Title: Method and apparatus for preparing cheese

The invention relates to a method for preparing cheese, in particular cheese prepared by means of acid curdling or enzymatic curdling. The invention also relates to an apparatus for preparing cheese. All this is especially directed towards the quality and the yield in industrial cheese preparation processes. In this context, the term "industrial" is understood to mean a cheese preparation process in which at least 1,000 kg/h of milk is processed into cheese. In more detail, the invention relates to such a method in which inline measurements are performed in order to derive, from the results, relevant information which can be utilized to adjust the cheese preparation process. "Inline" is understood to mean that the measurements can be performed directly in the process without valuable material, such as a sample of the cheese or an intermediate product in the processing of milk into cheese, having to be taken from the process and/or the production process having to be interrupted. Should such a measure be necessary, the term "off-line" is used. Methods and apparatuses for preparing cheese are generally known from practice. See, in this context, the standard manual "Cheese (Chemistry, Physics and Microbiology)", third edition, volume 2, P.F. Fox, P.L.H. McSweeney, T.M. Cogan, T.P. Guinee, 2004" in particular pages 23 - 51 on "General aspects of cheese technology", and, in particular, Fig. 1 on page 24. In general, cheese is prepared by forming a milk gel from milk with rennet, cutting this milk gel into a mixture of curd and whey, draining the whey and then forming young cheese from the curd. Thereupon, the young cheese can be subjected to a brining treatment by introducing it into a brine bath for some time. After the young cheese has undergone a brining treatment, it is usually stored in a cheese warehouse to undergo a controlled ripening process.

In general, in industrial cheese preparation processes, all sorts of factors can be recognized which may cause (undesired) process malfunctions, loss of yield and/or quality variation of the cheese. Premature detection of

deviations in the cheese preparation may help to reduce or prevent process malfunctions, loss of yield and/or quality variation of the cheese. For instance, by monitoring the curdling of milk into milk gel, cutting the milk gel into a whey-curd mixture and/or the syneresis of curd particles of the whey-curd mixture, important indications for process malfunctions, loss of yield and/or variation in quality can be obtained. It is known, inter alia, that the size of the curd particles after cutting of the milk gel is of great influence on the moisture content of the eventually produced cheese (Australian Journal of Dairy Technology 9 (1954) 103-107), usually an important quality parameter of cheese. In addition, it is known that inadequate execution of the cutting process can lead to the formation of many small curd particles called grit and the loss of fat with the whey (Journal of Dairy Research 58 (1991) 345-354). Both result in a reduced cheese yield.

However, the methods for determining the course of curdling, cutting and/or the syneresis of the curd known from the state of the art are typically off-line, time consuming and/or require withdrawing relatively much material from the cheese preparation process. Therefore, they are considered unattractive for (frequent) industrial use.

An example of such a method is a sieve analysis of the whey-curd mixture after cutting for determining the particle size of the curd particles. This particles size is deemed of importance for the yield of the cheese preparation process because of the influence on the moisture content in the cheese. This method is described, for instance, by Johnston et al. (Int. Dairy Journal 8 (1998) 281-288). To this end, a sample of the whey-curd mixture is taken in an amount such that approximately 145 g of curd is obtained. This mixture is separated by a stack of five stainless steel wire sieves (mesh sizes 12, 7.5, 5, 3 and < 3 mm). The amount of curd on each sieve is then defined as a percentage of the total amount of curd on the sieves. It will be clear that this method can only be used off-line, is time consuming and requires withdrawing a relatively large amount of valuable material from the cheese preparation

process. In addition, the result of the method is not fully objective since its execution involves factors which differ from person-to-person.

