EP4665651A1 - Label feeding system provided with a recovery mode and a validation mode of the sensor device - Google Patents

Abstract

There is described a label feeding system (15) for a labelling machine (1) comprising: a sensor device including an optical sensor (17) having a fixed field of view (FV); a feed roller (11) for longitudinally feeding a web (4) of labelling material along a feed path (P) and through said field of view (FV), the web (4) comprising a plurality of longitudinally joined labels (3), each label (3) having a plurality of longitudinal portions and an aesthetic feature (AF) arranged at a determined longitudinal portion; and a reference member (16) fixed with respect to the feed path (P) and to the field of view (FV) and arranged within the field of view (FV); the sensor device is configured for operating in a production mode according to which, sequentially for each label (3), the optical sensor (17) acquires an image (I) of the field of view (FV) at said determined longitudinal portion for detecting a linear distance between the respective aesthetic feature (AF) and the reference member (16); the sensor device is further configured for operating in a recovery mode according to which the optical sensor (17) acquires a number of images (I1, I2, 13) of the field of view (FV) for searching the aesthetic feature (AF), wherein each image (I1; I2; I3) of said number of images includes at least one longitudinal portion which is different from the other longitudinal portions included in the other images (I2, I3; I1, I3; I1, I2) of said number of images, so as to inspect all the different longitudinal portions during the recovery mode, by means of a single label (3) or cumulatively by means of a plurality of labels (3a, 3b, 3c).

