Vertical monitoring system for foam size and survival rule in flotation machine

Abstract

The invention provides a vertical monitoring system for the size and survival rule of foam in a flotation machine, and belongs to the technical field of automatic control. The vertical monitoring system for the size and survival rule of the foam in the flotation machine comprises a plurality of image acquisition devices used for acquiring foam images in real time, and the plurality of image acquisition devices are sequentially arranged along the height direction of a flotation machine groove; and the processor is used for identifying the plurality of foam images to obtain the foam sizes and the foam quantity of the plurality of heights, and obtaining the foam sizes and survival rules based on the foam sizes and the foam quantity of the plurality of heights. The problem that an existing foam layer is poor in monitoring accuracy can be solved.

Description

Technical Field

[0001] The present disclosure relates to the field of automatic control technology, and in particular to a vertical monitoring system for foam size and survival law in a flotation machine. Background Art

[0002] Coal flotation machine is an important equipment for coal sorting. It mainly utilizes the difference in physical and chemical properties of mineral surfaces. By adding specific flotation agents into the pulp, the target mineral particles are selectively attached to the bubbles and rise to the pulp surface with the bubbles to form a foam layer, thereby achieving separation from the gangue minerals. The characteristics of the foam are closely related to the flotation effect. Therefore, by monitoring the foam, the state of the flotation process can be understood in real time, and the process parameters of the flotation machine can be adjusted to ensure the effective separation of minerals.

[0003] Traditional flotation foam monitoring devices are mostly single-point or plane measurement foam layer monitoring fields, which have certain errors and blind spots in application and cannot achieve accurate monitoring of the foam layer. Summary of the invention

[0004] The disclosed embodiment provides a vertical monitoring system for foam size and survival law in a flotation machine to solve the problem of poor accuracy in existing foam layer monitoring.

[0005] In a first aspect, an embodiment of the present disclosure provides a vertical monitoring system for foam size and survival law in a flotation machine, comprising: An image acquisition device, used for acquiring foam images in real time, wherein there are a plurality of image acquisition devices, and the plurality of image acquisition devices are arranged in sequence along the height direction of the flotation machine tank; The processor is used to identify multiple foam images, obtain foam sizes and foam quantities at multiple heights, and obtain foam sizes and survival rules based on the foam sizes and foam quantities at multiple heights.

[0006] In an exemplary embodiment of the present disclosure, a vertical monitoring system for froth size and survival law in a flotation machine further includes: A laser scattering instrument, used for real-time monitoring of physical parameters of the foam, wherein there are multiple laser scattering instruments, and the laser scattering instruments are arranged corresponding to the image acquisition device; The processor is specifically used for: The foam image is identified based on the physical parameters of the foam to obtain the foam size and foam quantity.

[0007] In an exemplary embodiment of the present disclosure, the physical parameters of the foam include foam thickness, The processor is specifically used for: screening a first region from the foam image based on foam thickness; Image recognition is performed on the first area to obtain the foam size and foam quantity.

[0008] In an exemplary embodiment of the present disclosure, the processor is further configured to: The number of convolutional layers and the size of the convolutional kernel of the convolutional neural network are determined based on the change in light intensity and the slurry concentration in the flotation machine; Image recognition is performed on the first area based on the convolutional neural network to obtain the foam size and foam quantity.

[0009] In an exemplary embodiment of the present disclosure, the processor is further configured to: determining a first adjustment parameter and a second adjustment parameter based on an amount of change in light intensity in the flotation machine; determining a third adjustment parameter and a fourth adjustment parameter based on the slurry concentration in the flotation machine; Adjust the first reference value based on the first adjustment parameter and the third adjustment parameter to obtain the number of convolution layers; The second reference value is adjusted based on the second adjustment parameter and the fourth adjustment parameter to obtain the size of the convolution kernel.

