Detecting deviations in packaging containers for liquid foods

Abstract

A monitoring system implements a method for universal and efficient detection and classification of deviations in packaging containers for liquid food products in a manufacturing plant. The method comprises obtaining (31) image data of the packaging container or of a starting material used to produce the packaging container; analyzing (32) the image data for detecting a current deviation; processing (33) the current deviation with respect to a set of basis functions, the set of basis functions (33) being associated with a deviation type of the current deviation to obtain a current set of weights representing the current deviation; and determining (35) a current classification of the current deviation based on the current set of weights. The set of basis functions can be pre-calculated based on recurrences of packaging containers or starting materials including different degrees of deviation.

Description

Technical Field

[0001] The present disclosure generally relates to techniques for detecting deviations in packaging containers for liquid food products, particularly detection based on image data captured in a manufacturing plant during production of the packaging containers. Background Art

[0002] In production lines, for example in the production of sealed packaging containers for liquid food products (in filling machines or other machines used to produce such packaging containers), it is crucial to detect deviations (e.g., defects or other deviations from the expected product configuration) in order to configure optimal operating settings and ensure that the expected performance is achieved over time. Deviations in the produced packaging containers may result in changes in the packaging container's appearance, such as inconsistencies that may be of concern from a consumer's perspective, or inconsistencies that result in suboptimal performance (e.g., in terms of the integrity or stability of the packaging container). Efficient, automated, and reliable tools and procedures are needed for quality control and for the identification and grading of deviations in liquid food packaging containers in manufacturing plants. Summary of the Invention

[0003] It is an object of the present invention to at least partially overcome one or more limitations of the prior art.

[0004] One object is to provide automated detection and grading of deviations in liquid food packaging containers.

[0005] Another object is to provide for the detection of deviations of different types in packaging containers for liquid food products.

[0006] One or more of these objects, as well as other objects that may become apparent from the following description, are achieved at least in part by a method for detecting deviations in liquid food packaging containers in a manufacturing plant, a method for generating a set of basis functions associated with a deviation type, a computer-readable medium, and a system, embodiments thereof. A method for detecting deviations in liquid food packaging containers in a manufacturing plant comprises: obtaining image data of a packaging container or a starting material used to produce the packaging container; analyzing the image data to detect a current deviation; processing the current deviation with respect to a set of basis functions, the set of basis functions being associated with the deviation type of the current deviation to obtain a current group weight representing the current deviation; and determining a current ranking of the current deviation based on the current group weight. A method for generating a set of basis functions associated with a deviation type, for use in a method for detecting deviations in liquid food packaging containers in a manufacturing plant, comprises: obtaining a reproduction of the packaging container or the starting material used to produce the packaging container, the reproduction including deviations of the deviation type; processing the deviation by a basis function calculation algorithm to obtain one or more basis functions; and generating the set of basis functions based on the one or more basis functions. A computer-readable medium comprising computer instructions which, when executed by a processor, cause the processor to execute a method for detecting deviations of liquid food packaging containers in a manufacturing plant or a method for generating a set of basis functions associated with a deviation type. A system for detecting deviations of liquid food packaging containers in a manufacturing plant, the system comprising a processor configured to execute a method for detecting deviations of liquid food packaging containers in a manufacturing plant or a method for generating a set of basis functions associated with a deviation type. In one embodiment, the method for detecting deviations of liquid food packaging containers in a manufacturing plant further comprises: mapping the current set of weights to a hierarchical database, the hierarchical database associating the weight combination with the classification of the deviation type of the current deviation, wherein the current classification of the current deviation is determined based on the mapping. In one embodiment, the method for detecting deviations of liquid food packaging containers in a manufacturing plant further comprises: processing the current deviation to determine the deviation type. In one embodiment, the method for detecting deviations of liquid food packaging containers in a manufacturing plant further comprises: obtaining the set of basis functions from a basis function database according to the deviation type. In one embodiment, the method for detecting deviations of liquid food packaging containers in a manufacturing plant further comprises: determining a current eigenvector representing the current deviation; and calculating the current group weight based on the current eigenvector and the set of basis functions. In one embodiment, the calculating the current group weight comprises: determining a projection of the current eigenvector onto the set of basis functions; and determining the current group weight based on a scalar value of the projection. In one embodiment, the basis functions are linearly independent and / or mutually orthogonal. In one embodiment, the basis functions correspond to principal components given by principal component analysis.In one embodiment, the method for detecting deviations of liquid food packaging containers in a manufacturing plant further comprises: determining a timestamp of the current deviation; determining associated production parameters of the manufacturing plant based on the timestamp, and associating the timestamp, the current grade, and the deviation type with the production parameters. In one embodiment, the method for detecting deviations of liquid food packaging containers in a manufacturing plant further comprises: transmitting control instructions including production parameters modified according to the current grade and / or the deviation type to machines in the manufacturing plant. In one embodiment, the method for detecting deviations of liquid food packaging containers in a manufacturing plant further comprises: causing an alarm notification based on the current grade. In one embodiment, the deviation type comprises any of the following: wrinkles in the material of the packaging container or in the starting material, unsealed flaps of the packaging container, tears or turbid holes in the packaging container or the starting material, dents or protrusions in the packaging container, delamination of the packaging container or in the starting material, and defective patterns and / or colors and / or holographic or metallized films on the surface of the packaging container or the starting material.

