Intelligent optoelectronic sorting machine and product separation method

Abstract

Disclosed are an intelligent photoelectric sorting machine and a product separation method of the intelligent photoelectric sorting machine. The intelligent photoelectric sorting machine comprises a belt drive mechanism (4), a recognition device and a material collecting mechanism (9). The belt drive mechanism (4) is arranged to convey a plurality of products to be sorted. The recognition device is arranged to recognize the plurality of products to be sorted and define the plurality of products to be sorted one by one with different label numbers. The material collecting mechanism (9) comprises a plurality of material collecting channels. Each material collecting channel is electrically connected with the recognition device and is arranged to recognize the label number corresponding to the material collecting channel and control the product to be sorted with the corresponding label number to enter the material collecting channel.

Description

Technical Field

[0001] The present application relates to the field of dry sorting technology, for example, to an intelligent photoelectric sorting machine and a product separation method. Background Art

[0002] Intelligent dry sorting technology has been widely applied in numerous fields, including coal, minerals, food, waste, and building materials. The sorting process involves two steps: identification and sorting. Materials are laid flat on a moving spreader, and a recognition mechanism identifies each piece of material using image, color, X-ray transmission, and X-ray fluorescence. After identification, the material is separated by sorting or by redirecting its trajectory as it falls from the spreader using elastic impacts or instantaneous high-pressure air jets.

[0003] In related technologies, intelligent dry sorting can often only sort two products. For multiple products, the identification mechanism cannot effectively define them after identification, which affects the subsequent separation of multiple products.

[0004] Furthermore, the X-ray sorter, which uses a transmission principle of dual-energy X-rays, performs inspections. However, due to the use of continuous-spectrum X-rays, the distinction between high-energy and low-energy is unclear, with significant overlap. Furthermore, the thickness of the materials being inspected is uneven. While algorithms based on high-energy and low-energy information can partially mitigate the effects of thickness variations, this still results in significant detection errors, particularly for materials with similar atomic numbers. Summary of the Invention

[0005] The present application provides an intelligent photoelectric sorting machine and a product separation method.

[0006] An intelligent photoelectric sorting machine includes a belt transmission mechanism, a recognition device and a collection mechanism; the belt transmission mechanism is configured to convey a variety of products to be sorted; the recognition device is configured to distinguish the various products to be sorted and define the various products to be sorted one by one with different labels; the collection mechanism includes a plurality of collection channels, the number of the plurality of collection channels is the same as the number of categories of the various products to be sorted, each collection channel is electrically connected to the recognition device, and is configured to identify the label corresponding to the collection channel, and control the products to be sorted corresponding to the label to enter the collection channel.

[0007] A product separation method for an intelligent photoelectric sorting machine comprises the following steps:

[0008] Defining first to Nth labels for the N products to be sorted in order of proportion from most to least according to the respective proportions of the N products to be sorted in the N products to be sorted, wherein N is greater than 1;

[0009] Starting from the end closest to the belt transmission mechanism, N collecting channels are set to correspond to the first label to the Nth label in sequence;

[0010] Control the products to be sorted to pass through the flattening device and then enter the identification device;

[0011] Controlling the X-ray linear array detector to obtain an equivalent energy attenuation rate of an X-ray continuous energy spectrum of the product to be sorted, and obtaining a first probability that the product to be sorted corresponds to different labels based on the equivalent energy attenuation rate of the X-ray continuous energy spectrum;

[0012] controlling an X-ray fluorescence receiver to obtain fluorescence spectrum information of the product to be sorted, and obtaining a second probability that the product to be sorted corresponds to the different label according to the fluorescence spectrum information;

[0013] controlling the image recognition system to obtain image information of the product to be sorted, and obtaining a third probability that the product to be sorted corresponds to the different labels based on the image information;

[0014] Performing weighted summation on the obtained probabilities of the same label of the products to be sorted to obtain the probability of the products to be sorted corresponding to each label;

