WO2023066433A1

WO2023066433A1 - Method for classifying the surface texture of rubber products

Info

Publication number: WO2023066433A1
Application number: PCT/DE2022/200192
Authority: WO
Inventors: Marco Rudolph; Muizzuddin Faiz BIN MOHMAD FAUZI
Original assignee: Continental Reifen Deutschland Gmbh
Priority date: 2021-10-21
Filing date: 2022-08-19
Publication date: 2023-04-27
Also published as: DE102021211888A1

Abstract

The invention relates to a method and a device for classifying the surface texture of rubber products, in particular tyre components in the retreading of tyres; the method comprises the following method steps: a) producing or providing a rubber product, comprising an at least partially vulcanized rubber compound, b) working a surface of the rubber product by removing material with one or more tools, and c) classifying the surface texture of the worked surface with a classifying unit, wherein the classifying unit comprises: i) at least one illuminating device for illuminating the surface of the rubber product at an angle of incidence of more than 10°, ii) at least one image-recording device for recording the surface of the rubber product in the area illuminated by the illuminating device, and iii) at least one electronic data-processing device connected to the image-recording device.

Description

Method for grading the surface texture of rubber products.

Description

The invention relates to a method for classifying the surface texture of rubber products, in particular tire components in the retreading of tires and a method for retreading a tire and an apparatus for classifying the surface texture of rubber products. Also disclosed is a computer program product and a control device for controlling a device for classifying the surface texture of rubber products.

The subject matter of the invention is defined in the appended claims.

The manufacture of modern vehicle tires is a complicated and relatively cost-intensive process in which considerable amounts of valuable raw materials have to be used on a regular basis. For this reason, processes for retreading used tires have been developed, with which it is possible to recycle a large proportion of used tires. The known methods for retreading are essentially based on the fact that the old, usually worn tread is removed so that a new tread can be applied to the underlying carcass surface.

Among the processes for retreading, so-called cold retreading is important. In this process, the old tread of the old tire is removed from the carcass, for example by grinding. This leaves an old tire with its carcass surface covered with rubber roughened ("casing"). A characteristic of cold retreading is that the new treads are pre-vulcanized ("pre-cured") so that they can be applied to the casings in the cold retreading process, such as is carried out in cold retreading plants (so-called "retread shops"). The resilient connection between the old tire and the new tread is guaranteed by a suitable adhesive, which creates a material connection between the two components. This is done, for example, by vulcanizing a sulfur-curable adhesive, for example a rubber mixture.

In contrast to this, in so-called hot retreading, a vulcanizable rubber compound is applied to the carcass surface and then vulcanized in a vulcanization device in such a way that a permanent connection to the carcass surface is produced and at the same time a tread profile is created in the surface of the tread.

A specifically adjusted roughness is required on the exposed carcass surface or on the back of the pre-vulcanized treads so that a long-term and firm bond can develop between the tire components. This set roughness must be within a specific range that is neither too large nor too small. If this specific roughness is not set, the tread can come loose when using the retreaded tire, which must be avoided at all costs. This can be caused, for example, by the fact that the adhesion is insufficient in areas that are too smooth or that the contacting of the two components with the adhesive is insufficient, since this collects in the depressions of the rough structure of a surface that is too rough. For this reason, the surface of the old tire largely freed from the old tread or the surface of the new tread to be applied are roughened in the prior art using material-removing methods. In the prior art, the roughness produced is determined, ie the surface structure is evaluated, mostly by means of a visual inspection by the workers employed with the aid of standardized roughness tests, with which the surface textures produced can be assigned to predefined classes. These predefined reference surfaces are therefore used for the manual classification of the surface texture, with the workers employed often evaluating the covering of the surface with removed material in addition to the surface texture in order to improve it with further cleaning measures, if necessary, and gross material defects (e.g. macroscopic defects or non-roughened ones). To identify areas, so-called "shiny spots").

The method known from the prior art for classifying the surface texture is regularly felt to be disadvantageous. It places high demands on the qualifications of the workers used so that they can reliably classify the surface texture. In addition, manual visual inspection is very time-consuming and often requires a large number of workers to ensure an acceptable throughput rate. Thus, the method known from the prior art is regularly felt to be inefficient in terms of time and costs.

Despite all the care taken by the workers used, with the method known from the prior art it can still happen that tire components with unsuitable surface textures get into further processing, especially if the workers used are tired and/or unfocused and thus responsible for an incorrect evaluation become more vulnerable. Corresponding misclassifications can lead to unwanted rejects in production from retreaded tires. In particularly serious cases, if corresponding deficiencies are not recognized early, corresponding tires could show an increased susceptibility to errors later in operation, which must be avoided at all costs.

In addition, it is not possible with the method known from the prior art to automatically adapt the upstream processing of the surfaces of the tire components to the surface texture achieved if deviations in the surface classification from the intended classification range are observed.

It was the primary object of the present invention to eliminate or at least reduce the disadvantages known from the prior art.

In particular, it was therefore the object of the present invention to provide an improved method for classifying the surface texture of rubber products, in particular tire components in the retreading of tires, which enables reliable classification of the processed rubber products and thus a consistent quality of the retreaded tires manufactured with these rubber products can reliably guarantee.

In this respect, it was an object of the present invention that the classification method should be able to be carried out in a particularly cost-effective and time-efficient manner. It was particularly desirable that the method to be specified should also be able to be used in continuous operation.

In addition, it was an object to design the specified method so that the processing of the surfaces of the rubber product depending on the observed classification of the Surface texture should be controllable, preferably already when deviations from the range of acceptable classes are only indicated.

The method to be specified for the classification should ideally be able to support the workers used in the classification or even at least partially replace them.

One of the tasks of the method to be specified was that it should enable the most objective possible assessment of the surface texture and thus lead to an improvement in the testing process.

An additional specification was that the process to be specified should ideally be suitable for reliably identifying foreign objects lying on the surface of the rubber products, in particular residues of the removed material, as well as other defects, such as macroscopic defects or unprocessed areas.

It was an additional object of the present invention that the method to be specified should be designed in such a way that, in addition to a basic classification, information about the actual roughness or surface topography of the surface can also be obtained with the method.

It was a further object of the present invention that the method to be specified should be able to be carried out comparatively flexibly with regard to the classification of the surface texture in order to adapt the method to be specified to the needs of different rubber products. It was desirable here that the method to be specified should have a high level of operational reliability and ideally be executable as a contactless method.