[0010] In an exemplary embodiment of the present disclosure, the physical parameters of the foam include foam size distribution information. The processor is further specifically configured to: The size of the convolution kernel of the convolutional neural network is adjusted based on the bubble size distribution information.

[0011] In an exemplary embodiment of the present disclosure, a vertical monitoring system for froth size and survival law in a flotation machine further includes: The foam stabilization pipeline is arranged along the height direction of the flotation machine tank, and the bottom opening of the foam pipeline is connected with the slurry outlet of the stirring zone of the flotation machine tank.

[0012] In an exemplary embodiment of the present disclosure, a vertical monitoring system for froth size and survival law in a flotation machine further includes: An image acquisition board is arranged on the inner wall of the foam stabilization pipeline, and the image acquisition device is arranged on the outer wall of the foam stabilization pipeline, and the image acquisition board and the image acquisition device are arranged correspondingly.

[0013] In an exemplary embodiment of the present disclosure, a vertical monitoring system for froth size and survival law in a flotation machine further includes: The high-pressure cleaning pipe is arranged in the foam stabilizing pipeline.

[0014] In a second aspect, the present disclosure provides a method for vertically monitoring the size and survival law of foam in a flotation machine, comprising: Collecting foam images in real time based on multiple image acquisition devices, wherein the multiple image acquisition devices are arranged in sequence along the height direction of the flotation machine tank; Multiple foam images are identified to obtain foam sizes and foam quantities at multiple heights, and foam sizes and survival rules are obtained based on the foam sizes and foam quantities at multiple heights.

[0015] The beneficial effects of the vertical monitoring system for foam size and survival law in a flotation machine provided by the embodiment of the present disclosure are: The embodiments of the present disclosure take into account that during the flotation process, foam will run from bottom to top. By arranging image acquisition devices at different heights to acquire foam images from different heights and obtaining foam sizes and foam quantities at different heights based on the foam images, the generation, development and extinction process of the foam can be comprehensively analyzed, and the flotation process parameters can be adjusted accordingly to obtain better flotation effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0017] Figure 1 It is a schematic diagram of the working process of a vertical monitoring system for foam size and survival law in a flotation machine provided by an embodiment of the present disclosure; Figure 2 It is a structural schematic diagram of a vertical monitoring system for foam size and survival law in a flotation machine provided by an embodiment of the present disclosure; In the figure: 1 fixed bracket, 2 foam collector, 3 high pressure cleaning pipe, 4 image acquisition device, 5 laser scattering instrument, 6 image acquisition board, 7 foam stabilization pipeline, 8 foam inlet. DETAILED DESCRIPTION

[0018] In order to enable people in the technical field to better understand the present solution, the technical solution in the embodiment of the present solution will be clearly described below in conjunction with the drawings in the embodiment of the present solution. Obviously, the described embodiment is an embodiment of a part of the present solution, not all of the embodiments. Based on the embodiments in the present solution, all other embodiments obtained by ordinary technicians in the field without creative work should fall within the scope of protection of the present solution.

[0019] The term "including" and any other variations in the specification and claims of this solution and the above drawings mean "including but not limited to", and is intended to cover non-exclusive inclusions and is not limited to the examples listed in the text. In addition, the terms "first" and "second" are used to distinguish different objects, not to describe a specific order.

[0020] The following is a detailed description of the implementation of the present disclosure in conjunction with the specific drawings: Reference Figure 1-Figure 2 , the vertical monitoring system for froth size and survival in flotation machines includes: An image acquisition device 4 is used to acquire foam images in real time. There are multiple image acquisition devices 4, and the multiple image acquisition devices 4 are arranged in sequence along the height direction of the flotation machine tank; The processor is used to identify the foam image, obtain the foam size and the foam quantity, so as to monitor the foam size and survival law.

[0021] In this embodiment, the image acquisition device 4 may include a high-speed industrial camera and a foam collector 2, wherein the high-speed camera captures images of flotation foam at a high frame rate and transmits the image data to the image collector in real time. The image collector receives the image data, performs preprocessing, and then stores or transmits the data to the processor.

