Textile identification system and method for identifying textile object

Abstract

The invention relates to a textile identification system for identifying textile articles. Wherein the identification system comprises a conveyor for moving an associated textile article in a first direction, a visible light emitter for emitting light within a visible wavelength, a color image sensor for measuring light within the visible wavelength, an infrared emitter for emitting light within an infrared wavelength, and a controller for controlling the color image sensor to measure light within the infrared wavelength. The invention relates to a system for measuring radiation in an X-ray or gamma-ray spectrum, comprising an infrared sensor for measuring light in an infrared wavelength, a radiation emitter for emitting radiation in the X-ray or gamma-ray spectrum, a radiation sensor for measuring radiation in the X-ray or gamma-ray spectrum, a database system adapted to at least acquire, store and process data measured by the sensors. The invention further relates to a method for identifying a textile article, and to a computer program and a computer-implemented method for classifying a textile article.

Description

TECHNICAL FIELD

[0001] The present invention relates to a textile identification system for identifying textile objects, wherein the identification system comprises: - a conveyor for moving the associated textile objects in a first direction, - a visual or optical identification system, - an infrared emitter for emitting light in the infrared wavelength spectrum, preferably in the near infrared wavelength spectrum, - an infrared sensor for measuring light in the infrared wavelength spectrum, preferably in the near infrared wavelength spectrum, - a radiation emitter for emitting radiation in the X-ray spectrum or gamma-ray spectrum, - a database system adapted to at least acquire, store and process data measured by at least: - a color image sensor, and - an infrared sensor.

[0002] Furthermore, the present invention relates to a method for identifying textile objects, preferably for use in the above-mentioned identification system, which method comprises the following steps: - conveying the textile objects in a first direction, preferably by means of a conveyor, - identifying the textile objects using a visual or optical system, - emitting infrared radiation to the objects by means of at least one infrared emitter, - measuring the reflected light by means of at least one infrared sensor, - emitting X-ray or gamma-ray radiation to the textile objects by means of a radiation emitter, - collecting the data measured by the color image sensor and the infrared sensor, - identifying at least one or more characteristics of the textile objects, - classifying the textile objects, preferably into predefined classification groups.

[0003] Furthermore, the present invention relates to a computer program for classifying textile objects and a computer-implemented method. BACKGROUND

[0004] There is a global need for greener, more sustainable solutions, and a huge field in which sustainable development can be optimized and utilized lies in the field of textile recycling, in particular mechanical recycling and / or reuse of textiles. There is therefore a high demand for systems and methods for optimizing the possibilities for textile reuse.

[0005] Within the field of textile recycling, in particular within the field of textile fiber recycling, providing and performing such recycling and / or reusing processes can seem simple and straightforward. However, the need for an ideal preliminary treatment before various textiles can be mechanically treated into reusable fibers is both essential and extremely challenging.

[0006] Before textiles can be reasonably recycled and / or reused, one or more characteristics of all various textiles must be correctly identified in order for the selected textiles for recycling and / or reusing to be correctly identified, classified and / or sorted before the subsequent recycling process.

[0007] Today, solutions for the preliminary treatment of identifying, classifying and / or sorting textiles for recycling and / or reusing purposes are available, however, no known system and method provides the optimal solution.

[0008] The known problem is that even if a large amount of textiles is collected for the purpose of recycling textile fibers, most of the collected textiles never end up in the fiber recycling process simply because the characteristics of the textiles could not be correctly identified. This means that the textiles were not correctly sorted and, ultimately, the final process of making the textiles into reusable fibers could not be performed because if the fibers are not correctly opened in the recycling process, they are destroyed.

[0009] The undesirable consequence is that the textiles intended for recycling and / or reusing are either incinerated or landfilled, wasting resources and damaging the environment instead of providing reusable and more sustainable textile fibers.

[0010] A system and method for the classification and sorting of materials is described in US2022 / 0161298 A1, in particular the classification and sorting of plastic materials is described in detail. The sorting is performed in order to obtain a classification based on different chemical characteristics of the material pieces.

[0011] However, in general, the term "other materials" mentioned in this application includes textiles, but it is not indicated which emitters and detectors are used for other materials in order to establish a classification based on other parameters than different chemical characteristics.

[0012] US2022 / 0161298 A1 discloses a system and method which is defined by the abstract and as defined in the preamble of claims 1 and 8, respectively. The method and system utilize emitting sources and detectors to sense / detect one or more characteristics from the material pieces for classification of different chemical characteristics.

