WO2023073732A1

WO2023073732A1 - A system and method for automated inline inspection of blister/strip packages

Info

Publication number: WO2023073732A1
Application number: PCT/IN2022/050942
Authority: WO
Inventors: Sudeep SUNDARAM; Anupriya BALIKAI
Original assignee: Spookfish Innovations Private Limited
Priority date: 2021-10-25
Filing date: 2022-10-22
Publication date: 2023-05-04

Abstract

The present invention discloses a system and method for automated inspection of blister/strip packages (102). Stored attributes of ideal sealed packages are used to create an AI/ML model. One or more sensor(s) (122) are appropriately placed after rollers (112) positioned along a conveyor (110) of a production unit (104). The sensor(s) (122) capture images of a top foil (304) and/or a bottom foil (306) respectively, as well as attributes such as heat signatures of the captured images. Subsequently, the recorded attributes are compared with the stored attributes to determine the integrity of the sealed packages. In case the sealed package is determined to have integrity, the sealed blister/strip packages are passed along the conveyor (110) towards further packaging steps. Defective/damaged blister/strip packages are rejected and expelled into a rejection tray (114).

Description

A system and method for automated inline inspection of blister/strip packages

The field of invention generally relates to an automated inline inspection of blister/strip package seals, and, more specifically, the field of invention relates to a system and method of using heat sensors to determine integrity of the blister/strip package seals and to identify any damages in the blister/strip package seal.

The Blister and strip packaging are the easiest way of packaging products such as edible products and pharmaceutical products, where food products, tablets, capsules or pills are packed in a convenient and controlled manner. Blister/strip packaging is typically highly resistive to ingresses of water, air, or environmental contamination, and provides long-lasting and damage-proof cover. However, the integrity of the blister package seal is vital to retain the quality of the edible or pharmaceutical products placed inside the blister/strip packages.

During blister packaging of products, the application of very high heat and pressure may damage foils. Further, a good knurl pattern in the blister packaging may not be obtained, thereby resulting in poor seal integrity. Generally, package seal inspection tests are grouped under destructive method and non-destructive methods.

Conventional non-destructive tests may not identify microscopic integrity defects, which in turn degrade the quality of products inside the blister packages. Further, existing techniques are not employed to detect defects or damages in the blister packaging in parallel with sequential movement of blister packages in a production line. Most of the conventional techniques are offline where the blister packages are tested after passing out from the production line.

The destructive methods of testing integrity in the seal are common and inexpensive but ineffective and inefficient. Some systems use methylene blue based vacuum tests, which is a subjective test and a destructive inspection method.

Other systems use a helium-based test which is a non-destructive test that provides quantitative and accurate values of leakage measurements. However, the helium-based test is a time consuming process. Further, when compared to the methylene blue based vacuum test, the helium based test is highly expensive.

Therefore, in light of the foregoing description, there is a need for a system and method that is devoid of the problems mentioned above and provides real-time inspection of the blister and strip packages to efficiently analyze integrity of heat seals and thereby identify any defects or damages on the blister/strip package sealing.

Object of Invention

A principal object of the invention is to provide a system and method to analyze integrity of seals on blister/strip packages by processing heat signatures of the blister/strip packages.

Another object of the invention is to provide sensors to capture images of the blister and strip package surfaces to identify damaged seals.

Another object of the invention is to provide an in-line inspection to test the integrity of seal of blister/strip packages in-line on a production line.

Another object of the invention is to provide a seal inspection apparatus with comprising a seal inspection module to determine heat signatures through image processing and perform calibrations of heat signature values.

Yet another object of the invention is to provide a machine learning framework to create a model of an ideal package seal.

Yet another object of the invention is to identify heat seal defects on the blister/strip package sealing.

Yet another object of the invention is to identify any deviation in heat signature values of the captured images from the heat signature values of an ideal seal.

Yet another object of the invention is to identify exact matches, close matches, or deviations in a match of a captured image’s heat signature values by using a created model.

Yet another object of the invention is to track defectively sealed packages and provide a signal to the packaging machine at the appropriate instant, in order to re-direct the defective packages into a rejection bin.

This invention is illustrated in the accompanying drawings, throughout which, like reference letters indicate corresponding parts in the various figures.

