CN117943213A - Real-time monitoring and early warning system and method for micro-bubble flotation machine - Google Patents
Real-time monitoring and early warning system and method for micro-bubble flotation machine Download PDFInfo
- Publication number
- CN117943213A CN117943213A CN202410355884.6A CN202410355884A CN117943213A CN 117943213 A CN117943213 A CN 117943213A CN 202410355884 A CN202410355884 A CN 202410355884A CN 117943213 A CN117943213 A CN 117943213A
- Authority
- CN
- China
- Prior art keywords
- pulp
- sequence
- real
- training
- time monitoring
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000012544 monitoring process Methods 0.000 title claims abstract description 148
- 238000005188 flotation Methods 0.000 title claims abstract description 93
- 238000000034 method Methods 0.000 title claims abstract description 75
- 239000013598 vector Substances 0.000 claims abstract description 78
- 230000004927 fusion Effects 0.000 claims abstract description 62
- 239000007788 liquid Substances 0.000 claims abstract description 43
- 238000012549 training Methods 0.000 claims description 115
- 238000001914 filtration Methods 0.000 claims description 53
- 230000008569 process Effects 0.000 claims description 22
- 238000009826 distribution Methods 0.000 claims description 16
- 238000012545 processing Methods 0.000 claims description 12
- 230000002146 bilateral effect Effects 0.000 claims description 11
- 230000000903 blocking effect Effects 0.000 claims description 11
- 238000005457 optimization Methods 0.000 claims description 10
- 230000001419 dependent effect Effects 0.000 claims 2
- 239000006260 foam Substances 0.000 description 12
- 230000006870 function Effects 0.000 description 12
- 238000010586 diagram Methods 0.000 description 11
- 238000003860 storage Methods 0.000 description 11
- 229910052500 inorganic mineral Inorganic materials 0.000 description 9
- 239000011707 mineral Substances 0.000 description 9
- 230000008859 change Effects 0.000 description 8
- 239000012141 concentrate Substances 0.000 description 8
- 238000004590 computer program Methods 0.000 description 4
- 238000004140 cleaning Methods 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 238000009499 grossing Methods 0.000 description 3
- 239000002245 particle Substances 0.000 description 3
- 238000011084 recovery Methods 0.000 description 3
- 230000001105 regulatory effect Effects 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- 239000002002 slurry Substances 0.000 description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000033558 biomineral tissue development Effects 0.000 description 2
- 230000001427 coherent effect Effects 0.000 description 2
- 238000011010 flushing procedure Methods 0.000 description 2
- 230000003116 impacting effect Effects 0.000 description 2
- 238000007477 logistic regression Methods 0.000 description 2
- 238000002156 mixing Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000001902 propagating effect Effects 0.000 description 2
- 238000000926 separation method Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000036632 reaction speed Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B03—SEPARATION OF SOLID MATERIALS USING LIQUIDS OR USING PNEUMATIC TABLES OR JIGS; MAGNETIC OR ELECTROSTATIC SEPARATION OF SOLID MATERIALS FROM SOLID MATERIALS OR FLUIDS; SEPARATION BY HIGH-VOLTAGE ELECTRIC FIELDS
- B03D—FLOTATION; DIFFERENTIAL SEDIMENTATION
- B03D1/00—Flotation
- B03D1/02—Froth-flotation processes
- B03D1/028—Control and monitoring of flotation processes; computer models therefor
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B03—SEPARATION OF SOLID MATERIALS USING LIQUIDS OR USING PNEUMATIC TABLES OR JIGS; MAGNETIC OR ELECTROSTATIC SEPARATION OF SOLID MATERIALS FROM SOLID MATERIALS OR FLUIDS; SEPARATION BY HIGH-VOLTAGE ELECTRIC FIELDS
- B03D—FLOTATION; DIFFERENTIAL SEDIMENTATION
- B03D1/00—Flotation
- B03D1/14—Flotation machines
- B03D1/24—Pneumatic
- B03D1/26—Air lift machines
Landscapes
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biotechnology (AREA)
- General Engineering & Computer Science (AREA)
- Image Analysis (AREA)
Abstract
A real-time monitoring and early warning system and method for a microbubble flotation machine are disclosed. Firstly, acquiring a sequence of pulp real-time monitoring images acquired by a camera, then extracting multi-dimensional pulp characteristics of the sequence of pulp real-time monitoring images to obtain a sequence of multi-dimensional pulp characteristic fusion images, then extracting context correlation characteristics of the sequence of multi-dimensional pulp characteristic fusion images to obtain a pulp time sequence semantic context correlation characteristic vector, and finally, determining whether to send out target liquid level adjustment early warning based on the pulp time sequence semantic context correlation characteristic vector. Thus, the real-time performance of the target liquid level adjustment can be improved.
Description
Technical Field
The application relates to the field of micro-bubble flotation machines, and more particularly relates to a real-time monitoring and early warning system and method of a micro-bubble flotation machine.
Background
When the micro-bubble flotation machine is operated, ore feeding pulp is pumped into a high-pressure ore feeding pipeline by a high-pressure pump, and the pressure provided by the high-pressure pump forms high-speed jet flow at an outlet. The pulp jet flow impacting downwards at high speed is mixed with air provided by the self-priming air charging valve in the lower flushing pipe, and the mineralization process of minerals is completed. Mineralized bubbles carrying minerals are extruded at the bottom of the down-wash pipe, the bubbles carrying concentrate automatically rise, and tailing pulp is discharged downwards. The whole flotation process is completed in the lower wash pipe, concentrate and tailing separation is completed in the outer flotation machine barrel of the lower wash pipe, and all working areas are independent and do not interfere with each other.
Wherein, in the flotation process, the adjustment and control of the target liquid level affects the flotation efficiency and the product quality. The target liquid level refers to the liquid level in the flotation tank. The stability and enrichment of the flotation froth are different due to the different adhesion capacities of the different mineral particles to the bubbles. If the target liquid level is too high, excessive flotation froth is caused, resulting in a decrease in recovery efficiency of mineral particles, and also an increase in energy consumption and chemical consumption. If the target liquid level is too low, this results in too little flotation froth, resulting in poor separation of mineral particles and also in reduced throughput of the plant.
The existing adjustment mode is to manually observe the state of flotation foam and manually control according to experience or rules. This approach has the following problems: firstly, manual observation is influenced by subjective factors, and the limit of the amount of flotation foam is difficult to accurately distinguish; secondly, the manual control is limited by the skill level and the reaction speed of operators, and the response to the change of the flotation process is difficult in time.
Therefore, a real-time monitoring and early warning scheme of the micro-bubble flotation machine is expected.
Disclosure of Invention
In view of the above, the application provides a real-time monitoring and early warning system and method of a micro-bubble flotation machine, which can collect a sequence of pulp real-time monitoring images by using a camera, and excavate time sequence state change of pulp by an image processing technology based on deep learning, thereby realizing real-time monitoring of flotation foam, and simultaneously carrying out intelligent judgment on target liquid level adjustment and early warning based on the sequence.
According to one aspect of the application, there is provided a real-time monitoring and early warning method of a micro-bubble flotation machine, comprising:
acquiring a sequence of ore pulp real-time monitoring images acquired by a camera;
Extracting the multidimensional pulp characteristics of the sequence of the pulp real-time monitoring images to obtain a sequence of multidimensional pulp characteristic fusion images;
extracting context associated features of the sequence of the multidimensional pulp feature fusion images to obtain pulp time sequence semantic context associated feature vectors; and
And determining whether to send out a target liquid level adjustment early warning based on the ore pulp time sequence semantic context association feature vector.
