Sorting devices

Abstract

The sorting device (10) comprises a transport means (12) for transporting a material flow (14) through the sorting device (10), a multi-energy X-ray system (20) configured to radiograph the material flow (14) using at least two different energies and to detect radiographs based on the radiographs, wherein each radiograph includes, for each region, first information relating to density and / or atomic number and second structural information, and a processor (28) configured to detect one or more regions, including a component to be recycled (16), or a battery, in particular a lithium-ion battery, or a battery cell, in particular a lithium-ion battery cell, within each radiograph of the radiographs by using an AI algorithm, wherein the detection is based on a first feature (M1) derived from the first information and a second feature (M2) derived from the second structural information.

Description

Technical Field

[0001] Embodiments of the present invention relate to sorting devices or plants having a single or multi - energy X - ray system, as well as methods and computer programs for their respective recycling. Preferred embodiments relate to a system that describes a combination of multi - energy X - ray technology and image evaluation based on deep learning to detect lithium - ion batteries in a material stream.

Background Art

[0002] Lithium - ion batteries (LIBs) cause fires and enormous economic damage in sorting plants. LIBs can be damaged in the mechanical processing steps of sorting plants (e.g., bag openers or shredders) and subsequently catch fire. Fires caused by LIBs release a large amount of energy, and the decomposition process of LIBs generates oxygen, which makes it very difficult to extinguish the fire because it accelerates the fire or reignites a fire that has already gone out.

[0003] The detection of LIBs in recycling streams (e.g., yellow bags, recycling bins, paper, etc.) is by no means trivial. On the one hand, materials often pile up 20 - 30 cm high on conveyor belts, and on the other hand, LIBs are incorporated into electronic devices and thus often cannot be freely detected on the material stream. This fact completely rules out optical detection systems (such as cameras). Mechanical separation methods such as air - flow sorting can separate light objects from heavy objects, but cannot reliably detect and remove LIBs. The present invention applied to this field solves the detection of LIBs in complex material streams that could not be implemented until now.

[0004] So far, the problem has been addressed by treating the symptoms rather than the root cause. Sorting plants use thermographic cameras and fire suppression equipment to detect and extinguish temperature increases in the material flow. Such systems cost up to 5 million euros in individual sorting plants. On the one hand, the detection rate of the cameras against high material thickness on the conveyor belt is problematic, and on the other hand, the fact that the resulting fire suppression water must be collected and processed separately is also problematic. No other method or product is known that operates similarly to the proposal herein. Often, there is a manual sorting step at the end of the sorting plant, where workers grab objects from the material flow while they are passing along the conveyor belt. At this point, in most cases, it is too late to react to ignited or recently ignited LIBs. Therefore, an improved method is needed. [Overview of the project] [Problems that the invention aims to solve]

[0005] The object of the present invention is to provide a concept for improving the recycling of specific components such as batteries or lithium-ion batteries. [Means for solving the problem]

[0006] This objective is resolved by the subject matter of the independent claim.

[0007] Embodiments of the present invention provide a sorting device having a transport means, a single or multi-energy X-ray system, and a processor. The transport means is configured to transport a material flow through the sorting device and can be mounted, for example, on a conveyor belt. The material flow to be recycled, for example, waste or waste from yellow bags, is then transported on the conveyor belt. This material flow may also include components such as batteries or rechargeable batteries or lithium-ion batteries that are specifically to be recycled. The single or multi-energy X-ray system is configured to radiograph the material flow using at least one or two different energies and to detect radiographs based on the radiographs, each radiograph including first information relating to density and / or atomic number for each region, and second structural information. According to embodiments, the structural information may include information relating to the location of one or more regions of the components / electronic devices / batteries or battery cells to be recycled, or information relating to the location of electronic devices or wiring. Additionally or alternatively, the second structural information may include information relating to the geometry of one or more regions of the components / electronic devices or batteries or battery cells to be recycled. The processor is configured to detect one or more regions containing a battery, particularly a lithium-ion battery, or a battery cell, particularly a lithium-ion battery cell, within each radiograph of a radiograph by using an AI algorithm or an AI algorithm that is pre-trained or trained in operation. The detection is based on a first feature derived from first information and / or a second feature derived from second structural information.

