Coal flotation method, coal flotation device and coal flotation system based on images
Through image recognition technology, the foam composition during coal flotation is monitored and adjusted, and the problem of untimely adjustment of the agent caused by manual observation is solved, and the flotation effect is improved.
Patent Information
- Application Number
- CN202411243993.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-05
- Publication Date
- 2024-11-15
AI Technical Summary
In traditional coal flotation technology, manually observing foam monitoring of flotation production status results in untimely adjustment of the agent, which affects the flotation effect.
The image-based coal flotation method is adopted, by adding flotation agent to the ore slurry, forming foam, and monitoring the surface image of the ore slurry in real time, classifying and adjusting the agent components based on image recognition technology to achieve precise control.
It improves the accuracy and standard of foam component identification, realizes the timeliness of agent adjustment, and improves the flotation effect.
Smart Images

Figure CN118950267A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of coal flotation, and in particular to an image-based coal flotation method, a coal flotation device, a computer-readable storage medium and a coal flotation system. Background Art
[0002] At present, coal mines generally use flotation technology to achieve ore separation. Coal flotation is a commonly used coal mineral processing technology, which aims to separate coal and impurities such as rocks from raw coal through physical and chemical methods. It is one of the important coal purification methods, which can be used to improve the quality of coal, reduce the ash and sulfur content, and meet the coal quality requirements of industries such as coal combustion, metallurgy and chemical industry.
[0003] In the coal flotation production process, when the feed properties, process parameters and other factors change, the working conditions are easily affected and fluctuate. If they cannot be handled in a timely and effective manner, the flotation process will become abnormal, which will eventually lead to a decrease in the recovery rate of mineral resources. Traditional industries have always relied on workers to observe the foam with their naked eyes to monitor the flotation production status. This method is labor-intensive and lacks objective standards.
[0004] In summary, the traditional manual adjustment method cannot adjust the control variables in a timely and accurate manner, which often leads to large fluctuations in process indicators, high consumption of reagents, and high equipment failure rate. It can only be achieved through the optimization design of the automatic control system. Summary of the invention
[0005] The main purpose of the present application is to provide an image-based coal flotation method, a coal flotation device, a computer-readable storage medium, a processor and a coal flotation system, so as to at least solve the problem in the prior art of manually observing foam to monitor the flotation production status, resulting in untimely adjustment of reagents and affecting the flotation effect.
[0006] In order to achieve the above-mentioned purpose, according to one aspect of the present application, there is provided an image-based coal flotation method, comprising: an adding step of adding a flotation agent to a slurry to separate minerals from impurities in the slurry; a stirring step of stirring the slurry until foam is formed on the surface of the slurry, and real-time monitoring of the slurry surface image to obtain a first target image; an identification step of classifying the foam based on the first target image to obtain a target foam, discharging the target foam and adjusting the proportion of the components in the flotation agent based on the amount of the target foam to eliminate and reduce the target foam, wherein the target foam is the foam containing impurities; and repeating the adding step, the stirring step and the identification step in sequence at least once until the amount of the target foam is less than or equal to a first threshold value.
[0007] Optionally, classifying the foam based on the first target image to obtain target foam includes: preprocessing the first target image, converting the first target image into a vector representation, and obtaining a first target vector; and inputting the first target vector into an image recognition model for classification to obtain the target foam.
[0008] Optionally, the first target image is preprocessed and converted into a vector representation to obtain a first target vector, including: extracting eigenvalues of the first target image to obtain multiple target eigenvalues, wherein the multiple target eigenvalues are used to characterize different features of the first target image; representing the target eigenvalues in vector form to obtain multiple second target vectors; and denoising and optimizing the second target vector to obtain the first target vector.
[0009] Optionally, feature value extraction is performed on the first target image to obtain multiple target feature values, including: segmenting the first target image by a watershed algorithm to screen out the foam area in the first target image to obtain a second target image; determining the corresponding foam quantity based on each of the second target images to obtain a first target quantity; summing the pixel point areas of each of the second target images to obtain a first target area; summing the second target quantities to obtain a second target quantity, calculating the ratio of the first target area to the second target quantity to obtain a second target area; and determining the second target area as the first target feature value.
[0010] Optionally, performing eigenvalue extraction on the first target image to obtain multiple target eigenvalues also includes: calculating the hue value based on the first component, the second component and the third component of each pixel of the first target image to obtain the second target eigenvalue, the first component being the red component, the second component being the green component, and the third component being the blue component; calculating the mean based on each of the first component, the second component and the third component to obtain the corresponding third target eigenvalue; calculating the grayscale mean based on the third target eigenvalue of each pixel of the first target image to obtain the fourth target eigenvalue; calculating the ratio based on the third target eigenvalue and the fourth target eigenvalue to obtain the fifth target eigenvalue; converting the first target image into a grayscale image to obtain a third target image, converting the third target image into a neighborhood grayscale correlation matrix and calculating the coarseness to obtain the sixth target eigenvalue.
[0011] Optionally, the second target vector is denoised, including: performing fuzzy C-means clustering on the second target vector with the minimum value of the objective function as a termination condition to obtain a target cluster center and a target cluster cluster; based on the target cluster center and the target cluster cluster, the second target vector corresponding to the outlier is determined as a third target vector; and the third target vector is deleted to denoise the second target vector.
[0012] Optionally, optimizing the second target vector to obtain the first target vector includes: adding a target identifier to each of the second target vectors to obtain a target data set, wherein the target identifier is used to characterize whether each of the second target vectors is input into the image recognition model; obtaining a third target number based on the number of second target vectors corresponding to the target identifier in the target data set that are input into the image recognition model; determining, based on the target data set, an accuracy rate of the target data set input into the image recognition model to obtain a target ratio; constructing an optimization model based on the third target number, the target data set, and the target ratio, and solving the optimization model to obtain the first target vector: Among them, A cc (s) is the target ratio, s is the target data set, ||s||0 is the third target quantity, and n is the type of the second target vector.
[0013] According to another aspect of the present application, an image-based coal flotation device is provided, the device comprising: a first control unit, for executing an adding step, adding a flotation agent to a slurry to separate minerals from impurities in the slurry; a second control unit, for executing a stirring step, stirring the slurry until foam is formed on the surface of the slurry, and monitoring the slurry surface image in real time to obtain a first target image; a third control unit, for executing an identification step, classifying the foam based on the first target image to obtain a target foam, discharging the target foam and adjusting the proportion of the components in the flotation agent based on the amount of the target foam to eliminate and reduce the target foam, wherein the target foam is the foam containing impurities; and a repeating unit, for sequentially repeating the adding step, the stirring step and the identification step at least once until the amount of the target foam is less than or equal to a first threshold value.
[0014] According to another aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium includes a stored program, wherein when the program is executed, the method described in any one of the devices where the computer-readable storage medium is located is controlled.
[0015] According to another aspect of the present application, a coal flotation system is provided, comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the one or more programs include methods for executing any one of the described methods.