[0007] A first aspect of the present invention is a method for detecting deviations in packaging containers of liquid food in a manufacturing plant, the method comprising: obtaining image data of the packaging container or the starting material used to produce the packaging container; analyzing the image data for detecting a current deviation; processing the current deviation with respect to a set of basis functions, the set of basis functions being associated with the deviation type of the current deviation to obtain a current group weight representing the current deviation; and determining a current grade of the current deviation based on the current group weight.

[0008] A first aspect provides a general and efficient technology that is particularly suitable for the automatic detection and grading of deviations in liquid food packaging containers. The first aspect characterizes the detected deviation according to the weights of one or more basis functions known to represent the deviation type of the detected deviation. Since different basis function groups can be determined for different deviation types, the use of basis functions makes the detection universal. In addition, the basis functions of the deviation type can be pre-calculated based on the reproduction of the actual deviation, and the detection and grading can therefore be customized for the actual production environment. The calculation of the current set of weights produces a set of corresponding weight values, which provide a "fingerprint" of the current deviation, thereby enabling the current deviation to be simply and reliably assigned a quality grade, for example by using the relationship between the known and predetermined weight values ​​and the quality grade. By increasing the number of basis functions, and thereby increasing the number of combined weights in the fingerprint, the accuracy of the grading can be improved.

[0009] In one embodiment, the method further comprises mapping the current set of weights to a ranking database that associates weight combinations with a ranking of the deviation type of the current deviation, wherein a current ranking of the current deviation is determined based on the mapping.

[0010] In one embodiment, processing further comprises processing the current deviation to determine the deviation type.

[0011] In one embodiment, the method further comprises: obtaining the set of basis functions from a basis function database according to the deviation type.

[0012] In one embodiment, the method further comprises: determining a current eigenvector representing the current deviation; and calculating the current set of weights based on the current eigenvector and the set of basis functions.

[0013] In one embodiment, calculating the current group weight includes: determining a projection of the current eigenvector onto the set of basis functions; and determining the current group weight based on a scalar value of the projection.

[0014] In one embodiment, the basis functions are linearly independent and / or mutually orthogonal.

[0015] In one embodiment, the basis functions correspond to principal components given by principal component analysis (PCA).

[0016] In one embodiment, the method further comprises: determining a timestamp of the current deviation; determining an associated production parameter of the manufacturing plant based on the timestamp, and associating the timestamp, the current grade, and the deviation type with the production parameter.