The embodiments herein will be better understood from the following description with reference to the drawings, in which:

Fig. 1

depicts/illustrates an overview of a system for automated inline blister/strip package inspection, in accordance with an embodiment of the invention.

Fig. 2

illustrates a schematic functional block diagram of a process for identifying defects or damages on the blister/strip package sealing, in accordance with an embodiment of the invention.

Fig. 3A

depicts/illustrates schematic representation of various components in an inline blister and strip packages inspection system, in accordance with an embodiment of the invention.

Fig. 3B

depicts/illustrates the schematic representation of various components in the inline blister and strip packages inspection system, in accordance with an embodiment of the invention.

Fig. 3C

depicts/illustrates examples of defects or damages on the blister/strip package sealing determined by the inline blister/strip package inspection system, in accordance with an embodiment of the invention.

Fig. 4

depicts a flowchart illustrating a method of identifying defects or damages on the blister/strip package sealing.

Statement of Invention

The present invention discloses a system and method for automated inspection of blister/strip packages to determine the integrity of the seals.

One or more sensor(s) are suitably positioned after a sealing unit which seals the blister/strip packages. The sensors capture images of the top and/or bottom foils of the blister/strip packages moving in-line in a package sealing or package production , wherein the images capture the heat signatures of the packages. Stored attributes of ideal sealed packages are used to create an AI/ML model and is compared with the recorded attributes to determine the integrity of the sealed packages.

The degree or level of match between the recorded and stored attribute values determines the integrity of the sealed blister/strip packages. The sealed packages which are determined to have integrity are passed along the conveyor for further packaging steps whereas the defective/damaged blister/strip packages are rejected and expelled into a rejection tray.

The embodiments herein and the various features and advantageous details thereof are explained more fully with reference to the non-limiting embodiments that are illustrated in the accompanying drawings and/or detailed in the following description. Descriptions of well-known components and processing techniques are omitted so as to not unnecessarily obscure the embodiments herein. The examples used herein are intended merely to facilitate an understanding of ways in which the embodiments herein may be practiced and to further enable those of skill in the art to practice the embodiments herein. Accordingly, the examples should not be construed as limiting the scope of the embodiments herein.

The present invention discloses, by way of an illustrative embodiment, a system and method for automated inline inspection of blister and strip packages to analyse the integrity of heat seals and identify any defects or damages in the blister/strip package sealing.

The automated inline blister/strip package inspection system comprises a seal inspection apparatus. The seal inspection apparatus comprises at least one sensor which is suitably positioned after a sealing unit which seals the blister/strip packages. The sensors are configured to capture images of the top and/or bottom foils of the blister/strip packages moving in-line in a package sealing or package production unit. Further, the captured images are communicated to a seal inspection module to process heat signatures of the captured images to determine attributes of the captured images. These processed heat signatures and/or attributes are used to determine any damage or defect of the heat seals of the blister/strip packages.

In a preferred embodiment, the package sealing or package production unit may be an edible product or pharmaceutical product manufacturing unit.

The seal inspection apparatus further comprises a machine learning framework configured to train a machine learning model with heat signatures of blister/strip packages with ideal seals without defects. According to the present invention, the disclosed method of creating and training the model may be carried out by the seal inspection module by using any machine learning or artificial intelligence techniques and algorithms. Moreover, the present invention discloses well-defined data processing features that calibrate sensed attributes of the heat signature data, compare the heat signature of the blister/strip package images with ideal heat signatures in the trained model.

Further, the seal inspection apparatus may comprise a rejection tray to collect any defective or damaged blister/strip packages expelled from the system.

Throughout this description, a system and method for automatic inline inspection of blister/strip packages of edible or pharmaceutical products such as tablets or pills has been disclosed in various embodiments. The embodiments of the disclosed system and method should not be read as a limitation of this invention, and the scope of this description covers other embodiments where inspection of consumer and industrial goods based packages may be implemented.

Referring now to the drawings, where similar reference characters denote corresponding features consistently throughout the figures, there are shown preferred embodiments.

depicts/illustrates an overview of an automated inline blister/strip package inspection system 100, in accordance with an embodiment of the invention. Blister/strip packages 102 are manufactured by using a production unit 104. The rollers 112 in the production unit 104 seal the blister/strip package 102 by using one or more conventional sealing means. In an embodiment, the rollers 112 use heat sealing to produce the sealed blister/strip packages 102.