[0016] The technical solution of the present application is applied, in the above-mentioned image-based coal flotation method, including: adding step, adding flotation reagent to the pulp to separate the minerals and impurities in the pulp; stirring step, stirring the pulp until foam is formed on the surface of the pulp, real-time monitoring the surface image of the pulp, and obtaining a first target image; identification step, classifying the foam based on the first target image to obtain the target foam, and adjusting the proportion of the components in the flotation reagent based on the number of the target foam to eliminate and reduce the target foam, the target foam is the foam containing impurities; repeating the adding step, the stirring step and the identification step at least once in sequence until the number of the target foam is less than or equal to the first threshold. The present application identifies the foam components through images to determine the components of the foam, and adjusts the flotation reagent to separate impurities according to the distinguished foam components. Compared with manual observation of foam components, this method improves the accuracy and standard of identification and realizes precise control. This method solves the problem in the prior art that the flotation production status is monitored by manual observation of foam, resulting in untimely adjustment of reagents and affecting the flotation effect. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 A hardware structure block diagram of a mobile terminal for an image-based coal flotation method provided in an embodiment of the present application is shown;
[0018] Figure 2 A schematic flow chart of an image-based coal flotation method provided according to an embodiment of the present application is shown;
[0019] Figure 3 A structural block diagram of an image-based coal flotation device provided according to an embodiment of the present application is shown.
[0020] The above drawings include the following reference numerals:
[0021] 102, processor; 104, memory; 106, transmission device; 108, input and output devices. DETAILED DESCRIPTION
[0022] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of 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.
[0023] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.
[0024] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0025] As introduced in the background technology, the prior art relies on workers to observe foam with the naked eye to monitor the flotation production status, which is inefficient and lacks a unified standard. In order to solve the problem that the prior art uses manual observation of foam to monitor the flotation production status, resulting in untimely adjustment of reagents and affecting the flotation effect, the embodiments of the present application provide an image-based coal flotation method, a coal flotation device, a computer-readable storage medium, a processor and a coal flotation system.
[0026] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.
[0027] The method embodiments provided in the embodiments of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a mobile terminal as an example, Figure 1 1 is a hardware structure block diagram of a mobile terminal for an image-based coal flotation method according to an embodiment of the present invention. Figure 1 As shown, the mobile terminal may include one or more ( Figure 1 Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the mobile terminal may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the mobile terminal. Figure 1More or fewer components as shown, or with Figure 1 Different configurations are shown.
[0028] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the display method of device information in the embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, the above method is implemented. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the above-mentioned network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof. The transmission device 106 is used to receive or send data via a network. The above-mentioned specific examples of the network may include a wireless network provided by a communication provider of the mobile terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.
[0029] In this embodiment, an image-based coal flotation method running on a mobile terminal, a computer terminal or a similar computing device is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0030] Figure 2 FIG. 1 is a flow chart of an image-based coal flotation method according to an embodiment of the present application. Figure 2 As shown, the method comprises the following steps:
[0031] Step S201, adding a flotation agent to the slurry to separate the minerals from the impurities in the slurry;
[0032] Specifically, flotation reagents are added to the ore pulp to separate the minerals and impurities in the ore pulp. It can be understood that the flotation process is a process of separating minerals and gangue according to the difference in hydrophilic and hydrophobic properties of the minerals. The role of adding flotation reagents is to enhance the hydrophilicity of useful minerals and improve the hydrophobicity of gangue impurities.
[0033] In the specific implementation, the flotation reagents mainly include collecting reagents, inhibiting reagents and adjusting reagents. According to the actual production situation, the proportion of the three reagents and the amount of reagents to be prepared are set, and the addition speed of various raw materials is adjusted by controlling the start and stop of the delivery pump and the opening of the regulating valve, and the addition amount is monitored by detection instruments such as flow meters and liquid level meters. In this way, the addition of flotation reagents is realized.
[0034] Step S202, a stirring step, stirring the slurry until foam is formed on the surface of the slurry, and monitoring the slurry surface image in real time to obtain a first target image;
[0035] Specifically, the slurry is stirred to form foam on the upper layer of the slurry. It is understood that after adding the flotation reagent, air is introduced into the slurry in the flotation tank to form a large number of bubbles. Under the stirring of the stirring paddle, hydrophobic impurities such as gangue impurities adhere to the surface of the bubbles and float out with the bubbles into the foam layer, while useful minerals such as hydrophilic impurities remain in the slurry as underflow.
[0036] Step S203, a recognition step, classifying the foam based on the first target image to obtain target foam, discharging the target foam and adjusting the proportion of the components in the flotation reagent based on the amount of the target foam to eliminate and reduce the target foam, wherein the target foam is the foam containing impurities;
[0037] Specifically, the foams are classified through images to determine common foams and target foams with impurities, the target foams are discharged, and the flotation reagent components for the next round of flotation are adjusted based on the number of target foams.
[0038] Step S204, repeating the adding step, the stirring step and the identifying step at least once in sequence until the amount of the target foam is less than or equal to a first threshold.
[0039] Specifically, the above-mentioned adding step, the above-mentioned stirring step and the above-mentioned identification step are repeated multiple times, that is, the roughing, concentrating and scavenging stages are performed multiple times. After the raw material slurry is regrinded and concentrated, the slurry that meets certain particle size and concentration requirements is pumped to the flotation process, and the first sorting is carried out in the roughing stage. The bottom flow enriched with useful minerals enters the concentrating stage for further sorting, and the roughing foam is sent to the scavenging process. Then, the roughing bottom flow slurry passes through the concentrating stage to obtain a qualified product as the final concentrate, and the concentrating foam returns to the roughing stage. The roughing foam reduces the useful minerals in the final tailings through the scavenging stage, and the scavenging foam is discharged into the foam collector as the final tailings. The scavenging bottom flow and the concentrating foam are sent to the roughing stage together for another sorting.
[0040] In the above embodiment, before adding the flotation reagent to the slurry, the above method further includes:
[0041] The minerals in the slurry are ground until the particle size of the minerals is less than or equal to a preset particle size.
[0042] Specifically, the minerals in the slurry are ground so that the particle size of the minerals meets the preset conditions.
[0043] It is understandable that flotation requires not only the minerals to be fully dissociated into monomers, but also the appropriate particle size for selection. If the mineral particles are too coarse, even if the minerals have been dissociated into monomers, they often cannot float because they exceed the buoyancy capacity of the bubbles. If the grinding particles are too fine, it is also not conducive to coal flotation.
[0044] In the specific implementation, the closed-circuit grinding consisting of a ball mill and a cyclone is used to regrind the concentrate from the magnetic separation process, so that the particle size is increased from 89% of -200 mesh to 94% of -200 mesh, so as to meet the particle size requirements of the flotation process and thereby improve the flotation efficiency.
[0045] After grinding the minerals in the slurry until the particle size of the minerals is less than or equal to a preset particle size, the method further includes:
[0046] The concentration and flow rate of the slurry are adjusted by a thickener until the concentration reaches a preset concentration and the flow rate reaches a preset flow rate.
[0047] Specifically, a thickener is used to adjust the concentration and flow rate of the slurry, i.e., the concentration step.
[0048] It is understandable that the concentration process is an important part of the coal flotation process. The slurry transported from the magnetic separation process is concentrated and separated by a thickener to ensure that the slurry with appropriate concentration and flow rate enters the flotation process.
[0049] In specific implementation, appropriate pulp concentration is conducive to obtaining higher metal recovery rate and concentrate grade in the flotation process. The pulp flow rate determines the flotation time, which in turn affects the metal recovery rate and concentrate grade in the flotation process.