[0017] In one embodiment, the method further comprises transmitting control instructions including production parameters modified according to the current grade and / or the type of deviation to a machine in the manufacturing plant.

[0018] In one embodiment, the method further comprises causing an alarm notification based on the current rating.

[0019] In one embodiment, the types of deviations include any of the following: wrinkles in the material of the packaging container or in the starting material, unsealed flaps of the packaging container, tears or cloudy holes in the packaging container or the starting material, dents or bulges in the packaging container, delamination of the packaging container or in the starting material, and defective patterns and / or colors and / or holographic or metallized films on the surface of the packaging container or the starting material.

[0020] A second aspect of the present invention is a method for generating a set of basis functions associated with a deviation type, for use in the method of the first aspect or any embodiment thereof. The method of the second aspect comprises: obtaining a representation of a packaging container or a starting material used to produce the packaging container, the representation including deviations of the deviation type; processing the deviations using a basis function calculation algorithm to obtain one or more basis functions; and generating the set of basis functions based on the one or more basis functions.

[0021] A third aspect of the present invention is a computer readable medium comprising computer instructions which, when executed by a processor, cause the processor to perform the method of the first or second aspect or any embodiment thereof.

[0022] A fourth aspect of the present invention is a system for detecting deviations in packaging containers of liquid foods produced in a manufacturing plant, the system comprising a processor configured to perform the method of the first or second aspect or any embodiment thereof. The system may also include at least one imaging device configured to capture and provide image data. The processor may be included in a monitoring device and operably connected to a communication interface configured to be connected to the at least one imaging device.

[0023] Other objects, as well as features, aspects and advantages of the present invention will become apparent from the following detailed description and accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Embodiments of the invention will now be described, by way of example, with reference to the accompanying schematic drawings.

[0025] Figure 1 is a schematic diagram of a system for detecting deviations in packaging containers;

[0026] Figure 2 is a top view of the packaging container with a deviation in the upper right corner;

[0027] Figure 3 is a flow chart of an exemplary method for detecting bias;

[0028] Figure 4 is a schematic diagram of a system for detecting deviations when operating in a manufacturing plant;

[0029] Figure 5a shows an example of a set of basis functions for a particular deviation type, and Figure 5b is a scatter plot of the weight values ​​determined for the packer with that particular bias type, and uses Figure 5a The two basis functions in ;

[0030] Figure 6 is a flow chart of an exemplary method for generating basis functions;

[0031] Figure 7 is a functional block diagram of a system for generating basis functions;

[0032] Figure 8 is a block diagram of an apparatus that may implement methods according to some embodiments of the present disclosure. DETAILED DESCRIPTION

[0033] The embodiments will now be described more fully hereinafter with reference to the accompanying drawings, in which some, but not all, embodiments are shown. Indeed, the invention may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure may satisfy applicable legal requirements.

[0034] In addition, it should be understood that, where possible, any advantages, features, functions, devices and / or operational aspects of any embodiment described and / or contemplated herein may be included in any other embodiment described and / or contemplated herein, and / or vice versa. In addition, where possible, any term expressed in the singular herein is also meant to include the plural form and / or vice versa unless expressly stated otherwise. As used herein, "at least one" shall mean "one or more" and these phrases are intended to be interchangeable. Accordingly, the terms "a" and / or "an" shall mean "at least one" or "one or more", although the phrases "one or more" or "at least one" are also used herein. As used herein, unless the context requires otherwise due to the language of expression or necessary implication, the word "comprise" or variations such as "comprises" or "comprising" are used in an inclusive manner, i.e., specifying the presence of the stated features, but not excluding the presence or addition of additional features in various embodiments. As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items. Furthermore, a "set" of items is intended to imply the provision of one or more items.

