WO2023111450A2

WO2023111450A2 - Device for the quality control and classification of a stopper, and associated method

Info

Publication number: WO2023111450A2
Application number: PCT/FR2022/052343
Authority: WO
Inventors: Patrick Mounaix; Quentin CASSAR; Jean-Baptiste PERRAUD
Original assignee: Universite de Bordeaux; Institut Polytechnique De Bordeaux; Centre National De La Recherche Scientifique
Priority date: 2021-12-16
Filing date: 2022-12-13
Publication date: 2023-06-22
Also published as: FR3130973A1; WO2023111450A3

Abstract

The present invention relates to a device (1) for the quality control and classification of a stopper (3), comprising: - a source (2) suitable for generating electromagnetic radiation in order to illuminate the stopper, said radiation being emitted in a spectral range suitable for passing through the stopper; - a detector (4) suitable for detecting the radiation transmitted by the stopper; and - a computing unit (10) comprising a unit (11) for constructing an image from the signals originating from the detector, the greyscale level of a pixel of the image representing an item of information about the transmissivity of a measurement point on the stopper, and a classifier unit (12) suitable for assigning a quality grade to the stopper by comparing a population distribution of greyscale pixels extracted from said image to a reference population distribution of greyscale pixels.

Description

DEVICE FOR QUALITY CONTROL AND CLASSIFICATION OF A CORK STOPPER AND ASSOCIATED METHOD

Technical area

This disclosure falls within the field of quality control and classification of cork stoppers, called natural stoppers intended in particular to close wine bottles. More particularly, it relates to a device and a method which makes it possible to control the quality of cork stoppers and to classify them from the characterization of the porous internal structure of cork stoppers based on images obtained by terahertz electromagnetic waves. (THz).

Prior technique

[0002] Cork stoppers are made from the casing of cork oak bark and are used for sealing wine bottles, and in particular for the conservation and aging of the wine contained in the bottle.

[0003] The quality of a cork stopper, called a natural stopper, depends on its mechanical and elastic properties and also on the capacity of its internal structure to allow reasonable quantities of oxygen to pass from the outside to the inside of the cork. bottle, in contact with wine. Indeed, the tannin compositions of wines consume oxygen and the physical sealing properties of the cork will allow a more or less rapid and quantitative kinetics of oxygen passage, thus influencing the sensory profile of the wine after having been stored. in a bottle.

[0004] Currently, corks are visually sorted and classified into seven categories based solely on their apparent surface defects. The more lenticels there are on the surface, the more the cork is considered to be of poor quality. Corks classified as “Flower”, for example, have no apparent defect (no hole, woody lenticel, color defect, etc.). The seven quality classes are called by the following names: Flower (grade 0), Extra (grade 1), Super (grade 2), First class (grade 3), Second class (grade 4), Low (grade 5), defective (grade 6). This sorting is currently carried out visually by qualified operators or automatically using photographic image analysis.

[0005] However, such a surface control does not make it possible to guarantee the overall quality of a cork. Indeed, cork being a particularly heterogeneous material, the distribution of defects on the surface of the stoppers is not necessarily representative of the internal distribution of porosity of the cork stopper. The method of visual control used for the distribution of corks into different classes, can lead to an erroneous classification of corks. Thus, it is found that up to 10% of corks are classified at a class of higher quality than that to which they would be attributed in internal knowledge. An analysis based on the number of defects on the surface of the plugs therefore leads to an underestimation of the internal porosity. The visual sorting carried out by equipment, used by the producers of tubed cork stoppers, does not make it possible to classify the stoppers according to their internal porosity, that is to say their capacity in terms of oxygen passage. Misclassification of a cork stopper can affect the preservation of wines, and can have a significant sensory impact on the wine.

[0006] There is therefore a real need to be able to analyze the internal structure of corks, in order to offer winegrowers ranges of natural corks with a reliable classification.

[0007] The inventors note that terahertz (THz) waves whose wavelength is between 30 μm and 3 mm are particularly well suited for probing the internal structure of plugs. Indeed, according to the inventors, the terahertz electromagnetic waves have sufficient penetrating power to pass through the dielectric material constituting the cap. Its non-ionizing radiation also allows use in an unprotected environment unlike X-rays.