[0022] The foam in the flotation machine can be collected cyclically by the image acquisition device 4. According to the flotation speed and flotation thread of the foam, no less than three image acquisition devices 4 can be arranged in the height direction (i.e., vertical direction) of the flotation machine tank to ensure that a single cycle acquisition is for the entire life cycle of the same batch of foam.

[0023] The processor identifies the foam images at different heights and obtains the foam size and foam quantity at different heights. Based on the statistical analysis of the foam size and quantity at different heights, the model of foam generation, stability and collapse process is established through time series analysis method as the flotation height changes, which can be used to predict the foam survival time and stability. For example, the relationship curve between foam size and floating height, the relationship curve between foam size and foam quantity, the relationship curve between floating height and foam collapse rate, the relationship curve between initial foam size and air intake, the relationship curve between initial foam size and stirring rate, etc. can be drawn.

[0024] For the same slurry properties, when the real-time monitoring survival curve of the foam deviates from the optimal foam survival curve, the flotation machine can automatically adjust the aeration volume and stirring rate according to the corresponding control instructions. The general standard is: compare the initial foam size with the optimal survival law. If it is too large, the aeration volume is reduced and the stirring rate is increased; if it is too small, the aeration volume is increased and the stirring rate is reduced. The control change is gradient adjustment, and the air inlet gradient value is ±0.1m 3 / min, the stirring rate gradient is ±1r / min.

[0025] This embodiment takes into account that during the flotation process, the foam will run from bottom to top. By arranging the image acquisition device 4 at different heights to acquire foam images from different heights, and obtaining the foam size and foam quantity at different heights based on the foam image, the generation, development and extinction process of the foam can be comprehensively analyzed, and the flotation process parameters can be adjusted accordingly to obtain a better flotation effect.

[0026] Reference Figure 2 In one embodiment of the present disclosure, the vertical monitoring system for froth size and survival law in a flotation machine further includes: A laser scattering instrument 5, used for real-time monitoring of physical parameters of the foam, wherein there are multiple laser scattering instruments 5, and the laser scattering instruments 5 are arranged corresponding to the image acquisition device 4; The processor is specifically used to: The foam image is identified based on the physical parameters of the foam to obtain the foam size and foam quantity.

[0027] In this embodiment, it is considered that during the flotation process, the foam often has an irregular shape; at the same time, due to the uneven illumination around the flotation machine, the brightness of the foam in the image is inconsistent. Some foams may appear too bright due to excessive illumination, and the contrast between the edge and the background is reduced, making the edge unclear; while other foams may appear too dark due to insufficient illumination, and it is also difficult to distinguish their edges; if the shooting angle is not appropriate, the foam image may be deformed, resulting in inaccurate size judgment. The above factors will lead to certain errors in the recognition of foam size and foam quantity by relying solely on the image acquisition device 4.

[0028] In this embodiment, by setting up a laser scatterer 5, physical parameters such as foam thickness, bubble size distribution, and foam density can be monitored in real time. The industrial high-speed camera, laser scatterer 5, and foam collector 2 can adopt a modular integrated design. The industrial high-speed camera and laser scatterer 5 monitor the foam at the same height at the same time. The laser scatterer 5 provides more data support for accurately identifying the foam size and foam quantity. The processor recognizes the foam image based on the physical parameters of the foam, which can improve the accuracy and reliability of image recognition.

[0029] In one embodiment of the present disclosure, the physical parameters of the foam include foam thickness, The processor is specifically used to: screening a first region from the foam image based on foam thickness; Image recognition is performed on the first area to obtain the bubble size and bubble quantity.