[0035] As used herein, "liquid food" refers to any food that is non-solid, semi-liquid, or pourable at room temperature, including beverages such as juice, wine, beer, soda, as well as dairy products, sauces, oils, cream, custard, soups, etc., and solid foods in liquids such as beans, fruits, tomatoes, stews, etc.

[0036] As used herein, "packaging container" refers to any container suitable for sealing and containing liquid food, including but not limited to containers formed from packaging laminate materials (such as cellulose-based materials), and containers made of or containing plastic materials.

[0037] As used herein, "starting material" refers to any base material that is processed to form a part of a packaging container, including but not limited to sheets of packaging laminate, closures (caps, lids, covers, stoppers, foils, etc.) for closing packaging containers, and labels for attachment to sheets or packaging containers.

[0038] As used herein, the term "deformation" is intended to generally refer to any distortion or deviation from the acceptable or ideal appearance of a packaging container. Thus, deformation is not limited to changes in form or shape, but also includes changes in surface structure, surface pattern, surface color, etc.

[0039] As used herein, the term "basis function" is used in its ordinary sense and refers to linearly independent elements across a function space, such that every function in the function space can be expressed as a linear combination of the basis functions. The basis functions can be expressed as vectors, and the function space can be a vector space of arbitrary dimension.

[0040] Like reference numerals refer to like elements throughout.

[0041] Figure 1 Schematic diagram of a system 300 for detecting deviations in a packaging container 401 of a liquid food produced at a manufacturing plant. Container 401 is sealed to hold the liquid food and can be made, at least in part, of a laminated or non-laminated cardboard material or plastic material. For example, container 401 can be a carton or bottle as is known in the art.

[0042] The system 300 can be arranged to detect deviations in a factory upstream, within, or downstream of a machine 400. The machine 400 can be a machine for feeding and / or manipulating a starting material for a container 401 or a portion thereof, a filling machine, a capping machine, an accumulator machine, a straw application machine, a secondary packaging machine, or any other type of packaging machine deployed in a manufacturing plant for packaging liquid food products.

[0043] System 300 includes a monitoring or inspection device 301 configured to detect and signal deviations occurring during the production of packaging containers 401; and an imaging device 302 arranged and operative to capture image data of the containers 401 or starting materials for use by monitoring device 301. Imaging device 302 can be positioned along any portion of a production line in a manufacturing facility. It is also contemplated that multiple imaging devices 302 can be positioned to capture image data from different portions of the production line and / or from different angles relative to container 401 and / or using different exposure settings or image processing parameters. Thus, the image data can include multiple image data streams captured from such multiple imaging devices 302.

[0044] The image data may represent the appearance of the container 401 or the starting material or a portion thereof. In an alternative embodiment, the imaging device 302 may be configured to capture images representing internal features of the container 401, such as one or more cross-sectional images. The image data may be one-dimensional, two-dimensional, or three-dimensional and include any number of channels, such as a grayscale channel and / or any number of color channels.

[0045] System 300 can be deployed for quality monitoring, such as indicating that packaging containers and / or starting materials should be discarded due to insufficient quality, or sorting packaging containers 401 according to different quality grades. Alternatively or additionally, system 300 can be deployed to provide input data to the control system of one or more machines 400 in a manufacturing plant. For example, the input data can cause the control system to interrupt production in the machine or reconfigure the machine by adjusting one or more current settings of the machine.

[0046] Figure 2 is a top view of a packaging container 401 captured by an imaging device 302. The container 401 has a deviation 403 in its upper right corner, which in this example is a deformation in the form of a dent / bump. The monitoring device 301 is configured to process Figure 2 to determine a classification for the containers in that image.

[0047] One aspect of the present disclosure relates to a detection method that can be implemented by system 300 and includes: obtaining image data of a packaging container or a starting material used to produce the packaging container; analyzing the image data for detecting a current deviation; processing the current deviation with respect to a set of basis functions associated with a deviation type of the current deviation to obtain a set of current weights representing the current deviation; and determining a current grade of the current deviation based on the set of current weights.