In addition, it was desirable for the method to be specified to have an increased degree of automation compared to the prior art and should ideally be achievable through a combination of components available on the market.

Against this background, it was an additional object of the present invention to specify a method for retreading a tire and a device for classifying the surface texture of rubber products.

In addition, it was an object of the present invention to specify a computer program product and a control device for controlling a device for classifying the surface texture of rubber products.

The inventors of the present invention have now found that the above objects can be surprisingly achieved when a classification unit as defined in the claims is employed in a method for classifying the surface texture of rubber products.

The above objects are correspondingly achieved by methods and devices as defined in the claims. Preferred configurations according to the invention result from the dependent claims and the following statements.

Such features of methods and devices according to the invention, which are referred to below as preferred, are combined in particularly preferred embodiments with other features referred to as preferred. Combinations of two or more of the embodiments described below as particularly preferred are therefore very particularly preferred. Also preferred are embodiments in which a feature designated as preferred to some extent is combined with one or more other features designated as preferred to some extent become. Features of preferred devices according to the invention and the disclosed computer program products and control devices result from the features of preferred methods.

The invention relates to a method for classifying the surface texture of rubber products, in particular tire components in the retreading of tires, comprising the following method steps: a) producing or providing a rubber product comprising an at least partially vulcanized rubber mixture, b) processing a surface of the rubber product by removing it of material with one or more tools, and c) classifying the surface texture of the machined surface with a classifying unit, the classifying unit comprising: i) at least one lighting device for illuminating the surface of the rubber product with an angle of incidence of more than 10°, ii) at least one Image recording device for recording the surface of the rubber product in the area illuminated by the lighting device, and iii) at least one electronic data processing device connected to the image recording device.

In the method according to the invention, the rubber product to be processed, which can be, for example, a tread strip or a used tire at least partially freed from the old tread strip, can be produced or provided. The Provision can be made, for example, by purchasing the relevant component from a supplier. On the other hand, the production takes place, for example, by removing an old tread from a used tire or by producing a pre-profiled tread by vulcanization.

For the purposes of the present invention, however, the rubber product can also be a used tire whose old tread is only removed by the tools in the course of the method according to the invention.

According to the invention, the rubber product comprises an at least partially vulcanized rubber mixture. In accordance with professional understanding, this is not only the case for used tires but also for the pre-cured treads used in cold retreading. It has proven to be an advantage of the method according to the invention that the rubber mixture contained in the rubber products does not have to be completely vulcanized, even if this can be regarded as advantageous for most applications. As a result, it is also not necessary to examine whether the rubber product is already fully vulcanized, which can be difficult in view of the complex material structure of vulcanized rubber compounds.

In step b) a surface of the tire component is processed, i. H. at least one surface of the tire component. This is done by removing material with one or more tools in a material-removing process and at least also serves the purpose of roughening the surface of the rubber product.

The process steps a) and b) corresponding to the steps known to those skilled in the art, with which the surface of a prepared or provided rubber product is roughened. In process step c), the surface texture of the processed surface is classified, which in the prior art - as explained above - is usually carried out by workers by visual inspection, which are given defined surfaces of comparable rubber products as a comparison, the known classes of classification are assigned. In contrast to this, in the method according to the invention, classification takes place with a classification unit, so that the classification is automatic.

In the context of the present invention, the step of classifying or classifying means that a surface is assigned to one of several classes of a classification, which in turn was generated by means of classification. Such a surface assigned to a class of the classification is a classified surface or a surface with a specific classification within the meaning of the present invention.

In accordance with the understanding of those skilled in the art, the step of grading means that the processed surface as a whole or partial areas of the surface depending on the surface texture, i. H. in practice, the mean surface texture, an indicator suitable for classification is assigned. This means that a class is assigned to the processed surface or the partial areas, which represents a summary upper unit for surfaces with comparable properties with regard to the surface parameters.

Consistent with the understanding of the art, the variation between the surface textures of multiple machined surfaces classified by the same class is assigned become, on average, lower than between surfaces assigned to different classes. The concept of classifying surface textures of machined surfaces is fundamentally known to those skilled in the art of machining rubber products. In view of the diverse possible surface properties and the large number of surfaces to be evaluated per unit of time, a more precise characterization method, for example the precise determination of the surface roughness, is regularly not possible efficiently and, moreover, often not necessary with this level of precision.

The classification unit includes a lighting device for illuminating the surface of the rubber product. In addition, the classification unit includes an image pickup device that captures the surface of the rubber product in the illuminated area. In accordance with the understanding of those skilled in the art, a plurality of lighting devices and/or a plurality of image recording devices can be provided here.

In the method according to the invention, the surface of the rubber product is illuminated by the at least one illumination device with an angle of incidence of more than 10°. In accordance with the understanding of those skilled in the art, the angle of incidence corresponds to the angle between the incident beam, or the central beam path of the lighting device, and the normal of the surface at the point of incidence of the beam.

This expresses the fact that the surface of the rubber product is illuminated obliquely, which can be achieved, for example, by arranging the lighting device at an angle to the side of the rubber product. This form of side lighting creates on the rough surface of the rubber product a shadow gradient (bright and dark field effect; so-called "dark field effect" and "light field effect"). This shadow progression can be recorded with the image recording device. The image recording device is now connected to an electronic data processing device, by means of which the classification takes place depending on the recording of the image recording device.

As explained above, the total surface can be classified. However, a method in which a large number of partial areas of the surface are each subjected to a separate classification is preferred, particularly with regard to the computing power required and to obtaining a spatially resolved assessment of the surface quality. This can be achieved, for example, by dividing the surface of the rubber product by a theoretical grid pattern into a large number of theoretical tiles, in each of which the classification takes place. Alternatively, other classification criteria can also be used for dividing into partial areas. For example, it is possible to align the subdivision with the surface texture in order to delimit sub-areas of similar characteristics from one another and then to classify them.

With this procedure, an overall evaluation of the surface of the rubber product can advantageously be carried out by evaluating the partial classifications. For example, the overall rating when using numerical values assigned to the classes can be determined by the arithmetic mean or the median of the partial classification, with exclusion criteria also being provided which, for example, allow a maximum number of partial classifications of a specific class.