The system 100 comprises a seal inspection apparatus 106 which is used to inspect the heat seals of the sealed blister/strip packages moving along a conveyor 110 of the production unit 104. The seal inspection apparatus 106 determines the integrity of the heat seals and detects damaged sealed packages which can be rejected and displaced into a rejection tray 114 which may or may not be attached to the conveyor 110.

In an embodiment, the seal inspection apparatus 106 comprises a processor 116, memory module 118, communication module 120, sensors 122 and a seal inspection module 108.

In an embodiment, the processor 116 may comprise any data processing device, comprising but not limited to, microprocessors, application specific integrated chip, field programmable gate array, etc.

In an embodiment, the memory module 118 comprise one or more volatile and non-volatile memory components, comprising but not limited to, magnetic, flash or optically readable memory devices, EEPROM, or other data storage devices.

In an embodiment, the seal inspection apparatus 106 may communicate with the production unit 104 and any other external devices by using the communication module 120 which comprises, but is not limited to, Local Area Network (LAN), Wide Area Network (WAN), Internet, GPS, GSM, Bluetooth low energy (BLE), NFC, ZigBee, a short-range wireless communication such as UWB, a medium-range wireless communication such as WiFi or a long-range wireless communication such as 3G/4G or WiMAX, according to the usage environment.

In an embodiment, the seal inspection apparatus 106 is configured to determine the blister/strip package seal’s quality by using sensors 122 for recording images and/ or heat signatures of the sealed blister/strip packages.

In an embodiment, the sensor(s) 122 may incorporate cameras such as thermal imaging cameras, but not limited to, any infrared thermo-graphic cameras or hyperspectral cameras. In an embodiment, the sensor(s) 122 can be secured to the system 100 such that fields of view of the sensors are positioned along top and/or bottom foils of the in-line moving sealed blister/strip packages, respectively.

In an embodiment, the seal inspection module 106 is either integrated within the sensors 122/1 and 122/2, in an external device which is in continuous communication with the sensors 122/1 and 122/2 through wired or wireless manner, or can be configured within a cloud server or processor. In case the seal inspection module 108 is not comprised in a cloud server or processor, the seal inspection module 108 may comprise any data processing device, but not limited to, microprocessors, application specific integrated chip, field programmable gate array, etc.

In an embodiment, the seal inspection module 108 comprises a machine learning module 124, an image processing module 126, an analysis module 128 and a damage detection module 130, for determining the integrity of the seals and detecting defects or damages in the sealed blister/strip packages.

In an embodiment, the machine learning module 124 is used to train or create one or more models such as machine learning or artificial intelligence (ML/AI) models, to determine damages or defects in the sealed blister/strip packages. The ML/AI models (not shown in the figure) are trained or created by a machine learning framework in the machine learning module 124, by using images and/or heat signatures recorded and processed from any suitable combination of defect-free, ideal seals and defective seals.

Further, the memory module 118 may save one or more of: the trained AI/ML models; and images, heat signatures and image attributes of the recorded packages as well as the ideal packages. The memory module 118 further comprises a set of instructions to be executed by the seal inspection module 108. The instructions may be used for one or more of recording images using the sensors 122, processing the recorded images, calibrating heat signature values, feeding the heat signatures to the AI/ML model, identifying the integrity and determining defects or damages in the of sealed blister/strip packages.

Further, the instructions, when executed, may enable the seal inspection module 108 to:

receive heat signature values from the sensors 122,
perform a comparison between recorded heat signature values received from the sensors 122/2 with the heat signature values of the ideal heat signature values by using the AI/ML model,
determine a degree or level of match between the received and ideal heat signature values,
determining whether the match is within a threshold value,
determining whether to allow the inspected blister/strip packages to move along the conveyor unit 110, and
determining whether to share instructions to displace the inspected blister/strip packages into the rejection tray 114.

In an embodiment, the heat signatures of the ideal sealed packages are hereinafter referred to as ideal attributes or stored attributes. The heat signatures of the images captured from the in-line blister/strip packages are hereinafter referred to as recorded attributes.

In an embodiment, the analysis module 128 compares the recorded attributes with the ideal stored attributes. Further, a match determination module in the seal inspection module 108 indicates the degree or level of match between the recorded and stored attribute values.

