DK 163.382 B (Slagteriernes Forskningsinstitut) discloses a method of determining the quality of individual meat pieces, in which method a reflection measuring probe is introduced into the meat piece and a number of measurements of the light reflectivity along the scanning line of the probe is recorded. The data set of reflection values obtained is subjected to a statistical analysis computing how many times each reflection value appears, and the degree of frequency is inserted in a multivariable algorithm expressing a quality property. The reflection is measured in a wavelength band in the border range between the visible and the near-infrared range, for instance at 950 nm.

One of the quality properties, which can be deter- mined, is the juice holding capacity, which is the capability of the meat to retain water, for instance during storage. This capability may be determined with high accuracy in chilled meat (post rigor) by means of the method described. When measuring slaughter-warm meat (pre rigor), no correlation as to juice holding capacity has been found.

It is, however, desirable to be able to determine the juice holding capacity early in the slaughter process, as the part carcass may then be subjected to treatments which are more optimally adapted to the actual quality properties of the individual carcass or meat.

DK 172.774 B1 (Slagteriernes Forskningsinstitut)

discloses a method of predicting the drip loss of a meat piece or determining its juice holding capacity.

In the method two sets of light reflectivity measure- ments with timing difference are recorded on a slaugh- ter-warm muscle in the carcass, for instance with a timing difference of 1-20 minutes. The change of the reflection values are inserted in an algorithm express- ing the juice holding capacity of the meat. Measure- ments are made in the near-infrared area from 900 to 1800 nm. A drawback of the method is that it takes a long time, seen in relation to the advancing tempo of carcasses on a modern slaughter line. Moreover, a higher accuracy than the one found (1.8%) is desirable.

It is the object of the invention to provide a method and an apparatus for prediction of the drip loss of a part of a carcass by measuring a muscle in the part carcass after slaughtering, said method and apparatus obviating the above drawbacks.

The method according to the invention is charac- terized in that measurements are carried out while the part carcass is still warm from slaughter, that the light reflectivity of the muscle is measured in at least one wave range with a wave number below 1500 cm 1, that the resulting, possibly processed measurement data are inserted as a variable in an algorithm expressing a prediction of the drip loss as a function of the light reflectivity in one or more wave ranges with a wave number below 1500 cm 1, and that the prediction of the drip loss is automatically calculated in a calculation unit by means of the algorithm.

The method according to the invention is based on the surprising observation that by carrying out a reflection measurement on a muscle in a carcass, while the carcass is still warm after the slaughtering, in at

least one wave range with a wave number below 1500 cm 1, a quick measurement can be made predicting the drip loss, which makes it possible to measure the carcasses in the tempo, in which they are advanced on a common slaughter line, and it has moreover been found that the measurement is able to predict the drip loss with an accuracy which is substantially higher than the one obtained by the method using measurements with timing difference.

The wave range in question with a wave number below 1500 cm 1 (corresponding to a wavelength band with a wavelength of more than 6,667 nm) lies preferab- ly in which corresponds to the middle-infrared range, i. e. very far from the above wavelengths of 950 and 900 -1800 nm used up till now.

A preferred embodiment of the method according to the invention is characterized in that measurements are carried out in one or more wave ranges with a wave number in the interval 900-1500 cm 1. In this connec- tion a very good correlation between the reflection ability and the prediction of the drip loss has been found.

Preferably, measurements are carried out in one or more wave ranges with a wave number in the interval 900-1200 cm.

Preferably, measurements are carried out in more than one wave range, and the algorithm is preferably multivariable. Preferably, measurements are made in less than 100 wave ranges and/or an algorithm having less than 100 variables is used. In particular, meas- urements are carried out in less than 10 wave ranges and/or an algorithm having less than ten variables is used.

It is preferred to use only one piece of measuring equipment per slaughter line. The piece of equipment

may be a fully automatic device or a semi-automatic, operator controlled instrument. The piece of equipment may be adapted to measure each carcass passing the equipment, but may possibly only carry out random measurements if conditions so require. The equipment may moreover be designed for measuring chilled meat.

The time required for recording measurement data may vary in accordance with the specifications of the equipment and will thus typically lie from below one second and up to 10 seconds, thus making it possible for one single piece of equipment to measure in the tempo of the slaughter line. Preferably, measurement data are used which have been recorded within less than 1 minute, for instance less than 10 seconds.

In the method according to the invention, measure- ments are carried out while the part carcass is still in slaughter-warm condition, i. e. while the muscles undergo the transformation from living tissue in a recently slaughtered animal to chilled, edible meat.

The measuring is therefore preferably carried out while the muscle is in pre rigor condition, in particular before the chilling of the part carcass.

The measuring is preferably carried out within two hours from the slaughtering (drainage of blood), for instance 0. 5 to 2 hours after slaughtering.

