Method, computer program product and system for identifying recyclable products

Abstract

A method of uniquely identifying a target recyclable product in a continuous feed of mixed recyclable products, comprising the steps of: capturing a digital image of the recyclable product; creating a trained database of the digital images of the recyclable products; identifying the recyclable product present in the digital image; and matching information in the product database to the identified image of the target recyclable product.

Description

[0001] This application is a divisional application of Chinese Patent Application No. 201980034400.5, filed on November 20, 2020, entitled "Recycling Method and Label for Recyclable Products". Technical Field

[0002] This invention relates to a recycling method and taggant for recyclable products or materials. In particular, this invention relates to a recycling method and taggant for recyclable products that can be used to quickly and easily identify the original manufacturing source of post-consumer materials. This enables manufacturers to recycle known materials and rheological resources, achieving true closed-loop recycling. Background Technology

[0003] Many food products, household / commercial cleaning products, and personal care products are packaged in single-use polymer packaging for ease of transport and use.

[0004] In recent years, it has become clear that traditional recycling methods need to evolve into a circular economy approach that incorporates "production and reuse" rather than just "production and disposal." One area of ​​immediate concern is single-use post-consumer / industrial polymer-synthesized packaging. Recent public and governmental pressure is now driving the search for a more sustainable solution to this seemingly ever-growing problem.

[0005] Currently, the recycling of plastic bottles is not defined according to the manufacturer's polymer grade and color, thus limiting further recycling opportunities and alignment with the circular economy. To address this issue, a feasibility study funded by WRAP UK and Innovate UK in 2014 conducted a technical and commercial feasibility study using identification technology based on fluorescent pigments applied to plastic packaging as masterbatches or pigments within the label, to achieve automated separation of various target materials such as high-density polyethylene (HDPE), polyethylene terephthalate (PET), and polypropylene (PP), thereby achieving closed-loop recycling. While adding masterbatches can be used to separate different target materials, it was found to be too costly to scale up, and there were UV stability issues when used with products with long shelf lives. Perhaps more importantly, there are food contact concerns for food packaging plastics. Additionally, there have been proposals to use RFID chips to tag products for tracking.

[0006] These markings may be damaged or unreadable during the packing process, and given the harsh environment in which recyclable products are handled in recycling facilities (due to volume and throughput), reliable recycling data has been virtually impossible to obtain to date.

[0007] Other marking systems known in the art include chemical etching. Chemical etching can be used in molds to etch machine- or human-readable patterns, markings, or codes onto the surface of molded products; however, such patterns, markings, or codes will be fixed, thus limiting the rapid changes to the data and information required by the manufacturer, filler, or brand owner, and making it highly undesirable to alter or modify them without shutting down the production mold.

[0008] The solution proposed in this invention is quite different. It enables plastic household and personal care products to be separated at Material Recycling Facilities (MRFs) and / or Plastic Recycling Facilities (PRFs) based on their original manufacturing origin, using simple and inexpensive markings or points. This allows manufacturers to recycle known materials and rheological resources, achieving true closed-loop recycling.

[0009] It is understood that, in this application, the term "manufacturer" can refer to any manufacturer throughout the product lifecycle, including manufacturers of the product and manufacturers who use the product, such as manufacturers who use the product to sell their goods (e.g., a bottler). Therefore, the term "manufacturer" includes manufacturers who directly and indirectly source products throughout the product's supply chain and lifecycle.

[0010] One object of this invention is to provide a recycling method and labeling for recyclable products that overcomes or reduces the drawbacks associated with known products of this type. This invention provides a recycling method and labeling for recyclable materials implemented by assigning one or more UV, NIR, and / or IR-readable ink colors and shapes to each manufacturer, and further assigning additional colors and shapes to the manufacturer's brand, enabling material detection via MRF and / or PRF for separation and continued recycling back to the primary or original manufacturing source. Another object of this invention is to provide traceable packaging materials and products that can be recycled through the supply chain and allow materials to be returned to their original manufacturing source for recycling, thereby ensuring process compliance with due diligence and corporate governance policies. Another object of this invention is to reduce reliance on virgin polymers while significantly increasing recycling rates and reducing costs. Another object of this invention is to provide a fully automated method for separating recyclable materials from raw materials using artificial intelligence and then recycling them back to their original manufacturing source. This invention addresses many of the problems associated with single-use plastics by strengthening manufacturers' corporate social responsibility (CSR) policies and significantly improving environmental resource efficiency. Furthermore, using this invention can ensure compliance with other legislative drivers and strategies, such as Extended Producer Responsibility (EPR), Packaging Recycling Certificate (PRN), and Packaging Export Recycling Certificate (PERN). Summary of the Invention

[0011] The invention is described herein and in the claims.

[0012] According to the present invention, a method for marking products with machine-readable code is provided, the method comprising the following steps: Create a database of trained digital images of labeled products; Applying machine-readable code to at least a portion of a product or its packaging; Read and verify the code applied to the product; The product is placed under excitation conditions, causing machine-readable code to fluoresce, allowing for the recovery of the machine-readable code; and The system captures a first image of the fluorescent shape or color of machine-readable code and matches the captured first image against a trained database to allow at least the identification of the product's manufacturer or brand.

[0013] One advantage of this invention is that it can be used to identify post-consumer materials from blended raw materials based on the product's manufacturer or brand, and can also acquire and verify data to ensure regulatory compliance, and / or track product consumption and lifecycle, and / or identify patterns, trends and associations, and monitor sales and marketing activities and promotional events.

[0014] More preferably, the method further includes the following steps: The data recovered from the machine-readable code is associated with the manufacturer or brand of the identified product; and The associated data, along with timestamps and / or tracking information and / or product metadata, is securely stored in a remote database or cloud-based portal.

[0015] In use, this method may also include the following steps: Based on the fluorescent shape or color of the detected machine-readable code, the product is separated from the mixed raw materials for further recycling.

[0016] Preferably, the machine-readable code is a 1D, 2D, or 3D barcode, a data matrix, a QR code, or any other suitable encoding structure.

[0017] More preferably, the machine-readable code is excited using radiation with an excitation wavelength in the UV, IR, NIR, or visible light spectrum.

[0018] In use, the recovery of machine-readable code, as well as the shape or color of fluorescence, can be detected in the same or different optical detectors at the same or different excitation wavelengths.

[0019] Preferably, the recovery data in the machine-readable code includes production data and / or PRN and / or PERN and / or EPR compliance information.

[0020] More preferably, the method further includes the following steps: A second image of the product's shape is captured and matched against a trained database to allow identification of at least the product's manufacturer or brand. This database looks for remnants of the product label and / or the product's shape and color.

[0021] In use, the second image is captured at an excitation wavelength different from that of the first image.

[0022] Preferably, the machine-readable code is a 2D data matrix that emits red or orange fluorescence upon UV excitation.

[0023] Furthermore, according to the present invention, a system for tracking products marked with machine-readable code is provided, comprising: A product database is configured to associate a product with a unique machine-readable code applied to at least a portion of the product or its packaging, the database containing timestamps and / or tracking information and / or product metadata; A detection device is used to simultaneously place a product under an excitation condition to cause the machine-readable code to fluoresce, thereby allowing the machine-readable code to be read using a barcode reader; and A first camera device is configured to capture a first digital image of the fluorescent shape or color of the machine-readable code, and to match the captured first image with one of a plurality of digital images of the tagged product to allow identification of at least the product's manufacturer or brand; and A means for automatically updating a product database using timestamps and / or tracking information and / or product metadata at one or more stages of a product's lifecycle.

[0024] Preferably, the detection device further includes: A second camera device is used to capture a second digital image of the shape of the product and match the captured image with a trained database to allow at least the identification of the product's manufacturer or brand, the database being matched with remnants of the product label and / or the product's shape and color.

[0025] Similarly, according to the present invention, a method for uniquely identifying a product for subsequent recycling is provided, comprising the following steps: Mark the product’s first traceable signature on the exposed outer surface of the product and / or on the portion of the product below any card holder or label attached thereto.

[0026] One advantage of this invention is that it can be used to identify post-consumer materials based on the primary manufacturer of the product, thereby allowing the manufacturer to recycle known materials and rheological resources for upgrading and remanufacturing into new products.

[0027] Preferably, the first tracking signature is any chemical or physical marker that can be read at the detector.

[0028] More preferably, the first tracking signature is at least one ultraviolet (UV) and / or infrared (IR) readable point applied to the product using continuous inkjet printing or any other suitable marking or encoding system.

[0029] In use, the at least one readable dot may be a transparent fluorescent marker and may only be detected when it is illuminated by UV, NIR and / or IR light at the detector.

[0030] Preferably, the at least one readable dot is printed in pairs on generally opposite surfaces of the product.

[0031] More preferably, the at least one readable dot is printed randomly on the surface of the product.

[0032] In use, this fluorescent marker can be applied as a luminescent or fluorescent ink.

[0033] Preferably, the applied fluorescent marker has a base layer that contacts the product; a fluorescent layer on top of the base layer; and an uppermost protective layer on top of the fluorescent layer.

[0034] More preferably, the base layer, fluorescent layer, and top protective layer are applied by continuous online inkjet printing or any other suitable marking or coding system.

[0035] In use, the base layer can be opaque, and when used with products that are essentially transparent, it eliminates false detection.

[0036] Preferably, the fluorescent label is completely removed during subsequent recovery processes.

[0037] More preferably, the fluorescent marking will not obscure the brand and / or product information on the product.

[0038] In use, the first tracking signature can be a printed dot in one of a variety of shapes and colors that can be detected by the detector.

[0039] Preferably, the printed dots have triangles, squares, rectangles, pentagons, hexagons, octagons, cylinders or any suitable polygonal shape, or vertical or horizontal lines or bands.

[0040] More preferably, the first tracking signature can be detected based on its shape and / or visible color and / or alphanumeric identifier.

[0041] In use, the first trace signature can be applied to the product and / or the product cap or closure and / or the removable tear strip located between the product and the cap or closure.

[0042] Preferably, the first tracking signature is applied to a printed label, which is then affixed to the product.

[0043] More preferably, the label also includes the manufacturer's name and / or a RAL or Pantone color code representing the manufacturer of the product.

[0044] In use, the first trace signature can be applied as a masterbatch or polymer carrier to components of the product in particulate, liquid or powder form, and provided through gravimetric analysis or other compatible filling processes.

[0045] More preferably, the product is packaging.

[0046] In use, the packaging may be formed from materials selected from the group including, but not limited to, any of the following: polymers, cardboard, paper, cellophane, ferrous and non-ferrous metals, composite alloys, etc.

[0047] Preferably, the method further includes the following steps: A second traceable signature representing the product's brand or ingredients is marked on the product surface.

[0048] More preferably, the first tracking signature and the second tracking signature are detected separately.