[0008] Another essential parameter is the spatial resolution so as to be able to detect the structural defects present in the stopper. The spatial resolution of an imaging system is limited by the diffraction phenomenon and is calculated according to the Rayleigh criterion. The use of terahertz waves allows an imaging system to achieve a theoretical spatial resolution between 36.6 pm and 3.66 mm. The structural defects present in the plug, which are of the order of a few hundred micrometers, can therefore be detected at depth by terahertz waves.

[0009] The document WO2012/058567 describes a method which uses terahertz spectroscopy to detect the presence of impurities present in the stopper and to measure the diffusion coefficient of a liquid through a stopper. The technique described in this document does not make it possible to characterize the internal porosity of the plugs and to classify them into different categories of classes.

[0010] This disclosure therefore aims to overcome all of the disadvantages mentioned above.

[0011] An objective of the present disclosure is to propose a device which makes it possible to characterize internal structural defects which is reliable, while being non-destructive.

[0012] Another objective of the present disclosure is also that the device can be implemented easily and quickly and at reduced cost. [0013] Finally, another objective of the present disclosure is to propose a solution which can be easily integrated into a cork stopper production line.

Summary

[0014] This disclosure improves the situation.

[0015] A device for quality control and classification of a cork stopper comprising:

- a source adapted to generate electromagnetic radiation to illuminate the stopper, said radiation being emitted in a spectral range adapted to pass through the stopper;

- a detector adapted to detect the radiation transmitted by the plug; And

- a calculation unit comprising a unit for constructing an image from the signals coming from the detector, the gray level of a pixel of the image representing information on the transmissivity of a measurement point of the cork stopper, a classification unit adapted to assign a quality class to the cap by comparing a distribution of the populations of gray level pixels extracted from said image with a distribution of the reference gray level pixel populations.

[0016] The characteristics set out in the following paragraphs can optionally be implemented, independently of each other or in combination with each other:

The source is suitable for emitting radiation in a terahertz spectral range between 60 GHz and 6 THz, preferably between 60 GHz and 400 GHz.

The distribution is a Gaussian distribution.

[0019] The device further comprises a database stored in a memory, said database comprising a plurality of distributions of reference gray level pixel populations, each of said distributions being associated with a cap having a quality class known.

The device further comprises two translation plates capable of moving the source and the detector in translation relative to the stopper.

According to one embodiment, the device comprises an optical system configured to form the incident radiation into a spot on the cap.

According to another embodiment, the device comprises an optical system configured to form the incident radiation in a line on the plug. [0023] According to another aspect, there is proposed a process for quality control and classification of a cork stopper comprising the following steps:

- emission of electromagnetic radiation located in a defined spectral range, in a direction normal to a surface of the cap, to illuminate one face of the cap, said radiation being emitted in a spectral range adapted to pass through the cap;

- detection of the radiation transmitted by the cap;

- calculation of information on the transmissivity of the radiation transmitted at each measurement point, the transmissivity being the capacity of the plug at each measurement point to transmit the incident radiation;

- construction of an image representative of the transmissivity information calculated at each measurement point of the cork stopper, the gray level of a pixel of the image representing the transmissivity information of a measurement point of the cork stopper ;

- determination of a distribution of the gray level pixel populations from said image;

- comparing said pixel population distribution to a reference gray level pixel population distribution;

- classification of the cork in a quality class.

[0024] The characteristics set out in the following paragraphs can optionally be implemented, independently of each other or in combination with each other:

[0025] The distribution of the populations of pixels in gray level is a Gaussian distribution.

The spectral range is between 60 GHz and 6 THz, preferably between 60 GHz and 400 GHz.

The emitted radiation is amplitude modulated.

[0028] The frequency of the radiation is modulated over time with a waveform chosen from the following forms: saw wave, triangular wave.

According to one embodiment, the step of calculating the transmissivity information consists in calculating the maximum transmissivity for each waveform of the radiation at a measurement point.

[0030] The distribution of the populations of pixels in reference gray level is obtained on corks having a known quality class. According to one embodiment, the electromagnetic radiation illuminates the stopper by forming a spot, the plurality of measurement points resulting from a sequential illumination of spots.