The apparatus according to the invention is characterized in comprising -a light reflection meter adapted to measure the light reflectivity of a muscle in at least one wave range with a wave number below 1500 cm 1, -a calculation unit with a program and/or memory part adapted to automatically record measured, possibly processed light reflectivity data in an algorithm and to calculate the algorithm value, -an algorithm contained in the program and/or

memory part expressing a prediction of the drip loss as a function of the light reflectivity in one or more wave ranges with a wave number below 1500 cm, measured on a muscle in a part carcass after slaughter- ing, while the part carcass is still warm (pre rigor), and -a signal unit which after the insertion of light reflectivity data in the algorithm and calculation of the algorithm value emits a signal depending on the calculated prediction.

The apparatus may be stationary or portable and may be adapted to measure directly on a visible meat surface or on a cut surface in the meat provided by insertion of a measurement probe with a pointed end.

The measurement head of the apparatus is preferably in contact with the muscle during measurement. If desired, measurements may be performed in various different muscles.

By a slaugter part carcass is in the present case preferably to be understood a whole carcass, in par- ticular a carcass from which the bowels have been removed, a half carcass produced by backsplitting of a whole carcass, or a part carcass or a cut-out piece, for instance a fore-end, a middle piece or a ham of a pig carcass. Measurements are preferably carried out on a muscle in a whole or backsplit carcass.

A carcass is in particular a pig carcass.

By drip loss is in the present connection also to be understood the juice holding capacity.

In a wave range with a wave number below 1500 cm 1 is in the present connection to be understood what corresponds to a wavelength band with a wavelength of more than 6667 nm.

Example This example illustrates that the drip loss of slaughter carcasses can be predicted with high accuracy already on the slaughter line by measurement of the reflection of the meat at several wave ranges with a wave number below 1500 cm. The measuring is carried out within a so short period of time that no transform- ation of the meat takes place during the period of measurement.

In a test slaughterhouse 1/3 of a group of 46 pigs are exercised to provide a higher share than normal of carcasses having a high drip loss after chilling. Then all pigs are given an anaesthetic and slaughtered. 35 minutes after sticking, a hot meat sample (cutlet) is taken from the back muscle (longissimus dorsi). The temperature is 37 to 41°C. An ATR-IR spectrum is obtained with 64 scans in an interval from 900 to 1500 cm 1 (ATR = Attenuated Total Reflectance). The defini- tion is 4 cm 1. A spectrometer of the type Arid-Zone MB155S from Bomem, Canada, with an InAs detector, single-plate beamsplitter and SiC radiation source.

The drip loss of the meat from the same 46 pigs is recorded in a laboratory by measuring of the weight of a meat sample by cutting out after 24 hours and again after 72 hours (K. O. Honikel 1987. How to measure the water-holding capacity of meat? Recommendation of standardized methods. Evaluation and Control of Meat Quality in Pigs. Edited by P. V. Tarrangt, G. Eikelen- boom & G. Monin. Martinus Nijhoof Publishers, Holland: p. 129-142). The drip losses measured lie between 0.7 and 8.0%.

PLS regressions (PLS = Partial Least Squares) between the IR measurement values and the drip loss measured in the laboratory are determined in respect of the 46 pigs, partly directly on the IR measurement

values of the spectrum and partly on the first and second derivative of the spectrum. On basis of this a multivariable algorithm is developed of the type Wdryp = k0 + k1 . a1 + k2 . a2 + kn . an in which Wdryp is the predicted drip loss (in %) k0'kl, k2.... kn are constants, al, a2 an are reflection measurement values or the first or second derivative in a range with a given wave number, and n is an integer, for instance 100.

Data from the four outliers do not form part of the development.

On basis of the algorithm and the two data sets the correlation R and the accuracy of the prediction RMSEP (% drip loss) can be calculated, which will appear from the table below.

Spectral range (cm 1) Data pre-R RMSEP processing 900-1100 0.90 0.82 1000-1200 0.91 0.78 1000-1200 1st derived 0.94 0.64 1000-1200 2nd derived 0.92 0.73 It will be seen that the drip loss can be deter- mined with an inaccuracy of less than 1% by measuring of the reflection capability of the meat in several wave ranges in the interval from 900 to 1200 cm 1 and by insertion of the measurement values in a multivari- able algorithm developed with a standard measuring

method as reference. This inaccuracy is sufficiently small for making it realistic to sort the carcasses in quality classes already before the carcasses are chilled in the slaughterhouse, which makes it possible to use carcasses with a very low drip loss or a high drip loss for the manufacture of types of products suitable therefor, whereas carcasses with a normal drip loss for instance may be used for the production of fresh cut-outs.

By the development of production equipment the reflection measurement and the calculation can be carried out in the same tempo as the advancing of the carcasses on a slaughter line, i. e. without any need to wait for transformation processes of the meat like in the method according to Danish Patent No. 172.774 B1.

The correlation between the spectroscopically measured drip loss and the drip loss measured in the laboratory of the 42 samples is shown in Fig. 1. In the calculation of the drip loss by spectroscopy the 1st derived of the IR spectra is used.

The IR spectra in the interval 900-1500 cm 1 for meat from all 46 pigs are shown in Fig. 2. In particu- lar about 1000 cm 1, the spectra contain much informa- tion which is correlated to the drip loss.

Above 1500 cm 1 no correlation between drip loss and reflection in slaughter-warm meat has been found.