Sorting plastics

Abstract

Systems and methods for sorting and separating plastic materials utilizing a vision system and one or more sensor systems, which may implement a machine learning system to identify or classify each material, which may then be separated into separate groups based on such identification or classification.

Description

[Technical field]

[0001] This application claims priority to U.S. Provisional Patent Application No. 63 / 146,892 and U.S. Provisional Patent Application No. 63 / 173,301. This application is a continuation-in-part of U.S. Provisional Patent Application No. 17 / 495,291, which is a continuation-in-part of U.S. Provisional Patent Application No. 17 / 380,928, which is a continuation-in-part of U.S. Provisional Patent Application No. 17 / 227,245, which is a continuation-in-part of U.S. Provisional Patent Application No. 16 / 939,011, which is a continuation-in-part of U.S. Provisional Patent Application No. 16 / 375,675 (issued as U.S. Patent No. 10,722,922), which is a continuation-in-part of U.S. Provisional Patent Application No. 15 / 9 No. 63,755 (issued as U.S. Pat. No. 10,710,119), which claims priority to U.S. Provisional Patent Application No. 62 / 490,219, which is a continuation-in-part of U.S. Patent Application No. 15 / 213,129 (issued as U.S. Pat. No. 10,207,296), which claims priority to U.S. Provisional Patent Application No. 62 / 193,332, all of which are incorporated herein by reference. This application is also a continuation-in-part of U.S. patent application Ser. No. 17 / 491,415, which is a continuation-in-part of U.S. patent application Ser. No. 16 / 852,514, which is a divisional application of U.S. patent application Ser. No. 16 / 358,374 (issued as U.S. Patent No. 10,625,304), which is a continuation-in-part of U.S. patent application Ser. No. 15 / 963,755 (issued as U.S. Patent No. 10,710,119).

[0002] Government Licensing Rights This disclosure was made with U.S. Government support under Grant No. DE-AR0000422 awarded by the U.S. Department of Energy. The U.S. Government may have certain rights in this disclosure.

[0003] The present disclosure relates generally to solid waste sorting, and more particularly to the sorting of plastic debris from municipal or industrial solid waste. [Background technology]

[0004] This section is intended to introduce various aspects of technology that may be related to exemplary embodiments of the present disclosure. The discussion is believed to be helpful in providing a framework to facilitate a better understanding of certain aspects of the present disclosure. As such, it should be understood that this section is to be read in this light, and not necessarily as admissions of prior art.

[0005] Recycling is the process of collecting and treating materials that would otherwise be discarded as trash (e.g., waste streams) to turn them into new products or at least allow for more appropriate disposal. Recycling benefits local communities and the environment because it reduces the amount of waste sent to landfills, conserves natural resources such as wood, water, and minerals, increases economic security by utilizing domestic material sources, prevents pollution by reducing the need to collect new raw materials, and saves energy. After collection, recyclables may be sent to a material recovery facility ("MRF") to be sorted, cleaned, and processed into materials that can be used in manufacturing. As a result, a high-throughput automated sorting platform that economically separates highly mixed waste streams would be beneficial across a variety of industries. Thus, there is a need for a cost-effective sorting platform that can identify, analyze, and separate mixed industrial or municipal solid waste streams at high throughput and economically generate high-quality feedstock (sometimes with low levels of trace contaminants) for subsequent processing. MRFs typically cannot distinguish between many materials, limiting the market for sorted materials to lower quality, lower prices, or they are too time-consuming, labor-intensive and inefficient, limiting the amount of material that can be economically recycled or recovered.

[0006] Municipal solid waste (MSW) is a broad term that refers to waste streams that include household, commercial, and industrial sources. Within each of these categories, there are thousands of different materials and products. The EPA reported that 267.8 million tons of MSW were generated in 2017. 35.37 million tons, or 13.2% of that MSW’s total weight, consisted of plastic. Of the 35.37 million tons of plastic, 2.96 million tons (8.4%) were recycled, 5.59 million tons (15.8%) were combusted with energy recovery, and 26.82 million tons (75.8%) were landfilled. It is clear that more recycling of plastics is needed.

[0007] Plastic recycling is the reprocessing of plastic waste to make new, useful products. Recycling is necessary because almost all plastics are non-biodegradable and accumulate in the environment. Currently, almost all recycling is done by remelting used plastics and regenerating them into new items: so-called mechanical recycling. This can cause degradation of the polymers at a chemical level, and also requires the plastic waste to be sorted by both color and polymer type before reprocessing, which is complex and expensive. Failing this can lead to unstable material properties that are unattractive for the industry. An alternative approach, known as feedstock recycling, allows plastic waste to be reduced to its original chemicals, which can then be reprocessed back into new plastics. This promises greater recycling, but comes at the expense of higher energy and capital costs. Plastic waste can also be burned instead of fossil fuels as part of energy recovery.

[0008] Currently, only some plastics are recyclable. When plastics are recycled, they are usually separated into different types of plastic. Recycling rates also vary by type of plastic. Several types are commonly used, each with different chemical and physical properties. This makes them easier to separate and reprocess, which affects the value and market size of the recovered material. Plastic packaging and products made of a single material (e.g., polyethylene terephthalate ("PET"), high density polyethylene ("HDPE"), polypropylene ("PP")) are more easily recycled. Plastics that are sometimes or rarely recyclable include polyvinyl chloride ("PVC"), low density polyethylene ("LDPE"), linear low density polyethylene ("LLDPE"), and polystyrene ("PS"). Furthermore, plastics can only be recycled a limited number of times.

[0009] Modern single-stream MRFs and plastic reclaimers experience high volumes of material flowing through them, necessitating processing equipment that can move and sort the material at high speeds. At the same time, the highest value is obtained from the purest, least contaminated stream. To achieve these somewhat contradictory goals, today's single-stream MRFs and reclaimers employ automated equipment that sorts plastic packaging by its near-infrared ("NIR") signature, either through transmission or reflection. These sensors rely on the reflection of light from an external source and can only see the surface of the material. Furthermore, only polymer information is obtained from this sensor. For example, NIR spectroscopy can distinguish between #1 types of plastic, which are clear and light blue PET, and #2 HDPE, while rejecting other plastics such as #1 colored PET, #3 PVC, #4 LDPE, #5 PP, #6 PS, and #7 multi-layered polymers, composite polymers, acrylic, and nylon. Furthermore, NIR spectroscopy cannot accurately identify black or highly colored plastics, or composite materials such as plastic-coated paper and multi-layered packaging (made with polymer multi-layered films), which can result in misleading readings. Most black plastics are colored using carbon. Black plastic is widely used in the automotive industry, electronics, food packaging, plastic bags, etc. However, black plastic has the unfortunate side effect of not only absorbing visible light, but also the near-infrared part of the spectrum, making it invisible to NIR spectroscopy. Thus, "stealthily" black plastic goes undetected into a "miscellaneous" bin at the end of a conveyor to be burned for energy or dumped in a landfill.

[0010] In closed-loop, or primary, recycling, waste plastics are recycled into new items of similar quality and type (e.g., beverage bottles turned back into beverage bottles). However, continuously mechanically recycling plastics without loss of quality is very challenging due to the cumulative degradation of the polymers and the risk of contaminant build-up. Closed-loop recycling has been studied for many polymers, but so far only PET bottle recycling has been industrially successful.

[0011] In open-loop or secondary recycling (also known as downcycling), the quality of plastics decreases each time they are recycled, so the material cannot be recycled forever and eventually becomes waste. Recycling plastic bottles into fleece and other fibers is a common example and accounts for the majority of PET recycling. The decrease in polymer quality can be offset by blending recycled plastic with virgin materials or compatibilized plastics when making new products.

[0012] Thermoset polymers do not melt, but techniques have been developed to mechanically recycle them, which typically involves grinding the material and then mixing it with some kind of binder to form a new composite material.

[0013] In feedstock or tertiary recycling (also called chemical recycling), polymers are reduced to their chemical building blocks (monomers) which can then be polymerized back into new plastics. Thermal depolymerization and chemical depolymerization are two types of feedstock recycling.

[0014] Energy recovery, also known as energy recycling or quaternary recycling, involves burning plastic waste to produce energy instead of fossil fuels.

[0015] Processes have been developed that allow certain types of plastics to be used as a carbon source (instead of coke) in the recycling of scrap steel. Crushed plastics may be used as construction aggregate or filler in certain applications.

[0016] Plastic waste may simply be burned as refuse-derived fuel ("RDF") in waste-to-energy processes, or it may first be chemically converted into synthetic fuels. In either approach, PVC must be excluded or compensated for by implementing dichlorination technologies, because PVC produces large amounts of hydrogen chloride (HCl) when burned, which can corrode equipment and lead to undesirable chlorination of fuel products.

[0017] Mixed plastic waste can be depolymerized to produce synthetic fuels, which have a higher calorific value than the starting plastics and can be burned more efficiently, but still less efficiently than fossil fuels. Various conversion techniques are being investigated, of which pyrolysis is the most common. The use of catalysts in pyrolysis results in more defined and higher value products. Compared to the widespread incineration, plastic-to-fuel technologies have historically struggled to be economically viable due to the high costs of collecting and sorting plastics and the relatively low value of the fuels produced. [Prior art documents] [Patent documents]

[0018] [Patent Document 1] U.S. Published Patent Application No. 2022 / 0016675 [Non-patent literature]

[0019] [Non-Patent Document 1] Krizhevsky et al., "ImageNet Classification with Deep Convolutional Networks," Proceedings of the 25th International Conference on Neural Information Processing Systems, December 3-6, 2012, Lake Tahoe, NV [Non-Patent Document 2] LeCun et al., "Gradient-Based Learning Applied to Document Recognition," IEEE Proceedings, Institute of Electrical and Electronics Engineers (IEEE), November 1998. Summary of the Invention [Problem to be solved by the invention]

[0020] As a result of the above, improved processes for separating all types of plastics, the ability to separate #3 through #7 types of plastics, the ability to separate PVC, and the ability to separate mixtures of plastics into new categories or fractions so that they can be recycled more efficiently are desired. [Means for solving the problem]

[0021] An aspect of the present disclosure provides a method that includes capturing a first visual image of a first piece of material, resulting in a first image data packet for the first piece of material, capturing a second visual image of a second piece of material, resulting in a second image data packet for the second piece of material, the first piece of material having a first chemical characteristic and the second piece of material having a second chemical characteristic different from the first chemical characteristic, processing the first and second image data packets with a machine learning system that has been previously trained to visually distinguish between pieces of material having different chemical characteristics, and classifying the first and second pieces of material into two different classifications using the machine learning system in response to the learned visual discrimination between the pieces of material having different chemical characteristics. The method may further include separating the first piece of material from the second piece of material in response to the classification. The pieces of material may be plastic pieces. The first chemical characteristic may be spectral data measured by a plurality of different sensor systems from at least one sample of plastic pieces of the same type as the first plastic piece, and the second chemical characteristic may be spectral data measured by a plurality of different sensor systems from at least one sample of plastic pieces of the same type as the second plastic piece. The spectral data may relate to the non-visible spectrum. The plurality of different sensor systems may be selected from the group consisting of near infrared ("NIR"), mid-wave infrared ("MWIR"), and x-ray fluorescence ("XRF") systems.The different sensor systems may be selected from the group consisting of infrared ("IR"), Fourier transform IR ("FTIR"), forward looking infrared ("FLIR"), very near infrared ("VNIR"), near infrared ("NIR"), short wavelength infrared ("SWIR"), long wavelength infrared ("LWIR"), mid wavelength infrared ("MWIR" or "MIR"), x-ray transmission ("XRT"), gamma ray, ultraviolet ("UV"), x-ray fluorescence ("XRF"), laser induced breakdown spectroscopy ("LIBS"), Raman spectroscopy, anti-Stokes Raman spectroscopy, gamma ray spectroscopy, hyperspectral spectroscopy (e.g., beyond visible wavelengths), acoustic spectroscopy, NMR spectroscopy, microwave spectroscopy, terahertz spectroscopy, differential scanning calorimetry ("DSC"), thermogravimetric analysis ("TGA"), capillary and rotational rheometry, optical and scanning electron microscopy ("SEM"), and chromatography. The first chemical characteristic may include measurements of organic and inorganic elements or molecules from at least one sample of plastic pieces of the same type as the first plastic piece, and the second chemical characteristic may include measurements of organic and inorganic elements or molecules from at least one sample of plastic pieces of the same type as the second plastic piece. The plastic pieces may be selected from the group consisting of Type #1 Polyethylene Terephthalate ("PET"), Type #2 High Density Polyethylene ("HDPE"), Type #3 Polyvinyl Chloride ("PVC"), Type #4 Low Density Polyethylene ("LDPE"), Type #5 Polypropylene ("PP"), Type #6 Polystyrene ("PS"), Type #7 Other polymers. The first material piece may include polyvinyl chloride. The two different classifications may be different fractions.