[0008] Embodiments of the present invention are based on the finding that a combination of multi-energy radiographs is supplied to an artificial neural network (ANN) to enable early recognition of components to be recycled, allowing for the evaluation of radiographed materials with respect to density and atomic number, as well as structural information from the radiographs (e.g., shape, decay, information about installed electronic equipment, etc.). For example, the neural network can be trained to detect devices such as lithium-ion batteries (LIBs) or individual LIBs (cells) in these evaluated images. This has the advantage that the fire hazard caused by LIBs can be eliminated as early as the start of the recycling process. With high detection quality, it can be assumed that the fire hazard caused by LIBs can be significantly reduced or completely eliminated. In contrast to other detection methods (e.g., optical systems) that require a single location in the material flow, this is not required in the described method. In most sorting plants, it is difficult to generate a single location in the material flow, especially at the start of the material flow. Thus, embodiments of the present invention improve fire safety, and the cost of X-ray detection systems for fire detection and extinguishing is significantly lower than in current plants.

[0009] According to one embodiment, the processor is configured to identify one or more candidate regions of a component to be recycled or a battery or battery cell based on a first feature, and to identify the candidate regions as one or more regions based on a second feature. According to a further embodiment, one or more candidate regions of a component to be recycled, a battery, or a battery cell may be identified based on a second feature, and the candidate regions may be identified as one or more regions based on a first feature. According to a further modification, the processor may also be configured to identify one or more regions based on a combination of the first and second features.

[0010] According to one embodiment, the processor is configured to determine information regarding the location of one or more regions and / or the location or relative position of one or more regions in a material flow and send this information to, for example, a sorting plant. According to an embodiment, the sorting device may include a control device configured to control a sorting means. According to a preferred embodiment, the control of the sorting means is performed such that the sorting means is activated when the processor identifies one or more regions. According to a further preferred modification, the control device is configured to control the sorting means so that the sorting means sorts components / batteries / battery cells to be recycled based on predetermined locations or determined relative positions. Here, the sorting means can be positioned based on information regarding the location of one or more regions in a material flow and / or the location or relative position of one or more regions.

[0011] According to one embodiment, the sorting means comprises a pneumatic system, a pneumatic high-speed switching valve, a drive flap, a reversing belt, or a robotic gripping arm.

[0012] According to the embodiment, the X-ray system is configured to determine second structural information by analyzing homogeneous or semi-homogeneous regions associated with the recycled component / battery or battery cell in terms of its geometry. Here, the geometry is detected. A cylindrical geometry is typical for battery cells. Additional components, such as wiring or electronics, may also be detected during detection. Wiring often has a high aspect ratio and at least partially irregular bends. Electronics are characterized by carriers such as printed circuit boards, or electronic components such as capacitors, ICs, and / or resistors. These additional components, such as wiring or electronics, can, in combination with circular, cylindrical, or square cells, represent the second structural information. In the simplest case, the second structural information may include information about the geometry associated with a sealed volume or the recycled component. Here, according to the embodiment, pattern detection can be performed by accessing several exemplary / similar patterns in a database and detecting similarities to typical patterns. According to the preferred embodiment, this database is trained with AI. User input labels the relevant objects, and the respective training data is fed to the AI ​​algorithm for training.

[0013] According to the embodiment, training can also be performed in combination with first information, i.e., it includes labeled training data containing first and second structural information. The training data is then examined by an AI algorithm with respect to the first and second features. Preferably, a large amount of training data is used to train the AI ​​algorithm either in advance or during operation. According to the embodiment, it is also possible to collect and subsequently label a large amount of data during operation.

[0014] According to the embodiment, the X-ray system is positioned in front of the sorting means in the material flow. It should be noted that, with respect to position, the position in the material flow and the relative position are calculated and passed along the movement of the material flow, i.e., according to the direction and velocity of the material flow. According to the embodiment, the material flow may have several layers, and the components / batteries / battery cells to be recycled do not necessarily have to be located in the upper layer, but can also be located between two layers.

[0015] Further embodiments include a method for recycling, The sorting means transports the material flow through the sorting device, The method involves radiographically imaging a material flow at at least one or two different energies, and detecting a radiograph based on the radiography, wherein each radiograph includes, for each region, first information relating to density and / or atomic number, and second structural information. The detection of recyclable components, or one or more regions containing batteries, particularly lithium-ion batteries, or battery cells such as lithium-ion battery cells, within each radiograph by using an AI algorithm, The method provides for detection based on a first feature (M1) derived from first information and / or a second feature (M2) derived from second structural information.