[0049] In use, this method also includes the following steps: Multiple trace signatures are marked on the surface of the product, representing the manufacturing source and / or the base polymer manufacturer and / or the polymer material and / or the material grade and / or the product brand, and enabling subsequent separation of the product based on the detected product attributes.

[0050] More preferably, the multiple tracking signatures are printed as readable dot strings, or as 1D, 2D or 3D data matrices, barcodes or QR codes, or any other suitable industrial letter, number or alphanumeric encoding process.

[0051] Preferably, the readable dot string is printed with registration marks.

[0052] In use, the detector can detect the presence of illuminated UV and / or IR light and / or near-infrared and / or visible light and / or shape or pattern recognition.

[0053] Furthermore, according to the present invention, a recyclable product is also provided, the product including a mark on its outer surface, the mark being a first traceable signature representing the manufacturer of the product.

[0054] Similarly, according to the present invention, a method for detecting uniquely marked products for subsequent recycling is provided, comprising the following steps: Use a detector to read the outer surface of the product; and The detection process involves identifying the first traceable signature of the product's manufacturer.

[0055] Furthermore, according to the present invention, a method for closed-loop recycling of a target product, the target product being marked with a first tracking signature representing the manufacturer of the product, includes the following steps: The first tracking signature on the outer surface of the product is detected, and the detected target product is separated from the mixed raw materials based on the detection. Optionally, the target product may be further divided into subgroups based on its brand or ingredients; Cut the separated product into pieces; Wash the fragment; The washed fragments were blended; and New products are formed from the blended granules.

[0056] Furthermore, according to the present invention, a label for attaching to recyclable products is provided, the label having a first tracking signature printed on its outer surface, the first tracking signature representing the manufacturer of the product.

[0057] Furthermore, according to the present invention, a method for uniquely identifying a target recyclable product in a continuous feedstock of mixed recyclable products is provided, comprising the following steps: Capture digital images of the recyclable products; Create a trained database of the digital images of the recyclable products; Identify recyclable products present in digital images; and The information in the product database is matched with the images of the target recyclable products.

[0058] Preferably, the method further includes the following steps: The target recyclable product is separated from the raw materials for subsequent recycling.

[0059] More preferably, the step of separating the target recyclable product from the raw material for subsequent recycling is carried out at a conveyor detection speed of less than about 1 m / s up to 3 m / s and above.

[0060] In use, the target recyclable product can be separated from the raw material based on the manufacturer or brand of the target recyclable product.

[0061] More preferably, a neural network is used to implement the training and recognition steps.

[0062] Furthermore, according to the present invention, a computer program product is provided for uniquely identifying a target recyclable product in a continuous feedstock of mixed recyclable products, comprising: Computer program apparatus for capturing digital images of the recyclable product; A computer program apparatus for creating a trained database of digital images of the recyclable products; Computer program apparatus for identifying recyclable products present in digital images; and A computer program device for matching information in a product database with identified images of target recyclable products.

[0063] Furthermore, according to the present invention, a system is provided for uniquely identifying a target recyclable product in a continuous feedstock of mixed recyclable products, comprising: A device for capturing digital images of the recyclable product; A means for creating a trained database of digital images of the recyclable products; A device for identifying recyclable products present in digital images; and A device for matching information in a product database with images of identified target recyclable products.

[0064] Preferably, the system further includes: A device for separating the target recyclable product from the raw materials for subsequent recycling.

[0065] According to the present invention, a system for identifying recyclable products in a mixture of recyclable raw materials is provided, comprising: A processor is configured to learn from training data containing an image library of pre-consumer recyclable products and post-consumer recyclable products, such that the processor recognizes one or more of the brand, logo, shape, and geometric features of the recyclable product to be identified; and Machine vision cameras are positioned near the transported mixed recyclable raw materials to capture digital images of the transported recyclable raw materials and match the identified digital images with the recyclable products to achieve Extended Producer Responsibility (EPR) or Deposit Refund Scheme (DRS) compliance.

[0066] According to the present invention, an artificial intelligence-enabled vision system is provided, deployed above a conveyed mixture of recyclable product raw materials, for visual identification of different recyclable products. The vision system acquires image data of the recyclable products to be identified, the image data including one or more of the brand, logo, shape, and geometric features of the recyclable products to be identified. The vision system is trained and tested to classify the acquired image data to identify the different recyclable products, and the vision system collects and reports data for each identified recyclable product to achieve Extended Producer Responsibility (EPR) or Deposit Refund Program (DRS) compliance.

[0067] According to the present invention, a computer-implemented method is provided for identifying recyclable products in a mixture of recyclable product raw materials, the method comprising the following steps: A training dataset was generated from image libraries of pre-consumer recyclable products and post-consumer recyclable products; Train a neural network to identify one or more of the brand, logo, shape, and geometric features of each recyclable product to be identified; Capture digital images of the recyclable raw materials being transported; Matching the identified digital images with recyclable products; and Collect and report data on each identified recyclable product to achieve Extended Producer Responsibility (EPR) or Deposit Refund Program (DRS) compliance.

[0068] Optionally, the image library of pre-consumer recyclable products and post-consumer recyclable products is acquired from multiple sources, including images captured from an integrated camera or from external images of categorized pre-consumer recyclable products and post-consumer recyclable products.

[0069] Optionally, the training images of the post-consumer recyclable products depict damaged, crushed, or deformed recyclable products.

[0070] Optionally, the step of capturing at least a digital image of the transported recyclable product raw material is performed at a material recycling facility (MRF), plastic recycling facility (PRF), or other material processing facility.

[0071] Optionally, the steps of matching the identified digital images with recyclable products to achieve deposit refund program (DRS) compliance may further include steps of redeeming the deposit.

[0072] It is believed that the recycling method and labeling for recyclable products according to the present invention can at least solve the above-mentioned problems.

[0073] It will be apparent to those skilled in the art that variations of the invention are possible, and that the invention may be used in ways other than those specifically described herein. Attached Figure Description

[0074] The invention will now be described by way of example only and with reference to the accompanying drawings, wherein: Figure 1 This is a flowchart illustrating the steps of a method according to the present invention for recycling recyclable materials or products back to their original manufacturing source; Figure 2a and Figure 2b The illustrations show how to present the label of the present invention on the recyclable product itself, or on a separate label subsequently affixed to the product. Figure 3 This is a flowchart illustrating the steps of a method for recycling recyclable materials or products back to their original manufacturing source according to a second embodiment of the present invention, wherein the product has been additionally labeled with a second label representing the brand or polymer composition of the product. Figure 4a and Figure 4b The illustrations show how the label of the second embodiment can be presented on the recyclable product itself, or on a separate label subsequently affixed to the product. Figure 5 This describes an alignment pattern for aligning and printing multiple labels on a product. In this embodiment of the invention, the multiple labels represent the original manufacturing source, the product's brand, the base polymer manufacturer, the polymer composition, and the grade, and can be used to further separate recycled products based on one or more of these attributes of the product. Figure 6 How to Figure 5 Examples of various labels shown on recyclable products; Figure 7 This is a flowchart illustrating the steps of a method for recycling recyclable materials or products back to their original manufacturing source using artificial intelligence, according to a third embodiment of the present invention. Figure 8 Other examples of how the smart tracking tags or labels of the present invention can be applied to recyclable products are shown; Figure 9 This is a flowchart illustrating the steps of a method according to a fourth embodiment of the present invention, which is used to manage the flow of recyclable materials or products using acquired or recovered data, and optionally, to recover detected recyclable materials or products by using a combined optical detection system that utilizes spectral marker detection, a barcode reader and artificial intelligence to detect the shape of the recyclable product. Figure 10 A high-level schematic diagram of a data acquisition and detection unit is shown, which can be adapted into an existing conveyor system according to the present invention; Figure 11 It shows the result of Figure 10 A high-level diagram illustrating how data collected by the data acquisition and detection unit can be connected to local networks as well as remote enterprise networks or cloud-based systems; Figure 12 It shows the result of Figure 10 The optical inspection system captured the data images, which have read and verified the 2D data matrix codes applied to the outer surface of recyclable products; Figure 13 is... Figure 10 A series of example data images captured by the optical inspection system, which can use artificial intelligence capabilities to detect and identify the manufacturer of recyclable products based on the shape of spectral markers applied to the outer surface of the recyclable products. Figure 14 is from Figure 10 A series of example data images captured by the optical inspection system, which can use artificial intelligence capabilities to detect and identify the brand of recyclable products based on the shape of the detected recyclable products; Figure 15 shows a series of letter, number, and / or alphanumeric tags applied to the outer surface of recyclable products. These tags have been developed by... Figure 10 The optical inspection system in the middle uses artificial intelligence capabilities for detection and classification; and Figure 16 This further illustrates how the markings or tags of the present invention can be applied to recyclable products as 2D data matrices and alphanumeric machine-readable codes. Detailed Implementation

[0075] This invention employs a method of recycling and labeling recyclable materials, implemented by assigning one or more UV, NIR, and / or IR-readable ink colors and shapes to each manufacturer, and further assigning additional colors and shapes to the manufacturer's brand, enabling material detection via MRF and / or PRF for separation and continued recycling back to the primary or original manufacturing source. Advantageously, this invention provides traceable packaging materials and products that can be recycled through the supply chain and allow materials to be returned to their original manufacturing source for recycling, thereby ensuring process compliance with due diligence and corporate governance policies. Further advantageously, this invention reduces reliance on virgin polymers while significantly increasing recycling rates and reducing costs. Further advantageously, this invention also provides a fully automated method for separating recyclable materials from raw materials using artificial intelligence and then recycling them back to their original manufacturing source. Further advantageously, this invention strengthens manufacturers' CSR policies and significantly improves environmental resource efficiency, thereby addressing many of the problems associated with single-use plastics. Furthermore, using this invention ensures compliance with other legislative drivers and strategies, such as Extended Producer Responsibility (EPR), Packaging Recycling Certificates (PRN), and Packaging Export Recycling Certificates (PERN).

[0076] Now refer to the attached diagram, Figure 1 A method 10 according to the present invention for recycling recyclable materials or products back to their original manufacturing source is shown. Method 10 described herein is a closed-loop recycling method, and therefore those skilled in the art will understand that the following description of the method can begin at any point in the cycle. In the following description, Figure 1 Each step will be referred to as "S", followed by a step number, such as S12, S14, etc.

[0077] For illustrative purposes, method 10 begins at S12 with the production of recyclable materials, packaging, or products 100 by manufacturer 20. The term "recyclable materials, packaging, or products" should be understood to encompass any recyclable article, substance, or object. (Regarding...) Figure 1 In the illustrative method 10 described, product 100 is a blow-molded polymer bottle for containing consumables, although this is by no means limiting.