[0050] According to this embodiment, in the step of adding, flotation reagent is added to the pulp to separate the minerals from the impurities in the pulp; in the step of stirring, the pulp is stirred until foam is formed on the surface of the pulp, and the surface image of the pulp is monitored in real time to obtain a first target image; in the step of identifying, the foam is classified based on the first target image to obtain the target foam, the target foam is discharged and the proportion of the components in the flotation reagent is adjusted based on the amount of the target foam to eliminate and reduce the target foam, and the target foam is the foam containing impurities; the step of adding, the step of stirring and the step of identifying are repeated at least once in sequence until the amount of the target foam is less than or equal to the first threshold. The present application identifies the foam components through images to determine the components of the foam, and adjusts the flotation reagent to separate impurities according to the distinguished foam components. Compared with manual observation of the foam components, this method improves the accuracy and standard of identification and realizes precise control. This method solves the problem in the prior art of monitoring the flotation production status by manual observation of foam, which leads to untimely adjustment of reagents and affects the flotation effect.
[0051] In order to obtain the target foam, in an optional embodiment, the above step S203 includes:
[0052] Step S2031, preprocessing the first target image, converting the first target image into a vector representation, and obtaining a first target vector;
[0053] Specifically, the first target image is preprocessed and converted into a vector representation to obtain the first target vector. The preprocessing at least includes format conversion, denoising and optimization.
[0054] Step S2032: input the first target vector into an image recognition model for classification to obtain the target foam.
[0055] Specifically, feature extraction is performed based on the first target vector, and classification and identification are performed through a model based on the features to determine the foam containing impurities and obtain the target foam.
[0056] In order to obtain the first target vector, in an optional implementation, the above step S2031 includes:
[0057] Step S20311, extracting feature values from the first target image to obtain a plurality of target feature values, wherein the plurality of target feature values are used to characterize different features of the first target image;
[0058] Specifically, since the image is directly used as the input of the classifier, the dimension is too high, which leads to a long training time and may cause the model to fall into a local optimal point. Based on the above considerations, this application selects multiple features such as foam average area, load rate and color tone characteristics, and converts the multiple features into feature value representation to obtain the above target feature value.
[0059] Step S20312, expressing the target characteristic value in vector form to obtain a plurality of second target vectors;
[0060] Specifically, the target eigenvalue is converted into a vector representation to obtain the second target vector.
[0061] Step S20313, denoising and optimizing the second target vector to obtain the first target vector.
[0062] Specifically, denoising and optimization operations are performed in sequence based on the second target vector to obtain the first target vector.
[0063] In order to extract the average foam area, in an optional implementation, the above step S20311 includes:
[0064] Step S2031101, segmenting the first target image by a watershed algorithm, screening out the foam area in the first target image, and obtaining a second target image;
[0065] Specifically, the foam area in the image is segmented by a watershed algorithm to obtain the above-mentioned second target image.
[0066] Step S2031102, determining the corresponding foam quantity based on each of the second target images to obtain the first target quantity;
[0067] Specifically, each of the second target images is traversed to determine the number of bubbles contained therein, thereby obtaining the first target number.
[0068] Step S2031103, summing up the pixel areas of each of the second target images to obtain the first target area;
[0069] Specifically, the pixel areas of each of the second target images are summed to obtain the first target area Sp.
[0070] Step S2031104, summing up the second target quantities to obtain the second target quantity, and calculating the ratio of the first target area to the second target quantity to obtain the second target area;
[0071] Specifically, the number of bubbles in each region is added to get N b , the above second target quantity is obtained, based on the average foam area Smean Calculation formula S mean =S p / N b The second target area is calculated.
[0072] Step S2031105, determining the second target area as the first target characteristic value.
[0073] In order to calculate multiple target feature values, in an optional implementation, the above step S20311 further includes:
[0074] Step S2031106, calculating the hue value based on the first component, the second component and the third component of each pixel of the first target image to obtain a second target feature value, where the first component is a red component, the second component is a green component, and the third component is a blue component;
[0075] Specifically, when the distance between the foam and the light source is different, it is difficult to describe the color of the foam surface under the same light intensity standard, that is, there is a deviation in feature extraction based solely on the color in the image RGB space. Therefore, this application is set to implement feature extraction in the HSV space.
[0076] Furthermore, by calculating the image hue value, the image is converted from the RGB space to the HSV space to extract the hue features that are not affected by light intensity. The calculation method is:
[0077]
[0078] Among them, Hi is the hue value of the i-th pixel, that is, the second eigenvalue mentioned above, Ri is the first component mentioned above, Gi is the second component mentioned above, and Bi is the third component mentioned above.
[0079] Step S2031107, calculating the mean value based on each of the first component, the second component and the third component, to obtain a corresponding third target eigenvalue;
[0080] Specifically, the calculation formula of the third target eigenvalue of each of the first component, the second component and the third component is:
[0081]
[0082] Among them, Np is the number of pixels occupied by foam, R mean is the third target eigenvalue of the first component, i.e., the red mean. It can be understood that the third eigenvalue (green mean) corresponding to the second component G mean The third eigenvalue (blue mean) B corresponding to the third component mean The calculation method of the third target eigenvalue is the same as that of , and the corresponding components can be substituted.
[0083] Step S2031108, calculating the grayscale mean based on the third target feature value of each pixel point of the first target image to obtain a fourth target feature value;
[0084] Specifically, the grayscale mean of an image reflects the brightness of the image, and its calculation formula is:
[0085] Y mean =0.299R mean +0.587G mean +0.114B mean ;
[0086] Among them, Y mean is the fourth target eigenvalue mentioned above.
[0087] Step S2031109, calculating a ratio based on the third target eigenvalue and the fourth target eigenvalue to obtain a fifth target eigenvalue;
[0088] Specifically, taking the fifth target eigenvalue corresponding to the first component as an example, the fifth target eigenvalue is used to characterize the red degree in the foam. Similarly, the fifth target eigenvalue corresponding to the second component is used to characterize the green degree in the foam. The fifth target eigenvalue corresponding to the third component is used to characterize the blue degree in the foam. The calculation formula is:
[0089]
[0090] Among them, R r is the fifth target eigenvalue mentioned above. The calculation method of the fifth eigenvalue corresponding to the second component and the third component is similar.
[0091] Step S2031110, based on converting the first target image into a grayscale image, a third target image is obtained, based on converting the third target image into a neighborhood grayscale correlation matrix and calculating the coarseness, a sixth target eigenvalue is obtained.
[0092] Specifically, the roughness C o The roughness is calculated by converting the grayscale image of the original image into the neighborhood grayscale correlation matrix (NGLDM). The formula for calculating the roughness is:
[0093]
[0094] Among them, Nrn and Ncn are the number of rows and columns of the neighborhood grayscale correlation matrix, respectively, and M i,j is the element value of the i-th row and j-th column in the matrix, C o is the sixth target eigenvalue mentioned above.
[0095] In order to denoise the second target vector, in an optional implementation, the above step S20313 includes:
[0096] Step S203131, taking the minimum value of the objective function as the termination condition, the second objective vector is subjected to fuzzy C-means clustering to obtain the target cluster center and the target cluster cluster;
[0097] Specifically, since the two abnormal working conditions, sinking and bubbling, are visually less different from normal working conditions, there is a possibility of labeling errors in the process of image anomaly recognition.
