X-ray three-dimensional imaging and defect quantitative characterization method for ore internal defects in X-ray intelligent concentrating machine

Abstract

The invention discloses an X-ray three-dimensional imaging and defect quantitative characterization method for ore internal defects in an X-ray intelligent concentrating machine, and relates to the technical field of data processing, and the method comprises the following steps: acquiring an ore three-dimensional image by using an X-ray binocular imaging device; inputting the ore three-dimensional image into the ore internal defect characterization model to obtain ore internal defect characterization data, and dividing ore internal defect grades; acquiring current X-ray intelligent concentrator setting parameters, and constructing an expert experience sequence by using the ore internal defect characterization data and the current X-ray intelligent concentrator setting parameters; training the multi-agent reinforcement learning model to obtain an ore internal defect grade sorting parameter model; and the to-be-screened ore is conveyed to an X-ray intelligent concentrating machine, and ore with different ore internal defect grades is obtained through separation. According to the method, ores with different ore internal defect grades can be sorted, and the problem that the setting parameters of the X-ray intelligent concentrating machine are adaptively updated along with the internal defects of the ores is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method for X-ray three-dimensional imaging and defect quantitative characterization of internal defects of ore in an X-ray intelligent ore dressing machine. Background Art

[0002] Intelligent ore dressing machines can pre-select and discard ore in the ore production process, effectively improving the utilization rate of ore. The intelligent ore dressing machine can be structurally divided into three major mechanisms, namely the feeding mechanism, the detection and identification mechanism, and the sorting execution mechanism. The feeding mechanism is responsible for evenly feeding the ore into the detection area; the detection and identification mechanism uses an X-ray detector to scan the ore and performs task processing through an image processing algorithm; the sorting execution mechanism accurately controls the sorting device based on the information provided by the detection and identification mechanism, effectively separating the concentrate and waste ore, thereby improving the ore dressing efficiency and product quality. The entire system realizes the collaborative work between various mechanisms through a highly integrated control system to ensure the continuity and stability of the ore dressing process.

[0003] The Chinese patent application with the publication number CN 112221657 B discloses a phosphate ore photoelectric beneficiation and sorting process, wherein the phosphate ore is screened; the screened ore is transported to the photoelectric sorting machine via a belt conveyor, and the ore and material are identified by X-ray fluoroscopy, each piece of ore is identified by X-ray, and the ore and waste rock are distinguished by intelligent algorithm identification; after the photoelectric sorting machine has completed the identification, the concentrate and tailings are separated by precise striking with a high-speed air discharge gun, and the photoelectric beneficiation of the phosphate ore can be completed. This method can be used for photoelectric beneficiation of phosphate ore, but it does not take into account the problem that the internal defects of the beneficiated ore and the error of the beneficiation machine parameters will increase with the use time of the beneficiation machine.

[0004] The Chinese patent application publication number CN 110046653 B discloses a scheelite separation method based on XRT rays, including: transmitting the raw ore through the X-ray processing area and the injection area in sequence and obtaining the raw ore X-ray image; using an adversarial neural network to process the raw ore X-ray image to separate the scheelite ore and calculate the spatial coordinates of the scheelite ore; based on the spatial coordinates of the scheelite ore, calculating the falling time of the scheelite ore falling to the spray valve position in the injection area and controlling the spray valve to spray compressed air to the scheelite ore for separation based on the falling time. This method can accurately identify scheelite ore and reduce the throwing rate of the raw ore, but does not consider the impact of the multi-grade classification of ore and the change of the parameter error of the ore dressing machine on the ore dressing accuracy.

[0005] To this end, the present invention proposes an X-ray three-dimensional imaging and defect quantitative characterization method for internal defects of ore in an X-ray intelligent ore dressing machine. Summary of the invention

[0006] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention proposes an X-ray three-dimensional imaging and defect quantitative characterization method for internal defects of ore in an X-ray intelligent ore dressing machine, which can solve the problem of multi-grade classification of ore and the influence of ore dressing accuracy caused by ore dressing machine parameter error changes.

[0007] In view of the technical problems existing in the prior art, the present invention provides a method for X-ray three-dimensional imaging and defect quantitative characterization of internal defects of ore in an X-ray intelligent ore dressing machine, and the specific steps are as follows: S1. Obtain a three-dimensional image of the ore using an X-ray binocular imaging device; perform feature processing on the three-dimensional image of the ore to obtain a data set of internal defects of the ore; S2, inputting the ore three-dimensional image into the ore internal defect characterization model to obtain ore internal defect characterization data, and classifying the ore internal defect levels; the ore internal defect characterization model is trained by the ore internal defect data set; S3, obtaining the current setting parameters of the X-ray intelligent ore dressing machine, and constructing an expert experience sequence using the ore internal defect characterization data and the current setting parameters of the X-ray intelligent ore dressing machine; S4, storing the expert experience sequence in the expert experience pool, training the multi-agent reinforcement learning model to obtain the ore internal defect grade sorting parameter model; S5. The ore to be screened is conveyed to an X-ray intelligent ore concentrator, and ores with different internal defect grades are sorted based on an ore internal defect characterization model and an ore internal defect grade sorting parameter optimization model.

[0008] In particular, the X-ray binocular imaging device includes an X-ray source and two flat panel detector arrays.