[0078] At S14, the bottle is filled. At S16, a tracking mark or tag 102 representing the original manufacturing origin is applied to the bottle. The mark 102 can then be read, and the data is sent to the cloud in preparation for pairing within MRF / PRF 26, as described below. Then at S18, the bottle is distributed from manufacturer 20 directly or through a retail network to the end consumer.

[0079] Those skilled in the art will understand that S12, S14, S16 and S18 can all occur at or be coordinated from the manufacturer’s facility 20.

[0080] After use, at S22, consumers then return the bottles via local roadside recycling, and at S24, the collected bottles are received at a Material Recycling Facility (MRF) or Plastic Recycling Facility (PRF) 26 for sorting.

[0081] At S28, method 10 involves using a detector to detect a tracking mark or tag 102 on product 100, as further described below. This is a continuous conveying process in which clearly identified bottles (i.e., bottles identified by the presence of the tracking mark or tag 102) are discharged from the conveyor using air pulses from multiple nozzles located in an adjacent conveyor configuration.

[0082] Technicians will understand that after testing step S28, bottles from a manufacturer 20 can be transported or packaged for further processing / recycling, and this can be done at MRF / PRF 26 or at secondary processing facility 30.

[0083] At MRF / PRF 26 or at secondary processing facility 30, S32 involves using standard near-infrared (NIR) detection technology to further sort those already separated bottles from manufacturer 20 into their polymer components. Figure 1 For illustrative purposes, at S32, the previously sorted bottle of a manufacturer 20 is optically sorted into, for example, one of the following three polymer types: high-density polyethylene (HDPE) 34, polyethylene terephthalate (PET) 36, or polypropylene (PP) 38.

[0084] If in Figure 3 As described in the text, it can also be further separated.

[0085] use Figure 1 In the closed-loop recycling method 10 shown, the separated streams of polymer types 34, 36, and 38 can then be granulated / shredded at S40, followed by washing / drying at S42, thereby producing free-flowing flakes suitable for further compounding or extrusion into granules. The granules are then compounded, and a small amount of the original polymer may need to be added at S44 before they can be reused as new product 100 at S12.

[0086] Technicians will understand that once the tracking mark or tag 102 has been applied, product 100 will have a permanent and unique signature that is customized for the manufacturing origin 20 of product 100.

[0087] Method 10 of the present invention allows for the detection of bottles at the recycling facility (MRF / PRF 26) via a new or modified computer file and optical array upgrade to its current near-infrared (NIR) technology. This technology can automatically redirect the bottles to a hopper or a separate packer for further reprocessing at the primary manufacturer's manufacturing facility 20, thereby creating a "bottle-to-bottle" opportunity. It can then be ensured that the primary manufacturer 20 guarantees that the polymer pellets it receives at S42 have fundamental polymer rheologies known to them prior to proceeding with reprocessing.

[0088] The mixed bottle can be separated into a single polymer type at S32 using known near-infrared sorting techniques. This technique is programmed to ignore tracking markers or tags 102 and only identify the signature of the underlying polymer compound to be separated, such as PP, HDPE, PET, polyvinyl chloride (PVC), acrylonitrile butadiene styrene (ABS), etc. (For example, in combination...) Figures 3 to 6 Alternatively, a second tracking marker or tag 102 may be used to detect the signature of the underlying polymer compound.

[0089] Figure 2a This illustrates how the tracking mark or tag 102 of the present invention can be applied to a product as a UV, NIR, and / or IR readable point 50a-50n. In a preferred embodiment, the readable point is located at... Figure 1 At point S16, continuous inkjet printing technology is applied. The term "inkjet printing" should be understood to encompass any printing or marking technique that pushes ink droplets onto paper, plastic, or other substrates, such as Valvejet, thermal inkjet, drop-on-demand, and piezoelectric inkjet. Other continuous printing or etching techniques, such as laser photonic printing or digital watermarking, may alternatively be used to apply tracking marks or labels 102 and / or printed codes 88 to product 100.

[0090] Known recycling infrastructures are based on near-infrared (NIR) detection of different plastics. Therefore, IR readability points 50a-50n can be easily integrated into existing recycling infrastructures. Currently, most waste infrastructures (MRF / PRF 26) use NIR detection equipment for material separation, and NIR / IR inks will be easily accommodated in current detection infrastructure systems, potentially requiring only minor software and / or hardware upgrades.

[0091] For IR ink formulations, the current near-infrared (NIR) detection technology in MRF / PRF 26 operates between 1300 nm and 1800 nm, which is the standard operating detection window. Customized IR ink formulations can be provided as examples below: Brand A will emit yellow fluorescence at 1300nm Brand B will emit red fluorescence at 1400nm Brand C will emit green fluorescence at 1500nm, etc.

[0092] Furthermore, NIR / IR inks are less prone to degradation than UV inks, especially in environments where dots 50a-50n may be exposed to the external environment. Additionally, some brands of fabric softeners and detergents contain UV optical brighteners in their liquid product and packaging labels. This can lead to difficult readings when using UV spectral detection. In these cases, NIR / IR dot and spectral detection can be used. When covered with excess detergent, the IR dots will not be interfered with the signal because many household care products, detergents, and fabric softeners contain UV brighteners in their formulations that would otherwise mask the UV ink marking signal. Consider using UV, NIR, and IR inks as a combined labeling approach for recyclable products 100, as described below. Figure 3 As described.

[0093] In a preferred embodiment of the invention, points 50a-50n are luminescent or fluorescent markers that are transparent to the naked eye and are detectable only when excited by ultraviolet (UV), near-infrared (NIR), and / or infrared (IR) light at the detector at S28. Those skilled in the art will understand that the detection at S28 can alternatively be incorporated into the existing near-infrared detection step (S32) of the secondary processing facility 30.

[0094] Furthermore, the dots 50a-50n that uniquely identify the manufacturer 20 of product 100 can be any chemical or physical mark that can be read by a machine or a human. In a preferred embodiment, at S16, dots 50a-50n are applied in pairs to opposite surfaces of product 100 and printed using opposite inkjet printheads (not shown), so that when the collected product 100 is processed at MRF / PRF 26, it can be detected and blown out or picked up by a robot regardless of the orientation of the product 100 on the conveyor.

[0095] In addition to the points 50a-50n that can be detected as UV, NIR, and / or IR markings, technicians will recognize that external shape, visible color, and / or alphanumeric identifiers can also be provided as a means of uniquely marking product 100. Figure 2a In the example shown, point 50a is a blue circle associated with primary manufacturer A. Point 50b is an orange square associated with primary manufacturer B. Point 50c is a red triangle associated with primary manufacturer C, and point 50n is a gray circular sector associated with primary manufacturer N, and so on. In an alternative implementation, points 50a-50n may be provided in the form of a Quick Response (QR) or 2D data matrix code 90, as described below. Figure 8 As described in further detail.

[0096] Technicians will understand that if product 100 is not detected by the UV, NIR, and / or IR detectors at S28, for example, if it falls off the conveyor before being detected at S28, the operator at MRF / PRF 26 can manually separate product 100 by simply interpreting the shape and / or color of points 50a-50n. Clearly, Figure 2a The examples shown are only a subset of a larger set of shapes and colors that can be used and are for illustrative purposes only.

[0097] Figure 2b The point 50a can also be applied to the printed label 52, which is then affixed to the product 100 during use. The printed label 52 may also include the manufacturer's name 54 and a RAL or Pantone code 56, which can be a four- or six-digit code, allowing an operator to manually identify the product 100 without having to pass it through a detector. Those skilled in the art will recognize that the preferred recycling method 10 of the present invention is a fully automated process; however, as a "failure precaution," various additional optical and / or alphanumeric and / or RAL / Pantone codes 56 may also be included to allow an operator at the MRF / PRF 26 or secondary processing facility 30 to manually identify the product 100.

[0098] Points 50a-50n can also be printed on the cap or sash of product 100, as well as on the tear strip of the cap or sash, to ensure that every part of product 100 can be detected and recycled.

[0099] In addition, the masterbatch may be used alone in certain components of product 100 (such as caps, seals, tear strips, labels, etc.) or on the outer surface of the product, depending on the requirements of product 100 or manufacturer 20, as described below.

[0100] If the tracking technology is applied only to the outer surface of the bottle, the chemical properties of the polyolefin polymer require that the substrate surface be "wet," and a suitable technology in the art for this purpose is corona discharge. When configured as points 50a-50n, the ink technology used is completely removed during the recycling process and / or discharged as a gas or residue at the temperature consistent with the polymer compounding (S44).

[0101] This tracer 102 can be identified using optical and near-infrared detection technologies in current recycling infrastructure. These detection systems are common in MRF / PRF 26 and secondary processing facilities 30 and can detect a variety of material types, including polymers.

[0102] Figure 3A second embodiment of the closed-loop recycling method 10 is shown. This second embodiment of method 10 is very similar to the first embodiment, and the corresponding features are given the same reference numerals. The difference between the second and first embodiments is that, instead of simply separating product 100 at the MRF / PRF 26 based on the first points 50a-50n of the manufacturer's origin 20, product 100 is additionally marked with second UV, NIR, and / or IR readable points 70a-70n representing the brand of product 100. The advantage of this is that the MRF / PRF 26 will have the capability to detect the first points 50a-50n according to the manufacturer 20. Product 100 from a manufacturer 20 is then packaged for further separation at the manufacturer 20's facility or secondary processing facility 30, rather than based on the specific brand of product 100 or the polymer substrate used in its formation.

[0103] Figure 3 Further details of this two-stage detection method are shown. It only describes... Figure 1 This is part of the closed-loop recycling method 10 described herein, and replaces... Figure 1 S24-S44 are shown in the diagram.

[0104] exist Figure 3 In the process, at S58, bottles, along with other local roadside recyclables, are received at MRF / PRF26 for sorting. The primary detection unit can be modified to MRF / PRF 26, and at S60, the primary detection unit detects first points 50a-50n according to manufacturer 20, but critically, it does not detect second points 70a-70n. The primary unit's main function is to eject labeled bottles regardless of polymer type, but only according to manufacturer points 50a-50n. At the output of S60, bottles from a single-source manufacturer, whether HDPE, PP, or PET, are separated and can be packaged for further recycling.

[0105] After being detected by the primary detection unit, at S62, the mixed color and polymer packs can enter the secondary recycling facility 30, where the packs are opened and then sorted into the following material groups by standard NIR detection: 1. Mixed color HDPE 2. Mixed-color PET 3. Mixed-color PP And as mentioned above... Figure 1As described in S32. Alternatively or in addition, method 10 involves using a secondary marking detection unit at S64, which will be accordingly programmed to not recognize primary manufacturer points 50a-50n, but only secondary markings 70a-70n, thereby separating single-source material groups (from S62) into brands. For example, HDPE brand A bottles will be detected and popped at S66 based on UV, NIR, and / or IR readable orange squares. Similarly, HDPE brand B bottles will be detected and popped at S66 based on UV and / or IR readable red triangles. For example, PET brand C bottles will be detected and popped at S68 based on UV and / or IR readable gray circular fan shapes, and at S71, PP bottles are stored in silos.