In the overall assessment, it is particularly preferred, for example, that a maximum number of partial areas is defined, which is above and/or below a range of the classification defined as acceptable, ie outside the classes defined as acceptable, the local distribution of sub-areas with insufficient classes preferably also being used, for example to exclude undesired large clusters of sub-areas with insufficient classes.

This makes it possible, for example, to separate out those rubber products that already have a completely inadequate surface texture in a partial area, even if this only exists in a single partial area. A method according to the invention is therefore preferred, in which the surface texture of the machined surface is classified in method step c) by separate partial classification of a large number of partial areas of the surface, with each partial area being assigned a classification, with an overall evaluation of the surface preferably being carried out as a function of all partial classifications .

In order to link the method according to the invention as efficiently as possible to the method which is already used in the prior art, it is particularly advantageous to use a comparable classification. In the prior art, this classification is mostly done using six predefined surface textures, which are assigned to one of six classes of the BT classification published by the TRMG, so that it makes sense to retain this classification. In the simplest embodiment, however, it would also be conceivable to divide the surface structure into just two classes, for example into the classes “sufficient” and “insufficient”. A method according to the invention is therefore preferred in which the classification is carried out by assigning the surface structure to one of two or more, preferably three or more, particularly preferably four or more, very particularly preferably six classes. As explained above, the decisive criterion for the classification is basically the surface texture. Even if it is possible in principle to assume certain ordered structures and their presence on the surface, it is preferred in most cases to correlate the classification with the roughness of the surface. In this case, the surface coverage with removed material, ie the degree of contamination and/or the presence of macroscopic surface defects, for example cuts and cracks in the surface, can also be taken into account. Based on a classification that is favorable in relation to the roughness of the surface, excessive surface coverage and/or the presence of macroscopic surface defects, for example so-called “shiny spots”, can lead to a less favorable classification, for example. A method according to the invention is therefore preferred, in which the classification correlates with the roughness of the surface, the classification preferably also taking place as a function of the surface coverage with removed material and/or the number of macroscopic surface defects.

In accordance with the above statements, the simplest form of classification can probably be seen in the fact that the images recorded by the image recording device are compared with the images that were recorded on predefined reference surfaces with a known classification. In the simplest configurations, global parameters of the recordings can be used here, such as the mean brightness on the recordings obtained. A method according to the invention is therefore initially preferred, in which the classification in method step c) is carried out by comparing a recording of the surface of the rubber product taken by means of the image recording device with two or more recordings are made, which were previously recorded by the image recording device on predefined reference surfaces, with the comparison being carried out by the electronic data processing device.

Although the very simple procedure described above is basically suitable for rough classification, the inventors believe that the result obtained will support conventional manual classification rather than replace it, since the accuracy and reliability are probably not sufficient for an assessment alone, in particular since this method places relatively high demands on the comparability of the environmental conditions (e.g. with regard to the lighting conditions or the surface contamination). In addition, the information density obtained from such image comparisons is in many cases relatively low, so that it is more of a decision-making aid for process-leading operators who are thereby delimbed in their work, with the method also enabling automated documentation of the results.

The inventors of the present invention have now recognized that modern data processing technologies can be used particularly efficiently in the method according to the invention, through which the classification in the arrangement described above can be automated, fast and particularly reliable, with in particular a particularly high resolution the classification becomes possible, d. H. assignment to a large number of only slightly different classes is also possible.

In this particularly preferred embodiment, a neural network is used to evaluate and classify the recorded images. For this purpose, the neural network is trained to to assign the shadows cast detected by the image recording device to the various classes. This can expediently be done by training the neural network on reference surfaces with a defined surface texture, ie on reference surfaces with a known class. It has proven to be particularly advantageous if the data used for training were previously recorded using the same image recording device, ie in the same arrangement that is also used for the method according to the invention. It has proven to be particularly advantageous to use a "convolutional neural network" (abbreviated "CNN"), which is preferably a CNN with several "convolutional layers", so that it is a "deep convolutional neural network". acts.

The person skilled in the field of tire production can buy corresponding artificial networks, for example in a trained or untrained state, which are suitable for the purposes of the present invention, from specialized software companies or design them himself in a team together with a specialist in the field of IT. The development, optimization and training of corresponding neural networks is easily possible for specialists in the field of IT with access to rubber products with processed surfaces, such as can be obtained in the method according to the invention, and known classification, so that the specialist in the field of tire production from these experts can obtain the corresponding neural networks, for example in a trained or untrained state.

Against this background, a method according to the invention is preferred, wherein the classification in method step c) takes place by evaluating a recording of the surface of the rubber product taken by means of the image recording device, the evaluation being carried out by the electronic data processing device uses an artificial neural network, in particular a covolutional neural network, which is preferably trained using recordings that were previously recorded using the image recording device on defined reference surfaces and/or was trained using recordings that were provided and provided from a storage unit previously recorded on defined reference surfaces using an identical classification unit.

It can be seen as a great advantage of the method according to the invention that the basic structure also allows an alternative form of classification without making fundamental changes to the method according to the invention. This alternative procedure, which can be combined with the use of neural networks if necessary, is particularly suitable for applications in which the classification method described above does not provide satisfactory results or more comprehensive and detailed information about the surface texture is required. In this method, statistical characteristic height values are determined, at least for parts of the examined surface, which can be typical roughness parameters, for example.

These parameters can advantageously be obtained in the method according to the invention by arranging two or more lighting devices offset relative to one another. In this way, it is advantageously possible, with relatively minor changes in the structure, to determine the surface quality using what is known as “photometric stereo”. In this case, the three-dimensional surface topography is calculated by means of what is known as “shape from shading” using the electronic data processing device. The technology underlying "photometric stereo" and "shape from shading" has been around since the 1980s known in the last century. The person skilled in the art can therefore use software or algorithms that have been designed by third parties for this technology and have been published in specialist articles, for example. The statistical height parameters can be determined from the three-dimensional surface topography and used for classification.

A method according to the invention is therefore also preferred, in which the classification in method step c) is carried out by comparing statistical height characteristics, preferably the two-dimensional roughness of the surface, in particular the mean surface roughness R _a , and/or an area-related roughness, with predetermined limit values, the predetermined limit values are preferably set by means of defined reference surfaces.

In this respect, a method according to the invention is also preferred, in which the classification unit comprises two or more, preferably three or more, lighting devices which are arranged offset relative to one another and/or relative to the image recording device and illuminate the surface of the rubber product from different directions, the electronic data processing device being used for this purpose is set up to determine a three-dimensional surface topography from the known arrangement of the lighting devices and the recording of the image recording device.