[0030] In this embodiment, the foam usually forms a foam layer on the liquid surface of the flotation machine, and the foam thickness can be understood as the vertical distance from the upper surface of the foam layer to the interface with the liquid. The foam thickness information can provide prior information for foam image recognition. For example, if the foam thickness measured by the laser scatterer 5 is within a certain range, in image processing, the foam image can be preliminarily screened based on this, and a specific area containing foam (that is, the first area) can be determined. Image recognition is performed on the first area, which can reduce the amount of calculation for subsequent image processing and improve processing efficiency.

[0031] It can be concluded from the above that this embodiment screens the image recognition area based on the foam thickness information, which can focus more on the area containing the foam, reduce the impact of background interference and noise, and help improve the accuracy of foam size and foam quantity recognition.

[0032] In one embodiment of the present disclosure, the processor is further configured to: Determine the number of convolutional layers and the size of the convolutional kernel of the convolutional neural network based on the light intensity and slurry concentration in the flotation machine; Image recognition of the first area is performed based on a convolutional neural network to obtain the foam size and foam quantity.

[0033] In this embodiment, the features in the image are automatically learned based on a convolutional neural network, so as to accurately identify the size and quantity of the bubbles.

[0034] Consider that the lighting inside the flotation machine may be uneven, with some areas brighter and some areas darker; at the same time, the light intensity inside the flotation machine may change over time or due to changes in the external environment, for example, different light intensities during the day and at night, or unstable light intensity due to equipment failure, power fluctuations, etc. All of these factors will cause the color and brightness characteristics of the foam to differ between areas.

[0035] At the same time, considering that high-concentration slurry may make the foam denser and stick to each other, it is difficult to distinguish the characteristics of a single bubble. High-concentration slurry will also reduce the contrast of the image and reduce the distinction between the foam and the background.

[0036] Therefore, this embodiment determines the number of convolution layers and the size of the convolution kernel of the convolutional neural network based on the change in light intensity and the slurry concentration in the flotation machine, which can make the convolutional neural network better adapt to different flotation environments, thereby improving the recognition accuracy of foam size and foam quantity.

[0037] Optionally, the number of convolution layers and the size of the convolution kernel of the convolutional neural network are determined based on the change in light intensity and the slurry concentration in the flotation machine, which can be described in detail as follows: determining a first adjustment parameter and a second adjustment parameter based on an amount of change in light intensity in the flotation machine; determining a third adjustment parameter and a fourth adjustment parameter based on the slurry concentration in the flotation machine; Adjust the first reference value based on the first adjustment parameter and the third adjustment parameter to obtain the number of convolution layers; The second reference value is adjusted based on the second adjustment parameter and the fourth adjustment parameter to obtain the size of the convolution kernel.

[0038] In this embodiment, the first reference value may be an initially set number of convolution layers, or a benchmark value obtained based on experience or experiments. Similarly, the second reference value may be an initial convolution kernel size setting value.

[0039] Light intensity detection sensors can be arranged around the flotation machine to measure light intensity values ​​at different time points through the light intensity detection sensors, thereby calculating the light intensity change.

[0040] Optionally, the first adjustment parameter can be determined based on the positive correlation between the change in illumination intensity and the first adjustment parameter, and the second adjustment parameter can be determined based on the negative correlation between the change in illumination intensity and the second adjustment parameter. That is, when the change in illumination intensity is large, the number of convolution layers can be increased to better capture the characteristics of the foam under different illumination conditions. A deeper network can learn a richer feature representation and improve the robustness to illumination changes. At the same time, a smaller convolution kernel, such as 3x3 or 5x5, can be used to better capture local detail features and reduce the impact of illumination changes on feature extraction. When the illumination intensity change is small, it indicates that the illumination conditions are relatively stable, and the illumination difference of the foam image at different time points is small. At this time, there is no need for an overly deep network to learn complex illumination change features. A shallower network can reduce the amount of calculation and training time, while avoiding the risk of overfitting. Therefore, the number of convolution layers can be reduced, for example, by reducing one or two convolution layers. When the illumination intensity change is small, a larger convolution kernel can capture a wider range of contextual information, which helps to extract the overall characteristics of the foam. Therefore, a larger convolution kernel can be selected, such as a 7x7 or 9x9 convolution kernel. A larger convolution kernel can smooth small illumination changes in the image to a certain extent and reduce the impact of noise on feature extraction.