[0106] These separated bottles, now sorted by manufacturer and brand, are further grouped into material groups by standard NIR testing before granulation S72, washing / drying S74, and compounding S76, and then reused as new product 100. Although separation can be achieved through other properties or characteristics of the recycled product 100, the secondary testing at S64 does not need to be based solely on polymer type, but can be performed based on multiple UV, NIR, and / or IR-readable colors associated with the brand rather than the manufacturer 20, as described below. Figure 5 and Figure 6 As described.

[0107] Figure 4a This illustrates how the tracking mark or tag 102 of the present invention can be applied to product 100 as two points, namely primary manufacturer points 50a-50n and secondary points 70a-70n. In a preferred embodiment, the tracking mark or tag 102... Figure 1 Continuous inkjet printing technology was applied at point S16. Figure 4a In the example shown, primary point 50a is a blue circle, associated with primary manufacturer A. Secondary point 70a is an orange square, associated with brand A of primary manufacturer A. Secondary point 70b is a red triangle, associated with brand B of primary manufacturer A, and secondary point 70n is a gray circular sector, associated with brand C of primary manufacturer A.

[0108] Technicians will understand that if product 100 is not detected at S28 by UV, NIR, and / or IR detectors, the operator at MRF / PRF 26 will know to manually separate product 100 by simply interpreting the shape and / or color of primary points 50a-50n and secondary points 70a-70n. Similarly, Figure 4a The examples shown are only a subset of a larger set of shapes and colors that can be used and are for illustrative purposes only.

[0109] Figure 4b It is also shown that primary dots 50a and secondary dots 70a-70n can also be applied to printed label 52, which is then affixed to product 100 during use. The printed label 52 also includes the manufacturer's name 54 and a RAL or Pantone code 56, which can be a four-digit or six-digit code, and allows an operator to manually identify product 100 without having to pass it through a detector.

[0110] Figure 5 This demonstrates how the present invention uses multiple points to indicate the manufacturing source 20, the base polymer manufacturer, the polymer material, the material grade, and the brand of the product, rather than using only one or two uniquely identifiable trace markers or points, and enables subsequent separation of the recycled product 100 based on one or more attributes or characteristics of the product 100. Figure 5 This illustration shows an alignment pattern for aligning and printing a series of dots on product 100. In this embodiment of the invention, the dots represent the original source of manufacture 50a-50j, the product brand 70a-70d, the base polymer manufacturer 80a-80j, the polymer material 82a-82d, and the material grade 84a-84d. Furthermore, the recycled product 100 can be subsequently separated based on one or more attributes of product 100, such as… Figure 6 As shown.

[0111] Figure 5 A style of alignment mark 86 for aligning and printing multiple dots on product 100 is also shown. As discussed in conjunction with Figures 2 and 4, the dots may also include RAL or Pantone code 56, which may be a four-digit or six-digit code, and allows an operator to manually identify product 100 or its attributes without having to pass product 100 through a detector.

[0112] Figure 6 This shows how to configure it on product 100 when in use. Figure 5 An illustrative example of multiple points shown.

[0113] Figure 7 A third embodiment of the closed-loop recycling method 10 is shown. This third embodiment of method 10 is very similar to the methods of the first and second embodiments, and the corresponding features are given the same reference numerals. The difference between the third embodiment and the first and second embodiments is that, instead of separating product 100 at MRF / PRF 26 based on detecting primary points 50a-50n indicating the manufacturing origin 20 and then additionally detecting secondary points 70a-70n indicating the brand of product 100, product 100 is separated using artificial intelligence. The advantage of this is that it is not always necessary to apply tracking marks or tags 102 to product 100.

[0114] Figure 7 This shows more details about this fully automated detection method. It only describes... Figure 1 This is part of the closed-loop recycling method 10 described herein, and replaces... Figure 1 S24-S38 are shown in the diagram.

[0115] exist Figure 7 In step S110, the mixed bottle raw materials are received at PRF 26 or secondary processing facility 30 for sorting. Step S112 involves sorting the mixed bottle raw materials into their polymer components using standard near-infrared detection technology. For example... Figure 7 For illustrative purposes, at S112, the bottle is optically classified as, for example, one of the following three polymer types: HDPE114, PET 116, or PP 118.

[0116] In a continuous process, a single artificial intelligence (AI) unit detects pre-sorted bottles and acts as a primary detection and picking device by removing bottles according to the shape, brand, and color of a manufacturer 20. For example, at S120, HDPE bottle raw materials are then sorted into three streams based on the identification of brands A, B, and C of manufacturer A. S122 involves simultaneously sorting PET bottle raw materials into three streams based on the identification of brands D, E, and F of manufacturer A. As before, the polymer-type separated streams are then granulated / shredded before washing / drying, and the shredded fragments are compounded ( Figure 7 (Not shown in the image). Therefore, the sorting step defines the polymer groups and their physical properties, such as melt flow index, tensile strength, flexural modulus, etc., in order to upgrade and remanufacture them into new bottles.

[0117] Imagine providing an AI unit with a large number of photographic images or actual broken bottles, enabling the unit's cameras and processors to learn the geometry and characteristics of each bottle type. The unit's neural network processor then learns the key features and parameters of each bottle type. This trained neural network processor is then able to select target bottles and automatically, vertically, remove them from a conveyor at a high pick-up rate of one minute using an integrated robot as they pass beneath the camera. By capturing a large number of images of broken bottles, a more reliable selection process can be achieved.

[0118] The artificial intelligence inspection methods and systems described herein can be used to improve the inspection of relevant products by using UV, NIR, and / or IR points, and / or based on the product's geometry, size, shape, and / or logo design, brand, and alphanumeric codes. The artificial intelligence inspection methods and systems described herein can also inspect products marked with 1D, 2D, or 3D data matrices, barcodes, QR codes, or any other suitable industrial alphanumeric or alphanumeric coding processes, as described below. Figure 8 As described. Therefore, the AI ​​unit can process pre-sorted product raw materials based on the detected tracking marks or tags 102, as described above, or envision an integrated or combined detection method and system that, when the target bottle passes under a camera or detector, is able to pick out the target bottle from the mixed raw materials at a high pick-up rate of per minute based on a database of trained digital images of recyclable products and / or the detection of any optical UV / NIR or IR tracking signatures described herein.

[0119] Figure 8 This document illustrates how the tracking mark or label 102 of the present invention can be applied to recyclable product 100 in the form of a printed code 88. As described herein, such a printed code 88 can be detected at MRF / PRF 26 or secondary processing facility 30 for separation and continued recycling back to the original manufacturing source 20, and / or sorting based on one or more attributes of product 100, such as brand, polymer material, material grade, and / or color.

[0120] Figure 8 An example of a tracking mark or tag 102 is shown, which is used as printed code 88, and more specifically, is applied in the form of a 2D data matrix code 90. Figure 8 The recyclable product 100 shown on the left is an example. A 2D data matrix code 90 is applied to the product 100 using continuous inkjet printing. This is to facilitate detection using current near-infrared detection technology. Figure 8 The 2D data matrix code 90 shown is printed using UV ink and can be read using a standard NIR detector at the MRF / PRF 26 or secondary processing facility 30. Figure 8In the illustrative example shown, the UV ink emits red fluorescence during recycling to aid detection and reading. Other colors can be read, and the optical array needs to be upgraded to current near-infrared (NIR) detection technology. For example, some inks can possess both UV and IR properties. In particular, they can be excited at a lower UV spectrum for UV detection and reading of data matrix code 90, and then excited at a higher wavelength for detection via incumbent optical NIR detection at MRF / PRF 26, allowing for further separation by brand, manufacturer, etc., based on fluorescence color. Thus, separate information can be read from the same data matrix 90; that is, data stored in data matrix code 90 can be read at one excitation wavelength, and fluorescence color by manufacturer / brand can be read at a second excitation wavelength for pop-up, retrieval, and recycling, in which case only one data matrix code 90 is needed.

[0121] Those skilled in the art will understand that the data matrix 90 can also be read immediately after the product 100 is manufactured, or at any time during its transportation, use, and disposal, and at any time before the recycled product 100 is granulated / shredded. Data collected at any point in the product 100's lifecycle, when read at a detector (even a handheld reader), can be transmitted to an enterprise network or cloud-based system to provide manufacturer 20 with a large dataset that can be processed using various processing techniques to extract and transform information for further commercial use and planning during the manufacturing, transportation, distribution, use, and recycling of product 100.

[0122] Therefore, embodiments of the present invention are provided using a random pattern of printed UV / IR / NIR color data matrix codes 90 arranged on product 100. The data matrix 90 allows key analytical information to be stored and retrieved upon reading. The data matrix 90 itself is a UV / IR / NIR fluorescent patch, which can then be detected at MRF / PRF 26 by a new, existing, or upgraded optical system or via a reverse sales system for brand acquisition and continued recycling. This means that a combination of information can be stored on product 100, and a color mark can be read to retrieve the information stored in the data matrix 90. Then, by detecting the fluorescent color of the data matrix 90 itself, product 100 can be separated according to manufacturer / brand, etc.

[0123] Figure 8 The printed code 88 shown on the right may include a solid mark 94 that is generally square, with a size of approximately 30mm × 30mm on product 100. Although Figure 8Not shown, but printed codes 88 are applied in pairs to opposite surfaces of product 100. Those skilled in the art will understand that if each of the plurality of dots 92a-n forming the printed codes 88 is printed, a substantially solid mark 94 will be applied to product 100, and this mark 94 can be detected at MRF / PRF 26 or at secondary processing facility 30 to separate and continue recycling back to the original manufacturing source 20, and / or sorting based on one or more attributes of product 100, such as brand, polymer material, material grade, and / or color, as combined with... Figures 1 to 6 As mentioned above.

[0124] Furthermore, when manifested as a 1D, 2D, or 3D data matrix, barcode, QR code, or any other suitable industrial alphanumeric or alphanumeric encoding process, the printed data code 88 can be used in conjunction with color and shape identification, branding, material spectral density, or label markings as described herein to provide additional data representing one or more attributes or characteristics of product 100. This data contained in the printed code 88 may include, for example, manufacturer 20, brand, color, polymer composition, place of manufacture, date of manufacture, expiry date and / or other relevant date stamps (Julian or Gregorian format), anti-counterfeiting measures, regulatory compliance, etc., and the information contained therein can be transmitted from the MRF / PRF 26 to an enterprise network or cloud-based system when read on a detector. This dataset is extremely useful in managing the flow of recycled material 100, and primarily, the data can be used for resource planning because manufacturer 20 can, for example, quantify in near real-time how much raw material (total quantity, type, brand, etc.) it possesses at various MRF / PRF sites 26 or in secondary processing facilities 30 for subsequent reuse. In addition, the dataset contained in Smart Print Code 88 can also be used to monitor sales and marketing activities and promotions, and how they affect the consumption and lifecycle of Product 100.