In the above method for real-time monitoring and early warning of a micro-bubble flotation machine, extracting the multi-dimensional pulp characteristics of the sequence of pulp real-time monitoring images to obtain a sequence of multi-dimensional pulp characteristic fusion images, including:
calculating gradient direction histograms of each ore pulp real-time monitoring image in the sequence of the ore pulp real-time monitoring images to obtain a sequence of edge information characteristics;
carrying out multistage filtering treatment on each ore pulp real-time monitoring image in the ore pulp real-time monitoring image sequence to obtain a characteristic sequence after multistage filtering treatment;
After each ore pulp real-time monitoring image in the sequence of the ore pulp real-time monitoring images is subjected to blocking treatment to obtain a plurality of ore pulp image blocks, calculating the gray maximum value of each ore pulp image block to obtain a sequence of local gray maximum value characteristics; and
And fusing the sequence of the edge information features, the sequence of the features subjected to the multistage filtering treatment and the sequence of the local gray maximum features to obtain the sequence of the multi-dimensional ore pulp feature fused image.
In the above method for real-time monitoring and early warning of a microbubble flotation machine, calculating a gradient direction histogram of each pulp real-time monitoring image in the sequence of pulp real-time monitoring images to obtain a sequence of edge information features, including:
uniformly dividing the ore pulp real-time monitoring image to obtain a plurality of cell spaces;
Calculating gradients of pixel points in each cell space in the cell spaces, and generating a plurality of cell direction gradient histograms according to gradient distribution; and
The gradient direction histogram is generated based on the plurality of cell direction gradient histograms.
In the above real-time monitoring and early warning method of the micro-bubble flotation machine, the multistage filtering treatment comprises gaussian filtering and bilateral filtering.
In the above method for real-time monitoring and early warning of a micro-bubble flotation machine, extracting context-related features of the sequence of multi-dimensional pulp feature fusion images to obtain a pulp time sequence semantic context-related feature vector, including:
And passing the sequence of the multidimensional pulp characteristic fusion image through a pulp time sequence semantic characteristic extractor based on a converter to obtain the pulp time sequence semantic context associated characteristic vector.
In the above method for real-time monitoring and early warning of a micro-bubble flotation machine, passing the sequence of the multi-dimensional pulp feature fusion image through a pulp time sequence semantic feature extractor based on a converter to obtain the pulp time sequence semantic context associated feature vector, including:
image blocking is carried out on the sequence of the multi-dimensional pulp characteristic fusion image respectively to obtain a sequence of a plurality of multi-dimensional pulp characteristic fusion image blocks;
Embedding each multi-dimensional pulp characteristic fusion image block in the sequence of the multi-dimensional pulp characteristic fusion image blocks by using an image block embedding layer to obtain a sequence of embedding vectors of the multi-dimensional pulp characteristic fusion image blocks; and
And enabling the sequence of the embedded vectors of the multi-dimensional pulp feature fusion image blocks to pass through the pulp time sequence semantic feature extractor based on the converter to obtain the pulp time sequence semantic context associated feature vector.
In the above method for real-time monitoring and early warning of a micro-bubble flotation machine, determining whether to send out a target liquid level adjustment early warning based on the pulp timing semantic context correlation feature vector includes:
and the ore pulp time sequence semantic context associated feature vector is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether a target liquid level adjustment early warning is sent out.
In the above method for monitoring and early warning of a microbubble flotation machine in real time, the method further comprises the training steps of: training the converter-based pulp timing semantic feature extractor and the classifier.
In the above method for real-time monitoring and early warning of a micro-bubble flotation machine, the training step includes:
acquiring training data, wherein the training data comprises a sequence of training pulp real-time monitoring images and a true value of whether a target liquid level adjustment early warning is sent out;
calculating a gradient direction histogram of each training pulp real-time monitoring image in the training pulp real-time monitoring image sequence to obtain a training edge information characteristic sequence;
Carrying out multistage filtering treatment on each training pulp real-time monitoring image in the training pulp real-time monitoring image sequence to obtain a characteristic sequence after training multistage filtering treatment, wherein the multistage filtering treatment comprises Gaussian filtering and bilateral filtering;
After each training pulp real-time monitoring image in the training pulp real-time monitoring image sequence is subjected to blocking processing to obtain a plurality of training pulp image blocks, calculating the gray maximum value of each training pulp image block to obtain a training local gray maximum value characteristic sequence;
Fusing the sequence of the training edge information characteristics, the sequence of the characteristics after the training multi-level filtering treatment and the sequence of the training local gray maximum characteristics to obtain a sequence of a training multi-dimensional ore pulp characteristic fusion image;
Passing the sequence of training multidimensional pulp feature fusion images through the transducer-based pulp time sequence semantic feature extractor to obtain training pulp time sequence semantic context associated feature vectors;
clustering optimization is carried out on the training ore pulp time sequence semantic context associated feature vectors so as to obtain optimized training ore pulp time sequence semantic context associated feature vectors;
the optimized training ore pulp time sequence semantic context association feature vector passes through the classifier to obtain a classification loss function value; and
Training the converter-based pulp timing semantic feature extractor and the classifier with the classification loss function values.
According to another aspect of the application, a real-time monitoring and early warning system of a micro-bubble flotation machine is provided, wherein the real-time monitoring and early warning system of the micro-bubble flotation machine operates according to the real-time monitoring and early warning method of the micro-bubble flotation machine.
In the application, firstly, a sequence of pulp real-time monitoring images collected by a camera is obtained, then, the multi-dimensional pulp characteristics of the sequence of pulp real-time monitoring images are extracted to obtain a sequence of multi-dimensional pulp characteristic fusion images, then, the context correlation characteristics of the sequence of multi-dimensional pulp characteristic fusion images are extracted to obtain a pulp time sequence semantic context correlation characteristic vector, and finally, whether a target liquid level adjustment early warning is sent out is determined based on the pulp time sequence semantic context correlation characteristic vector. Thus, the real-time performance of the target liquid level adjustment can be improved.
Other features and aspects of the present application will become apparent from the following detailed description of the application with reference to the accompanying drawings.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate exemplary embodiments, features and aspects of the application and together with the description, serve to explain the principles of the application.
Fig. 1 shows a flow chart of a real-time monitoring and early warning method of a micro-bubble flotation machine according to an embodiment of the present application.
Fig. 2 shows a schematic architecture diagram of a real-time monitoring and early warning method of a micro-bubble flotation machine according to an embodiment of the present application.
Fig. 3 shows a flow chart of substep S120 of the real-time monitoring and early warning method of the micro-bubble flotation machine according to an embodiment of the present application.
Fig. 4 shows a schematic structural diagram of a real-time monitoring and early warning system of a micro-bubble flotation machine according to an embodiment of the present application.
Fig. 5 shows a block diagram of a system of a real-time monitoring and early warning system of a micro-bubble flotation machine according to an embodiment of the present application.
Fig. 6 shows an application scenario diagram of a real-time monitoring and early warning method of a micro-bubble flotation machine according to an embodiment of the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some, but not all embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are also within the scope of the application.