[0009] In particular, the step of performing feature processing on the three-dimensional image of the ore to obtain the ore internal defect data set is: Step S11: normalizing, denoising and enhancing voxel data of the three-dimensional image of the ore; Step S12: using an edge detection algorithm to locate the edges of internal defects of the ore and obtain edge data of the defects; Step S13: segmenting the defect region using a region growing algorithm to obtain geometric data of the defect; Step S14: marking the internal defects of the processed ore three-dimensional image to obtain an ore internal defect data set; The ore is a particle-sized ore obtained after the original ore is crushed and vibrated; the original ore is a concentrate containing only one type of ore after pre-selection and discarding.

[0010] In particular, the steps of inputting the ore three-dimensional image into the ore internal defect characterization model to obtain the ore internal defect characterization data and classifying the ore internal defect levels are as follows: Step S21: dividing the ore internal defect data set into a training set and a validation set, inputting the data into a 3D CNN network model for training, and obtaining an ore internal defect characterization model; The ratio of the training set to the validation set is 7:3; the 3D-CNN consists of an input layer, a convolution layer, a pooling layer, a fully connected layer and an output layer, and the output layer outputs the internal defect characterization data of the ore; Step S22: inputting the ore three-dimensional image into the ore internal defect characterization model to obtain ore internal defect characterization data; The ore internal defect characterization data includes defect classification results and internal defect volume ratio, the defect classification results include two classification results: defective and non-defective, and the internal defect volume ratio refers to the ratio of the sum of the volumes of all internal defects in the ore to the volume of the ore; Step S23: classifying the internal defect levels of the ore according to the internal defect volume ratio; The method for classifying the internal defect levels of the ore is: Set a first threshold value X1 and a second threshold value X2; If the internal defect volume ratio is greater than and equal to the first threshold value X1, the internal defect level of the ore is level A; If the internal defect volume ratio is greater than the second threshold value X2 and less than the first threshold value X1, the internal defect level of the ore is level B; If the internal defect volume ratio is less than and equal to the second threshold value X2, the internal defect level of the ore is level C.

[0011] In particular, the current X-ray intelligent ore dressing machine setting parameters are obtained, and an expert experience sequence is constructed using the ore internal defect characterization data and the current X-ray intelligent ore dressing machine setting parameters, wherein: The current X-ray intelligent ore dressing machine setting parameters include feed belt speed, body belt speed, vibration amplitude, vibration frequency, vibration direction angle, and nozzle injection pressure; The expert experience sequence includes a set of ore internal defect states, a set of intelligent ore dressing machine parameter actions, a set of reward values, and a set of next states; The ore internal defect state set includes defect classification result data and internal defect volume ratio data; The intelligent ore dressing machine parameter action set includes feed belt speed data, body belt speed data, vibration amplitude data, vibration frequency data, vibration direction angle data, and nozzle injection pressure data.

[0012] In particular, the steps of storing the expert experience sequence in the expert experience pool and training the multi-agent reinforcement learning model to obtain the ore internal defect grade sorting parameter model are: Step S31: constructing an expert experience sequence using the ore internal defect characterization data and the current X-ray intelligent ore dressing machine setting parameters; The above is a set of internal defect states of the ore, a set of parameter actions of the intelligent ore dressing machine, a set of sorting accuracy reward values, and a set of internal defect states of the ore at the next moment; Step S32: storing the expert experience sequence into the expert experience pool, completing the initialization of the expert experience pool; Step S33: Using the expert experience pool, the multi-agent reinforcement learning model is trained to obtain the ore internal defect grade sorting parameter model.

[0013] Furthermore, the steps of storing the expert experience sequence in the expert experience pool and training the multi-agent reinforcement learning model to obtain the ore internal defect grade sorting parameter model are: Step S41: constructing a prediction network and a target network; Step S42: Construct N agents based on the prediction network, target network and experience pool, where is a variable and ; Step S43: Initialize the experience pool as an expert experience pool; the experience pool has a capacity of M; Step S44: Use random prediction network weights Initialize the prediction network with random target network weights Initialize the target network and initialize the parameter state set , batch n and step size m; Step S45: Randomly obtain expert experience sequences from the experience pool in batches n Conduct training, including ; The prediction network inputs the internal defect state set of the ore And intelligent ore dressing machine parameter action set , output prediction value function And the next moment intelligent ore dressing machine parameter action set ; The target network inputs the set of internal defect states of the ore at the next moment , output target value function , The calculation formula is: ; In the formula, is the set of defect states inside the ore Intelligent ore dressing machine parameter action set The operation effect reward value, is the discount factor and , is the set of internal defect states of the ore at the next moment The next moment intelligent concentrator parameter action set is selected The maximum predicted value function obtained; Step S46: Every m steps, the weight parameters of the prediction network are Weight parameters copied to the target network ; Step S47: Calculate the joint prediction value function of multiple agents , the calculation formula is: ; Calculate the joint objective value function of multiple agents , the calculation formula is: ; In the formula, Function and g function are calculated by BP neural network model; The multi-agent loss function is updated as , the calculation formula is: ; Step S48: Minimize the MSE loss function using gradient descent , update the network model parameters and Until convergence, the ore internal defect grade sorting parameter model is obtained; The prediction network and the target network have the same network structure, both using BP neural network as the basic network structure; the BP neural network consists of an input layer, a hidden layer and an output layer; The number of neurons in the input layer is equal to the number of defect state sets inside the ore; The number of neurons in the output layer is equal to the number of actions in the intelligent ore dressing machine parameter action set.

