Multi-mode heavy metal sorting system and method based on dual-energy X rays

Abstract

The invention provides a multi-mode heavy metal sorting system based on dual-energy X-rays. The multi-mode heavy metal sorting system comprises a material conveying subsystem, an X-ray signal detection subsystem, a computer subsystem and a material separation subsystem. According to the technical scheme, the problems that in the existing sorting technology, materials are limited, the sorting efficiency is limited, the fine foreign matter sorting capacity is insufficient, the machine maintenance cost is high, the environment is polluted, and complex mixtures cannot be sorted are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of material detection, in particular to the technical field of X-ray transmission material sorting, and can be used in scenarios such as the recycling industry of waste metals, mining quarries, and geological exploration. Specifically, it relates to a dual-energy X-ray multimodal heavy metal sorting system and method. Background Art

[0002] The process of metal resource utilization includes four steps: collection, crushing, sorting, and smelting. Among them, the sorting step is a precondition for smelting, responsible for separating different target metals in the mixed materials, and ensuring the relative purity of the materials before smelting as much as possible, which directly determines the quality and efficiency of smelting. In practical applications, related industries such as ore mining and waste metal recycling have application requirements for metal sorting. In these application scenarios, the materials to be sorted are diverse, including non-metallic impurities, alloys, or materials with different metals intermingled. Therefore, how to solve the problems of high-speed and high-precision target recognition and screening in the sorting process for different application scenarios and different materials is the key to improving the comprehensive utilization efficiency of metal resources and realizing its energy-saving, environmental protection, and economic development.

[0003] X-ray sorting is a new sorting technology, mainly divided into two types: The first is the X-ray fluorescence spectroscopy (XRF) sorting technology, which uses the excitation effect of X-rays on atomic substances and judges the metal components of substances by the spectra of the excited fluorescence and then screens. The X-ray fluorescence spectroscopy technology has a low coupling degree with metal types and can generally identify most metals, but the spectral analysis equipment usually does not have spatial identification ability, and only single-piece materials can be identified in one sampling, with low efficiency; The second is the X-ray transmission sorting technology, which uses the penetration effect of X-rays and classifies according to the different absorption rates of materials for X-rays. It can collect the shape, position, and internal structure information of materials more clearly, and the dual-energy X-ray transmission technology (DE-XRT) can also use the ratio of the absorption rates of materials for X-rays at different energy levels to reduce the influence of material size or thickness attributes on recognition. However, the calibration of characteristic information depends on the composition of the sorting object, lacks robustness, and has low sensitivity to heavy metal materials.

[0004] Related companies at home and abroad have already used X-ray transmission recognition technology for material sorting and put it into actual production. For example, patent CN202123065029.2 uses dual-energy X-rays. By taking advantage of the different attenuation rates of X-rays transmitted through metals at different energy levels, it identifies the types of metals and then uses the method of blowing to sort. However, this invention only uses X-ray transmission technology to identify metal blocks and cannot utilize the fluorescence excitation effect of X-rays on metals, so its recognition ability is limited. Patent CN202210710985.1 also uses X-rays to identify the types of metals and then sort. The difference is that this sorting method uses a mechanical structure to organize the passing blocks into a single-line channel for separation, and the sorting efficiency is limited. Another research on the key algorithm and system for high-speed sorting of metal crushed materials by Huang Jing (Master's thesis of Xiamen University, 2023) proposes a sorting method based on dual-energy X-ray transmission linear array imaging technology and pulse spectrum analysis technology. However, there are problems in the fusion between the linear array image and the spectrum in this method, and spectral data matching loss is likely to occur during the process of image frame switching.

[0005] However, the current X-ray sorting technology still has the following problems:

[0006] 1) Limited sorting effect: The current X-ray sorting technology still has limitations in the sorting effect of certain substances. For example, it performs well in the sorting of aluminum materials, but has a poor sorting effect on other heavy metals.

[0007] 2) High particle size requirements: Some X-ray sorting equipment has relatively high requirements for the particle size range of materials, which reduces its versatility and applicable range.

[0008] 3) Difficulty in balancing efficiency and accuracy: The single fluorescence spectroscopy technology does not have the ability of spatial positioning. Therefore, the object block recognition technology based on fluorescence spectroscopy can only recognize a single object block each time it samples. Adopting a multi-channel recognition scheme can solve the problem of spatial identification to a certain extent, but it is still difficult to distinguish object blocks at different positions within the channel, lacking sufficient spatial identification ability.