As used in the specification and in the claims, the terms "a," "an," "the," and/or "the" are not specific to a singular, but may include a plurality, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.
Various exemplary embodiments, features and aspects of the application will be described in detail below with reference to the drawings. In the drawings, like reference numbers indicate identical or functionally similar elements. Although various aspects of the embodiments are illustrated in the accompanying drawings, the drawings are not necessarily drawn to scale unless specifically indicated.
In addition, numerous specific details are set forth in the following description in order to provide a better illustration of the application. It will be understood by those skilled in the art that the present application may be practiced without some of these specific details. In some instances, well known methods, procedures, components, and circuits have not been described in detail so as not to obscure the present application.
Aiming at the technical problems, the technical conception of the application is as follows: the sequence of the real-time monitoring images of the ore pulp is collected by the camera, and the time sequence state change of the ore pulp is mined by the image processing technology based on deep learning, so that the real-time monitoring of the flotation foam is realized, meanwhile, the intelligent judgment of the target liquid level adjustment early warning can be carried out based on the sequence, the change of the flotation process is quickly reflected, and the real-time performance of the target liquid level adjustment is improved.
Based on this, fig. 1 shows a flow chart of a real-time monitoring and early warning method of a micro-bubble flotation machine according to an embodiment of the present application. Fig. 2 shows a schematic architecture diagram of a real-time monitoring and early warning method of a micro-bubble flotation machine according to an embodiment of the present application. As shown in fig. 1 and 2, the real-time monitoring and early warning method of the micro-bubble flotation machine according to the embodiment of the application comprises the following steps: s110, acquiring a sequence of ore pulp real-time monitoring images acquired by a camera; s120, extracting the multidimensional pulp characteristics of the sequence of the pulp real-time monitoring images to obtain a sequence of multidimensional pulp characteristic fusion images; s130, extracting context associated features of the sequence of the multi-dimensional pulp feature fusion images to obtain pulp time sequence semantic context associated feature vectors; and S140, determining whether to send out a target liquid level adjustment early warning or not based on the ore pulp time sequence semantic context correlation feature vector.
It should be appreciated that the purpose of step S110 is to acquire a real-time monitoring image sequence of the slurry via a camera for subsequent processing and analysis. In step S120, multi-dimensional pulp characteristics, such as color, texture, shape, etc., are extracted from the sequence of pulp real-time monitoring images, and then these characteristics are fused together to form a sequence of multi-dimensional pulp characteristic fusion images. In step S130, context-related features are extracted from the multi-dimensional pulp feature fusion image sequence, and these features can capture timing information and semantic context association relationship of pulp in the image sequence, and the extracted features are organized into pulp timing semantic context-associated feature vectors. Step S140 judges whether the current pulp liquid level needs to be adjusted or not by utilizing the pulp time sequence semantic context associated feature vector, and decides whether to send out the early warning of target liquid level adjustment or not according to the judging result. In general, the steps form a real-time monitoring and early warning method of the micro-bubble flotation machine, and the real-time monitoring and adjustment of the pulp liquid level are realized by collecting pulp images, extracting multidimensional features, analyzing context associated features and early warning judgment.
Specifically, in the technical scheme of the application, firstly, a sequence of ore pulp real-time monitoring images acquired by a camera is acquired. Here, the sequence of acquiring the pulp real-time monitoring images acquired by the camera is to realize monitoring and analysis of the flotation froth state in the operation process of the microbubble flotation machine. The form, distribution and change conditions of the foam in the flotation process can be obtained by continuously collecting and recording the sequence of the ore pulp real-time monitoring images, so that the comprehensive observation of the flotation foam state is provided. Specifically, the real-time monitoring image collected by the camera can intuitively display the flotation foam condition in the ore pulp, including the size, shape, density and the like of the foam. Such visual information helps determine whether the target level needs to be adjusted. In addition, the sequence of the ore pulp real-time monitoring images is acquired in real time, so that the change of the foam state in the flotation process can be reflected in time. This is very important for real-time adjustment of the target liquid level, enabling a more accurate control of the flotation process.
Then, calculating gradient direction histograms of each ore pulp real-time monitoring image in the sequence of the ore pulp real-time monitoring images to obtain a sequence of edge information characteristics; carrying out multistage filtering treatment on each ore pulp real-time monitoring image in the sequence of the ore pulp real-time monitoring images to obtain a characteristic sequence after the multistage filtering treatment, wherein the multistage filtering treatment comprises Gaussian filtering and bilateral filtering; and meanwhile, after each ore pulp real-time monitoring image in the sequence of the ore pulp real-time monitoring images is subjected to blocking treatment to obtain a plurality of ore pulp image blocks, calculating the gray maximum value of each ore pulp image block so as to obtain a sequence of local gray maximum value characteristics.
Here, calculating the gradient direction histogram of each pulp real-time monitoring image in the sequence of pulp real-time monitoring images may be used to extract the edge information features. The edge is an area with obvious change of pixel gray level in the image, and for the ore pulp real-time monitoring image, the edge information can provide a boundary between foam and background to help judge the position and form of flotation foam. In addition, the noise information in the image can be removed by carrying out multistage filtering treatment on each ore pulp real-time monitoring image in the sequence of the ore pulp real-time monitoring images, so that smoother and stable characteristics are extracted. Specifically, gaussian Filter (Gaussian Filter) is a linear smoothing Filter that achieves filtering by weighted averaging of a neighborhood around each pixel point in an image. The gaussian filter gives different weights according to the distance from the center pixel, and the closer the distance is, the larger the pixel weight is, and the farther the distance is, the smaller the pixel weight is. The filter can effectively reduce high-frequency noise in the image, and meanwhile, the integral structure of the image is reserved. While bilateral filtering (Bilateral Filter) is also a nonlinear filter, it also preserves edge information while smoothing the image, as compared to gaussian filters. The bilateral filter takes into account the difference between the spatial distance between pixels and the pixel value, by smoothing in the region where the difference in pixel value is small, to preserve the sharpness of the edge. This allows the bilateral filter to smooth the image while maintaining good edge detail. By carrying out multistage filtering treatment on each ore pulp real-time monitoring image, noise information in the ore pulp real-time monitoring image can be gradually reduced, and smoother image characteristics are obtained. Secondly, the sequence of the ore pulp real-time monitoring images is subjected to blocking treatment, and the gray maximum value of each ore pulp image block is calculated, so that the local gray maximum value characteristic can be extracted. This feature helps to capture the brightness variation and regional shading differences in the pulp real-time monitoring image, thereby providing local information about the flotation froth, and thus knowing the evolution and trend of the froth region.
And then, fusing the sequence of the edge information features, the sequence of the features after the multistage filtering treatment and the sequence of the local gray maximum features to obtain a sequence of the multi-dimensional ore pulp feature fused image. That is, feature information of different aspects is comprehensively utilized, namely, edge information features can capture boundary and contour information of an object, the features after multistage filtering processing can provide smooth and stable image features, and local gray maximum features can reflect brightness changes of areas so as to provide more comprehensive and accurate image description.