[0048] In one embodiment, the detection method is deterministic and operates on pre-computed basis functions for a specific type or class of deviations (denoted herein as "deviation types"). The deviation type can be defined by a specific location and / or a specific deformation on the container (or starting material), examples of which include: dents, wrinkles, unsealed flaps, torn or cloudy holes, delamination, imperfect color and / or pattern of the surface, defects in holographic or metallized films attached to or otherwise included on the surface, imperfect embossing or folding, etc. In one example, the deviation type can be a specific deformation regardless of location. In another example, the deviation type can be any deformation at a specific location. Many variations are conceivable and readily understood by those skilled in the art.

[0049] The set of basis functions can be pre-calculated using any suitable basis function calculation algorithm that presents linearly independent basis functions based on eigenvectors representing the types of deviations in multiple containers (or starting material items). Examples of such calculation algorithms include, but are not limited to, principal component analysis (PCA), independent component analysis (ICA), wavelet analysis, non-negative matrix factorization (NMF), Fourier analysis, autoregressive analysis, factor analysis, common spatial patterns (CSP), canonical correlation analysis (CCA), and the like. Such calculation algorithms define a model function for the observations based on a set of basis functions. A general linear model function is typically assumed: X = ∑(wi·φi), where X is the observation value, φi is the corresponding basis function, and wi is the corresponding weight or basis function coefficient. The basis functions are linearly independent. For better regulation, some calculation algorithms may also impose orthogonality between the basis functions. Below, an example will be given regarding PCA, a statistical procedure that uses a transformation to convert the observations of a set of correlated variables into a set of linearly uncorrelated basis function values, called "principal components." The number of principal components is less than or equal to the number of original variables. The transformation is defined in such a way that the first principal component has the largest possible variance, i.e., it explains as much of the variability in the data as possible, and each subsequent component, in turn, has the largest possible variance subject to the constraint that it is independent of the previous component. Thus, PCA results in multiple principal components for the set of observations and a variance for each principal component. This variance is or corresponds to the weight of the corresponding basis function described above.

[0050] Now refer to Figure 3 Flowchart and Figure 4The detection method is illustrated by an exemplary monitoring system of FIG. The monitoring system includes an imaging system 302 arranged to capture images of a container 401 and / or a starting material 401' (illustrated as a sheet of packaging material) in a manufacturing plant. The monitoring device 301 includes a deviation detector 40, a weight generator 41, and a classifier 42. The monitoring device 301 is arranged to access a first database 43 storing a dictionary of pre-calculated basis functions and a second database 44 storing classification data. As shown, the databases 43, 44 can be part of the monitoring device 301, or located on one or more external devices and accessed by the monitoring device 301 via a wired or wireless connection.

[0051] The detection method 30 shown comprises a step 31 of obtaining current image data of one or more containers 401 or starting materials 401'. The current image data may be obtained by Figure 4 , which depicts a container 401 having a deviation 403. Step 32 analyzes the current image data Ic to detect the current deviation. In one example, step 32 compares the current image data Ic or a selected subset thereof with a reference image of a non-deviation container or starting material to detect the presence of the current deviation. In another example, step 32 extracts values ​​of one or more feature parameters from the current image data and compares these values ​​with one or more reference values ​​to detect the presence of the current deviation. If step 32 does not detect a deviation, method 30 returns to step 31. Figure 4 In the system, steps 31 and 32 may be performed by the deviation detector 40 .