[0009] 4) High cost: Due to the high technical complexity, the cost of X-ray sorting equipment is higher than that of traditional mechanical sorting equipment.

[0010] Therefore, it is necessary to develop a new type of intelligent sorting equipment that can efficiently and accurately sort materials, reduce maintenance costs and the impact on the environment, and solve the problem that the current sorting technology cannot be well applied to the sorting of heavy metals in scenarios such as the recycling industry of waste metals, mines and quarries, and geological exploration. Summary of the Invention

[0011] To achieve the above object, the present invention discloses a heavy metal sorting system and method integrating dual-energy X-ray transmission (DE-XRT) and X-ray fluorescence spectroscopy (XRF) technologies, which can solve the problems existing in the existing sorting technologies, such as the limitation of materials, limited sorting efficiency, insufficient sorting ability for fine foreign matters, high machine maintenance cost, environmental pollution, and inability to cope with the sorting of complex mixtures.

[0012] According to one aspect of the present invention, there is provided a dual-energy X-ray multimodal heavy metal sorting system, comprising: a material conveying subsystem for conveying the material to be sorted to the detection area; an X-ray signal detection subsystem for collecting dual-energy transmission images (DE-XRT) and X-ray fluorescence spectrum data (XRF) by irradiating the material transported to the detection area with X-rays; a computer subsystem for performing data exchange and control operations with the X-ray signal detection subsystem, the material conveying subsystem, and the material classification subsystem, and performing sorting processing based on multimodal data to obtain the composition information of the material to be sorted; and a material separation subsystem for separating the material to be sorted according to the composition information of the material to extract the target heavy metal components.

[0013] Optionally, the X-ray signal detector subsystem in the above heavy metal sorting system includes: a dual-energy X-ray linear array detector unit for converting the obtained transmission image signal of the material to be sorted into a digital signal; and an X-ray fluorescence spectrum detector unit for collecting the X-ray fluorescence spectrum data generated after the metal material is excited by X-rays and preprocessing and transmitting it to the computer subsystem.

[0014] Optionally, the UDP transmission protocol is adopted to transmit the multimodal data to be processed to the computer subsystem in the above heavy metal sorting system.

[0015] Optionally, the dual-energy X-ray linear array detector unit in the above heavy metal sorting system is a multi-channel high-low energy X-ray transmission data acquisition system.

[0016] Optionally, the X-ray fluorescence spectrum detector unit in the above heavy metal sorting system is a multi-channel X-ray spectrum data receiving system.

[0017] Optionally, the dual-energy X-ray transmission source in the above heavy metal sorting system is an open X-ray tube or a conical X-ray tube.

[0018] Optionally, the material conveying subsystem in the above heavy metal sorting system includes a hopper and a belt. The hopper evenly scatters the material on the belt by a vibration motor for transportation by the belt, and the belt material is sent to the material blowing port by the conveyor belt through the discharge port.

[0019] Optionally, the computer subsystem in the above heavy metal sorting system includes an industrial control computer.

[0020] According to another aspect of the present invention, there is provided a dual-energy X-ray multi-modal heavy metal sorting method, including: a material conveying subsystem conveys the material to be sorted to a detection area; an X-ray signal detection subsystem emits X-rays to the material conveyed to the detection area, and acquires dual-energy transmission images (DE-XRT) and fluorescence spectrum data (XRF);

[0021] A computer subsystem performs data exchange and control operations with the X-ray signal detection subsystem, the material conveying subsystem, and the material separation subsystem, and performs sorting processing based on multi-modal data to obtain the composition information of the material to be sorted; the material separation subsystem separates the material to be sorted according to the material composition to extract the target heavy metal components.

[0022] Optionally, the above heavy metal sorting method further includes that the sorting processing based on multi-modal data includes linear array image framing, fluorescence information processing, multi-modal fusion, and deep learning model recognition.

[0023] Optionally, the above heavy metal sorting method further includes that during the framing process, two buffer areas are opened in the computer memory to store two adjacent frames of images. Each frame of image includes three attributes: the number of rows, the number of columns, and the pixel gray value. The number of rows m of the combined two-dimensional image is divided into a front transition area m pro 、a main area m main and a rear transition area m post from top to bottom. The content of the rear transition area of the previous frame is the same as that of the front transition area of the next frame.