According to another embodiment, the electromagnetic radiation illuminates the plug forming a line, the plurality of measurement points resulting from a sequential illumination of lines.

Brief description of the drawings

[0033] Other characteristics, details and advantages will appear on reading the detailed description below, and on analyzing the appended drawings, in which:

Fig. 1

[0034] [Fig. 1] represents a quality control and classification device according to an embodiment in which the incident radiation illuminates the stopper in the form of a spot on the stopper.

Fig. 2

[0035] [Fig. 2] shows a top view of a quality control and classification device according to another embodiment in which the incident radiation illuminates the stopper in the form of a line.

Fig. 3

[0036] [Fig. 3] shows a terahertz image obtained from thirty-eight corks of different quality obtained with the device of figure 1.

Fig. 4

[0037] [Fig. 4] shows an amplitude curve of a waveform acquired at an XY measurement point.

Fig. 5

[0038] [Fig. 5] shows on the upper part: a histogram of the populations of pixels in gray level for corks of the "Flower" quality class or grade 0 and different Gaussian distributions of the associated pixel populations, on the lower part: a histogram of the grayscale pixel populations for caps of the "defective" or grade 6 quality class and different distributions of associated pixel populations.

Fig. 6 [0039] [Fig. 6] shows a diagram detailing the main steps of the quality control and classification process implementing the device of Figure 1.

Description of embodiments

In the rest of the description, the electromagnetic radiation is called "incident beam" over the entire path from the source to the stopper to be analyzed, and "transmitted beam" over the entire path from the stopper to the detector.

[0041] In the remainder of the description, the cork stopper is referred to as a “stopper”.

In the remainder of the description, the terms “upstream” and “downstream” are defined according to the direction of the radiation.

[0043] Reference is now made to Figure 1 which is a schematic representation of a quality control and classification device 1 of a cork stopper 3 according to one embodiment.

The device comprises a source 2 adapted to generate electromagnetic radiation 17, a detector 4, an optical system comprising optical components 5, 6, 7, 8 placed on the optical path of the radiation between the source 2 and the cap 3 and between the stopper 3 and the detector 4 and a calculation unit 10 (COMPUTATION UNIT).

[0045] Figure 1 indicates a z axis as the direction of propagation of the radiation emitted from the source 1 in the direction of the plug, an X axis as the transverse direction perpendicular to the Z axis and a Y axis as the direction vertical perpendicular to the Z axis.

In the example illustrated in Figure 1, the cork stopper 3 to be analyzed is presented alone. The radiation emitted by the source only passes through the middle of the plug before being detected by the detector.

[0047] In general, the plug has a cylindrical shape with a substantially circular section. In FIG. 1, the incident radiation is oriented along a direction normal to a surface of the cap.

The cork has visible grooves and the stopper has an anisotropic medium. Preferably, two series of measurements are carried out for each stopper in order to avoid any artefact which could induce errors in the interpretation of the results. For this, a first series of measurements is carried out by orienting the plug so that its growth plane is parallel to the incident terahertz beam. In a second series of measurements, the plug is oriented so that its growth plane is perpendicular to the incident terahertz beam. The source 2 is a source suitable for generating terahertz radiation in a terahertz spectral range suitable for passing through a dielectric medium such as the middle of the cap. The terahertz domain designates so-called millimetric electromagnetic waves located between the far infrared and radar waves (microwaves). In this frequency range, composite materials and plastic materials are transparent to terahertz radiation. The spectral range of terahertz radiation is between 60 GHz and 6 THz. Preferably, the spectral domain is between 60 GHz and 400 GHz for the radar sources.

According to one embodiment, the radiation emitted by the source can be a continuous wave (CW). According to another embodiment, the radiation emitted is a frequency modulated continuous wave.

The radiation generated by the source illuminates the stopper by forming a spot. For this, the radiation emitted by the source 2 is collimated by a first lens 5 and focused on the stopper by a second lens 6. The radiation, after having passed through the stopper, is again collimated by a third lens 7 then focused on the detector by a fourth lens 8. The detector located downstream of the stopper, detects the signal and outputs, to the calculation unit 10, in real time, detection signals corresponding to the intensities of the transmitted radiation. The first lens and the second lens are configured to form on the stopper a spot with a diameter substantially between 3 mm and 1 mm depending on the frequency of the source.