[0125] Those skilled in the art will also understand that the necessary printing quality of the data matrix code 90 and the assurance that the read data is checked or verified are defined by various international standards, including ISO / IEC 15415 and ISO / IEC 16022. In a preferred embodiment, it is important that the UV / IR / NIR data matrix 90 is printed on bottle 100 as an A or B grade quality data matrix 90, as required by the manufacturer and retailer. This allows for a degree of redundancy throughout the consumer and recycling cycle, as the data matrix code 90 may drop to grade C when the label is damaged during this cycle and the product 100 reaches an optical data matrix detector / reader at MRF / PRF 26 or any other suitable location 30. The built-in verification of the data matrix 90 ensures compliance with these industry standards, which is crucial for reading and decoding the data matrix 90 to comply with PRN, PERN, and / or EPR regulations. The decoded data can be transferred from our data storage system / cloud to a blockchain network, allowing for further transfer to regulatory bodies in the manufacturer's home country and overseas.

[0126] Similarly, one can also envision an array 96 of points 92a-n with arbitrary shapes and configurations, such as... Figure 8 As shown on the left. As described above, such an array of points 92a-n allows for detection by shape and color based on manufacturer 20, brand, rheology, and color, etc., and the position and incidence of points 92a-n in the data matrix (at specific X, Y coordinates of product 100) carry additional data representing one or more attributes or characteristics of the recyclable product 100.

[0127] As described herein, the detection of the printed code 88 can be read on a standalone system or may be used in conjunction with an optical detection system for spectral label separation. The aforementioned code 88 is also colored with UV / IR / NIR to associate with brands, for example, yellow for brand A and blue for brand B, as in the preferred embodiments of the invention.

[0128] As described herein, smart tags or printed codes 88, which can be used in conjunction with standard shape and color markings for brands, manufacturers 20, etc., can also be provided by upconversion phosphors of microscopic ceramic particles that provide a color response when excited by invisible light at 980 nm. When these upconversion particles are irradiated with infrared light under NIR / IR conditions, they emit colored light visible to the human eye and the in-situ optical detector at MRF / PRF 26.

[0129] Furthermore, if FMCG / manufacturer 20 wishes to recycle their bottles / packaging 100 regardless of brand, polymer type, or color, and they only require manufacturer-specific mixed packs, it is possible to use the UV or IR data matrix code 90 in the following manner. First, the data matrix 90 is fluorescently processed in one area or location of the conveyor to recover valuable data such as geographic location, consumer habits, anti-counterfeiting, PRN, PERN, and / or EPR regulatory compliance, which will be retrieved through a cloud-based portal. Further down, in the next area or location of the conveyor, in-situ optical inspection at MRF / PRF 26 uses fluorescent blocks of squares in the data matrix code 90 for pop-up, retrieval, and recycling of manufacturer-specified colors; in these cases, only one tag 90 is required.

[0130] The present invention also provides the opportunity to further label packaged manufacturer / brand specific recyclable products 100 after sorting. This involves applying machine-readable code to packaged products 100, which emits fluorescence under excitation conditions to allow detection and recovery of coded data. The fluorescent shape or color of the machine-readable code also allows for rapid identification of the manufacturer / brand of product 100 and sharing of the data with a cloud-based portal.

[0131] Figure 9 A fourth embodiment of the invention is illustrated. The method 10 of the fourth embodiment is very similar to that of the first, second, and third embodiments, and the same reference numerals are used for corresponding features. The fourth embodiment differs from the first, second, and third embodiments in that, instead of separating the product 100 at the MRF / PRF 26 based on the detected manufacturing source 20 and / or the brand of the recycled product 100 and / or other detected identifiable trace marks or attributes, data is obtained from the product 100 in a variety of different ways, including through the use of artificial intelligence. The obtained data can be transmitted back to the manufacturer 20 in real-time or near real-time to allow the manufacturer 20 to make rational decisions regarding the return of its materials 100 to the circular economy.

[0132] The method 10 begins at S124, where a neural network processor connected to an AI-enabled camera learns key features and parameters for each product type 100. Branded and branded packaging 100 is presented to an AI-enabled bottle-shaped camera to develop an image library for pre- and post-consumer use. Furthermore, during the pre- and post-recycling stages, the provided image library will display both high-quality products and products 100 from the post-consumer and recycling stages. Those skilled in the art will understand that products in the pre- and post-recycling stages are typically damaged, crushed, and twisted; and the camera will train a neural network to recognize the brand, logo, general features, and geometry of the twisted and damaged packaging 100 of the cooperating manufacturer 20.

[0133] At S126, the manufacturer 20 or filler fills the recyclable packaging 100. In a preferred embodiment, a red or orange UV 2D data matrix 90 is applied to the visible outer surface of the product 100, and / or a visible ink data matrix 90 is consistently applied to the overlay surface (e.g., below a card sleeve or label that is often removed before recycling). Using the same process, designated spectral marking labels 102 are applied in the form of UV ink on the visible surface and / or visible ink on the overlay surface, according to brand, shape, color, letter, number, or alphanumeric code 88. All markings 102 are applied to the outer surface of the bottle 100 through an encoding process, or applied to card sleeves and labels during the printing stage, or applied by encoding the card sleeves or labels.

[0134] At S128, the data matrix 90 and the specified spectral marker tag 102 are verified. This is achieved using a machine vision camera that reads the data applied to the custom data matrix 90 and forwards that information to a cloud-based portal in a ready state for subsequent pairing at MRF and / or PRF 26 during the recycling and recovery phase (at S134).

[0135] Then, at S130, the marked product 100 is distributed to the end consumer directly from manufacturer 20 or through a retail network. Those skilled in the art will understand that S126, S128, and S130 can all occur at or be coordinated from manufacturer facility 20.

[0136] After use, at S132, the consumer then returns bottle 100 via local roadside recycling, and at S134, the collected products are received at MRF 24 or PRF 26 for sorting.

[0137] At S134, product 100 is received at MRF 26. At this point, packaging 100 is separated using a conventional mixed recyclable separation method; the polymer portions of the mixed polymers, namely HDPE, PET, and PP, are packaged at PRF 26 for further recycling. Reading and removal of brand-specific packaging can also be performed at MRF 26.

[0138] At S136, product 100 is received at PRF 26, and data is obtained from product 100. Figure 9 The detection method 10 is performed using a modifiable optical detection system 160, which utilizes spectral marker detection, a barcode reader, and artificial intelligence to detect the shape of the recyclable product 100, such as... Figure 10 and Figure 11 As further detailed in the text.

[0139] At PRF 26, the packaged goods are opened and any unwanted contaminants, such as metal, paper, and cardboard, are removed. Bottles 100 are then optically sorted into their desired single polymer streams using conventional NIR sorting technology or any other suitable separation method: natural HDPE, PET clear, and mixed-color HDPE and PP streams. These materials fall from the optical sorter onto three separate conveyors, preferably operating at speeds below 2 meters per second, but not limited to this.

[0140] At S136, the three-level detection unit 140 identifies the package 100. A UV or white light camera 152 reads the data matrix 90 of UV red, orange, or visible ink. The reading is then correlated with the data applied during the marking stage (previously at S128 at the manufacturer 20 or filler). If the data cannot be obtained from the data matrix 90 due to damage, the detection unit 140 attempts to identify the package 100 using a UV spectral marking camera 154 based on the brand-specified UV or visible ink shape or color mark or label 102 applied during or concurrently with the filling stage at the manufacturer 20. An AI-enabled camera 158 supports this information collection and processing by attempting to identify label residues or by recognizing them based on a previously trained image database; all this information is continuously transmitted to a cloud-based portal, enabling the manufacturer 20 to access the information and allowing appropriate choices to be made in the journey of the recycled package 100 back into the circular economy, such as in conjunction with... Figure 10 and Figure 11 This will be explained in further detail.

[0141] At S138, package / product 100 is retrieved. As described herein, under ambient light, white light, or UV light conditions, a robotic picker operated by machine vision or artificial intelligence retrieves the product according to brand-specified shape and color tags 102, or alphanumeric or alphanumeric codes 88, or data matrix 90. Recyclable product 100 can be retrieved from conveyor 144 using a generally vertical extraction technique that minimizes the risk of collision with non-target materials, a risk present in horizontally actuated actuator types known in the art.

[0142] The brand-specified color and shape markings 102 can also be identified by a high-speed in-situ optical sorter operating under UV or white light conditions, which allows for the acquisition of product 100. Material 100 can then be sent to manufacturer 20’s recycling and compounding partners to remove ink markings, clean, reduce size and compound it into its technical specifications for reuse in new packaging.

[0143] Figure 10A schematic diagram of a data acquisition and detection unit 140 is shown. This unit 140 can be adapted to an existing conveyor system and can be used in the MRF / PRF 26 according to the invention. The detection unit 140 forms part of a detection and data acquisition system 160, which can be connected to a local network and a remote enterprise network of the manufacturer 20 or a cloud-based system, such as... Figure 11 This is shown in further detail below.

[0144] At MRF / PRF 26, the detection unit 140 is a housing 142 located above the conveyor 144, on which the recyclable product 100 is conveyed. In a preferred embodiment, the recyclable product 100 is conveyed into a first detection area 146, and then sequentially into a second detection area 148. This is not intended to be limiting, as the order in which the shape and color tags 102 detected inside the detection unit 140, or the letter, number, or alphanumeric code 88 or data matrix 90 applied to the outer surface of the product 100, and the shape of the detected recyclable product 100 can be varied.

[0145] The inspection unit 140 uses a machine vision (optical) camera inspection system 150. In a preferred embodiment, this system will include two inspection areas: a first inspection area 146 operating under UV conditions, which can read the UV data matrix 90 and brand-specified colors and shapes 50, 70, 80, 82, 84, 102, as described herein. The machine vision camera inspection system 150 includes at least one UV or white light camera 152 that reads the data matrix 90 of UV red or orange or visible ink. In a preferred embodiment, the UV or white light camera 152 is a 2D barcode reader. The readings are then correlated with the data applied during the marking stage (previously at S128 of the manufacturer 20 or the filler). If the data cannot be obtained from the data matrix 90 due to damage, the machine vision system 150 attempts to identify the package 100 using a second UV optical detector 154 based on the brand-specified UV or visible ink shape or color 50, 70, 80, 82, 84, 102 applied at the manufacturer 20 during or concurrently with the filling stage (S126). In a preferred embodiment, the first detection area 14 is illuminated with UV light using a UV strip lamp 156.