Accordingly, in step S120, as shown in fig. 3, the extracting the multi-dimensional pulp characteristics of the sequence of pulp real-time monitoring images to obtain a sequence of multi-dimensional pulp characteristic fusion images includes: s121, calculating gradient direction histograms of each ore pulp real-time monitoring image in the sequence of the ore pulp real-time monitoring images to obtain a sequence of edge information characteristics; s122, carrying out multistage filtering treatment on each ore pulp real-time monitoring image in the sequence of the ore pulp real-time monitoring images to obtain a characteristic sequence after the multistage filtering treatment; s123, after each ore pulp real-time monitoring image in the sequence of the ore pulp real-time monitoring images is subjected to blocking treatment to obtain a plurality of ore pulp image blocks, calculating the gray maximum value of each ore pulp image block so as to obtain a sequence of local gray maximum value characteristics; and S124, fusing the sequence of the edge information features, the sequence of the features after the multistage filtering treatment and the sequence of the local gray maximum features to obtain the sequence of the multi-dimensional ore pulp feature fused image.
It should be understood that the purpose of step S121 is to extract edge information features by calculating a gradient direction histogram for each image in the image sequence, where the gradient direction histogram may describe the direction distribution of the edges in the image, thereby capturing edge information in the image. In step S122, a multi-stage filtering process is performed on each image in the image sequence, where the multi-stage filtering process may extract features of different frequencies through filters of different scales, so as to obtain a feature sequence after the multi-stage filtering process. The purpose of step S123 is to block each image in the image sequence, calculate the gray maximum value of each image block, and capture the gray variation of different areas in the image by calculating the local gray maximum value feature. In step S124, the edge information feature sequence, the feature sequence after the multi-stage filtering process and the local gray maximum feature sequence obtained in the previous step are fused to form a multi-dimensional ore pulp feature fusion image sequence, and the fusion of these features can provide more comprehensive and rich ore pulp feature information for subsequent analysis and processing. The method aims at extracting multidimensional features of the ore pulp real-time monitoring image sequence, and fusing the features into a sequence of multidimensional ore pulp feature fusion images so as to facilitate subsequent context-related feature extraction and liquid level adjustment early warning judgment.
In step S121, calculating a gradient direction histogram of each pulp real-time monitoring image in the sequence of pulp real-time monitoring images to obtain a sequence of edge information features, including: uniformly dividing the ore pulp real-time monitoring image to obtain a plurality of cell spaces; calculating gradients of pixel points in each cell space in the cell spaces, and generating a plurality of cell direction gradient histograms according to gradient distribution; and generating the gradient direction histogram based on the plurality of cell direction gradient histograms.
Wherein, in step S122, the multi-stage filtering process includes gaussian filtering and bilateral filtering.
It will be appreciated that the conditions of the slurry are time-varying, and that the conditions at a previous time will generally affect the conditions at a subsequent time. And obviously, the sequence of the multi-dimensional pulp characteristic fusion images lacks of a time sequence association relationship of the context among the multi-dimensional pulp characteristic fusion images, and lacks of comprehensive and dynamic expression and description of pulp characteristics. Therefore, in the technical scheme of the application, the sequence of the multi-dimensional pulp characteristic fusion image is used for establishing time sequence context association between different key frames in the sequence of the multi-dimensional pulp characteristic fusion image through a pulp time sequence semantic characteristic extractor based on a converter, and dynamic change of flotation foam is captured, so that a pulp time sequence semantic context association characteristic vector is obtained.
Accordingly, in step S130, extracting the context-related features of the sequence of multi-dimensional pulp feature fusion images to obtain a pulp timing semantic context-related feature vector, including: and passing the sequence of the multidimensional pulp characteristic fusion image through a pulp time sequence semantic characteristic extractor based on a converter to obtain the pulp time sequence semantic context associated characteristic vector.
It should be noted that the converter refers to a model or a component for converting input data from one form or representation to another form or representation, and in the step of extracting the pulp time sequential semantic context associated feature vector, the pulp time sequential semantic feature extractor based on the converter may be understood as a model or a component capable of converting the multi-dimensional pulp feature fusion image sequence into the pulp time sequential semantic context associated feature vector. Specifically, the converter can adopt a machine learning or deep learning method, map the multidimensional pulp characteristic fusion image sequence into a semantic space through the inherent characteristics and the association relation of learning data, and extract time sequence information and semantic context association characteristics in the pulp image sequence in the mapping process so as to obtain a pulp time sequence semantic context association characteristic vector. Common converters include convolutional neural networks (Convolutional Neural Networks, CNN), cyclic neural networks (Recurrent Neural Networks, RNN), self-encoders (Autoencoders), transducers, and the like.
Specifically, the sequence of the multidimensional pulp feature fusion image is passed through a pulp time sequence semantic feature extractor based on a converter to obtain the pulp time sequence semantic context associated feature vector, which comprises the following steps: image blocking is carried out on the sequence of the multi-dimensional pulp characteristic fusion image respectively to obtain a sequence of a plurality of multi-dimensional pulp characteristic fusion image blocks; embedding each multi-dimensional pulp characteristic fusion image block in the sequence of the multi-dimensional pulp characteristic fusion image blocks by using an image block embedding layer to obtain a sequence of embedding vectors of the multi-dimensional pulp characteristic fusion image blocks; and enabling the sequence of the multi-dimensional ore pulp feature fusion image block embedded vectors to pass through the ore pulp time sequence semantic feature extractor based on the converter to obtain the ore pulp time sequence semantic context associated feature vector.
Further, the ore pulp time sequence semantic context associated feature vector is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether to send out target liquid level adjustment early warning.
Accordingly, in step S140, determining whether to issue a target liquid level adjustment pre-warning based on the pulp timing semantic context associated feature vector includes: and the ore pulp time sequence semantic context associated feature vector is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether a target liquid level adjustment early warning is sent out.
Specifically, the ore pulp time sequence semantic context associated feature vector is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether to send out a target liquid level adjustment early warning or not, and the method comprises the following steps: performing full-connection coding on the ore pulp time sequence semantic context associated feature vectors by using a full-connection layer of the classifier to obtain coded classification feature vectors; and inputting the coding classification feature vector into a Softmax classification function of the classifier to obtain the classification result.
That is, in the technical solution of the present application, the tag of the classifier includes sending out a target liquid level adjustment pre-warning (first tag) and not sending out a target liquid level adjustment pre-warning (second tag), where the classifier determines, through a soft maximum function, to which classification tag the pulp timing semantic context associated feature vector belongs. It should be noted that the first tag p1 and the second tag p2 do not include a manually set concept, and in fact, during the training process, the computer model does not have a concept of "whether to send out the target level adjustment warning", which is just two kinds of classification tags, and the probability that the output feature is the sum of the two classification tags sign, that is, p1 and p2 is one. Therefore, the classification result of whether the target liquid level adjustment early warning is sent out is actually converted into the classified probability distribution conforming to the two classifications of the natural law through classifying the labels, and the physical meaning of the natural probability distribution of the labels is essentially used instead of the language text meaning of whether the target liquid level adjustment early warning is sent out.
It should be appreciated that the role of the classifier is to learn the classification rules and classifier using a given class, known training data, and then classify (or predict) the unknown data. Logistic regression (logistics), SVM, etc. are commonly used to solve the classification problem, and for multi-classification problems (multi-class classification), logistic regression or SVM can be used as well, but multiple bi-classifications are required to compose multiple classifications, but this is error-prone and inefficient, and the commonly used multi-classification method is the Softmax classification function.