[0022] An aspect of the disclosure provides a system comprising a camera configured to capture a first visual image of a first piece of material resulting in a first image data packet for the first piece of material and a second visual image of a second piece of material resulting in a second image data packet for the second piece of material, the first piece of material having a first chemical characteristic and the second piece of material having a second chemical characteristic different from the first chemical characteristic, a data processing system configured to process the first and second image data packets with a machine learning system previously trained to visually distinguish between pieces of material having different chemical characteristics, the machine learning system classifying the first and second pieces of material into two different fractions in response to the learned visual distinction between the pieces of material having the different chemical characteristics, and a sorting device configured to separate the first piece of material from the second piece of material in response to the fractions. The pieces of material may be plastic pieces. The first chemical characteristic may be spectral data for the non-visible spectrum measured by a plurality of different sensor systems from at least one sample of plastic piece of the same type as the first plastic piece, and the second chemical characteristic may include spectral data for the non-visible spectrum measured by a plurality of different sensor systems from at least one sample of plastic piece of the same type as the second plastic piece. The plurality of different sensor systems may be from the group consisting of near infrared ("NIR"), mid-wave infrared ("MWIR"), and x-ray fluorescence ("XRF") systems.The different sensor systems may be from the group consisting of infrared ("IR"), Fourier transform IR ("FTIR"), forward looking infrared ("FLIR"), very near infrared ("VNIR"), near infrared ("NIR"), short wavelength infrared ("SWIR"), long wavelength infrared ("LWIR"), mid wavelength infrared ("MWIR" or "MIR"), x-ray transmission ("XRT"), gamma ray, ultraviolet ("UV"), x-ray fluorescence ("XRF"), laser induced breakdown spectroscopy ("LIBS"), Raman spectroscopy, anti-Stokes Raman spectroscopy, gamma ray spectroscopy, hyperspectral spectroscopy (such as beyond visible wavelengths), acoustic spectroscopy, NMR spectroscopy, microwave spectroscopy, terahertz spectroscopy, differential scanning calorimetry ("DSC"), thermogravimetric analysis ("TGA"), capillary and rotational rheometry, optical and scanning electron microscopy ("SEM"), and chromatography. The first chemical characteristic may include measurements of organic and inorganic elements or molecules from at least one sample of plastic piece of the same type as the first plastic piece, and the second chemical characteristic may include measurements of organic and inorganic elements or molecules from at least one sample of plastic piece of the same type as the second plastic piece, the plastic piece being selected from the group consisting of type #1 polyethylene terephthalate ("PET"), type #2 high density polyethylene ("HDPE"), type #3 polyvinyl chloride ("PVC"), type #4 low density polyethylene ("LDPE"), type #5 polypropylene ("PP"), type #6 polystyrene ("PS"), and type #7 other polymers.

[0023] Aspects of the present disclosure provide a method that includes using a plurality of different sensor systems to determine a chemical signature of each of a mixture of different plastic pieces, capturing a visual image of each plastic piece, digitally associating the chemical signature of each plastic piece with the visual image, determining a specific fraction for separating the plastic pieces, using the visual image to identify which plastic pieces in the mixture have a chemical signature that corresponds to the specific fraction, and training a machine learning system to visually identify the plastic pieces that correspond to the specific fraction, where the training is performed using a control group created from the identified plastic pieces. The control group may be composed of the captured visual image data of each of the identified plastic pieces. The fraction may be composed of a specific combination of organic and inorganic elements or molecules. The plurality of different sensor systems may be selected from the group consisting of near infrared ("NIR"), mid-wave infrared ("MWIR"), and x-ray fluorescence ("XRF") systems. The mixture of different plastic pieces may be selected from the group consisting of type #1 polyethylene terephthalate ("PET"), type #2 high density polyethylene ("HDPE"), type #3 polyvinyl chloride ("PVC"), type #4 low density polyethylene ("LDPE"), type #5 polypropylene ("PP"), type #6 polystyrene ("PS"), and type #7 other polymers.The different sensor systems may be selected from the group consisting of infrared ("IR"), Fourier transform IR ("FTIR"), forward looking infrared ("FLIR"), very near infrared ("VNIR"), near infrared ("NIR"), short wavelength infrared ("SWIR"), long wavelength infrared ("LWIR"), mid wavelength infrared ("MWIR" or "MIR"), x-ray transmission ("XRT"), gamma ray, ultraviolet ("UV"), x-ray fluorescence ("XRF"), laser induced breakdown spectroscopy ("LIBS"), Raman spectroscopy, anti-Stokes Raman spectroscopy, gamma ray spectroscopy, hyperspectral spectroscopy (e.g., beyond visible wavelengths), acoustic spectroscopy, NMR spectroscopy, microwave spectroscopy, terahertz spectroscopy, differential scanning calorimetry ("DSC"), thermogravimetric analysis ("TGA"), capillary and rotational rheometry, optical and scanning electron microscopy ("SEM"), and chromatography. [Brief description of the drawings]

[0024] [Figure 1] FIG. 1 is a schematic diagram of a sorting system configured in accordance with certain embodiments of the present disclosure. [Diagram 2] FIG. 1 illustrates an example representation of a control set of pieces of material used in the training phase of a machine learning system. [Diagram 3] FIG. 1 illustrates a flow chart arranged in accordance with certain embodiments of the present disclosure. [Figure 4] FIG. 1 is a simplified schematic diagram configured in accordance with certain embodiments of the present disclosure. [Diagram 5] FIG. 1 shows examples of chemical features. [Figure 6] FIG. 1 shows examples of chemical features. [Figure 7] FIG. 1 illustrates a flow chart arranged in accordance with certain embodiments of the present disclosure. [Figure 8] FIG. 1 illustrates a flow chart arranged in accordance with certain embodiments of the present disclosure. [Figure 9] FIG. 1 is a block diagram of a data processing system configured in accordance with certain embodiments of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0025] Various detailed embodiments of the present disclosure are disclosed herein. However, it should be understood that the disclosed embodiments are merely exemplary of the present disclosure, which may be embodied in various alternative forms. The numerical values ​​are not necessarily to scale, and some features may be exaggerated or minimized to show the details of specific components. Therefore, the specific structural and functional details disclosed herein should not be interpreted as limiting, but merely as a representative basis for teaching those skilled in the art to use various embodiments of the present disclosure.

[0026] As used herein, "material" includes metals (ferrous and non-ferrous), alloys, plastics (including, but not limited to, those disclosed herein, known in the industry, or newly created in the future), rubber, foam, glass (including, but not limited to, borosilicate glass or soda lime glass, and various colored glasses), ceramics, paper, cardboard, Teflon, PE, bundled wire, insulated wire, rare earth elements, leaves, wood, plants, plant parts, fibers, biowaste, packaging, electronic waste, batteries and accumulators, end-of-life vehicles, mining, construction and demolition waste, agricultural crop waste, forest residues, cultivated grasses, wood energy, etc. The waste stream may include any items or objects, including, but not limited to, energy crops, microalgae, urban food waste, food waste, hazardous chemical and biomedical waste, construction debris, farm waste, objects of biogenic origin, objects of non-biogenic origin, objects with a particular carbon content, other objects that may be found in municipal solid waste, and any other objects, articles, or materials disclosed herein, including further types or classes of any of the foregoing that can be distinguished from one another by one or more sensor systems, including, but not limited to, any of the sensor technologies disclosed herein.

[0027] "Material" may include any item or object composed of a chemical element, a compound or mixture of chemical elements, or a compound or mixture of compounds or mixtures of chemical elements, where the compound or mixture may range in complexity from simple to complex. As used herein, "element" means a chemical element of the periodic table of the elements, including elements that may be discovered after the filing date of this application. Within this disclosure, the terms "scrap," "scrap pieces," "material," and "material pieces" may be used interchangeably.

[0028] As is well known in the art, a "polymer" is a substance or material made up of very large molecules or macromolecules made up of many repeating subunits. Polymers can be natural polymers found in nature or synthetic polymers.

[0029] A "multilayer polymer film" is a film made up of two or more different compositions, with a maximum of about 7.5 -8 ×10 -4 The layer may have a thickness of about 100 nm. The layer is at least partially contiguous, and preferably, but optionally, is coextensive.

[0030] As used herein, the terms "plastic," "plastic piece," and "plastic material piece" (which may all be used interchangeably) refer to any object that includes or is composed of a polymeric composition of one or more polymers and / or multilayer polymeric films.

[0031] As used herein, the term "chemical signature" refers to a unique pattern (such as a fingerprint spectrum) generated by one or more analytical instruments that indicates the presence of one or more particular elements or molecules (including polymers) in a sample. The elements or molecules may be organic and / or inorganic. Such analytical instruments include any of the sensor systems disclosed herein. According to embodiments of the present disclosure, one or more sensor systems disclosed herein may be configured to generate a chemical signature of a piece of material (e.g., a piece of plastic).

[0032] As used herein, "fraction" refers to a particular combination of organic and / or inorganic elements or molecules, polymer type, plastic type, polymer composition, chemical characteristics of a plastic, physical properties of a plastic piece (e.g., color, clarity, strength, melting point, density, shape, size, type of manufacture, uniformity, response to stimuli, etc.), including all of the various classifications and types of plastics disclosed herein. Non-limiting examples of fractions are LDPE with a relatively high percentage of aluminum; LDPE and PP with a relatively low percentage of iron; PP with zinc; combinations of PE, PET, and HDPE; any type of red LDPE plastic piece; any combination of plastic pieces except PVC; black plastic pieces; combinations of #3-#7 types of plastics containing specified combinations of organic and inorganic molecules; combinations of one or more different types of multilayer polymer films; combinations of certain plastics that do not contain certain contaminants or additives; any type of plastic with a melting point above a specified threshold; multiple specific types of thermosetting plastics; certain plastics that do not contain chlorine; combinations of plastics with similar densities; combinations of plastics with similar polarity; and one or more different types of plastic pieces, including PET bottles without caps or vice versa.

[0033] "Catalytic pyrolysis" involves the decomposition of a polymeric material by heating the polymeric material in the presence of a catalyst in the absence of oxygen.

[0034] The term "predetermined" refers to something that is established or determined in advance.

[0035] "Spectral imaging" is imaging that uses multiple bands across the electromagnetic spectrum. While a typical camera captures light across three wavelength bands of the visible spectrum, red, green, and blue (RGB), spectral imaging includes a variety of techniques including but beyond RGB. Spectral imaging may use infrared, visible, ultraviolet, and / or x-ray spectrum, or combinations of the above. Spectral data, or spectral image data, is a digital data representation of a spectral image. Spectral imaging may include simultaneous acquisition of spectral data in visible and non-visible bands, illumination from outside the visible range, or use of optical filters to capture specific spectral ranges. It is also possible to capture hundreds of wavelength bands for each pixel of a spectral image.

[0036] As used herein, the term "image data packet" refers to a packet of digital data relating to a captured spectral image of an individual piece of material.

[0037] As used herein, the terms "identify" and "classify", "identification" and "classification", and their derivatives, may be used interchangeably. As used herein, "classifying" a piece of material means determining (i.e., identifying) the type or class of material to which the piece of material belongs. For example, according to certain embodiments of the present disclosure, a sensor system (described further herein) may be configured to collect and analyze any type of information to classify materials, and classification may be utilized within a sorting system to selectively separate pieces of material according to a set of one or more physical and / or chemical properties (e.g., user definable), including, but not limited to, color, texture, hue, shape, brightness, weight, density, composition, size, uniformity, manufacturing type, chemical characteristics, predetermined percentages, radioactive characteristics, transmittance of light, sound, or other signals, and response to stimuli such as various fields, including radiative and / or reflected electromagnetic radiation ("EM") of the piece of material. As used herein, "production type" refers to the type of manufacturing process by which a piece of material is produced, such as a metal part formed by a forging process, a cast metal part (including but not limited to expendable die casting, permanent die casting, and powder metallurgy), a forging, a material removal process, etc.

[0038] The material types or classes (i.e., classifications) are user definable and are not limited to known material classifications. The granularity of the types or classes can range from very coarse to very fine. For example, the types or classes can include relatively coarse types or classes of plastics, ceramics, glass, metals, and other materials, finer types or classes of various metals and metal alloys, such as zinc, copper, brass, chrome plate, aluminum, etc., or relatively fine types or classes of plastics of a particular type. Thus, the types or classes can be configured to distinguish between materials of significantly different composition, such as, for example, different types of plastics (e.g., any of types #1-#7 of plastics), or to distinguish between materials of nearly identical composition, such as, for example, different subclasses of plastics that may fall within a particular plastic type. It should be understood that the methods and systems discussed herein can be applied to accurately identify / classify materials whose composition is completely unknown before being classified.

[0039] Embodiments of the present disclosure improve plastic sorting capabilities through the fusion of multiple sensor technologies with machine learning systems. The limitations of sensor-based sorting technology arise from the use of a single sensor, as each sensor can only detect a narrow range of signals. The most common sorter sensor types are eddy current, visible camera, x-ray transmission, near infrared, and x-ray fluorescence ("XRF"). These are summarized in the following table:

[0040] [Table 1]

[0041] However, the plastic pieces of MSW may be composed of one or more organic polymers, one or more inorganic elements, and come in a variety of colors, shapes, and sizes. Examples of these plastics include potato chip bags, squeezable juice boxes, some beverage containers, and electromagnetic sensitive packaging of electronic devices. Embodiments of the present disclosure generate novel fractions from this waste stream using sensor-based techniques that can achieve the separation of these different types of plastics into unique classifications that can account for their organic polymer composition and / or their inorganic element composition. For example, a conversion chemist with a keen interest in the relative composition of polymers and inorganic elements can select one or more novel fractions and create specific products from recycled plastics separated into such fractions. As a result, a sorting system configured according to embodiments of the present disclosure can generate fractions beyond those possible with existing state-of-the-art sorting techniques.