[0098] It is understandable that the above-mentioned incorrectly labeled data appears as an outlier among similar data, and the incorrectly labeled data will mislead the training of the classifier. Therefore, the present application sets up C-means clustering for the above-mentioned second target vector to perform denoising.
[0099] In the specific implementation, n sample data are divided into c groups, and the jth sample and the i-th cluster center point v are calculated. i The distance d ij When the membership function u is added ij To indicate the degree to which the jth sample belongs to the ith class, its objective function J m The calculation formula of (u, v) is as follows:
[0100]
[0101] Wherein, m represents the degree of blur and is set to 2 in this embodiment, but other parameter values may also be used. Wherein, the cluster center point v i And the membership function u ij The calculation formula is as follows:
[0102]
[0103] This application sets the goal of minimizing the objective function, iterates, and searches for the optimal membership function and cluster center.
[0104] Step S203132, determining the second target vector corresponding to the outlier point as a third target vector based on the target cluster center and the target cluster;
[0105] Specifically, after clustering is completed, outliers are determined based on the optimized cluster centers and the corresponding cluster clusters, and then the second target vector corresponding to the outlier is determined as the third target vector.
[0106] Step S203133, delete the third target vector and denoise the second target vector.
[0107] Specifically, the third target vector is the noise in the data set, and the third target vector is deleted to denoise the second target vector.
[0108] In order to obtain the first target vector, in an optional implementation, the above step S20313 includes:
[0109] Step S203134, adding a target identifier to each of the second target vectors to obtain a target data set, wherein the target identifier is used to indicate whether each of the second target vectors is input into the image recognition model;
[0110] Step S203135, obtaining a third target quantity based on the number of second target vectors in the target data set corresponding to the target identifier as input to the image recognition model;
[0111] Step S203136, based on the target data set, determining the accuracy of the target data set input into the image recognition model to obtain a target ratio;
[0112] Step S203137, constructing an optimization model based on the third target quantity, the target data set and the target ratio, solving the optimization model to obtain the first target vector:
[0113]
[0114] Among them, A cc (s) is the target ratio, s is the target data set, ||s||0 is the third target quantity, and n is the type of the second target vector.
[0115] In the above embodiments, since the present application sets multiple related features, the information contained is redundant and even interferes with each other, which increases the training time and difficulty of the classifier and affects the final recognition result.
[0116] It can be understood that in order to eliminate the above-mentioned interference, the present application sets up a feature optimization method based on a binary state transfer algorithm to screen out the most effective features.
[0117] In the specific implementation, the feature optimization problem is an optimization problem with a large search space, high dimension, and nonlinearity. Its goal is to make the feature subset as small as possible and the recognition accuracy as high as possible, which is expressed by the above formula.
[0118] Furthermore, the optimal feature subset can be quickly found through the binary state transfer algorithm, and the state transfer process is completed by four operators: the exchange transformation operator si and sj, the move transformation operator moves si to the front of sj, the symmetry transformation operator performs a center rotation transformation on the sequence from si to sj, and the substitution transformation operator performs an inversion operation on si.
[0119] The feature optimization method of the above binary state transfer algorithm is as follows:
[0120] S1: Randomly determine an initial feature subset s0;
[0121] S2: Use four operators to transform sk respectively, input it into the classifier, and take the feature subset with the highest accuracy as sk+1;
[0122] S3: Compare the current optimal state with sk+1 and update the optimal feature subset sbest.
[0123] S4: If the maximum number of iterations is reached, the iteration is terminated to obtain the final feature subset sbest, otherwise it returns to S1.
[0124] In the above embodiment, the steps of the multi-objective learning algorithm of the image recognition model of the present application are as follows:
[0125] S11: Calculate the loss function of all image recognition tasks and derive the shared layer weights and the specific task layer weights respectively;
[0126] S21: Update each specific task layer weight;
[0127] S31: Update shared layer weights.
[0128] In the above embodiment, a multi-task learning problem from input space X to output space Γ is considered, and the function expression of each task output is obtained through the neural network architecture:
[0129] f t (x;θ sh ,θ t ): X→Γ;
[0130] Among them, x is the input vector representing the image features; θsh is the weight of the shared layer; θt is the weight of the specific task layer.
[0131] The loss function corresponding to the above formula is as follows:
[0132] L t (θ st ,θ t )=loss(f t (x;θ sh ,θ t ), y t );
[0133] Among them, y is the category to which the image belongs; loss is the cross entropy loss function.
[0134] Furthermore, the multi-objective optimization problem can be expressed as follows:
[0135]
[0136] For multi-objective optimization problems, the final solution should be the Pareto optimal solution.
[0137] When the Pareto optimal solution is obtained, the Karush-Kuhn-Tucker (KKT) condition derives:
[0138]
[0139] Then, based on the Frank-Wolfe constrained optimization method, the above equation is solved quickly and effectively. The specific steps include:
[0140] Step 1: Initialize the weight vector
[0141] Step 2: Calculate the parameter matrix M,
[0142] Step 3: Pass Find the optimal index value Construct 1 A row vector e with T columns containing 1 and other columns containing 0 i .
[0143] Step 4: The optimal solution is obtained by
[0144]
[0145] Step 5: Pass Update alpha.
[0146] Step 6: If γ is within the allowable error range or the algorithm reaches the maximum number of iterations, terminate the iteration, otherwise return to step 3.
[0147] After obtaining the weight vector, we can further give the process of the multi-objective learning algorithm based on multi-gradient descent:
[0148] Step 1: Calculate the loss function L for all tasks t (θ sh ,θ t ) and compare them to θ sh and θ t Derivation, get and
[0149] Step 2: Pass Update the parameters of each specific task layer, where η is a pre-set learning rate.
[0150] Step 3: Use the Frank-Wolfe algorithm to obtain α, through Update shared layer parameters.
[0151] Step 4: If the maximum number of iterations is reached, terminate the iteration; otherwise, return to step 1.
[0152] Step 5: Save the updated parameters θ sh and θ t The coal flotation condition identification model is proposed.
[0153] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0154] The embodiment of the present application also provides an image-based coal flotation device. It should be noted that the image-based coal flotation device of the embodiment of the present application can be used to execute the image-based coal flotation method provided by the embodiment of the present application. The device is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware for a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceived.
[0155] The following is an introduction to the image-based coal flotation device provided in the embodiments of the present application.
[0156] Figure 3 is a structural block diagram of an image-based coal flotation device according to an embodiment of the present application. Figure 3 As shown, the device comprises:
[0157] The first control unit 10 is used to perform an adding step, adding a flotation agent to the slurry to separate the minerals from the impurities in the slurry;
[0158] Specifically, flotation reagents are added to the ore pulp to separate the minerals and impurities in the ore pulp. It can be understood that the flotation process is a process of separating minerals and gangue according to the difference in hydrophilic and hydrophobic properties of the minerals. The role of adding flotation reagents is to enhance the hydrophilicity of useful minerals and improve the hydrophobicity of gangue impurities.