Further, in the technical scheme of the application, the real-time monitoring and early warning method of the micro-bubble flotation machine further comprises the training steps of: training the converter-based pulp timing semantic feature extractor and the classifier. It should be understood that the training step plays a key role in the real-time monitoring and early warning method of the micro-bubble flotation machine, and through training the converter-based ore pulp time sequence semantic feature extractor and classifier, the ore pulp data can be effectively processed and analyzed. Training the converter-based pulp timing semantic feature extractor is aimed at learning timing and semantic context-related features in pulp data, by training the extractor can extract useful features from an input multi-dimensional pulp feature fusion image sequence, capture timing relationships and semantic associations in pulp data, and can be used for subsequent classification and prediction tasks. The purpose of training the classifier is to classify and predict pulp data according to extracted time sequence semantic features, the classifier can be set according to specific problems, for example, whether the liquid level of pulp needs to be adjusted for early warning is judged, and the classifier is trained to enable the classifier to have the capability of accurately classifying and predicting the pulp data, so that the operation of the micro-bubble flotation machine is monitored and controlled. The key of the training step is to train a model by using the labeled training data, and by providing pulp data of known classes or labels as training samples, the model can be represented by learning the relevance and characteristics among the samples, thereby improving the generalization capability of unknown data. The training step can adjust the parameters of the model through an optimization algorithm and a loss function, so that the model can better adapt to task requirements. In summary, the training step is very important for a real-time monitoring and early warning method of the micro-bubble flotation machine, and the method can enable the ore pulp time sequence semantic feature extractor and the classifier based on the converter to have the capability of processing and analyzing ore pulp data, so that the operation of the micro-bubble flotation machine is monitored and controlled.
Wherein, in one example, the training step comprises: acquiring training data, wherein the training data comprises a sequence of training pulp real-time monitoring images and a true value of whether a target liquid level adjustment early warning is sent out; calculating a gradient direction histogram of each training pulp real-time monitoring image in the training pulp real-time monitoring image sequence to obtain a training edge information characteristic sequence; carrying out multistage filtering treatment on each training pulp real-time monitoring image in the training pulp real-time monitoring image sequence to obtain a characteristic sequence after training multistage filtering treatment, wherein the multistage filtering treatment comprises Gaussian filtering and bilateral filtering; after each training pulp real-time monitoring image in the training pulp real-time monitoring image sequence is subjected to blocking processing to obtain a plurality of training pulp image blocks, calculating the gray maximum value of each training pulp image block to obtain a training local gray maximum value characteristic sequence; fusing the sequence of the training edge information characteristics, the sequence of the characteristics after the training multi-level filtering treatment and the sequence of the training local gray maximum characteristics to obtain a sequence of a training multi-dimensional ore pulp characteristic fusion image; passing the sequence of training multidimensional pulp feature fusion images through the transducer-based pulp time sequence semantic feature extractor to obtain training pulp time sequence semantic context associated feature vectors; clustering optimization is carried out on the training ore pulp time sequence semantic context associated feature vectors so as to obtain optimized training ore pulp time sequence semantic context associated feature vectors; the optimized training ore pulp time sequence semantic context association feature vector passes through the classifier to obtain a classification loss function value; and training the converter-based pulp timing semantic feature extractor and the classifier with the classification loss function values.
In the technical scheme of the application, when the sequence of the training multi-dimensional pulp feature fusion image passes through the pulp time sequence semantic feature extractor based on the converter, corresponding time sequence semantic feature encoding is carried out based on the respective feature context relations of the sequence of the edge information features, the sequence of the features after the multi-stage filtering processing and the sequence of the local gray maximum features in the sequence of the training multi-dimensional pulp feature fusion image, so that the feature distribution information saliency of the respective feature distribution information based on the specific feature time sequence distribution is influenced inevitably while the feature representation relations among the edge information features, the features after the multi-stage filtering processing and the local gray maximum features are promoted, and the training pulp time sequence semantic context relation feature vectors are difficult to focus on the local salient feature distribution stably in the training process, that is, and have the salient local feature distribution discreteness.
Therefore, when the training pulp timing sequence semantic context associated feature vector is used as a whole to be classified by the classifier, convergence of the class regression by the classifier facing to the preset class probability is difficult due to the local feature distribution discreteness of the pulp timing sequence semantic context associated feature vector, so that the training speed of the classifier and the accuracy of a finally obtained classification result are affected. Based on the above, the application performs cluster optimization on the training pulp time sequence semantic context associated feature vectors, namely, firstly, clusters each feature value of the training pulp time sequence semantic context associated feature vectors, for example, clusters based on the distance between feature values, and then performs optimization based on the clustered feature class inner and class outer appearance.
Accordingly, in one example, performing cluster optimization on the training pulp time sequence semantic context associated feature vector to obtain an optimized training pulp time sequence semantic context associated feature vector, including: clustering each characteristic value of the training ore pulp time sequence semantic context associated characteristic vector; optimizing based on the clustered feature intra-class and inter-class features to obtain optimized training pulp time sequence semantic context associated feature vectors.
Specifically, optimizing based on the clustered feature intra-class and inter-class appearance to obtain optimized training pulp time sequence semantic context associated feature vectors, including: optimizing based on the clustered feature class inner and class outer appearances by using the following clustering optimization formula to obtain optimized training ore pulp time sequence semantic context associated feature vectors; wherein, the clustering optimization formula:
Wherein, Is each characteristic value of the training pulp time sequence semantic context associated characteristic vector,/>Is the number of feature sets corresponding to the training pulp time sequence semantic context associated feature vectors,/>Is the number of cluster features,/>Representing a set of clustering features,/>Is each characteristic value of the optimized training pulp time sequence semantic context associated characteristic vector.
Specifically, the intra-class feature and the extra-class feature of the training pulp time sequence semantic context associated feature vector serve as different example roles to perform cluster proportion distribution-based class example description, and cluster response histories based on intra-class and extra-class dynamic contexts are introduced to keep a coordinated global view of the intra-class distribution and the extra-class distribution of the overall feature of the training pulp time sequence semantic context associated feature vector, so that the feature clustering operation of optimization of the training pulp time sequence semantic context associated feature vector can maintain coherent and consistent responses of the intra-class feature and the extra-class feature, the regression convergence paths based on feature clustering keep coherent and consistent in the class regression process, and the convergence effect of the training pulp time sequence semantic context associated feature vector facing to the preset class probability is improved, so that the training speed of a classifier and the accuracy of a classification result are improved.
In summary, the real-time monitoring and early warning method of the microbubble flotation machine based on the embodiment of the application can improve the real-time performance of target liquid level adjustment.
Further, the application also provides a real-time monitoring and early warning system of the micro-bubble flotation machine, which operates in the real-time monitoring and early warning method of the micro-bubble flotation machine.
Specifically, as shown in fig. 4, the micro-bubble flotation machine comprises a device main body, and a mineral feeding circulating pump 1, a concentrate outlet 2, a cleaning water regulating valve 3, a down-wash pipe regulating valve 4, a mineral feeding pressure gauge 5, a tailings outlet 6, an air flow meter 7, a gas flow regulating valve 8, a tailings control valve 9, a vacuum pressure gauge 10 and a cleaning water interface 11 which are arranged on the device main body.