[0024] Optionally, the above heavy metal sorting method further includes that during the framing process, it is necessary to separate the background information and the object block information and perform gray value balancing.

[0025] Optionally, the above heavy metal sorting method further includes that the fluorescence information processing process includes target object block spectral wavelength calibration and timestamp recording.

[0026] Optionally, the above heavy metal sorting method further includes that the multi-modal fusion includes spectral signal storage, time-space alignment, and image channel fusion.

[0027] Optionally, the above heavy metal sorting method further includes that the deep learning model recognition includes data set construction and model training.

[0028] Optionally, the above heavy metal sorting method further includes that the material separation subsystem selects a rectangle as the feature of the material contour to locate the position of the material to be sorted.

[0029] According to the dual-energy X-ray multi-modal heavy metal sorting system and method of the embodiments of the present invention, various heavy metals can be sorted, not just aluminum materials, and it has a strong sorting ability for fine foreign matters, greatly reducing the limitations on the types and particle sizes of materials, and has high flexibility. Moreover, through automated operations, the need for manual intervention is reduced, greatly improving the sorting efficiency. At the same time, by using the heavy metal sorting system with programmable software and hardware collaboration in the embodiments of the present invention to process and identify multi-modal data, high-precision information on the components of the materials to be sorted can be obtained, thereby achieving more accurate sorting results. By applying this technical solution to the sorting scenarios of heavy metals in waste metal recycling, mine quarrying, geological exploration, etc., not only the efficiency and accuracy are improved, but also the maintenance cost and environmental impact are reduced at the same time, thus bringing higher economic and environmental benefits.

[0030] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it used to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The drawings incorporated into the specification and constituting a part of the specification illustrate the embodiments of the present invention and, together with the related written description, are used to explain the principles of the present invention. In these drawings, like reference numerals are used to represent like elements. The drawings in the following description are some embodiments of the present invention, not all embodiments. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0032] Figure 1 is a block diagram of the dual-energy X-ray multi-modal heavy metal sorting system provided by the embodiments of the present disclosure;

[0033] Figure 2 is a schematic diagram of the nozzle array provided by the embodiments of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0034] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention. It should be noted that, without conflict, the embodiments and features in the embodiments of the present application can be combined with each other arbitrarily.

[0035] The dual-energy X-ray multimodal heavy metal sorting system provided by the embodiments of the present invention is applicable to heavy metal sorting in scenarios such as waste metal recycling, mine quarrying, and geological exploration, which can improve sorting flexibility, efficiency, and accuracy, and has advantages in reducing maintenance costs and environmental impacts.

[0036] Figure 1 It is a structural diagram of the dual-energy X-ray multimodal heavy metal sorting system provided by the embodiments of the present invention. Figure 1 The heavy metal sorting system in it includes:

[0037] a. The material conveying subsystem includes a hopper (1) and a belt (2). Exemplarily, the hopper evenly scatters the material on the belt through a vibrating motor, and then the belt transports the material to be measured from the feeding port through the detection area conveyor belt to the material blowing port for blowing;

[0038] b. The X-ray detection subsystem includes a dual-energy X-ray linear array detector (DE-XRT, 3), an X-ray fluorescence detector (XRF, 4), an X-ray isolation protection housing (5), and a dual-energy X-ray source (6). Exemplarily, the dual-energy X-ray source (6) is used to project X-rays onto the material, and then the dual-energy X-ray linear array detector (3) collects and preprocesses the X-ray transmission multimodal data and transmits the multimodal data to the computer subsystem. At the same time, the fluorescence detector (4) synchronously collects the fluorescence signal and transmits it to the computer. In a preferred embodiment, the X-ray linear array detection system (3) and the X-ray fluorescence detector (4) use multiple devices in parallel to collect multi-channel data, so as to achieve wide-range sampling.

[0039] c. The computer subsystem includes a PLC (9) and an industrial control computer (10). Exemplarily, the industrial control computer (9) exchanges data and performs control operations with the dual-energy X-ray transmission source, the linear array detector, the material conveying subsystem, and the material separation subsystem through UDP and serial ports, and obtains the component information of the material to be sorted based on the multimodal data material recognition method, and controls the action of the separation subsystem through the PLC.