[0014] Another aspect of the present application provides an X-ray three-dimensional imaging and defect quantitative characterization system for internal defects of ore in an X-ray intelligent ore dressing machine, the system comprising: Ore image processing module: used to obtain ore three-dimensional images using X-ray binocular imaging devices; perform feature processing on the ore three-dimensional images to obtain ore internal defect data sets; Ore defect characterization module: used to input the ore three-dimensional image into the ore internal defect characterization model, obtain the ore internal defect characterization data, and classify the ore internal defect level; the ore internal defect characterization model is trained by the ore internal defect data set; Mining machine parameter acquisition module: used to obtain the current X-ray intelligent ore dressing machine setting parameters, and construct an expert experience sequence using the ore internal defect characterization data and the current X-ray intelligent ore dressing machine setting parameters; Mining machine parameter optimization module: used to store expert experience sequences into the expert experience pool, train the multi-agent reinforcement learning model to obtain the ore internal defect grade sorting parameter model; Intelligent ore sorting module: It is used to transfer the ore to be screened to the X-ray intelligent ore concentrator, and sort ores with different internal defect grades based on the ore internal defect characterization model and the ore internal defect grade sorting parameter optimization model.

[0015] Another aspect of the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in the above-mentioned method for X-ray three-dimensional imaging and quantitative characterization of internal defects of ore in the X-ray intelligent ore dressing machine are implemented.

[0016] Another aspect of the present application provides a readable storage medium storing a computer program, wherein the computer program is suitable for loading by a processor to implement the steps in the above-mentioned X-ray three-dimensional imaging and defect quantitative characterization method of internal defects of ore in the X-ray intelligent ore dressing machine.

[0017] Beneficial technical effects of the present invention: The present invention uses binocular imaging combined with X-ray technology to obtain a three-dimensional image of the ore, which can accurately reflect the internal structure of the ore compared to a two-dimensional image, and improves the accuracy of the characterization of the internal defects of the ore; using a trained ore internal defect characterization model, the key features of the internal defects of the ore can be extracted, and the internal defect grade can be automatically divided. At the same time, the ore internal defect characterization model is obtained by training the ore internal defect data set, realizing a closed-loop design from data to model; the introduction of the expert experience pool combines artificial experience with intelligent algorithms, so that the system can continuously accumulate and optimize sorting parameters, the construction of the expert experience sequence associates the ore defect characterization data with the ore dressing machine parameters, forming a complete workflow from perception to decision-making, and the multi-agent reinforcement learning model enables the system to continuously optimize the ore sorting parameters based on experience and real-time feedback, on the one hand, realizing accurate sorting based on defect grade, and on the other hand, dynamically adapting to different ore conditions, improving the sorting efficiency and accuracy of different internal defect grades of ores, thereby improving resource utilization and reducing energy consumption and cost in the ore dressing process. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart of a method for X-ray three-dimensional imaging and defect quantitative characterization of internal defects of ore in an X-ray intelligent ore dressing machine according to Example 1 of the present invention.

[0019] Figure 2 This is a step of classifying the internal defect levels of ore in Example 1 of the present invention.

[0020] Figure 3 This is a step of training a multi-agent reinforcement learning model in Example 1 of the present invention.

[0021] Figure 4 It is an X-ray three-dimensional imaging and defect quantitative characterization system for internal defects of ore in an X-ray intelligent ore dressing machine provided in Example 2 of the present invention.

[0022] Figure 5 This is a schematic diagram of the structure of an electronic device provided in Example 3 of the present invention.

[0023] Figure 6 It is a schematic diagram of the structure of a computer-readable storage medium provided in Example 4 of the present invention. DETAILED DESCRIPTION

[0024] In order to better understand the present application, a more detailed description will be made of various aspects of the present application with reference to the accompanying drawings. It should be understood that these detailed descriptions are only descriptions of exemplary embodiments of the present application, and are not intended to limit the scope of the present application in any way. Throughout the specification, the same reference numerals refer to the same elements. The expression "and / or" includes any and all combinations of one or more of the associated listed items.

[0025] In the accompanying drawings, the size, dimensions and shapes of the elements have been slightly adjusted for ease of illustration. The drawings are for illustration only and are not strictly drawn to scale. As used herein, the terms "substantially", "approximately" and similar terms are used as terms of approximation, not as terms of degree, and are intended to illustrate the inherent deviations in measurements or calculations that will be recognized by those of ordinary skill in the art. In addition, in this application, the order in which the steps are described does not necessarily represent the order in which these processes occur in actual operation, unless otherwise specified or can be derived from the context.

[0026] It should also be understood that expressions such as "include", "including", "have", "contain" and / or "comprising" are open rather than closed expressions in this specification, which indicate the presence of the stated features, elements and / or components, but do not exclude the presence of one or more other features, elements, components and / or combinations thereof. In addition, when expressions such as "at least one of..." appear after a list of listed features, they modify the entire list of features rather than just the individual elements in the list. In addition, when describing embodiments of the present application, "may" is used to mean "one or more embodiments of the present application". And, the term "exemplary" is intended to refer to an example or illustration.

[0027] Unless otherwise defined, all words (including engineering terms and scientific and technological terms) used in this article have the same meaning as those commonly understood by ordinary technicians in the field to which this application belongs. It should also be understood that unless there is a clear explanation in this application, words defined in commonly used dictionaries should be interpreted as having the same meaning as their meaning in the context of the relevant technology, and should not be interpreted in an idealized or overly formal sense.