In an embodiment, the degree or level of match may be based on a scale with a range of 1 to 3, wherein a scale value of 3 indicates an exact match, a scale value of 2 indicates a close match, and a scale value of 1 indicates a deviation in the recorded attribute values of the captured image from the stored attribute values in the created model.

In an embodiment, the analysis module 128 may determine whether the degree or level of match is either exact or in close range. Further, based on the output from the analysis module 128, a decision is made to either pass the sealed blister/strip package along the conveyor line 110, or to expel the rejected sealed blister/strip package into the rejection tray 114.

In an embodiment, an alert is communicated to any external devices (not shown in the figure), in case any deviations in recorded attribute values are detected by the seal inspection apparatus 106. The alert may comprise one or more of: audio or video signals, audio beacon, visual display, text message, push notification, phone calls, and the like.

illustrates a schematic functional block diagram 200 of a process for identifying any defects or damages on the blister/strip package sealing, in accordance with an embodiment of the invention. The sensors 122 may continuously capture recorded images 202 of the sealed packages. Further, the image processing module 126 may process the recorded images 202 to determine one or more recorded attributes 204.

The stored attributes 206 may be used by the machine learning module 124 to create a trained AI/ML model 208. The analysis module 128 is used to compare the recorded attributes 204 and the stored attributes 206 by using the trained ML model 208. Subsequently, the analysis module 128 indicates the degree or level of match between the sensed attributes 204 and the stored attributes 206.

Further, the analysis module 128 may determine whether the degree or level of match is either in exact or close range. Consequently, the analysis module 128 either passes the respective sealed blister/strip packages or expels defective/damaged blister/strip package into the rejection tray 114.

Figures 3A and 3B depict/illustrate a schematic representation 300 of various components in an inline blister/strip package inspection system 100, in accordance with an embodiment of the invention. The production unit 104 comprises the conveyor 110 along which the blister/strip packages 102 move in-line towards other sections (not shown in the figure) of the production unit 104. Further, heat is applied to a top foil 304 and to a bottom foil 306 using one or more rollers 112/1 and 112/2. As the heat is applied, the top foil 304 and the bottom foil 306 stick to each other by forming a knurl pattern, thereby resulting in a sealed blister/strip package 102. Subsequently, the blister/strip packages 102 are sequentially passed along the conveyor 110.

In an embodiment, the sensors 122/1 and/or 122/2 of the seal inspection apparatus 106 are suitably positioned after the rollers 112/1 and 112/2 in the production unit 104. The sensor 122/1 may be secured at a top position, as the top foils 304 of the blister/strip packages 102 are sequentially moved within a field of view of the sensor 122/1.

Similarly, the bottom foils 306 of the blister/strip packages 102 are sequentially moved within a field of view of the sensor 122/2. The sensor 122/2 is secured at a bottom position after the rollers 112/1 and 112/2. The sensors 122/1 and 122/2 capture images of the top foil 304 and the bottom foil 306, respectively. Further, recorded attributes 204 such as the recorded heat signatures are captured by using the sensors 122.

Further, stored attributes 206 are sent to the seal inspection module 108 and are compared with the recorded attributes 204. Subsequently, during comparison, in case values of the recorded attributes 204 show any deviation from the value of the stored attributes 206, the blister/strip package 102 is expelled into the rejection tray 114. Alternatively, in case the values of the recorded attributes 204 show close or exact match with the values of the stored attributes 206, the blister/strip package 102 is allowed to move along the conveyor unit 110 towards further packaging steps.

In an embodiment, the sensors 122/1 and 122/2 capture the images from the surfaces of the blister/strip package 102. The sensors 122/1 and 122/2 may be operated in a sensing region of electromagnetic wavelength that ranges from 0.7μm to 14μm.

depicts/illustrates examples of defects or damages 304/1, 304/2, 304/3 on the blister/strip package sealing determined by the inline blister/strip package inspection system, in accordance with an embodiment of the invention.

depicts a flowchart illustrating a method 400 of identifying any defects or damages on the blister/strip package sealing, in accordance with an embodiment of the invention. The method starts at step 402 where images of blister/strip packages with ideal seal are captured. Subsequently, at step 404, attributes such as heat signatures are generated for the captured images of the blister/strip package with ideal seal. At step 406, the ideal attributes from ideal sealed packages are received by the seal inspection module. The machine learning framework learns the heat signatures and attributes of the ideal sealed packages and creates a model, as depicted at step 408.