In a combination of the classification methods described, the three-dimensional surface topography obtained by means of "photometric stereo" or even just an intermediate product obtained in the course of the calculation, in particular a calculated image of the reflection behavior over the surface under consideration (so-called "albedo"), can be analyzed by an artificial neuronal network, preferably as above described, evaluated, whereby the training of the neural network should take place accordingly on three-dimensional surface topographies or corresponding intermediate products, which were recorded using "photometric stereo" on defined reference surfaces.

In order to provide a method according to the invention, in which there is a particularly high degree of flexibility with regard to the available information and the technology to be used, it is preferable to provide two or more lighting devices so that "photometric stereo" can be used, but the lighting devices can be designed to be controlled separately from one another , so that if necessary only one lighting device can be used, the shadow cast is evaluated with the artificial neural network as described above.

In order to further increase the information density when using the technology described above for "photometric stereo" and in particular to further optimize the applicability of the method according to the invention in continuous operation for "photometric stereo", the inventors of the present invention propose lighting devices with different spectral use characteristics. By using suitable filters, for example bandpass filters, different lighting situations can thus be created efficiently and quasi-simultaneously, which can be evaluated separately from one another by the image recording device. A method according to the invention is therefore preferred, in which the classification unit comprises two or more lighting devices whose light has the maximum intensity at different wavelengths, the wavelength difference preferably being more than 50 nm, particularly preferably more than 100 nm, very particularly preferably more than 200 nm and/or wherein the lighting devices and/or the image recording device, preferably the image recording device, comprises a filter, preferably a bandpass filter, particularly preferably a multi-bandpass filter, in particular a multi-bandpass filter with two or three band ranges, the band ranges preferably being separated by more than 50 nm, more preferably more than 100 nm, most preferably more than 200 nm.

It has proven expedient to also provide a storage unit in the classification unit, whereby the storage unit can advantageously not only archive the image information recorded during operation, but can also include the neural network used or the software used to evaluate the "photometric stereo". A method according to the invention is preferred, wherein the classification unit comprises a memory unit connected to the electronic data processing device, wherein the memory unit preferably comprises a computer program product disclosed below.

It has proven to be particularly advantageous to also provide means with which the information obtained can be passed on directly to the workers involved in the process. This can be, for example, signaling devices with which a warning signal is output if the classification of the surface texture, ie the class assigned to the class, is outside a predetermined class range, ie if the classification is not sufficient for the products in a later product to process. This can be a warning siren or a signal light, for example. Accordingly, a method according to the invention is preferred, wherein the classification unit is a display and/or Signaling device, in particular a signaling light or a signaling column, which is set up to emit a display or a signal when the classification of the surface texture of the machined surface is within and/or outside of a predetermined class range.

However, it is particularly preferred for the support of the workers deployed if a visual representation of the classification of the surface or the partial classification of the partial areas on the surface is shown on a display device with a display. In this way, for example, a type of "heat map" can be obtained by suitable coloring, i. H. a representation of the spatial resolution of the classifications obtained or the assigned classes. This means, for example, that the final decision as to whether a classified overall surface should be discarded can be left to the workers who make the decision based on the information displayed. A method according to the invention is therefore preferred, in which the display device comprises a display, in which case the display comprises a visual representation of the classification of the surface or the partial classification of partial regions of the surface.

For all of the embodiments described above, it is particularly preferred if the image recording device is a line camera. Appropriate cameras are available on the market and, in the method according to the invention, advantageously allow a particularly efficient evaluation of the shadow cast along the recorded line, for example by combining several of these line recordings into one recording. Line scan cameras are particularly suitable for the components that are moved relative to the image recording device or their surfaces, which are to be classified during the retreading of tires. As a special advantage of the line camera can It can be considered that this is particularly suitable for use in a continuous process and, with appropriate dimensioning, allows a uniform analysis of most rubber products over the entire width, with the comparatively narrow recording area advantageously being able to ensure particularly uniform conditions for each recording , whereby, for example, the influence of any lens distortion can be minimized. In addition, line scan cameras can regularly produce recordings with a very high frequency, which advantageously allows the process to be carried out quickly. A method according to the invention is preferred in which the image recording device is a line camera, the image of the surface of the rubber product recorded with the line camera preferably having lighter and darker areas as a result of the oblique illumination of the rough surface and the resulting shadows cast. Within the scope of the present invention, such regular cameras are also regarded as line cameras which are used in the sense of a line camera by setting or limiting the evaluation to a low number of pixel lines, for example three pixel lines. The image recording device is preferably controlled in such a way, for example via an incremental encoder, that the generated recordings of the surface are at a predefined, preferably equidistant, distance from one another. Alternatively, the image recording device is controlled in such a way that the images of the surface are recorded at predefined time intervals.

The person skilled in the art selects suitable lighting devices for the method according to the invention depending on the actual structure. However, directional lighting devices are particularly preferred, ie lighting devices that emit the light in comparatively focused beams. In the opinion of the inventors, spotlights in particular are particularly suitable for casting suitable shadows on the surface of the rubber products. It is particularly preferred that all lighting devices illuminate the area of the surface of the rubber product captured by the image recording device. A method according to the invention is therefore preferred, in which the one or more lighting devices are directed lighting devices, preferably headlights, particularly preferably spotlights, and/or in which the one or more image recording devices are cameras. A method according to the invention is also preferred, in which the one or more lighting devices in the classification unit are arranged in such a way that the light beam(s) generated by the lighting devices impinge on the surface of the rubber product in the area covered by the image recording device.

In the combination of the above statements, a method according to the invention is particularly preferred, in which the image recording device is a line camera and/or in which the lighting device is a spotlight.

As explained above, the angle of incidence to be used according to the invention can advantageously be achieved in that the lighting device is arranged laterally offset from the rubber product. The inventors of the present invention have identified preferred angles of incidence, with angles of incidence of approximately 20° or 70° being particularly well suited for implementing the bright field or dark field effect in connection with the artificial neural networks. Particularly high angles of incidence, on the other hand, are particularly suitable for using the concept of the Implement "photometric stereo" in the process of the invention. A method according to the invention is therefore preferred, in which the one or more lighting devices in the classification unit are arranged to the side of the surface of the rubber product, so that the central beam path to the surface of the rubber product preferably has an angle of incidence of more than 45°, preferably more than 60°. more preferably of more than 75°, or alternatively an angle of incidence in the range from 15° to 25° or in the range from 65° to 75°.