[0052] Otherwise, the method proceeds to step 33, which determines current deviation data representing the current deviation and calculates a set of current weights which, when applied to a linear combination of a set of predefined basis functions (i.e. corresponding to the model function X = ∑(wi·φi)), produce the best approximation of the current deviation according to a predefined criterion. Step 33 can calculate any number of weights, including a single weight. The current deviation data is determined in the same format as the observations used to determine the set of predefined basis functions. The current deviation data represents the current deviation by a plurality of numerical values ​​extracted from the current image data Ic. The current deviation data can therefore be represented as a current "feature vector" of one or more dimensions. In one embodiment, the feature vector represents the geometry of the deviation, such as its shape and / or topography. As used herein, "topography" is the distribution of height values ​​relative to a geometric reference (e.g. a two-dimensional geometric plane). In Figure 2In the example of the indentation 403, the topography can specify a change in height perpendicular to a geometric plane parallel to the top surface of the container 401. The topography can be expressed as height values ​​along a line (e.g., along a deviation, such as a wrinkle; or along a predefined reference line). Alternatively or additionally, the topography can be expressed as a two-dimensional distribution of height values. In another example, the feature vector can define the curvature at different locations within the deviation, such as along a line or in two dimensions. Any suitable definition of curvature can be used, including conventional measurements of intrinsic and extrinsic curvature in two or three dimensions. In another example, the feature vector can represent the calculated difference in one or more dimensions between the image data and a reference template corresponding to a container (or starting material) without the deviation. It is also conceivable that the feature vector includes a combination of different metrics (e.g., any of the above).

[0053] In one embodiment, step 33 determines the projection of the current eigenvector on a set of basis functions and determines a set of current weights based on the scalar value of the projection. For example, each weight can be calculated as the dot product (scalar product) between the eigenvector and each basis function.

[0054] exist Figure 4 In a system of , step 33 may be performed by the weight generator 41 and the set of basis functions may be retrieved from the first database 43. Figure 4 , the current group weight calculated by the weight generator 41 for the current image data Ic is represented by [W]c.

[0055] Since a pre-computed set of basis functions is generally applicable only to a specific deviation type, the foregoing description assumes that the method 30 is customized to detect a specific deviation type. In a more general embodiment of the method 30, step 33 may process the current deviation to assign the current deviation to one of a plurality of predefined deviation types and retrieve a corresponding set of basis functions for the current deviation type, for example from the first database 43. Figure 4 As shown, the first database 43 can store different sets of basis functions, denoted by [BF]i, for different deviation types, denoted by TYPEi. Figure 4 In the example system of , step 33 may involve determining a current deviation type TYPEc for the current deviation, and accessing the first database 43 to retrieve a current set of basis functions ([BF]c) associated with the current deviation type (TYPEc).

[0056] Step 34 maps the weight group calculated in step 33 to the hierarchical data in the hierarchical database. Should be understood that the hierarchical database can store the hierarchical data of every kind of deviation type. In one embodiment, the hierarchical data is associated with the value of one or more weights with the hierarchical. For example, the hierarchical data can define "weight space" and associate the different areas in the weight space with corresponding hierarchical, wherein each weight in this group of weights defines a dimension of the weight space. For example, if this group of weights includes four weights, then the weight space has four dimensions, and defines the above-mentioned area in the four-dimensional weight space. Step 35 determines the current grading or grade of the current deviation based on the mapping in step 34. The grading can indicate the degree (severity) of the current deviation of the outward appearance and / or function of the packaging container 401. The grading can be assigned with any number of gradings, levels or grades. In a non-limiting example, grading is binary and the packaging container can be designated as acceptable or unacceptable.

[0057] exist Figure 4 In a system, steps 34-35 may be performed by a classifier 42, and the classification database corresponds to a second database 44, wherein the aforementioned regions are denoted by [W]i, the respective associated classifications by Gi, and the current classification by Gc.