[0146] The second detection area 148 of unit 140 also includes a machine vision (optical) camera detection system 150, which includes an AI camera system 158 for searching for label residues and possible bottle shapes and colors. Those skilled in the art will know that most labels fall off during recycling, and bottles 100 are often crushed beyond recognition, but the AI ​​video system 158 effectively acts as a fail-safe device, capable of effectively identifying the shape of broken, transported recyclable product 100, as will be described in further detail with reference to FIG14. In a preferred embodiment, the second detection area is illuminated by white diffused strip lights 160.

[0147] Although in the preferred embodiment, AI-enabled cameras 154 and 158 operate under UV light in the first detection area 146 and white diffuse light in the second detection area 148, respectively, this is by no means a limitation, as AI-enabled cameras 154 and 158 can operate within a single detection unit 140 under ambient light, white light or diffuse light, or UV light or a combination of the above spectra.

[0148] Reference Figure 11 The detection unit 140 forms part of the detection and data acquisition system 160, which can be connected to a local network and the manufacturer 20's remote enterprise network or cloud-based system. Figure 11 As schematically shown, the various camera systems 152, 154, and 158 are connected to a control unit 162 that can be housed in a housing or enclosure 164. Those skilled in the art will understand that... Figure 11 This is a schematic diagram of the hardware configuration, and for clarity, many other circuit elements are not shown.

[0149] Control unit 162 controls the power supply to the individual camera systems 152, 154, 158 and illumination systems 156, 160. Control unit 162 includes a local personal computer (PC) 166. Figure 11 As schematically illustrated, the local PC 166 receives multiple inputs from the various camera systems 152, 154, and 158 via the GigE interface switch 168. The local PC 166 can be considered as a standalone system with a CPU, memory, and peripherals, which can be used to process the data received from the various camera systems 152, 154, and 158 and output the information to the MRF / PRF 26 and / or the manufacturer 20 via multiple outputs.

[0150] Connectivity to other input / output peripherals and / or other wirelessly connected devices will be achieved using wireless transmission protocols (e.g., Wi-Fi (IEEE 802.11 standard), Bluetooth, or cellular telecommunications networks) and / or by using the Near Field Communication (NFC) protocol. Additionally, those skilled in the art will understand that the control unit 162 can be connected to other external devices via a wired network connection 172.

[0151] Updates or further content to the software used to control the detection and data acquisition system 160 can be wirelessly downloaded to a local PC 166 via a local USB port interface, or using Wi-Fi, Bluetooth, a cellular telecommunications network, or an NFC antenna, or via network connection 172. In a preferred embodiment, the control unit 162 is powered by an external AC power supply 170 and connected to a local network via connection 172. Any number of interfaces and communication protocols known in the art can be used by... Figure 11 The data collected by the detection and data acquisition system 160 can be sent in real time or near real time to a remote enterprise network or a cloud-based system (not shown).

[0152] As described above, the information obtained from the detection unit 140 will enable the manufacturer 20 and / or brand owner to gain a better understanding of consumer habits, product cycles, the geographic location of the recyclable 100, demographic strengths and weaknesses, etc., in almost real time.

[0153] The data obtained will also contribute to PRN and EPR regulatory compliance. Recently, there has been significant focus and emphasis on the greater contribution of manufacturers to waste recycling costs, which is entirely appropriate. Data will be crucial for government agencies, brand owners, and recycling and strategic partners to help quantify and understand, and importantly, demonstrate, their products' place in the supply, consumption, and recycling lifecycle.

[0154] Figure 12 It shows the result of Figure 10 The machine vision camera inspection system 150 uses a UV or white light camera and a 2D barcode reader 152 to capture data images (screenshots) that have read and verified the 2D data matrix code 90 applied to the outer surface of the recyclable product 100. Figure 12 In the example shown, camera 152 has read the data matrix 90 of red or orange or visible ink directly illuminated by UV emission or white light source 156, and it has been read, checked and verified as described above.

[0155] Figure 12The image shows a data image captured by a UV camera 2D barcode reader 152, along with resulting data obtained from a data matrix 90 on a broken card sleeve PET package 100 using red and / or orange UV inks. These red and orange UV inks have been found to significantly address the problem of UV optical brighteners, which has been observed on the labels of some products 100 and depends on residues sometimes encountered from household and personal care products, causing noise problems and color interference. Figure 12 As can be seen, the captured red or orange UV data matrix 90, as well as similar data matrices 90, provide reliable and readable data matrices 90 even when deformed, damaged, or in different directions on the conveyor 144, utilizing their built-in error correction capabilities.

[0156] In a preferred embodiment of the invention, the detection unit 140 of the pre- or post-in-situ NIR optical sorter, which can be installed within the MRF / PRF 26, incorporates multiple cameras 152, 154, and 158 in each unit 140 for use with both detection technologies. The phrase "pre-in-situ NIR optical sorter" should be understood as placing the detection unit 140 before using in-situ NIR detection technology to sort the recyclable product 100 into individual polymer components, such as... Figure 1 As shown at S28. Alternatively, it is entirely possible to do so after the product has been 100% sorted into individual polymer streams using in-situ NIR detection technology (i.e. Figure 1 (After S32) the detection unit 140 is placed, which is the meaning of the phrase "post-in-situ NIR optical sorter".

[0157] The first detection area 146 within the detection unit 140 incorporates multiple 1D and 2D barcode readers 152, preferably of an arcuate design, which operate under UV light to acquire data from 1D or 2D barcodes, QR codes, or data matrix 90 for sharing with a suitable database or cloud-based technology. Within the same first detection area 146, a UV optical detector 154 with artificial intelligence capabilities is positioned to analyze bottles or packages 100 on the conveyor 144 below it to match the bottles or packages 100 with UV shapes and colors, letters, numbers, or alphanumeric codes 50, 70, 80, 82, 84, 88, 102 specified by the manufacturer 20 or brand, as described below.

[0158] The barcode reader 152 is placed in an arc shape, where the surface is flat when the bottle 100 is initially marked (S126). When the bottle is packaged during recycling, the data matrix 90 or 2D code is not damaged, but creases or folds within the bottle or package 100 change its orientation to the left or right. Therefore, placing two or more cameras 152 within the arc shape will improve the detection rate.

[0159] The detection unit 140 can be placed before or after the optical sorting technology, or installed as part of a new processing facility, or modified within existing infrastructure.

[0160] Another implementation shows three units 140 housed in the rear optical sorter within the MRF / PRF 26, and the output is analyzed as follows: Line 1 - HDPE Natural - Food Packaging Grade (mainly baby bottles) Line 2 - PET transparent and light blue (mainly for water bottles) Line 3 - HDPE and PP colored bottles (mainly for household and personal care products) Figure 13 shows the result of Figure 10 The machine vision camera inspection system 150 captures a series of example data images using its UV optical detector 154. The UV optical detector 154 can detect and identify the manufacturer or brand of the recyclable product 100 based on the shape or color of spectral marking tags 50, 70, 80, 82, 84, 102 applied to the outer surface of the recyclable product 10. Located in the first detection area 146, the optical spectral marking detector 154 has artificial intelligence capabilities and analyzes bottles or packages 100 on the conveyor 144 below to match them against alphanumeric or alphanumeric codes within the UV shape and UV color 50, 70, 80, 82, 84, 88, 102 specified by the manufacturer or brand.

[0161] Figure 13 includes three images: a) a data image obtained from UV blue square labels or marks 102 applied to the outer surface of product 100; b) a data image obtained from UV red square labels or marks 102 applied to the outer surface of product 100; and c) a data image obtained from UV red circular labels or marks 102 applied to the outer surface of product 100. Each image shows the degree of certainty that the spectral mark 102 has been classified as a percentage.

[0162] To identify different fluorescent shapes 102, a neural network is first trained using an initial set of training images, and then the tracking markers 102, which may be deformed or damaged, are identified. In the example shown in Figure 13, a relatively small number of training images are used, and the neural network is trained using red and blue UV inks printed in circles, squares, and triangles on the outer surfaces of various test packages 100.

[0163] Figure 13 shows a large number of samples obtained by the optical spectral label detector 154 and system 160, which can distinguish most shapes 102, with very few misclassifications in the initial trial. As expected, Figure 13a (Displayed as blue square 102 on the packaging) has been correctly categorized by system 160 100% Figure 13b The red square 102 was also correctly classified by system 160, even though packaging 100 was greatly deformed. Figure 13c It has been correctly classified as red circle 102, although the certainty is less than 100%. Observation Figure 13c It is possible that the deformation of bottle 100 causes mark 102 to fluoresce at different intensities (i.e., shadows) along the straight edge inside circle 102. Any crushing or deformation of product 100 along the straight edge could explain this behavior. Even with a limited number of samples and training (the example shown in Figure 13 is for illustrative purposes only), a reliable system can be established to determine the manufacturer 20 or brand of bottle 100 based on the detected UV shape and color, or alphanumeric or alphanumeric codes.

[0164] Figure 14 shows the result of Figure 10 The machine vision camera inspection system 150 captures a series of sample data images with its AI-enabled camera 158. The AI-enabled camera 158 can use artificial intelligence capabilities to detect and identify the brand of the recyclable product 100 based on the shape of the detected recyclable product 100.

[0165] Within the second detection area 148, an AI-enabled camera 158 (hardware identical to the optical spectral label detector 154) operating under white light, diffused white light, or ambient light conditions analyzes the packaging or bottle 100 to obtain label residues, markings, or branding on the packaging and bottle, including features or geometry, and matches them against a trained image database. This part of the detection system 160 can also inspect for manufacturer- or brand-specified shapes, letters, numbers, or alphanumeric markings 102 within the visible spectrum.

[0166] In the initial trials, images were captured to verify that System 160 could identify the brand of Bottle 100 based on its appearance. Fifteen different brands of Bottle 100 were tested in this trial. After obtaining a set of images, System 160 was trained using these image samples to recognize the different brands. Once trained, the system was tested using images not used for training. Subsequently, a final set of images was captured, but with greater distortion of the various labels. These images were then used to test System 160 against images it had never seen before.

[0167] The test results for the unseen product 100 were completely accurate, and the system 160 was able to accurately identify the brand, logo, and general features and geometry of the distorted and damaged packaging 100 of the partner manufacturer 20. Figures 14a to 14d Various images of products never seen before are shown, with the identified brand and degree of certainty displayed in the lower left portion of each image. The cross-sections are the result of removing the background from the images. Those skilled in the art will note that even highly damaged product 100 has been correctly classified with a high degree of certainty.