Description

LABEL FEEDING SYSTEM PROVIDED WITH A RECOVERY MODE AND A VALIDATION MODE OF THE SENSOR DEVICE TECHNICAL FIELD The present invention relates to a label feeding system for a labelling machine configured to label containers, such as bottles, jars, flacons or the like, adapted to contain a pourable product, preferably a pourable food product. BACKGROUND ART Labelling machines are known, which are commonly used to prepare, transport and apply labels, in particular labels obtained from a web of labelling material, onto containers, such as bottles, flacons, jars or the like, destined to be filled with a pourable product, in particular a pourable food product. Labelling machines of this sort typically operate with well-known glued labels, which are cut from the web at appropriate lengths, sprinkled with glue and then applied on the containers. In other embodiments, labelling machines can operate with so-called “sleeve-labels”, which are cut from a web of heat-shrinking film, are applied with a certain clearance on the containers, and then heated in an oven to obtain their shrinking and perfect adhesion to the lateral surfaces of the containers. Regardless of the type of label used, a known labelling machine typically comprises: - a carousel rotatable around a vertical axis and configured to convey a plurality of containers along a horizontal, arc-shaped labelling path; and - a labelling module, peripherally arranged relatively to the carousel (i.e. radially with respect to the aforementioned vertical axis) and configured to prepare, transport and feed a plurality of labels to the carousel at an application station, in order to apply such labels to the respective containers. Typically, the labelling module essentially comprises: - at least one support shaft around which the web of labelling material is initially wound in form of a reel; - a label feeding system for feeding the web of labelling material along a feeding path, usually including a feed roller for unwinding the web off the reel and advancing it along the feed path; - a cutting unit for repeatedly cut the web at a cutting station so as to separate a sequence of labels from the web itself; and - a label transfer device, for example a known vacuum drum arranged peripherally to the carousel, configured to receive, retain and advance each label previously cut and to feed each label to the carousel, at the application station. In case glued labels are used, the labelling module further comprises a gluing roller arranged peripherally to the vacuum drum, for sequentially sprinkling with glue each label retained and transferred by the drum itself. It is known in the field the need for: - constantly monitoring the positioning of the web of labelling material during the unwinding thereof from the respective reel; and - correcting such positioning, in case a deviation from a nominal positioning is detected. In fact, the positioning of the web along the feed path, for example the longitudinal positioning, during the unwinding thereof affects the correct nominal separation of each label from the web itself at the cutting station. More specifically, the web must pass through the cutting station with its cutting portions (i.e. the joined transversal edges between each pair of longitudinally adjacent labels) in the correct position, so as to ensure a nominal cut of the web itself, and subsequently a nominal application of the labels on the containers. To this end, a typical labelling module comprises a sensor system comprising an optical sensor and a control unit. The optical sensor is configured to sequentially detect a reference element on each label, such as an aesthetic pattern or motive or design printed on the label, prior to the cutting of the web In order to achieve the above, the optical sensor is configured to acquire images of a detection area, which corresponds to a field of view of the optical sensor, with a predetermined frequency. The predetermined frequency is related to the production speed. For instance, if five labels per second are supposed to transit through the field of view (i.e. the detection area), then the acquisition frequency will be set to 5Hz, namely five acquisitions per second. The Applicant has observed that the optical sensor can in some occasions “loose” the reference element. More specifically, the element which was originally implemented in the sensor system as a reference and which was originally correctly inside the detection area and the field of view, can be situated, for a certain number of consecutive labels, outside of such detection area or the field of view. For example, the above situation can occur in case the phase displacement of the web is too large. In this case, it is necessary to stop the labelling machine and to manually recalibrate and reinitialize the sensor system, so as to re-implement the reference element in the sensor system (i.e. so that the reference element is again within the field of view and/or within the reference region). The above operation is necessary also upon starting the labelling process or upon a label format change (in which case the manual implementation of a new reference element is necessary). In addition, once the reference element is chosen and implemented, it can occur that the optical sensor erroneously detects a feature on the label which is similar to the reference element, although not being the reference element. DISCLOSURE OF INVENTION It is therefore an object of the present invention to provide a label feeding system for a labelling machine, which is designed to overcome at least one of the above- mentioned drawbacks in a straightforward and low-cost manner. This object is achieved by a label feeding system as claimed in claim 1. BRIEF DESCRIPTION OF THE DRAWINGS Non-limiting embodiments of the present invention will be described by way of example with reference to the accompanying drawings, in which: Figure 1 is a schematic top view, with parts