[0015] Determine the label corresponding to the product to be sorted according to a preset probability threshold;

[0016] Transmitting the determined labels to the N collecting channels, and comparing them with the labels corresponding to the N collecting channels to determine the collecting channels corresponding to the products to be sorted;

[0017] controlling the material collecting channel to transmit a signal to the solenoid valve according to a label corresponding to the material collecting channel;

[0018] The solenoid valve controls the air nozzle to blow the products to be sorted into the collecting channel. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a structural diagram of an intelligent photoelectric sorting machine provided by this application;

[0020] Figure 2 This is a flow chart of a sorting method of an intelligent photoelectric sorting machine provided in this application.

[0021] In the picture:

[0022] 1. Cleaning mechanism; 2. X-ray system; 21. X-ray emitting mechanism; 22. X-ray linear array detector; 23. X-ray fluorescence receiver; 3. Image recognition system; 4. Belt drive mechanism; 5. Feeding mechanism; 6. Laying device; 7. Air storage tank; 8. Air nozzle; 9. Collecting mechanism. DETAILED DESCRIPTION

[0023] The technical solution of the present application is described below in conjunction with the accompanying drawings and implementation methods. The specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. For ease of description, only the parts related to the present application are shown in the accompanying drawings, not all of them.

[0024] In the description of this application, unless otherwise specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they can refer to fixed or detachable connections; mechanical or electrical connections; direct or indirect connections through an intermediate medium; and internal communication between two components. The meanings of these terms in this application can be understood based on the circumstances.

[0025] like Figure 1 As shown, the present application provides an intelligent photoelectric sorting machine, including a belt transmission mechanism 4, a recognition device and a collection mechanism 9; the belt transmission mechanism 4 is configured to convey a variety of products to be sorted; the recognition device is configured to recognize a variety of products to be sorted and define the multiple products to be sorted one by one with different labels, the collection mechanism 9 includes a plurality of collection channels, the number of the plurality of collection channels is the same as the number of categories of the multiple products to be sorted, each collection channel is electrically connected to the recognition device, and is configured to identify the label corresponding to the collection channel, and control the products to be sorted with the corresponding label to enter the collection channel.

[0026] In one embodiment, the number of the plurality of collecting channels is greater than or equal to the number of preset product categories.

[0027] In one embodiment, the sorting machine further includes a spreading device 6 , which is disposed below the belt drive mechanism 4 . The spreading device 6 causes the conveyor belt of the belt drive mechanism 4 to vibrate, thereby spreading the products to be sorted flatly on the conveyor belt. This prevents accumulation of the products to be sorted, which could lead to inaccurate subsequent identification and sorting. In this embodiment, the spreading device 6 is a vibrator that vibrates the conveyor belt. In other embodiments, other types of vibrating mechanisms may be used, as long as they can vibrate the conveyor belt to spread the products to be sorted flatly on the conveyor belt.

[0028] In one embodiment, the cleaning mechanism 1 is arranged above the end of the belt transmission mechanism 4 where the product falls. The cleaning mechanism 1 includes a dust hood and a vacuum cleaner connected to the dust hood. The vacuum cleaner can absorb the dust generated during the product sorting process through the dust hood to prevent the generated dust from polluting the air.

[0029] In one embodiment, the sorting machine further includes a feeding mechanism 5 , which is disposed above the belt drive mechanism 4 . Products being conveyed for sorting slide onto the belt drive mechanism 4 via the feeding mechanism 5 . The feeding mechanism 5 allows the products to slide smoothly onto the conveyor belt, preventing them from sticking or overlapping on the conveyor belt, which could prevent subsequent effective identification and sorting. The feeding mechanism 5 in this embodiment is a vibrating feeder; in other embodiments, it could alternatively be a chute or a slide, etc., without further limitation.

