Product labelling

0001This invention relates to a product labelling apparatus according to the preamble of claim 1, and a method according to the preamble of claim 17, and known from <patcit id="pcit0001" dnum="WO02102669A"><text>WO 02/102669 A</text></patcit>.

0002Products to be sold are commonly labelled. In this regard, automatic labelling apparatus may be employed where the products are smaller and processed in large volumes. One approach in this regard is to wipe a label onto each product as its passes a labelling head. This approach, however, is only well suited for labelling products of uniform dimensions. Where products have irregular dimensions, such as agricultural produce, the distance between a given product and the labelling head will vary. To label such products, tamping labellers are typically used. <patcit id="pcit0002" dnum="US6257294B"><text>US6,257,294</text></patcit> to Weisbeck discloses a tamping labeller. In Weisbeck, a turret carries a number of reciprocating pick up heads about its periphery. The turret has a vacuum plenum and a positive pressure plenum. The turret rotates each head, consecutively, to a labelling station. A head normally communicates with the vacuum plenum which keeps it in a retracted position; also, due to end perforations in the head, the negative pressure holds a label at the end of the head. However, when the head reaches the labelling station, it is coupled to the positive pressure plenum which causes the head to rapidly extend until it tamps a product below. The force of the tamping forms an adhesive bond between the pressure sensitive adhesive of the label and the product. Labels are fed to each pick-up head from a label cassette with a label web comprising serially arranged labels on a release tape.

0003The labelling apparatus of Weisbeck is suited to label a continuous line of products passing under the labeller. However, more typically, agricultural produce which is to be labelled arrives in trays, each tray having an arrangement of cup-like depressions which hold the products. In order to label products in a tray, a bank of tamping labellers may be used and the trays conveyed underneath this bank of labellers. However, with this set-up, some mechanism is required to ensure that the labellers, when tamping, do not miss the products. One approach in this regard is to use a limited number of types of trays to hold the products, where each type of tray has predefined pattern of cup-like depressions.

0004The labelling apparatus may then be configured to expect products to be arranged in a certain pattern, with the expected pattern being based on the type of tray that will next pass under the labellers. With such a system, a vision system may be used to detect the type of tray.

0005A drawback with this approach is that products may not be present in each of the tray cups. A further drawback is that some types of products, such as vine ripened tomatoes, may have obstructions (the vines) which may end up being labelled rather than the product itself.

0006Therefore, there remains a need for more accurate product labelling apparatus. The present invention aims to provide an improved product labelling apparatus and method in a simple yet effective manner and to alleviate at least some of the problems of the prior art.

0007Various aspects of the invention are set out in the independent claims. Various optional features are set out in the dependent claims.

0008A product labelling apparatus according to claim 1 has inter alia a plurality of labellers, an imager for imaging products, and a processor responsive to an output of the imager and operatively connected to a control input of each of the labellers. The processor processes an image received from the imager to identify a portion of a product which portion will pass a target area of a given labeller. The processor then tracks progress of that portion of the product and controls an appropriate one of the labellers to label the portion of the product when that portion of the product is at the target area of the given labeller.

0009Optionally, the imager may be a colour camera. In such instance, the image may be filtered to leave a first range of colours which may represent the colours of the products. The filtered image may be processed to obtain a plurality of groups of blobs, each blob comprising an area of the first range of colours and each group of blobs representing a product. A blob may then be selected from a given group of blobs which blob represents a portion of a product which will pass a target area of a given labeller. The progress of the product represented by the given group of blobs may be tracked and the given labeller is controlled to label the noted portion of the product.

0010In a preferred construction, there is provided product labelling apparatus, comprising: <ul id="ul0001" list-style="none" compact="compact"><li>a labeller for labelling products; a camera for capturing an image of a product; a processor responsive to receiving said image from said camera and operatively connected to a control input of said labeller for: processing said image to reduce said image to a representation of a plurality of blobs; analysing said representation to select a one of said plurality of blobs within a labelling area of said labeller; and</li><li>controlling said labeller such that said labeller applies a label to a target area of said product; where said target area of said product corresponds to said one of said plurality of blobs within said labelling area of said labeller.</li></ul>

0011The method of the invention set out in Claim 17 may in a further preferred embodiment include: <ul id="ul0002" list-style="none" compact="compact"><li>filtering said image to leave a first range of colours representative of colours of said products; obtaining a plurality of groups of blobs, each blob comprising an area of the first range of colours and each group of blobs representing one of said products; selecting a blob from a given group of blobs, which blob represents a portion of a given product which will pass a target area of a given labeller; tracking said given product represented by said given group of blobs and controlling said given labeller to label said portion of said given product.</li></ul>