[0040] d. The material separation subsystem (a blowing head (7) and an air pump (8)) is used to separate the material to be sorted according to the material component information to extract the target metal component. The total length of the blowing head array (7) is set to 1 m, which is the same as the width of the transport belt. It consists of 198 jet valves. The 198 jet valves are divided into 5 groups. Each group of valves can flexibly adjust the blowing range, blowing time, blowing angle, and blowing air pressure in a combined manner through a pneumatic motor, and finally blow the material with airflows at different pressures and angles. Figure 2 It shows a schematic diagram of the nozzle array provided by the embodiments of the present disclosure.

[0041] Exemplarily, the multimodal data material recognition method of the computer subsystem in c. includes the following steps:

[0042] (a) Linear array acquisition of dual-energy X-ray projection image I m×n×2 , where n represents the number of pixels in the row pixel sequence obtained by each sampling of the linear array, m represents the number of sampling rows included in one frame of the image, and 2 indicates that a pixel point includes two signals of high energy and low energy. Since each pixel contains two values of low-energy signal and high-energy signal, the specific representation form of the pixel point is (p h , p l ), where p h represents the high-energy signal intensity, and p l represents the low-energy signal intensity;

[0043] (b) Fluorescence detector collects fluorescence spectrum data;

[0044] (c) Align the linear array and fluorescence sampling times to generate a sampling picture containing transmission information and fluorescence spectrum information, and achieve multimodal fusion;

[0045] (d) Identify the types of blocks in the image through the yolov8 algorithm;

[0046] (e) Encode the information into a PLC control signal to control the spraying of the spray head.

[0047] The method for obtaining the dual-energy X-ray projection image in step (a) includes the following steps:

[0048] (a1) The linear array samples one row of one-dimensional pixel data each time. During the operation of the device, the sensor continuously samples and sends the data to the computer. The computer combines multiple rows of one-dimensional pixel data from the top to the bottom of the picture in chronological order and with the specified maximum number of rows to form a frame of two-dimensional image. The abscissa of each pixel in the image represents its spatial position, forming a mapping with the spray head number, indicating which spray heads need to spray, and the ordinate contains the moment information of the block passing through the linear array, indicating when the spray head should spray;

[0049] (a2) The computer separates the block information and background information in the image and balances the gray value differences at different positions in the image;

[0050] Among them, the process of combining the two-dimensional image in step (a1) specifically refers to combining a two-dimensional image formed by several rows of one-dimensional linear array pixel sequences. Since the block may span frames, that is, the upper part is in the previous frame while the lower part is in the next frame, in order to ensure the integrity of the block information during the image processing process, the number of rows m of the combined two-dimensional image is divided into a front transition region m pro , a main region m main and a rear transition region m post, the content of the post-transition region of the previous frame is the same as that of the pre-transition region of the next frame. In this way, for the object blocks in the image, as long as it is ensured that the object blocks do not span two regions, the object blocks must retain a complete image in one of the two consecutive frames.

[0051] The frame grouping process in step (a1) includes the following steps:

[0052] (a1-1) Allocate two buffer areas in the computer memory to store two adjacent frames of images. Each frame of image includes three attributes: the number of rows, the number of columns, and the pixel gray level. After the linear array sensor starts sampling, every time the sensor collects a row of pixels, it fills a row of pixels into the space of one or two frames of images in the buffer area. After completing the data collection and content filling of the specified number of rows, the picture is output to complete the image sampling, and at the same time, the time of the first row of data input is saved as the time stamp of this frame of picture.

[0053] (a1-2) Set the number of transition rows. For each frame of image, save the data to the buffer area according to the order of the pre-transition region, the main region, and the post-transition region;

[0054] (a1-3) When starting to combine the content of the post-transition region after completing the frame grouping of the main region of the previous frame of image, at the same time copy the row information being combined to the pre-transition region of the next frame of image for frame grouping, ensuring that the content of the post-transition region of the previous frame is the same as that of the pre-transition region of the next frame.

[0055] Among them, step (a2) of separating the object block information and the background information in the image specifically refers to: statistically calculating the background gray level, and separating the object block information and the background information from the image through differential calculation. Since the X-ray light source is a point light source, there is a gray level difference between the two sides far from the light source and the middle close to the light source under the state of high-speed sampling. Therefore, it is necessary to balance the color difference at each position.