removed for clarity, of a labelling machine having a label feeding system according to the present invention, for feeding a web of labelling material through the labelling machine; Figure 2 is a schematic lateral view, with parts removed for clarity, of the web fed by the system of Figure 1 and of some parts of the system of Figure 1, during a normal operative condition; Figures 3a-3c are schematic lateral views, with parts removed for clarity, of the web and of some parts of the system of Figure 1, during three distinct operative conditions according to a recovery mode of operation of the system; Figures 4a-4c are schematic lateral views, with parts removed for clarity, of the web and of some parts of the system of Figure 1, during three distinct operative conditions of a recovery mode of operation of the system according to a second preferred embodiment of the invention; and Figures 5a-5c are schematic lateral views, with parts removed for clarity, of the web and of some parts of the system of Figure 1, during three distinct operative conditions of a validation mode of operation of the system. BEST MODE FOR CARRYING OUT THE INVENTION With reference to Figure 1, number 1 indicates as a whole a labelling machine for labelling containers 2, such as bottles, flacons, cans or containers of this sort, adapted to contain a pourable product, preferably a pourable food product. Labelling machine 1 is configured to apply labels 3 obtained from a web 4 of labelling material onto containers 2. According to this preferred and non-limiting embodiment, labels 3 are glued labels, i.e. strips of labelling material that are cut at predetermined lengths from web 4 and then sprinkled with glue before their application on the respective containers 2. Web 4 is provided in the form of a continuous strip initially wound in a reel 5 and is progressively unwound, in use, off reel 5, along a feed path P according to a known manner. As schematically shown in Figure 2, web 4 comprises a plurality of longitudinally joined labels 3. More specifically, labels 3 are joined together at respective transversal edges thereof, at which edges the web 4 is to be cut to obtain the separated labels 3. Each label 3 has a plurality of longitudinal portions, i.e. portions of labelling material arranged at specific and distinct longitudinal coordinates, relative to the longitudinal extension of the label along feed path P. Hence, each longitudinal portion has a longitudinal extension smaller than the total longitudinal length of the respective label 3. Each label 3 has an aesthetic feature AF arranged at a determined longitudinal portion, as schematically shown in Figure 2. Aesthetic feature AF defines a reference element on each label 3. Aesthetic feature AF can be defined by a pattern or an aesthetic motive or design printed on a surface of each label 3. As visible in Figure 1, labelling machine 1 comprises: - a conveyor device, preferably a carousel 6 rotatable around a central axis X, preferably vertical, and configured to advance a plurality of containers 2 along an arc-shaped labelling path, preferably horizontal; - a labelling module 7 (only schematically shown), arranged peripherally relatively to carousel 6 and configured to prepare a plurality of labels 3 and feed them to the carousel 6 at an application station A, for the application thereof onto respective containers 2. Labelling module 7 comprises: - a fixed frame 8; - a support shaft 10 for supporting one reel 5 at a time in an unwindable manner; - a feed roller 11 fixed in a rotatable manner to frame 8, actuatable in rotation for feeding web 4 along feed path P, which extends within and through labelling module 7, feed roller 11 being conveniently arranged downstream of support shaft 10 along feed path P; - a plurality of guide rollers 12 (only one shown), arranged operatively downstream of support shaft 10 and of feed roller 11, and configured to support the web 4 progressively unwound from reel 5 and to guide web 4 along feed path P; - a cutting unit 13 (known per se and not described in detail) arranged downstream of guide rollers 12, along feed path P, and configured to repeatedly cut web 4 at a cutting station C thereby separating a sequence of labels 3 therefrom; and - a transfer drum 14, preferably a vacuum drum of the known type, arranged downstream of cutting unit 13, mounted on frame 8 in a rotatable manner about a central axis which is preferably parallel to axis X, and configured to transfer labels 3 to application station A for the application thereof onto the respective containers 2. The feed roller 11 is supported by said frame 8. The feed roller 11 is configured for feeding the web 4 by rotating on itself and with respect to the frame 8. Since in this preferred non-limiting embodiment labels 3 require the use of glue, labelling module 7 further comprises at least one gluing roller 50 arranged substantially tangent to transfer drum 13 and configured for spreading glue on at least the extremities of each individual label 3, before the application thereof onto the containers 2. Feed roller 11 is of the known type and is rotatable about a central axis W preferably parallel to axis X. Feed roller 11 is part of a label feeding system 15 configured for feeding web 4 through labelling module 7, i.e. along feed path P. System 15 further comprises: - a sensor device including an optical sensor 17 which has a field of view FV, and which is configured for repeatedly inspecting field of view FV; - a control unit 18 operatively connected to optical sensor 17 and to feed roller 11; and - a reference member 16 fixed with respect to feed path P and to field of view FV and arranged within field of view FV so as to be detected by optical sensor 17. The frame 8 supports the sensor device, so that the field of view FV can be fixed with respect to the frame 8. The reference member 16 comprises at least part of a physical body, such as a mechanical rod or a bar, said physical