[0041] Optionally, the first adjustment parameter and the second adjustment parameter may be calculated using the following formula:

[0042] in, represents the first adjustment parameter, represents the second adjustment parameter, and These are preset scale parameters. Indicates the change in light intensity.

[0043] Optionally, the third adjustment parameter can be determined based on the positive correlation between the slurry concentration and the third adjustment parameter, and the fourth adjustment parameter can be determined based on the positive correlation between the slurry concentration and the fourth adjustment parameter. That is, for high-concentration slurry, a larger convolution kernel, such as 7x7 or 9x9, can be used to better integrate the information between the bubbles and reduce the impact of the mutual adhesion of bubbles on feature extraction. At the same time, the number of convolution layers can be increased to better learn the feature representation of the bubbles and improve the robustness to high-concentration slurry. For low-concentration slurry, which often has higher fluidity, smaller convolution kernels and fewer convolution layers can be used to improve the real-time performance of the network.

[0044] For example, the third adjustment parameter and the fourth adjustment parameter may be calculated using the following formula:

[0045] in, represents the third adjustment parameter, represents the fourth adjustment parameter, and These are preset scale parameters. Indicates the slurry concentration.

[0046] On the basis of obtaining the first adjustment parameter and the third adjustment parameter, a weighted sum may be performed on the first adjustment parameter and the third adjustment parameter to obtain a fifth adjustment parameter, and the fifth adjustment parameter and the first reference value may be multiplied to obtain the number of convolution layers.

[0047] Similarly, based on the second adjustment parameter and the fourth adjustment parameter, the second adjustment parameter and the fourth adjustment parameter can be weightedly summed to obtain a sixth adjustment parameter, and the sixth adjustment parameter and the second reference value can be multiplied to obtain the size of the convolution kernel.

[0048] It can be concluded from the above that this embodiment determines the number of convolution layers and the size of the convolution kernel of the convolutional neural network based on the change in light intensity and the slurry concentration, while ensuring the feature extraction effect, reducing the amount of calculation as much as possible, and improving the calculation efficiency and the stability of the model.

[0049] In one embodiment of the present disclosure, the physical parameters of the foam include foam size distribution information. The processor is also specifically used to: Adjust the convolution kernel size of the convolutional neural network based on the bubble size distribution information.

[0050] In this embodiment, the foam size distribution information can be used to determine the size range of bubbles in the foam. If the foam size distribution is relatively concentrated, that is, most of the foams are of similar size, then a smaller convolution kernel can be selected to better capture the local features of these foams. For example, when the foam is mainly small in size, using a 3x3 convolution kernel can focus more on extracting the detailed features of small foams. If the foam size distribution is wider, including foams of various sizes, then a larger convolution kernel may be required to capture both the overall features of large foams and the local features of small foams. For example, using a 5x5 or 7x7 convolution kernel can take into account the feature extraction requirements of foams of different sizes to a certain extent.

[0051] Optionally, the foam size can be segmented according to the foam size distribution information, and the ratio of the number of foams in each segment to the total number of foams is calculated; the first foam segment with the largest size and a ratio greater than a preset first threshold is selected; the seventh adjustment parameter is determined based on the average size of the first foam segment; and the size of the convolution kernel is adjusted based on the seventh adjustment parameter.

[0052] For example, if the proportion of the number of foam segments of the largest size is greater than the preset first threshold, the seventh adjustment parameter is determined based on the average size of the foam segments of the largest size; if the proportion of the number of foam segments of the largest size is less than the preset first threshold, and the proportion of the number of foam segments ranked second is greater than the preset first threshold, the seventh adjustment parameter is determined based on the average size of the foam segments ranked second.