[0056] Step (a2) includes the following steps:

[0057] (a2-1) In the initialization stage before blanking, collect several rows of linear array pixel data g i，j , where i represents the index of the pixel in the row sequence, j represents the serial number of each row, and calculate the average gray level of each pixel Add a certain constant quantity to the average gray level as the threshold to distinguish background pixels and object block pixels;

[0058] (a2-2) According to the Lambert-Beer law, the X-ray transmission energy is proportional to the incident energy. Therefore, the parts with lower gray levels on both sides of the image can be kept consistent with the gray level of the middle part through linear correction. According to the average gray level of each pixel obtained in step (a-1) Calculate the average value of the entire row of linear array data Calculate each pixel and The difference between As the grayscale correction value, during subsequent adoption, this value is added to each pixel for correction, and finally the corrected image is output.

[0059] The specific steps of step (b) include the following steps:

[0060] (b1) Before implementing the blowing operation, several blocks are sampled from the material to be measured, including positive samples and negative samples. Then, a calibration experiment without blowing is implemented, the response spectral curves of each channel are sampled, and the target frequency range is determined;

[0061] (b2) During the sorting operation, the sensor collects the sampling number at a certain sampling frequency and sends it to the computer. The information sent includes the spectral signal and the timestamp of the signal acquisition moment. The computer analyzes the distribution of the light intensity in the spectral signal, sets a threshold, and when the light intensity in the specified spectral frequency range reaches the threshold, it is determined that there is a block in the sampling channel at this moment.

[0062] The specific steps of step (c) include the following steps:

[0063] (c1) Set two spectral signal buffer spaces S pro and S post in the computer memory, where the S pro space directly stores the spectral signals sent by the sensor, and S post is used to store the spectral signals that may span two frames of images;

[0064] (c2) When the computer completes the framing process of a frame of line array image, it preferentially traverses the spectral cache data in S post , compares the spectral signal time with the line array image time, matches the spectral data that lags behind the image timestamp and is in the pre-transition and main body intervals with the blocks in the transmission image, and discards other time cache information in S post ;

[0065] (c3) Then traverse the cache data in S pro , match the spectral data that lags behind the image timestamp and is in the pre-transition and main body intervals with the blocks in the transmission image, discard the data whose time is earlier than the line array sampling timestamp, and save the data that lags behind the image timestamp and is in the post-transition region to the S post cache.

[0066] (c4) Expand the dimension of each pixel point in step (a1), from the original (p l , p h ) to (p h , p l, f, s), where f represents the median value of the target frequency range obtained in step (b1), and s represents the characteristic signal strength, thereby converting the original high- and low-energy linear array image I m×n×2 into I m×n×4 in the form of a high-dimensional fusion image.

[0067] In steps (c2) and (c3), the matching of spectral data with the object blocks in the transmission image specifically includes the following steps:

[0068] (c2-1) According to the object block information and background information separated in step (a2), calculate the rectangular box of each object block in the linear array image. In step (a1), each pixel represents two kinds of information, time and space. Therefore, the upper edge of the rectangular box represents the moment when the spray head starts spraying, the length represents the spraying duration, the left edge represents the serial number of the spray head that needs to start spraying, and the width represents how many spray heads need to spray.

[0069] (c2-2) After the computer retrieves the fluorescence data of a certain channel from the cache, map it to the linear array image space, calculate the difference between the fluorescence signal and the time represented by the midpoint of the specified object block rectangular box. When the difference is less than a certain threshold, it is determined that the fluorescence signal is preliminarily matched with the object block, and the time alignment is completed. Then, it is judged whether the object block is entirely or partially located within the sensor channel that collects the signal. If it is within the channel, it represents spatial alignment.

[0070] Mapping the spectral signal to the image space in step (c2-2) specifically includes the following steps:

[0071] (c2-21) Vertically divide the picture into several vertical strips according to the number of channels, and each vertical strip represents the sensing area of a fluorescence sensor.

[0072] (c2-22) Convert the fluorescence information into a similar rectangular box in the form of picture pixels. The upper edge of the rectangular box represents the response moment of the fluorescence signal, the length represents the response duration, and the left and right edges of the rectangle are located on the dividing lines of the response intervals of the corresponding channels.

[0073] Step (d) specifically includes the following steps:

[0074] (d1) Collect the data set required for the training of the deep learning model;

[0075] (d2) Divide the data set into a training set, a validation set, and a test set, and then train the yolov8 model;

[0076] (d3) Deploy the model to the computer on-site for classification.