[0042] For example, certain embodiments of the present disclosure may be configured to sort and / or separate predetermined fractions from bales of #3-#7 type plastics to create new products (e.g., by recycling methods) and / or fuels. Examples of end uses for such fractions include, but are not limited to, gases (e.g., C1-C4, etc.), fuels (e.g., gasoline, diesel, etc.), and vacuum gas oil. However, sorting #3-#7 type plastics based on organic and inorganic elemental composition has not been successful to date.

[0043] Embodiments of the present disclosure may be configured to classify pieces of plastic material according to various different pre-defined fractions or combinations of characteristics or types as disclosed below and elsewhere within this disclosure.

[0044] Plastics are classified into three types according to their characteristics, such as chemical structure, polarity, and applications.

[0045] According to their chemical structure and temperature behavior, plastics can be classified as thermoplastics, thermosets, and elastomers.

[0046] Regarding polarity, when atoms of different nature are present, electrons move towards the most electronegative atom in a covalent bond, resulting in a dipole. Polymers containing extremely electronegative atoms such as CI, O, N, and F become polar compounds, which affect the properties of the material. Higher polarity improves mechanical resistance, hardness, rigidity, heat resistance, water absorption and moisture absorption, and chemical resistance, as well as permeability to polar compounds such as water vapor, and adhesion and adhesion to metals. At the same time, increased polarity leads to thermal expansion, electrical insulation ability, tendency to accumulate electrostatic charges, and the ability to form polar molecules (O 2 , N 2 In this way, different families such as polyolefins, polyesters, acetals, halogenated polymers, etc. can be distinguished.

[0047] Depending on the application, a third classification applies to thermoplastic materials. In this third classification there are four types of plastics:

[0048] Standard Plastic or Commodity: A plastic that is produced and used in large quantities due to its price and various favorable properties. Examples are polyethylene ("PE"), polypropylene ("PP"), polystyrene ("PS"), polyvinyl chloride ("PVC"), or acrylonitrile butadiene styrene ("ABS") copolymers.

[0049] Engineering Plastics: Used where superior structure, transparency, self-lubrication, and thermal properties are required. Examples include polyamide ("PA"), polyacetal ("POM"), polycarbonate ("PC"), polyethylene terephthalate ("PET"), polyphenylene ether ("PPE"), and polybutylene terephthalate ("PBT").

[0050] Specialty Plastics: Plastics with very specific properties, such as polymethyl methacrylate ("PMMA"), which has high transparency and light resistance, and polytetrafluoroethylene (Teflon(R)), which has excellent resistance to temperature and chemicals.

[0051] High-Performance Plastics: Mainly high-temperature resistant thermoplastics, i.e. with good mechanical resistance to high temperatures, especially up to 150°C. Polyimides ("PI"), polysulfones ("PSU"), polyethersulfones ("PES"), polyarylsulfones ("PAS"), polyphenylene sulfide ("PPS"), and liquid crystal polymers ("LCP") are high-performance plastics.

[0052] Many plastic products are marked with a symbol that indicates the type of polymer from which they are made. These resin identification codes, often abbreviated as RIC, are used internationally. There are seven codes in total, six of which correspond to the most common types of general-purpose plastics and one that is a catch-all for all others. These types are also referred to herein as polymer types #1 through #7. Polymer type #1 refers to polyethylene terephthalate ("PET"), #2 refers to high density polyethylene ("HDPE"), #3 refers to polyvinyl chloride ("PVC"), #4 refers to low density polyethylene ("LDPE"), #5 refers to polypropylene ("PP"), #6 refers to polystyrene ("PS"), and #7 refers to other polymers not included in polymer types #1 through #6 (e.g., acrylic, polycarbonate ("PC"), polyactive fiber, polylactic acid, nylon, fiberglass, ABS). The EU maintains a similar nine-code list, which also includes ABS and polyamide.

[0053] PET plastics are used to make many common household items, such as beverage bottles, medicine bottles, rope, clothing, and carpet fibers. HDPE plastics are often used to make containers for milk, motor oil, shampoo and conditioner, soap bottles, detergent, and bleach. PVC is used for all kinds of pipes and tiles, most commonly used in plumbing pipes. LDPE products include plastic wrap, sandwich bags, squeezable bottles, plastic grocery bags, and more. PP is used to make lunch boxes, margarine containers, yogurt pots, syrup bottles, medicine bottles, and plastic bottle caps. Polystyrene products include disposable coffee cups, plastic food boxes, plastic cutlery, and packaging foam. Polycarbonate is used in baby bottles, CDs, medical storage containers, and more. Thus, according to an embodiment of the present disclosure, a vision system implementing a machine learning system can be trained to identify and separate these different types of plastics based on the type of product produced.

[0054] Plastic pieces can be classified according to the types of additives they may contain. Additives are compounds that are blended into plastics to improve their performance, and include stabilizers, fillers and dyes. Transparent plastics will be the most valuable as they may still be dyed, while black or strongly colored plastics will be much less valuable as their inclusion may discolor the product. Thus, to obtain material suitable for recycling, it may be necessary to separate plastics by both polymer type and color.

[0055] Plastics can also be classified and separated based on density. Certain polymers have similar density ranges (e.g. PP and PE, or PET, PS, and PVC). If a piece of plastic contains a high percentage of filler, this can affect the density.

[0056] Plastic waste can be broadly divided into two categories: industrial scrap (also known as post-industrial resin) and post-consumer waste.

[0057] Plastic pieces may be sorted / separated depending on how they may be recycled. During mechanical recycling, plastics may be reprocessed at temperatures between 150-320°C depending on the type of polymer, which may result in undesirable chemical reactions that cause degradation of the polymer. This may reduce the physical properties and overall quality of the plastic, produce volatile low molecular weight compounds, produce undesirable tastes and odors, or cause thermal discoloration. Thus, embodiments of the present disclosure may be configured to sort and separate plastic pieces such that such undesirable chemical reactions are avoided. Additives present within plastics may accelerate this degradation. For example, oxo-biodegradable additives intended to improve the biodegradability of plastics may increase the degree of thermal degradation. Similarly, flame retardants may also have undesirable effects. Thus, embodiments of the present disclosure may be configured to sort and separate plastic pieces such that plastic pieces containing certain of such additives are discarded.

[0058] The quality of the product can also depend heavily on how well the plastics are separated. Many polymers are immiscible with each other when melted and phase separate during reprocessing (like oil and water). Products made from such blends contain many boundaries between different types of polymers and have poor cohesion across these boundaries, resulting in poor mechanical properties. Thus, embodiments of the present disclosure can be configured to sort and separate plastic pieces so that certain immiscible plastic pieces are not sorted together in the same group.

[0059] The systems and methods described herein in accordance with certain embodiments of the present disclosure receive a heterogeneous mixture of multiple pieces of material (e.g., any combination of various plastics disclosed herein), where at least one piece of material within the heterogeneous mixture includes a different elemental composition (e.g., chemical signature) than one or more other pieces of material and / or at least one piece of material within the heterogeneous mixture is distinguishable from other pieces of material (e.g., visually identifiable properties or characteristics, different chemical signatures, etc.), and the systems and methods are configured to identify / classify / separate the piece of material into a group separate from such other pieces of material. Embodiments of the present disclosure may be utilized to separate any type or class of material or fraction as defined herein.

[0060] Embodiments of the present disclosure are described herein as separating pieces of material into separate groups by physically placing (e.g., diverting or discharging) the pieces of material into separate containers or bins according to user-defined groupings (e.g., material type classifications or fractions). As an example, in certain embodiments of the present disclosure, pieces of material may be separated into separate bins to separate pieces of material having physical characteristics (e.g., visually identifiable characteristics or features, different chemical characteristics, etc.) that are distinguishable from the physical characteristics of other pieces of material.

[0061] FIG. 1 illustrates an example of a system 100 configured in accordance with various embodiments of the present disclosure. A conveyor system 103 may be implemented to transport one or more streams of individual pieces of material 101 through the system 100, so that each of the individual pieces of material 101 can be tracked, classified, and separated into predetermined desired groups. Such a conveyor system 103 may be implemented using one or more conveyor belts along which the pieces of material 101 typically move at a predetermined constant speed. However, certain embodiments of the present disclosure may be implemented with other types of conveyor systems, including systems in which the pieces of material free-fall through various components of the system 100 (or any other type of vertical separator), or vibratory conveyor systems. Hereinafter, where applicable, the conveyor system 103 may also be referred to as a conveyor belt 103. In one or more embodiments, some or all of the operations of communicating, stimulating, detecting, classifying, and separating may be performed automatically, i.e., without human intervention. For example, in system 100, one or more stimulus sources, one or more radiation detectors, a classification module, a fractionator, and / or other system components may be configured to perform these and other operations automatically.

[0062] Additionally, while FIG. 1 illustrates a single stream of pieces of material 101 on the conveyor system 103, embodiments of the present disclosure may be implemented such that multiple such streams of pieces of material pass through various components of the system 100 in parallel with one another. For example, as further described in U.S. Pat. No. 10,207,296, the pieces of material may be distributed into two or more parallel, singulated streams moving on a single conveyor belt, or a set of parallel conveyor belts. Thus, certain embodiments of the present disclosure may simultaneously track, sort, and separate multiple such parallel moving streams of pieces of material. In accordance with certain embodiments of the present disclosure, the incorporation or use of a singulator is not required. Instead, a conveyor system (e.g., conveyor system 103) may simply transport a mass of pieces of material deposited on the conveyor system 103 in a random manner.

[0063] According to certain embodiments of the present disclosure, some suitable feeder mechanism (e.g., another conveyor system or hopper 102) may be utilized to feed the pieces of material 101 onto the conveyor system 103, which may then transport the pieces of material 101 through various components within the system 100. After the pieces of material 101 are received by the conveyor system 103, an optional tumbler / vibrator / singulator 106 may be utilized to separate individual pieces of material from the collection of pieces of material. In certain embodiments of the present disclosure, the conveyor system 103 is operated to move at a predetermined speed by a conveyor system motor 104. This predetermined speed may be programmable and / or adjustable by an operator in any known manner. Alternatively, monitoring of the predetermined speed of the conveyor system 103 may be performed using a position detector 105. In certain embodiments of the present disclosure, control of the conveyor system motor 104 and / or the position detector 105 may be performed by an automatic control system 108. Such an automated control system 108 may operate under the control of a computer system 107 and / or the functions for performing the automated control may be implemented in software within the computer system 107 .

[0064] The conveyor system 103 may be a conventional endless belt conveyor that uses a conventional drive motor 104 suitable for moving the belt conveyor at a predetermined speed. The position detector 105 may be a conventional encoder and may be operatively coupled to the conveyor system 103 and the automatic control system 108 to provide information corresponding to the movement (e.g., speed) of the conveyor belt. Thus, as described further herein, through the use of control over the conveyor system drive motor 104 and / or the automatic control system 108 (or including the position detector 105), once each of the pieces of material 101 moving on the conveyor system 103 are identified, they may be tracked by position (relative to the various components of the system 100) and time, such that various components of the system 100 may be activated / deactivated as each piece of material 101 passes nearby. As a result, the automatic control system 108 may track the position of each of the pieces of material 101 as they move along the conveyor system 103.

[0065] Referring again to FIG. 1 , certain embodiments of the present disclosure may utilize a vision or optical recognition system 110 and / or a piece of material tracker 111 as a means of tracking each of the pieces of material 101 moving on the conveyor system 103. The vision system 110 may utilize one or more still or live action cameras 109 to record the position (i.e., location and timing) of each of the pieces of material 101 on the moving conveyor system 103. The vision system 110 may additionally or alternatively be configured to perform a particular type of identification (e.g., classification) of all or a portion of the pieces of material 101, as described further herein. For example, such a vision system 110 may be utilized to capture or obtain information about each of the pieces of material 101. For example, the vision system 110 may be configured (e.g., using a machine learning system) to capture or gather any type of information from the pieces of material available in the system 100 in order to classify and / or selectively separate the pieces of material 101 according to a set of one or more characteristics (e.g., physical and / or chemical and / or radioactive materials, etc.) as described herein. According to certain embodiments of the present disclosure, the vision system 110 may be configured to capture visual images (including one-dimensional, two-dimensional, three-dimensional, or holographic images) of each of the pieces of material 101, for example, by using optical sensors used in common digital cameras and video equipment. Such visual images captured by the optical sensors are stored in a memory device as image data (e.g., initialized as image data packets). According to certain embodiments of the present disclosure, such image data may represent images captured within optical wavelengths of light (i.e., wavelengths of light observable by a typical human eye). However, alternative embodiments of the present disclosure may utilize sensor systems configured to capture images of materials comprised of wavelengths of light outside the visual wavelengths of the human eye.