0012The present invention will now be explained in more detail by the following nonlimiting description of preferred embodiments and with reference to the accompanying drawings, in which: <ul id="ul0003" list-style="none" compact="compact"><li><figref idref="f0001">Figure 1</figref> is a plan schematic view of a labelling apparatus according to an embodiment of the invention,</li><li><figref idref="f0002">Figure 2</figref> is a perspective view of a preferred configuration for each labeller in the apparatus of claim 1,</li><li><figref idref="f0003">Figure 3</figref> is a flow diagram illustrating the operation of a processor of the apparatus of <figref idref="f0001">Figure 1</figref>, and</li><li><figref idref="f0004">Figure 4</figref> is a schematic view of a preferred construct of the processor.</li></ul>

0013Turning to <figref idref="f0001">Figure 1</figref>, a labelling apparatus 10 comprises labellers 12a to 12h (referred to individually as labellers 12) mounted by mounts 14 at a fixed position above a conveyer 16, which moves in a downstream direction D. The labellers 12 are arranged as an upstream bank 18a of labellers (12a to 12d) and a downstream bank 18d of labellers (12e to 12h). Each bank 18a, 18d of labellers extends transversely of the conveyor 16. The labellers in a bank are equally spaced and the labellers of the downstream bank 18d are offset from those of the upstream bank 18a so that each labeller has a different transverse position over the conveyor. Further, the labellers 12 extend substantially across the width of the conveyor so as to provide eight distinct transverse positions over the conveyor. The labellers 12 are operatively connected to a processor 22 on paths 20. The processor has an associated memory 23 and user interface 36. Memory 23 is loaded with software so that the processor may operate as hereafter described from a computer readable medium which may be, for example, a disk 34, a CD-ROM, a solid state memory chip, or a file downloaded from a remote source.

0014The labellers 12 are downstream of an imager 24, which in this embodiment is a colour camera; a filter 25 may be positioned in front of the camera. The camera is arranged to image an area of the conveyor and output this image to the processor 22. In this regard, products 26 may be carried in trays 28 and the camera may image an area which captures one such tray. A photocell 29 may detect the leading edge of a tray when the tray is within the field of view of the camera and output a detect signal to the camera 24 which prompts the camera to capture the image of the tray. The photocell may also output directly to processor 22. A conveyor speed indicator 32 (which, for example, may be a rotary encoder, a sensor which senses marks on the conveyor, or, where the conveyor moves at a known constant speed, simply a timer) also outputs to the processor.

0015Referencing <figref idref="f0002">Figure 2</figref>, an example labeller 12 has a rotatably mounted turret 40. A timing belt 42 connects the turret 40 to a stepper motor 44. A label cassette (not shown) has a cassette magazine (not shown) to which is wound a label web 56. The web comprises a release tape 58 carrying a plurality of labels backed with a pressure sensitive adhesive. The label web extends from the cassette magazine along a tongue 74 to a label pick-up station 70, with the release tape 58 returning. A communication path 20 from the processor 22 (<figref idref="f0001">Figure 1</figref>) terminates at stepper motor 44.

0016The turret 40 has a stationary core 80 with a port 82 for connection to a vacuum source (not shown) and a port 84 for connection to a source of positive pressure (not shown). A bellows 60 fabricated of flexible material, such as rubber or silicone, is stretched over a lip of each air diffuser (not shown) extending from the turret 40. The tamping end 62 of each bellows is perforated with pin holes. Further details of example labeller 12 may be had from <patcit id="pcit0003" dnum="WO02102669A"><text>WO 02/102669 published December 27, 2002</text></patcit>.

0017Another exemplary tamping labeller is a piston-type tamping labeller, such as the afore-referenced labeller of <patcit id="pcit0004" dnum="US6257294B"><text>US 6,257,294</text></patcit> to Weisbeck. Also, it will be appreciated that if the products are of a reasonably uniform nature, other types of labellers may be suitable, such as a labeller which wipes labels onto the products.

0018Tray 28 may have a pattern of cup-like depressions, however, as illustrated in <figref idref="f0001">Figure 1</figref>, not all of the depressions may hold a product. Thus, the products are unpredictably positioned in the tray. For example, as illustrated, the products may be vine ripened tomatoes which remain attached to vines 30 such that the products are irregularly spaced.

0019With reference to <figref idref="f0003">Figure 3</figref> along with <figref idref="f0001">Figure 1</figref>, in operation, a user, through interface 36, may input the type of products that will be held by trays 28 placed on conveyor 16. With this information, the processor may retrieve from memory 23 a range of foreground colours indicative of the predominant colour of the products, a range of colours of any obstructions, and a range of background colours indicative of the colour of the trays (S110). In this regard, the trays may be manufactured so as to uniformly have a colour which is distinct from the colour of any product that will be labelled by labelling apparatus 10. For example, the trays may be blue in colour and, if so, memory 23 stores a range of blue colours as the background colour.

0020If, for example, the user indicates that the products to be labelled are vine-ripened tomatoes, then the range of foreground colours may be reds. Further a range of greens may be retrieved as indicating the colour of the obstructing vines.