When the micro-bubble flotation machine is operated, ore feeding pulp is pumped into a high-pressure ore feeding pipeline by a high-pressure pump, and the pressure provided by the high-pressure pump forms high-speed jet flow at an outlet. The pulp jet flow impacting downwards at high speed is mixed with air provided by the self-priming air charging valve in the lower flushing pipe, and the mineralization process of minerals is completed. Mineralized bubbles carrying minerals are extruded at the bottom of the down-wash pipe, the bubbles carrying concentrate automatically rise, and tailing pulp is discharged downwards. The whole flotation process is finished in the lower wash pipe, concentrate and tailings are separated in the outer flotation machine barrel of the lower wash pipe, all working areas are independent of each other and do not interfere with each other, and the conditions can be optimized respectively, so that the flotation efficiency in the lower wash pipe is guaranteed to be optimal. Each working area of the microbubble flotation machine comprises a bubble generation area, a pulp and bubble mixing area, a concentrate recovery area, a tailing discharge area and the like which are relatively independent, and a certain feasibility is provided for optimizing working conditions of each working area. The average diameter of mineralized bubbles generated by mixing the slurry in the micro-bubble flotation machine with air after passing through the jet flow is about 0.3mm, which is about one tenth of the bubble diameter of a conventional flotation machine and a flotation column. When the ore is floated, the number of mineralized bubbles generated by the former is more than 50 times of that generated by the latter under the same condition, which is one of key factors for improving the ore dressing recovery rate and the ore concentrate grade simultaneously by the micro-bubble flotation machine.
Accordingly, the method of operating a microbubble flotation machine includes: starting an electric control cabinet power supply, and checking whether the display of each instrument is normal; setting a target liquid level; the feeding pump starts feeding; starting the circulating pump at a low frequency when feeding to the current liquid level display value; after the current liquid level display is consistent with the target liquid level display, adjusting the frequency of the circulating pump; the frequency of the circulating pump is adjusted to the target feeding pressure; adjusting the target liquid level according to the flotation froth phenomenon; adjusting the air inflow; adjusting the cleaning water quantity; repeatedly adjusting the steps until target concentrate and tailings are stably produced; recording various parameters.
Fig. 5 shows a block diagram of a real-time monitoring and early warning system 100 of a micro-bubble flotation machine according to an embodiment of the present application. As shown in fig. 5, a real-time monitoring and early warning system 100 of a micro-bubble flotation machine according to an embodiment of the present application includes: an image acquisition module 110, configured to acquire a sequence of pulp real-time monitoring images acquired by the camera; the multi-dimensional pulp characteristic extraction module 120 is used for extracting multi-dimensional pulp characteristics of the sequence of the pulp real-time monitoring images to obtain a sequence of multi-dimensional pulp characteristic fusion images; the context-associated feature extraction module 130 is configured to extract context-associated features of the sequence of multi-dimensional pulp feature fusion images to obtain a pulp timing semantic context-associated feature vector; and the early warning analysis module 140 is used for determining whether to send out the target liquid level adjustment early warning based on the ore pulp time sequence semantic context association feature vector.
Here, it will be understood by those skilled in the art that the specific functions and operations of the respective units and modules in the real-time monitoring and early warning system 100 of the above-described micro-bubble flotation machine have been described in detail in the above description of the real-time monitoring and early warning method of the micro-bubble flotation machine with reference to fig. 1 to 3, and thus, repetitive descriptions thereof will be omitted.
As described above, the real-time monitoring and early warning system 100 of the micro-bubble flotation machine according to the embodiment of the present application may be implemented in various wireless terminals, for example, a server or the like having a real-time monitoring and early warning algorithm of the micro-bubble flotation machine. In one possible implementation, the real-time monitoring and early warning system 100 of the micro-bubble flotation machine according to an embodiment of the present application may be integrated into the wireless terminal as a software module and/or a hardware module. For example, the real-time monitoring and early warning system 100 of the micro-bubble flotation machine may be a software module in the operating system of the wireless terminal, or may be an application developed for the wireless terminal; of course, the real-time monitoring and early warning system 100 of the micro-bubble flotation machine can also be one of a plurality of hardware modules of the wireless terminal.
Alternatively, in another example, the real-time monitoring and early warning system 100 of the micro-bubble flotation machine and the wireless terminal may be separate devices, and the real-time monitoring and early warning system 100 of the micro-bubble flotation machine may be connected to the wireless terminal through a wired and/or wireless network and transmit interactive information according to an agreed data format.
Fig. 6 shows an application scenario diagram of a real-time monitoring and early warning method of a micro-bubble flotation machine according to an embodiment of the present application. As shown in fig. 6, in this application scenario, first, a sequence of pulp real-time monitoring images collected by a camera (for example, D illustrated in fig. 6) is acquired, and then the sequence of pulp real-time monitoring images is input to a server (for example, S illustrated in fig. 6) in which a real-time monitoring and early warning algorithm of a micro-bubble flotation machine is deployed, where the server can process the sequence of pulp real-time monitoring and early warning images using the real-time monitoring and early warning algorithm of the micro-bubble flotation machine to obtain a classification result for indicating whether to issue a target liquid level adjustment early warning.
In an exemplary embodiment, a non-transitory computer readable storage medium is also provided, such as a memory including computer program instructions executable by a processing component of an apparatus to perform the above-described method.
The present application may be a system, method, and/or computer program product. The computer program product may include a computer readable storage medium having computer readable program instructions embodied thereon for causing a processor to implement aspects of the present application.
The computer readable storage medium may be a tangible device that can hold and store instructions for use by an instruction execution device. The computer readable storage medium may be, for example, but not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium would include the following: portable computer disks, hard disks, random Access Memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static Random Access Memory (SRAM), portable compact disk read-only memory (CD-ROM), digital Versatile Disks (DVD), memory sticks, floppy disks, mechanical coding devices, punch cards or in-groove structures such as punch cards or grooves having instructions stored thereon, and any suitable combination of the foregoing. Computer-readable storage media, as used herein, are not to be construed as transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., optical pulses through fiber optic cables), or electrical signals transmitted through wires.
The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logical function(s). In some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The foregoing description of embodiments of the application has been presented for purposes of illustration and description, and is not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the various embodiments described. The terminology used herein was chosen in order to best explain the principles of the embodiments, the practical application, or the improvement of technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.
Claims (10)
1. The real-time monitoring and early warning method for the micro-bubble flotation machine is characterized by comprising the following steps of:
acquiring a sequence of ore pulp real-time monitoring images acquired by a camera;
Extracting the multidimensional pulp characteristics of the sequence of the pulp real-time monitoring images to obtain a sequence of multidimensional pulp characteristic fusion images;
extracting context associated features of the sequence of the multidimensional pulp feature fusion images to obtain pulp time sequence semantic context associated feature vectors; and
And determining whether to send out a target liquid level adjustment early warning based on the ore pulp time sequence semantic context association feature vector.
2. The method of claim 1, wherein extracting the multi-dimensional pulp characteristics of the sequence of pulp real-time monitoring images to obtain the sequence of multi-dimensional pulp characteristic fusion images comprises:
calculating gradient direction histograms of each ore pulp real-time monitoring image in the sequence of the ore pulp real-time monitoring images to obtain a sequence of edge information characteristics;
carrying out multistage filtering treatment on each ore pulp real-time monitoring image in the ore pulp real-time monitoring image sequence to obtain a characteristic sequence after multistage filtering treatment;
After each ore pulp real-time monitoring image in the sequence of the ore pulp real-time monitoring images is subjected to blocking treatment to obtain a plurality of ore pulp image blocks, calculating the gray maximum value of each ore pulp image block to obtain a sequence of local gray maximum value characteristics; and
And fusing the sequence of the edge information features, the sequence of the features subjected to the multistage filtering treatment and the sequence of the local gray maximum features to obtain the sequence of the multi-dimensional ore pulp feature fused image.