[0053] The seventh adjustment parameter may be determined using the following formula:

[0054] in, represents the seventh adjustment parameter, is the preset scale parameter, represents the average size of the first foam segment.

[0055] On the basis of obtaining the seventh adjustment parameter, the seventh adjustment parameter is multiplied by the size of the convolution kernel obtained in the above embodiment to obtain the adjusted size of the convolution kernel.

[0056] It can be concluded from the above that this embodiment adjusts the convolution kernel size according to the bubble size distribution, which can reduce the calculation cost and improve the response speed of the system while ensuring accuracy.

[0057] In one embodiment of the present disclosure, during the training of the convolutional neural network, the loss function selected is:

[0058] in, represents the loss function, represents the bubble area obtained by image recognition, represents the actual bubble area, Indicates area and Region The intersection and union ratio of Indicates area and Region The shape matching degree, Indicates area and Region Hausdorff distance of the contour, Indicates area The diameter of Indicates area The diameter of Indicates area and Region The maximum value of the diameter; Indicates area and Region The position distance weight, is the preset scaling factor, Indicates area The center position coordinates of Indicates area The center position coordinates of .

[0059] In this embodiment, considering that there will be overlap between bubbles, it is not comprehensive to measure the accuracy of image recognition only by the area overlap ratio, so the relative position weight and shape matching degree are introduced. and Region The center position coordinates of the two regions reflect the impact of the relative layout relationship on the overlap, because overlapping regions with close center distances have a closer overlap relationship in practical terms. The shape matching degree is obtained by calculating the Hausdorff distance between the contours of the two regions. Regions with similar shapes and overlap have a higher overlap. Finally, the above three factors are multiplied to obtain the loss function for training the convolutional neural network, which enables it to measure the region more comprehensively and accurately. and Region The overlap degree is also the accuracy of image recognition.

[0060] In one embodiment of the present disclosure, a vertical monitoring system for froth size and survival law in a flotation machine further includes: The foam stabilization pipeline 7 is arranged along the height direction of the flotation machine tank, and the bottom opening of the foam pipeline is connected to the slurry outlet of the stirring zone of the flotation machine tank.

[0061] In this embodiment, the foam stabilization pipeline 7 can reduce the flow rate of the foam in the pipeline, thereby reducing the shear force of the fluid on the foam. The life of the foam is extended and it is more stable. Furthermore, the slurry coming out of the stirring zone is in the state of the first time after the mineral and the bubble are mixed. This is the initial stage of bubble growth, the smallest shape and the highest density. The foam stabilization pipeline 7 is set at the outlet of the slurry in the stirring zone of the flotation machine tank, which can ensure that the foam enters the stabilization pipeline in the shortest time and reduce the breakage and merging of the foam.

[0062] In one embodiment of the present disclosure, a vertical monitoring system for froth size and survival law in a flotation machine further includes: The image acquisition board 6 is arranged on the inner wall of the foam stabilization pipeline 7, and the image acquisition device 4 is arranged on the outer wall of the foam stabilization pipeline 7, and the image acquisition board 6 and the image acquisition device 4 are arranged correspondingly.

[0063] In this embodiment, considering that the slurry environment is black during the foam rising process, an image acquisition board 6 is arranged on the inner wall of the foam stabilization pipeline 7. The image acquisition board 6 and the image acquisition device 4 are arranged correspondingly. The image acquisition board 6 is equivalent to a background board, which can make the edge and shape of the image taken by the image acquisition device 4 clear, which is conducive to improving the accuracy of image recognition.

[0064] In one embodiment of the present disclosure, a vertical monitoring system for froth size and survival law in a flotation machine further includes: The high-pressure cleaning pipe 3 is arranged in the foam stabilizing pipeline 7 .

[0065] In this embodiment, a high-pressure cleaning pipe 3 is added to the foam stabilization pipeline 7 to clean the foam image acquisition board 6, thereby achieving the purpose of collecting images for real-time calibration.