The device 1 further comprises two translation plates 15, 16 allowing the source 2 and the detector 4 to be moved synchronously in the transverse direction X and the vertical direction Y in an orthonormal frame (X, Y, Z) . The translation plate 15 makes it possible to move the source 2 in both directions so as to direct the incident beam and to illuminate one side of the cork stopper in a sequential manner in the form of spots. In other words, the plurality of measurement points results from a sequential illumination of spots. The translation stage 16 makes it possible to move the detector in both directions so as to sequentially detect the beam transmitted by the plurality of measurement points. The turntables are controlled by a motor which is controlled by a control module. The control module is able to determine a series of displacements in the two X and Y directions to move the source and the detector sequentially. The detector 4 transmits the signals representative of the intensity of the beam transmitted to the calculation unit 10.

[0053] Referring to Figure 2, a top view of a quality control and classification device 50 according to another embodiment is illustrated. The device 50 comprises a source 2 adapted to generate electromagnetic radiation, a detector 54, and a calculation unit 10 (COMPUTION UNIT).

The device 50 comprises an optical system configured to form the incident beam in the form of a line on the cap. By way of example, the optical system is a cylindrical lens 19 which is adapted to generate a beam line. The cylindrical lens 19 is placed between the source and the stopper. The device comprises a detector 54 configured to collect a set of measurement points from the beam line transmitted by the plug. In this configuration, the translation stage 15 associated with the source 2 makes it possible to move the source 2 in a vertical direction so as to direct the incident beam and to illuminate sequentially in the form of a plurality of lines on the plug in cork. In other words, the plurality of measurement points results from a sequential illumination of lines.

[0056] The calculation unit 10 comprises a two-dimensional and three-dimensional image construction unit 11 if one comes to visualize the cork from other angles of view (IMAG), a classification unit 12 (CLASS) and a data storage memory 13 (MEM).

The image construction unit 11 is configured to construct a two-dimensional or three-dimensional image from the signals transmitted by the detector. The two-dimensional image represents transmissivity information associated with each measurement point. In other words, each level of gray of an image pixel represents a value of transmissivity information associated with a measurement point of the traffic jam. The transmissivity corresponds to the capacity of the cap at each measurement point to transmit the incident radiation. It is evaluated by calculating the transmission coefficient of the plug at each measurement point.

The construction unit 11 receives the signals from the detector and calculates the maximum transmissivity for each waveform acquired by the detector for each measurement point.

Referring to Figure 4, a waveform acquired by the detector 4, 54 at a measurement point is shown. The curve represented in FIG. 4 represents the amplitude of the signal transmitted at a specific measurement point, through a sample of the cork stopper. This curve is obtained by analyzing the signal detected by the detector and transmitted to the calculation unit. The maximum amplitude value associated with this measurement point is selected and stored in order to be able to be used to construct the two-dimensional image. [0060] The construction unit 11 generates a two-dimensional image from the maximum amplitude values selected for each waveform. Each level of gray or chrominance of a pixel of the image represents the transmissivity associated with a measurement point of the traffic jam. The construction unit assigns each pixel in the image a gray level scale value proportional to the transmissivity of the radiation transmitted by the cap.

[0061] Referring to Figure 3, a two-dimensional image of thirty-eight plugs is shown. According to one embodiment, the frequency of acquisition of the measurement points is 300 GHz with an acquisition step every 500 pm. For example, the acquisition time for a plug with a transverse dimension of 28 mm and a vertical dimension of the order of 48 mm is 120 seconds.

The two-dimensional image is saved in a memory 13. It is also transmitted to a display unit 14 allowing the generated 2D image to be viewed.

The 2D image comprises a distribution of pixels each having a given transmissivity value. Cork being a dielectric material relatively transparent to terahertz radiation, the transmissivity value of the beam transmitted through a zone of a cork reflects the internal structure. The applicants note that a good transmissivity whose gray level is close to 1 (one) reflects the absence of a diffusing and absorbing cavity. Conversely, an area of a plug whose gray level is close to 0 (zero) reflects the presence of one or more diffusing and absorbing defects. It is thus possible to correlate the distribution of pixels having a given gray level and the quality of a traffic jam. In particular, it is possible to model the distributions in the form of Gaussian curves whose parameters depend on the quality of the closure to classify the closures into different categories.