In view of the cost efficiency, it is preferable to provide only one lighting device, especially for the implementation with the artificial neural network. In accordance with the understanding of those skilled in the art, this lighting device can also be composed of a plurality of lighting devices, for example separate LEDs, the beam paths of which then run essentially parallel to one another. A method according to the invention is therefore preferred in which the classification unit comprises precisely one lighting device or a plurality of lighting devices whose central beam path runs parallel to one another.

With a view to the quality of the compounds to be obtained in the retreading of tires, the inventors of the present invention have identified particularly suitable surface roughnesses on which the classification should be based. A method according to the invention is preferred in which the machined surface has an average roughness Ra in the range from 10 to 70 μm, preferably in the range from 15 to 45 μm, particularly preferably in the range from 20 to 30 μm.

It can be seen as a particular advantage of the method according to the invention that it can be operated as a continuous method, resulting in particularly high cost and time efficiency is reached. Advantageously, comparatively high belt speeds can be achieved in this way. Namely, a process according to the invention is preferred in which the process is operated as a continuous process, the relative speed of the rubber product to the tool and/or to the classification unit being in the range from 5 to 30 m/min, preferably in the range from 10 to 25 m/min .

Even if the present method, as described above, can be used for a large number of rubber components, it has proven to be particularly advantageous in particular in the classification of tire components, in particular tread strips or refurbished used tires. A method according to the invention is therefore preferred, in which the rubber product is a tire component, preferably a tread strip or a used tire with an at least partially worn, old tread strip, particularly preferably a tread strip.

In particular when only one camera is used, which is preferred in view of the acquisition costs and the required computing capacity, it has proven to be advantageous not to select too broad a range observed in the classification process. Thus, it is advantageous if the width of the surface of the processed rubber product is in a similar size range. More precisely, a method according to the invention is preferred, wherein the width of the processed surface of the rubber product is in the range from 100 to 500 mm, preferably in the range from 150 to 450 mm, particularly preferably in the range from 180 to 400 mm.

In a particularly preferred embodiment of the method according to the invention, the rubber product to be processed, for example a tread strip, is provided as a strand of material which is only cut to the required length in the method. One is therefore preferred Method according to the invention, wherein the rubber product in method step a) is provided from a roll as a strand of material from which individual rubber products are separated in the method by a separating tool.

When using the method according to the invention with prepared treads, such as those used for cold retreading, the method according to the invention should be carried out on the side facing away from the profile, since the mostly very pronounced profile information can lead to a strong shadow cast, which reduces the accuracy of the Classification method could alienate in terms of surface texture. Accordingly, a method according to the invention is preferred, wherein the rubber product is a tread strip, wherein the tread strip has a tire tread on one side and wherein the processing and the grading take place exclusively on the surface opposite the tire tread.

Although the rubber product may advantageously also comprise an only partially vulcanized rubber compound, in practice the most relevant case for most applications is when the rubber compound is substantially fully vulcanized. Analogously, the rubber product can in principle also be a composite material which, for example, comprises reinforcing cords made of metal or plastic. For most applications in the retreading of tires, however, it is advantageous if the rubber product essentially consists entirely of the at least partially vulcanized rubber mixture or of a plurality of at least partially vulcanized rubber mixtures. A method according to the invention is therefore preferred in which the rubber product comprises a completely vulcanized rubber mixture and/or in which the rubber product consists of at least partially, preferably completely, vulcanized rubber mixtures. It can be seen as an advantage of the method according to the invention that, according to the inventors, the classification of the surface texture essentially works on all rubber products. However, the inventors were able to identify a particularly suitable vulcanized rubber mixture whose surface can be classified particularly efficiently using the method according to the invention. Namely, a method according to the invention is preferred in which the vulcanized rubber mixture was obtained by vulcanization from a rubber mixture which comprises a filler and at least one rubber selected from the group consisting of styrene-butadiene rubber, butadiene rubber and isoprene rubber.

In addition, the inventors of the present invention were able to identify material-removing methods with which the processing in method step b) can be carried out particularly efficiently, with wire brushes in particular allowing comparatively inexpensive processing of the materials and adequate processing of the surface of rubber products. A method according to the invention is preferred, in which the processing of the surface in method step b) takes place with one or more tools for material-removing processing, which are selected from the group consisting of lasers, milling tools and brushes, preferably selected from the group consisting of lasers and Wire brushes, in particular wire brushes, the wire brushes preferably being designed as segmented wire brushes.

In this respect, the inventors propose that several of these tools can be arranged one behind the other. A method according to the invention is therefore preferred, in which the processing of the surface in method step b) takes place using a large number of tools arranged one behind the other. For a particularly efficient procedure in the method according to the invention, the inventors propose moving the rubber products to be processed by means of a conveyor along different processing stations, for example material removal and classification. In the case of refurbished scrap tires, this can also involve a rotatable suspension with which the surface of the scrap tire is moved relative to the tools and the classification unit. Preference is therefore given to a method according to the invention, in which the rubber product is conveyed in the method on a conveyor device or is rotated about its own axis in a device, or in which the one or more tools and the classification unit are moved relative to the rubber product, preferably by a robotic arm .

Even if it would be possible in principle to have the processing in method step b) carried out manually by workers, it is preferred if this is controlled by a control device .

A process according to the invention is also preferred, in which process steps b) and c) are each carried out on at least two surfaces of the rubber product, preferably on both surfaces simultaneously.

With regard to the quality of the surfaces obtained and to improve the classification quality in the method according to the invention, it is advantageous if the surface is also cleaned in step b), which preferably takes place after the material-removing processing. Here, the inventors were particularly suitable cleaning devices for removing vulcanized Identify rubber compound from a rubber product. Against this background, a method according to the invention is preferred, additionally comprising the method step after method step b): b2) removing the removed material from the processed surface of the rubber product with a cleaning device.

In this respect, a method according to the invention is preferred, in which the cleaning device comprises a rotating brush and/or a blower.