[0028] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0029] Example 1 like Figure 1 to Figure 3 As shown, this embodiment provides an X-ray three-dimensional imaging and defect quantitative characterization method for internal defects of ore in an X-ray intelligent ore dressing machine; like Figure 1 As shown, the steps of the X-ray three-dimensional imaging and defect quantitative characterization method of ore internal defects in the X-ray intelligent ore dressing machine include: S1. Obtain a three-dimensional image of the ore using an X-ray binocular imaging device; perform feature processing on the three-dimensional image of the ore to obtain a data set of internal defects of the ore; In particular, the X-ray binocular imaging device includes an X-ray source and two flat panel detector arrays.

[0030] X-ray two-dimensional imaging can only obtain the plane projection information of the ore, which is difficult to fully reflect the internal structure and defects of the ore. At the same time, the two-dimensional projection is easily affected by the shape, rotation angle and complex internal characteristics of the ore, resulting in inaccurate detection results of internal defects in the ore. Using an X-ray source and two flat-panel detector arrays, it is possible to obtain two-dimensional projection images of the ore from different angles, and generate three-dimensional images of the ore through reconstruction algorithms, which can clearly express the internal volume, cracks, holes and other structural features of the ore, avoiding misjudgment caused by the loss of two-dimensional projection information.

[0031] In a preferred embodiment, the internal defects of the ore, such as cracks, pores, inclusions, etc., have three-dimensional characteristics, and X-ray three-dimensional imaging can express and extract the complete spatial distribution information of the internal defects of the ore, providing data support for the subsequent quantitative characterization of the internal defects of the ore. The X-ray three-dimensional imaging in the X-ray intelligent ore dressing machine uses, including but not limited to, a filtered back-projection algorithm and an iterative reconstruction algorithm to reconstruct the two-dimensional projection data into a three-dimensional ore image, which improves the accuracy and robustness of the internal defect detection of the ore when the internal defects of the ore are complex and unevenly distributed.

[0032] In a preferred embodiment, an X-ray binocular imaging device using one X-ray source and two flat-panel detector arrays collects ore projection data from different angles without adding additional ray sources. The hardware structure is simpler and suitable for different ore sizes and shapes. It has strong adaptability to the input direction and position of the ore. While reducing the hardware complexity and operating costs of the X-ray intelligent ore dressing machine, it can generate high-resolution three-dimensional ore images, and is a key component of the X-ray intelligent ore dressing machine.

[0033] Using an X-ray binocular imaging device to obtain a three-dimensional image of the ore is the starting point of the entire process, which determines the accuracy of subsequent internal defect characterization and intelligent sorting results of the ore. The three-dimensional image of the ore can fully describe the spatial distribution of internal defects in the ore, so that the internal defect characterization model of the ore can more accurately extract features and divide different internal defect levels, thereby achieving the effect of improving the accuracy of quantitative characterization of internal defects in the ore and the accuracy of intelligent sorting of different internal defect levels of the ore.

[0034] In particular, the step of performing feature processing on the three-dimensional image of the ore to obtain the ore internal defect data set is: Step S11: normalizing, denoising and enhancing voxel data of the three-dimensional image of the ore; Step S12: using an edge detection algorithm to locate the edges of internal defects of the ore and obtain edge data of the defects; Step S13: segmenting the defect region using a region growing algorithm to obtain geometric data of the defect; Step S14: marking the internal defects of the processed ore three-dimensional image to obtain an ore internal defect data set; The ore is a particle-sized ore obtained after the original ore is crushed and vibrated; the original ore is a concentrate containing only one type of ore after pre-selection and discarding.

[0035] S2, inputting the ore three-dimensional image into the ore internal defect characterization model to obtain ore internal defect characterization data, and classifying the ore internal defect levels; the ore internal defect characterization model is trained by the ore internal defect data set; In particular, see Figure 2 The steps of inputting the ore three-dimensional image into the ore internal defect characterization model to obtain the ore internal defect characterization data and classifying the ore internal defect levels are as follows: Step S21: dividing the ore internal defect data set into a training set and a validation set, inputting the data into a 3D CNN network model for training, and obtaining an ore internal defect characterization model; The ratio of the training set to the validation set is 7:3; the 3D-CNN consists of an input layer, a convolution layer, a pooling layer, a fully connected layer and an output layer, and the output layer outputs the internal defect characterization data of the ore; Step S22: inputting the ore three-dimensional image into the ore internal defect characterization model to obtain ore internal defect characterization data; The ore internal defect characterization data includes defect classification results and internal defect volume ratio, the defect classification results include two classification results: defective and non-defective, and the internal defect volume ratio refers to the ratio of the sum of the volumes of all internal defects in the ore to the volume of the ore; Step S23: classifying the internal defect levels of the ore according to the internal defect volume ratio; The method for classifying the internal defect levels of the ore is: Set a first threshold value X1 and a second threshold value X2; If the internal defect volume ratio is greater than and equal to the first threshold value X1, the internal defect level of the ore is level A; If the internal defect volume ratio is greater than the second threshold value X2 and less than the first threshold value X1, the internal defect level of the ore is level B; If the internal defect volume ratio is less than and equal to the second threshold value X2, the internal defect level of the ore is level C.

[0036] By dividing the ore internal defect data set into a training set and a validation set, and inputting them into the 3D-CNN network model for training, the ore internal defect characterization model is obtained, which can effectively learn and extract the feature representation of the ore internal defects, thereby achieving accurate characterization of the ore internal defects. The three-dimensional convolutional neural network 3D-CNN can effectively process the three-dimensional image data of the ore, capture the ore spatial information and defect structure characteristics in the image, adapt to ore data of different types and sizes, and incorporate the volume information of the ore internal defects into the model training and characterization process. Through multi-layer convolution and pooling operations, it learns and extracts high-level feature representations in the data, thereby improving the model's ability to identify and characterize ore internal defects.