[0016] According to the embodiments, this method can be implemented in a computer. Embodiments of the present invention will be described with reference to the accompanying drawings. [Brief explanation of the drawing]

[0017] [Figure 1] This is a schematic diagram of a sorting plant according to a basic embodiment. [Figure 2] This is a schematic diagram of a flow chart illustrating a recycling method according to an embodiment. [Figure 3] These are schematic diagrams of recycled components (batteries) to illustrate the first and second features or first and second (structural) information applied in the embodiments. [Modes for carrying out the invention]

[0018] Before describing embodiments of the present invention below with reference to the accompanying drawings, it should be noted that the same elements and structures are given the same reference numerals, and their descriptions are mutually applicable or interchangeable.

[0019] Figure 1 shows a sorting plant 10 (generally a sorting device) having a transport means 12 and an X-ray system, in this case a multi-energy X-ray system 20. The sorting plant includes, for example, a radiation source 22 and a radiation detector 24. Furthermore, the sorting plant 10 includes a processor 28. The sorting plant is, for example, informationally coupled to the (multi-energy) X-ray system and receives radiographs from the X-ray detector 24. According to an optional embodiment, the sorting plant 10 may also include sorting means 30.

[0020] The conveying means 12 is configured here as a conveyor belt and conveys the material flow 14 along the direction of movement 12b. The object to be detected, such as a battery or lithium-ion battery 16, may be contained within the material flow 14. The sorting plant 10 is configured to identify the object to be detected 16 and to sort this object by sorting means 30 according to an optionally selected embodiment. Next, the identification of the object 16 according to a basic embodiment will be described.

[0021] The multi - energy X - ray system 20 radiographs the material stream 14, and thus also the object 16 to be recycled, by means of the radiation source 22. For this purpose, two or more radiographic energies E1 and E2 are used and are then detected by the X - ray detector 24 after the radiography of the material stream 14 or the object 16 to be recycled. The X - ray detector 24 outputs radiographs associated with the energies E1 and E2, for example, to a processor. According to a preferred variant, the radiographs can generally exist as multi - energy radiographs or generally as radiographs. The radiographs have the advantage that they can determine information such as the density of the object being radiographed, thus the material stream 14 or the object 16 to be recycled, and also the atomic number of the material stream 14 or the object 16 to be recycled. The density and / or the atomic number are regarded as the first information. Here, it should be noted that batteries such as lithium - ion batteries (referring to the object 16 to be recycled) often have a specific density and / or a specific atomic number due to their material properties. This first information I1 is determined by the processor 28 or obtained from each radiograph. When the entire material stream 14 including the object 16 to be recycled is radiographed, this information I1 can be obtained from each radiograph associated with different regions, for example, associated with different pixels or different clustered pixels. Apart from determining I1, the processor 28 is also configured to determine I2. I2 represents structural information, for example, the geometry or location or position of the object 16 within the material stream 14. In this connection, the processor 28 is configured to detect / mark, for example, a continuous region consisting of several pixels within each radiograph and analyze this region with respect to position, location, size, and geometry. For example, the geometry of the object 16 to be recycled or the battery can be detected. Batteries often have a cylindrical shape. The processor 28 can detect and mark such a typical geometry.

[0022] The first feature M1 is derived from information I1 regarding density or atomic number, and the second feature M2 is derived from information I2 regarding geometric shape or generally structural information. These two features in combination enable a conclusion regarding the presence or absence of the object 16, or the object 16 being searched for such as a battery, a lithium-ion battery, or a battery cell, or a lithium-ion battery cell. The features M1, M2 can each have different signs, and detection is made possible by combining the signs according to an embodiment. The detection is performed according to an AI algorithm or a trained algorithm. The algorithm is implemented on the processor 28 and is trained by training data either beforehand or during operation. In this way, according to an embodiment, the processor 28 can access a database, for example, a database stored in an internal memory or an external memory (server). The external database offers the advantage of being able to increase a large-scale database for training the AI algorithm by several linked AI algorithms or similar screening plans.

[0023] Each method for controlling the screening plan is described below with reference to FIGURE 2, and optional steps are also described.

[0024] Method 100 includes three basic steps 110, 120, and 130. Subsequently, an optional step 140 can be provided. In step 140, for example, a material flow 14 containing the lithium-ion battery 16 is conveyed through the screening plant 10 by the conveying means 12.