[0159] In the specific implementation, the flotation reagents mainly include collecting reagents, inhibiting reagents and adjusting reagents. According to the actual production situation, the proportion of the three reagents and the amount of reagents to be prepared are set, and the addition speed of various raw materials is adjusted by controlling the start and stop of the delivery pump and the opening of the regulating valve, and the addition amount is monitored by detection instruments such as flow meters and liquid level meters. In this way, the addition of flotation reagents is realized.
[0160] The second control unit 20 is used to perform a stirring step, stir the slurry until foam is formed on the surface of the slurry, monitor the slurry surface image in real time, and obtain a first target image;
[0161] Specifically, the slurry is stirred to form foam on the upper layer of the slurry. It is understood that after adding the flotation reagent, air is introduced into the slurry in the flotation tank to form a large number of bubbles. Under the stirring of the stirring paddle, hydrophobic impurities such as gangue impurities adhere to the surface of the bubbles and float out with the bubbles into the foam layer, while useful minerals such as hydrophilic impurities remain in the slurry as underflow.
[0162] A third control unit 30 is used to perform a recognition step, classify the foam based on the first target image, obtain target foam, discharge the target foam and adjust the proportion of the components in the flotation reagent based on the amount of the target foam to eliminate and reduce the target foam, wherein the target foam is the foam containing impurities;
[0163] Specifically, the foams are classified through images to determine common foams and target foams with impurities, the target foams are discharged, and the flotation reagent components for the next round of flotation are adjusted based on the number of target foams.
[0164] The repeating unit 40 is used to sequentially repeat the adding step, the stirring step and the identifying step at least once until the target foam quantity is less than or equal to a first threshold.
[0165] Specifically, the above-mentioned adding step, the above-mentioned stirring step and the above-mentioned identification step are repeated multiple times, that is, the roughing, concentrating and scavenging stages are performed multiple times. After the raw material slurry is regrinded and concentrated, the slurry that meets certain particle size and concentration requirements is pumped to the flotation process, and the first sorting is carried out in the roughing stage. The bottom flow enriched with useful minerals enters the concentrating stage for further sorting, and the roughing foam is sent to the scavenging process. Then, the roughing bottom flow slurry passes through the concentrating stage to obtain a qualified product as the final concentrate, and the concentrating foam returns to the roughing stage. The roughing foam reduces the useful minerals in the final tailings through the scavenging stage, and the scavenging foam is discharged into the foam collector as the final tailings. The scavenging bottom flow and the concentrating foam are sent to the roughing stage together for another sorting.
[0166] In the above embodiment, before adding the flotation reagent to the slurry, the method further includes:
[0167] The minerals in the slurry are ground until the particle size of the minerals is less than or equal to a preset particle size.
[0168] Specifically, the minerals in the slurry are ground so that the particle size of the minerals meets the preset conditions.
[0169] It is understandable that flotation requires not only the minerals to be fully dissociated into monomers, but also the appropriate particle size for selection. If the mineral particles are too coarse, even if the minerals have been dissociated into monomers, they often cannot float because they exceed the buoyancy capacity of the bubbles. If the grinding particles are too fine, it is also not conducive to coal flotation.
[0170] In the specific implementation, the closed-circuit grinding consisting of a ball mill and a cyclone is used to regrind the concentrate from the magnetic separation process, so that the particle size is increased from 89% of -200 mesh to 94% of -200 mesh, so as to meet the particle size requirements of the flotation process and thereby improve the flotation efficiency.
[0171] After grinding the minerals in the slurry until the particle size of the minerals is less than or equal to the preset particle size, the method further includes:
[0172] The concentration and flow rate of the slurry are adjusted by a thickener until the concentration reaches a preset concentration and the flow rate reaches a preset flow rate.
[0173] Specifically, a thickener is used to adjust the concentration and flow rate of the slurry, i.e., the concentration step.
[0174] It is understandable that the concentration process is an important part of the coal flotation process. The slurry transported from the magnetic separation process is concentrated and separated by a thickener to ensure that the slurry with appropriate concentration and flow rate enters the flotation process.
[0175] In specific implementation, appropriate pulp concentration is conducive to obtaining higher metal recovery rate and concentrate grade in the flotation process. The pulp flow rate determines the flotation time, which in turn affects the metal recovery rate and concentrate grade in the flotation process.
[0176] Through this embodiment, the first control unit performs an adding step to add a flotation agent to the slurry to separate the minerals from the impurities in the slurry; the second control unit performs a stirring step to stir the slurry until foam is formed on the surface of the slurry, monitors the surface image of the slurry in real time, and obtains a first target image; the third control unit performs an identification step to classify the foam based on the first target image to obtain the target foam, discharge the target foam and adjust the proportion of the components in the flotation agent based on the amount of the target foam to eliminate and reduce the target foam, wherein the target foam is the foam containing impurities; the repeating unit repeats the adding step, the stirring step, and the identification step at least once in sequence until the amount of the target foam is less than or equal to the first threshold. The present application identifies the foam components through images to determine the components of the foam, and adjusts the flotation agent to separate impurities according to the distinguished foam components. Compared with manual observation of the foam components, this method improves the accuracy and standard of identification and realizes precise control. This method solves the problem in the prior art of monitoring the flotation production status by manual observation of the foam, which leads to untimely adjustment of the agent and affects the flotation effect.
[0177] In order to obtain the target foam, in an optional embodiment, the third control unit comprises:
[0178] A first processing subunit is used to pre-process the first target image, convert the first target image into a vector representation, and obtain a first target vector;
[0179] Specifically, the first target image is preprocessed and converted into a vector representation to obtain the first target vector. The preprocessing at least includes format conversion, denoising and optimization.
[0180] The second processing subunit is used to input the first target vector into the image recognition model for classification to obtain the target foam.
[0181] Specifically, feature extraction is performed based on the first target vector, and classification and identification are performed through a model based on the features to determine the foam containing impurities and obtain the target foam.
[0182] In order to obtain the first target vector, in an optional implementation, the first processing subunit includes:
[0183] A first processing module is used to extract feature values from the first target image to obtain a plurality of target feature values, wherein the plurality of target feature values are used to characterize different features of the first target image;
[0184] Specifically, since the image is directly used as the input of the classifier, the dimension is too high, which leads to a long training time and may cause the model to fall into a local optimal point. Based on the above considerations, this application selects multiple features such as foam average area, load rate and color tone characteristics, and converts the multiple features into feature value representation to obtain the above target feature value.
[0185] A second processing module is used to express the target feature value in a vector form to obtain a plurality of second target vectors;
[0186] Specifically, the target eigenvalue is converted into a vector representation to obtain the second target vector.
[0187] The third processing module is used to denoise and optimize the second target vector to obtain the first target vector.
[0188] Specifically, denoising and optimization operations are performed in sequence based on the second target vector to obtain the first target vector.
[0189] In order to extract the average foam area, in an optional embodiment, the first processing module includes:
[0190] A first processing submodule is used to segment the first target image by a watershed algorithm, filter out the foam area in the first target image, and obtain a second target image;
[0191] Specifically, the foam area in the image is segmented by a watershed algorithm to obtain the above-mentioned second target image.