Everard et al. (J. Dairy Sci. 90, 2007, 316203170) describe a method for monitoring syneresis of curd particles on the basis of color analysis of images. The images were recorded with exposure to fluorescent light in a cheese vat with a 10 liter content. As the authors have already put forward, this method is unsuitable for in-line measurements in industrial cheese preparation processes because of the occurrence of settlement of curd particles. The object of the present invention is to provide an inline measuring method which can monitor the course of the process of an industrial cheese preparation process. With the obtained data, the process can be adapted, in order that for instance process malfunctions, loss of yield and/or variation in quality can be prevented or minimized. The object of the present invention is in particular in-line monitoring of the course of the industrial cheese preparation process in the time period from the start of the cutting of the milk gel until the forming of the young cheese.

In the present invention, this object is achieved through inline registration of photographic images of at least one intermediate product in the processing of milk into cheese, whereupon the photographic images can be processed and analyzed by means of image analysis techniques, such as available in, for instance, commercial software packages for image analysis such as Halcon, of MVTec Software GmbH, Germany (www.mvtec.com) (see for instance "Computer vision" by D.H. Ballard and CM. Brown (Prentice -Hall Inc., Englewood Cliffs, New Yersey, 1982) and "The Image Processing Handbook" by J. C. Russ (CRC Press).

It will be clear that the photographic image involves, for instance, a two-dimensional depiction of a registered field of view on the intermediate product. The photographic image can be registered with the aid of, for instance, a (digital) camera. The term "photographic image" also comprises data or information associated with a photographic image. It will be clear that,

broadly speaking, an optical property, such as a light intensity, registered by a single light sensitive diode or quadrant sensor, cannot be regarded as a photographic image.

An image analysis of a photographic image can consist of, for instance, one or more of the following steps:

1. Contrast enhancement of the image

2. Background correction through filtering

3. Digitalization of the image utilizing a threshold value

4. Identification of particles (also called "blobs") 5. Identification of composite particles (also called "sub-blobs") by monitoring the form in successive photographic images.

6. Determination of the surface of each particle in the field of view

7. Calculation of the (equivalent) diameter of the particle.

By processing and analyzing the photographic images, process or product parameters relevant for cheese preparation can be derived. The knowledge of these relevant parameters can then be applied for adjusting the process. The advantage which can thus be achieved is that process malfunctions, loss of yield and/or variation in quality can be minimized. Further advantages of the method used are that the data are obtained rapidly and automatically, without great effort, highly accurately, independently of individuals and without loss of valuable product. The total cheese preparation process can thus be continuously monitored and the process needs no longer be controlled on the basis of incidental measurements.

In a first aspect, the invention concerns a method for industrial preparation of cheese, comprising curdling milk into a milk gel, cutting the milk gel into a whey -curd mixture and forming and optionally pressing the curd to form a young cheese, and further comprising inline registration of at least one photographic image of at least one intermediate product in the processing of milk into cheese, and processing and analyzing the at least one image into at least one process or product parameter. In addition, the

invention concerns an apparatus with which this method can be utilized, and a use of a system for inline registration of photographic images in a method for industrial cheese preparation and the subsequent processing and analyzing of the images into at least one process or product parameter. Examples of process and product parameters that can thus be obtained are curd size, amount of grit in the whey, extent of coagulation of curd particles, extent of syneresis of the curd particles (based on, for instance, measurements successive over time within a batch), whey/curd ratio in whey/curd mixtures (for instance as an indication of the extent of breaking of whey and curd), contrast between whey and curd (for instance as an indication of the amount of protein and fat in the whey), and variation of these parameters over time and/or per process line. Further, from the extent of difference in contrast between curd particles and surrounding liquid a parameter can be derived about the efficiency of the cheese preparation process in regard to protein and/or fat losses.

Utilizing measuring methods in the cheese preparation without cheese being lost ("nondestructive measurement) is generally known from the state of the art. From an earlier publication of applicant, it is known to perform nondestructive moisture measurements on young cheese before brining, for regulating the cheese preparation in time, in a manner with which the moisture content of the end product can be controlled within narrow boundaries (EP1388291). However, what is involved here is not the registration of a photographic image such as in the present invention, but a different measuring principle and method than in the present invention, that is to say reflection-near-infrared measurement.