0021The conveyor 16 may then be advanced in downstream direction D to convey trays 28, loaded with the indicated products, toward labelling apparatus 10. When the leading edge of a tray 28 reaches photocell 29, the photocell prompts the camera 24 to image the tray. The camera then sends this image to processor 22 (S112). The processor can then process this image as follows. With knowledge of the range of colours representative of the product, the processor can electronically filter out from the image all but this range of colours to obtain a first (product colour) filtered image (S114). The processor can also electronically filter out the range of colours representative of the background colours, i.e. the colour of the trays, in order to obtain a second (background colour) filtered image (S116). Further, if the memory 23 has an indication that there is a range of colours associated with obstructions, with knowledge of this range of colours, the processor can electronically filter out from the camera image all but this range of colours in order to obtain a third (obstruction colour) filtered image (S118). As an alternative to the processor electronically filtering the camera image, physical filters 25 may be placed in front of the camera. In this instance, the camera may take up to three (rapid) consecutive images and the processor may control which of the filters is in front of the camera while each image is taken. (The control path to the optional filters 25 is not shown).

0022The processor may then establish groups of blobs, each group representing a product. In doing so, the processor may overlay the second filter on the first filter in order to assist in establishing the perimeter of each group of blobs. Further, the processor may overlay the third filter on the first filter in order to better delineate the boundary between the blobs and obstructions. Additionally, the processor may connect separated blobs in a group, at least where such orphan blobs are not separated by areas represented in the third filtered image (S120).

0023The resulting groups 226 of blobs 230 for the tray 28 illustrated in <figref idref="f0001">Figure 1</figref> are illustrated in <figref idref="f0004">Figure 4</figref>. Each labeller 12 (<figref idref="f0001">Figure 1</figref>) can label a product which lies within a certain range of transverse positions on the conveyor 16. The processor may therefore overlay "swaths" (or paths) 212 on the groups 230 of blobs where each swath represents the range of transverse portions over which one labeller can label a product. Thus, for example, swath 212b represents the transverse positions over which labeller 12b may label a product, and so on. For each group of blobs, the processor may then select a blob that is comfortably within a given swath 212. The selection process may involve looking for the largest blob that is comfortably within a given swath. For example, for group 226a (which represents product 26a of <figref idref="f0001">Figure 1</figref>), the processor may note that blob 230b is comfortably within swath 212b and that blob 230a is comfortably within swath 212f. In this instance, the processor may select blob 230a, as it is the larger of the two blobs.

0024Once the processor has identified an appropriate swath 212 for a given group of blobs, it chooses the labeller 12 associated with that swath as the labeller to label the product which is represented by the given group of blobs (S122).

0025When the photocell 29 detects the leading edge of a tray, the tray is a known distance from labellers 12. This detection signal may be input from the photocell directly to processor 22. Alternatively, this signal may be indirectly received by the processor as the image signal from camera 24. With the processor knowing when the leading edge of a tray is at the photocell and knowing the speed of the conveyor from speed indicator 32, the processor will be aware when each product 26 in tray 28 reaches one of the banks 18 of labellers 12. Thus, the processor can track a product represented by a given group of blobs reaches each bank of labellers. Therefore, the processor can signal the labeller which it chose to label a product represented by the given group of blobs at an appropriate time (S124). Put another way, the processor can track the progress of the tray by notionally progressing the image of the groups of blobs with respect to notional banks of labellers. In this way, the processor will know when a given group of blobs reaches each notional bank of labellers and can fire the chosen labeller for the given group of blobs at the appropriate time.

0026Optionally, the processor may establish groups of blobs with only a filtered image leaving the first range of colours representing a product. However, such an approach is not likely to be as robust as one which also uses a filtered image leaving the background colours. And, where there are obstructions, the approach becomes even more robust if use is made of a filtered image leaving the obstruction colours.

0027Optionally, rather than using colour-based blob analysis, a monochrome blob analysis may be used. More particularly, the imager 24 may be a monochrome camera and different grey-scales may be considered to be indicative of different colours. More particularly, the processor may retrieve from memory 23 a range of grey-scales indicative of the predominant colour of the products, a range of grey-scales indicative of background colours (i.e., the colours of the trays), and a range of grey-scales indicative of obstructions. Mechanical or electronic filtering may be used to obtain images of the different ranges of grey-scales which are indicative of the selected colours. Blob-based analysis may then proceed as described hereinbefore in order to target products for labelling.

0028As an option to a blob-based analysis, with an appropriate imager 24, processor 22 may obtain and analyse topographic images. For example, the processor 22 may be configured to generate a topographic image (without colour information) from output received from stereoscopic cameras (as, for example, infra-red cameras), ultrasonic imagers, sonar imagers, or radar imagers. Processor 22 may then be configured to analyse the topographic image to identify topographies indicative of products and then select a suitable high point on each product for labelling. Product recognition may be accomplished in any suitable fashion, such as with a neutral network. Where there are obstructions (stems), the processor must also be configured to identify, and avoid labelling, these.

0029Other modifications will be apparent to those skilled in the art and, therefore, the invention is defined in the claims.