[0077] Step (d1) specifically includes the following steps:

[0078] (d1-1) Manually sort out some calibrated materials from the target materials, and independently conduct detection experiments according to the categories by imitating the real operation process to ensure that each picture contains only positive samples or negative samples.

[0079] (d1-2) Use the segmentation algorithm to segment the multi-modal block sub-pictures in the positive sample pictures and negative sample pictures, and enhance them by flipping, rotating, and scaling.

[0080] (d1-3) Set the maximum image resolution. Expand the pictures smaller than this resolution to the maximum resolution by padding, and reduce the pictures larger than this resolution to the same resolution by scaling.

[0081] (d1-4) Save all the sub-pictures as the model training data set.

[0082] Step (d3) specifically includes the following steps:

[0083] (d3-1) Through the segmentation algorithm, crop out each block from the input I m×n×4 picture, unify the sub-pictures of each block to the same resolution in the same way as in step (d1=3), and achieve efficient block classification through parallel inference of the computer graphics card in a batch processing manner.

[0084] Step (e) specifically includes the following steps:

[0085] (e1) Obtain the candidate box coordinates of each block, and through the formula

[0086]

[0087] map the pixel point coordinates to the nozzle number, where i represents the nozzle number corresponding to the pixel coordinates, h 1 represents the distance from the light source to the belt, h 2 represents the distance from the light source to the linear array, d 1 represents the physical distance from the projection of the light source on the belt to the leftmost side of the belt, d 2 represents the physical distance from the projection point of the light source on the belt to the leftmost side of the linear array imaging, n represents the number of nozzles, X represents the pixel abscissa; at the same time, through the formula

[0088] t = T diff ×y

[0089] solve how long it takes for the pixel coordinates to be blown, where T diff is the sampling period of the linear array camera, and y represents the pixel ordinate;

[0090] (e2) Obtain the block type information obtained by model inference, encode the block type in the form of numbers, and at the same time encode the obtained blow head serial number and blow time in step (e1) into a binary array. Then, the PLC program receives this array to complete the blow hardware control and implement the blow.

[0091] The dual-energy X-ray multi-modal heavy metal sorting system and method proposed in the embodiments of the present invention can sort a variety of heavy metals, not just aluminum materials, and have strong sorting ability for fine foreign matters, greatly reducing the limitations on the types and particle sizes of materials, and having high flexibility. Moreover, the need for manual intervention is reduced through automated operations, greatly improving the sorting efficiency. At the same time, by using the programmable software-hardware collaborative heavy metal sorting system in the embodiments of the present invention to process and identify multi-modal data, high-precision information on the components of the materials to be sorted can be obtained, thereby achieving more accurate sorting results. By applying this technical solution to the sorting scenarios of waste metal recycling, mine quarrying, etc., the problems existing in the existing sorting technologies, such as limitations on materials, limited sorting efficiency, insufficient sorting ability for fine foreign matters, high machine maintenance costs, environmental pollution, and inability to handle the sorting of complex mixtures, are solved. It not only improves the efficiency and accuracy but also reduces the maintenance costs and environmental impact at the same time, thus bringing higher economic and environmental benefits.

[0092] Those skilled in the art should understand that the embodiments of the present invention can be provided as a system and a method. Therefore, the present invention can be implemented in the form of a complete hardware embodiment, or in the form of an embodiment combining software and hardware aspects. Moreover, the present invention can be implemented in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0093] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, and the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0094] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce a manufacture including an instruction device that implements the function specified in one process Figure 1 or more processes and / or blocks Figure 1 or the functions specified in a block or more blocks.

[0095] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, so that the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in one process Figure 1 or more processes and / or blocks Figure 1 or the functions specified in a block or more blocks.

[0096] It should be noted that, in the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of other elements or steps not listed in the claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention can be implemented by means of hardware including several different elements and by means of a suitably programmed computer. In the unit claims listing several means, several of these means can be embodied by one and the same item of hardware. The use of the words first, second, and third, etc. does not denote any order. These words can be interpreted as names.

[0097] Although the preferred embodiments of the present invention have been described, additional changes and modifications can be made by those skilled in the art once they learn of the basic inventive concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications falling within the scope of the present invention.