[0168] Those skilled in the art will understand the significant advantages provided by the three-level detection unit 140 described above. The UV camera 152 continuously reads the UV data matrix 90 on the transported recyclable product 100. If data cannot be obtained from the data matrix 90 due to damage, the detection unit 140 attempts to identify the product 100 using artificial intelligence capabilities and a UV spectral marking camera 154, based on the shape or color of the brand-specified UV ink applied at the manufacturer 20 or filler. The AI-enabled camera 158 effectively acts as a fail-safe measure, efficiently identifying the shape of the broken, transported recyclable product 100 by recognizing label residue or by using a database of previously trained images of broken products. This detection method and system ensures that all products 100 are detected, and this information can be continuously transmitted to a cloud-based portal, enabling the manufacturer 20 to access it.

[0169] Those skilled in the art will understand that the current AI-enabled camera systems 154, 158 cannot decode any 2D or data matrix 90 information, and without applying shape or color spectral marking labels 102 to the outer surface of the recyclable product 100, the AI-enabled camera 158 itself cannot distinguish, for example, clear or colored PET "contract bottles" used to hold beverages and bottled water from many brands and supermarkets simply from the broken shape. However, marking the recyclable product 100 with UV / NIR / IR ink shape or color labels 50, 70, 80, 82, 84, 102 allows the product 100 to be separated by its manufacturer 20 or brand upon recycling by detecting the ink. If the tracking mark or label 102 is applied as a printed code, such as a 2D data matrix 90, more information can be stored. As described herein, the ink label 102 can combine the characteristics of both UV / IR / NIR, allowing individual information to be read from the same printed code 88 or data matrix 90, and different cameras 152, 154 ensure that the information is read correctly, and an AI-enabled camera system 158 serves as fault protection to verify the information obtained from cameras 152, 154.

[0170] Therefore, the optical detection unit 140 of the present invention allows for the reliable acquisition of data that can be reported to the brand owner in real time or near real time.

[0171] The manufacturer 20 or the bottler can then retrieve their bottles 100 from the waste stream using their assigned UV shape and color labels 102, which can be achieved using a current detection system with modified lighting or a robotic picker operating under UV light conditions.

[0172] Then, the single or mixed material packages “separated” by the manufacturer or brand will be shipped to the relevant reprocessing facility to be manufactured into new materials, enabling brand owners to enter the circular economy.

[0173] In a preferred embodiment, the detection unit 140 will contain two detection areas 146, 148 within a separate modular housing. This is by no means a limitation, as one or more additional detection areas 146, 148 can be added in a modular manner, or the first and second detection areas 146, 148 can be housed in completely different housings and connected together using known interfaces and communication protocols. For example, one unit 140 may read the UV orange / red data matrix 90 separately, while another completely independent unit 140 may read or collect the color and shape of the spectral marker tags 102 along the same conveyor 144.

[0174] As outlined above regarding Figure 13, although Figure 10 The optical detection unit 140 can use artificial intelligence capabilities to detect and identify the manufacturer of the recyclable product based on the shape or color of the spectral marker tag 102 applied to the outer surface of the recyclable product 100. However, it can also use artificial intelligence to detect the shape and color of the letters, numbers, or alphanumeric codes 88 applied to the outer surface of the product 100.

[0175] Figure 15 shows the results already... Figure 10 The detection unit 140 shown detects and sorts a series of letters, numbers, and / or alphanumeric tags 102 that have been applied to the outer surface of the recyclable product 100. Figure 15 illustrates the process by which… Figure 10 A series of example data images captured by the UV optical detector 154 of the machine vision camera inspection system 150, which has been trained to detect and identify the manufacturer or brand of the recyclable product 100 based on the detected letters, numbers and / or alphanumeric tags 102 applied to the outer surface of the recyclable product 10.

[0176] Figures 15a to 15dThese are various data images obtained from previously unseen tagged products 100, and the identified codes and degrees of certainty are shown in the lower left portion of each image. Figures 15a to 15d In the image shown, the UV optical detector 154 has correctly read and classified the UV red ink letter, number, or alphanumeric code 88.

[0177] Figure 16 This demonstrates how to use a label 102 configured to combine a UV data matrix 90 and UV alphanumeric codes 88 to mark product 100, thereby allowing accurate and repeatable detection of the manufacturer or brand of product 100, and also allowing data acquisition for recycling. Such an alphanumeric coding system can be configured along the route of a first manufacturing partner 100 designated with the letter U, whose brand is marked with the following alphanumeric code 88: Brand A–U1 Brand B–U2 Brand C-U3, etc.

[0178] The second manufacturing partner was designated with the letter W, and its brand is marked with the alphanumeric code 88: Brand A–W1 Brand B–W2 Brand C–W3, etc.

[0179] This invention is not intended to be limited to the details of the embodiments described herein, which are described only by way of example. The inventors anticipate that various substitutions, changes, and modifications can be made to the invention without departing from the spirit and scope of the invention as defined in the claims. It should be understood that features described with respect to any particular embodiment can be characterized in combination with other embodiments. Such examples include the following: As an example, although a particular implementation mentions the use of closed-loop recycling method 10 with polymer products, this is by no means intended to limit it, as any number of different types of recyclable products, packaging, materials and articles can be tagged and recycled.

[0180] It is also conceivable that the present invention, particularly providing a means for traceable packaging materials and product 100 that can be recycled through the supply chain, can be identified or detected and aggregated into a dataset, rather than being separated for recycling. Besides separating product 100 for subsequent recycling, this dataset is extremely useful to manufacturers because it can reveal patterns, trends, and correlations, especially related to the use, recycling behavior, shelf life, and lifecycle of product 100 from its manufacturing to its consumption and disposal.

[0181] All data written to, and read from, the 1D, 2D, or 3D data matrix 90, barcodes or QR codes, and / or appropriate alphanumeric or alphanumeric encoding processes 88, or from marked shape or color labels 50, 70, 80, 82, 84, 102, is encrypted and authenticated to prevent impersonation and fraud. Those skilled in the art will understand that various technologies can be used to achieve secure tracking of the product 100 throughout its lifecycle, including blockchain technology.

[0182] Data collected at any point in the product 100's lifecycle can be transmitted to an enterprise network or cloud-based system to provide Manufacturer 20 with a large dataset. This dataset can be processed using various processing technologies to extract and transform information for further commercial use and planning during the manufacturing, transportation, distribution, use, and recycling of Product 100. All communications between Manufacturer 20, MRF / PRF 26, or secondary processing facilities 30, or to strategic partners in the supply chain, and through a cloud-based portal, are encrypted and authenticated.

[0183] It is also envisioned that product 100 information can be accessed via remote application software or a user interface, which may reside on a remote computing device and / or mobile communication device and be securely connected to the portal. As those skilled in the art will know, all available means of protecting data from fraud and hacking should be in place.

[0184] The detection methods and systems described herein are not limited to separation at MRF / PRF 26 or at secondary processing facility 30 for further recycling. The invention can also be embodied in recycling centers or reverse sale systems and schemes. For example, data represented in Printed Code 88 can be read and shared to demonstrate compliance with national and international reverse sale legislation, including deposit refund systems, local authority collection containers, recycling points, and schemes.

[0185] Furthermore, although embodiments of the present invention involve illuminating and subsequently detecting the tracking marks or labels 102 and / or printed codes 88 on product 100 with light invisible to the human eye, those skilled in the art will understand that the detection methods and systems described herein can be achieved through visible markings on bottle / package 100. While the brand themes and aesthetics employed by manufacturer 20 in the course of trade to market its product 100 may indeed be altered or diluted due to the inclusion of various visible colored inks and shapes for subsequent brand / manufacturer recycling and / or printing of data matrix codes, such visible markings do not depart from the scope of the present invention.

[0186] In this regard, it is understood that the machine vision (optical) camera inspection system 150, positioned in the first inspection area 146, can be modified to operate in the visible light spectrum rather than under UV conditions. In this regard, the visible ink data matrix 90 and visible ink shapes or colors 50, 70, 80, 82, 84, 102 can be applied at the manufacturer's location during or concurrently with the filling stage (S126). The visible ink data matrix 90 and / or shape or color labels 50, 70, 80, 82, 84, 102 are applied to the outer surface of the product 100 or under a sleeve or label, which can be intentionally separated from the product 100 during recycling or detached during packaging and / or shipping. Alternatively, in its current form, the manufacturer 20 may wish to promote its green certification by affixing visible recycling markings to the product 100.

[0187] Although various different colors were envisioned for printing code 88 or data matrix 90 and / or marking shapes or labels 50, 70, 80, 82, 84, 102 (Figures 2 and 4 to 102), Figure 6 (Including exemplary examples), but using only one ink color to mark product 100, such as a red or orange UV data matrix 90 and / or combined with a red or orange UV alphanumeric or alphanumeric code 88, will allow for accurate and repeatable identification of the manufacturer or brand of product 100, while also allowing for the acquisition of additional data for recycling.

[0188] It is also envisioned that, if needed, marks, shapes, or patterns formed by primary and secondary dots 50, 70, 80, 82, 84, 102, alignment marks 86, and / or printed codes 88 can be applied to the contouring of the bottle / package 100. The term "contouring" can refer to any mechanical feature or facet located on the product 100, including such features or facets to enhance the product's use or style, such as fingerprints or dents. Furthermore, the package / bottle 100 can be marked before its brand label is applied, so that if the product label falls off during recycling, the primary and secondary dots 50, 70, 80, 82, 84, 102, alignment marks 86, and / or printed codes 88 are exposed underneath for data reading and recovery, thus facilitating recycling.

[0189] Furthermore, it is envisioned that the printed code 88 can be configured or embedded in the form of printed electronic packaging, or applied as a smart tag that uses electromagnetic induction to write and read the code 88 on the recyclable product 100.

[0190] Clause The following clauses define preferred embodiments of the present invention.

[0191] 1. A method for marking a product, the method comprising the following steps: A machine-readable code is applied to at least a portion of a product or its packaging. The machine-readable code fluoresces under excitation conditions to allow detection and recovery of coded data. The fluorescent shape or color of the machine-readable code allows detection of the product's manufacturer or brand.

[0192] 2. The method according to Clause 1, wherein the product is separated from the mixed raw materials based on the fluorescent shape or color of the detected machine-readable code for further recycling. 3. The method according to Clause 1, wherein the machine-readable code is a 1D, 2D or 3D barcode, data matrix or QR code or any other suitable encoding structure.

[0193] 4. The method according to Clause 3, wherein the machine-readable code is excited using radiation having an excitation wavelength in the UV, IR, NIR or visible light spectrum.

[0194] 5. The method according to Clause 4, wherein the detection of the encoded data and the fluorescence shape or color are detected in the same or different optical detectors at the same or different excitation wavelengths.

[0195] 6. The method described in Clause 1, wherein the recovered data includes production data and / or PRN and / or PERN and / or EPR compliance information.

[0196] 7. A method for uniquely identifying a product for subsequent recycling, comprising the following steps: The product's surface is marked with the first traceable signature representing the product's manufacturer.