[0013] The application does not teach the use of emitters emitting visible light, which have a spectrum within 300-1000 nm. The application also does not disclose an optical recognition system in the form of a color image sensor for measuring color to sort textile items according to color.

[0014] Furthermore, the application does not teach the use of X-ray emitters and the use of radiation sensors arranged below a transport machine on which textile items are transported, which are set up to measure radiation within the X-ray spectrum or gamma spectrum by using absorption spectroscopy.

[0015] Therefore, the application also does not teach the use of a database system to acquire, store and process data from the radiation sensors or color image sensors for measuring color.

[0016] WO 2022 / 028746 A1 discloses a textile recognition system using X-ray recognition.

[0017] US 7,564,942 B2 discloses an X-ray CT device for CT scanning of a person lying on a movable bed.

[0018] KONSTANTINIDIS FOTIOS K et al.: "Multi-sensor cyber-physical sorting system (CPSS) based on industry 4.0 principles: a versatile approach", PROCEEDINGS OF COMPUTER SCIENCE, ELSEVIER, AMSTERDAM, NL, vol. 217, January 2023 (2023-01), pages 227-237, XP087246859, ISSN: 1877-0509, DOI: 10.1016 / J.PROCS.2022.12.218 discloses a sorting system which utilizes a robot to receive signals from a camera about the position of an item on a transport machine.

[0019] Therefore, an improved system and method, preferably an efficient and reliable system and method, which is suitable for recognizing, classifying and / or sorting textiles would be advantageous and, in particular, a more optimized, reliable, efficient, precise, economic and sustainable system and method would be advantageous.

[0020] Object of the invention It is an object of the present invention to provide an alternative to the prior art, which relates to a system and method which efficiently sorts textile items based on more parameters.

[0021] In particular, it can be seen as an object of the present application to provide a system and a method that solves the above-mentioned problems by providing an optimized preliminary treatment of the textile to be mechanically recycled and / or reused, which can include the identification of at least the color, the material composition and the presence of hard materials such as buttons and zippers, and which is optimized to improve the fiber yield and the fiber quality in the subsequent recycling process. SUMMARY

[0022] The above-mentioned object, as well as several other objects, is thus achieved, according to a first aspect of the present application, by providing a textile identification system as described in the abstract and as defined in the preamble of claim 1, characterized in that the system further comprises: - a visible light emitter for emitting light in the visible wavelength spectrum, the visible light emitter having a spectrum in the range of 300-1000 nm, and - a radiation sensor for measuring radiation in the X-ray spectrum or the gamma spectrum by using absorption spectroscopy, such as X-ray or gamma-ray absorption spectroscopy, the radiation sensor being located below the transport machine, and the part of the transport machine above the radiation sensor being made of a suitable material that does not block X-rays or gamma rays, and - a visual or optical identification system comprising a color image sensor for measuring light in the visible wavelength spectrum, preferably an RGB sensor and / or a CMOS sensor and / or a CCD sensor, and - a database system adapted to acquire, store and process data measured by the radiation sensor, wherein the radiation sensor is located below the transport machine, and the part of the transport machine above the radiation sensor is made of a suitable material that does not block X-rays or gamma rays.

[0023] Furthermore, the above-mentioned object, as well as several other objects, is achieved, according to a second aspect of the present application, by providing a method for identifying a textile object as described in the abstract and as defined in the preamble of claim 8, characterized in that the method further comprises the steps of: - emitting light to the textile object by at least one first light emitter, the first light emitter emitting light in the visible light spectrum, the visible light emitter having a spectrum in the range of 300-1000 nm, - providing a visual or optical identification system in the form of a color image sensor for measuring light in the visible wavelength spectrum, preferably an RGB sensor and / or a CMOS sensor and / or a CCD sensor, - measuring the color distributed over the textile object by at least one color image sensor, preferably an RGB sensor and / or a CMOS sensor and / or a CCD sensor, - providing a radiation sensor at a position below the conveyor (10), - providing a conveyor or a part of the conveyor which is located above the radiation sensor (45) made of a suitable material which does not block X-rays or gamma rays, - measuring the radiation within the X-ray spectrum or gamma spectrum by using absorption spectroscopy, such as X-ray or gamma absorption spectroscopy, by using the radiation sensor, - collecting the data measured by the radiation sensor, - storing the measured and collected data in a database, - identifying at least: - the color, - the material composition, and - the presence of hard materials, such as buttons and zippers, and in the method - the steps of the method can be performed in any order and / or simultaneously.