In a preferred embodiment of the present invention, the at least one photographic image is registered in the time period from the start of the cutting of the milk gel until the forming of the young cheese. An example of a parameter that can be obtained by registering photographic images in the time period from the start of the cutting of the milk gel until the forming of the

young cheese is the size of the curd particles. The size of the curd particles can vary, as a result of, for instance, changes in the composition of the milk or as a result of deviating process conditions, such as curdling temperature or curdling time. Here, not only the initial size of the curd particles after cutting can be of influence, but also the change of the size under the influence of, for instance, syneresis. It is known that deviations in the size distribution of the curd particles can be accompanied by quality deviations of the cheese such as, for instance, a deviant moisture content. In addition, deviations in the size distribution of the curd particles can have adverse affects on the yield of the cheese preparation process. The formation of many small curd particles, also called grit, can for instance lead to loss of dry matter with the whey. As a rule, such deviations are detected only with the passage of time. This may lead to, for instance, the production of large amounts of cheese in which particular deviations occur before the disturbance causing this deviation is detected and remedied, which, in turn, as a rule entails high costs. It is therefore important that deviations in the particle size of the curd particles be detected at an early stage. The invention described here has as an advantage that the particle size of curd particles can be determined rapidly and automaticaUy, without much effort and without loss of valuable product. As measurements can simply be often repeated, the reliability of the measurements is furthermore much higher. Customary methods such as the sieve analysis do not have these advantages.

For an accurate determination of the size of the curd particles the quality of image formation is important; the curd particles should be well- identifiable in the image. Here, one or more of the following aspects can play a part:

1. Image definition: the curd particles should be depicted as clearly as possible. This is influenced both by the depiction optics (clarity, optical depth of penetration) and by the suppression of movement blur.

2. Contrast: the difference between curd particles and background (whey) should be as sharp as possible.

3. Homogeneity: the curd-whey mixture should be exposed as homogeneously as possible in the field of view.

Surprisingly, applicant has found that a good balance between the above-mentioned aspects can be achieved utilizing exposure to infrared light (wavelength between approximately 750 and approximately 1000 nm). Therefore, in a further preferred embodiment, the invention provides a method for registering at least one photographic image with exposure to infrared light. Lighting can also be carried out with the aid of so-called Light Emitting Diodes (LEDs). LEDs have as an additional advantage that flashes with a higher light intensity can be obtained, so that the recording time can be relatively short and movement blur in the images can be suppressed. Surprisingly, it has been found that it is well possible to measure on strongly moving systems, such as for instance in pipes in which whey-curd mixtures are transported, in that relatively short recording times for recording the photographical images can be utilized. This also allows for registration of photographical images of whey- curd mixtures such as for the purpose of the determination of the curd size distribution. The fact is that under less strongly moving conditions obtaining photographic images of whey-curd mixture can be hindered by the settlement of the curd particles. Therefore, a preferred embodiment of the invention relates to a method for industrial preparation of cheese which furthermore comprises a method in which the at least one photographic image is registered utilizing recording times of less than 10 milliseconds, more preferred less than 5 milliseconds, such as 1 or 2 milliseconds. A preferred embodiment of the invention further relates to a method for the industrial preparation of cheese in which further the at least one photographic image comprises an image of a whey-curd mixture. In a further preferred embodiment, the at least one

photographic image is processed and analyzed for determining the particle size distribution of the curd particles.

Preferably, the invention relates to a method for the industrial preparation of cheese which further comprises adjustment of the industrial cheese preparation process. With it, process malfunctions, loss of yield and/or quality variations can be minimized.

The invention further relates to an apparatus for the industrial preparation of cheese, comprising a vat for curdling milk, a cutting system for cutting a milk gel, at least one mold and optionally a pressing system for the preparation of young cheese and a system for inline registration of at least one photographic image of at least one intermediate product in the processing from milk into cheese and the processing and analyzing of the at least one image to at least one process or product parameter. Fig. 1 shows a schematic representation of an apparatus according to the invention. In addition, the invention relates to the use of a system for inline registration of at least one photographic image of at least one intermediate product in the processing of milk into cheese.