[0192] A first determination submodule is used to determine the corresponding foam quantity based on each of the second target images to obtain a first target quantity;
[0193] Specifically, each of the second target images is traversed to determine the number of bubbles contained therein, thereby obtaining the first target number.
[0194] A first calculation submodule is used to sum the pixel areas of each of the second target images to obtain a first target area;
[0195] Specifically, the pixel areas of each of the second target images are summed to obtain the first target area Sp.
[0196] A second calculation submodule is used for summing the second target quantities to obtain the second target quantity, and calculating the ratio of the first target area to the second target quantity to obtain the second target area;
[0197] Specifically, the number of bubbles in each region is added to get N b , the above second target quantity is obtained, based on the average foam area S meanCalculation formula S mean =S p / N b The second target area is calculated.
[0198] The second determining submodule is used to determine the second target area as the first target characteristic value.
[0199] In order to calculate multiple target feature values, in an optional implementation, the second processing subunit further includes:
[0200] A third calculation submodule is used to calculate the hue value based on the first component, the second component and the third component of each pixel point of the first target image to obtain a second target feature value, wherein the first component is a red component, the second component is a green component, and the third component is a blue component;
[0201] Specifically, when the distance between the foam and the light source is different, it is difficult to describe the color of the foam surface under the same light intensity standard, that is, there is a deviation in feature extraction based solely on the color in the image RGB space. Therefore, this application is set to implement feature extraction in the HSV space.
[0202] Furthermore, by calculating the image hue value, the image is converted from the RGB space to the HSV space to extract the hue features that are not affected by light intensity. The calculation method is:
[0203]
[0204] Among them, Hi is the hue value of the i-th pixel, that is, the second eigenvalue mentioned above, Ri is the first component mentioned above, Gi is the second component mentioned above, and Bi is the third component mentioned above.
[0205] A fourth calculation submodule, used to calculate the mean value based on each of the first component, the second component and the third component to obtain a corresponding third target eigenvalue;
[0206] Specifically, the calculation formula of the third target eigenvalue of each of the first component, the second component and the third component is:
[0207]
[0208] Among them, Np is the number of pixels occupied by foam, R mean is the third target eigenvalue of the first component, i.e., the red mean. It can be understood that the third eigenvalue (green mean) corresponding to the second component G mean The third eigenvalue (blue mean) B corresponding to the third component mean The calculation method of the third target eigenvalue is the same as that of , and the corresponding components can be substituted.
[0209] A fifth calculation submodule, configured to calculate a grayscale mean based on the third target feature value of each pixel point of the first target image to obtain a fourth target feature value;
[0210] Specifically, the grayscale mean of an image reflects the brightness of the image, and its calculation formula is:
[0211] Y mean =0.299R mean +0.587G mean +0.114B mean ;
[0212] Among them, Y mean is the fourth target eigenvalue mentioned above.
[0213] a sixth calculation submodule, configured to calculate a ratio based on the third target eigenvalue and the fourth target eigenvalue to obtain a fifth target eigenvalue;
[0214] Specifically, taking the fifth target eigenvalue corresponding to the first component as an example, the fifth target eigenvalue is used to characterize the red degree in the foam. Similarly, the fifth target eigenvalue corresponding to the second component is used to characterize the green degree in the foam. The fifth target eigenvalue corresponding to the third component is used to characterize the blue degree in the foam. The calculation formula is:
[0215]
[0216] Among them, R r is the fifth target eigenvalue mentioned above. The calculation method of the fifth eigenvalue corresponding to the second component and the third component is similar.
[0217] The second processing submodule is used to convert the first target image into a grayscale image to obtain a third target image, convert the third target image into a neighborhood grayscale correlation matrix and calculate the coarseness to obtain a sixth target eigenvalue.
[0218] Specifically, the roughness C o The roughness is calculated by converting the grayscale image of the original image into the neighborhood grayscale correlation matrix (NGLDM). The formula for calculating the roughness is:
[0219]
[0220] Among them, Nrn and Ncn are the number of rows and columns of the neighborhood grayscale correlation matrix, respectively, and M i,j is the element value of the i-th row and j-th column in the matrix, C o is the sixth target eigenvalue mentioned above.
[0221] In order to denoise the second target vector, in an optional implementation, the third processing module includes:
[0222] A seventh calculation submodule is used to perform fuzzy C-means clustering on the second target vector with the minimum value of the target function as the termination condition to obtain a target cluster center and a target cluster cluster;
[0223] Specifically, since the two abnormal working conditions, sinking and bubbling, are visually less different from normal working conditions, there is a possibility of labeling errors in the process of image anomaly recognition.
[0224] It is understandable that the above-mentioned incorrectly labeled data appears as an outlier among similar data, and the incorrectly labeled data will mislead the training of the classifier. Therefore, the present application sets up C-means clustering for the above-mentioned second target vector to perform denoising.
[0225] In the specific implementation, n sample data are divided into c groups, and the jth sample and the i-th cluster center point v are calculated. i The distance d ij When the membership function u is added ij To indicate the degree to which the jth sample belongs to the ith class, its objective function J m The calculation formula of (u, v) is as follows:
[0226]
[0227] Wherein, m represents the degree of blur and is set to 2 in this embodiment, but other parameter values may also be used. Wherein, the cluster center point v i And the membership function u ij The calculation formula is as follows:
[0228]
[0229] This application sets the goal of minimizing the objective function, iterates, and searches for the optimal membership function and cluster center.
[0230] A third determination submodule, configured to determine the second target vector corresponding to the outlier point as a third target vector based on the target cluster center and the target cluster cluster;
[0231] Specifically, after clustering is completed, outliers are determined based on the optimized cluster centers and the corresponding cluster clusters, and then the second target vector corresponding to the outlier is determined as the third target vector.
[0232] The third processing submodule is used to delete the third target vector and perform denoising on the second target vector.
[0233] Specifically, the third target vector is the noise in the data set, and the third target vector is deleted to denoise the second target vector.
[0234] In order to obtain the first target vector, in an optional implementation, the third processing module further includes:
[0235] A fourth processing submodule is used to add a target identifier to each of the second target vectors to obtain a target data set, wherein the target identifier is used to indicate whether each of the second target vectors is input into the image recognition model;
[0236] An eighth calculation submodule, configured to obtain a third target quantity based on the quantity of second target vectors corresponding to the target identifier in the target data set as input to the image recognition model;
[0237] a ninth calculation submodule, configured to determine, based on the target data set, the accuracy of the target data set input into the image recognition model, and obtain a target ratio;
[0238] A tenth calculation submodule is used to construct an optimization model based on the third target quantity, the target data set and the target ratio, and solve the optimization model to obtain the first target vector:
[0239]
[0240] Among them, A cc (s) is the target ratio, s is the target data set, ||s||0 is the third target quantity, and n is the type of the second target vector.
[0241] The image-based coal flotation device includes a processor and a memory. The first control unit, the second control unit, the third control unit and the repeating unit are stored in the memory as program units, and the processor executes the program units stored in the memory to implement corresponding functions. The modules are all located in the same processor; or, the modules are located in different processors in any combination.
[0242] The processor includes a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and the efficiency and accuracy of the coal flotation process can be improved by adjusting the kernel parameters.
[0243] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.
[0244] An embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the image-based coal flotation method.