Those skilled in the art, in light of the foregoing, will clearly understand that the continued use of a rubber product with a graded surface texture is dependent on how graded it was. Those skilled in the art will determine a predetermined class range depending on the surface texture requirements of subsequent applications. This means that the person skilled in the art selects one or more of the classes he uses which he considers sufficient. Such rubber products, which are within these predetermined classes, ie the predetermined class range, are transferred to a downstream processing method, which can be, for example, a retreading process. Those rubber products that do not fall within the predetermined class range are discarded. Depending on how the classification deviates from the desired classification, discarded rubber products can be re-entered into the process according to the invention, particularly if the overall surface roughness was assessed as insufficient, ie not sufficiently pronounced, and an improvement can be expected from repeated processing. Preference is therefore given to a method according to the invention, additionally comprising, after method step c), one of the method steps: c2) transferring the rubber product to a downstream processing method, in particular to a tire retreading method, if the surface texture classification of the treated surface is within a predetermined class range, or c3) discarding the rubber product if the surface texture classification of the treated surface is outside of a predetermined class range.

A very particular advantage of the method according to the invention can be seen in the fact that digital information about the classification of the surface texture and thus about the surface condition is automatically obtained by means of the classification unit. This makes it possible, by observing the development of the classification over time, to recognize early on when the classification of the surface of the observed rubber products of the same type is moving in the direction of the border areas of the previously defined class range. Corresponding tendencies can be identified particularly early, particularly when using a neural network, as explained above. Such tendencies to leave the predetermined class range can be caused, for example, by the deterioration or wear of the tools used or a change in the material composition in the rubber product. It can be seen as a particularly great advantage of the method according to the invention that the information obtained by the classification unit can be used to automatically adapt the operating parameters of the tools used for surface processing in order to counteract a deterioration in the surface texture at an early stage. For example For example, the pressing force and/or the rotational speed of wire brushes used for roughening may be changed when the physicochemical properties of the processed rubber product, such as surface hardness, change depending on a changing composition of the rubber product. Preference is therefore given to a method according to the invention, additionally comprising the method step after method step c): d) adjustment of one or more operating parameters of the tool or tools used in method step b) for processing the surface of the rubber product if the classification of the surface texture of the processed surface is outside a predetermined one class range, with the adaptation preferably being carried out automatically by the classification unit, in particular the electronic data processing device of the classification unit, with two or more class ranges preferably being predetermined in order to be able to detect early on with a narrower class range a departure from the broader class range and to be able to counteract this.

The invention also relates to a method linked to the method according to the invention for retreading a tire, comprising the method steps: e) producing or providing a used tire with an at least partially worn, old tread strip and a roughened carcass surface, and either the method step: f) attaching a tread with a roughened surface to the roughened carcass surface by means of an adhesive by creating a material connection, or g) applying a vulcanizable rubber mixture to the roughened carcass surface and vulcanizing the compound obtained in a vulcanizing device to form a tread profile, the carcass surface, and optionally the roughened surface of the tread strip, is classified with the method according to the invention for classifying the surface texture of rubber products and the classification of the surface texture lies in a predetermined class range.

The method described above is particularly advantageous because with it it is possible to obtain particularly reliable retreaded tires in which the risk of the occurrence of a material defect, which could be caused by an unsuitable roughness of the individual components, is advantageously minimized.

It can be seen as a particular advantage of the method according to the invention that the tread strip can already be removed from the old tire in the method according to the invention, namely by the material-removing tools. A method according to the invention for retreading a tire is therefore also preferred, the old tire being produced by mechanically separating part of the old tread from the carcass of the tire, in particular by grinding off the old tread.

It has proven to be particularly advantageous when assembling the components, the carcass surface or the machined surface of the tread, as this leads to a particularly strong bond and thus to particularly durable and error-free retreaded tires. A method according to the invention for retreading a tire is preferred, the carcass surface and/or the treated surface of the tread strip being heated to a temperature of more than 50° C., preferably more than 60° C., particularly preferably more than 70° C., before attachment is tempered.

In this respect, the inventors of the present invention have recognized that the method according to the invention is not limited in terms of the adhesive used, with the inventors proposing a particularly proven adhesive. Namely, a method according to the invention for retreading a tire is preferred, wherein the adhesive comprises at least one styrene-butadiene rubber, wherein the adhesive preferably comprises a solvent which evaporates when the prepared tread strip is attached to the roughened carcass surface.

The invention also relates to a device for classifying the surface texture of rubber products, preferably in a method according to the invention, comprising: v) one or more tools for material-removing processing of the surface of the rubber product, w) a movement unit which is set up to move the surface of the rubber product and move the one or more tools relative to each other, and x) a classification unit comprising: i) at least one lighting device that is set up to illuminate the surface of the rubber product with an angle of incidence of more than 10°, ii) at least one image recording device that is set up to capture the surface of the rubber product in the area illuminated by the lighting device , and iii) at least one electronic data processing device connected to the image recording device.

Here, the movement unit can move the rubber product relative to a stationary tool, for example by using a conveyor belt or a rotating suspension, as is also described above, for example. Alternatively, however, the tools can also be designed to be movable and, for example, guided by a robotic arm along a stationary rubber product. In an analogous manner, the rubber product can also be moved relative to the classification unit, it also being possible for the classification unit to be moved relative to a stationary rubber product, analogously to the tools. It can be seen as an advantage of the present invention that the person skilled in the art is free to design the device in this regard. Corresponding devices according to the invention are particularly preferred because the method according to the invention can be carried out with them and the advantages described above result accordingly. Analogously, devices according to the invention are particularly preferred with which the method according to the invention can be carried out in preferred embodiments.

First of all, a device according to the invention is preferred, wherein the movement unit is a conveyor device for conveying a rubber product through the device or a device for rotating a rubber product, preferably a conveyor device for conveying a rubber product through the device.

A device according to the invention is also preferred, the device comprising a receiving device for a roller, via which the rubber product can be made available as a strand of material.

A device according to the invention is also preferred, the device comprising separating tools for separating individual rubber products from a strand of material.

Preference is also given to a device according to the invention, the device comprising a plurality of tools for material-removing processing of the surface of the rubber product, which are preferably selected from the group consisting of lasers, milling tools and brushes, preferably selected from the group consisting of lasers and wire brushes, in particular wire brushes , The wire brushes are preferably designed as segmented wire brushes.

Furthermore, a device according to the invention is preferred, wherein the device comprises a plurality of tools for material-removing machining, which are arranged along the conveyor device.

A device according to the invention is also preferred, the device additionally comprising a control device for the automatic control of the tools for material-removing machining.

