column is provided by the upper edge (dorsal) of the spinal canal (8), and a vertical line (10) is provided thereon at the height of the front (head) end (11) of the gluteus medius-MGM (5), which line has a length corresponding to the shortest distance between the front end (11) of the gluteus medius (5) and the upper (dorsal) edge of the spinal canal (8), corresponding to the mass of meat (F) as the thickness of the psoas muscle, and a connecting line (12) from the contour of the gluteus medius MGM (5) to the outer contour (2.2) is determined at the height of the thinness layer on the side of the gluteus medius MGM (5), which line length represents the mass of fat (S), wherein the lean muscle Mass (MF) is calculated by two numerical terms (F) and (S) using a specific lean official formula according to ZP-method, a ranking of the trade classification is then performed.

4. Method according to claim 1 or 2 and 3, characterized in that the distances of the other vertical lines (13) on the line (9) to the outer contour line (2.2) are calculated in parallel to the vertical line (10), their starting points on the line (9) in each case falling in an imaginary vertical extension of the layer between the vertebrae (6), wherein the vertical line (13) is interrupted by the inner contour line (14) of the fat region (3) such that part of the line segments lies in the muscle and part of the line segments lies in the fat, their lengths being used as the length of the fat segments and the length of the muscle segments and their length ratio to one another for the evaluation of the rib row.

5. A method as claimed in claims 2, 3 and 4, characterized in that the weight of the individual parts is determined directly as in hip thigh or rib steak by means of image-analyzed measurements.

6. The method as claimed in claim 1 or 2 and 3 to 5, characterized in that the average fat thickness through the gluteus medius (5) side in the region between the extension of the vertical line (10) to the outer contour line (2.2) and another vertical line (15) on the straight line (9) at the height of the posterior (caudal) end (16) of the gluteus medius (5) is used for evaluation of the gluteal thigh and for determination of the trade value.

7. Method according to claim 1 or 2 and from 3 to 6, characterized in that the abdomen is represented in terms of the average subcutaneous fat layer thickness (17) in the rib row area, and in that the shoulder area is represented in terms of the hip thighs, the rib row and the abdomen in the image area (1) from above the head end (11) of the gluteus medius muscle (5).

8. Method as in claims 1 or 2 and 3 to 7, characterized in that it has a self-learning effect, implemented using a self-consistency check of the data volume, when applied in a splitting operation, by correcting the weighing results of the blocks during the machining process with the values present in the data volume and, if necessary, supplemented with other data.

9. Method as claimed in claim 1 or 2 and 3 to 8, characterized in that the data volume expanded by self-learning is used as an upgrade in small-scale slaughtering operations.