CN115578612B - Blast furnace top distribution stage identification method and device based on marker target detection - Google Patents
Blast furnace top distribution stage identification method and device based on marker target detection Download PDFInfo
- Publication number
- CN115578612B CN115578612B CN202211241185.6A CN202211241185A CN115578612B CN 115578612 B CN115578612 B CN 115578612B CN 202211241185 A CN202211241185 A CN 202211241185A CN 115578612 B CN115578612 B CN 115578612B
- Authority
- CN
- China
- Prior art keywords
- blast furnace
- furnace top
- distribution stage
- target detection
- marker
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- C—CHEMISTRY; METALLURGY
- C21—METALLURGY OF IRON
- C21B—MANUFACTURE OF IRON OR STEEL
- C21B5/00—Making pig-iron in the blast furnace
- C21B5/006—Automatically controlling the process
-
- C—CHEMISTRY; METALLURGY
- C21—METALLURGY OF IRON
- C21B—MANUFACTURE OF IRON OR STEEL
- C21B5/00—Making pig-iron in the blast furnace
- C21B5/008—Composition or distribution of the charge
-
- C—CHEMISTRY; METALLURGY
- C21—METALLURGY OF IRON
- C21B—MANUFACTURE OF IRON OR STEEL
- C21B7/00—Blast furnaces
- C21B7/24—Test rods or other checking devices
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- C—CHEMISTRY; METALLURGY
- C21—METALLURGY OF IRON
- C21B—MANUFACTURE OF IRON OR STEEL
- C21B2300/00—Process aspects
- C21B2300/04—Modeling of the process, e.g. for control purposes; CII
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- General Physics & Mathematics (AREA)
- Organic Chemistry (AREA)
- Metallurgy (AREA)
- Materials Engineering (AREA)
- Manufacturing & Machinery (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Radiation Pyrometers (AREA)
Abstract
The invention discloses a blast furnace top distribution stage identification method and device based on marker target detection. According to the method, a furnace top infrared thermal imager is utilized, markers are selected according to characteristics of infrared images of the furnace top of the blast furnace in different material distribution stages, a furnace top marker target detection model is established, and a blast furnace top material distribution stage judgment condition is provided by combining target detection results and a blast furnace material distribution rule, so that the recognition of the blast furnace material distribution stage is realized. The invention converts the identification problem of the furnace top material distribution stage into the detection problem of the marker target, and realizes the automatic identification of the blast furnace material distribution stage. In the production and operation process of the blast furnace, the method provided by the invention can realize high-precision effective dynamic identification of the material distribution stage, and provides important furnace top material distribution information for ensuring the normal operation of the blast furnace.
Description
Technical Field
The invention relates to the technical field of computer vision, in particular to a blast furnace top material distribution stage identification method and device based on marker target detection.
Background
With the increasing world steel demand and steel production, steel production and quality have become important indicators for the development and economic viability of a country. The steel industry is one of the pillar industries in modern countries, and is also an important source of energy use, resource consumption and environmental pollution, and has a critical influence on the implementation of sustainable development strategies.
The blast furnace is one of the core links of the steel industry, and the main working procedure of the blast furnace is to reduce iron elements from iron ores, so as to provide raw materials for the subsequent steelmaking links, and the quality of produced molten iron directly influences the quality of products of the subsequent links such as steelmaking, steel processing and the like. Blast furnace burden distribution is an important operation in the blast furnace process, and determines burden distribution at the furnace top, affects reaction and heat transfer between gas and solid phases in the furnace, and finally affects quality and yield of produced molten iron. Therefore, the method has very important significance for monitoring the blast furnace burden distribution process.
In order to realize the monitoring of the blast furnace burden distribution process, the existing methods can be divided into two types: direct measurement and indirect measurement. The direct measurement method directly obtains the distribution condition of the furnace top charge level through a sensor, so as to obtain the distribution information, for example, a mechanical stock rod, a radar stock rod and the like are used in a blast furnace. The method can effectively acquire the furnace charge information of the position of the stock rod, but because the method is limited by the number of measurement points, the method can only acquire the information of limited points, cannot reflect the real furnace top charge level distribution and has certain limitation. The indirect measurement method is used for measuring the temperature distribution at the throat, the distribution condition of the furnace top is indirectly deduced through the temperature distribution, and the common indirect measurement method is a throat thermocouple cross temperature measurement technology. The method has good dynamic performance, but the sensor is easy to be broken down or damaged due to direct contact with high-temperature airflow, especially the sensor at the central part. And because the maintenance period of the blast furnace is longer, once the temperature sensor is damaged or fails, the maintenance period is longer and the replacement difficulty is higher. Therefore, the normal blast furnace throat temperature monitoring is difficult to maintain, the related blast furnace state monitoring, control and optimization are affected, and the normal production is not facilitated.