body being supported by the frame 8 so that the physical body can be fixed with respect to the frame 8. In Figure 1, for clarity purposes, only a small part of the frame 8 is showed, which is the part of the frame 8 supporting the reference member 16. Feed roller 11 is configured to feed web 4 through field of view FV. Aesthetic feature AF is detectable by optical sensor 17, in particular when passing through field of view FV thereof. Preferably, optical sensor 17 is fixable, in particular fixed, to frame 8. Preferably, reference member 16 is fixable, in particular fixed, to frame 8. Hence, reference member 16 is fixed with respect to web 4 advancing along path P. Preferably, reference member 16 comprises, in particular is defined by, a mechanical rod or a bar which is mounted on frame 8 in a fixed position and which is operatively interposed between feed roller 11 and cutting unit 13, in particular between feed roller 11 and guide rollers 12. In detail, optical sensor 17 is arranged opposite to reference member 16 with respect to feed path P, and therefore of web 4 advancing along it, i.e. on the other side of feed path P (and of web 4 fed along it) with respect to reference member 16. Hence, the aforesaid inspection is performed by optical sensor 17 on labels 3 still joined together to form web 4. Preferably, optical sensor 17 comprises, in particular is defined by, a camera. The sensor device is configured for operating in a production mode (Figure 2) according to which sequentially for each label 3, optical sensor 17 acquires an image I of field of view FV at said determined longitudinal portion for detecting a linear distance between the respective aesthetic feature AF and reference member 16. Opportunely, optical sensor 17 is configured to acquire each image I when the determined longitudinal portion of each label 3 transits through field of view FV, sequentially for each label 3. More precisely, the sensor device is configured so that for all the images I acquired during the production mode the longitudinal position of each respective label 3 with respect to field of view FV is the same. In other words, during the production mode, all the images I are acquired when the same longitudinal portion, for each label, transits through field of view FV. For example, with reference to Figure 2, optical sensor 17 acquires each image I when the terminal longitudinal portion of each label 3 (at which aesthetic feature AF is arranged to, according to the example shown) transits through field of view FV. Preferably, the detected linear distance is a distance along the longitudinal direction, i.e. along path P. Linear distance can also be a transversal distance, relative to path P. Based on the detected linear distance between aesthetic feature AF and reference member 16, control unit 18 controls the rotational velocity of feed roller 11, so as to adjust the phase of web 4. According to an aspect of the present invention, the sensor device is further configured for operating in a recovery mode according to which optical sensor 17 acquires a number of images I1, I2, I3 of field of view FV for searching aesthetic feature AF, wherein each image I1, I2, I3 of said number of images includes at least one longitudinal portion which is different from the other longitudinal portions included in the other images I1, I2, I3 of said number of images, so as to inspect all the different longitudinal portions during the recovery mode, by means of a single label 3 or cumulatively by means of a plurality of labels 3a, 3b, 3c. More in detail, according to a first preferred embodiment schematically shown in Figures 4a, 4b, 4c, the sensor device is configured so that, during the recovery mode, optical sensor 17 sequentially acquires, for a single label 3 to be inspected, a number of images including at least a first image I1 and a second image I2. In the example shown, optical sensor 17 acquires a first image I1, a second image I2 and a third image I3 for a single label 3. According to such first preferred embodiment, in each image I1, I2, I3 of said number of images label 3 has a respective longitudinal position, relative to field of view FV, which is different from the longitudinal position of the same label 3 in all the other images I1, I2, I3 of said number of images. For example, in image I1 (Figure 4a) label 3 has a longitudinal position, i.e. it is at a longitudinal position, with respect to field of view FV, which is different from the longitudinal position of the same label 3 in image I2 (Figure 4b) and in image I3 (Figure 4c), and so on. In practice, the sensor device is configured so that optical sensor 17 acquires images I1, I2, I3 of all the longitudinal portions of a single label 3 in order to search for aesthetic feature AF, thereby defining the recovery mode. If aesthetic feature AF is not found in that particular label 3, the recovery mode is performed for the subsequent label 3. Such recovery mode is for example expedient when, for any cause, optical sensor 17 “looses” aesthetic feature AF, i.e. its reference element based on which control unit 18 monitors the unwinding of web 4 along path P. Advantageously, during the aforementioned recovery mode, control unit 18 is configured to reduce the rotational velocity of feed roller 11 so as to cause a deceleration of web 4 along feed path P and, therefore, a deceleration of each label 3 through field of view FV. In this way, optical sensor 17 is able to perform the acquisition of the different images I1, I2, I3 for all the longitudinal portions of label 3, without unduly increasing the frequency of acquisition. This ensures the effectiveness and the precision of the acquisition and, therefore, of the search of aesthetic feature AF. Furthermore, it is ensured that the upper limit frequency of acquisition of optical sensor 17 is not reached or overcome. Optical sensor 17 is configured to acquire images I of field of view FV with a first acquisition frequency