[0037] According to the internal defect characterization model of the ore, the internal defect characterization data, including defect classification results and internal defect volume ratio data, are obtained according to the preset threshold value X 1 and X 2 The internal defects of the ore are classified into different grades. The classification method based on the volume ratio of internal defects can objectively evaluate the internal defects of the ore. Classifying and processing the ore according to different internal defect grades is conducive to adopting different sorting strategies for different grades of ore, improving sorting efficiency and accuracy. At the same time, the classification of internal defect grades also helps to monitor and control quality problems in the ore production process, providing quantitative internal defect characterization and grade evaluation for the overall solution, which is beneficial to improving the intelligence level and production management efficiency of the ore sorting system.

[0038] S3, obtaining the current setting parameters of the X-ray intelligent ore dressing machine, and constructing an expert experience sequence using the ore internal defect characterization data and the current setting parameters of the X-ray intelligent ore dressing machine; In particular, the current X-ray intelligent ore dressing machine setting parameters are obtained, and an expert experience sequence is constructed using the ore internal defect characterization data and the current X-ray intelligent ore dressing machine setting parameters, wherein: The current X-ray intelligent ore dressing machine setting parameters include feed belt speed, body belt speed, vibration amplitude, vibration frequency, vibration direction angle, and nozzle injection pressure; The expert experience sequence includes a set of ore internal defect states, a set of intelligent ore dressing machine parameter actions, a set of reward values, and a set of next states; The ore internal defect state set includes defect classification result data and internal defect volume ratio data; The intelligent ore dressing machine parameter action set includes feed belt speed data, body belt speed data, vibration amplitude data, vibration frequency data, vibration direction angle data, and nozzle injection pressure data.

[0039] In a preferred embodiment, the ore internal defect characterization data and the intelligent concentrator setting parameters are combined to construct an expert experience sequence, which realizes the comprehensive analysis and modeling of the ore internal defect state and the concentrator parameters. As an empirical model, the expert experience sequence integrates the ore internal defect data and the intelligent concentrator parameter data, and can provide data-driven decision support for the subsequent multi-agent reinforcement learning of the X-ray intelligent concentrator, strengthen the coupling relationship between ore quality and machine parameters, and improve the intelligence level and production management efficiency of the ore sorting system.

[0040] S4, storing the expert experience sequence in the expert experience pool, training the multi-agent reinforcement learning model to obtain the ore internal defect grade sorting parameter model; In particular, see Figure 3 The steps of storing the expert experience sequence into the expert experience pool and training the multi-agent reinforcement learning model to obtain the ore internal defect grade sorting parameter model are as follows: Step S31: construct an expert experience sequence using the ore internal defect characterization data and the current X-ray intelligent ore dressing machine setting parameters ; Said is the set of internal defect states of the ore, It is a set of intelligent ore dressing machine parameter actions. is the set of sorting accuracy reward values, is the set of internal defect states of the ore at the next moment; Step S32: Expert experience sequence Store into the expert experience pool to complete the initialization of the expert experience pool; Step S33: Using the expert experience pool, the multi-agent reinforcement learning model is trained to obtain the ore internal defect grade sorting parameter model.

[0041] Further, see Figure 3 The steps of storing the expert experience sequence into the expert experience pool and training the multi-agent reinforcement learning model to obtain the ore internal defect grade sorting parameter model are as follows: Step S41: constructing a prediction network and a target network; Step S42: Construct N agents based on the prediction network, target network and experience pool, where is a variable and ; Step S43: Initialize the experience pool as an expert experience pool; the experience pool has a capacity of M; Step S44: Use random prediction network weights Initialize the prediction network with random target network weights Initialize the target network and initialize the parameter state set , batch n and step size m; Step S45: Randomly obtain expert experience sequences from the experience pool in batches n Conduct training, including ; The prediction network inputs the internal defect state set of the ore And intelligent ore dressing machine parameter action set , output prediction value function And the next moment intelligent ore dressing machine parameter action set ; The target network inputs the set of internal defect states of the ore at the next moment , output target value function , The calculation formula is: ; In the formula, is the set of defect states inside the ore Intelligent ore dressing machine parameter action set The operation effect reward value, is the discount factor and , is the set of internal defect states of the ore at the next moment The next moment intelligent concentrator parameter action set is selected The maximum predicted value function obtained; Step S46: Every m steps, the weight parameters of the prediction network are Weight parameters copied to the target network ; Step S47: Calculate the joint prediction value function of multiple agents , the calculation formula is: ; Calculate the joint objective value function of multiple agents , the calculation formula is: ; In the formula, Function and g function are calculated by BP neural network model; The multi-agent loss function is updated as , the calculation formula is: ; Step S48: Minimize the MSE loss function using gradient descent , update the network model parameters and Until convergence, the ore internal defect grade sorting parameter model is obtained; Furthermore, the prediction network and the target network have the same network structure, both using BP neural network as the basic network structure; The BP neural network consists of an input layer, a hidden layer and an output layer; The number of neurons in the input layer is equal to the number of defect state sets inside the ore; The number of neurons in the output layer is equal to the number of actions in the intelligent ore dressing machine parameter action set.

[0042] By storing the expert experience sequence in the expert experience pool and using the expert experience pool for model training, the expert experience pool is used as the initial transfer learning weight of the multi-agent reinforcement learning model to improve the training efficiency of the model. The expert experience sequence contains the initial information on the internal defect state of the ore and the parameter action of the intelligent ore dressing machine, which helps the model to better understand the relationship between the internal defect state of the ore and the parameters of the intelligent ore dressing machine, and improves the generalization ability of the model. The multi-agent reinforcement learning model is used to obtain the ore internal defect grade sorting parameter model, and the intelligent ore dressing machine parameters for different ore internal defect states are optimized. Through the training model, the optimal parameter setting that is more suitable for the current ore internal defect state can be obtained, which improves the efficiency and accuracy of ore sorting and realizes the optimization of the parameters of the X-ray intelligent ore dressing machine, thereby providing a beneficial contribution to the intelligent sorting system of different internal defects of the ore.

[0043] S5. The ore to be screened is conveyed to an X-ray intelligent ore concentrator, and ores with different internal defect grades are sorted based on an ore internal defect characterization model and an ore internal defect grade sorting parameter optimization model.

[0044] In a preferred embodiment, the X-ray intelligent ore dressing machine is started, and the ore to be screened is sent to the X-ray binocular imaging device area through a conveyor belt, and the X-ray three-dimensional imaging technology is used to perform high-precision scanning of the inside of the ore. The imaging data is transmitted to the data processing unit in real time, and the ore internal defect characterization model quickly analyzes the data to identify the type and grade of internal defects in the ore. The ore internal defect grade sorting parameter optimization model dynamically adjusts the sorting parameters of the ore dressing machine according to the analysis results. Finally, ores of different defect grades are collected into designated containers respectively to complete the entire sorting process. Through this series of steps, while realizing accurate identification and efficient sorting of internal defects in the ore, the parameters of the X-ray intelligent ore dressing machine for processing different internal defects of the ore are dynamically and adaptively adjusted, and the parameter setting custom update is completed, which improves the sorting efficiency and accuracy of different internal defect grades of the ore, thereby improving resource utilization and reducing energy consumption and cost in the ore dressing process.

[0045] Example 2 Figure 4 An embodiment of the present application provides an X-ray three-dimensional imaging and defect quantitative characterization system for internal defects of ore in an X-ray intelligent ore dressing machine. The system as a whole can be divided into: Ore image processing module: used to obtain ore three-dimensional images using X-ray binocular imaging devices; perform feature processing on the ore three-dimensional images to obtain ore internal defect data sets; Ore defect characterization module: used to input the ore three-dimensional image into the ore internal defect characterization model, obtain the ore internal defect characterization data, and classify the ore internal defect level; the ore internal defect characterization model is trained by the ore internal defect data set; Mining machine parameter acquisition module: used to obtain the current X-ray intelligent ore dressing machine setting parameters, and construct an expert experience sequence using the ore internal defect characterization data and the current X-ray intelligent ore dressing machine setting parameters; Mining machine parameter optimization module: used to store expert experience sequences into the expert experience pool, train the multi-agent reinforcement learning model to obtain the ore internal defect grade sorting parameter model; Intelligent ore sorting module: It is used to transfer the ore to be screened to the X-ray intelligent ore concentrator, and sort ores with different internal defect grades based on the ore internal defect characterization model and the ore internal defect grade sorting parameter optimization model.

[0046] Example 3 Figure 5 1 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Figure 5As shown, according to another aspect of the present application, an electronic device 100 is also provided. The electronic device 100 may include one or more processors and one or more memories. The memories store computer readable codes, and when the computer readable codes are executed by one or more processors, the X-ray three-dimensional imaging and defect quantitative characterization method of ore internal defects in the X-ray intelligent ore dressing machine as described above can be executed.

[0047] The method or device according to the embodiment of the present application can also be used by Figure 5 The electronic device architecture shown in FIG. Figure 5 As shown, the electronic device 100 may include a bus 101, one or more CPUs 102, a read-only memory (ROM) 103, a random access memory (RAM) 104, a communication port 105 connected to a network, an input / output component 106, a hard disk 107, etc. The storage device in the electronic device 100, such as the ROM 103 or the hard disk 107, may store the X-ray three-dimensional imaging and defect quantitative characterization method of ore internal defects in the X-ray intelligent ore dressing machine provided in the present application. The method for X-ray three-dimensional imaging and defect quantitative characterization of internal defects of ore in an X-ray intelligent ore dressing machine may, for example, include the following steps: using an X-ray binocular imaging device to obtain a three-dimensional image of the ore; performing feature processing on the three-dimensional image of the ore to obtain an internal defect data set of the ore; inputting the three-dimensional image of the ore into an internal defect characterization model of the ore to obtain internal defect characterization data of the ore, and classifying the internal defects of the ore into grades; the internal defect characterization model of the ore is trained by the internal defect data set of the ore; obtaining the current setting parameters of the X-ray intelligent ore dressing machine, and constructing an expert experience sequence using the internal defect characterization data of the ore and the current setting parameters of the X-ray intelligent ore dressing machine; storing the expert experience sequence in an expert experience pool, and training a multi-agent reinforcement learning model to obtain an ore internal defect grade sorting parameter model; transmitting the ore to be screened to the X-ray intelligent ore dressing machine, and sorting ores with different internal defect grades of the ore based on the internal defect characterization model of the ore and the internal defect grade sorting parameter optimization model of the ore.

[0048] Furthermore, the electronic device 100 may further include a user interface 108. Figure 5 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different devices. Figure 5 One or more components of an electronic device are shown.

[0049] Example 4 Figure 6 Schematic diagram of a computer-readable storage medium structure provided by an embodiment of the present application. Figure 6As shown, it is a computer-readable storage medium 200 according to an embodiment of the present application. Computer-readable instructions are stored on the computer-readable storage medium 200. When the computer-readable instructions are executed by the processor, the X-ray three-dimensional imaging and defect quantitative characterization method of internal defects of ore in the X-ray intelligent ore dressing machine according to the embodiment of the present application described with reference to the above figures can be executed. The computer-readable storage medium 200 includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0050] In addition, according to the implementation of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, the non-transitory machine-readable storage medium stores machine-readable instructions, the machine-readable instructions can be run by a processor to execute instructions corresponding to the method steps provided by the present application, and when the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0051] The methods, apparatuses, and devices of the present application may be implemented in many ways. For example, the methods, apparatuses, and devices of the present application may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above order of steps for the method is for illustration only, and the steps of the method of the present application are not limited to the order specifically described above unless otherwise specified. In addition, in some embodiments, the present application may also be implemented as programs recorded in a recording medium, which include machine-readable instructions for implementing the method according to the present application. Therefore, the present application also covers recording media storing programs for executing the method according to the present application.

[0052] In addition, the parts of the above-mentioned technical solutions provided in the embodiments of the present application that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive redundancy.

[0053] The specific implementation modes as described above further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation mode of the present invention and is not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0054] The above preset parameters or preset thresholds are all set by technicians in this field according to actual conditions or obtained through large-scale data simulation.

[0055] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. X-ray three-dimensional imaging and defect quantitative characterization method of ore internal defects in X-ray intelligent ore dressing machine, characterized in that: The following steps are involved: S1. Obtain a three-dimensional image of the ore using an X-ray binocular imaging device; Perform feature processing on the ore three-dimensional image to obtain the ore internal defect data set; S2, inputting the ore three-dimensional image into the ore internal defect characterization model to obtain ore internal defect characterization data, and classifying the ore internal defect levels; the ore internal defect characterization model is trained by the ore internal defect data set; S3, obtaining the current setting parameters of the X-ray intelligent ore dressing machine, and constructing an expert experience sequence using the ore internal defect characterization data and the current setting parameters of the X-ray intelligent ore dressing machine; S4, storing the expert experience sequence in the expert experience pool, training the multi-agent reinforcement learning model to obtain the ore internal defect grade sorting parameter model; S5. The ore to be screened is conveyed to an X-ray intelligent ore concentrator, and ores with different internal defect grades are sorted based on an ore internal defect characterization model and an ore internal defect grade sorting parameter optimization model.

2. The X-ray three-dimensional imaging and defect quantitative characterization method of ore internal defects in an X-ray intelligent ore dressing machine according to claim 1 is characterized in that: The X-ray binocular imaging device includes an X-ray source and two flat panel detector arrays.

3. The X-ray three-dimensional imaging and defect quantitative characterization method of ore internal defects in an X-ray intelligent ore dressing machine according to claim 2 is characterized in that: The step of performing feature processing on the three-dimensional image of the ore to obtain the ore internal defect data set is: Step S11: normalizing, denoising and enhancing voxel data of the three-dimensional image of the ore; Step S12: using an edge detection algorithm to locate the edges of internal defects of the ore and obtain edge data of the defects; Step S13: segmenting the defect region using a region growing algorithm to obtain geometric data of the defect; Step S14: marking the internal defects of the processed ore three-dimensional image to obtain an ore internal defect data set; The ore is a particle-sized ore obtained after the original ore is crushed and vibrated; the original ore is a concentrate containing only one type of ore after pre-selection and discarding.

4. The X-ray three-dimensional imaging and defect quantitative characterization method of ore internal defects in an X-ray intelligent ore dressing machine according to claim 3 is characterized in that: The steps of inputting the ore three-dimensional image into the ore internal defect characterization model to obtain the ore internal defect characterization data and classifying the ore internal defect levels are as follows: Step S21: dividing the ore internal defect data set into a training set and a validation set, inputting the data into a 3D CNN network model for training, and obtaining an ore internal defect characterization model; The ratio of the training set to the validation set is 7:3; the 3D-CNN consists of an input layer, a convolution layer, a pooling layer, a fully connected layer and an output layer, and the output layer outputs the internal defect characterization data of the ore; Step S22: inputting the ore three-dimensional image into the ore internal defect characterization model to obtain ore internal defect characterization data; The ore internal defect characterization data includes defect classification results and internal defect volume ratio, the defect classification results include two classification results: defective and non-defective, and the internal defect volume ratio refers to the ratio of the sum of the volumes of all internal defects in the ore to the volume of the ore; Step S23: classifying the internal defect levels of the ore according to the internal defect volume ratio; The method for classifying the internal defect levels of the ore is as follows: Set a first threshold value X1 and a second threshold value X2; If the internal defect volume ratio is greater than and equal to the first threshold value X1, the internal defect level of the ore is level A; If the internal defect volume ratio is greater than the second threshold value X2 and less than the first threshold value X1, the internal defect level of the ore is level B; If the internal defect volume ratio is less than and equal to the second threshold value X2, the internal defect level of the ore is level C.