[0064] The classification unit 12 notably comprises an image processing unit making it possible to quantify the number of pixels having a given gray level value. It is then possible to represent this quantification in the form of a histogram. On the ordinate of this histogram are counted the number of pixels, of the same attenuation (level of gray), the attenuation being represented on the abscissa. From this histogram, it is possible to model a distribution, for example a Gaussian distribution of the populations of pixels of a given gray level representative of the quality of the plug. By comparing this Gaussian distribution with a reference Gaussian distribution determined on a cork having a known quality class, it is therefore possible to assign a quality class to the analyzed cork. [0065] With reference to FIG. 5, Gaussian distributions are represented for two categories of plugs: Flower and Defective.

On the upper part of FIG. 5 is represented a histogram of populations of pixels in gray level for corks of the “Flower” quality class or grade 0. The curve referenced A1 represents a Gaussian distribution associated with pixels having a gray level close to 1. The curve referenced A2 represents a Gaussian distribution associated with pixels having a gray level close to 0.

In the lower part of FIG. 5 is shown a histogram of populations of pixels in gray level for caps of the “defective” quality class or grade 6. The curve referenced B1 represents a Gaussian distribution associated with pixels having a gray level close to 1. The curve referenced B2 represents a Gaussian distribution associated with pixels having a gray level close to 0.

From these distributions, the inventors have determined that the height of the Gaussian distributions varies according to the quality of the corks and constitutes a classification parameter for classifying the corks. Indeed, they find that the height of Gaussian distributions associated with populations of conforming pixels (gray level close to 1) decreases from Flower to Defective categories. Conversely, the height of the Gaussian distributions associated with populations of nonconforming pixels (gray level close to 0) increases from Flower to Defective categories.

A process 100 for controlling and classifying the quality of a cork is illustrated in FIG. 6.

The method comprises a step 101 of emitting an incident beam located in a spectral range between 60 GHz and 6 THz, in a direction normal to a surface of the cap. The beam is collimated and sequentially focused on the cap.

According to an exemplary embodiment, the incident beam illuminates the stopper in the form of a spot. According to another exemplary embodiment, the incident beam illuminates the plug in the form of a beam line.

[0072] During a step 102, the beam transmitted by the plug is collimated and focused sequentially on a detector. The transmitted beam is transformed into an electrical signal by the detector and transmitted to the calculation unit 10.

In the case where the incident beam forms a spot on the stopper, the detector 4 is adapted to detect a transmitted beam corresponding to each spot. Each measurement point corresponds to a spot. In the case where the incident beam forms a line on the cap, the detector 54 is suitable for detecting several transmitted beams coming from the line. Thus, each line generates several measurement points.

During a step 103, the transmissivity information of the beam transmitted at each measurement point is calculated and stored in a memory 13.

The step of calculating the transmissivity information resulting from the interaction between the THz waves and the material of the plug comprises the following steps:

- binary encoding;

- binary decoding;

- calculation of the maximum transmissivity for each waveform at each point.

During a step 104, a two-dimensional image is constructed from the plurality of calculated maximum transmissivity values associated with measurement points. The gray level of each pixel of the image represents the transmissivity information of a measurement point of the traffic jam.

Advantageously, it is also possible to correlate the density of the cork with the transmission determined for each stopper. Density is calculated by taking the ratio of mass to volume. The presence of cavities and lack of material therefore directly influences the density. Similarly, the measurement of the effective index of the medium is correlated to the density.

During a step 105, a distribution of gray level pixel populations is determined from the two-dimensional image. According to an exemplary embodiment, this distribution can be a Gaussian distribution.

During a step 106, the distribution of determined pixel populations is compared with a reference distribution.

According to an exemplary embodiment, the distribution of reference pixel populations is obtained on cork stoppers classified with a known quality class. The quality class of the reference cork stoppers is obtained by another imaging technique, for example by X-rays. Preferably, the memory 13 comprises a database storing a plurality of the reference distributions.