[0058] To further promote the understanding of Method 30, Figure 5a The diagram illustrates a set of basis functions (BFs) pre-calculated for a particular deviation type by operating a PCA algorithm on eigenvectors obtained from a large number of image data of containers having deviations 403 of the particular deviation type. Thus, basis functions BF1-BF4 are the principal components produced by the PCA algorithm. In this particular example, the eigenvectors and basis functions are one-dimensional and represent the curvature at points along a line in the image data. Figure 5b The diagram shows Figure 5a The weight space is determined by the basis functions BF1 and BF2 in . Therefore, the weight space is defined by the weight W1 of BF1 and the weight W2 of BF2. Figure 5b Each point in is calculated for the deviation of an individual container detected in the image data. Figure 5b The grading regions in the weight space are also shown, where the different grading regions are assigned respective gradings II-V, where the degree of deviation decreases from II to V. Figure 5b In this particular example, the corresponding containers can be assigned a grade based on the weights W1 and W2. Figure 5bIn FIG, cross 403′ shows the result of step 34 mapping the current deviation to the weight space, where the current deviation has been detected by step 32 and has been assigned the current values ​​of weights W1, W2 by step 33. In this example, step 35 may assign the grade IV to the current deviation.

[0059] It will be appreciated that the use of basis functions provides a great deal of freedom in adapting the method 30 to classify deviations of different deviation types. For each deviation type, different definitions of the eigenvector can be tested, for example by calculating Figure 5b to achieve the desired accuracy and specificity in grading.

[0060] Back to Figure 3 Method 30 may further include step 36 of providing feedback to an operator and / or control system in the manufacturing facility. In one example, step 36 signals that the current container 401 or starting material 401' should be discarded. In another example, step 36 generates an alarm notification indicating a production error and optionally shuts down one or more machines. The alarm notification may depend on and / or indicate the current grade and / or type of deviation.

[0061] In a further example, step 36 causes or prompts the reconfiguration of one or more machines in the manufacturing plant. In one such embodiment, step 36 includes a substep of determining a timestamp for the current deviation. Optionally, this substep is performed only when the current grade exceeds a grade limit. The timestamp can be provided with reference to a master clock within the manufacturing plant. Step 36 may also include a substep of determining associated production parameters of the manufacturing plant based on the timestamp. Accordingly, when the current deviation is detected and an associated timestamp is defined, step 36 is configured to obtain production data including parameters of the production process at or before the timestamp. The production data can be obtained from a control system within the manufacturing plant. The production parameters may include any parameters associated with the production chain of the packaging container 401, such as settings and / or sensor data in one or more machines, and / or characteristics of the starting material 401' or the liquid food to be sealed therein. Step 36 may include another substep of correlating the timestamp, current grade, and deviation type with the production parameters. Through this correlation, step 36 can accurately characterize the origin and circumstances of the current deviation. This can facilitate production line optimization and provide a reliable deviation detection tool. In one embodiment, step 36 may also transmit control instructions including modified production parameters to machines in the production plant according to the current grade and / or deviation type.

[0062] Another aspect of the present disclosure relates to a method for generating a set of basis functions associated with a deviation type. Figure 6An embodiment of the method 60 is depicted in FIG. Step 61 obtains a plurality of representations of the container 401 or of the starting material 401' used to produce the container 401. The representations include deviations of a particular deviation type and may be in the form of image data, e.g. Figure 7 The reconstruction preferably depicts containers with varying degrees of deviation. Step 62 processes the deviations in the reconstruction by a basis function calculation algorithm to calculate basis functions, such as Figure 5a The illustrated basis functions BF1-BF4. Step 63 generates a set of predefined basis functions [BF] to be used by the detection method 30. For example, step 63 may involve selecting a subset of the basis functions generated by step 62, scaling the basis functions, etc. Figure 7 As shown, the generation method 60 can be performed by a dedicated basis function calculation device 70, which is configured to store the set of basis functions [BF] in the first database 43. The calculation device 70 can perform the generation method 60 for different deviation types and store the corresponding set of basis functions ([BF]i) associated with the deviation type (TYPEi) in the first database 43.