5. The X-ray three-dimensional imaging and defect quantitative characterization method of ore internal defects in an X-ray intelligent ore dressing machine according to claim 4 is characterized in that: The current X-ray intelligent ore dressing machine setting parameters are obtained, and an expert experience sequence is constructed using the ore internal defect characterization data and the current X-ray intelligent ore dressing machine setting parameters, wherein: The current X-ray intelligent ore dressing machine setting parameters include feed belt speed, body belt speed, vibration amplitude, vibration frequency, vibration direction angle, and nozzle injection pressure; The expert experience sequence includes a set of ore internal defect states, a set of intelligent ore dressing machine parameter actions, a set of reward values, and a set of next states; The ore internal defect state set includes defect classification result data and internal defect volume ratio data; The intelligent ore dressing machine parameter action set includes feed belt speed data, body belt speed data, vibration amplitude data, vibration frequency data, vibration direction angle data, and nozzle injection pressure data.

6. The X-ray three-dimensional imaging and defect quantitative characterization method of ore internal defects in an X-ray intelligent ore dressing machine according to claim 5, characterized in that: The steps of storing the expert experience sequence in the expert experience pool and training the multi-agent reinforcement learning model to obtain the ore internal defect grade sorting parameter model are as follows: Step S31: construct an expert experience sequence using the ore internal defect characterization data and the current X-ray intelligent ore dressing machine setting parameters ; Said is the set of internal defect states of the ore, It is a set of intelligent ore dressing machine parameter actions. is the set of sorting accuracy reward values, is the set of internal defect states of the ore at the next moment; Step S32: Expert experience sequence Store into the expert experience pool to complete the initialization of the expert experience pool; Step S33: Using the expert experience pool, the multi-agent reinforcement learning model is trained to obtain the ore internal defect grade sorting parameter model.

7. The X-ray three-dimensional imaging and defect quantitative characterization method of ore internal defects in an X-ray intelligent ore dressing machine according to claim 6 is characterized in that: The steps of storing the expert experience sequence in the expert experience pool and training the multi-agent reinforcement learning model to obtain the ore internal defect grade sorting parameter model are as follows: Step S41: constructing a prediction network and a target network; Step S42: Construct N agents based on the prediction network, target network and experience pool, where is a variable and ; Step S43: Initialize the experience pool as an expert experience pool; the experience pool has a capacity of M; Step S44: Use random prediction network weights Initialize the prediction network with random target network weights Initialize the target network and initialize the parameter state set , batch n and step size m; Step S45: Randomly obtain expert experience sequences from the experience pool in batches n Conduct training, including ; The prediction network inputs the internal defect state set of the ore And intelligent ore dressing machine parameter action set , output prediction value function And the next moment intelligent ore dressing machine parameter action set ; The target network inputs the set of internal defect states of the ore at the next moment , output target value function , The calculation formula is: ； In the formula, is the set of defect states inside the ore Intelligent ore dressing machine parameter action set The operation effect reward value, is the discount factor and , is the set of internal defect states of the ore at the next moment The next moment intelligent concentrator parameter action set is selected The maximum predicted value function obtained; Step S46: Every m steps, the weight parameters of the prediction network are Weight parameters copied to the target network ; Step S47: Calculate the joint prediction value function of multiple agents , the calculation formula is: ； Calculate the joint objective value function of multiple agents , the calculation formula is: ； In the formula, Function and g function are calculated by BP neural network model; The multi-agent loss function is updated as , the calculation formula is: ； Step S48: Minimize the MSE loss function using gradient descent , update the network model parameters and Until convergence, the ore internal defect grade sorting parameter model is obtained; The prediction network and the target network have the same network structure, both using BP neural network as the basic network structure; the BP neural network consists of an input layer, a hidden layer and an output layer; The number of neurons in the input layer is equal to the number of defect state sets inside the ore; The number of neurons in the output layer is equal to the number of actions in the intelligent ore dressing machine parameter action set.

8. An X-ray three-dimensional imaging and defect quantitative characterization system for internal defects of ore in an X-ray intelligent ore dressing machine, which is based on the X-ray three-dimensional imaging and defect quantitative characterization method for internal defects of ore in an X-ray intelligent ore dressing machine according to any one of claims 1 to 7, characterized in that: The system comprises: Ore image processing module: used to obtain ore three-dimensional images using X-ray binocular imaging devices; perform feature processing on the ore three-dimensional images to obtain ore internal defect data sets; Ore defect characterization module: used to input the ore three-dimensional image into the ore internal defect characterization model, obtain the ore internal defect characterization data, and classify the ore internal defect level; the ore internal defect characterization model is trained by the ore internal defect data set; Mining machine parameter acquisition module: used to obtain the current X-ray intelligent ore dressing machine setting parameters, and construct an expert experience sequence using the ore internal defect characterization data and the current X-ray intelligent ore dressing machine setting parameters; Mining machine parameter optimization module: used to store expert experience sequences into the expert experience pool, train the multi-agent reinforcement learning model to obtain the ore internal defect grade sorting parameter model; Intelligent ore sorting module: It is used to transfer the ore to be screened to the X-ray intelligent ore concentrator, and sort ores with different internal defect grades based on the ore internal defect characterization model and the ore internal defect grade sorting parameter optimization model.

9. An electronic device, characterized in that: It comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the method for X-ray three-dimensional imaging and defect quantitative characterization of internal defects of ore in an X-ray intelligent ore dressing machine as described in any one of claims 1 to 7.

10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which is suitable for loading by a processor to execute the steps in the method for X-ray three-dimensional imaging and defect quantitative characterization of internal defects of ore in an X-ray intelligent ore dressing machine as described in any one of claims 1-7.

Images