[0098] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

[0099] In the description of the present invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and cannot be construed as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, "a plurality of" means two or more unless otherwise specifically defined.

[0100] In the present invention, unless otherwise clearly defined or limited, terms such as "installed", "connected", "joined", "fixed", etc. shall be construed in a broad sense. For example, it may be a fixed connection, a detachable connection, or integrated; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the internal communication of two components or the interaction relationship between two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0101] In the present invention, unless otherwise clearly defined or limited, the first feature being "on" or "under" the second feature may be that the first and second features are in direct contact, or the first and second features are indirectly in contact through an intermediate medium. Moreover, the first feature being "above", "over" and "on top of" the second feature may be that the first feature is directly above or obliquely above the second feature, or merely indicates that the first feature has a higher horizontal height than the second feature. The first feature being "under", "below" and "beneath" the second feature may be that the first feature is directly below or obliquely below the second feature, or merely indicates that the first feature has a lower horizontal height than the second feature.

[0102] In the description of this specification, the description with reference to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic descriptions of the above terms should not be understood as necessarily referring to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0103] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A dual-energy X-ray multi-modal heavy metal sorting system, characterized in that: include: The material conveying subsystem is used to convey the materials to be sorted to the detection area; An X-ray signal detection subsystem, used to collect dual energy transmission images (DE-XRT) and fluorescence spectrum data (XRF) by projecting X-rays to the material transmitted to the detection area; A computer subsystem, used to exchange data and perform control operations with the X-ray signal detection subsystem, the material conveying subsystem, and the material separation subsystem, and to perform sorting processing based on the multimodal data to obtain composition information of the material to be sorted; The material separation subsystem is used to separate the material to be sorted according to the material composition information to extract the target heavy metal components.

2. The multimodal heavy metal separation system according to claim 1, characterized in that: The X-ray signal detection subsystem includes: a dual-energy X-ray linear array detector unit, which is used to convert the acquired transmission image signal of the material to be sorted into a digital signal; and an X-ray fluorescence spectrum detector unit, which is used to collect the fluorescence spectrum data generated by the metal material after being excited by X-rays and transmit it to the computer subsystem after pre-processing.

3. The multimodal heavy metal separation system according to claim 2, characterized in that: The dual-energy X-ray linear array detector unit is a multi-channel high- and low-energy X-ray transmission data acquisition system.

4. The multimodal heavy metal separation system according to claim 2, characterized in that: The X-ray fluorescence spectrum detector unit is a multi-channel X-ray spectrum data receiving system.

5. The multimodal heavy metal separation system according to claim 1, characterized in that: The material conveying subsystem includes a hopper and a belt. The hopper scatters the material on the belt through a vibrating motor and the belt is transported. The belt material is transported from a discharge port to a material injection port by a conveyor belt.

6. A dual-energy X-ray multi-modal heavy metal sorting method, characterized in that: include: The material conveying subsystem conveys the materials to be sorted to the detection area; The X-ray signal detection subsystem collects dual-energy transmission images (DE-XRT) and fluorescence spectrum data (XRF) by projecting X-rays to the materials transmitted to the detection area; the computer subsystem exchanges data and performs control operations with the X-ray signal detection subsystem, the material conveying subsystem, and the material separation subsystem, and performs sorting processing based on the multimodal data to obtain the composition information of the materials to be sorted; The material separation subsystem separates the material to be sorted according to the material components to extract the target heavy metal components.

7. The multimodal heavy metal separation method according to claim 6, characterized in that: The sorting process based on the multimodal data includes linear array image framing, fluorescence information processing, multimodal fusion, and deep learning model recognition.

8. The multimodal heavy metal separation method according to claim 6, characterized in that: The linear array framing process opens two cache areas in the computer memory for storing two adjacent frames of images. Each frame of image includes three attributes: the number of rows, the number of columns, and the pixel grayscale. The number of rows m of the combined two-dimensional image is divided into the front transition area m from top to bottom. pro , main area m main and the post-transition region m post , the content of the post-transition area of ​​the previous frame is the same as the content of the pre-transition area of ​​the next frame.

9. The multimodal heavy metal separation method according to claim 7, characterized in that: The fluorescence information processing process includes target object block spectrum wavelength calibration and time stamp recording.

10. The multimodal heavy metal separation method according to claim 7, characterized in that: The multimodal fusion includes spectral signal storage, time-space alignment and image channel fusion.

Images