[0066] According to certain embodiments of the present disclosure, the system 100 may be implemented with one or more sensor systems 120, which may be utilized alone or in combination with the vision system 110 to classify / identify the pieces of material 101. The sensor systems 120 may be any sensor system that utilizes emitted or reflected electromagnetic radiation (e.g., infrared ("IR"), Fourier transform IR ("FTIR"), forward looking infrared ("FLIR"), very near infrared ("VNIR"), near infrared ("NIR"), short wavelength infrared ("SWIR"), long wavelength infrared ("LWIR"), mid wavelength infrared ("MWIR" or "MIR"), x-ray transmission ("XRT"), gamma ray, ultraviolet ("UV"), x-ray fluorescence ("XRF"), laser induced breakdown spectroscopy ("LIBS"), Raman spectroscopy, and the like). The sensor system 120 may be configured with any type of sensor technology to determine the chemical signature of the plastic pieces and / or to classify the plastic pieces for separation, including utilizing Stokes Raman spectroscopy, gamma ray spectroscopy, hyperspectral spectroscopy (e.g., any range above visible wavelengths), acoustic spectroscopy, NMR spectroscopy, microwave spectroscopy, terahertz spectroscopy, including one-dimensional, two-dimensional, or three-dimensional imaging using any of the above), or by any other type of sensor technology including, but not limited to, chemical or radioactive materials. Implementations of exemplary XRF systems (e.g., for use as the sensor system 120 herein) are further described in U.S. Pat. No. 10,207,296. XRF may be used within embodiments of the present disclosure to identify inorganic materials within the plastic pieces (e.g., for inclusion within the chemical signature).

[0067] The following sensor systems may also be used within certain embodiments of the present disclosure to determine the chemical signature of plastic pieces and / or to classify plastic pieces for separation.

[0068] Various forms of infrared spectroscopy as previously disclosed can be used to obtain a unique chemical signature for each plastic piece that provides information about the base polymer of any plastic material as well as other components present in the material (mineral fillers, copolymers, polymer blends, etc.).

[0069] Differential Scanning Calorimetry (DSC) is a thermal analysis technique that captures the material-specific thermal transitions that are produced during heating of the material being analyzed.

[0070] Thermogravimetric analysis ("TGA") is another thermal analytical technique that provides quantitative information about the composition of plastic materials with respect to the proportion of polymers, other organic components, mineral fillers, carbon black, etc.

[0071] Capillary and rotational rheometry can determine the rheological properties of polymeric materials by measuring their creep resistance and resistance to deformation.

[0072] Optical microscopy and scanning electron microscopy ("SEM") can provide information about the structure of the analyzed material with respect to the number and thickness of layers in a multilayer material (such as a multilayer polymer film), the dispersion size of pigment or filler particles in a polymer matrix, defects in a coating, the interfacial morphology between components, etc.

[0073] Chromatography (LC-PDA, LC-MS, LC-LS, GC-MS, GC-FID, HS-GC, etc.) can quantify trace components of plastic materials such as residual monomers, residual solvents in inks and adhesives, and degradation products, as well as UV stabilizers, antioxidants, plasticizers, and anti-slip agents.

[0074] It should be noted that while FIG. 1 is shown with a combination of a vision system 110 and one or more sensor systems 120, embodiments of the present disclosure may be implemented with any combination of sensor systems utilizing any of the sensor technologies disclosed herein or other sensor technologies currently available or developed in the future. Although FIG. 1 is shown including one or more sensor systems 120, such sensor system implementation is optional within certain embodiments of the present disclosure. In certain embodiments of the present disclosure, a combination of both a vision system 110 and one or more sensor systems 120 may be used to classify the piece of material 101. In certain embodiments of the present disclosure, any combination of one or more different sensor technologies disclosed herein may be used to classify the piece of material 101 without utilizing a vision system 110. Additionally, embodiments of the present disclosure may include any combination of one or more sensor systems and / or vision systems where the output of the sensor / vision system may be processed within a machine learning system (as further disclosed herein) to classify / identify materials from a heterogeneous mixture of materials, which may then be separated from one another.

[0075] According to alternative embodiments of the present disclosure, the vision system 110 and / or sensor system may be configured to identify which pieces of material 101 are not of the type to be separated by the system 100 (e.g., plastic pieces that contain certain contaminants, additives, or undesirable physical characteristics (e.g., an attached container cap made of a different type of plastic than the container)) and transmit a signal to reject such pieces of material. In such a configuration, the identified pieces of material 101 may be diverted / ejected utilizing one of the mechanisms described below for physically diverting separated pieces of material into individual bins.

[0076] In certain embodiments of the present disclosure, the material piece tracker 111 and associated control system 112 may be utilized and configured to measure the position (i.e., location and timing) of each material piece 101 on the moving conveyor system 103 as well as the size and / or shape of each of the material pieces 101 passing in proximity to the material piece tracker 111. Exemplary operation of such material piece tracker 111 and control system 112 is further described in U.S. Pat. No. 10,207,296. Alternatively, as previously disclosed, the vision system 110 may be utilized to track the position (i.e., location and timing) of each of the material pieces 101 transported by the conveyor system 103. Accordingly, certain embodiments of the present disclosure may be implemented without a material piece tracker (e.g., material piece tracker 111) for tracking the material pieces.

[0077] In certain embodiments of the present disclosure implementing one or more sensor systems 120, the sensor systems 120 may be configured to assist the vision system 110 in identifying the chemical composition, relative chemical composition, and / or manufacturing type of each of the pieces of material 101 as they pass in the vicinity of the sensor system 120. The sensor system 120 may include, for example, an energy emitting source 121 that may be powered by a power source 122 to stimulate a response from each of the pieces of material 101.

[0078] According to certain embodiments of the present disclosure implementing an XRF system as the sensor system 120, the source 121 may include an in-line X-ray fluorescence ("IL-XRF") tube as further described in U.S. Pat. No. 10,207,296. Such an IL-XRF tube may include a separate X-ray source dedicated to one or more streams (e.g., singulated) of conveyed pieces of material. In such a case, the one or more detectors 124 may be implemented as an XRF detector that detects X-ray fluorescence from the pieces of material 101 within each of the singulated streams. Examples of such XRF detectors are further described in U.S. Pat. No. 10,207,296.

[0079] In certain embodiments of the present disclosure, the sensor system 120 may emit an appropriate sensing signal towards the piece of material 101 as each piece of material 101 passes in proximity to the radiation source 121. The one or more detectors 124 may be positioned and configured to sense / detect one or more characteristics from the piece of material 101 in a format appropriate for the type of sensor technology employed. The one or more detectors 124 and associated detector electronics 125 may capture these received sensed characteristics and perform signal processing to generate digitized information (e.g., spectral data) representative of the sensed characteristics, which may then be analyzed in accordance with certain embodiments of the present disclosure and used to assist the vision system 110 in classifying each of the pieces of material 101. This classification can be performed within the computer system 107 and then utilized by the automatic control system 108 to operate one of the N (N>1) sorting devices 126...129 of the sorting device to sort (e.g., redirect / discharge) the pieces of material 101 according to the determined classification into one or more N (N>1) sorting bins 136...139. Four sorting devices 126...129 and four sorting bins 136...139 associated with these sorting devices are shown in FIG. 1 by way of non-limiting example only.

[0080] Existing sorters for plastics are designed to separate materials in a binary fashion, with air nozzles at the end of a conveyor discharging the identified class of plastic into one of two bins. For example, if four classes of plastics need to be separated, the entire stream would need to be conveyed through such a binary sorter four separate times, which would take four times as long as removing a single object in the stream. According to an embodiment of the present disclosure, the system 100 allows for the separation of multiple classes of plastics in a single pass.

[0081] The sorting device may include any known mechanism for redirecting selected pieces of material 101 toward a desired location, including, but not limited to, diverting the pieces of material 101 from a conveyor belt system to multiple sorting bins. For example, the sorting device may utilize air jets, with each air jet assigned to one or more classifications. When one of the air jets (e.g., 127) receives a signal from the automated control system 108, that air jet emits an air stream that redirects / discharges the pieces of material 101 from the conveyor system 103 into the sorting bin (e.g., 137) that corresponds to that air jet.

[0082] Other mechanisms may also be used to redirect / eject the pieces of material, such as robotically removing the pieces of material from the conveyor belt, pushing the pieces of material off the conveyor belt (e.g., using a paintbrush-type plunger), creating an opening in the conveyor system 103 through which the pieces of material may fall (e.g., a trap door), or using air jets to redirect the pieces of material to another bin as they fall off the end of the conveyor belt. The term pusher device, as used herein, may refer to any form of device that can be actuated to dynamically move objects on or off the conveyor system / apparatus using pneumatic, mechanical, or other means, such as an appropriate type of mechanical pushing mechanism (e.g., ACME screw drive), pneumatic pushing mechanism, or air jet pushing mechanism. Some embodiments may include multiple pusher devices located at different positions along the path of the conveyor system and / or with different redirection path orientations. In various different implementations, these sorting systems described herein may determine which pusher device (if any) to activate depending on the classification of the pieces of material performed by a machine learning system. Additionally, the determination of which pusher device to activate may be based on the detected presence and / or characteristics of other objects that may be in the diversion path of the pusher device at the same time as the target items. Additionally, even in installations where singulation along the conveyor system is not perfect, the disclosed sorting system may recognize that multiple objects are not properly separated and dynamically select which pusher device to activate from the multiple pusher devices based on which pusher device provides the optimal diversion path that may separate the closely spaced objects. In some embodiments, objects identified as target objects may represent material that should be diverted from the conveyor system. In other embodiments, objects identified as target objects represent material that should be allowed to remain on the conveyor system so that non-target objects are diverted instead.

[0083] In addition to the N sorting bins 136...139 into which the pieces of material 101 are diverted / discharged, the system 100 may also include a container or bin 140 that receives pieces of material 101 that have not been diverted / discharged from the conveyor system 103 into any of the aforementioned sorting bins 136...139. For example, the pieces of material 101 may not be diverted / discharged from the conveyor system 103 into one of the N sorting bins 136...139 if the classification of the pieces of material 101 has not been determined (or simply because the sorting device has not been able to properly divert / discharge the pieces of material). Thus, the bin 140 may function as a default container into which unsorted pieces of material are dumped. Alternatively, the bin 140 may be used to receive one or more classifications of pieces of material that have not been purposefully assigned to any of the N sorting bins 136...139. These pieces of material may be further sorted according to other characteristics and / or by another sorting system.

[0084] Depending on the various classifications of the desired pieces of material, multiple classifications can be mapped to a single sorting device and associated sorting bins. In other words, there need not be a one-to-one correlation between classifications and sorting bins. For example, a user may desire to separate certain classifications of material into the same sorting bin (e.g., different plastic types contained in fractions). To achieve this separation, when pieces of material 101 are classified as falling into a given classification group (e.g., fraction), the same sorting device can operate to separate them into the same sorting bin. Such combination sorting can be applied to generate any desired combination of separated pieces of material. The mapping of classifications may be programmed by the user (e.g., using a sorting algorithm (e.g., see FIG. 7) operated by the computer system 107) to generate such desired combinations. Furthermore, the classifications of the pieces of material are user-definable and are not limited to any particular known classifications of pieces of material (e.g., fractions as disclosed herein).

[0085] Conveyor system 103 may include a carousel (not shown) such that unsorted pieces of material are returned to the beginning of system 100 and passed through system 100 again. Additionally, because system 100 can specifically track each piece of material 101 moving on conveyor system 103, some sorting device (e.g., sorting device 129) may be implemented to guide / eject the pieces of material 101 such that system 100 does not fail to sort after a predetermined number of cycles by system 100 (or after the pieces of material 101 have been collected in bins 140).

[0086] In certain embodiments of the present disclosure, the conveyor system 103 may be divided into multiple belts configured in series, e.g., two belts, where a first belt carries pieces of material past the vision system 110 and a second belt carries a particular sorted piece of material past a sensor system 120 implemented for a second sorting. Furthermore, such a second conveyor belt may be at a lower height than the first conveyor belt such that the pieces of material fall from the first belt onto the second belt.

[0087] In certain embodiments of the present disclosure implementing the sensor system 120, the light emission source 121 may be located above the detection area (i.e., above the conveyor system 103), although certain embodiments of the present disclosure may place the light emission source 121 and / or detector 124 in other locations that still produce an acceptable sensed / detected physical characteristic.

[0088] The systems and methods described herein may be applied to sort and / or separate individual pieces of material having any of a variety of sizes and shapes. Although the systems and methods described herein are described primarily in the context of separating individual pieces of material, the systems and methods described herein are not limited thereto. Such systems and methods may be used to stimulate and / or detect emissions from multiple materials simultaneously. For example, as opposed to singulated material streams being conveyed serially along one or more conveyor belts, multiple singulated streams may be conveyed in parallel. Each stream may be on the same belt or on different belts arranged in parallel. Additionally, the pieces of material may be randomly distributed on (e.g., across and along) one or more conveyor belts. Thus, the systems and methods described herein may be used to stimulate and / or detect emissions from multiple pieces of material simultaneously. In other words, multiple pieces of material may be treated as a single part, rather than each piece of material being considered individually. Thus, multiple pieces of material may be sorted and separated together (e.g., redirected / discharged from a conveyor system).