[0024] The advantages of the present invention are in particular, but not only, that a system is provided in which the information of the textile for recycling and / or re-use is not only provided by a single sensor, but by at least three independent sensors, each providing essential information of the textile item. Furthermore, and most characteristically - the synergy between these three specific sensors has not been seen before and proves that this synergy provides unique and exclusive information in the identification, classification and sorting of textiles. The device provides a system which is suitable for identifying any necessary information of the textile which is needed for optimizing the subsequent mechanical recycling.

[0025] Thus, the system consists of a combination of three specific sensor types for the classification of the textile item: an X-ray sensor, an infrared sensor and a color image sensor. The textile item is transported on a conveyor and passes below these sensors, each of which collects an "image" of the textile under different electromagnetic spectra.

[0026] For example, the system with the three specific sensors enables to conclude with a certain probability that the textile is a pair of jeans, which none of the sensors alone could do.

[0027] With this unique combination of sensors, the identification, classification and sorting of the textile is enabled and the fiber yield and quality in the subsequent recycling process is optimized.

[0028] It is to be understood that the specific order of the sensors is not relevant in the context of the present invention, only the presence of the sensors is essential.

[0029] Furthermore, it should be understood that the system can be combined with more sensors and / or other more devices, the invention is not limited to having to contain only these three sensors, as long as these three sensors are present in the system, more sensors of the same type and / or other types and / or other devices can be present in the system.

[0030] In the context of the invention, a "textile item" can be any type of textile material, such as: garments / clothing / attire / wearables / decorations, home textiles, outdoor equipment, non-woven materials, cleaning textiles, workwear, footwear, etc. The type of textile should therefore not be considered as limiting to the invention.

[0031] By exposing the textile item to three different types of electromagnetic radiation and measuring its reflectance and / or absorbance using dedicated sensors, the following sensor outputs can be provided: - a color image sensor focused on visible light.

[0032] This sensor allows color detection by measuring the color distribution on the textile.

[0033] In addition, this sensor can extract some features from the image, such as reflective materials, polymer prints, accessories (buttons, zippers, other attachments, etc.) or fabric types (woven, knitted, terry, denim, etc.), patterns, and / or preferably also provide information on the material composition properties (at the molecular level).

[0034] It is possible to use machine learning algorithms to optimize the information extraction.

[0035] - an infrared sensor focused on the infrared spectrum, preferably using reflectance / diffuse reflectance spectroscopy.

[0036] This sensor enables material detection by measuring the reflected light spectrum.

[0037] This sensor can detect the material composition of the textile item.

[0038] In addition, the infrared sensor can detect moisture or specific pollutant chemicals, such as oil stains on workwear, mold or dirt on old textiles.

[0039] It is possible to use machine learning-driven classification algorithms in a multi-level classification scheme.

[0040] - an X-ray or gamma-ray sensor focused on the X-ray or gamma-ray spectrum, preferably using absorption spectroscopy, such as X-ray or gamma-ray absorption spectroscopy.

[0041] The radiation source emits X-rays or gamma rays that penetrate the textile item. The rays are measured by a ray sensor located below the textile item.

[0042] The ray sensor allows to detect density and thus also to detect hard material parts, such as accessories.

[0043] Furthermore, the sensor can be used to infer the quality of the textile item.

[0044] It is possible to train a machine learning algorithm to detect common accessories, such as buttons, zippers, pens or tools.

[0045] The database system is adapted to interpret the sensor output and to perform a classification using a probabilistic machine learning model.

[0046] The system thus also provides optimal conditions for inferring information using a probabilistic model.

[0047] By using different sensor outputs, several different characteristics of the textile item can be derived with different degrees of certainty. Using a probabilistic machine learning model (Bayesian network), it is possible to estimate the probability that the textile item has certain characteristics.

[0048] In the context of the present invention, a "color image sensor" can be understood as a sensor that is capable of measuring visible light along a straight line (line scan) or in an area (area scan) to provide a digital color image in RGB format or a similar digital color image presentation.

[0049] An example of a color image sensor that is preferably used in the present invention is an RGB sensor (a color sensor for identifying / detecting the color of a material in RGB (red, green, blue) format) and / or a CMOS sensor (a color sensor for acquiring a color image of a textile item in RGB (red, green, blue) format to identify / detect the color of a material).

[0050] Furthermore, the system provides the opportunity to utilize all or at least as many textiles as possible for recycling and / or reusing, since all necessary characteristics of the textiles are correctly identified, the textiles can also be correctly classified and ultimately an optimized and reliable sorting of the textiles is achieved, so that the sorted textiles are ready for recycling and / or reusing by means of opening without destroying them.