[0063] Figure 8 yes Figure 1 and 4 Monitoring device 301 or Figure 7 81 . A block diagram of an example structure of a computing device 70 is shown. In the example shown, the device 301 / 70 includes a processing system 80 and a memory 81. The processing system 80 may include any commercially available processing device, such as a CPU, DSP, GPU, microprocessor, ASIC, FPGA, or other electronic programmable logic device, or any combination thereof. The processing system 80 may be configured to read executable computer program instructions 81A from the memory 81 and execute these instructions to control the operation of the device 301 / 70, for example, to perform any of the methods described herein. The program instructions 81A may be provided to the device 301 / 70 on a computer-readable medium 85, which may be a tangible (non-transitory) product (e.g., magnetic media, optical disk, read-only memory, flash memory, etc.) or a transient product, such as a propagating signal. The memory 81 may be, for example, one or more of a buffer, flash memory, a hard drive, a removable medium, volatile memory, non-volatile memory, random access memory (RAM), or other suitable device. As Figure 8 As shown, the memory 81 may also store data 81B for use by the processing system 80, such as the first and second databases 43, 44. The device 301 / 70 further includes one or more communication interfaces 82 for operatively connecting the device 30 / 701 to an external device via a wired or wireless connection, for example, to access the first database 43 and / or the second database 44, to receive image data from the imaging device 302, to provide alarm notifications and / or control instructions, etc.

Claims

1. A method for detecting deviations in packaging containers of liquid food in a manufacturing plant, the method comprising: Obtaining image data of the packaging container; analyzing the image data to detect current deviations; processing the current deviation with respect to a set of basis functions associated with a deviation type of the current deviation to obtain a current set of weights representing the current deviation; as well as determining a current ranking of the current deviation based on the current group weight; The method further comprises: determining a timestamp of the current deviation; determining associated production parameters of the manufacturing plant based on the timestamp, and associating the timestamp, the current grade and the deviation type with the production parameters.

2. The method according to claim 1, further comprising: The current set of weights is mapped to a ranking database that associates weight combinations with a ranking of the deviation type for the current deviation, wherein a current ranking of the current deviation is determined based on the mapping.

3. The method according to claim 1 or 2, further comprising: The current deviation is processed to determine the deviation type.

4. The method according to claim 1 or 2, further comprising: The set of basis functions is obtained from a basis function database according to the deviation type.

5. The method according to claim 1 or 2, further comprising: A current eigenvector representing the current deviation is determined; and the current set of weights is calculated based on the current eigenvector and the set of basis functions.

6. The method according to claim 5, wherein the calculating the current group weight comprises: A projection of the current eigenvector onto the set of basis functions is determined; and the current set of weights is determined based on a scalar value of the projection.

7. The method according to claim 1 or 2, wherein the basis functions are linearly independent and / or mutually orthogonal.

8. The method according to claim 1 or 2, wherein the basis functions correspond to principal components given by principal component analysis.

9. The method according to claim 1, further comprising: Control instructions including modified production parameters according to the current grade and / or the type of deviation are transmitted to machines in the manufacturing plant.

10. The method according to claim 1 or 2, further comprising: An alert notification is caused based on the current rating.

11. The method according to claim 1 or 2, wherein the deviation type includes any of the following: wrinkles in the material of the packaging container, unsealed flaps of the packaging container, tears or cloudy holes in the packaging container, dents or bulges in the packaging container, delamination of the packaging container, and defective patterns and / or colors and / or holographic or metallized films on the surface of the packaging container.

12. A method for generating a set of basis functions associated with a deviation type, for use in a method according to any one of claims 1 to 11, the method comprising: obtaining a representation of the packaging container, the representation including deviations of the deviation type; processing the deviations through a basis function calculation algorithm to obtain one or more basis functions; as well as The set of basis functions is generated based on the one or more basis functions.

13. A computer-readable medium comprising computer instructions, which, when executed by a processor, cause the processor to perform the method according to any one of claims 1 to 12.

14. A system for detecting deviations in packaging containers of liquid food in a manufacturing plant, the system comprising a processor configured to execute the method according to any one of claims 1-12.

Images