[0030] In one embodiment, the identification device includes an X-ray system 2 and / or an image recognition system 3. In this embodiment, the identification device includes an X-ray system 2 and an image recognition system 3. The X-ray system 2 can perform radiographic identification of the products to be sorted based on their photoelectric properties, while the image recognition system 3 can perform image identification of the products to be sorted based on their image properties. In one embodiment, the X-ray system 2 uses the atomic number of the product equivalent in the pixels of the captured product image. When X-rays penetrate objects with low tissue density, less X-rays are absorbed, leaving more X-rays, resulting in a brighter image. However, when X-rays penetrate objects with high tissue density, more X-rays are absorbed, leaving less X-rays, resulting in a darker image. Ultimately, a black-and-white contrast and a light-dark contrast image are formed on the screen. Different atomic numbers per unit volume represent different products.

[0031] The two systems work together to more accurately identify the category of the product. The X-ray system 2 includes an X-ray emitting mechanism 21, an X-ray fluorescence receiver 23, and an X-ray linear array detector 22. The X-ray emitting mechanism 21 emits X-rays with a continuous energy spectrum, which are received by the X-ray linear array detector 22 after being transmitted by the product to be sorted, and are received by the X-ray fluorescence receiver 23 after being reflected by the product to be sorted. The image recognition system 3 obtains image information of the product to be sorted, and obtains the category information of the product to be sorted according to the voting weight mechanism based on the above-mentioned X-ray photoelectron energy spectrum and X-ray fluorescence spectrum to determine the category of the product, and defines multiple products one by one with different labels, and transmits the labels to the aggregate channel. In one embodiment, an X-ray light source emits a continuous energy spectrum. After being transmitted by the product to be sorted, the X-rays received by the X-ray linear array detector 22 are used to calculate the equivalent energy attenuation rate of the X-ray continuous energy spectrum of the product to be sorted due to different product densities, and the probability of which label the product is labeled is determined. The X-ray light source emits a continuous energy spectrum. After being reflected by the product to be sorted, it stimulates a fluorescence spectrum, which is received by the X-ray fluorescence receiver 23, and the probability of which label the product is labeled is determined based on the received fluorescence spectrum.

[0032] In other embodiments, in places where the requirements for identification accuracy are not high, the identification device may also include only the X-ray system 2, which performs X-ray identification on the products to be sorted according to the photoelectric properties of the products to be sorted, defines multiple products one by one with different labels, and transmits the labels to the aggregate channel.

[0033] In one embodiment, the belt drive mechanism 4 includes a driving member and spaced-apart, parallel active and passive rollers. A conveyor belt is wound around these rollers, which are connected by the conveyor belt. The output shaft of the driving member is in driving connection with the active roller, which can be driven by a belt, coupling, or gear assembly. The driving member drives the active roller to rotate, thereby driving the conveyor belt to transport the products to be sorted. In this embodiment, the driving member is a motor, such as a speed-adjustable servo motor, stepper motor, or reduction motor. The driving member controls the speed of the conveyor belt, thereby controlling the conveying speed of the products and preventing interference between adjacent products that fall at excessive speeds.

[0034] In one embodiment, the classification component includes an air nozzle 8, an air tank 7 and an air compressor. The air nozzle 8 is arranged below and / or obliquely above the belt transmission mechanism 4. The setting position and setting quantity are designed according to the actual needs on site. The air tank 7 is connected to the air nozzle 8 and can provide high-pressure gas to the air nozzle 8. The air nozzle 8 is located at one end where the product falls. The air nozzle 8 is provided with a solenoid valve, which is electrically connected to the aggregate channel. When products with the same label are transmitted, the aggregate channel controls the solenoid valve to control the air nozzle 8 to blow air, so that the product enters the corresponding aggregate channel. The air compressor is connected to the air tank 7. The air compressor can replenish high-pressure gas to the air tank 7. A pressure sensor is provided in the air tank 7. Whether to replenish high-pressure gas to the air tank 7 is determined according to the detection value of the pressure sensor. In this embodiment, the air nozzles 8 are provided in two rows, one row of large nozzles and the other row of small nozzles. The gas ejected by the large nozzles has a high intensity, and the product is sprayed over a long distance. The gas ejected by the small nozzles has a low intensity, and the product is sprayed over a short distance. The products that are not sprayed move forward the smallest distance under the action of the inertial force of the belt transmission mechanism 4.