[0089] Although the systems and methods described herein are primarily described in relation to fractionating pieces of material, such systems and methods are not limited to that application, and they may be used for other applications, such as identifying elements (e.g., contaminants, etc.) within a piece of material or determining the composition of a piece of material.

[0090] As previously mentioned, certain embodiments of the present disclosure may implement one or more vision systems (e.g., vision system 110) to identify, track, and / or sort pieces of material. According to embodiments of the present disclosure, such vision systems may operate alone to identify and / or sort and separate pieces of material, or may operate in combination with one or more sensor systems (e.g., sensor system 120) to identify and / or sort and separate pieces of material. If a sorting system (e.g., system 100) is configured to operate with only such vision system 110, sensor system 120 may be omitted from system 100 (or may simply be turned off).

[0091] Regardless of the sensed characteristics / type of information captured of the piece of material, the information (e.g., image data packets) may then be transmitted to a computer system (e.g., computer system 107) and processed by a machine learning system to identify and / or classify each piece of material. Such machine learning systems may implement any well-known machine learning system, including those that implement neural networks (e.g., artificial neural networks, deep neural networks, convolutional neural networks, recurrent neural networks, autoencoders, reinforcement learning, etc.), fuzzy logic, artificial intelligence ("AI"), deep learning algorithms, deep structured learning hierarchical learning algorithms, support vector machines ("SVMs") (e.g., linear SVMs, non-linear SVMs, SVM regression, etc.), decision tree learning (e.g., classification and regression trees ("CART"), ensemble techniques (e.g., ensemble learning, random forests, bagging and pasting, patching and subspace, boosting, stacking, etc.), dimensionality reduction (e.g., projection, manifold learning, principal component analysis, etc.) and / or deep machine learning algorithms, such as those described and published at the deeplearning.net website (including hyperlinks to all software, publications, and available software referenced within this website), which is incorporated herein by reference.Non-limiting examples of publicly available machine learning software and libraries that can be utilized within embodiments of the present disclosure include Python, OpenCV, Inception, Theano, Torch, PyTorch, Pylearn2, Numpy, Blocks, TensorFlow, MXNet, Caffe, Lasagne, Keras, Chainer, Matlab Deep Learning, CNTK, MatConvNet (a MATLAB® toolbox implementing convolutional neural networks for computer vision applications), DeepLearnToolbox (a Matlab toolbox for deep learning (by Rasmus Berg Palm)), BigDL, Cuda-Convnet (a fast C++ / CUDA implementation of convolutional (more generally feedforward) neural networks), Deep Belief Networks, RNNLM, RNNLIB-RNNLIB, matrbm, deeplearning4j, Eblearn.lsh, deepmat, MShadow, Matplotlib, SciPy, CXXNET, Nengo-Nengo, Eblearn, cudamat, Gnumpy, 3-Way Factorized RBM and mcRBM, mPoT (Python code to train a model for natural images using CUDAMat and Gnumpy), ConvNet, Elektronn, OpenNN, NeuralDesigner, Theano Generalized Hebbian Learning, Apache Singa, Lightnet, and SimpleDNN.

[0092] According to certain embodiments of the present disclosure, machine learning may be performed in two stages. For example, training is performed first, which may be performed offline in that the system 100 is not utilized to perform the actual classification / segregation of the material pieces (see, e.g., FIGS. 3-4). The system 100 may be utilized to train the machine learning system such that a homogenous set (also referred to herein as a control sample) of material pieces (i.e., having the same type or class of material or falling within the same predefined fraction) is passed through the system 100 (e.g., by conveyor system 103) and all such material pieces may not be separated but may be collected in a common bin (e.g., bin 140, etc.). Alternatively, training may be performed at another location remote from the system 100, including using some other mechanism to collect sensory information (characteristics) of a control set of material pieces. During this training phase, algorithms within the machine learning system extract features from the captured information (e.g., using image processing techniques well known in the art). Non-limiting examples of training algorithms include, but are not limited to, linear regression, gradient descent, feed forward, polynomial regression, learning curve, regularized learning model, and logistic regression. It is during this training phase that the algorithms in the machine learning system learn the relationships between materials and their features / characteristics (e.g., captured by the vision and / or sensor systems) to create a knowledge base for later classifying heterogeneous mixtures of material pieces that may be received by the system 100 and then separated by a desired classification. Such a knowledge base may include one or more libraries, each library including parameters (e.g., neural network parameters) utilized by the machine learning system in classifying the material pieces. For example, one particular library may include parameters configured by the training phase to recognize and classify a particular type or class of material, or one or more materials that fall into a given fraction.According to certain embodiments of the present disclosure, such libraries can be input into a machine learning system, and a user of the system 100 may be able to adjust certain of the parameters to tune the operation of the system 100 (e.g., adjusting a threshold effectiveness for how well the machine learning system recognizes a particular piece of material from a heterogeneous mixture of materials).

[0093] As shown in FIG. 2, in a training phase, multiple pieces of material 201 of one or more particular types, classes, or fractions of material that are control samples may be delivered (e.g., by conveyor system 203) through the vision system and / or one or more sensor systems such that algorithms in the machine learning system detect, extract, and learn features representative of such types or classes of material. For example, each of the pieces of material 201 may be an individual piece of plastic of a particular type, class, or predefined fraction that has been through such a training phase such that algorithms in the machine learning system "learn" (train) how to detect, recognize, and classify such plastic pieces. When training a vision system (e.g., vision system 110), it is trained to visually identify the pieces of material. This creates a library of parameters specific to one or more particular types, classes, or fractions of plastic material. The same process can then be performed on different types, classes, or fractions of plastic pieces to create a library of parameters specific to that type, class, or fraction, and so on. For each type, class, or fraction of plastic classified by the machine learning system, any number of exemplary plastic pieces of that type, class, or fraction of plastic may be passed through the system. Given the captured sensory information as input data, an algorithm within the machine learning system uses N classifiers, each testing one of the N different material types, classes, or fractions. It is noted that the machine learning system may be "educated" (trained) to detect any type, class, or fraction of material, including any type, class, or fraction of material found in MSW, or any other material disclosed herein.

[0094] After the algorithms are established and the machine learning system has sufficiently learned (trained) the differences (e.g., visually discernible differences) in the material classifications (e.g., within a user-defined statistical confidence level), the library of different material classifications may then be implemented into a material classification / separation system (e.g., system 100) that is used to identify and / or classify material pieces from a heterogeneous mixture of material pieces (e.g., as contained within MSW) and, if separation is performed, to separate such classified material pieces.

[0095] Techniques for building, optimizing, and utilizing machine learning systems are known to those skilled in the art as described in the relevant literature, examples of which include the following publications: Krizhevsky et al., "ImageNet Classification with Deep Convolutional Networks," Proceedings of the 25th International Conference on Neural Information Processing Systems, December 3-6, 2012, Lake Tahoe, NV, and LeCun et al., "Gradient-Based Learning Applied to Document Recognition," IEEE Transactions on Machine Learning, Institute of Electrical and Electronics Engineers (IEEE), November 1998, both of which are incorporated herein by reference in their entirety.

[0096] In one exemplary technique, data captured by a vision or sensor system for a particular piece of material may be processed (within a data processing system (configured) to implement a machine learning system (e.g., data processing system 3400 of FIG. 9 )) as an array of data values. For example, the data may be spectral data captured by a digital camera or other type of sensor system for a particular piece of material and processed as an array of data values ​​(e.g., image data packets). Each data value may be represented by a single number or as a series of numbers representing a value. These values ​​may be multiplied by weight parameters of a neuron (e.g., using a neural network) and a bias may be added. This may affect the nonlinearity of the neuron. The resulting number output by a neuron may be treated the same as a value by multiplying this output by the weight value of the subsequent neuron, optionally adding a bias, again affecting the nonlinearity of the neuron. Each iteration of such processing is known as a “layer” of the neural network. The final output of the final layer may be interpreted as a probability that material is present or absent in the captured data associated with the piece of material. Examples of such processing are detailed in both the aforementioned references, "ImageNet Classification with Deep Convolutional Networks" and "Gradient-Based Learning Applied to Document Recognition."

[0097] According to certain embodiments of the present disclosure in which a neural network is implemented as the final layer ("classification layer"), a final set of neuron outputs is trained to represent the likelihood that the piece of material is associated with the captured data. In operation, if the likelihood that the piece of material is associated with the captured data exceeds a user-specified threshold, it is determined that the piece of material is indeed associated with the captured data. These techniques can be extended to determine not only the presence of a type of material associated with a particular captured data, but also whether a sub-region of a particular captured data belongs to one type of material or another. This process is known as segmentation, and there are techniques in the literature that use neural networks, such as neural networks known as "fully convolutional" neural networks and networks that are not fully convolutional but include convolutional parts (i.e., are partially convolutional). This allows the location and size of the material to be determined.

[0098] It should be understood that the present disclosure is not limited to only machine learning techniques. Other common approaches for material classification / identification can also be used. For example, a sensor system may utilize spectroscopy techniques using a multispectral or hyperspectral camera to provide a signal indicative of the presence or absence of a type, class, or fraction of material by examining the material's spectral emission (i.e., spectral imaging). Spectral images of the material pieces may also be used in a template matching algorithm, where a database of spectral images is compared to the captured spectral images to detect the presence or absence of a particular type of material from the database. A histogram of the captured spectral image may also be compared to a database of histograms. Similarly, a bag-of-words model may be used in conjunction with feature extraction techniques such as scale-invariant feature transform ("SIFT") to compare extracted features between the captured spectral image and the spectral images in the database.

[0099] Thus, as disclosed herein, certain embodiments of the present disclosure provide for identification / classification of one or more different types, classes, or fractions of material to determine which pieces of material should be diverted from the conveyor system within a defined group. According to certain embodiments, machine learning techniques are utilized to train (i.e., configure) a neural network to identify various one or more different types, classes, or fractions of material. Spectral images or other types of sensory information are captured from the material (e.g., moving on the conveyor system), and based on such material identification / classification, the system described herein can determine which pieces of material should remain on the conveyor system and which pieces of material should be diverted / removed from the conveyor system (e.g., either placed in a collection bin or diverted to another conveyor system).

[0100] According to certain embodiments of the present disclosure, a machine learning system for an existing facility (e.g., system 100) can be dynamically reconfigured to identify / classify the properties of a new type, class, or fraction of material by replacing the current set of neural network parameters with a new set of neural network parameters.

[0101] One point to mention here is that according to certain embodiments of the present disclosure, the detected / captured features / characteristics (e.g., spectral images) of the pieces of material do not necessarily have to be simply particularly identifiable or distinguishable physical properties, they may be abstract formulas that can only be expressed mathematically, or may not be expressible mathematically at all, and nevertheless the machine learning system may be configured to analyze the spectral data to look for patterns that can classify the control samples during the training phase. Furthermore, the machine learning system may take subsections of the captured information (e.g., spectral images, etc.) of the pieces of material and try to find correlations between predefined classifications.

[0102] According to certain embodiments of the present disclosure, instead of utilizing a training phase in which control samples of pieces of material are passed by a vision and / or sensor system, training of the machine learning system can be performed using labeling / annotation techniques, whereby data / information of the pieces of material is captured by the vision / sensor system and a user inputs labels or annotations that identify each piece of material and are used to create a library for use by the machine learning system in classifying pieces of material within a heterogeneous mix of pieces of material.

[0103] With reference to FIGS. 3-6, embodiments of the present disclosure combine or fuse multiple sensor technologies (e.g., any combination of visual ("VIS"), XRF, NIR, and MWIR) in a manner that uniquely identifies various types, classes, or fractions of plastics so that they may be separated by their organic and inorganic chemical composition. However, because these plastic pieces in MSW come in a variety of sizes and shapes, there may be significant variance in the signals generated from these various sensors between sensors. Thus, combining machine learning with the fusion of various sensor technologies improves the classification accuracy of these signals even in the presence of such large variance. Because implementing multiple different sensors in a system can increase the cost of the system and also reduce the speed of separation, certain embodiments of the present disclosure implement a system (e.g., system 100) with a smaller number of sensor systems (and consequently lower capital and operating costs) to increase economic viability while still being able to sufficiently separate materials.

[0104] 4 shows a simplified schematic diagram of a system (e.g., system 100) in which pieces of material (e.g., plastic pieces) 401 are transported by a conveyor system 403 past a sensor system that captures spectral data from each piece of material 401. In this non-limiting example, the sensor systems are a camera 410 (e.g., vision system 110), an XRF system 411, an NIR system 412, and an MWIR system 413 that capture visible image data of each piece of material 401. However, it should be noted that any of the other sensor systems disclosed herein may be utilized in any combination.