In this respect, a device according to the invention is also preferred, the device being set up so that the material-removing processing and the subsequent classification can take place simultaneously on two or more surfaces of the rubber product. A device according to the invention is also preferred, wherein the device comprises a cleaning device for removing removed material from the processed surface of the rubber product, preferably a rotating brush and/or a blower.

A device according to the invention is also preferred, wherein the classification unit comprises a memory unit connected to the electronic data processing device.

A device according to the invention is also preferred, wherein the classification unit comprises a display and/or signal device connected to the electronic data processing device.

A device according to the invention is also preferred, wherein the classification unit, in particular the electronic data processing device, is set up to carry out method steps c) and/or d) of the method according to the invention.

In principle, a device according to the invention is also preferred, in which the one or more lighting devices are directed headlights, preferably spotlights, and/or in which the one or more image recording devices are cameras.

A device according to the invention is also preferred, wherein the one or more lighting devices in the classification unit are arranged in such a way that the light beam(s) generated by the lighting devices impinge on the movement unit in the area covered by the image recording device.

A device according to the invention is also preferred, wherein the one or more lighting devices are in the classification unit are arranged to the side of the movement unit, so that the central beam path to the surface of a rubber product guided in the movement unit has an angle of incidence of more than 45°, preferably more than 60°, particularly preferably more than 75°.

A device according to the invention is also preferred, in which case the classification unit comprises exactly one lighting device or a plurality of lighting devices whose central beam path runs parallel to one another.

A device according to the invention is also preferred, wherein the image recording device is a line camera.

A device according to the invention is also preferred, wherein the classification unit comprises two or more lighting devices which are arranged offset relative to one another and/or relative to the image recording device and illuminate the rubber product to be classified from different directions, the electronic data processing device being set up to do so, from the known arrangement of the lighting devices and the recording of the image recording device to determine a three-dimensional surface topography.

A device according to the invention is also preferred, wherein the classification unit comprises two or more lighting devices whose light has the maximum intensity at different wavelengths, the wavelength difference preferably being more than 50 nm, particularly preferably more than 100 nm, very particularly preferably more than 200 nm, and / or wherein the lighting devices and / or the image recording device, preferably the image recording device, a Filters, preferably a bandpass filter, particularly preferably a multi-bandpass filter, in particular a multi-bandpass filter with two or three band ranges, the band ranges preferably having a distance of more than 50 nm, particularly preferably more than 100 nm, most preferably of more than 200 nm.

Also preferred is a device according to the invention, wherein the conveying device is set up to press a conveyed rubber product against the one or more tools for material-removing processing.

A device according to the invention is also preferred, the device additionally comprising a height gauge and/or thickness gauge for the spatially resolved determination of the thickness of a rubber product, in particular a light section sensor.

Finally, a device according to the invention is also preferred, wherein the classification unit has rollers with which the classification unit can be moved at a predetermined distance over the surface of the rubber product, for example by means of a robot arm.

Also disclosed is a computer program product, comprising instructions which, when the program is executed by the electronic data processing device of a classification unit, cause the latter to carry out method steps c) and/or d) of the method according to the invention, preferably comprising an artificial neural network, in particular a covolutional neural network , which was particularly preferably trained on the basis of recordings that were previously recorded by means of the image recording device of a device according to the invention on defined reference surfaces of rubber products with known classes of the surface. Finally, a control device is disclosed for controlling a device for classifying the surface texture of rubber products, preferably a device according to the invention, comprising a memory unit on which the disclosed computer program product is stored, which is executed by the control device.

Preferred embodiments of the invention are explained and described in more detail below with reference to the accompanying drawings. It shows:

1 shows a schematic representation of part of an exemplary device according to the invention when carrying out the method according to the invention in a preferred embodiment on a tread; and

2 shows a schematic representation of a part of an exemplary device according to the invention when carrying out the method according to the invention in a preferred embodiment on an at least partially of a tread.

1 shows a schematic representation of part of an exemplary device 20 according to the invention when carrying out the method according to the invention in a preferred embodiment. 1 shows the embodiment of the method according to the invention using an artificial neural network in order to carry out a classification from the course of the shadow on the recording.

Shown is a device 20 for classifying the surface texture of a rubber product 10, which is a tread strip in the present case. The device 20 comprises a movement unit 22 which is set up to move the surface 12 of the rubber product 10 to move. In the illustrated embodiment of the device 20, this movement takes place relative to the tools for material-removing processing (not shown here) and relative to the classification unit 14.

In FIG. 1, the classification unit 14 comprises an illumination device 16a which illuminates the surface 12 of the rubber product 10 at an angle of incidence of approximately 45° with respect to the central beam path. Also shown is the image recording device 18 which, as part of the classification unit 14, is set up to record the surface 12 of the rubber product 10 at the location, i. H. to record, which is illuminated by the lighting device 16a. The image recording device 18 is connected to an electronic data processing device, which is not shown in FIG. The surface 12 of the tread, which was previously roughened by machining with tools, runs on the movement unit 22, which is designed here as a conveyor belt, under the classification unit 14.

Due to its roughness, the oblique arrangement of the lighting device 16a causes a shadow progression on the surface 12 which correlates with the surface texture and can be classified by an artificial neural network and the electronic data processing device. In the example shown, the classification is in six classes and essentially correlates with the roughness of the surface 12. In the evaluation by the neural network, there is a separate classification of sub-areas that are tiled over the surface 12 of the tread. In this case, a convolutional neural network is used, which is previously trained using recordings that were recorded with the same arrangement of image recording devices 18 and lighting devices 16a on defined reference surfaces with a known class. The electronic In this case, the data processing device comprises a memory unit on which the artificial neural network is stored, which is part of a computer program product which carries out the classification in the method according to the invention.

In the example shown, the electronic data processing device has a display device, via which the classification of the surface texture of the classified partial areas is visualized with a color code. The image recording device 18 is a line camera, whereas the lighting device is designed as a spotlight.

The device 20 shown is suitable for use in a continuous process and can be operated, for example, with a relative speed of the rubber product 10 to the classification unit 14 of 20 m/min. In the exemplary method, the processed tread is provided as a strand of material and only separated into a tread by a separating tool immediately before passing through the classification unit 14 . A tire tread is provided on the rear side of the tread, not shown.