[0035] In one embodiment, since the storage point is fixed, specific types of products must be stored. However, the proportion of multiple types of products in the products to be sorted is different in different batches of products. In order to save energy and reduce sorting errors, products with a larger proportion generally fall into the collection channel close to the belt drive mechanism 4, products with the smallest proportion are blown into the collection channel farthest from the belt drive mechanism 4, and products with an intermediate proportion are blown into the middle collection channel. The products in each collection channel are transported to the corresponding storage point by the conveying mechanism. In this embodiment, the product with the first label enters the collection channel close to the belt drive mechanism 4; the product with the second label enters the middle collection channel under the drive of the small nozzle, and the product with the third label enters the collection channel farthest from the belt drive mechanism 4 under the drive of the large nozzle. In order to ensure that the first label has the largest proportion of products, it is necessary to define the product with the largest proportion as the first label before sorting.

[0036] The intelligent photoelectric sorting machine provided by the present application has a belt transmission mechanism 4 on which transports the products to be sorted to the bottom of the identification device. The identification device identifies the products to be sorted and defines multiple products one by one with different labels. The aggregate channel is electrically connected to the identification device, which can identify the label corresponding to the aggregate channel and control the air nozzle 8 on the classification component to blow the product with the corresponding label to the corresponding aggregate channel. By using different labels to define the products, the identification of multiple products is facilitated, and the products enter the corresponding aggregate channel for separation.

[0037] like Figure 2 As shown, this embodiment also provides a product separation method for an intelligent photoelectric sorting machine, which is applicable to the intelligent photoelectric sorting machine described above, and includes the following steps:

[0038] S10. The products to be sorted are defined as the first label, the second label to the Nth label in sequence.

[0039] S20, marking the aggregate channel in order from the end closest to the belt transmission mechanism 4 according to the number of labels from most to least.

[0040] S30: The products to be sorted pass through the flattening device 6 and enter the identification device.

[0041] S40: The X-ray linear array detector 22 obtains the equivalent energy attenuation rate of the X-ray continuous energy spectrum of the product to be sorted, and determines the probability that the item to be sorted has the above-defined label.

[0042] S50: The image recognition system 3 obtains image information of the product to be sorted and determines the probability that the product to be sorted has the above-defined label.

[0043] S60, the X-ray fluorescence receiver 23 obtains the fluorescence spectrum information of the product to be sorted, and judges the probability of the product to be sorted being the label defined above.

[0044] S70, the probabilities of the same label obtained in the above steps S40-S60 are weighted and added, and the probability of each label is obtained.

[0045] S80, according to the probability preset threshold value set for the product to be sorted, the product is judged to be the label.

[0046] S90, the determined label signal is transmitted to the material collecting channel and compared with the label marked by the material collecting channel.

[0047] S100, the material collecting channel transmits the signal to the sorting mechanism according to the label corresponding to the channel.

[0048] S110, the sorting mechanism controls the electromagnetic valve to control the air nozzle 8 to blow the product to be sorted into the corresponding channel.