Fig. 2 shows a schematic representation of a system for inline registration of a photographic image of an intermediate product. It is further preferred that the apparatus is equipped with infrared lighting equipment and/or Light Emitting Diodes for exposure of the field of view.

The invention is presently further elucidated on the basis of the following non-limitative examples.

Fig. 1: Flow chart of a method according to the invention. In the embodiment shown, image registration of the whey-curd mixture takes place. The interrupted lines with arrows indicate that the product or process parameter can be utilized for (real time) adjustment or fine-tuning of the cheese preparation process, for instance curdling and/or cutting.

Fig. 2: Schematic representation of the apparatus according to the invention. The continuous lines between elements 2 — 5 represent transport means. The image processing system 6 can record one or several photographic images of at least one intermediate product. In the Figure, the interrupted lines with arrow indicate that images can be recorded of, for instance, curdled milk 2 in curdling vat 2 and/or of the whey-curd mixture in cutting system 3.

Fig. 3: Schematic representation of a set up for inline registration of photographic images of a whey-curd mixture. The exposure is carried out with a ring exposure of 64 infrared LEDs, placed in a ring having a diameter of 175 mm.

Fig. 4: Curd size distribution over time (for the time intervals 0-10, 10-20, 20-30, 30-40 and 40-50 minutes, the fraction is a fraction based on the number of particles).

Example 1

Comparison of the curd size distribution according to the photographic method and the sieve analysis.

The sieve analysis consisted of the following steps:

- representative sample taking of a whey-curd mixture, at a moment after cutting the curd and before forming the young cheese, this sample contains ~ 174 grams of curd; - separating the whey-curd mixture by means of 5 sieves with different mesh-width (11.0, 7.0, 5.0, 3.5 and 1.0 mm);

- determining the weight of curd per sieve and expressing this weight as percentage of the total weight of curd which remained behind on the 5 sieves; - calculating the average curd size.

Per curd maker one sieve analysis was performed. The test setup for the photographic method consisted of a monochrome camera with a 16 mm objective and a field of view with a diameter of approximately 80 mm as represented in Fig. 3. The lighting was a ring lighting of 64 infrared LEDs with a wavelength of 950 nm. The LEDs were flashed with a time duration of 0.5 ms and a current of approximately 0.5A, in a manner such that the flashes occurred at the same time as the photographic recordings. This test setup was placed before a window and used for depicting a whey -curd mixture flowing behind the window. Per curd maker, 480 images were recorded and read in by a computer. With the aid of a suitable software package for digital image processing, and as described in this application, these images were processed into histograms of the (equivalent) diameter of the curd particles. From the histograms, a surface average diameter of 5.00 mm was calculated. The weight average diameter based on the sieve analysis was 4.9 mm. From this it appears that the results of the two methods correspond well.

Example 2

Monitoring the curd size over time during syneresis.

The test setup for the photographic method described in Example 1 was used for monitoring the course of the curd size over time during cheese preparation. Per time interval of 10 minutes, the curd size distribution was determined, the result of which is represented in Fig. 4.

Also, the course of the total particle surface in the photographic image over time can be determined for gaining insight in the ratio of curd and whey. Such results can be used to verify whether any breaking in the whey-curd mixture occurs.

Example 3

Detecting process deviations.

In a cheese preparation process, the particle size of the curd particles was followed with the photographic method as described in Example 1. During the process, the average diameter suddenly dropped from 6.2 mm, the normal value for the respective type of cheese, to 5.5 mm. Examination of the cheese preparation process taught that due to a malfunction in the heating circuit the post-heating temperature had become too high, with the risk of the moisture content in the cheese being too low. This timely detection allowed for rapid intervention and the correct operation could be restored. The respective batch portion could thus also be rapidly isolated and be given a suitable use.

Description of the numbers in Figure 3: 1 = camera

2 = objective

3 = infrared (IR) LED 4. = field of view

5 = window 6 = curd