[0105] 3 and 4, the chemical signature of the material is determined in process block 301 using one or more sensor systems. The sensed / detected / captured signals from the sensor systems are combined (e.g., in a multi-dimensional data array) for each material to create a chemical signature. Recall that an XRF sensor system can determine the presence of inorganic elements or molecules in a plastic piece, while one or more other sensor systems, such as NIR or MWIR, can be combined to determine the presence of organic elements or molecules in a plastic piece. In process block 302, a visible image of each material piece is captured. In process block 303, the captured visible image of each material piece (i.e., its associated image data) is associated with its determined chemical signature (i.e., spectral image data). FIGS. 5 and 6 show non-limiting exemplary representations of the chemical signatures and associated image data of two different types of plastic materials (a bag of potato chips and an electronics package). As can be readily seen, different types or classes of plastic pieces have different (unique) chemical signatures, which are utilized within embodiments of the present disclosure to generate a separation and / or classification (which may be user-defined) of plastic waste. According to embodiments of the present disclosure, a control group of plastic pieces of a particular type or class can be run through the system shown in FIG. 4 to train a machine learning system to associate certain chemical signatures with certain types or classes of plastic pieces.

[0106] For example, with respect to the example shown in FIG. 5, images captured from multiple bags of potato chips (which may include bags of different physical conditions or orientations, or even bags associated with a brand and / or manufacturer of chips) may be processed to train a machine learning system.

[0107] Processing block 304 may include separating the plastic pieces into one or more fractions. There are various ways to create these fractions. One way is to create a first layer based on the major elements, then a second and even a third layer based on the minor elements. For example, fractions can be determined first by type of polymer, then bifurcated into inorganic elements such as aluminum and zinc. Other exemplary fractions can then be created for blends of polymers, bifurcated into their inorganic elemental composition. There are also computational techniques, such as principal component analysis, K-means clustering, unsupervised learning, and semi-supervised learning, to perform this type of clustering to determine fractions. Fractions are further defined herein.

[0108] In process block 305, after the fractions have been determined, the plastic pieces associated with the fractions can be separated (e.g., manually) to create a control group for each fraction. Because each fraction was measured with a sensor system, each control group contains chemical information about the pieces. A vision system (e.g., vision system 110) can be used to train a machine learning system to identify those fractions. Using this method, the chemical data in the plastic is converted into visual features that the machine learning system can learn and classify. Using system 100 to perform classification based on visual images also separates plastics by chemical composition. This method works when two objects look different and have different chemical compositions. When two objects are the same or very similar but have different chemical compositions, two or more sensor systems (e.g., VIS and XRF, etc.) may be used to perform the classification.

[0109] Because the determined fractions may comprise any desired variety of specific organic and / or inorganic elements or molecules, process 300 may be utilized to train a machine learning system implemented within the sorting system to be configured to separate a heterogeneous mixture of different plastic pieces to generate at least one fraction comprising one or more different types or classes of plastic pieces. For example, if the machine learning system is trained to identify any plastic pieces comprising a specified combination of organic and / or inorganic elements or molecules, then once sorting is complete, the sorted fractions may contain plastic pieces that are not all identical (i.e., multiple plastic chip bags associated with different brands of chips, since each plastic chip bag is comprised of organic and / or inorganic elements or molecules defined by a predetermined fraction).

[0110] FIG. 7 illustrates a flow chart diagram illustrating an exemplary embodiment of a process 3500 for sorting / separating material pieces utilizing a vision system and / or one or more sensor systems, according to certain embodiments of the present disclosure. The process 3500 may be performed to sort a heterogeneous mixture of plastic pieces into any combination of predefined types, classes, and / or fractions. The process 3500 may be configured to operate within any embodiment of the present disclosure described herein, including the system 100 of FIG. 1. The operations of the process 3500 may be performed by hardware and / or software, including within a computer system (e.g., computer system 3400 of FIG. 9) that controls the system (e.g., computer system 107, vision system 110, and / or sensor system 120 of FIG. 1). In process block 3501, material pieces may be deposited on a conveyor system. In process block 3502, the position of each material piece on the conveyor system is detected to track each material piece as it moves through the system 100. This may be performed by the vision system 110 (e.g., by distinguishing the piece of material from the underlying conveyor system material while communicating with the conveyor system position detector (e.g., position detector 105)). Alternatively, the material tracker 111 may be used to track the piece of material. Or, a system may be used that can generate a light source (including but not limited to visible light, UV, and IR) and has a detector that can be used to identify the location of the piece of material. In processing block 3503, sensed information / characteristics of the piece of material are captured / obtained as the piece of material moves near one or more of the vision system and / or sensor system. In processing block 3504, the vision system as described above (e.g., implemented in the computer system 107) may perform pre-processing of the captured information, and the captured information may be utilized to detect (extract) information of each of the pieces of material (e.g., from the background (e.g., conveyor belt); in other words, pre-processing may be utilized to identify the difference between the piece of material and the background). Well-known image processing techniques such as dilation, thresholding, contouring, etc. may be utilized to identify the piece of material as distinguishable from the background.In processing block 3505, segmentation may be performed. For example, the captured information may include information about one or more pieces of material. Furthermore, when the image is captured, a particular piece of material may be located at a seam of a conveyor belt. Therefore, in such cases, it may be desirable to separate the image of the individual piece of material from the background of the image. In an exemplary technique of processing block 3505, the first step is to apply high contrast to the image, in this way the background pixels are reduced to substantially all black pixels and at least some pixels related to the piece of material are lightened to substantially all white pixels. The white image pixels of the piece of material are expanded to cover the size of the entire piece of material. Upon completing this step, the location of the piece of material is a high contrast image of all white pixels placed on a black background. A contour algorithm may then be utilized to detect the boundary of the piece of material. The boundary information is saved and the boundary location is transferred to the original image. Then, segmentation is performed on an area of ​​the original image that is larger than the previously defined boundary. In this way, the piece of material is identified and separated from the background.

[0111] In optional processing block 3506, the pieces of material may be transported along a conveyor system near a piece of material tracking device and / or sensor system to track each piece of material and / or determine the size and / or shape of the piece of material, which may be useful if an XRF system or other spectroscopic sensor is also implemented in the sorting system. In processing block 3507, post-processing may be performed. Post-processing may include resizing the captured information / data in preparation for use by a machine learning system. This post-processing may also include modifying certain characteristics (e.g., enhancing image contrast, changing the background of the image, applying filters, etc.) in a manner that enhances the machine learning system's ability to classify the pieces of material. In processing block 3509, the data may be resized. Under certain circumstances, it may be desirable to resize the data to match the data input requirements of a particular machine learning system, such as a neural network. For example, a neural network may require an image size (e.g., 225×255 pixels or 299×299 pixels) that is much smaller than the size of an image captured by a typical digital camera. Furthermore, the smaller the size of the input data, the less processing time required to perform the classification. Therefore, smaller data sizes may ultimately increase the throughput of the system 100, making it more valuable.

[0112] In processing blocks 3510 and 3511, each piece of material is identified / classified based on the sensed / detected features. For example, processing block 3510 may consist of a neural network using one or more machine learning algorithms that compare the extracted features to features stored in a previously generated knowledge base (e.g., generated during a training phase) and, based on such comparison, assigns the most matching classification to each piece of material. The algorithms of the machine learning system may process the captured information / data hierarchically using automatically trained filters. The filter responses are successfully combined at the next level of the algorithm until a probability is obtained in a final step. In processing block 3511, these probabilities can be used for each of the N classifications to determine which of the N sorting bins the respective piece of material should be sorted into. For example, each of the N classifications can be assigned to one sorting bin and the piece of material under consideration is sorted into that bin corresponding to the classification that returns a probability greater than a predefined threshold. In an embodiment of the present disclosure, such a predetermined threshold may be pre-set by a user. If neither probability is greater than a predetermined threshold, the particular piece of material may be classified as an outlier bin (eg, sorting bin 140).

[0113] Next, in process block 3512, a sorting device corresponding to the classification of the piece of material is activated. Between the time the image of the piece of material was captured and the time the sorting device was activated, the piece of material moved from near the vision system and / or sensor system to a position downstream of the conveyor system (e.g., at the conveying speed of the conveyor system). In an embodiment of the present disclosure, activation of the sorting device is timed such that the sorting device is activated when the piece of material passes a sorting device mapped to the classification of the piece of material, and the piece of material is diverted / discharged from the conveyor system to an associated sorting bin. In an embodiment of the present disclosure, activation of the sorting device may be timed by a respective position detector that detects when the piece of material passes in front of the sorting device and transmits a signal enabling activation of the sorting device. In process block 3513, a sorting bin corresponding to the activated sorting device receives the diverted / discharged piece of material.

[0114] FIG. 8 illustrates a flow chart diagram illustrating an example embodiment of a process 800 for separating pieces of material, according to certain embodiments of the present disclosure. Process 800 may be configured to operate within any embodiment of the present disclosure described herein, including system 100 of FIG. 1. Process 800 may be configured to operate in conjunction with process 3500. For example, according to certain embodiments of the present disclosure, process blocks 803 and 804 may be incorporated into process 3500 (e.g., operating in series or parallel with process blocks 3503-3510) to combine the efforts of vision system 110 implemented in conjunction with a machine learning system with a sensor system (e.g., sensor system 120) that is not implemented in conjunction with a machine learning system to classify and / or separate pieces of material.

[0115] The operations of process 800 may be performed by hardware and / or software, including within a computer system (e.g., computer system 3400 of FIG. 9) that controls a system (e.g., computer system 107 of FIG. 1). In process block 801, pieces of material may be deposited on a conveyor system. Then, in optional process block 802, the pieces of material may be transported along the conveyor system within the vicinity of a piece of material tracker and / or an optical imaging system to track each piece of material and / or determine the size and / or shape of the piece of material. In process block 803, as the pieces of material move near the sensor system, they may be interrogated or stimulated with EM energy (waves) or other types of stimuli appropriate for the particular type of sensor technology utilized in the sensor system. In process block 804, physical properties of the pieces of material are sensed / detected and captured by the sensor system. In process block 805, for at least some of the pieces of material, the type of material is identified / classified based (at least in part) on the captured properties, and may be combined with classification by a machine learning system in conjunction with the vision system 110.

[0116] Next, in process block 806, if separation of the material pieces is to be performed, a separation device corresponding to the classification of the material pieces is activated. Between sensing the material piece and activating the separation device, the material piece has moved from near the sensor system to a position downstream of the conveyor system at the conveying speed of the conveyor system. In certain embodiments of the present disclosure, activation of the separation device is timed such that when the material piece passes a separation device mapped to the classification of the material piece, the separation device is activated and the material piece is diverted / discharged from the conveyor system to its associated separation bin. In certain embodiments of the present disclosure, activation of the separation device may be timed by a respective position detector that detects when the material piece passes in front of the separation device and transmits a signal enabling activation of the separation device. In process block 807, a separation bin corresponding to the activated separation device receives the diverted / discharged material piece.

[0117] According to certain embodiments of the present disclosure, at least a portion of the systems 100 may be linked together in series to perform multiple iterations or layers of sorting. For example, when two or more systems 100 are coordinated in this manner, the conveyor system may be implemented with a single conveyor belt or multiple conveyor belts to convey the material pieces past a first vision system (and according to certain embodiments, a sensor system) configured to separate a first set of material pieces of a heterogeneous mixture of materials by a sorting device (e.g., a first automated control system 108 and associated one or more sorting devices 126...129) into a first set of one or more containers (e.g., sorting bins 136...139), and then convey the material pieces past a second vision system (and according to certain embodiments, another sensor system) configured to separate a second set of material pieces of a heterogeneous mixture of materials by a second sorting device into a second set of one or more sorting bins. Further discussion of such multi-stage sorting is described in U.S. Patent Application Publication No. 2022 / 0016675, which is incorporated herein by reference.

[0118] Such a series of systems 100 may include any number of such systems linked together in such a manner, and according to certain embodiments of the present disclosure, each successive system may be configured to separate a different category or type of material than the previous system.

[0119] According to various embodiments of the present disclosure, different types, classes, or fractions of material can be classified by different types of sensors, each for use in a machine learning system, and combined to classify pieces of material in a scrap or waste stream.

[0120] According to various embodiments of the present disclosure, data (e.g., spectral data) from two or more sensors can be combined using a single or multiple machine learning systems to perform classification of the pieces of material.

[0121] According to various embodiments of the present disclosure, multiple sensor systems can be mounted on a single conveyor system, and each sensor system can utilize a different machine learning system. According to various embodiments of the present disclosure, multiple sensor systems can be mounted on different conveyor systems, and each sensor system can utilize a different machine learning system.

[0122] Certain embodiments of the present disclosure may be configured to produce chunks of material after fractionation that have less than a predetermined weight or volume percent content of a particular element or material.

[0123] According to various embodiments of the present disclosure, any combination of different types of sensor systems can be utilized to identify / classify and possibly separate the materials disclosed herein. For example, each imaging or spectroscopic sensor disclosed herein can be used to generate data from information / characteristics sensed from a piece of material to be processed by a machine learning system specific to that sensor system. Alternatively, any sensor system can be used without processing by a machine learning system, with processing by a machine learning system, or a combination of both.

[0124] According to various embodiments of the present disclosure, different types, classes, and / or fractions of material may be classified by different types of sensor systems, each for use in a machine learning system, and combined to classify pieces of material in a waste stream.

[0125] 9, a block diagram is shown illustrating a data processing ("computer") system 3400 in which aspects of embodiments of the present disclosure may be implemented. (The terms "computer", "system", "computer system", and "data processing system" may be used interchangeably herein). The computer system 107, the automatic control system 108, aspects of the sensor system 120, and / or the vision system 110 may be configured similarly to the computer system 3400. The computer system 3400 may use a local bus 3405 (e.g., a peripheral component interconnect ("PCI") local bus architecture). Any suitable bus architecture may be utilized, such as Accelerated Graphics Port ("AGP") or Industry Standard Architecture ("ISA"), among others. One or more processors 3415, volatile memory 3420, and non-volatile memory 3435 may be connected to the local bus 3405 (e.g., via a PCI bridge (not shown)). An integrated memory controller and cache memory may be coupled to the one or more processors 3415. The one or more processors 3415 may include one or more central processor units and / or one or more graphics processor units 3401 and / or one or more tensor processing units. Additional connections to the local bus 3405 may be made through direct component interconnections or add-in boards. In the illustrated example, a communications (e.g., network (LAN)) adapter 3425, an I / O (e.g., small computer system interface ("SCSI") host bus) adapter 3430, and an expansion bus interface (not shown) may be connected to the local bus 3405 by component direct connections. An audio adapter (not shown), a graphics adapter (not shown), and a display adapter 3416 (coupled to a display 3440) may be connected to the local bus 3405 (e.g., by an add-in board inserted into an expansion slot).

[0126] User interface adapter 3412 may provide a connection for a keyboard 3413 and mouse 3414, a modem (not shown), and additional memory (not shown). I / O adapter 3430 may provide a connection for a hard disk drive 3431, a tape drive 3432, and a CD-ROM drive (not shown).

[0127] An operating system may run on one or more processors 3415 and be used to coordinate and control various components within the computer system 3400. In FIG. 9, the operating system may be a commercially available operating system. An object-oriented programming system (e.g., Java, Python, etc.) may run in conjunction with the operating system and provide calls to the operating system from programs (e.g., Java, Python, etc.) running on the system 3400. Instructions for the operating system, object-oriented operating system, and programs may be located on a non-volatile memory 3435 storage device, such as a hard disk drive 3431, and may be loaded into volatile memory 3420 for execution by the processor 3415.

[0128] Those skilled in the art will appreciate that the hardware in FIG. 9 may vary depending on the implementation. Other internal hardware or peripheral devices, such as flash ROM (or equivalent non-volatile memory) or optical disk drives, may be used in addition to or in place of the hardware depicted in FIG. 9. Also, any of the processes disclosed herein may be applied to a multi-processor computer system or performed by multiple such systems 3400. For example, training of a machine learning system may be performed by a first computer system 3400, while the operations of system 100 for classification may be performed by a second computer system 3400.

[0129] As another example, computer system 3400 may be a stand-alone system configured to be bootable independent of some type of network communication interface, regardless of whether computer system 3400 includes any type of network communication interface. As a further example, computer system 3400 may be an embedded controller configured with ROM and / or flash ROM that provides non-volatile memory for storing operating system files or user generated data.

[0130] 9 and above-described examples are not meant to imply architectural limitations. Moreover, the computer program format of the disclosed aspects may reside on any computer readable storage medium (i.e., floppy disk, compact disk, hard disk, tape, ROM, RAM, etc.) for use by a computer system.

[0131] As described herein, embodiments of the present disclosure may be implemented to perform various functions described for identifying, tracking, sorting, and / or separating pieces of material. Such functions may be implemented in hardware and / or software, such as in one or more data processing systems (e.g., data processing system 3400 of FIG. 9), such as aspects of computer system 107, vision system 110, sensor system 120, and / or automated control system 108 described above. However, the functions described herein are not limited to implementation on any particular hardware / software platform.

[0132] As will be appreciated by those skilled in the art, aspects of the present disclosure may be embodied as a system, process, method, and / or computer program product. Accordingly, various aspects of the present disclosure may take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, microcode, etc.), or an embodiment combining software and hardware aspects, generally referred to herein as a "circuit," "circuitry," "module," or "system." Additionally, aspects of the present disclosure may take the form of a computer program product embodied in one or more computer-readable storage media having computer-readable program code embodied therein. (However, any combination of one or more computer-readable media may be utilized. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium.)

[0133] A computer-readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, biological, atomic, or semiconductor system, apparatus, controller, or device, or any suitable combination thereof, and the computer-readable storage medium itself is not a transitory signal. More specific examples (non-exhaustive list) of computer-readable storage media may include an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory ("RAM") (e.g., RAM 3420 of FIG. 9), a read-only memory ("ROM") (e.g., ROM 3435 of FIG. 9), an erasable programmable read-only memory ("EPROM" or flash memory), an optical fiber, a portable compact disk read-only memory ("CD-ROM"), an optical storage device, a magnetic storage device (e.g., hard drive 3431 of FIG. 9), or any suitable combination of the foregoing. In the context of this document, a computer-readable storage medium may be any tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, controller, or device. The program code embodied on a computer readable signal medium may be transmitted using any suitable medium, including but not limited to wireless, wired, fiber optic cable, RF, etc., or any suitable combination of the foregoing.

[0134] A computer-readable signal medium may include a propagated data signal having computer-readable program code embodied therein, for example as part of a baseband or carrier wave. Such a propagated signal may take any of a variety of forms, including but not limited to electromagnetic, optical, or any suitable combination thereof. A computer-readable signal medium is not a computer-readable storage medium, but may be any computer-readable medium that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, controller, or device.

[0135] The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods, processes, and program products according to various embodiments of the present disclosure. In this regard, each block in the flowcharts or block diagrams may represent a module, segment, or portion of code, including one or more executable program instructions for implementing a specified logical function. It should also be noted that, depending on the implementation, the functions shown in the blocks may be executed in a different order than the order shown in the figures. For example, two blocks shown in succession may in fact be executed substantially simultaneously, or the blocks may be executed in the reverse order, depending on the functionality involved.

[0136] In the description herein, the flow charted techniques may be described as a series of sequential operations. The order of operations and the parties performing the operations may be freely modified without departing from the scope of the teachings. Operations may be added, removed, or modified in some manner. Similarly, the order of operations may be changed or looped. Furthermore, while processes, methods, algorithms, etc. may be described in a sequential order, such processes, methods, algorithms, or any combination thereof, may be operable to be performed in another order. Furthermore, some actions within a process, method, or algorithm may be performed simultaneously at least at some point (e.g., actions performed in parallel), or may be performed in whole, in part, or in any combination thereof.

[0137] Modules implemented in software executed by various types of processors (e.g., GPU 3401, CPU 3415) may include one or more physical or logical blocks of computer instructions that may be organized as, for example, an object, procedure, or function. However, executable files of identified modules need not be physically located together and may include disparate instructions stored in different locations that, when logically combined, incorporate the module and achieve the specified purpose of the module. In fact, a module of executable code may be a single instruction or many instructions and may be distributed across multiple different code segments, different programs, and multiple memory devices. Similarly, operational data (e.g., a material classification library described herein) may be identified and illustrated in modules herein and may be embodied in any suitable form and organized in any suitable type of data structure. The operational data may be collected as a single data set or distributed in different locations, including different storage devices. The data may provide electronic signals over a system or network.

[0138] These program instructions can be provided to one or more processors and / or controllers of a general purpose computer, special purpose computer, or other programmable data processing apparatus (e.g., a controller) such that the instructions, which execute via a processor (e.g., GPU 3401, CPU 3415) of the computer or other programmable data processing apparatus, create circuits or means for implementing the functions / acts specified in the blocks of the flowcharts and / or block diagrams to produce a machine.

[0139] It should also be noted that each block of the block diagrams and / or flowchart illustrations, and combinations of blocks in the block diagrams and / or flowchart illustrations, can be implemented by special purpose hardware-based systems (which may include, for example, one or more graphics processing units (such as GPU3401)) that perform the specified functions or operations, or by a combination of special purpose hardware and computer instructions. For example, a module can be implemented as a hardware circuit comprising custom VLSI circuits or gate arrays, off-the-shelf semiconductors such as logic chips, transistors, controllers, or other discrete components. A module may also be implemented with programmable hardware devices such as field programmable gate arrays, programmable array logic, programmable logic devices, etc.

[0140] Computer program code, i.e., instructions, for carrying out the operations of aspects of the present disclosure can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, Smalltalk, Python, C++, the "C" programming language or similar programming languages, or traditional procedural programming languages ​​such as any of the machine learning software disclosed herein. The program code may run entirely on the user's computer system as a standalone software package, partially on the user's computer system, partially on the user's computer system (e.g., the computer system used for sorting), partially on a remote computer system (e.g., the computer system used for training the sensor system, etc.), or entirely on a remote computer system or server. In the latter scenario, the remote computer system may be connected to the user's computer system via any type of network, including a local area network ("LAN") or wide area network ("WAN"), or may be connected to an external computer system (e.g., via the Internet using an Internet Service Provider).

[0141] These program instructions may also be stored on a computer-readable storage medium that can cause a computer system, other programmable data processing device, controller, or other device to function in a particular manner, such that the instructions stored on the computer-readable medium produce an article of manufacture that includes instructions that implement the functions / acts specified in the flowchart and / or block diagram blocks.

[0142] The program instructions may also be loaded into a computer, other programmable data processing device, controller, or other device to cause the computer, other programmable device, or other device to perform a series of operational steps to generate a computer-implemented process, such that the instructions, which execute on the computer or other programmable device, provide a process for implementing the functions / operations specified in the flowchart and / or block diagram blocks.

[0143] A host may include one or more databases to store and provide access to data for various implementations. Those skilled in the art will also appreciate that for security reasons, any database, system, or component of the present disclosure may include any combination of databases or components in a single location or multiple locations, and each database or system may include any of a variety of suitable security features, such as firewalls, access codes, encryption, decryption, etc. The databases may be any type of database, such as relational, hierarchical, object-oriented, etc. Common database products that may be used to implement the databases include IBM's DB2, database products available from Oracle Corporation, Microsoft Access from Microsoft Corporation, or other database products. The databases may be organized in any suitable manner, such as data tables and lookup tables.

[0144] Association of specific data (e.g., for each of the pieces of material processed by the separation system described herein) can be accomplished through any data association technique known and practiced in the art. For example, association can be performed manually or automatically. Automatic association techniques can include, for example, database lookup, database merge, GREP, AGREP, SQL, etc. The association step can be accomplished by a database merge function that uses key fields of the respective data tables of the manufacturer and retailer. The key fields divide the database according to high-level classes of objects defined in the key fields. For example, a particular class can be designated as a key field in both the first data table and the second data table, and the two data tables can be merged based on the class data in the key field. In these embodiments, the data corresponding to the key fields in each of the merged data tables is preferably the same. However, data tables that are not identical but have similar data in the key fields can also be merged using, for example, AGREP.

[0145] As used herein, reference is made to "configuring" a device or a device being "configured" to perform some function. It should be understood that this may include selecting predefined logic blocks and logically associating them to provide a particular logical function, including monitoring or control functions. It may also include programming computer software-based logic of custom controllers, wiring of discrete hardware components, or a combination of any or all of the above.

[0146] In the description herein, numerous specific details are provided, such as examples of programming, software modules, user selections, network transactions, database queries, database structures, hardware modules, hardware circuits, hardware chips, controllers, etc., to provide a thorough understanding of embodiments of the present disclosure. However, those skilled in the art will recognize that the present disclosure may be practiced without one or more of the specific details, or with other methods, components, materials, etc. In other cases, well-known structures, materials, or operations may not be shown or described in detail to avoid obscuring aspects of the present disclosure.

[0147] Those skilled in the art will appreciate that the various settings and parameters of the components of system 100 (including the neural network parameters) can be customized, optimized, and reconfigured over time based on the type of material being sorted and separated, the desired sorting and separation results, the type of equipment being used, empirical results of previous sortings, data that becomes available, and other factors.

[0148] Reference throughout this specification to "one embodiment," "embodiment," or similar terms means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present disclosure. Thus, appearances throughout this specification of "one embodiment," "in an embodiment," "embodiment," "particular embodiment," "various embodiments," and similar terms may all refer to the same embodiment, but not necessarily the same. Furthermore, the described features, structures, aspects, and / or characteristics of the present disclosure may be combined in any suitable manner in one or more embodiments. Accordingly, in some cases, one or more features may be deleted from a combination described in a claim, even if the features are initially claimed to function in a particular combination, and the combination described in the claims may be subject to subcombinations or variations of the subcombinations.

[0149] Benefits, advantages, and solutions to problems have been described herein with respect to specific embodiments. However, the benefits, advantages, solutions to problems, and elements by which the benefits, advantages, solutions occur or become more pronounced may not be construed as critical, necessary, or essential features or elements of some or all of the claims. Moreover, no element described herein is required for the practice of the disclosure unless expressly described as essential or critical.

[0150] Although this specification contains many details, these should not be construed as limiting the scope of the disclosure or the claims, but rather as descriptions of features specific to particular implementations of the disclosure. The headings in this specification may not be intended to limit the disclosure, the embodiments of the disclosure, or other matters disclosed under the heading.

[0151] As used herein, the term "or" is intended to be inclusive, such that "A or B" includes A or B, and also includes both A and B. As used herein, the term "and / or," when used in the context of a list of entities, refers to the entities present alone or in combination. Thus, for example, the phrase "A, B, C, and / or D" includes A, B, C, and D individually, but also all combinations and subcombinations of A, B, C, and D.

[0152] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. As used herein, the singular forms "a," "an," and "the" may be intended to include the plural forms as well, unless the context clearly indicates otherwise.

[0153] Corresponding structures, materials, acts, and equivalents of all means- or step-plus-function elements in the following claims may be intended to include any structure, material, or act for performing a function in combination with other claim elements that are specifically claimed.

[0154] As used herein, the terms "controller," "processor," "memory," "neural network," "interface," "sorter," "device," "pushing mechanism," "pusher device," "image sensor," "bin," "receptacle," "system," "circuit," and the like each refer to a non-common device element that would be recognized and understood by one of ordinary skill in the art, and are not used herein as ordinarily or provisionally for purposes of invoking 35 U.S.C. 112(f).

[0155] As used herein with respect to a specified characteristic or condition, "substantially" refers to a degree of deviation that is small enough so as not to measurably impair the specified characteristic or condition. The exact degree of deviation that is permitted may depend on the particular situation in some cases.

[0156] As used herein, for convenience, a plurality of items, structural elements, components, illustrative fractions, and / or materials may be presented in common lists. However, these lists should be construed as if each member of the list were individually identified as a separate and unique member. Thus, the individual members of such lists should not be construed as being de facto equivalents to other members of the same list solely based on their appearance within a common grouping, absent any indication to the contrary.

[0157] Unless otherwise defined, all technical and scientific terms used herein (such as acronyms used for polymers or chemical elements in the periodic table) have the same meaning as commonly understood by those skilled in the art to which the subject matter disclosed herein belongs. All publications, patent applications, patents, and other references mentioned herein are incorporated by reference in their entirety unless a specific text is cited. In case of conflict, the present specification, including definitions, controls. Additionally, materials, methods, and examples (e.g., listed fractions, plastics) are for illustrative purposes only and are not intended to be limiting.

[0158] Unless otherwise described herein, many of the details regarding specific materials, process acts, and circuits are conventional and can be found in computing, electronics, and software technology textbooks and other sources.

[0159] Unless otherwise indicated, all numbers expressing quantities of ingredients, reaction conditions, and so forth used in the specification and claims are to be understood as being modified in all instances by the term "about." Accordingly, unless indicated to the contrary, the numerical parameters set forth in the specification and appended claims are approximations that may vary depending upon the desired properties sought to be obtained by the presently disclosed subject matter. [Explanation of symbols]

[0160] 100 Systems 101 Material pieces 102 Hopper 103 Conveyor systems, conveyor belts 104 Conveyor system motor 105 Position detector 106 Tumbler / Vibrator / Singulator 107 Computer Systems 108 Automatic Control System 109 Live-action camera 110 Optical Recognition Systems, Vision Systems 111 Material piece tracking device, material tracking device 112 Control System 120 Sensor System 121 Radiation source 122 Power supply 124 Detector 125 Detector Electronics 126…129 Sorting device 136…139 Sorting bin 140 Bottles, waste sorting containers 201 Material pieces 203 Conveyor System 401 Plastic Pieces 403 Conveyor System 410 VIS 411 XRF 412 NIR 413 MWIR 3400 Data processing systems, computer systems 3401 Graphics Processor Unit, GPU 3405 Local Bus 3412 User Interface Adapter 3413 Keyboard 3414 Mouse 3415 Processor (CPU) 3416 Display Adapter 3420 Volatile memory, RAM 3425 Network (LAN) Adapter 3430 I / O Adapter 3431 Hard Disk Drive 3432 Tape Drive 3435 Non-volatile memory, ROM 3440 Display

Claims

1. capturing a first visual image of a first piece of material, resulting in a first image data packet relating to the first piece of material; capturing a second visual image of a second piece of material, resulting in a second image data packet relating to the second piece of material, the first piece of material having a first chemical characteristic and the second piece of material having a second chemical characteristic different from the first chemical characteristic; processing the first and second image data packets with a machine learning system that has been pre-trained to visually identify pieces of material having the distinct chemical characteristics; classifying, using the machine learning system, the first and second pieces of material into two different classifications in response to the learned visual discrimination between the pieces of material having the different chemical characteristics; separating the first pieces of material from the second pieces of material in response to the classification; Including, the piece of material is a piece of plastic; measuring the first chemical characteristic comprising spectral data measured from at least one sample of the same type of plastic piece as the first plastic piece by a plurality of different sensor systems; measuring the second chemical characteristic by the plurality of different sensor systems, the second chemical characteristic comprising spectral data measured from at least one sample of a plastic piece of the same type as the second plastic piece. method.

2. The method of claim 1 , wherein the spectral data relates to the non-visible spectrum.

3. The method of claim 1 , wherein the plurality of different sensor systems is selected from the group consisting of near infrared ("NIR"), mid-wave infrared ("MWIR"), and x-ray fluorescence ("XRF") systems.

4. The plurality of different sensor systems may include infrared ("IR"), Fourier transform IR ("FTIR"), forward looking infrared ("FLIR"), very near infrared ("VNIR"), near infrared ("NIR"), short wavelength infrared ("SWIR"), long wavelength infrared ("LWIR"), mid wavelength infrared ("MWIR" or "MIR"), x-ray transmission ("XRT"), gamma ray, ultraviolet ("UV"), x-ray fluorescence ("XRF"), laser induced breakdown spectroscopy ("LIBS"), and the like. "), Raman spectroscopy, anti-Stokes Raman spectroscopy, gamma ray spectroscopy, hyperspectral spectroscopy (e.g., beyond visible wavelengths), acoustic spectroscopy, NMR spectroscopy, microwave spectroscopy, terahertz spectroscopy, differential scanning calorimetry ("DSC"), thermogravimetric analysis ("TGA"), capillary and rotational rheometry, optical and scanning electron microscopy ("SEM"), and chromatography.

5. A method of manufacturing a device, comprising: capturing a first visual image of a first piece of material, resulting in a first image data packet relating to the first piece of material; capturing a second visual image of a second piece of material, resulting in a second image data packet relating to the second piece of material, the first piece of material having a first chemical characteristic and the second piece of material having a second chemical characteristic different from the first chemical characteristic; processing the first and second image data packets with a machine learning system that has been pre-trained to visually identify pieces of material having the distinct chemical characteristics; classifying, using the machine learning system, the first and second pieces of material into two different classifications in response to the learned visual discrimination between the pieces of material having the different chemical characteristics; separating the first pieces of material from the second pieces of material in response to the classification; Including, the piece of material is a piece of plastic; measuring said first chemical characteristics including organic and inorganic elements or molecules from at least one sample of plastic pieces of the same type as the first plastic piece; measuring said second chemical characteristics, including organic and inorganic elements or molecules, from at least one sample of plastic pieces of the same type as the second plastic piece. method.

6. 2. The method of claim 1, wherein the plastic pieces are selected from the group consisting of: Type #1 Polyethylene Terephthalate ("PET"), Type #2 High Density Polyethylene ("HDPE"), Type #3 Polyvinyl Chloride ("PVC"), Type #4 Low Density Polyethylene ("LDPE"), Type #5 Polypropylene ("PP"), Type #6 Polystyrene ("PS"), and Type #7 Other polymers.

7. The method of claim 1 , wherein the first piece of material comprises polyvinyl chloride.

8. The method of claim 1 , wherein the two different classes are different fractions.

9. a camera configured to capture a first visual image of a first piece of material resulting in a first image data packet for the first piece of material and a second visual image of a second piece of material resulting in a second image data packet for a second piece of material, the first piece of material having a first chemical characteristic and the second piece of material having a second chemical characteristic different from the first chemical characteristic; a data processing system configured to process the first and second image data packets with a machine learning system previously trained to visually distinguish between pieces of material having the different chemical characteristics, the machine learning system classifying the first and second pieces of material into two distinct fractions in response to the trained visual discrimination between the pieces of material having the different chemical characteristics; and a separation device configured to separate the first pieces of material from the second pieces of material according to the fraction; Equipped with the piece of material is a piece of plastic; a plurality of different sensor systems configured to measure the first chemical characteristic comprising spectral data for a non-visible spectrum of at least one sample of a plastic piece of the same type as the first plastic piece; the plurality of different sensor systems are configured to measure the second chemical characteristic comprising spectral data for a non-visible spectrum measured from at least one sample of a plastic piece of the same type as the second plastic piece; system.

10. 10. The system of claim 9, wherein the plurality of different sensor systems are selected from the group consisting of near infrared ("NIR"), mid-wave infrared ("MWIR"), and x-ray fluorescence ("XRF") systems.

11. The plurality of different sensor systems may include infrared ("IR"), Fourier transform IR ("FTIR"), forward looking infrared ("FLIR"), very near infrared ("VNIR"), near infrared ("NIR"), short wavelength infrared ("SWIR"), long wavelength infrared ("LWIR"), mid wavelength infrared ("MWIR" or "MIR"), x-ray transmission ("XRT"), gamma ray, ultraviolet ("UV"), x-ray fluorescence ("XRF"), laser induced breakdown spectroscopy ("LIBS"), and the like. ), Raman spectroscopy, anti-Stokes Raman spectroscopy, gamma ray spectroscopy, hyperspectral spectroscopy (e.g., beyond visible wavelengths), acoustic spectroscopy, NMR spectroscopy, microwave spectroscopy, terahertz spectroscopy, differential scanning calorimetry ("DSC"), thermogravimetric analysis ("TGA"), capillary and rotational rheometry, optical and scanning electron microscopy ("SEM"), and chromatography.

12. A camera configured to capture a first visual image of a first piece of material resulting in a first image data packet relating to the first piece of material and a second visual image of the second piece of material resulting in a second image data packet relating to the second piece of material, the first piece of material having a first chemical characteristic and the second piece of material having a second chemical characteristic different from the first chemical characteristic; a data processing system configured to process the first and second image data packets with a machine learning system previously trained to visually distinguish between pieces of material having the different chemical characteristics, the machine learning system classifying the first and second pieces of material into two distinct fractions in response to the trained visual discrimination between the pieces of material having the different chemical characteristics; and a separation device configured to separate the first pieces of material from the second pieces of material according to the fraction; Equipped with the piece of material is a piece of plastic; a plurality of different sensor systems configured to measure the first chemical characteristic comprising organic and inorganic elements or molecules from at least one sample of plastic pieces of the same type as the first plastic piece; the plurality of different sensor systems are configured to measure the second chemical characteristics comprising organic and inorganic elements or molecules from at least one sample of a plastic piece of the same type as the second plastic piece; The plastic pieces are selected from the group consisting of: Type #1 Polyethylene Terephthalate ("PET"), Type #2 High Density Polyethylene ("HDPE"), Type #3 Polyvinyl Chloride ("PVC"), Type #4 Low Density Polyethylene ("LDPE"), Type #5 Polypropylene ("PP"), Type #6 Polystyrene ("PS"), and Type #7 Other polymers; system.

13. determining a chemical signature of each of the different plastic piece mixtures using a plurality of different sensor systems; capturing a visual image of each said plastic piece; digitally associating the chemical signature of each plastic piece with the visual image; Determining a specific fraction for separating the plastic pieces; using the visual image to identify which of the plastic pieces in the mixture have a chemical signature corresponding to the particular fraction; training a machine learning system to visually identify plastic pieces that fall into the particular fraction, said training being performed using a control group created from the identified plastic pieces; The method includes:

14. The method of claim 13 , wherein the control group is comprised of captured visual image data of each of the identified plastic pieces.

15. The method of claim 13 , wherein the fraction is composed of a specific combination of organic and inorganic elements or molecules.

16. 14. The method of claim 13, wherein the plurality of different sensor systems are selected from the group consisting of near infrared ("NIR"), mid-wave infrared ("MWIR"), and x-ray fluorescence ("XRF") systems.

17. 17. The method of claim 16, wherein the mixture of different plastic pieces is selected from the group consisting of: Type #1 Polyethylene Terephthalate ("PET"), Type #2 High Density Polyethylene ("HDPE"), Type #3 Polyvinyl Chloride ("PVC"), Type #4 Low Density Polyethylene ("LDPE"), Type #5 Polypropylene ("PP"), Type #6 Polystyrene ("PS"), and Type #7 Other polymers.

18. The plurality of different sensor systems may include infrared ("IR"), Fourier transform IR ("FTIR"), forward looking infrared ("FLIR"), very near infrared ("VNIR"), near infrared ("NIR"), short wavelength infrared ("SWIR"), long wavelength infrared ("LWIR"), mid wavelength infrared ("MWIR" or "MIR"), x-ray transmission ("XRT"), gamma ray, ultraviolet ("UV"), x-ray fluorescence ("XRF"), laser induced breakdown spectroscopy ("LIBS"), and the like. "), Raman spectroscopy, anti-Stokes Raman spectroscopy, gamma ray spectroscopy, hyperspectral spectroscopy (e.g., beyond visible wavelengths), acoustic spectroscopy, NMR spectroscopy, microwave spectroscopy, terahertz spectroscopy, differential scanning calorimetry ("DSC"), thermogravimetric analysis ("TGA"), capillary and rotational rheometry, optical and scanning electron microscopy ("SEM"), and chromatography.

Images