[0049] In an embodiment, the product separation method of the intelligent photoelectric sorting machine comprises: defining first label to Nth label for N kinds of products in order of the proportion of each kind of product from large to small according to the preset proportion of each kind of product; setting N material collecting channels corresponding to the first label to the Nth label from the nearest end of the belt drive mechanism 4; controlling the product to be sorted to enter the identification device after passing through the flattening device 6; controlling the X-ray linear array detector 22 to obtain the X-ray continuous energy spectrum equivalent energy attenuation rate of the product to be sorted, and obtaining the first probability of the product to be sorted corresponding to different label numbers according to the X-ray continuous energy spectrum equivalent energy attenuation rate; controlling the X-ray fluorescence receiver 23 to obtain the fluorescence spectrum information of the product to be sorted, and obtaining the second probability of the product to be sorted corresponding to the different label numbers according to the fluorescence spectrum information; controlling the image identification system 3 to obtain the image information of the product to be sorted, and obtaining the third probability of the product to be sorted corresponding to the different label numbers according to the image information; weighting and adding the probabilities of the same label number of the product to be sorted obtained, and obtaining the probability of the product to be sorted corresponding to each label number; judging the label number corresponding to the product to be sorted according to the preset probability threshold value; transmitting the determined label number to the N material collecting channels, and comparing the label number corresponding to the N material collecting channels to determine the material collecting channel corresponding to the product to be sorted; controlling the material collecting channel to transmit the signal to the electromagnetic valve according to the label corresponding to the material collecting channel; and blowing the product to be sorted into the material collecting channel through the electromagnetic valve controlling the air nozzle 8.

[0050] Optionally, the parameter μ is the equivalent energy attenuation rate of the X-ray continuous energy spectrum of the product to be sorted, and the calculation method of μ is as follows:

[0051] μ=(μ1 / μn)×[n / (1+2+...+n)]+μ2 / μ(n-1)×[(n-1) / (1+2+...+n)]+...+(μn / 2) / (μn / 2+1)×[(n / 2+1) / (1+2+...+n)];

[0052] Among them, μ1 is the energy attenuation rate of the X-ray after passing through the material at the 1Kev energy level; μ2 is the energy attenuation rate of the X-ray after passing through the material at the 2Kev energy level; μn is the energy attenuation rate of the X-ray after passing through the material at the nKev energy level; n is the maximum energy level of the X-ray system, and n is an even number.

[0053] In the related technology, high and low energy spectra are used for transmission. This method adopts a transmission full energy spectrum detection algorithm to calculate the transmission X-rays of multiple energy levels. The boundaries of multiple energy levels are clear and the data dimensions are rich, which solves the problem of unclear high and low energy boundaries in continuous spectrum X-ray identification; through multi-dimensional data, the influence of thickness differences of the tested products on the test results is effectively eliminated.

[0054] In one embodiment, since the products that account for a larger proportion of the products to be sorted are not exactly the same each time, for example, in coal sorting, when coal accounts for a larger proportion, coal is sorted and gangue is discharged, and when gangue accounts for a larger proportion, gangue is selected and coal is discharged. In the field of grain color sorting or solid waste sorting, the types of products to be sorted change, and the products that account for a larger proportion in each batch of products change. Therefore, it is necessary to define the product with the largest quantity as the first label to save the number of blowing times during sorting.

[0055] For products near the threshold, misjudgment often occurs, especially in the field of coal preparation. The X-ray equivalent energy attenuation rate of coal blocks containing gangue and gangue containing coal will be close to or even the same as that of coal or gangue due to the different content of coal or gangue. In this case, misjudgment will occur in the sorting. In order to avoid such situations, this application uses image recognition system and X-ray fluorescence for auxiliary identification, and sets up a voting mechanism. Since X-ray continuous photoelectron spectroscopy is an identification of the atomic sequence of the product, it has higher accuracy than the other two identification methods, so its voting ratio is higher. The weight ratio is set to 0.7, the weight ratio of the image recognition system is set to 0.2, and the weight ratio of X-ray fluorescence is set to 0.1. The setting of weights is not limited to the above values, and is set and adjusted according to the characteristics of different products to be selected in different fields.