3. The method of claim 2, wherein calculating a gradient direction histogram of each pulp real-time monitoring image in the sequence of pulp real-time monitoring images to obtain the sequence of edge information features comprises:
uniformly dividing the ore pulp real-time monitoring image to obtain a plurality of cell spaces;
Calculating gradients of pixel points in each cell space in the cell spaces, and generating a plurality of cell direction gradient histograms according to gradient distribution; and
The gradient direction histogram is generated based on the plurality of cell direction gradient histograms.
4. The method of claim 3, wherein the multistage filtering process comprises gaussian filtering and bilateral filtering.
5. The method of claim 4, wherein extracting the context-dependent features of the sequence of multi-dimensional pulp feature fusion images to obtain the pulp timing semantic context-dependent feature vector comprises:
And passing the sequence of the multidimensional pulp characteristic fusion image through a pulp time sequence semantic characteristic extractor based on a converter to obtain the pulp time sequence semantic context associated characteristic vector.
6. The method of claim 5, wherein passing the sequence of multi-dimensional pulp feature fusion images through a transducer-based pulp timing semantic feature extractor to obtain the pulp timing semantic context-associated feature vector, comprises:
image blocking is carried out on the sequence of the multi-dimensional pulp characteristic fusion image respectively to obtain a sequence of a plurality of multi-dimensional pulp characteristic fusion image blocks;
Embedding each multi-dimensional pulp characteristic fusion image block in the sequence of the multi-dimensional pulp characteristic fusion image blocks by using an image block embedding layer to obtain a sequence of embedding vectors of the multi-dimensional pulp characteristic fusion image blocks; and
And enabling the sequence of the embedded vectors of the multi-dimensional pulp feature fusion image blocks to pass through the pulp time sequence semantic feature extractor based on the converter to obtain the pulp time sequence semantic context associated feature vector.
7. The method of claim 6, wherein determining whether to issue a target level adjustment pre-warning based on the pulp timing semantic context associated feature vector comprises:
and the ore pulp time sequence semantic context associated feature vector is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether a target liquid level adjustment early warning is sent out.
8. The method for real-time monitoring and early warning of a micro-bubble flotation machine according to claim 7, further comprising the training step of: training the converter-based pulp timing semantic feature extractor and the classifier.
9. The method of claim 8, wherein the training step comprises:
acquiring training data, wherein the training data comprises a sequence of training pulp real-time monitoring images and a true value of whether a target liquid level adjustment early warning is sent out;
calculating a gradient direction histogram of each training pulp real-time monitoring image in the training pulp real-time monitoring image sequence to obtain a training edge information characteristic sequence;
Carrying out multistage filtering treatment on each training pulp real-time monitoring image in the training pulp real-time monitoring image sequence to obtain a characteristic sequence after training multistage filtering treatment, wherein the multistage filtering treatment comprises Gaussian filtering and bilateral filtering;
After each training pulp real-time monitoring image in the training pulp real-time monitoring image sequence is subjected to blocking processing to obtain a plurality of training pulp image blocks, calculating the gray maximum value of each training pulp image block to obtain a training local gray maximum value characteristic sequence;
Fusing the sequence of the training edge information characteristics, the sequence of the characteristics after the training multi-level filtering treatment and the sequence of the training local gray maximum characteristics to obtain a sequence of a training multi-dimensional ore pulp characteristic fusion image;
Passing the sequence of training multidimensional pulp feature fusion images through the transducer-based pulp time sequence semantic feature extractor to obtain training pulp time sequence semantic context associated feature vectors;
clustering optimization is carried out on the training ore pulp time sequence semantic context associated feature vectors so as to obtain optimized training ore pulp time sequence semantic context associated feature vectors;
the optimized training ore pulp time sequence semantic context association feature vector passes through the classifier to obtain a classification loss function value; and
Training the converter-based pulp timing semantic feature extractor and the classifier with the classification loss function values.
10. A real-time monitoring and early warning system of a micro-bubble flotation machine, characterized in that the real-time monitoring and early warning system of the micro-bubble flotation machine operates in the real-time monitoring and early warning method of the micro-bubble flotation machine according to claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410355884.6A CN117943213B (en) | 2024-03-27 | 2024-03-27 | Real-time monitoring and early warning system and method for micro-bubble flotation machine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410355884.6A CN117943213B (en) | 2024-03-27 | 2024-03-27 | Real-time monitoring and early warning system and method for micro-bubble flotation machine |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117943213A true CN117943213A (en) | 2024-04-30 |
| CN117943213B CN117943213B (en) | 2024-06-04 |
Family
ID=90799845
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410355884.6A Active CN117943213B (en) | 2024-03-27 | 2024-03-27 | Real-time monitoring and early warning system and method for micro-bubble flotation machine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117943213B (en) |
Citations (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB201814982D0 (en) * | 2018-09-14 | 2018-10-31 | Reliable Data Systems International Ltd | Vehicle Position Identification |
| CN109772593A (en) * | 2019-01-25 | 2019-05-21 | 东北大学 | A kind of mineral pulp level prediction technique based on flotation froth behavioral characteristics |
| CN110899005A (en) * | 2019-12-24 | 2020-03-24 | 江西理工大学 | Intelligent monitoring device and monitoring method for flotation tank pulp and foam full tank |
| CN211838492U (en) * | 2019-12-24 | 2020-11-03 | 江西理工大学 | Full groove intelligent monitoring device of flotation cell ore pulp and foam |
| KR102358769B1 (en) * | 2021-08-27 | 2022-02-08 | 한국지질자원연구원 | Method for controlling operation of flotation process |
| EP3957712A1 (en) * | 2020-08-21 | 2022-02-23 | Sartorius Stedim Data Analytics AB | Automated foam detection |
| CN114383985A (en) * | 2021-12-24 | 2022-04-22 | 鞍钢集团矿业有限公司 | Flotation concentrate grade detection method based on multi-parameter feature fusion |
| US20220156490A1 (en) * | 2020-11-16 | 2022-05-19 | Kyocera Document Solutions Inc. | Method and system for extracting information from a document image |
| CN115711128A (en) * | 2022-11-11 | 2023-02-24 | 深圳市金航深海矿产开发集团有限公司 | Real-time abundance detection system and method for deep-sea polymetallic nodule mineral products |