With the development of non-contact detection technology, infrared thermal imaging technology is increasingly applied to temperature detection tasks. The infrared thermal imaging is a novel non-contact temperature measurement technology, is based on the principle that the radiant energy of an object changes along with the temperature, does not directly contact with the object to be measured, can be used for detecting the temperature of the surfaces of a moving object and a high-temperature object, and does not damage the temperature field of the object to be measured. Infrared thermal imaging has a longer lifetime than contact temperature measurement and can provide instant temperature field information. However, blast furnace burden distribution is a continuous and complex process. In the burden distribution process, iron ore and coke are alternately conveyed into a blast furnace as burden materials, and due to the influence of physical and chemical reactions between air flow in the furnace and the burden materials, the temperature in the furnace can be periodically lifted, and dust carried on the burden materials can move along with the air flow in the furnace. At the moment, the furnace top material level information can not be effectively obtained through the furnace top infrared thermal imager, and the outline of the material level can be seen through the furnace top infrared thermal imager only when the material distribution is stopped. Therefore, the premise of effectively monitoring the blast furnace burden distribution process by using the furnace top infrared thermal imager is to effectively identify the burden distribution stage, and a method for effectively identifying the blast furnace burden distribution stage through the blast furnace top infrared image is needed.
Computer vision is a technology that has studied how to use cameras and computers instead of human eyes to accomplish the task of identifying, detecting and tracking targets. The furnace top infrared thermal imager can acquire temperature information in the furnace in real time in the form of infrared thermal images, and when the blast furnace is in different material distribution stages, the infrared images of the furnace top of the blast furnace show different characteristics. Selecting a proper marker, performing image processing on an infrared image of the blast furnace top by adopting a computer vision technology, detecting the marker, and judging different material distribution stages according to detection results, thereby being a reliable method. Therefore, the invention provides a blast furnace top burden distribution stage identification method based on marker target detection.
Disclosure of Invention
The invention aims to provide a blast furnace top distribution stage identification method based on marker target detection, aiming at the defects in the existing blast furnace top distribution stage identification technology. The method samples the infrared video of the furnace top of the blast furnace, and performs image preprocessing operation on the infrared image of the furnace top according to the characteristics of the infrared image of the furnace top. According to the visibility difference of all parts of the infrared view field inside the furnace throat at the furnace top of the blast furnace in different material distribution stages, selecting objects with different visibility characteristics at the different material distribution stages of the furnace top of the blast furnace as markers. And establishing a marker infrared image data set, training a target detection model by using the marker infrared image data set, and establishing a furnace top marker target detection model. Meanwhile, the invention provides the judgment conditions of the blast furnace top burden distribution stage, and combines the detection results of the marker targets to realize high-precision effective dynamic identification of the burden distribution stage, thereby providing important furnace top burden distribution information for ensuring the normal operation of the blast furnace.
In order to achieve the above purpose, the invention adopts the following technical scheme: a blast furnace top burden distribution stage identification method based on marker target detection comprises the following steps:
s1: acquiring an infrared image of the top of the blast furnace, and performing image preprocessing operation according to the characteristics of the infrared image of the top of the blast furnace;
s2: according to the visibility difference of all parts of the infrared view field inner furnace throat at the top of the blast furnace in different material distribution stages, selecting objects with different visibility characteristics at the top of the blast furnace in different material distribution stages as markers;
s3: building a furnace top marker target detection model based on a target detection algorithm, training the model, and inputting an infrared image of the furnace top of the blast furnace after model training is completed to obtain a marker target detection result;
s4: according to a blast furnace burden distribution rule, as the physical properties of iron ore and coke are different, the definition difference exists in the selected markers in the infrared images of the two furnace tops, according to the definition difference of the selected markers in the infrared images of the two furnace tops, and the target detection results of the markers are combined, the blast furnace top burden distribution stage judgment conditions are respectively formulated for the two furnace burden conditions, and the blast furnace top burden distribution stage identification is realized according to the blast furnace top burden distribution stage judgment conditions.