During a step 107, the cork is classified in a quality class associated with the reference Gaussian distribution if a correspondence is established between the determined Gaussian distribution and the reference Gaussian distribution.

[0083] Although the detailed description above relates to an embodiment in which the source and the detector are moved and the plug is held fixed facing the source, it is clear that the invention also applies to an alternative embodiment in which the source and detector assembly is kept fixed and the stopper is moved in translation in two orthogonal dimensions. In another embodiment, the cap can be rotated.

Claims

[Claim 1] Device (1) for quality control and classification of a cork stopper (3) comprising:

- a source (2) adapted to generate electromagnetic radiation to illuminate the stopper, said radiation being emitted in a spectral range adapted to pass through the stopper;

- a detector (4) adapted to detect the radiation transmitted by the plug; And

- a calculation unit (10) comprising a construction unit (11) of an image from the signals coming from the detector, the gray level of a pixel of the image representing information on the transmissivity of a point of measurement of the cork stopper, a classification unit (12) adapted to assign a quality class to the stopper by comparing a distribution of gray level pixel populations extracted from said image with a distribution of gray level pixel populations of reference.

[Claim 2] Device according to claim 1, in which the source (2) is suitable for emitting radiation in a terahertz spectral range between 60 GHz and 6 THz, preferably between 60 GHz and 400 GHz.

[Claim 3] Apparatus according to claim 1 or 2, wherein the distribution is a Gaussian distribution.

[Claim 4] Device according to one of Claims 1 to 3, further comprising a database stored in a memory (13), said database comprising a plurality of distributions of the reference gray level pixel populations, each of said distributions being associated with a cork having a known quality class.

[Claim 5] Device according to one of claims 1 to 4, further comprising two translation plates (15, 16) able to move the source (2) and the detector (4) in translation relative to the stopper.

[Claim 6] Device according to one of Claims 1 to 5, comprising an optical system (5, 6) configured to form the incident radiation into a spot on the cap.

[Claim 7] Device according to one of Claims 1 to 5, comprising an optical system (19) configured to form the incident radiation in a line on the cap.

[Claim 8] Method (100) for quality control and classification of a cork stopper (3) comprising the following steps:

- emission of electromagnetic radiation located in a defined spectral range, in a direction normal to a surface of the cap, to illuminate one face of the cap, said radiation being emitted in a spectral range adapted to pass through the cap (101);

- detection of the radiation transmitted by the cap (102);

- calculation of information on the transmissivity of the radiation transmitted at each measurement point (103), the transmissivity being the capacity of the cap at each measurement point to transmit the incident radiation;

- construction of an image representative of the transmissivity information calculated at each measurement point of the cork stopper, the gray level of a pixel of the image representing the transmissivity information of a measurement point of the cork stopper (104);

- determining a distribution of gray level pixel populations from said image (105);

- comparing said pixel population distribution to a reference gray level pixel population distribution (106);

- classification of the cork in a quality class (107).

[Claim 9] A method according to claim 8, wherein the gray level pixel population distribution is a Gaussian distribution.

[Claim 10] A method according to claim 8 or 9, wherein the spectral range is between 60 GHz and 6 THz, preferably between 60 GHz and 400 GHz.

[Claim 11] Method according to one of Claims 8 to 10, in which the emitted radiation is amplitude modulated.

[Claim 12] Method according to one of Claims 8 to 11, in which the frequency of the radiation is modulated in time with a waveform chosen from the following forms: saw wave, triangular wave.

[Claim 13] Method according to one of Claims 8 to 12, in which the step of calculating the transmissivity information consists of calculating the maximum transmissivity for each waveform of the radiation at a measurement point.

[Claim 14] Method according to one of Claims 8 to 13, in which the distribution of the reference grayscale pixel populations is obtained on corks having a known quality class.

[Claim 15] Method according to one of Claims 8 to 14, in which the electromagnetic radiation illuminates the stopper by forming a spot, the plurality of measurement points resulting from a sequential illumination of spots. -15-

[Claim 16] Method according to one of Claims 8 to 14, in which the electromagnetic radiation illuminates the stopper in a line, the plurality of measurement points resulting from sequential illumination of lines.