The created model is stored in the memory which is either embedded in the seal inspection module or an external memory device in continuous communication with the seal inspection module, as depicted at step 410. Further, at step 412, images of the the blister/strip package for which seal integrity is to be determined, are captured. Further, attributes such as heat signatures of the captured images are recorded, as depicted at step 414. Subsequently, the seal inspection module receives the sensed attributes, as shown at step 416.

At step 418, the sensed attributes and the saved attributes of the model are compared. Further, a match determination indicates the degree or level of match between the sensed and saved attributes, as depicted at step 420. The degree or level of match may be based on a scale range of 1 to 3, wherein a scale value of 3 indicates exact match, scale value of 2 indicates closest match, and scale value of 1 indicates deviation of the sensed attribute values when compared with the saved attributes of the created model.

At step 422, the analysis module of the seal inspection module may check whether the degree or level of match is either in exact or close range. Further, based on the output from the analysis module, a decision is made either to accept the blister/strip package and move along the conveyor line or to expel the blister/strip package into the rejection tray. The analysis module performs the function of deciding whether the blister/strip package to be accepted and allowed to move along the conveyor line, as depicted at step 424 or to reject and expel into the rejection tray, as shown in step 426.

An advantage of the present invention is that the blister/strip package seal inspection system provides a real-time monitoring of defects in blister/strip packages prepared in a production unit.

An additional advantage of the invention is that both the surfaces such as top and bottom foils of the blister/strip packages may be inspected to identify the integrity of seal between the surfaces/foils and to determine any defects in the blister/strip package seals.

A further advantage of the invention is to facilitate inspection of the blister/strip packages in line with said blister/strip package`s manufacture in a production unit.

A furthermore advantage of the invention is to provide a non-destructive inspection of blister/strip package`s seal defect through heat signatures without affecting the physical or chemical properties of objects sealed inside said packages.

A furthermore advantage of the invention is that 100% of all blister packages packaged in the said production unit are inspected for seal integrity, as opposed to the existing method of random sampling for seal integrity testing.

The foregoing description of the specific embodiments will so fully reveal the general nature of the embodiments herein that others can, by applying current knowledge, readily modify and/or adapt for various applications such specific embodiments without departing from the generic concept, and, therefore, such adaptations and modifications should and are intended to be comprehended within the meaning and range of equivalents of the disclosed embodiments.

It is to be understood that the phraseology or terminology employed herein is for the purpose of description and not of limitation. Therefore, while the embodiments herein have been described in terms of preferred embodiments, those skilled in the art will recognize that the embodiments herein can be practiced with modification within the scope of the embodiments as described here.