[0051] This means that the potential of the sorting system is fully utilized, since almost all textiles can be optimally processed and thus can be mechanically recycled and / or reused in the most sustainable way possible. In this way, an extremely environmentally friendly, sustainable and economically viable system is provided.

[0052] The recycling process of the fibers, for example the opening in the mechanical recycling process, should not be considered part of the present invention, but rather a process that is suitable to be performed after the textile items have been sorted by the system. However, the present invention can also be used to sort textiles for other recycling processes, such as chemical recycling.

[0053] In one embodiment of the invention, the database system is adapted to identify at least one predetermined characteristic of the associated textile item.

[0054] The advantage of this embodiment is, inter alia, but not exclusively, to provide a system capable of reliably identifying characteristics of a textile item.

[0055] In one embodiment of the invention, the identification system further comprises an algorithm adapted to classify the associated textile item based on the at least one predetermined characteristic.

[0056] The advantage of this embodiment is, inter alia, but not exclusively, to provide an optimized system for classifying a textile item into a predetermined classification group.

[0057] The data is processed by at least one algorithm to extract useful information from the data measured by the sensors - i.e. to transform from raw data to information that can serve as a basis for classification.

[0058] In one embodiment of the invention, the identification system is adapted to provide a command for sorting the associated textile item based on the classification.

[0059] The advantage of this embodiment is, inter alia, but not exclusively, to provide an optimized system for sorting a textile item into a predetermined sorting group.

[0060] In one embodiment of the invention, the algorithm is a machine learning algorithm adapted to be trained by data collected from the database.

[0061] The advantage of this embodiment is, inter alia, but not exclusively, to provide a system that improves itself every time a textile item passes through the system. By employing a machine learning algorithm, the system will be able to identify characteristics of a textile item, classify a textile item and / or sort a textile item with increasing accuracy over time until it can reach an accuracy of approximately 100% correct and reliable.

[0062] In one embodiment of the invention, the radiation sensor is arranged below the conveyor, and wherein the part of the conveyor located above the radiation sensor is made of a suitable material that does not block X-rays.

[0063] The advantage of this embodiment is in particular, but not only, that it allows extracting information about the textile article through the material of the conveyor (belt) and still be able to provide an optimized and usable information from the X-ray or gamma-ray sensor, which means that even if the measurement of the sensor is first through other materials and then the textile article, it is only possible to obtain a measurement result about the textile, so that the conveyor data of the measurement does not influence the textile data. The material of the conveyor (belt) forms a "constant noise" in the data / pictures collected by the sensor, and this constancy makes it easier to remove the "noise", so that even if the measurement is through the conveyor material, it is possible to obtain an optimized image of the textile through the radiation sensor.

[0064] In one embodiment of the invention, the conveyor comprises a gap, and wherein the radiation sensor is arranged below the gap of the conveyor.

[0065] The advantage of this embodiment is in particular, but not only, that it allows extracting information about the textile article transported on the conveyor, but without the need to measure the information through the material of the conveyor (belt), since the measurement is made at the gap, which means that there is no interfering data in the measurement.

[0066] In one embodiment of the invention, the position of the textile article on the conveyor is measured by at least one sensor, and wherein a robot, preferably a robotic arm, is adapted to collect the article for sorting.

[0067] The advantage of this embodiment is in particular, but not only, that it provides a reliable sorting after the textile article is identified and classified.

[0068] In one embodiment of the invention, the system further comprises a sorting air system, such as a fan system and / or air nozzles with pressurized air, adapted to sort the textile article.

[0069] The advantage of this embodiment is in particular, but not only, that it provides a reliable sorting after the textile article is identified and classified.

[0070] In one embodiment of the invention, the system comprises a plurality of reversible, overlapping sorting conveyors adapted to receive the textile article from at least one main conveyor and transport the article for sorting.

[0071] The advantage of this embodiment is in particular, but not only, that it provides a reliable sorting after the textile article is identified and classified.

[0072] In one embodiment of the invention, the position of the textile article is manually sorted.

[0073] The advantage of this embodiment is in particular, but not only, that it provides a reliable sorting after the textile article is identified and classified.

[0074] In one embodiment of the invention, the system further comprises shielding means for shielding the radiation of the radiation emitter.

[0075] The advantages of this embodiment are, among others, but not only, to provide a system that is protected from X-ray and gamma radiation.

[0076] In one embodiment of the invention, - the infrared emitter is a halogen or LED bulb, and / or - the infrared sensor is a near infrared (NIR) or mid-infrared (MIR) sensor.