[0025] In the subsequent step 120, the material flow is radiographed at two different energies to obtain a radiograph. According to the embodiment, these two steps 110 and 120 are repeated continuously, i.e., always for a new portion of the material flow, and always a new multi-energy radiograph is taken from a further or shifted portion. In the next step 130, either one multi-energy radiograph or multiple multi-energy radiographs associated with several samples can be analyzed. This step is denoted by reference no. 130 and includes detecting one or more regions containing recyclable components such as batteries, lithium-ion batteries, or battery cells, lithium-ion battery cells, etc., by using an AI algorithm. Detection 130 is performed based on a first or second feature M1 / M2, as already described above. Here, the first and second features can be used in combination. The combination means that both features are equivalent, i.e., evaluated together. Alternatively, evaluation by the first feature and confirmation by the second feature, or evaluation by the second feature and confirmation by the second feature, is possible. According to further embodiments, obviously, additional features can be added.

[0026] According to the embodiment, each feature is characterized by one or more parameters. In the case of the first feature, this would be the atomic number or density. A lithium-ion battery has a specific atomic number or a range to which that atomic number belongs. This could be, for example, 3 or 1 to 30. Exemplarily, the density may also be 0.1 to 5 g / cm³. 3 Or 0.5 g / cm³ 3 ~12g / cm 3 It can be within the range of

[0027] There are also parameters for the second feature M2, which can be used to describe the structure. For example, it may include geometric parameters that characterize the shape or geometric parameters that characterize the volume. This second structural feature M2 may include information about whether it is connected to further components, such as electronic devices. Combinations of sub-features are possible for both the first and second features M1 / M2 (e.g., atomic number + density for the first feature, and volume + shape factor and / or + further components to be detected for the second feature). Detection is performed based on the combination of features or the combination of features of sub-features.

[0028] Detected recyclable objects, such as lithium-ion batteries, are marked; that is, information is output indicating a high probability of the presence of each recyclable object, such as lithium-ion batteries. Additionally, the position of object 16 in the material flow 14 can also be shown, and the position based on the movement of the material flow 12b can include information such as the speed and direction of movement.

[0029] This information is used in the following optional step 140, in which the detected object 16 is sorted accordingly, i.e., separated from the rest of the material flow 14 by, for example, a pneumatic device or a gripping arm. In Figure 1, these sorting means are given reference numeral 30.

[0030] In summary, this means that the system 10 in Figure 1 uses multi-energy X-ray technology 20 to radiograph, for example, a waste flow 14 on a conveyor belt 12 and detect LIBs 16 within the material flow 14, even within different devices with large material thicknesses. A machine learning (ML) based method is used to identify the LIBs 16 (exposed or within devices). This method uses at least a first feature M1 and a second structural feature M2 or their respective sub-features. For this purpose, the algorithm is trained with multiple training data. This step is optional and is referenced with reference number 135 in Figure 2. During training, multiple multi-energy radiographs associated with different material flows or different material flows with objects to be detected, such as LIBs, are provided and labeled pre- or post-.

[0031] In this way, it is possible to classify individual different substances within the material flow 14 based on features M1 and M2, and consequently to detect the LIB 16 or, in general, the object 16 to be detected. The first feature M1 is determined based on first information, and feature M2 is determined based on second structural information. As already mentioned above, several pieces of information can also be used for each feature M1 and M2. It is also possible to divide each feature into sub-features.

[0032] In some embodiments, a further embodiment allows an ML processor or processor trained by ML28 to classify the discovered devices 16 into classes (e.g., power banks, mobile phones). This also means that it can distinguish between individually detected objects 16. In this way, the algorithm can also be configured to detect and distinguish different objects. This can also be done by linking information from multi-energy radiographs (X-ray data based on density and atomic number) with object detection (shape, attenuation, information, e.g., information about installed electronic equipment).

[0033] Herein, it should be noted that, according to the embodiment, the multi-energy radiography technology is installed as far forward as possible in the material flow direction in the sorting plant in order to detect hazardous elements such as LIBs as early as possible (see 12b).

[0034] Regarding Figure 3, the feature combinations will be explained based on the figure. Figure 3 shows a two-dimensional diagram with features M1 and M2. A higher value indicates a higher level of fit to each feature. For example, a high M1 value indicates that the atomic number and / or density are close to the typical density / typical atomic number of the object being searched for, e.g., LIB. Feature M2 determines a combination of different shape factors. For example, a high M2 value indicates that the size is within the respective selection range, i.e., the detected object has a volume that matches the volume of the searched object, e.g., LIB, i.e., it is neither significantly larger nor smaller. The figure is divided into three parts, part A of the figure showing that LIB is probably not present in the searched region, and part B of the figure showing the average probability. In region C, the probability of finding LIB in the searched region is high. According to further embodiments, individual features such as the second structural feature M2 can also be divided, for example, to produce a multidimensional, e.g., three-dimensional feature space.