Claims

A seal inspection apparatus (106) for automated inline inspection of blister/strip packages (102), comprising:
a seal inspection module (108) for determining integrity of seals and detecting defects in the sealed blister/strip packages (102) in a production unit (104), wherein the seal inspection module (108) comprises:
at least one sensor (122) configured to capture recorded images (202) of the sealed packages;
an image processing module (126) configured to process the recorded images (202) to determine recorded attributes (204);
a machine learning module (124) configured to create a trained AI/ML model (208) by using the recorded attributes (204) and stored attributes (206);
an analysis module (128) configured to analyze the recorded attributes (204) and the stored attributes (206) by using the trained AI/ML model (208); and
a damage detection module (130) configured to detect any damage and expel defective blister/strip packages, based on the analysis of the recorded attributes (204) and the stored attributes (206).
The seal inspection apparatus (106) as claimed in claim 1, wherein the at least one sensor (122) is positioned near a production unit (104) such that fields of view of the least one sensor (122) is positioned along top and/or bottom foils of the sealed blister/strip packages (102) moving in-line in the production unit (104).
The seal inspection apparatus (106) as claimed in claim 1, wherein the sensor (122) comprises at least one of thermal imaging camera, infrared thermo-graphic camera, and hyperspectral camera operating in the wavelength range of 0.7 micrometers to 14.0 micrometers.
The seal inspection apparatus (106) as claimed in claim 1, comprising a machine learning framework in the machine learning module (124), wherein a trained ML/AI model (208) is trained or created by the machine learning framework, by using images and/or heat signatures recorded and processed from defect-free, ideal seals.
The seal inspection apparatus (106) as claimed in claim 1, wherein the analysis module (128) is configured to:
compare the recorded attributes (204) with the stored attributes (206);
use a match determination module to indicate a degree or level of match between the recorded attributes (204) and stored attributes (206), wherein the degree or level of match is based on a scale;
determine whether the degree or level of match is either exact or in close range; and
determine whether to either pass the sealed blister/strip package, or to expel a rejected sealed blister/strip package.
The seal inspection apparatus (106) as claimed in claim 5, wherein the degree or level of match is based on a scale with a range of 1 to 3; and
wherein a scale value of 3 indicates an exact match, a scale value of 2 indicates a close match, and a scale value of 1 indicates a deviation in recorded attribute values of the recorded image (202) from the stored attribute values in the trained ML model (124).
The seal inspection apparatus (106) as claimed in claim 1, wherein the seal inspection module (108) communicates with the production unit (104) comprising a conveyor line (110) and a rejection tray (114), to pass the sealed blister/strip package (102) along the conveyor line (110), or to expel a rejected sealed blister/strip package into the rejection tray (114), based on inputs from the analysis module (128) and the damage detection module (130).
A method for automated inline inspection of blister/strip packages (102), comprising:
determining integrity of seals of the blister/strip packages (102) and detecting defects in the sealed blister/strip packages (102);
capturing recorded images (202) of the sealed packages;
processing the recorded images (202) to determine recorded attributes (204);
creating a trained AI/ML model (208) by using the recorded attributes (204) and stored attributes (206);
analyzing the recorded attributes (204) and the stored attributes (206) by using the trained AI/ML model (208);
detecting any damage and expel defective blister/strip packages, based on the analysis of the recorded attributes (204) and the stored attributes (206).
The method as claimed in claim 7, comprising:
determining integrity of seals of the blister/strip packages (102) and detecting defects in the sealed blister/strip packages (102) in a production unit (104), by using a seal inspection module (108);
capturing recorded images (202) of the sealed packages, by using at least one sensor (122);
processing the recorded images (202) to determine recorded attributes (204), by using an image processing module (126);
creating a trained AI/ML model (208) by using the recorded attributes (204) and stored attributes (206), by using a machine learning module (124);
analyzing the recorded attributes (204) and the stored attributes (206) by using the trained AI/ML model (208), by using an analysis module (128);
detecting any damage and expel defective blister/strip packages, based on the analysis of the recorded attributes (204) and the stored attributes (206), by using a damage detection module (130).
The method as claimed in claim 8, comprising positioning the sensor (122) near a production unit (104) such that fields of view of the sensors are positioned along top and/or bottom foils of the in-line moving sealed blister/strip packages (102).
The method as claimed in claim 8, wherein the sensor (122) comprises at least one of thermal imaging camera, infrared thermo-graphic camera, and hyperspectral camera.
The method as claimed in claim 7, comprising training or creating a trained ML/AI model (208) by using images and/or heat signatures recorded and processed from defect-free, ideal seals, by using a machine learning framework.
The method as claimed in claim 7, comprising:
comparing the recorded attributes (204) with the stored attributes (206);
using a match determination module to indicate a degree or level of match between the recorded attributes (204) and stored attributes (206), wherein the degree or level of match is based on a scale;
determining whether the degree or level of match is either exact or in close range; and
determining whether to either pass the sealed blister/strip package, or to expel a rejected sealed blister/strip package.
The method as claimed in claim 7, wherein the degree or level of match is based on a scale with a range of 1 to 3; and
wherein a scale value of 3 indicates an exact match, a scale value of 2 indicates a close match, and a scale value of 1 indicates a deviation in recorded attribute values of the recorded image (202) from the stored attribute values in the trained ML model (124).
The method as claimed in claim 8, comprising passing the sealed blister/strip package (102) along a conveyor line (110), or expelling a rejected sealed blister/strip package into a rejection tray (114), based on inputs from the analysis module (128) and the damage detection module (130), by communicating with the production unit (104) comprising the conveyor line (110) and the rejection tray (114).