[0077] The advantages of the second aspect of the invention are, among others, but not only, to provide a method of identifying textile items, wherein the information of the textile for recycling and reuse is not provided by a single sensor only, but by at least three independent sensors, each providing basic information about the textile. Furthermore, and most characteristically, the synergy between these three specific sensors is unprecedented so far and it has proven that this synergy is able to provide unique and exclusive information in the identification, classification and sorting of textiles. The method provides a process suitable for identifying any necessary information of a textile.

[0078] The method provides the possibility to conclude with a certain probability that the textile is a pair of jeans, which none of the sensors alone is able to do.

[0079] Furthermore, the method provides the opportunity to utilize all or at least as many as possible of the textiles to be recycled and / or reused, since all necessary characteristics of the textile are correctly identified, the textile is correctly classified and finally an optimized and reliable sorting of the textile is achieved, so that the sorted textile is ready for recycling and / or reuse by mechanical opening of the fibers without damaging them or by using other recycling methods, such as chemical recycling.

[0080] The recycling process of the fibers, such as opening of the fibers, should not be considered as part of the invention, but as a process that is suitably performed after the textile items are sorted by the system.

[0081] In one embodiment of the invention, the method further comprises the following steps: - removing the "background noise" of the conveyor from the radiation sensor measurement data.

[0082] The advantages of this embodiment are, among others, but not only, that it is achieved that the measurements are only taken on the textile, where the transport does not influence the data. The material of the transport (belt) forms a "constant noise" in the data / pictures collected by the sensors, and this constancy makes it easier to remove the noise, so that even if the measurements are taken through the transport material, an optimized image of the textile is obtained by the radiation sensors.

[0083] In the present invention, either system or method, different transports can be provided for different sensors, so that the removal of background noise is optimized for different sensor types.

[0084] In an embodiment of the present invention, the method further comprises the step of: - storing the measured and collected data in a database.

[0085] The advantages of this embodiment are, among others, but not only, that it is possible to use all the stored information to advance the development of the system.

[0086] In an embodiment of the present invention, the method further comprises the step of: - providing at least one algorithm.

[0087] The advantages of this embodiment are, among others, but not only, that a method is provided that allows a reliable identification of the characteristics of the textile.

[0088] The data is preferably, but not necessarily, processed by various algorithms to obtain useful information from the data measured by the sensors - i.e. to transform the raw data into information that can serve as a basis for the identification.

[0089] In an embodiment of the present invention, the method further comprises the step of: - using the data stored in the database for machine learning by training at least one algorithm, Preferably, the algorithm is trained for at least one of: - optimizing the identification of the characteristics of the textile item, - optimizing the classification of the item, and / or - optimizing the sorting of the item.

[0090] The advantages of this embodiment are, among others, but not only, that the possibility is provided to use machine learning on (at least) three levels, for example: - the identification optimization can be used to see specific characteristics (e.g. identify a button on an X-ray image).

[0091] By using different sensor outputs, several different characteristics of the textile item can be derived with different degrees of certainty. Using a probabilistic machine learning model (Bayesian network), given the measured data, it is possible to estimate the probability that the textile item has certain characteristics.

[0092] - sorting optimization can be used to allow as many textiles as possible to be recycled and / or reused.

[0093] - sorting optimization can be used to allow as many textiles as possible to be recycled and / or reused.

[0094] In one embodiment of the invention, the classification of the item is based on at least one algorithm.

[0095] The advantages of this embodiment are, among others, but not only, that an improved classification method is provided.

[0096] In one embodiment of the invention, the method further comprises the steps of: - selecting a predetermined sorting group, - sorting the item based on one or more predetermined sorting groups.

[0097] In one embodiment of the invention, the sorting of the item is based on at least one algorithm.

[0098] The advantages of this embodiment are, among others, but not only, that a method is provided which makes the sorting of textile items more precise and accurate than the state of the art, taking the potential of textile reuse to a new level.

[0099] In one embodiment of the invention, the method further comprises the steps of: - using a machine learning algorithm for at least one of: - identifying features of the textile item, - classifying the item into a predetermined classification group, and / or - sorting the item into a predetermined sorting group.

[0100] The advantages of this embodiment are, among others, but not only, that an improved identification, classification and / or sorting method is provided and also a method is provided which continuously improves itself after each execution.