[0245] Specifically, the image-based coal flotation method includes:
[0246] Step S201, adding a flotation agent to the slurry to separate the minerals from the impurities in the slurry;
[0247] Specifically, flotation reagents are added to the ore pulp to separate the minerals and impurities in the ore pulp. It can be understood that the flotation process is a process of separating minerals and gangue according to the difference in hydrophilic and hydrophobic properties of the minerals. The role of adding flotation reagents is to enhance the hydrophilicity of useful minerals and improve the hydrophobicity of gangue impurities.
[0248] In the specific implementation, the flotation reagents mainly include collecting reagents, inhibiting reagents and adjusting reagents. According to the actual production situation, the proportion of the three reagents and the amount of reagents to be prepared are set, and the addition speed of various raw materials is adjusted by controlling the start and stop of the delivery pump and the opening of the regulating valve, and the addition amount is monitored by detection instruments such as flow meters and liquid level meters. In this way, the addition of flotation reagents is realized.
[0249] Step S202, a stirring step, stirring the slurry until foam is formed on the surface of the slurry, and monitoring the slurry surface image in real time to obtain a first target image;
[0250] Specifically, the slurry is stirred to form foam on the upper layer of the slurry. It is understood that after adding the flotation reagent, air is introduced into the slurry in the flotation tank to form a large number of bubbles. Under the stirring of the stirring paddle, hydrophobic impurities such as gangue impurities adhere to the surface of the bubbles and float out with the bubbles into the foam layer, while useful minerals such as hydrophilic impurities remain in the slurry as underflow.
[0251] Step S203, a recognition step, classifying the foam based on the first target image to obtain target foam, discharging the target foam and adjusting the proportion of the components in the flotation reagent based on the amount of the target foam to eliminate and reduce the target foam, wherein the target foam is the foam containing impurities;
[0252] Specifically, the foams are classified through images to determine common foams and target foams with impurities, the target foams are discharged, and the flotation reagent components for the next round of flotation are adjusted based on the number of target foams.
[0253] Step S204, repeating the adding step, the stirring step and the identifying step at least once in sequence until the amount of the target foam is less than or equal to a first threshold.
[0254] Specifically, the above-mentioned adding step, the above-mentioned stirring step and the above-mentioned identification step are repeated multiple times, that is, the roughing, concentrating and scavenging stages are performed multiple times. After the raw material slurry is regrinded and concentrated, the slurry that meets certain particle size and concentration requirements is pumped to the flotation process, and the first sorting is carried out in the roughing stage. The bottom flow enriched with useful minerals enters the concentrating stage for further sorting, and the roughing foam is sent to the scavenging process. Then, the roughing bottom flow slurry passes through the concentrating stage to obtain a qualified product as the final concentrate, and the concentrating foam returns to the roughing stage. The roughing foam reduces the useful minerals in the final tailings through the scavenging stage, and the scavenging foam is discharged into the foam collector as the final tailings. The scavenging bottom flow and the concentrating foam are sent to the roughing stage together for another sorting.
[0255] An embodiment of the present invention provides a processor, and the processor is used to run a program, wherein the image-based coal flotation method is executed when the program is run.
[0256] Specifically, the image-based coal flotation method includes:
[0257] Step S201, adding a flotation agent to the slurry to separate the minerals from the impurities in the slurry;
[0258] Specifically, flotation reagents are added to the ore pulp to separate the minerals and impurities in the ore pulp. It can be understood that the flotation process is a process of separating minerals and gangue according to the difference in hydrophilic and hydrophobic properties of the minerals. The role of adding flotation reagents is to enhance the hydrophilicity of useful minerals and improve the hydrophobicity of gangue impurities.
[0259] In the specific implementation, the flotation reagents mainly include collecting reagents, inhibiting reagents and adjusting reagents. According to the actual production situation, the proportion of the three reagents and the amount of reagents to be prepared are set, and the addition speed of various raw materials is adjusted by controlling the start and stop of the delivery pump and the opening of the regulating valve, and the addition amount is monitored by detection instruments such as flow meters and liquid level meters. In this way, the addition of flotation reagents is realized.
[0260] Step S202, a stirring step, stirring the slurry until foam is formed on the surface of the slurry, and monitoring the slurry surface image in real time to obtain a first target image;
[0261] Specifically, the slurry is stirred to form foam on the upper layer of the slurry. It is understood that after adding the flotation reagent, air is introduced into the slurry in the flotation tank to form a large number of bubbles. Under the stirring of the stirring paddle, hydrophobic impurities such as gangue impurities adhere to the surface of the bubbles and float out with the bubbles into the foam layer, while useful minerals such as hydrophilic impurities remain in the slurry as underflow.
[0262] Step S203, a recognition step, classifying the foam based on the first target image to obtain target foam, discharging the target foam and adjusting the proportion of the components in the flotation reagent based on the amount of the target foam to eliminate and reduce the target foam, wherein the target foam is the foam containing impurities;
[0263] Specifically, the foams are classified through images to determine common foams and target foams with impurities, the target foams are discharged, and the flotation reagent components for the next round of flotation are adjusted based on the number of target foams.
[0264] Step S204, repeating the adding step, the stirring step and the identifying step at least once in sequence until the amount of the target foam is less than or equal to a first threshold.
[0265] Specifically, the above-mentioned adding step, the above-mentioned stirring step and the above-mentioned identification step are repeated multiple times, that is, the roughing, concentrating and scavenging stages are performed multiple times. After the raw material slurry is regrinded and concentrated, the slurry that meets certain particle size and concentration requirements is pumped to the flotation process, and the first sorting is carried out in the roughing stage. The bottom flow enriched with useful minerals enters the concentrating stage for further sorting, and the roughing foam is sent to the scavenging process. Then, the roughing bottom flow slurry passes through the concentrating stage to obtain a qualified product as the final concentrate, and the concentrating foam returns to the roughing stage. The roughing foam reduces the useful minerals in the final tailings through the scavenging stage, and the scavenging foam is discharged into the foam collector as the final tailings. The scavenging bottom flow and the concentrating foam are sent to the roughing stage together for another sorting.
[0266] An embodiment of the present invention provides a coal flotation system, the coal flotation system comprising a processor, a memory, and a program stored in the memory and executable on the processor, and when the processor executes the program, at least the following steps are implemented:
[0267] Step S201, adding a flotation agent to the slurry to separate the minerals from the impurities in the slurry;
[0268] Step S202, a stirring step, stirring the slurry until foam is formed on the surface of the slurry, and monitoring the slurry surface image in real time to obtain a first target image;
[0269] Step S203, a recognition step, classifying the foam based on the first target image to obtain target foam, discharging the target foam and adjusting the proportion of the components in the flotation reagent based on the amount of the target foam to eliminate and reduce the target foam, wherein the target foam is the foam containing impurities;
[0270] Step S204, repeating the adding step, the stirring step and the identifying step at least once in sequence until the amount of the target foam is less than or equal to a first threshold.