[0035] As already mentioned above, a comparison with a typical "target value" is always performed for the characteristics, that is, the probability of the existence of each object being searched for, such as LIBs, is presented when the atomic number or density is very high.

[0036] It should be noted that, according to this embodiment, multi-energy radiography can be realized not only in one-dimensional radiography, i.e., in a single radiography direction, but also in multi-dimensional radiography directions using CT.

[0037] The following describes a preferred embodiment in its entirety. A multi-energy X-ray system is installed (as early as possible) in a suitable location within the sorting plant. The system radiographs the material flow on a conveyor belt, generating radiographs. These radiographs enable evaluation of the radiographed material in terms of density and atomic number, as well as the supply of structural information (shape, decay, information, e.g., information about installed electronic equipment, etc.) from the radiographs to an artificial neural network (ANN). This network is trained to detect devices or individual LIBs within these evaluated images. Thus, LIBs can be detected, discovered, and separated from the material flow in the early stages of the recycling process. Different methods (e.g., pneumatic high-speed switching valves, driven flaps, reversing belts, or (robot) grippers) can be used for sorting.

[0038] Embodiments of the present invention are primarily used in the recycling industry. Here, different material flows within the sector can be addressed. These include, for example, lightweight packaging, electrical and electronic waste (WEEE), industrial waste, or general waste. Additionally, it would be possible to extend the patent to other application areas, such as the detection of LIBs in paper waste. This is relevant not only to sorting plants but also to, for example, paper processing plants, temporary storage facilities, or packaging and compression plants.

[0039] While several embodiments have been described in the context of apparatus, it is clear that these embodiments also represent descriptions of corresponding methods, so that blocks or devices of the apparatus also correspond to their respective method steps or features of a method step. Similarly, embodiments described in the context of a method step also represent descriptions of corresponding blocks, details, or features of the corresponding apparatus. Some or all of a method step may be performed by (or using) hardware devices such as a microprocessor, a programmable computer, or an electronic circuit. In some embodiments, some or more of the most important method steps may be performed by such devices.

[0040] Depending on specific implementation requirements, embodiments of the present invention can be implemented in hardware or software. The implementations can be carried out using digital storage media, such as floppy disks, DVDs, Blu-ray discs, CDs, ROMs, PROMs, EPROMs, EEPROMs, or flash memory, hard drives, or other magnetic or optical memory storing electronically readable control signals, which may or may not cooperate with a programmable computer system to perform their respective methods. Therefore, the digital storage media may be computer-readable.

[0041] Some embodiments of the present invention include a data carrier containing electronically readable control signals that can cooperate with a programmable computer system so that one of the methods described herein can be performed.

[0042] Generally, embodiments of the present invention can be implemented as a computer program product having program code, the program code being operable to perform one of the methods when the computer program product is executed on a computer.

[0043] The program code may be stored, for example, in a machine-readable carrier.

[0044] Other embodiments include a computer program for performing one of the methods described herein, the computer program being stored on a machine-readable carrier. In other words, one embodiment of the method of the present invention is a computer program that, when the computer program is executed on a computer, includes program code for performing one of the methods described herein.

[0045] Accordingly, a further embodiment of the method of the present invention is a data carrier (or digital storage medium or computer-readable medium) on which a computer program for performing one of the methods described herein is recorded. The data carrier, digital storage medium, or computer-readable medium is typically tangible or non-volatile.

[0046] Therefore, a further embodiment of the method of the present invention is a data stream or sequence of signals representing a computer program for performing one of the methods described herein. The data stream or sequence of signals may be configured to be transmitted, for example, over a data communication connection, such as the Internet.

[0047] Further embodiments include processing means configured or adapted to perform one of the methods described herein, such as a computer or a programmable logic device.

[0048] Further embodiments include a computer on which a computer program for performing one of the methods described herein is installed.

[0049] Further embodiments of the present invention include an apparatus or system configured to transmit a computer program for performing at least one of the methods described herein to a receiver. The transmission may be, for example, electronic or optical. The receiver may be, for example, a computer, a mobile device, a memory device, or a similar device. The apparatus or system may include, for example, a file server for transmitting the computer program to the receiver.