[0101] In one embodiment of the invention, the textile item features identified from the measurements of the one or more sensors are one or more of: - color distribution, such as color gradient - fabric type, such as woven, knit, terry and / or denim, - material composition, - density, - mass, - condition, such as cleanliness, water content and / or mold content, - presence of reflective material, - the presence of a pattern, such as stripes and / or checks, - the presence of a polymer print, - the presence of a hard accessory, such as a button, a zipper, a pen and / or a tool, - the presence of a contaminant chemical, such as an oil stain, - the presence of a multi-layer textile, such as a jacket, a pillow and a duvet, - the type of textile item, such as trousers, jeans, a shirt, a bag or shoes, or - an item that is not a textile item.

[0102] In one embodiment of the invention, within the system or within the method: - the visible light spectrum is within 300-1000 nm, and / or - the infrared spectrum is within 1300-4000 nm, and / or - the X-ray or gamma-ray spectrum is within 0.01-10 nm.

[0103] In a third aspect, the invention also relates to a computer program comprising instructions which, when the program is executed by a computer, cause the computer to at least perform the identifying and classifying, preferably also the sorting.

[0104] In a fourth aspect, the invention also relates to a computer-implemented textile item classifying method, comprising: - receiving an input dataset comprising at least one identified feature of a textile item according to the measuring step; and - generating an output dataset comprising a classification of the textile item.

[0105] Each of the first, second, third and fourth aspects of the invention can be combined with any of the other aspects. These and other aspects of the invention will become apparent from and be elucidated with reference to the embodiments described hereinafter.

[0106] Embodiments from the system / product should be considered applicable to the method, and embodiments from the method should be considered applicable to the product / system. BRIEF DESCRIPTION OF DRAWINGS

[0107] Figure 1 A simplified version of a textile identifying system is described, which is suitable for performing a method for identifying a textile item.

[0108] Figure 2 A gap of a conveyor belt is described. DETAILED DESCRIPTION

[0109] Figure 1A simplified version of a textile identification system 1 for identifying textile objects 2 is described, wherein the identification system comprises: - a transport machine 10 for moving the associated textile objects 2 in a first direction D. In Figure 1 only one main transport machine 10 is shown, however, the system can comprise a plurality of main transport machines. Furthermore, the system can comprise a sorting transport machine (not shown) at the end of at least one main transport machine, suitable for sorting the classified textile objects.

[0110] - a visible light emitter 20 for emitting light in the visible wavelength spectrum. Figure 1 Two visible light emitters 20 are shown in. In the present invention, there can also be only one visible light emitter, or more than two visible light emitters. The visible light spectrum is preferably within 300-1000 nm.

[0111] - a colour image sensor 25 for measuring light in the visible wavelength spectrum, preferably an RGB sensor. Figure 1 One colour image sensor 25 is shown in. In the present invention, there can also be a plurality of colour image sensors.

[0112] - an infrared emitter 30 for emitting light in the infrared wavelength spectrum, preferably in the near infrared wavelength spectrum. Figure 1 Two infrared emitters 30 are shown in. In the present invention, there can also be only one infrared emitter, or more than two infrared emitters. The infrared emitter can be a halogen or LED bulb. The infrared spectrum is preferably within 1300-4000 nm.

[0113] - an infrared sensor 35 for measuring light in the infrared wavelength spectrum, preferably in the near infrared spectrum. Figure 1 One infrared sensor is shown in. In the present invention, there can also be a plurality of infrared sensors. The infrared sensor is preferably a near infrared (NIR) sensor and / or a mid infrared (MIR) sensor.

[0114] - a radiation emitter 40 for emitting radiation in the X-ray or gamma-ray spectrum. In Figure 1 One radiation emitter 40 is shown in. In the present invention, there can also be only one radiation emitter, or a plurality of radiation emitters. The X-ray or gamma-ray spectrum is preferably within 0.01-10 nm.

[0115] - a radiation sensor 45 for measuring radiation in the X-ray or gamma-ray spectrum. In Figure 1 One radiation sensor 45 is shown in. In the present invention, there can also be a plurality of radiation sensors.

[0116] - a database system 50 adapted to at least acquire, store and process data measured by at least: - a color image sensor 25, - an infrared sensor 35, and - an X-ray or gamma ray sensor 45.

[0117] In Figure 1 the radiation sensor 45 is arranged below the conveyor 10 and the conveyor or the part of the conveyor above which the radiation sensor is located is made of a suitable material which does not block X-rays or gamma rays.

[0118] It is to be understood that the sensors can be arranged below the conveyor and the emitters above the conveyor. Figure 1 The position in

[0119] The system further comprises shielding means 60 for shielding the radiation of the radiation emitters.