[0271] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program for initializing at least the following method steps:
[0272] Step S201, adding a flotation agent to the slurry to separate the minerals from the impurities in the slurry;
[0273] Step S202, a stirring step, stirring the slurry until foam is formed on the surface of the slurry, and monitoring the slurry surface image in real time to obtain a first target image;
[0274] Step S203, a recognition step, classifying the foam based on the first target image to obtain target foam, discharging the target foam and adjusting the proportion of the components in the flotation reagent based on the amount of the target foam to eliminate and reduce the target foam, wherein the target foam is the foam containing impurities;
[0275] Step S204, repeating the adding step, the stirring step and the identifying step at least once in sequence until the amount of the target foam is less than or equal to a first threshold.
[0276] Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order than here, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.
[0277] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.
[0278] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0279] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0280] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0281] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0282] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.
[0283] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.
[0284] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity or device including the element.
[0285] From the above description, it can be seen that the above embodiments of the present application achieve the following technical effects:
[0286] 1) The image-based coal flotation method of the present application includes an adding step, adding a flotation agent to the slurry to separate the minerals from the impurities in the slurry; a stirring step, stirring the slurry until foam is formed on the surface of the slurry, monitoring the surface image of the slurry in real time, and obtaining a first target image; an identification step, classifying the foam based on the first target image to obtain the target foam, discharging the target foam and adjusting the proportion of the components in the flotation agent based on the amount of the target foam to eliminate and reduce the target foam, wherein the target foam is the foam containing impurities; repeating the adding step, the stirring step, and the identification step at least once in sequence until the amount of the target foam is less than or equal to the first threshold. The present application identifies the foam components through images to determine the components of the foam, and adjusts the flotation agent to separate impurities according to the distinguished foam components. Compared with manual observation of the foam components, this method improves the accuracy and standard of identification and realizes precise control. This method solves the problem in the prior art of manually observing the foam to monitor the flotation production status, resulting in untimely adjustment of the agent and affecting the flotation effect.
[0287] 2) In the image-based coal flotation device of the present application, the first control unit performs an adding step to add a flotation agent to the slurry to separate the minerals from the impurities in the slurry; the second control unit performs a stirring step to stir the slurry until foam is formed on the surface of the slurry, and monitors the surface image of the slurry in real time to obtain a first target image; the third control unit performs an identification step to classify the foam based on the first target image to obtain the target foam, discharge the target foam and adjust the proportion of the components in the flotation agent based on the amount of the target foam to eliminate and reduce the target foam, wherein the target foam is the foam containing impurities; the repeating unit repeats the adding step, the stirring step and the identification step at least once in sequence until the amount of the target foam is less than or equal to the first threshold. The present application identifies the foam components through images to determine the components of the foam, and adjusts the flotation agent to separate impurities according to the distinguished foam components. Compared with manual observation of the foam components, this method improves the accuracy and standard of identification and realizes precise control. This method solves the problem in the prior art of monitoring the flotation production status by manual observation of the foam, which leads to untimely adjustment of the agent and affects the flotation effect.
[0288] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims
1. An image-based coal flotation method, characterized in that: include: An adding step of adding a flotation agent to the ore pulp to separate the minerals from the impurities in the ore pulp; A stirring step, stirring the slurry until foam is formed on the surface of the slurry, and monitoring the slurry surface image in real time to obtain a first target image; an identification step, classifying the foam based on the first target image to obtain target foam, discharging the target foam and adjusting the proportion of the components in the flotation reagent based on the amount of the target foam to eliminate and reduce the target foam, wherein the target foam is the foam containing impurities; The adding step, the stirring step and the identifying step are sequentially repeated at least once until the target foam quantity is less than or equal to a first threshold.
2. The method according to claim 1, characterized in that Classifying foam based on the first target image to obtain target foam includes: Preprocessing the first target image, converting the first target image into a vector representation, and obtaining a first target vector; The first target vector is input into an image recognition model for classification to obtain the target foam.
3. The method according to claim 2, characterized in that Preprocessing the first target image, converting the first target image into a vector representation, and obtaining a first target vector, includes: Extracting feature values from the first target image to obtain a plurality of target feature values, wherein the plurality of target feature values are used to characterize different features of the first target image; Representing the target feature value in a vector form to obtain a plurality of second target vectors; The second target vector is denoised and optimized to obtain the first target vector.
4. The method according to claim 3, characterized in that Extracting feature values from the first target image to obtain multiple target feature values includes: Segmenting the first target image by using a watershed algorithm, screening out the foam area in the first target image, and obtaining a second target image; Determine the corresponding foam quantity based on each of the second target images to obtain a first target quantity; Summing the pixel areas of each of the second target images to obtain a first target area; summing up the second target quantities to obtain a second target quantity, and calculating a ratio of the first target area to the second target quantity to obtain a second target area; The second target area is determined as the first target characteristic value.
5. The method according to claim 4, characterized in that Extracting feature values from the first target image to obtain a plurality of target feature values also includes: Calculate the hue value based on the first component, the second component and the third component of each pixel point of the first target image to obtain a second target feature value, wherein the first component is a red component, the second component is a green component, and the third component is a blue component; Calculate the mean value based on each of the first component, the second component and the third component to obtain a corresponding third target eigenvalue; Calculate the grayscale mean based on the third target feature value of each pixel point of the first target image to obtain a fourth target feature value; Calculate a ratio based on the third target eigenvalue and the fourth target eigenvalue to obtain a fifth target eigenvalue; Based on converting the first target image into a grayscale image, a third target image is obtained, and based on converting the third target image into a neighborhood grayscale correlation matrix and calculating the coarseness, a sixth target eigenvalue is obtained.
6. The method according to claim 3, characterized in that: Denoising the second target vector includes: Taking the minimum value of the objective function as the termination condition, the second objective vector is subjected to fuzzy C-means clustering to obtain the target cluster center and the target cluster cluster; Determining the second target vector corresponding to the outlier as a third target vector based on the target cluster center and the target cluster; The third target vector is deleted to denoise the second target vector.
7. The method according to claim 3, characterized in that Optimizing the second target vector to obtain the first target vector includes: Adding a target identifier to each of the second target vectors to obtain a target data set, wherein the target identifier is used to indicate whether each of the second target vectors is input into the image recognition model; Obtaining a third target quantity based on the number of second target vectors corresponding to the target identifier in the target data set as input to the image recognition model; Based on the target data set, determining the accuracy of the target data set input into the image recognition model to obtain a target ratio; An optimization model is constructed based on the third target quantity, the target data set, and the target ratio, and the optimization model is solved to obtain the first target vector: Among them, A cc (s) is the target ratio, s is the target data set, ||s||0 is the third target quantity, and n is the type of the second target vector.
8. An image-based coal flotation device, characterized in that: The device comprises: A first control unit is used to perform an adding step, adding a flotation agent to the ore pulp to separate the minerals from the impurities in the ore pulp; a second control unit, configured to execute a stirring step, stir the slurry until foam is formed on the surface of the slurry, monitor the slurry surface image in real time, and obtain a first target image; a third control unit, configured to perform a recognition step, classify the foam based on the first target image to obtain target foam, discharge the target foam and adjust the proportion of the components in the flotation reagent based on the amount of the target foam to eliminate and reduce the target foam, wherein the target foam is the foam containing impurities; A repeating unit is used to sequentially repeat the adding step, the stirring step and the identifying step at least once until the amount of the target foam is less than or equal to a first threshold.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 7.
10. A coal flotation system, characterized in that: include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the one or more programs include methods for executing any one of claims 1 to 7.