| CN116213105A (en) * | 2023-03-31 | 2023-06-06 | 浙江艾领创矿业科技有限公司 | Gold ore floatation process and equipment |
| WO2023178798A1 (en) * | 2022-03-25 | 2023-09-28 | 平安科技（深圳）有限公司 | Image classification method and apparatus, and device and medium |
| WO2023179111A1 (en) * | 2022-03-23 | 2023-09-28 | 中国矿业大学 | Intelligent flotation chemicals loading system and method based on flotation tailing ore pulp detection |
| CN116832968A (en) * | 2023-05-25 | 2023-10-03 | 宁波春泥新材料有限公司 | Intelligent floatation process and system for lepidolite |
| CN116889933A (en) * | 2023-08-16 | 2023-10-17 | 江西江钒科技实业有限公司 | Enrichment beneficiation method for vanadium-containing stone coal mine |
| CN117115193A (en) * | 2023-08-25 | 2023-11-24 | 大连海事大学 | Infrared ship positioning method based on line inhibition |
| CN117244678A (en) * | 2023-10-11 | 2023-12-19 | 浙江艾领创矿业科技有限公司 | Intelligent monitoring control system and method for sand mill |
| CN117274059A (en) * | 2023-10-11 | 2023-12-22 | 杭州一隅千象科技有限公司 | Low-resolution image reconstruction method and system based on image coding-decoding |
| CN117292148A (en) * | 2023-11-27 | 2023-12-26 | 中交第二公路勘察设计研究院有限公司 | Tunnel surrounding rock level assessment method based on directional drilling and test data |
| CN117579513A (en) * | 2024-01-16 | 2024-02-20 | 北京中科网芯科技有限公司 | Visual operation and maintenance system and method for convergence and diversion equipment |
| CN117764994A (en) * | 2024-02-22 | 2024-03-26 | 浙江首鼎视介科技有限公司 | biliary pancreas imaging system and method based on artificial intelligence |
-
2024
- 2024-03-27 CN CN202410355884.6A patent/CN117943213B/en active Active
Patent Citations (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB201814982D0 (en) * | 2018-09-14 | 2018-10-31 | Reliable Data Systems International Ltd | Vehicle Position Identification |
| CN109772593A (en) * | 2019-01-25 | 2019-05-21 | 东北大学 | A kind of mineral pulp level prediction technique based on flotation froth behavioral characteristics |
| CN110899005A (en) * | 2019-12-24 | 2020-03-24 | 江西理工大学 | Intelligent monitoring device and monitoring method for flotation tank pulp and foam full tank |
| CN211838492U (en) * | 2019-12-24 | 2020-11-03 | 江西理工大学 | Full groove intelligent monitoring device of flotation cell ore pulp and foam |
| EP3957712A1 (en) * | 2020-08-21 | 2022-02-23 | Sartorius Stedim Data Analytics AB | Automated foam detection |
| US20220156490A1 (en) * | 2020-11-16 | 2022-05-19 | Kyocera Document Solutions Inc. | Method and system for extracting information from a document image |
| KR102358769B1 (en) * | 2021-08-27 | 2022-02-08 | 한국지질자원연구원 | Method for controlling operation of flotation process |
| CN114383985A (en) * | 2021-12-24 | 2022-04-22 | 鞍钢集团矿业有限公司 | Flotation concentrate grade detection method based on multi-parameter feature fusion |
| WO2023179111A1 (en) * | 2022-03-23 | 2023-09-28 | 中国矿业大学 | Intelligent flotation chemicals loading system and method based on flotation tailing ore pulp detection |
| WO2023178798A1 (en) * | 2022-03-25 | 2023-09-28 | 平安科技（深圳）有限公司 | Image classification method and apparatus, and device and medium |
| CN115711128A (en) * | 2022-11-11 | 2023-02-24 | 深圳市金航深海矿产开发集团有限公司 | Real-time abundance detection system and method for deep-sea polymetallic nodule mineral products |
| CN116213105A (en) * | 2023-03-31 | 2023-06-06 | 浙江艾领创矿业科技有限公司 | Gold ore floatation process and equipment |
| CN116832968A (en) * | 2023-05-25 | 2023-10-03 | 宁波春泥新材料有限公司 | Intelligent floatation process and system for lepidolite |
| CN116889933A (en) * | 2023-08-16 | 2023-10-17 | 江西江钒科技实业有限公司 | Enrichment beneficiation method for vanadium-containing stone coal mine |
| CN117115193A (en) * | 2023-08-25 | 2023-11-24 | 大连海事大学 | Infrared ship positioning method based on line inhibition |
| CN117244678A (en) * | 2023-10-11 | 2023-12-19 | 浙江艾领创矿业科技有限公司 | Intelligent monitoring control system and method for sand mill |
| CN117274059A (en) * | 2023-10-11 | 2023-12-22 | 杭州一隅千象科技有限公司 | Low-resolution image reconstruction method and system based on image coding-decoding |
| CN117292148A (en) * | 2023-11-27 | 2023-12-26 | 中交第二公路勘察设计研究院有限公司 | Tunnel surrounding rock level assessment method based on directional drilling and test data |
| CN117579513A (en) * | 2024-01-16 | 2024-02-20 | 北京中科网芯科技有限公司 | Visual operation and maintenance system and method for convergence and diversion equipment |
| CN117764994A (en) * | 2024-02-22 | 2024-03-26 | 浙江首鼎视介科技有限公司 | biliary pancreas imaging system and method based on artificial intelligence |
Non-Patent Citations (2)
| Title |
|---|
| J・亚纳托斯等: "工业型浮选回路自动控制的流体动力学特性和选矿指标研究", 国外金属矿选矿, no. 05, 25 May 2002 (2002-05-25), pages 31 - 37 * |
| 何桂春等: "浮选泡沫图像处理技术研究现状与进展", 有色金属科学与工程, no. 02, 15 April 2011 (2011-04-15), pages 57 - 63 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117943213B (en) | 2024-06-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110175613A (en) | Street view image semantic segmentation method based on Analysis On Multi-scale Features and codec models | |
| CN108830196A (en) | Pedestrian detection method based on feature pyramid network | |
| CN109101108B (en) | Method and system for optimizing human-computer interaction interface of intelligent cabin based on three decisions | |
| CN111862093A (en) | Corrosion grade information processing method and system based on image recognition | |
| CN102033964A (en) | Text classification method based on block partition and position weight | |
| CN109284779A (en) | Object detection method based on deep full convolution network | |
| CN104268600B (en) | A kind of mineral flotation foam image texture analysis and operating mode's switch method based on Minkowski distances | |
| Mehrshad et al. | New image-processing algorithm for measurement of bubble size distribution from flotation froth images | |
| CN108446616A (en) | Method for extracting roads based on full convolutional neural networks integrated study | |
| CN110738132B (en) | Target detection quality blind evaluation method with discriminant perception capability | |
| CN107067022B (en) | Method, device and equipment for establishing image classification model | |
| CN111062383A (en) | Image-based ship detection depth neural network algorithm | |
| WO2019049060A1 (en) | Froth segmentation in flotation cells | |
| CN117611932A (en) | Image classification method and system based on double pseudo tag refinement and sample re-weighting | |
| CN105354547A (en) | Pedestrian detection method in combination of texture and color features | |
| CN117943213B (en) | Real-time monitoring and early warning system and method for micro-bubble flotation machine | |
| CN105069450A (en) | Quick multi-character recognition method | |
| CN115601747A (en) | Method and system for calculating confluency of adherent cells | |
| CN115588191A (en) | Cell sorting method and system based on image acoustic flow control cell sorting model | |
| CN115661904A (en) | Data labeling and domain adaptation model training method, device, equipment and medium | |
| Lan et al. | PCB defect detection algorithm of improved YOLOv8 | |
| CN109447136A (en) | A kind of conspicuousness detection method for 360 degree of images | |
| CN115424250A (en) | License plate recognition method and device | |
| CN116597258B (en) | Ore sorting model training method and system based on multi-scale feature fusion | |
| CN110807456A (en) | Method and device for positioning bank card number |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |