A blast furnace condition prediction method based on charge surface flame image and related equipment

Abstract

The present application discloses a method for predicting blast furnace conditions based on flame images of charge surfaces and related equipment, the prediction method comprising obtaining an in-furnace image of the blast furnace to be predicted, and determining a dynamic boundary threshold corresponding to the in-furnace image based on an image histogram of the in-furnace image; determining a flame region in the in-furnace image based on the dynamic boundary threshold; obtaining regional parameters of the flame region and image parameters of the in-furnace image; and predicting the blast furnace condition of the blast furnace to be predicted through a trained classification model based on the regional parameters and the image parameters. The present application improves the prediction accuracy of the blast furnace condition by obtaining an in-furnace image, determining the flame region corresponding to the in-furnace image based on a dynamic boundary threshold, and then predicting the blast furnace condition through a classification model based on the regional parameters of the flame region and the image parameters of the in-furnace image.

Description

Technical Field

[0001] The present application relates to the field of control science and engineering technology, and in particular to a blast furnace condition prediction method based on material surface flame images and related equipment. Background Art

[0002] As my country's basic raw material industry and an important pillar industry of the national economy, the steel industry has long faced major common problems such as low resource utilization and heavy environmental load. The blast furnace is a key equipment for the steel manufacturing process, the core unit for the conversion of ferrous material flow, and the link with the highest energy consumption and production cost in the steel manufacturing process, accounting for about 60%-70% of the comprehensive energy consumption of steel and the total cost of steel manufacturing. In the blast furnace production process, charge distribution is the most active and critical control method, which determines the distribution of charge and gas flow in the furnace. Among them, the important reference basis for the blast furnace charge distribution operation is the situation in the blast furnace and the smelting state.

[0003] However, as a large-scale countercurrent reactor with high temperature, high pressure, strong dust and closed lightlessness, the internal environment of the blast furnace is constantly undergoing complex and violent physical and chemical reactions, which makes most detection devices in the industry unable to be effectively used for monitoring the internal furnace conditions of the blast furnace, and thus the blast furnace conditions generally rely on human observation and recording. However, due to the typical characteristics of the harshness and complexity of the internal environment of the blast furnace, as well as the inconsistent judgment standards of different people on the status of the blast furnace, the on-site operators often fail to monitor the blast furnace conditions in a timely and accurate manner, making it difficult to assist in achieving fixed-point and quantitative material distribution. Summary of the invention

[0004] The technical problem to be solved by the present application is to provide a blast furnace condition prediction method based on charge surface flame images and related equipment in view of the deficiencies in the prior art.

[0005] In order to solve the above technical problems, the first aspect of the embodiment of the present application provides a blast furnace condition prediction method based on a material surface flame image, and the prediction method comprises:

[0006] Acquire an in-furnace image of a blast furnace to be predicted, and determine a dynamic boundary threshold corresponding to the in-furnace image based on an image histogram of the in-furnace image;

[0007] Based on the dynamic boundary threshold, determining a flame area in the furnace image;

[0008] Acquire the regional parameters of the flame region and the image parameters of the image in the furnace;

[0009] Based on the regional parameters and the image parameters, the blast furnace condition of the blast furnace to be predicted is predicted by a trained classification model.

[0010] The blast furnace condition prediction method based on the material surface flame image, wherein the image inside the furnace is obtained by taking pictures with an industrial high-temperature endoscope covering the material surface of the blast furnace to be predicted.

[0011] The blast furnace condition prediction method based on the material surface flame image, wherein the regional parameters at least include the regional perimeter and regional area of ​​the flame area, and the image parameters include image clarity and image change rate.

[0012] The blast furnace condition prediction method based on the material surface flame image, wherein the process of obtaining the image change rate is specifically as follows:

[0013] A preceding image of the in-furnace image is acquired, and an image change rate corresponding to the in-furnace image is determined based on the preceding image and the in-furnace image.

[0014] The blast furnace condition prediction method based on the material surface flame image, wherein the blast furnace condition includes a normal furnace condition, a suspended material furnace condition, a tilted material furnace condition or a collapsed material furnace condition.

[0015] The blast furnace condition prediction method based on the charge surface flame image, wherein determining the dynamic boundary threshold corresponding to the image in the furnace based on the image histogram of the image in the furnace specifically includes:

[0016] Determining a plurality of initial thresholds corresponding to the image in the furnace based on an image histogram of the image in the furnace;

[0017] Performing Gaussian curve fitting on the image histogram based on the initial thresholds to obtain a histogram fitting curve;

[0018] Based on the histogram fitting curve, several segmentation thresholds corresponding to the furnace image are determined, and a dynamic boundary threshold is selected from the several segmentation thresholds.

[0019] The blast furnace condition prediction method based on the material surface flame image, wherein the several image areas include the flame area, the material surface area and the furnace wall area; the several segmentation thresholds include the dynamic boundary threshold for dividing the flame area and the material surface area and the reference boundary threshold for the user to divide the material surface area and the furnace wall area.

[0020] The method for predicting blast furnace conditions based on material surface flame images, wherein the method further comprises:

[0021] Dividing the furnace image into a plurality of image regions based on the plurality of segmentation thresholds;

[0022] Each image region is linearly stretched to enhance the image in the furnace and present the enhanced image in the furnace, wherein the linear stretching corresponding to each image region has a different stretching degree.

[0023] A second aspect of the embodiment of the present application provides a blast furnace condition prediction device based on a material surface flame image, the prediction device comprising:

[0024] A first acquisition module is used to acquire an in-furnace image of a blast furnace to be predicted, and determine a dynamic boundary threshold corresponding to the in-furnace image based on an image histogram of the in-furnace image;

[0025] A determination module, configured to determine a flame region in the furnace image based on the dynamic boundary threshold;

[0026] A second acquisition module is used to acquire the regional parameters of the flame area and the image parameters of the image in the furnace;

[0027] A prediction module is used to predict the blast furnace condition of the blast furnace to be predicted based on the regional parameters and the image parameters through a trained classification model.

[0028] A third aspect of an embodiment of the present application provides a computer-readable storage medium, which stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in any of the above-described methods for predicting blast furnace conditions based on material surface flame images.

[0029] A fourth aspect of the embodiments of the present application provides a terminal device, comprising: a processor, a memory, and a communication bus; the memory stores a computer-readable program that can be executed by the processor;

[0030] The communication bus realizes the connection and communication between the processor and the memory;

[0031] When the processor executes the computer-readable program, the steps in any of the above-mentioned methods for predicting blast furnace conditions based on material surface flame images are implemented.

[0032] Beneficial effects: Compared with the prior art, the present application provides a method and related equipment for predicting the blast furnace condition based on the flame image of the charge surface, the prediction method comprising obtaining the in-furnace image of the blast furnace to be predicted, and determining the dynamic boundary threshold corresponding to the in-furnace image based on the image histogram of the in-furnace image; determining the flame area in the in-furnace image based on the dynamic boundary threshold; obtaining the regional parameters of the flame area and the image parameters of the in-furnace image; and predicting the blast furnace condition of the blast furnace to be predicted through a trained classification model based on the regional parameters and the image parameters. The present application improves the prediction accuracy of the blast furnace condition by obtaining the in-furnace image, determining the flame area corresponding to the in-furnace image based on the dynamic boundary threshold, and then predicting the blast furnace condition through a classification model based on the regional parameters of the flame area and the image parameters of the in-furnace image. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without inventive work.

[0034] Figure 1 This is a flow chart of a blast furnace condition prediction method based on charge surface flame images provided in this application.

[0035] Figure 2 A schematic flow chart of a blast furnace condition prediction method based on charge surface flame images provided in this application.

[0036] Figure 3 This is an example flow chart of the blast furnace condition prediction method based on the charge surface flame image provided in this application.

[0037] Figure 4 A schematic diagram of the process of obtaining regional parameters in the blast furnace condition prediction method based on the charge surface flame image provided in this application.

[0038] Figure 5 This is a structural principle diagram of a blast furnace condition prediction device based on charge surface flame images provided in this application.

[0039] Figure 6 This is a schematic diagram of the structure of the terminal device provided in this application. DETAILED DESCRIPTION

[0040] The present application provides a blast furnace condition prediction method based on a charge surface flame image and related equipment. In order to make the purpose, technical solution and effect of the present application clearer and more specific, the present application is further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0041] It will be understood by those skilled in the art that, unless expressly stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present application refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be an intermediate element. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein includes all or any unit and all combinations of one or more associated listed items.

[0042] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art to which this application belongs. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with those in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless specifically defined as here.

[0043] It should be understood that the sequence numbers and sizes of the steps in this embodiment do not mean the order of execution. The execution order of each process is determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present application.

[0044] The inventors have found through research that the steel industry, as my country's basic raw material industry and an important pillar industry of the national economy, has long faced major common problems such as low resource utilization and heavy environmental load. The blast furnace is a key equipment for the steel manufacturing process, the core unit for the conversion of ferrous material flow, and the link with the highest energy consumption and production cost in the steel manufacturing process, accounting for about 60%-70% of the comprehensive energy consumption of steel and the total cost of steel manufacturing. In the blast furnace production process, the distribution of charge and gas flow in the furnace is the most active and critical control method, which determines the distribution of charge and gas flow in the furnace. Among them, the important reference basis for the blast furnace distribution operation is the situation in the blast furnace and the smelting state.

[0045] However, as a large-scale countercurrent reactor with high temperature, high pressure, strong dust and closed lightlessness, the internal environment of the blast furnace is constantly undergoing complex and violent physical and chemical reactions, which makes most detection devices in the industry unable to be effectively used for monitoring the internal furnace conditions of the blast furnace, and thus the blast furnace conditions generally rely on human observation and recording. However, due to the typical characteristics of the harshness and complexity of the internal environment of the blast furnace, as well as the inconsistent judgment standards of different people on the status of the blast furnace, the on-site operators often fail to monitor the blast furnace conditions in a timely and accurate manner, making it difficult to assist in achieving fixed-point and quantitative material distribution.

[0046] In order to solve the above problems, in an embodiment of the present application, an in-furnace image of a blast furnace to be predicted is obtained, and a dynamic boundary threshold corresponding to the in-furnace image is determined based on an image histogram of the in-furnace image; based on the dynamic boundary threshold, a flame area in the in-furnace image is determined; regional parameters of the flame area and image parameters of the in-furnace image are obtained; based on the regional parameters and the image parameters, a trained classification model is used to predict the blast furnace condition of the blast furnace to be predicted. The present application improves the prediction accuracy of the blast furnace condition by obtaining an in-furnace image, determining the flame area corresponding to the in-furnace image based on a dynamic boundary threshold, and then predicting the blast furnace condition through a classification model based on the regional parameters of the flame area and the image parameters of the in-furnace image.

[0047] The application content is further explained below through the description of embodiments in conjunction with the accompanying drawings.

[0048] This embodiment provides a blast furnace condition prediction method based on the charge surface flame image. Figure 1 , Figure 2 , Figure 3 as well as Figure 4 As shown, the method includes:

[0049] S10, obtaining an in-furnace image of a blast furnace to be predicted, and determining a dynamic boundary threshold corresponding to the in-furnace image based on an image histogram of the in-furnace image.

[0050] Specifically, an industrial high-temperature endoscope is provided on the blast furnace to be predicted, and the industrial high-temperature endoscope covers the material surface of the blast furnace to be predicted, so as to obtain the furnace image of the blast furnace to be predicted through the industrial high-temperature endoscope. Among them, the industrial high-temperature endoscope collects the furnace image once at a preset time interval through the endoscope hole of the blast furnace, and the furnace image can at least include the image area formed by the flame in the blast furnace to be predicted and the image area formed by the material surface. In addition, the preset time can be set according to the time requirement, which is used to control the acquisition frequency of the industrial high-temperature endoscope, so that the industrial high-temperature endoscope can obtain an image of the furnace at a preset time interval. For example, the preset time is 1 second, 0.5 seconds, etc.

[0051] The dynamic boundary threshold is used as a segmentation threshold for image segmentation, and the flame area in the furnace image can be determined based on the dynamic boundary threshold. The dynamic boundary threshold is determined based on the furnace image, and the furnace image taken by the industrial high-temperature endoscope at each shooting moment is different, so that the dynamic boundary threshold corresponding to each furnace image can be different, thereby improving the accuracy of the flame area subsequently determined based on the dynamic boundary threshold.

[0052] In an implementation of this embodiment, determining the dynamic boundary threshold corresponding to the image in the furnace based on the image histogram of the image in the furnace specifically includes:

[0053] Determining a plurality of initial thresholds corresponding to the image in the furnace based on an image histogram of the image in the furnace;

[0054] Performing Gaussian curve fitting on the image histogram based on the initial thresholds to obtain a histogram fitting curve;

[0055] Based on the histogram fitting curve, several segmentation thresholds corresponding to the furnace image are determined, and a dynamic boundary threshold is selected from the several segmentation thresholds.

[0056] Specifically, the image histogram is a histogram for reflecting the brightness distribution in the furnace image, and the number of pixels of each brightness value in the image is marked. It can be understood that the image histogram is a manifestation of the brightness distribution in the furnace image, and each reflects the number of pixels of each brightness value. For example, assuming that the brightness value of the furnace image is reflected by the grayscale value of the furnace image, the grayscale value of the furnace image ranges from 0 to 255, and 0-255 is divided into several sub-areas, and the difference between the upper limit value and the lower limit value of each sub-area is 15, and then the number of pixels of each sub-area is counted to obtain the image histogram corresponding to the furnace image.

[0057] After obtaining the image histogram, several initial thresholds can be obtained based on the image histogram, wherein the number of thresholds of the several initial thresholds corresponds to the number of regions of the image region to be divided in the furnace image. It can be understood that the image region to be divided can be obtained based on several initial thresholds, wherein the number of thresholds of the several initial thresholds is 1 less than the number of regions of the image region to be divided. In a specific implementation, the image in the furnace is obtained by photographing the inside of the blast furnace to be predicted, and the industrial high-temperature endoscope used to photograph the image in the furnace covers the material surface of the blast furnace to be predicted, so that the image in the furnace obtained includes the flame area, the material surface area and the furnace wall area, and the brightness of the flame area, the material surface area and the furnace wall area is different. Therefore, after determining the image histogram, some local extreme values ​​can be eliminated by smoothing with a small window, and then the local maximum and minimum values ​​of the image histogram that eliminates the local extreme values ​​are calculated, and finally the extreme values ​​whose distance is less than the constraint value are deleted, and the k extreme values ​​and their boundaries with the largest coverage area are screened out as the initial threshold, wherein the constraint value can be set according to actual needs.

[0058] In a typical implementation of this embodiment, the plurality of image regions include a flame region, a material surface region, and a furnace wall region, and the plurality of initial thresholds include a first initial threshold and a second initial threshold, wherein the first initial threshold is used to determine the furnace wall region, and the second initial threshold is used to determine the flame region. That is, when the image in the furnace is divided based on the first initial threshold, the furnace wall region and the first reference image region can be obtained, and when the image in the furnace is divided based on the second initial threshold, the flame region and the second reference region can be obtained. Then, when the image in the furnace is divided based on the first initial threshold and the second initial threshold, the image to be processed can be divided into the furnace wall region, the material surface region, and the flame region.

[0059] In one implementation of this embodiment, after obtaining several initial thresholds, the image histogram can be fitted with a Gaussian curve using an STA algorithm, a particle swarm algorithm, or a genetic algorithm to obtain a histogram fitting curve, and then the thresholds between different peaks are determined based on the fitted histogram fitting curve to obtain several segmentation thresholds.

[0060] In one implementation of the present embodiment, since the image inside the furnace includes the furnace wall area, the material surface area and the flame area, and the furnace wall area, the material surface area and the flame area have different brightness characteristics, after the furnace wall area, the material surface area and the flame area are determined based on the segmentation threshold, the furnace wall area, the material surface area and the flame area can be linearly stretched to different degrees respectively. While retaining the authenticity of the pixel texture of the furnace wall area, the material surface area and the flame area, the contrast and texture information of the image inside the furnace can also be enhanced, thereby improving the clarity of the image inside the furnace presented to the user and facilitating the user to identify the furnace wall area, the material surface area and the flame area.

[0061] Based on this, in one implementation of this embodiment, the blast furnace condition prediction method based on the material surface flame image also includes:

[0062] Dividing the furnace image into a plurality of image regions based on the plurality of segmentation thresholds;

[0063] Each image region is linearly stretched to enhance the image inside the furnace, and the enhanced image inside the furnace is presented.

[0064] Specifically, the several image areas include a flame area, a material surface area and a furnace wall area; the several segmentation thresholds include a dynamic boundary threshold for dividing the flame area and the material surface area and a reference boundary threshold for the user to divide the material surface area and the furnace wall area, wherein the degree of linear stretching corresponding to each image area is different.

[0065] S20. Determine a flame area in the furnace image based on the dynamic boundary threshold.

[0066] Specifically, the flame area is included in the in-furnace image and is an image area formed by the flame in the blast furnace to be predicted. When the flame area is determined based on the dynamic boundary threshold, a binary segmentation method can be used to segment the in-furnace image to divide the in-furnace image into a flame area and a background area, and the background area includes a material surface area and a furnace wall area.

[0067] S30, acquiring the regional parameters of the flame region and the image parameters of the image in the furnace.

[0068] Specifically, the region parameters at least include the region perimeter and region area of ​​the flame region, and the image parameters include image clarity and image change rate. The region perimeter can be determined based on the edge length of the flame region, that is, after the flame region is acquired, the flame edge of the flame region can be extracted, and then the edge length of the flame edge can be determined to obtain the region perimeter. In an implementation of this embodiment, the flame region can include connected regions, the region perimeter is the sum of the region perimeters of each connected region, and the region area is the sum of the region areas of each connected region.

[0069] In an implementation of this embodiment, the process of acquiring the image change rate is specifically as follows:

[0070] A preceding image of the in-furnace image is acquired, and an image change rate corresponding to the in-furnace image is determined based on the preceding image and the in-furnace image.

[0071] Specifically, the preceding image and the furnace image are taken by the same image acquisition device, for example, they are both acquired by an industrial high-temperature endoscope installed on the blast furnace to be predicted. The acquisition time of the preceding image is earlier than the acquisition time of the furnace image, and the preceding image and the furnace image are adjacent images. It can be understood that the furnace image is not the first image, and the preceding image is an image that is adjacent to the furnace image and is located before the furnace image in the acquisition order, that is, the preceding image is the furnace image at the previous acquisition moment of the furnace image.

[0072] S40: Based on the region parameters and the image parameters, predict the blast furnace condition of the blast furnace to be predicted by using a trained classification model.

[0073] Specifically, the classification model is a trained network model, the input items of the classification model are the region parameters and the image parameters, and the output items are the blast furnace conditions. That is to say, the blast furnace conditions of the blast furnace to be predicted can be predicted by the classification model, and the blast furnace conditions include normal conditions, suspended material conditions, inclined material conditions, or collapsed material conditions. Figure 3 As shown, the training process of the classification model can be specifically to obtain a training sample set, wherein each training sample in the training sample set includes regional parameters of the flame area, image clarity and image change rate, and then input the training sample into a preset classification model, predict the predicted blast furnace condition corresponding to the training sample through the preset classification model, and then train the preset classification model based on the predicted blast furnace condition and the actual blast furnace condition corresponding to the training sample to obtain a trained classification model. Among them, the prediction classification model can adopt classification algorithms such as decision tree, support vector machine, Bayes, etc.

[0074] Based on the blast furnace condition prediction method based on the material surface flame image, this embodiment provides a blast furnace condition prediction device based on the material surface flame image, such as Figure 5 As shown, the prediction device comprises:

[0075] The first acquisition module 100 is used to acquire an in-furnace image of a blast furnace to be predicted, and determine a dynamic boundary threshold corresponding to the in-furnace image based on an image histogram of the in-furnace image;

[0076] A determination module 200, configured to determine a flame region in the furnace image based on the dynamic boundary threshold;

[0077] A second acquisition module 300 is used to acquire the regional parameters of the flame region and the image parameters of the furnace image;

[0078] The prediction module 400 is used to predict the blast furnace condition of the blast furnace to be predicted based on the regional parameters and the image parameters through a trained classification model.

[0079] Based on the above-mentioned blast furnace condition prediction method based on material surface flame images, this embodiment provides a computer-readable storage medium, and the computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in the blast furnace condition prediction method based on material surface flame images as described in the above-mentioned embodiment.

[0080] Based on the above-mentioned blast furnace condition prediction method based on material surface flame image, the present application also provides a terminal device, such as Figure 6 As shown, it includes at least one processor (processor) 20; display screen 21; and memory (memory) 22, and may also include a communications interface (Communications Interface) 23 and a bus 24. Among them, the processor 20, the display screen 21, the memory 22 and the communication interface 23 can communicate with each other through the bus 24. The display screen 21 is configured to display a preset user guide interface in the initial setting mode. The communication interface 23 can transmit information. The processor 20 can call the logic instructions in the memory 22 to execute the method in the above embodiment.

[0081] In addition, the logic instructions in the memory 22 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product.

[0082] The memory 22 is a computer-readable storage medium that can be configured to store software programs, computer executable programs, such as program instructions or modules corresponding to the methods in the embodiments of the present disclosure. The processor 20 executes functional applications and data processing by running the software programs, instructions or modules stored in the memory 22, that is, implementing the methods in the above embodiments.

[0083] The memory 22 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal device, etc. In addition, the memory 22 may include a high-speed random access memory and may also include a non-volatile memory. For example, a variety of media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, may also be a transient storage medium.

[0084] In addition, the specific process of loading and executing the plurality of instructions in the storage medium and the processor in the mobile terminal has been described in detail in the method above, and will not be described one by one here.

[0085] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application 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 application.

Claims

1. A method for predicting blast furnace conditions based on charge surface flame images, characterized in that: The prediction method includes: Acquire an in-furnace image of a blast furnace to be predicted, and determine a dynamic boundary threshold corresponding to the in-furnace image based on an image histogram of the in-furnace image, wherein the in-furnace image at least includes an image area formed by the flame in the blast furnace to be predicted and an image area formed by the material surface, and the image histogram is used to reflect the brightness distribution in the in-furnace image, and marks the number of pixels of each brightness value in the in-furnace image; Based on the dynamic boundary threshold, determining a flame area in the furnace image; Acquire the regional parameters of the flame region and the image parameters of the image in the furnace; Based on the regional parameters and the image parameters, predicting the blast furnace condition of the blast furnace to be predicted by a trained classification model, wherein the blast furnace condition includes a normal condition, a suspended charge condition, a tilted charge condition, or a collapsed charge condition; Wherein, determining the dynamic boundary threshold corresponding to the image in the furnace based on the image histogram of the image in the furnace specifically includes: Determine a number of initial thresholds corresponding to the image in the furnace based on the image histogram of the image in the furnace, wherein the number of the initial thresholds corresponds to the number of image regions to be divided into the image in the furnace, and the process of determining the number of initial thresholds is to eliminate some local extreme values ​​of the image histogram through window smoothing, calculate local maximum values ​​and minimum values ​​of the image histogram that eliminates the local extreme values, delete extreme values ​​whose distance is less than the constraint value, and screen out k extreme values ​​with the largest coverage area and their boundaries to obtain the number of initial thresholds; Performing Gaussian curve fitting on the image histogram based on the initial thresholds to obtain a histogram fitting curve; Based on the histogram fitting curve, several segmentation thresholds corresponding to the furnace image are determined, and a dynamic boundary threshold is selected from the several segmentation thresholds.

2. The blast furnace condition prediction method based on the charge surface flame image according to claim 1, characterized in that: The furnace image is captured by an industrial high-temperature endoscope covering the blast furnace charge surface to be predicted.

3. The blast furnace condition prediction method based on charge surface flame image according to claim 1, characterized in that: The region parameters at least include the region perimeter and the region area of ​​the flame region, and the image parameters include image clarity and image change rate.

4. The method for predicting blast furnace conditions based on charge surface flame images according to claim 3, characterized in that: The process of obtaining the image change rate is specifically as follows: A preceding image of the in-furnace image is acquired, and an image change rate corresponding to the in-furnace image is determined based on the preceding image and the in-furnace image.

5. The method for predicting blast furnace conditions based on charge surface flame images according to claim 1, characterized in that: The blast furnace conditions include normal conditions, suspended charge conditions, inclined charge conditions or collapsed charge conditions.

6. The method for predicting blast furnace conditions based on charge surface flame images according to claim 1, characterized in that: The image regions include a flame region, a material surface region and a furnace wall region; the segmentation thresholds include a dynamic boundary threshold for dividing the flame region and the material surface region and a reference boundary threshold for dividing the material surface region and the furnace wall region by a user.

7. The method for predicting blast furnace conditions based on charge surface flame images according to claim 6, characterized in that: The method further comprises: Dividing the furnace image into a plurality of image regions based on the plurality of segmentation thresholds; Each image region is linearly stretched to enhance the image in the furnace and present the enhanced image in the furnace, wherein the linear stretching corresponding to each image region has a different stretching degree.

8. A blast furnace condition prediction device based on charge surface flame image, characterized in that: The prediction device comprises: A first acquisition module is used to acquire an in-furnace image of a blast furnace to be predicted, and determine a dynamic boundary threshold corresponding to the in-furnace image based on an image histogram of the in-furnace image, wherein the in-furnace image at least includes an image area formed by the flame in the blast furnace to be predicted and an image area formed by the material surface, and the image histogram is used to reflect the brightness distribution in the in-furnace image, and the number of pixels of each brightness value in the in-furnace image is marked; A determination module, configured to determine a flame region in the furnace image based on the dynamic boundary threshold; A second acquisition module is used to acquire the regional parameters of the flame area and the image parameters of the image in the furnace; A prediction module, for predicting a blast furnace condition of the blast furnace to be predicted by a trained classification model based on the regional parameters and the image parameters, wherein the blast furnace condition includes a normal condition, a suspended charge condition, a tilted charge condition, or a collapsed charge condition; Wherein, determining the dynamic boundary threshold corresponding to the image in the furnace based on the image histogram of the image in the furnace specifically includes: Determine a number of initial thresholds corresponding to the image in the furnace based on the image histogram of the image in the furnace, wherein the number of the initial thresholds corresponds to the number of image regions to be divided into which the image in the furnace is to be divided, and the process of determining the number of initial thresholds is to eliminate some local extreme values ​​by window smoothing, calculate local maximum and minimum values ​​for the image histogram that eliminates the local extreme values, delete extreme values ​​whose distance is less than the constraint value, and screen out k extreme values ​​with the largest coverage area and their boundaries to obtain the number of initial thresholds; Performing Gaussian curve fitting on the image histogram based on the initial thresholds to obtain a histogram fitting curve; Based on the histogram fitting curve, several segmentation thresholds corresponding to the furnace image are determined, and a dynamic boundary threshold is selected from the several segmentation thresholds.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in the blast furnace condition prediction method based on the material surface flame image as described in any one of claims 1-7.

10. A terminal device, characterized in that: include: Processor, memory and communication bus; The memory stores a computer-readable program executable by the processor; The communication bus realizes the connection and communication between the processor and the memory; When the processor executes the computer-readable program, the steps of the blast furnace condition prediction method based on the material surface flame image as described in any one of claims 1 to 7 are implemented.

Images