during the production mode of the sensor device. “Acquisition frequency” means an acquisition time frequency, for example in Hz. Optical sensor 17 is configured to acquire images I1, I2, I3 of field of view FV with a second acquisition frequency during the recovery mode of the sensor device, wherein the second acquisition frequency is different from the first acquisition frequency. Conveniently, the second acquisition frequency is higher than the first acquisition frequency. In detail, control unit 18 is configured to command an increase in the frequency of acquisition of optical sensor 17 during the recovery mode with respect to the production mode. In this way, the rotational velocity of feed roller 11 does not have to be unduly reduced, so that the labelling process is not slowed too much during the recovery mode. Thanks to the label feeding system 15 according to the configuration described above, it is no longer necessary to stop labelling machine 1 and to manually recalibrate and reinitialize the sensor device in case optical sensor 17 looses aesthetic feature AF. In fact, the re-implementation of aesthetic feature AF in the sensor device can be performed automatically and on-the-fly (i.e. during production and without stopping the labelling process). The above configuration is also particularly advantageous upon starting the labelling process or upon a label format change (in which case the implementation of a new reference element is necessary). Furthermore, thanks to the above-described configuration, the total number of discarded labels 3 is reduced, since the search of aesthetic feature AF can be performed by means of a single label 3. As schematically shown in Figures 2 and 4a-4c, optical sensor 17 is configured for analyzing, for each image I, I1, I2, I3 acquired during the production mode and during the recovery mode, a preestablished reference region RR of field of view FV, to detect aesthetic feature AF. In this way the computational efforts are reduced, because the reference region corresponds to a smaller area with respect to the field of vision FV. The reference region RR is preferably set in the machine by an operator. Preferably, field of view FV has a longitudinal extension, relative and along feed path P, which is a fraction of the longitudinal length of each label 3. Conveniently, the minimum number of images I1, I2, I3 to be acquired during the recovery mode for searching aesthetic feature AF is at least equal to the ratio between the longitudinal length of a label 3 and the longitudinal extension of field of view FV. Therefore, the sensor device is configured so that the number of images I1, I2, I3 which are acquired during the recovery mode for searching aesthetic feature AF is greater or equal to the ratio between the longitudinal length of a label 3 and the longitudinal extension of field of view FV. More precisely, the minimum number of images I1, I2, I3 to be acquired during the recovery mode for searching aesthetic feature AF is at least equal to the first integer number following the numeric value of the aforementioned ratio. For example, as in the case depicted in Figures 4a- 4c wherein the longitudinal extension of field of view FV is 1/3 of the longitudinal length of each label 3, the minimum number of images to be acquired in order to inspect all the longitudinal portions is three. In this way, it is ensured that all the aforementioned longitudinal portions are inspected by optical sensor 17. Figures 3a, 3b and 3c schematically illustrate a second preferred embodiment of the sensor device according to the present invention. Being the sensor device according to the second embodiment similar to the sensor device according to the first embodiment, only the differences between these two will be described below. In particular, the sensor device according to the second embodiment is configured so that, during the recovery mode and for a group of labels 3 including at least a first label 3a and a second label 3b, the optical sensor is configured to acquire at least one single said image I1, I2, I3 for each label 3a, 3b, 3c of the group of labels 3. In the example shown, the group of labels 3 includes a first label 3a, a second label 3b and a third label 3c. Therefore, optical sensor 17 is configured to acquire a first image I1, a second image I2 and a third image I3. More specifically, optical sensor 17 acquires a single first image I1 for the first label 3a, a single second image I2 for the second label 3b and a single third image I3 for the third label 3c. According to such second preferred embodiment of the sensor device according to the invention, the sensor device is configured so that in each acquired image I1, I2, I3 the longitudinal position of the respective label 3a, 3b, 3c of the group of labels 3, relative to field of view FV, is different from the longitudinal position of the other labels 3a, 3b, 3c of the group of labels 3 in the respective images I1, I2, I3. For example, optical sensor 17 acquires first image I1 at a first longitudinal portion of first label 3a (Figure 3a); then optical sensor 17 acquires second image I2 at a second longitudinal portion of second label 3b (Figure 3b); and then optical sensor 17 acquires third image I3 at a third longitudinal portion of third label 3c (Figure 3c). Hence, in first image I1 (Figure 3a) first label 3a has a longitudinal position, i.e. it is at a longitudinal position, with respect to field of view FV, which is different from the longitudinal position of second label 3b in second image I2 (Figure 3b), and from the longitudinal position of third label 3c in third image I3 (Figure 3c), and so on. In practice, the sensor device is configured so that optical sensor 17 acquires images I1, I2, I3 of all the longitudinal portions of a label 3, by means of a plurality of labels 3a, 3b, 3c in order to search for aesthetic feature AF, thereby defining the recovery mode. Thanks to the above configuration, the search and re-implementation of aesthetic feature AF can be performed without stopping labelling machine 1 and the labelling process, and furthermore without slowing down the labelling process. In other words, the advancing speed of web 4 along feed path P can be maintained constant. This is particularly advantageous in terms of inertial conditions of the system, since feed roller 11 does not have to be slowed down and then accelerated again to reestablish production speed. Moreover, complex control of the speed of feed roller 11 can be avoided. Accordingly, the aforementioned second acquisition frequency is conveniently lower than the first acquisition frequency, so that a complete inspection of all the longitudinal portions is possible by means of a plurality of labels 3. Compared to the first embodiment described above, more labels are needed to re-implement aesthetic feature AF in sensor device, thereby increasing the number of discarded labels 3. However, the advantages on inertial conditions and on labelling process speed are clear with respect to the first embodiment. With reference to Figures 5a, 5b and 5c, a further operative mode of sensor device according to the present invention will be described, using where possible the same reference numbers for corresponding elements. Optical sensor 17 is configured to acquire said images I, I1, I2, I3 with a first level of accuracy. In one embodiment, said level of accuracy is defined by a level of contrast of the image in field of view FV. In particular, the sensor device according to the invention is configured for operating in a validation mode according to which optical sensor 17, sequentially for a number of labels 3 (for example one single label 3, as in the case shown in Figures 5a-5c), acquires a number of images I1, I2, I3 with a second level of accuracy which is lower than the first level of accuracy, to detect whether on each label 3 of said number of labels 3 there is a further aesthetic feature AF’ which can be confused with aesthetic feature AF. More in particular, the sensor device is configured so that, during the validation mode, optical sensor 17 sequentially acquires, for each label 3 of said number of labels 3 and with said second level of accuracy, a number of images including at least a first image I1 and a second image I2. In the example shown, optical sensor 17 acquires a first image I1, a second image I2 and a third image I3 for a single label 3. According to the invention, the sensor device is configured so that in each image I1, I2, I3 of said number of images the label 3 has a respective longitudinal position, relative to field of view FV, which is different from the longitudinal position of the same label 3 in all the other images I1, I2, I3 of said number of images. In practice, during the validation mode, optical device 17 operates as in the first embodiment described above, with the difference that the level of accuracy changes from the first level of accuracy to the second level of accuracy. Accordingly, in image I1 (Figure 5a) label 3 has a longitudinal position, i.e. it is at a longitudinal position, with respect to field of view FV, which is different from the longitudinal position of the same label 3 in image I2 (Figure 5b) and in image I3 (Figure 5c), and so on. In practice, the sensor device is configured so that optical sensor 17 acquires images I1, I2, I3 of all the longitudinal portions of the label 3 in order to search for further aesthetic feature AF’, thereby defining the validation mode. If further aesthetic feature AF’ as defined above is not detected in label 3, the validation mode is successful. In fact, if a further aesthetic feature AF’ as defined above is not detected with a level of contrast lower than the one used during production mode, then it is rather sure that with the nominal level of contrast optical sensor 17 will not confuse aesthetic feature AF with other aesthetic features printed on label 3. If further aesthetic feature AF’ as defined above is indeed detected in label 3, validation mode is not successful and another aesthetic feature distinct from aesthetic feature AF has to be implemented as reference element, in order to avoid that optical sensor 17 could confuse aesthetic feature AF with further aesthetic feature AF’ during production mode. The above configuration allows to avoid that optical sensor 17 erroneously detect a feature on the label which is similar to the reference element during the production mode. In this way, an incorrect calculation of the linear distance between the reference element and the reference member 16 is avoided and, therefore, an incorrect cutting of the web 4 is avoided. The advantages of label feeding system 15 according to the present invention will be clear from the foregoing description. In particular, thanks to the recovery mode of the sensor device, the search and re-implementation of aesthetic feature AF can be performed without stopping labelling machine 1 and the labelling process. More in particular, it is no longer necessary to stop labelling machine 1 and to manually recalibrate and reinitialize the sensor device if optical sensor 17 looses aesthetic feature AF. In fact, the re-implementation of aesthetic feature AF in the sensor device can be performed automatically and on-the-fly (i.e. during production and without stopping the labelling process). The above configuration is also particularly advantageous upon staring the labelling process or upon a label format change (in which case the implementation of a new reference element is necessary). Furthermore, thanks to the validation mode of the sensor device, it is avoided that optical sensor 17 erroneously detect a feature on the label which is similar to the reference element during the production mode. Clearly, changes may be made to label feeding system 15 as described herein without, however, departing from the scope of protection as defined in the accompanying claims.

Claims

CLAIMS 1.- Label feeding system (15) for a labelling machine (1) configured to label containers (2) adapted to contain a pourable product, the label feeding system (15) comprising: - a sensor device including an optical sensor (17), the optical sensor (17) having a field of view (FV) and being configured for repeatedly inspecting the field of view (FV); - a frame (8) supporting said sensor device, so that the field of view (FV) can be fixed with respect to the frame (8); - a feed roller (11) for longitudinally feeding a web (4) of labelling material along a feed path (P) and through said field of view (FV), the web (4) comprising a plurality of longitudinally joined labels (3), each label (3) having a plurality of longitudinal portions and an aesthetic feature (AF) arranged at a determined longitudinal portion, said aesthetic feature (AF) being detectable by the optical sensor (17), the feed roller (11) being supported by said frame (8) and being configured for feeding the web (4) by rotating on itself and with respect to the frame (8); and - a reference member (16) fixed with respect to the field of view (FV) and arranged within the field of view (FV) so as to be detected by the optical sensor (17); the sensor device being configured for operating in a production mode according to which, sequentially for each label (3), the optical sensor (17) acquires an image (I) of the field of view (FV) at said determined longitudinal portion for detecting a linear distance between the respective aesthetic feature (AF) and the reference member (16); the sensor device being further configured for operating in a recovery mode according to which the optical sensor (17) acquires a number of images (I1, I2, I3) of the field of view (FV) for searching the aesthetic feature (AF), wherein each image (I1; I2; I3) of said number of images includes at least one longitudinal portion which is different from the other longitudinal portions included in the other images (I2, I3; I1, I3; I1, I2) of said number of images, so as to inspect all the different longitudinal portions during the recovery mode, by means of a single label (3) or cumulatively by means of a plurality of labels (3a, 3b, 3c).

2.- Label feeding system as claimed in claim 1, wherein the sensor device is configured so that, during the production mode, for all the images (I) acquired, the longitudinal position of each respective label (3) with respect to the field of view (FV) is the same.

3.- Label feeding system as claimed in claim 2, wherein the sensor device is configured so that, during the recovery mode, the optical sensor (17) sequentially acquires for a single label (3) a number of images (I1, I2, I3) including at least a first image (I1) and a second image (I2), and wherein the sensor device is configured so that in each image (I1; I2; I3) of said number of images the label (3) has a respective longitudinal position, relative to the field of view (FV), which is different from the longitudinal position of the same label (3) in all the other images (I2, I3; I1, I3; I1, I2) of said number of images.

4.- Label feeding system as claimed in claim 3, wherein the sensor device comprises a control unit (18) operatively connected to the optical sensor (17) and to the feed roller (11); and wherein the control unit (18) is configured to reduce the rotational velocity of the feed roller (11) so as to cause a deceleration of the web (4) along the feed path (P) and through the field of view (FV), during the recovery mode.

5.- Label feeding system as claimed in claim 2, wherein the sensor device is configured so that, during the recovery mode and for a group of labels (3a, 3b, 3c) including at least a first label (3a) and a second label (3b), the optical sensor (17) is configured to acquire at least one single said image (I1, I2, I3) for each label (3a, 3b, 3c) of the group of labels; and wherein the sensor device is configured so that in each acquired image (I1, I2, I3) the longitudinal position of the respective label (3a; 3b; 3c) of the group of labels, relative to the field of view (FV), is different from the longitudinal position of the other labels (3b, 3c; 3a, 3c; 3a, 3b) of the group of labels in the respective images.

6.- Label feeding system as claimed in any one of the foregoing claims, wherein the optical sensor (17) is configured for analyzing, for each image (I; I1, I2, I3) acquired during the production mode and during the recovery mode, a preestablished reference region (RR) of the field of view (FV), to detect the aesthetic feature (FV).

7.- Label feeding system as claimed in any one of the foregoing claims, wherein the field of view (FV) has a longitudinal extension, relative to the feed path (P), which is a fraction of the longitudinal length of each label (3); and wherein the sensor device is configured so that the number of images (I1, I2, I3) which are acquired during the recovery mode for searching the aesthetic feature (AF) is greater or equal to the ratio between the longitudinal length of a label (3) and the longitudinal extension of the field of view (FV).

8.- Label feeding system as claimed in claim 7, wherein the said number of images which are acquired during the recovery mode for searching the aesthetic feature (AF) is greater or equal to the first integer number following the numeric value of the ratio between the longitudinal length of a label (3) and the longitudinal extension of the field of view (FV).

9.- Label feeding system as claimed in any one of the foregoing claims, wherein the reference member (16) comprises at least part of a physical body, such as a mechanical rod or a bar, said physical body being supported by the frame (8) so that the physical body can be fixed with respect to the frame (8).

10.- Label feeding system as claimed in any one of the foregoing claims, wherein the optical sensor (17) is configured to acquire said images (I) of the field of view (FV) with a first acquisition frequency during the production mode; and wherein the optical sensor (17) is configured to acquire said images (I1, I2, I3) of the field of view (FV) with a second acquisition frequency during the recovery mode, the second acquisition frequency being different from the first acquisition frequency.

11.- Label feeding system as claimed in claims 3 and 10, wherein the second acquisition frequency is higher than the first acquisition frequency.

12.- Label feeding system as claimed in claims 5 and 10, wherein the second acquisition frequency is lower than the first acquisition frequency.

13.- Label feeding system as claimed in any one of the foregoing claims, wherein the optical sensor (17) is configured to acquire said images (I; I1, I2, I3) with a first level of accuracy; and wherein the sensor device is configured for operating in a validation mode according to which the optical sensor (17), sequentially for a number of labels (3) or for one single label (3), acquires a number of images (I1, I2, I3) with a second level of accuracy which is lower than the first level of accuracy, to detect whether on each label (3) of said number of labels (3) there is a further aesthetic feature (AF’) which can be confused with said aesthetic feature (AF).

14.- Label feeding system as claimed in claim 13, wherein the sensor device is configured so that, during the validation mode, the optical sensor (17) sequentially acquires, for each label (3) and with said second level of accuracy, a number of images (I1, I2, I3) including at least a first image (I1) and a second image (I2), and wherein the sensor device is configured so that in each image (I1, I2, I3) of said number of images the label (3) has a respective longitudinal position, relative to the field of view, which is different from the longitudinal position of the same label (3) in all the other images (I2, I3; I1, I3; I1, I2) of said number of images.

15.- Label feeding system as claimed in claim 13 or 14, wherein the level of accuracy is defined by a level of contrast of the image in the field of view (FV).

16.- Labelling machine (1) for labelling containers (2) adapted to contain a pourable product by means of labels (3) obtained from a web (4) of labelling material which comprises a plurality of longitudinally joined labels (3), each label (3) having a plurality of longitudinal portions and an aesthetic feature (AF) arranged at a determined longitudinal portion, the labelling machine (1) comprising a labelling module (7) and a label feeding system (15) for feeding the web (4) through the labelling module (7) as claimed in any one of the foregoing claims, the labelling module (7) comprising: - a fixed frame (8); - a plurality of guide rollers (12) arranged on the fixed frame (8) and configured to support the web (4) and to guide it along the feed path (P); - a cutting unit (13) for repeatedly cutting the web (4) at a cutting station (C) so as to separate a sequence of labels (3) from the web (4) itself; and - a label transfer device (14) configured to receive, retain and advance each label (3) and to feed each label (3) toward the containers (2) for the application thereon.