[0056] In one embodiment, the X-ray continuous energy spectrum, image recognition system, and X-ray fluorescence are compared and label probabilities are defined separately according to the set thresholds. Then, the label probabilities obtained by each method are voted by weight. For example, the X-ray continuous energy spectrum determines that the probability of the product being the first label is 80%, and the probability of being the second label is 20%. The image recognition system determines that the probability of the product being the first label is 70%, and the probability of being the second label is 30%. The X-ray fluorescence determines that the probability of the product being the first label is 40%, and the probability of being the second label is 60%. The weight of the X-ray continuous energy spectrum is set to 0.7, the weight ratio of the image recognition system is set to 0.2, and the weight ratio of the X-ray fluorescence is set to 0.1. After weighted voting, the first The label votes are 0.8*0.7+0.2*0.7+0.1*0.4=0.74, and the second label votes are 0.7*0.2+0.2*0.3+0.1*0.6=0.26. The preset threshold value defines that when the first label votes exceed 0.6, it is judged to be the first label, and the above votes judge that the item is the first label. If the preset threshold value defines that when the first label votes exceed 0.75, it is judged to be the first label, then the above votes judge that the item is the second label. The preset threshold is defined according to the product characteristics, and the setting of the above threshold is set according to the product characteristics and is not limited to the above setting. This application adopts three identification methods to identify separately, and finally calculates and compares according to the weighted voting method to obtain the category of the product.

[0057] This application uses X-ray identification, image recognition system, X-ray fluorescence spectroscopy to assist in identification, and adopts a voting weight mechanism. After the weights are added, the label with the largest proportion is identified as the product, which greatly improves the sorting precision and reduces the error rate.

Claims

1. An intelligent photoelectric sorting machine, comprising a belt transmission mechanism (4), a distinguishing device and a collecting mechanism (9); The belt transmission mechanism (4) is configured to convey a plurality of products to be sorted; The identification device is configured to identify the multiple products to be sorted and define the multiple products to be sorted one by one with different label numbers; The collecting mechanism (9) comprises a plurality of collecting channels, each of which is electrically connected to the identification device and is configured to identify the label number corresponding to the collecting channel and control the products to be sorted corresponding to the label number to enter the collecting channel; The identification device includes at least one of the following: an X-ray system (2), an image identification system (3); The X-ray system (2) is configured to perform radiographic identification on the plurality of products to be sorted; The image recognition system (3) is configured to perform image recognition on the plurality of products to be sorted; The X-ray system (2) includes an X-ray emitting mechanism (21), an X-ray linear array detector (22) and an X-ray fluorescence receiver (23); in, The X-ray emitting mechanism (21) is configured to emit X-rays with a continuous energy spectrum; The X-ray linear array detector (22) is configured to receive X-rays generated after the continuous energy spectrum of the X-rays is transmitted through each product to be sorted, and to determine a first probability that the product to be sorted corresponds to the different label numbers based on the received X-rays; The X-ray fluorescence receiver (23) is configured to receive X-ray fluorescence rays generated after X-rays of a continuous energy spectrum are reflected by each product to be sorted, and to determine a second probability that the product to be sorted corresponds to the different label numbers based on the received X-ray fluorescence rays; The image recognition system (3) is configured to obtain image information of each product to be sorted, and determine a third probability that the product to be sorted corresponds to the different label numbers based on the image information; The identification device is configured to obtain a label number corresponding to each product to be sorted according to a weighting mechanism based on at least one of the following, and transmit the label number to the plurality of collecting channels: the first probability and the second probability; the third probability; the first probability, the second probability, and the third probability; The weights are set and adjusted according to the characteristics of different products to be selected in different fields.

2. The intelligent photoelectric sorting machine according to claim 1, wherein: The number of the plurality of aggregate channels is greater than or equal to the number of preset product categories.

3. The intelligent photoelectric sorting machine according to claim 1 further includes a flattening device (6), which is arranged below the belt transmission mechanism (4), and the flattening device (6) is configured to drive the conveyor belt of the belt transmission mechanism (4) to vibrate so that the multiple products to be sorted are flattened on the conveyor belt.