[0050] In some embodiments, programmable logic devices (e.g., field-programmable gate arrays, FPGAs) may be used to perform some or all of the functionality of the methods described herein. In some embodiments, a field-programmable gate array may cooperate with a microprocessor to perform one of the methods described herein. Generally, the methods are preferably performed by any hardware device, which may be universally applicable hardware such as a computer processor (CPU) or method-specific hardware such as an ASIC.

[0051] The embodiments described above are merely illustrative of the principles of the present invention. Modifications and variations of the configurations and details described herein will be apparent to those skilled in the art. Accordingly, the present invention is intended to be limited only by the appended claims and not by the specific details presented herein.

Claims

1. A sorting device (10), A transport means (12) for transporting a material flow (14) through the sorting device (10), A single or multi-energy X-ray system (20) is configured to radiograph the material flow (14) using at least one energy or at least two different energies, and to detect radiographs based on the radiographs, wherein each radiograph includes, for each region, first information relating to density and / or atomic number and second structural information. A processor (28) configured to detect one or more regions, including a recyclable component (16), or electronic equipment, or a battery, particularly a lithium-ion battery, or a battery cell, particularly a lithium-ion battery cell, within each of the radiographs, by using an AI algorithm, A sorting device in which detection is performed based on a first feature (M1) derived from the first information and / or a second feature (M2) derived from the second structural information.

2. The sorting device (10) according to claim 1, wherein detection is performed based on the first feature (M1) in combination with the second feature (M2).

3. The second structural information includes information regarding the location of the recycled components (16) or one or more areas of the battery or battery cell, information regarding the location of electronic equipment or wiring, and / or The sorting device (10) according to any one of the preceding claims, wherein the second structural information includes information relating to the geometric shape of one or more regions of the recycled component (16) or the battery or the battery cell.

4. The sorting device (10) according to any one of the preceding claims, wherein the processor (28) is configured to identify one or more candidate regions of the recyclable component (16) or the battery or the battery cell based on the first feature (M1), and to identify the candidate regions as one or more regions based on the second feature (M2).

5. A sorting device (10) is configured such that the processor (28) identifies one or more candidate regions of the recyclable component (16) or the battery or the battery cell based on the second feature (M2), and identifies the candidate regions as one or more regions based on the first feature (M1).

6. The sorting device (10) according to any one of the preceding claims, wherein the processor (28) is configured to identify one or more regions based on a combination of the first feature and the second feature (M2).

7. The sorting device (10) according to any one of the preceding claims, wherein the processor (28) is configured to determine the position of one or more regions in the material flow (14) and / or information relating to the position or relative position of one or more regions.

8. A sorting device (10) according to any one of the prior claims, further comprising a control device configured to control a sorting means.

9. The sorting device (10) according to claim 8, wherein the control device is configured to activate the sorting means when the processor (28) identifies one or more regions.

10. The sorting device (10) according to claim 8, further comprising a control device configured to sort the recyclable components (16) or the battery or battery cells by the sorting means based on the determined position or determined relative position, and / or to position the sorting means based on the position of one or more regions in the material flow (14) and / or information regarding the position or relative position of one or more regions.

11. The sorting means comprises a pneumatic system, a pneumatic high-speed switching valve, a drive flap, a reversing belt, or a robotic gripping arm, according to any one of claims 8, 9, and 10, the sorting device (10).

12. The sorting device (10) according to any one of claims 8, 9, 10, or 11, wherein the single or multi-energy X-ray system (20) is positioned in front of the sorting means in the material flow direction.

13. The sorting device (10) according to any one of claims 7 to 12, wherein the processor (28) is configured to calculate the position or relative position of the material flow (14) along the movement of the material flow (14).

14. The sorting device (10) according to any one of the preceding claims, wherein the material flow (14) has several layers, and / or the components to be recycled (16) or the battery or battery cell are arranged between two layers.

15. A method (100) for recycling, The transport means (12) transports the material flow (14) through the sorting device (10) (110), The material flow (14) is radiographed (120) at at least two different energies, and a radiograph is detected based on the radiograph, wherein each radiograph includes, for each region, first information relating to density and / or atomic number and second structural information. The method includes detecting (130) one or more regions within each of the radiographs of the radiographs that include a component to be recycled (16), or an electronic device, or in particular a lithium-ion battery, or a battery cell such as a lithium-ion battery cell, by using an AI algorithm, A method in which detection is performed based on a first feature (M1) derived from the first information and / or a second feature (M2) derived from the second structural information.

16. A computer program for performing the method step described in claim 15.

Images