[0120] Figure 1 The system in

[0121] In Figure 1 the dashed line 70 represents the data flow from the sensors to the database 50. The data collected in the database 50 and connected to the algorithm 55 provides a unique solution. While certain information / features of the textile item are specifically provided by a single sensor, other information / feature types can only be inferred by using all three sensors 25, 35, 45 and using the information from the sensors in the algorithm 55, e.g. using a probabilistic model. In the system 1 it is now possible to identify features of the textile item 2 with a certain probability and based on these features classify the textile item as e.g. a pair of jeans, which none of the sensors alone could do.

[0122] Figure 1 The illustrated system is adapted to perform a method of identifying a textile item 2, wherein the method comprises the following steps: - transporting the textile item in a first direction D, preferably on a conveyor 10, - emitting light to the textile item by at least one visible light emitter, the first light emitter emitting light within the visible light spectrum, - measuring the color distributed on the textile item by at least one color image sensor, preferably an RGB sensor, - emitting infrared light to the item by at least one second light emitter, - measuring the reflected light by at least one infrared sensor, - emitting X-ray or gamma-ray radiation by a radiation emitter against the textile, - measuring the radiation by at least one radiation sensor, - collecting data measured by the color image sensor, the infrared sensor and the radiation sensor, - identifying at least one or more features of the textile item, including: - color, - material composition, and - presence of hard materials (such as buttons and zippers), and - classifying the item, preferably into a predetermined classification group, wherein the steps of the method can be performed in any order and / or simultaneously. Furthermore, the method can comprise more steps.

[0123] Furthermore, Figure 1 a computer program is indirectly described, comprising instructions which, when the computer executes the program, cause the computer to perform at least the identifying and classifying steps of the method, preferably also the sorting step.

[0124] Furthermore, Figure 2 a computer-implemented textile item classification method is also indirectly described, comprising: - receiving an input data set comprising at least one identified feature of a textile item according to the measuring step; and - generating an output data set comprising a classification of the textile item.

[0125] Figure 1 the same system is described, with the difference that the system comprises a gap on the transport machine, wherein the radiation sensor is arranged below the gap of the transport machine. ​

[0126] Although the present invention has been described in connection with specific embodiments thereof, it is not to be understood that the application is limited in any way to the examples presented. The scope of the present invention is defined by the appended claims. In the context of the claims, the term "comprising" does not exclude other possible elements or steps. Also, the mention of a reference sign in the claims shall not be construed as limiting the scope of the present invention to the reference sign. Furthermore, the single features mentioned in the different claims can also be combined with each other, and the combination of features mentioned in the different claims is possible even if this combination is not explicitly mentioned in the claims.​

Claims

1. A textile identification system (1) for identifying textile objects (2), wherein the identification system comprises: – A transport aircraft (10) for moving associated textile items along a first direction (D), -Visual or optical recognition systems – An infrared emitter (30) is used to emit light within the infrared wavelength spectrum, preferably within the near-infrared wavelength spectrum. – An infrared sensor (35) is used to measure light within the infrared wavelength spectrum, preferably within the near-infrared wavelength spectrum. – Radiation emitter (40), used to emit radiation within the X-ray or gamma-ray spectrum. – A database system (50) suitable for acquiring, storing, and processing data measured by at least the following: – Color image sensor (25), and – Infrared (35), characterized in that the system further includes – A visible light emitter (20) for emitting light within the visible wavelength spectrum, the visible light emitter (20) having a spectrum in the range of 300–1000 nm, and – A radiation sensor (45) for measuring radiation within an X-ray or gamma spectrum using absorption spectroscopy (e.g., X-ray or gamma-ray absorption spectroscopy), the radiation sensor (45) being located below the transport vehicle (10), and the portion of the transport vehicle (10) above the radiation sensor (45) being made of a suitable material that does not obstruct X-rays or gamma rays, and – A visual or optical recognition system, comprising a color image sensor (25) for measuring light within the visible wavelength spectrum, preferably an RGB sensor and / or a CMOS sensor and / or a CCD sensor, and – Database system (50), the database system (50) being adapted to acquire, store and process data measured by radiation sensor (45). – wherein the radiation sensor (45) is located below the transport (10), and the portion of the transport or the portion of the transport above the radiation sensor is made of a suitable material that does not block X-rays or gamma rays.

2. The identification system (1) according to claim 1, characterized in that, The database system (50) is adapted to identify at least one predetermined feature of an associated textile object (2).

3. The identification system (1) according to claim 2, characterized in that, The identification system also includes an algorithm (55) that is suitable for classifying associated textile objects (2) based on at least one predetermined feature.