[0066] The above embodiments are only used to illustrate the technical solutions of the present disclosure, rather than to limit them. Although the present disclosure has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present disclosure.

Claims

1. A vertical monitoring system for froth size and survival law in a flotation machine, characterized in that: include: An image acquisition device, used for acquiring foam images in real time, wherein there are a plurality of image acquisition devices, and the plurality of image acquisition devices are arranged in sequence along the height direction of the flotation machine tank; The processor is used to identify multiple foam images, obtain foam sizes and foam quantities at multiple heights, and obtain foam sizes and survival rules based on the foam sizes and foam quantities at multiple heights.

2. A vertical monitoring system for foam size and survival law in a flotation machine as claimed in claim 1, characterized in that: Also includes: A laser scattering instrument, used for real-time monitoring of physical parameters of the foam, wherein there are multiple laser scattering instruments, and the laser scattering instruments are arranged corresponding to the image acquisition device; The processor is specifically used for: The foam image is identified based on the physical parameters of the foam to obtain the foam size and foam quantity.

3. A vertical monitoring system for foam size and survival law in a flotation machine as claimed in claim 1, characterized in that: The physical parameters of the foam include foam thickness, The processor is specifically used for: screening a first region from the foam image based on foam thickness; Image recognition is performed on the first area to obtain the foam size and foam quantity.

4. A vertical monitoring system for foam size and survival law in a flotation machine as claimed in claim 3, characterized in that: The processor is further specifically configured to: The number of convolutional layers and the size of the convolutional kernel of the convolutional neural network are determined based on the change in light intensity and the slurry concentration in the flotation machine; Image recognition is performed on the first area based on the convolutional neural network to obtain the foam size and foam quantity.

5. A vertical monitoring system for foam size and survival law in a flotation machine as claimed in claim 4, characterized in that: The processor is further specifically configured to: determining a first adjustment parameter and a second adjustment parameter based on an amount of change in light intensity in the flotation machine; determining a third adjustment parameter and a fourth adjustment parameter based on the slurry concentration in the flotation machine; Adjust the first reference value based on the first adjustment parameter and the third adjustment parameter to obtain the number of convolution layers; The second reference value is adjusted based on the second adjustment parameter and the fourth adjustment parameter to obtain the size of the convolution kernel.

6. A vertical monitoring system for foam size and survival law in a flotation machine as claimed in claim 4, characterized in that: The physical parameters of the foam include foam size distribution information, The processor is further specifically configured to: The size of the convolution kernel of the convolutional neural network is adjusted based on the bubble size distribution information.

7. A vertical monitoring system for foam size and survival law in a flotation machine as claimed in claim 1, characterized in that: Also includes: The foam stabilization pipeline is arranged along the height direction of the flotation machine tank, and the bottom opening of the foam pipeline is connected with the slurry outlet of the stirring zone of the flotation machine tank.

8. A vertical monitoring system for foam size and survival law in a flotation machine as claimed in claim 7, characterized in that: Also includes: An image acquisition board is arranged on the inner wall of the foam stabilization pipeline, and the image acquisition device is arranged on the outer wall of the foam stabilization pipeline, and the image acquisition board and the image acquisition device are arranged correspondingly.

9. A vertical monitoring system for foam size and survival law in a flotation machine as claimed in claim 7, characterized in that: Also includes: The high-pressure cleaning pipe is arranged in the foam stabilizing pipeline.

10. A method for vertically monitoring the size and survival of foam in a flotation machine, characterized in that: include: Collecting foam images in real time based on multiple image acquisition devices, wherein the multiple image acquisition devices are arranged in sequence along the height direction of the flotation machine tank; Multiple foam images are identified to obtain foam sizes and foam quantities at multiple heights, and foam sizes and survival rules are obtained based on the foam sizes and foam quantities at multiple heights.

Images