The tread consists entirely of an essentially fully vulcanized rubber mixture. In the example, this rubber mixture comprises amorphous silicon dioxide as a filler and styrene-butadiene rubber. The surface is processed with a plurality of wire brushes arranged one behind the other and is controlled by a control device which is set up to control the operating parameters of the tools used in method step b) depending on the feedback from the classification unit 14 . After the surface 12 of the tread has been processed and before the tread is introduced into the classification unit 14, removed material is removed from the surface 12 by means of a further brush removed. The further use of the tread strip after passing through the classification unit 14 depends on the classification assigned to the surface. In the present example, the tread strip is only used to produce a retreaded tire if the average classification is in class 2 of 6 possible classes and no sub-class has a classification of 3 or more. Alternatively, however, other criteria can also be used for the overall assessment of the treads, as explained above.

On the other hand, treads that are considered insufficient as a result of the classification are discarded. If the classification unit 14 detects that the development of the classification is moving in the direction of the border areas of the predetermined class range, the operating parameters of the previously used tools are automatically adjusted, whereby in the illustrated embodiment the operating parameters are also adjusted by an artificial neural network, which trained for this purpose.

FIG. 2 shows a structure comparable to FIG. In the device 20 of FIG. 2, however, a total of three lighting devices 16a, 16b, 16c are provided, which are arranged offset relative to one another. In the embodiment shown, the three lighting devices 16a, 16b, 16c are arranged approximately as an isosceles triangle, in the center of which is the image recording device 18, which has approximately the same height relative to the moving unit 22, ie the conveyor belt, as the lighting devices 16a, 16b , 16c and itself is arranged centrally in the middle above the tread strip passed underneath. This structure makes it possible to illuminate the surface 12 of the rubber product 10 from different directions. The electronic data processing device is now set up to use the known arrangement of the lighting device 16a, 16b, 16c from the recording Image recording device 18 to determine a three-dimensional surface topography, from which statistical height characteristics can be determined, which can then be used to classify the surface or for partial classification of partial areas of the surface. In the illustrated embodiment, the three lighting devices 16a, 16b, 16c have different spectral characteristics and have the maximum intensity of the emission at different wavelengths. The image recording device 18 is equipped with suitable filters so that it can generate different lighting scenarios with each recording using appropriate filters.

Reference List

10 rubber product

12 surface

14 classification unit 16a-c lighting device

18 image pickup device

20 device

22 movement unit

Claims

44 claims

1 . Method for classifying the surface texture of rubber products (10), in particular of tire components in the retreading of tires, comprising the following method steps: a) producing or providing a rubber product (10) comprising an at least partially vulcanized rubber mixture, b) processing a surface (12 ) of the rubber product (10) by removing material with one or more tools, and c) classifying the surface texture of the machined surface (12) with a classification unit (14), the classification unit (14) comprising: i) at least one lighting device (16a , 16b, 16c) for illuminating the surface (12) of the rubber product (10) with an angle of incidence of more than 10°, ii) at least one image recording device (18) for capturing the surface of the rubber product (10) in the area illuminated by the lighting device (16a , 16b, 16c) illuminated area, and iii) at least one electronic data processing device connected to the image recording device (18).

2. The method according to claim 1, wherein the classification of the surface texture of the machined surface (12) in method step c) takes place by separate partial classification of a plurality of partial areas of the surface (12), each partial area being classified 45 is assigned, with an overall evaluation of the surface (12) preferably being carried out as a function of all partial classifications.

3. The method according to any one of claims 1 or 2, wherein the classification is carried out by assigning the surface structure to one of two or more, preferably three or more, particularly preferably four or more, very particularly preferably six classes.

4. The method according to any one of claims 1 to 3, wherein the image recording device (18) is a line camera and / or wherein the lighting device (16a, 16b, 16c) is a spotlight.

5. The method according to any one of claims 1 to 4, wherein the classification in method step c) is carried out by evaluating a recording of the surface (12) of the rubber product taken by the image recording device (18), the evaluation by the electronic data processing device using an artificial neural Network, in particular a covolutional neural network, which has preferably been trained using recordings that were previously recorded using the image recording device (18) on defined reference surfaces.

6. The method according to any one of claims 1 to 4, wherein the classification in method step c) by comparing statistical height parameters, preferably the two-dimensional roughness of the surface (12), in particular the average surface roughness R _a , 46 and/or an area-related roughness, with predetermined limit values, the predetermined limit values preferably being established by means of defined reference surfaces.

7. The method according to claim 6, wherein the classification unit (14) comprises two or more lighting devices (16a, 16b, 16c) which are arranged offset relative to each other and illuminate the surface (12) of the rubber product (10) from different directions, the electronic data processing device is set up to determine a three-dimensional surface topography from the known arrangement of the lighting devices (16a, 16b, 16c) and the recording of the image recording device (18).

8. The method according to any one of claims 1 to 7, additionally comprising, after method step c), the method step: d) adjusting one or more operating parameters of the tool or tools used in method step b) for processing the surface (12) of the rubber product (10), if the classification of the surface texture of the machined surface (12) lies outside a predetermined class range, with the adaptation preferably being carried out automatically by the classification unit (14), in particular the electronic data processing device of the classification unit (14).

9. A method for retreading a tire, comprising the steps of: e) producing or providing a used tire with an at least partially worn, old tread and a roughened carcass surface, and either the method step: f) attaching a tread with a roughened surface to the roughened carcass surface using an adhesive by creating a material connection, or g) orders a vulcanizable rubber mixture on the roughened carcass surface and vulcanizing the compound obtained in a vulcanizing device for forming a tread profile, the carcass surface and optionally the roughened surface (12) of the tread with the method for classifying the surface texture of rubber products (10) according to one of Claims 1 to 8 is classified and the classification of the surface texture is in a predetermined class range.

10. Device (20) for classifying the surface texture of rubber products (10), preferably in a method according to any one of claims 1 to 8, comprising: v) one or more tools for material-removing processing of the surface (12) of the rubber product (10), w) a moving unit (22) arranged to move the surface (12) of the rubber product (10) and the one or more tools relative to each other, and x) a classification unit (14) comprising: i) at least one lighting device (16a, 16b, 16c) adapted to illuminate the surface (12) of the rubber product (10) with an angle of incidence greater than 10°, ii) at least one image recording device (18) which is set up to record the surface (12) of the rubber product (10) in the area illuminated by the lighting device (16a, 16b, 16c), and iii) at least one with the image recording device (18) connected electronic data processing device.