[0197] 8. The method according to Clause 7, wherein the first tracking signature is any chemical or physical marker that can be read at the detector.

[0198] 9. The method according to Clause 7 or 8, wherein the first trace signature is at least one ultraviolet (UV), NIR and / or infrared (IR) readable point applied to the product using continuous inkjet printing or any other suitable marking or encoding system.

[0199] 10. The method according to Clause 9, wherein the at least one readable dot is a transparent fluorescent marker and is detectable only when it is illuminated at a detector by UV, NIR and / or IR light.

[0200] 11. The method according to Clause 10, wherein the at least one readable dot is printed in pairs on generally opposite surfaces of the product.

[0201] 12. The method according to Clause 10, wherein the at least one readable dot is printed on the surface of the product in a random manner.

[0202] 13. The method according to Clause 10, wherein the fluorescent mark is applied as a luminescent or fluorescent ink.

[0203] 14. The method according to Clause 13, wherein the applied fluorescent mark has a base layer in contact with the product; a fluorescent layer on top of the base layer; and an uppermost protective layer on top of the fluorescent layer.

[0204] 15. The method according to Clause 13, wherein the base layer, the fluorescent layer and the uppermost protective layer are applied by a continuous online inkjet printing process or any other suitable marking or coding system.

[0205] 16. The method described in accordance with Clause 14 or 15, wherein the base layer is opaque and, when used with a substantially transparent product, eliminates false detection.

[0206] 17. The method according to any one of Clauses 13 to 16, wherein the fluorescent label is completely removed during subsequent recovery.

[0207] 18. The method according to any one of Clauses 13 to 17, wherein the fluorescent marking does not obscure the brand and / or product information on the product.

[0208] 19. The method according to Clause 8, wherein the first tracking signature is a dot printed in one of a variety of shapes and colors detectable by the detector.

[0209] 20. The method according to Clause 19, wherein the printed dots have a triangle, square, rectangle, pentagon, hexagon, octagon, cylinder or any suitable polygonal shape, or a vertical or horizontal line or band.

[0210] 21. The method according to Clause 8, wherein the first tracking signature is detectable based on its shape and / or visible color and / or alphanumeric identifier.

[0211] 22. The method according to any one of the preceding clauses, wherein the first trace signature is applied to the product and / or the cap or closure of the product and / or a removable tear strip located between the product and the cap or closure.

[0212] 23. The method according to any one of the preceding clauses, wherein the first tracking signature is applied to a printed label, and then the printed label is affixed to the product.

[0213] 24. The method described in Clause 23, wherein the label also includes the manufacturer’s name and / or a RAL or Pantone code representing the manufacturer of the product.

[0214] 25. The method according to Clause 1, wherein the first trace signature is applied as a masterbatch or polymer carrier in particulate, liquid or powder form and provided by gravimetric analysis or other compatible dispensing process.

[0215] 26. The method according to Clause 25, wherein the first tracking signature is applied to the outer surface of the product.

[0216] 27. The method according to any one of the preceding clauses, wherein the product is packaging.

[0217] 28. The method according to Clause 27, wherein the packaging is formed of a material selected from the group including, but not limited to, any of the following materials: polymers, cardboard, paper, cellophane, ferrous and non-ferrous metals, composite alloys, etc.

[0218] 29. The method according to any one of the foregoing clauses further includes the following steps: A second traceable signature representing the product's brand or ingredients is marked on the product's surface.

[0219] 30. The method according to Clause 29, wherein the first tracking signature and the second tracking signature are detected respectively.

[0220] 31. The method described in Clause 7 further includes the following steps: Multiple trace signatures are marked on the surface of the product, representing the manufacturing source and / or the base polymer manufacturer and / or the polymer material and / or the material grade and / or the product brand, and enabling subsequent separation of the product based on the detected product attributes.

[0221] 32. The method according to Clause 31, wherein the plurality of trace signatures are printed as readable dot strings, or as 1D, 2D or 3D data matrices, barcodes or QR codes, or any other suitable industrial alphanumeric or alphanumeric encoding process.

[0222] 33. The method according to Clause 32, wherein the readable dot string is printed with alignment marks.

[0223] 34. The method according to Clause 8, wherein the detector detects the presence of illuminated UV and / or IR light and / or near-infrared and / or visible light and / or shape or pattern recognition.

[0224] 35. A recyclable product comprising a mark on its outer surface that is a first traceable signature representing the manufacturer of the product.

[0225] 36. A method for detecting a uniquely marked product for subsequent recycling, comprising the following steps: Use a detector to read the surface of the product; and The detection represents the first traceable signature of the product's manufacturer.

[0226] 37. A method for closed-loop recycling of a target product, the target product being marked with a first traceability signature representing the manufacturer of the product, comprising the following steps: Detect the first tracking signature, and based on the detection, separate the detected target product from the mixed raw materials; Optionally, the target product may be further divided into subgroups based on its brand or ingredients; Cut the separated product into pieces; Wash the fragment; The washed fragments were blended; and New products are formed from the blended granules.

[0227] 38. A label for attaching to a recyclable product, the label having a first trace signature representing the manufacturer of the product printed on its outer surface.

[0228] 39. A method for uniquely identifying a target recyclable product in a continuous feedstock of mixed recyclable products, comprising the following steps: Capture digital images of recyclable products; Create a database of trained digital images of recyclable products; Identify recyclable products present in digital images; and The information in the product database is matched with images of the identified target recyclable products.

[0229] 40. The method according to Clause 39 further includes the following steps: Separate the target recyclable product from the raw materials for subsequent recycling.

[0230] 41. The method according to Clause 40, wherein the step of separating the target recyclable product from the raw material for subsequent recycling is carried out at a conveyor detection speed of less than about 1 m / s up to 3 m / s and above.

[0231] 42. The method described in Clause 40 or 41, wherein the target recyclable product is separated from the raw material according to the manufacturer or brand of the product.

[0232] 43. The method according to any one of clauses 39 to 42, wherein the training and recognition steps are implemented using a neural network.

[0233] 44. A computer program product for uniquely identifying a target recyclable product in a continuous feedstock of mixed recyclable products, comprising: Computer program apparatus for capturing digital images of recyclable products; A computer program apparatus for creating a database of trained digital images of recyclable products; Computer program apparatus for identifying recyclable products present in digital images; and A computer program device for matching information in a product database with images of identified target recyclable products.

[0234] 45. A system for uniquely identifying a target recyclable product in a continuous feedstock of mixed recyclable products, comprising: Device for capturing digital images of recyclable products; A device for creating a database of trained digital images of recyclable products; Device for identifying recyclable products present in digital images; and A device for matching information in a product database with images of identified target recyclable products.

[0235] 46. ​​The system described in Clause 45 further includes: A device used to separate a target recyclable product from raw materials for subsequent recycling.

Claims

1. A method for uniquely identifying a target recyclable product in a continuous feedstock of mixed recyclable products, comprising the following steps: Capture digital images of the recyclable products; Create a trained database of the digital images of the recyclable products; Identify recyclable products present in digital images; as well as The information in the product database is matched with the images of the target recyclable products.

2. The method according to claim 1, further comprising the following step: The target recyclable product is separated from the raw material for subsequent recycling.

3. The method according to claim 2, wherein, The step of separating the target recyclable product from the raw material for subsequent recycling is carried out at a conveyor detection speed of less than about 1 m / s up to 3 m / s and above.

4. The method according to claim 2 or 3, wherein, The target recyclable product is separated from the raw material according to the manufacturer or brand of the target recyclable product.

5. The method according to any one of claims 1 to 4, wherein, The training and recognition steps are implemented using neural networks.

6. A computer program product for uniquely identifying a target recyclable product in a continuous feedstock of mixed recyclable products, comprising: Computer program apparatus for capturing digital images of the recyclable product; A computer program apparatus for creating a trained database of digital images of the recyclable products; Computer program apparatus for identifying recyclable products present in digital images; as well as A computer program device for matching information in a product database with identified images of target recyclable products.

7. A system for uniquely identifying a target recyclable product in a continuous feedstock of mixed recyclable products, comprising: A device for capturing digital images of the recyclable product; A means for creating a trained database of digital images of the recyclable products; A device for identifying recyclable products present in digital images; as well as A device for matching information in a product database with images of identified target recyclable products.

8. The system according to claim 7, further comprising: An apparatus for separating the target recyclable product from the raw material for subsequent recycling.

9. A system for identifying recyclable products in a mixture of recyclable raw materials, comprising: A processor is configured to learn from training data containing an image library of pre-consumer recyclable products and post-consumer recyclable products, such that the processor recognizes one or more of the brand, logo, shape, and geometric features of the recyclable product to be identified; and Machine vision cameras are positioned near the transported mixed recyclable raw materials to capture digital images of the transported recyclable raw materials and match the identified digital images with the recyclable products to achieve Extended Producer Responsibility (EPR) or Deposit Refund Scheme (DRS) compliance.

10. An AI-enabled vision system deployed above a conveyed mixture of recyclable product raw materials for visual identification of different recyclable products, the vision system acquiring image data of the recyclable products to be identified, the image data including one or more of the brand, logo, shape, and geometric features of the recyclable products to be identified, the vision system being trained and tested to classify the acquired image data to identify the different recyclable products, and the vision system collecting and reporting data for each identified recyclable product to achieve Extended Producer Responsibility (EPR) or Deposit Refund Program (DRS) compliance.

11. A computer-implemented method for identifying recyclable products in a mixture of recyclable product raw materials, the method comprising the steps of: A training dataset was generated from image libraries of pre-consumer recyclable products and post-consumer recyclable products; Train a neural network to identify one or more of the brand, logo, shape, and geometric features of each recyclable product to be identified; Capture digital images of the recyclable raw materials being transported; Matching the identified digital images with recyclable products; and Collect and report data on each identified recyclable product to achieve Extended Producer Responsibility (EPR) or Deposit Refund Program (DRS) compliance.

12. The computer-implemented method according to claim 11, wherein, The image library of pre-consumer recyclable products and post-consumer recyclable products is acquired from multiple sources, including images captured from integrated cameras or from external images of categorized pre-consumer and post-consumer recyclable products.

13. The computer-implemented method according to claim 11 or 12, wherein, The training images of post-consumer recyclable products depict damaged, crushed, or deformed recyclable products.

14. The computer-implemented method according to any one of claims 11 to 13, wherein, At least the step of capturing digital images of the transported recyclable product raw materials is carried out at a Material Recycling Facility (MRF), Plastic Recycling Facility (PRF), or other material processing facility.

15. The computer-implemented method according to any one of claims 11 to 14, wherein, The steps to match the identified digital images with recyclable products to achieve deposit refund program (DRS) compliance further include steps to redeem the deposit.

Images