4. The identification system (1) according to claim 3, characterized in that, The identification system is used to provide sorting commands for associated textile items (2) based on classification.

5. The identification system (1) according to claim 3, characterized in that, The algorithm is a machine learning algorithm that is suitable for training with data collected from a database (50).

6. The identification system (1) according to any one of the preceding claims, characterized in that, The position of the textile item (2) on the conveyor is measured by at least one sensor (25, 35, 45), and the identification system includes a robot, preferably a robotic arm, adapted to collect the item for sorting.

7. The identification system (1) according to any one of the preceding claims, characterized in that, The system also includes a shielding device (60) for shielding the radiation emitted by the radiator (40).

8. A method for identifying textile objects, preferably applicable to the identification system according to claim 1, the method comprising the following steps: – The textile item (2) is transported along the first direction (D), preferably on a transport machine (10). – Use a visual or optical system to identify the textile object. –Emit infrared light to the object via at least one infrared emitter (30), – The reflected light is measured by at least one infrared sensor (35). – X-rays or gamma rays are emitted onto the textile object via a radiation emitter (40). – Collect data measured by color image sensors and infrared sensors. – Identify at least one or more features of the textile object (2), – Classify textile items, preferably into pre-defined classification groups. The method is characterized by further comprising the following steps: – Light is emitted onto the textile object by at least one first light emitter (20), the first light emitter emitting light within the visible light spectrum, the visible light emitter (20) having a spectrum in the range of 300–1000 nm. – Provide a visual or optical recognition system in the form of a color image sensor (25) for measuring light within the visible wavelength spectrum, preferably an RGB sensor and / or a CMOS sensor and / or a CCD sensor. – The color distribution across the entire textile object is measured using at least one color image sensor (25), preferably an RGB sensor and / or a CMOS sensor and / or a CCD sensor. – A radiation sensor (45) is installed below the transport aircraft (10). – Provide a transport aircraft or transport aircraft section, located above the radiation sensor (45), the transport aircraft or transport aircraft section being made of a suitable material that does not block X-rays or gamma rays, –The radiation within the X-ray or gamma spectrum is measured using absorption spectroscopy, such as X-ray or gamma-ray absorption spectroscopy (45), wherein the radiation sensor (45) is used to, – Collect data measured by the radiation sensor (45), – The measured and collected data are stored in a database (50). –Based on the collected data, identify at least one or more features of the textile object (2), preferably at least: -color, – Material composition, and – The presence of rigid materials, such as buttons and zippers, and in this method – The steps of the method can be performed in any order and / or simultaneously.

9. The identification method according to claim 8, characterized in that, The method further includes the following steps: - Provide at least one algorithm (55).

10. The identification method according to claim 9, characterized in that, The method further includes the following steps: – By training at least one algorithm (55), machine learning is performed using data stored in the database (50). Preferably, the algorithm is trained for at least one of the following: –Optimize the identification of features of textile objects (2), –Optimize the classification of textile items (2), and / or –Optimize the sorting of textile items (2).

11. The identification method according to claim 10, characterized in that, The method also Includes the following steps: – Use the machine learning algorithm to perform at least one of the following: – Identify the characteristics of textile objects – Sort items into predefined category groups, and / or - Sort items into predetermined sorting groups.

12. The method according to any one of claims 8 to 11, characterized in that, The method further includes the following steps: – Select the pre-selected sorting group, - Items are sorted based on one or more predetermined sorting groups, preferably based on at least one algorithm.

13. The identification system according to any one of claims 1-7, characterized in that, The textile object features identified from measurements by one or more sensors are one or more of the following: - Color distribution, such as color gradient - Fabric type, such as woven, knitted, terry cloth, and / or denim. – material composition, -density, -quality, – Conditions, such as cleanliness, moisture content, and / or mold content, -The presence of reflective materials – The presence of patterns, such as stripes and / or checks, –The existence of polymer printing – The presence of rigid accessories, such as buttons, zippers, pens and / or tools, -The presence of contaminant chemicals, such as oil stains – The presence of multiple layers of textiles, such as jackets, pillows, and bedding, – The type of textile item, such as trousers, jeans, shirts, bags, or shoes, or - An object that is not a textile item.

14. A computer program comprising instructions that, when executed by a computer, cause the computer to perform at least the identification and classification steps of the method according to claim 8.

15. A computer-implemented method for classifying textile items, comprising: – The measurement step according to claim 8 receives an input dataset, the input dataset comprising at least one identified feature of the textile object according to claim 13; as well as – Generate an output dataset that contains classifications of textile objects.

Images