4. The intelligent photoelectric sorting machine according to claim 1 further includes a cleaning mechanism (1), which is arranged above one end where the multiple products to be sorted fall, and the cleaning mechanism (1) is configured to remove dust generated by the multiple products to be sorted.

5. The intelligent photoelectric sorting machine according to claim 1 further includes a feeding mechanism (5), which is arranged above the belt transmission mechanism (4), and the feeding mechanism (5) is configured to transport the multiple products to be sorted to the belt transmission mechanism (4).

6. The intelligent photoelectric sorting machine according to claim 3, wherein: The belt transmission mechanism (4) comprises a driving member, an active roller and a driven roller, the conveyor belt is annular and is respectively sleeved outside the active roller and the driven roller, the active roller is connected to the driven roller through the conveyor belt, and the driving member is in transmission connection with the active roller.

7. The intelligent photoelectric sorting machine according to claim 1, further comprising a sorting component, wherein the sorting component is configured to blow each product to be sorted into a collecting channel corresponding to the product to be sorted; The classification component includes an air nozzle (8), an air storage tank (7) and an air compressor; The air nozzle (8) is arranged at one end of the belt transmission mechanism (4) where the multiple products to be sorted fall, and the air nozzle (8) is provided with a solenoid valve, which is electrically connected to the multiple collecting channels; The gas storage tank (7) is configured to provide high-pressure gas to the air nozzle (8); The air compressor is configured to replenish the gas storage tank (7) with high-pressure gas.

8. A product separation method for an intelligent photoelectric sorting machine, applicable to the intelligent photoelectric sorting machine according to any one of claims 1 to 7, comprising: Define first to Nth labels for the N categories of products in descending order of proportions corresponding to the N categories of products, where N is greater than 1; Starting from the end closest to the belt transmission mechanism (4), N collecting channels are provided corresponding to the first label to the Nth label in sequence; Controlling the products to be sorted to pass through the paving device (6) and then enter the identification device; Controlling the X-ray linear array detector (22) to obtain an equivalent energy attenuation rate of an X-ray continuous energy spectrum of the product to be sorted, and obtaining a first probability that the product to be sorted corresponds to different label numbers based on the equivalent energy attenuation rate of the X-ray continuous energy spectrum; controlling an X-ray fluorescence receiver (23) to obtain fluorescence spectrum information of the product to be sorted, and obtaining a second probability that the product to be sorted corresponds to the different label numbers based on the fluorescence spectrum information; Controlling the image recognition system (3) to obtain image information of the product to be sorted, and obtaining a third probability that the product to be sorted corresponds to the different label numbers based on the image information; Performing weighted summation on the obtained probabilities of the same label number of the products to be sorted to obtain the probability of the products to be sorted corresponding to each label number; Determine the label number corresponding to the product to be sorted according to a preset probability threshold; Transmitting the determined label number to the N collecting channels, and comparing it with the label numbers corresponding to the N collecting channels to determine the collecting channel corresponding to the product to be sorted; controlling the material collecting channel to transmit a signal to the solenoid valve according to a label corresponding to the material collecting channel; The solenoid valve controls an air nozzle (8) to blow the product to be sorted into the collecting channel.

9. The method according to claim 8, wherein μ is the equivalent energy attenuation rate of the X-ray continuous energy spectrum, and the calculation method of μ is as follows: μ=(μ1 / μn)×[n / (1+2+...+n)]+μ2 / μ(n-1)×[(n-1) / (1+2+...+n)]+...+(μn / 2) / (μn / 2+1)×[(n / 2+1) / (1+2+...+n)]; Among them, μ1 is the energy attenuation rate of the X-ray after passing through the material at the 1Kev energy level; μ2 is the energy attenuation rate of the X-ray after passing through the material at the 2Kev energy level; μn is the energy attenuation rate of the X-ray after passing through the material at the nKev energy level; n is the maximum energy level of the X-ray system, and n is an even number.

Images