Further, the step S1 specifically includes the following sub-steps:
s1-1, screenshot sampling is carried out on infrared video of the blast furnace top at intervals of a certain frame number, and a sampled infrared image dataset of the blast furnace top is obtained;
s1-2, redundant color information is contained in the sampled infrared image of the blast furnace top, the sampled infrared image of the blast furnace top is subjected to gray-scale treatment according to the following formula, and the color image is converted into a gray-scale image so as to reduce the complexity of subsequent calculation;
Gray(i,j)＝αR(i,j)+βG(i,j)+γB(i,j)
wherein Gray (i, j) represents a Gray value of the pixel having coordinates (i, j) after graying; r (i, j), G (i, j) and B (i, j) represent luminance values of R channel, G channel and B channel before graying of the pixel having coordinates (i, j), respectively; alpha, beta and gamma are the proportion of the three-channel brightness value respectively;
s1-3, the internal environment of the blast furnace is bad, so that the imaging effect of the furnace top infrared thermal imager is bad, the obtained infrared image of the furnace top of the blast furnace contains a lot of noise information, wherein the most dominant noise is salt and pepper noise, the noise in the infrared image of the furnace top of the blast furnace is filtered by adopting a median filtering method with good filtering effect on the salt and pepper noise, and the expression is as follows:
wherein F is i,j Is the gray value of the pixel with coordinates (i, j) after median filtering; a is a two-dimensional filter window with a size of m x m centered on a pixel with coordinates (i, j);in order to order the gray values of the pixels in the filter window from small to large.
Further, in step S2, in the non-burden distribution stage, the following features can be clearly seen in the infrared image of the blast furnace roof: temperature measuring rod, chute, material hammer and central gas flow; in the material distribution stage, dust carried on furnace burden moves upwards irregularly in the furnace under the influence of air flow in the furnace, so that the furnace top infrared thermal imager cannot clearly acquire the characteristics; at least one of the above features is selected as a marker.
Further, the step S3 specifically includes the following substeps:
s3-1, marking the markers in the infrared image dataset according to the format requirements of the selected target detection model and establishing the marker infrared image dataset;
s3-2, training the selected target detection model by using a marker infrared image data set, and establishing a furnace top marker target detection model;
s3-3, detecting the infrared image of the furnace top of the real-time blast furnace by using a furnace top marker target detection model to obtain a marker target detection result, wherein the marker target detection result comprises the size and the position of each marker and the corresponding confidence level.
Further, the step S4 specifically includes the following substeps:
s4-1 blast furnace top burden distribution stage determination conditions are as follows, and when the current time is the Nth frame, the blast furnace top burden distribution stage determination result of the Nth frame image is usedIndicating (I)>Determined as follows:
wherein S is 1 Is the material distribution stage position of the blast furnace top,indicating that the blast furnace is distributing during the N-1 frame,indicating that the blast furnace is not distributing in the N-1 frame; s is S 2 Is the type of furnace burden>Indicating that the burden being charged at frame N-1 is iron ore, +>Indicating that the charge being charged at frame N-1 is coke; s is S 3 Is the target detection result bit,/->The value of (2) is the number of the detected markers in the Nth frame of image; />Is the result of the summation of the target detection result bits of the images from the nth frame to the n+a frame; a. b, c and d are four parameters to be determined in the model, a is the number of frames of infrared images of the blast furnace top which are required to be inspected when judging the blast furnace top distribution stage of the Nth frame image, the larger the value of a is, the more the number of frames which are required to be inspected is, the more accurate the judging result of the corresponding blast furnace top distribution stage is, meanwhile, the larger the value of a means that the model has higher time delay, so that the two factors are required to be comprehensively considered when determining the value of a, and the value of a is as small as possible on the premise of ensuring the accuracy of the model; b is +.A. judging whether the blast furnace top burden distribution stage of the Nth frame image is changed from the non-burden distribution stage to the burden distribution stage>The value of b is determined by the value of a, the number of the selected throat markers and the visibility difference of the selected throat markers in different distribution stages; and is to judge whether the blast furnace top burden distribution stage of the Nth frame image is changed from the burden distribution stage to the non-burden distribution stage +.>The values of c and d are determined by the value of a, the number of selected throat markers and the difference in definition of the selected markers under different furnace charges; />A blast furnace top material distribution stage identification result of the Nth frame image;
s4-2 asWhen N frame image is in "non-cloth phase", when +.>When the Nth frame image is in a cloth stage; after obtaining the N-th frameAfter the recognition result, according to the N frame +.>And->And the subsequent marker target detection result realizes the identification of the blast furnace top material distribution stage of the subsequent frame according to the blast furnace top material distribution stage judgment condition.
The invention also provides a blast furnace top burden distribution stage identification device based on the marker target detection, which comprises a memory and one or more processors, wherein executable codes are stored in the memory, and the processors are used for realizing the blast furnace top burden distribution stage identification method based on the marker target detection when executing the executable codes.
The invention also provides a computer readable storage medium, wherein a program is stored on the computer readable storage medium, and when the program is executed by a processor, the method for identifying the blast furnace top material distribution stage based on the marker target detection is realized.
The invention has the beneficial effects that: the effective identification of the blast furnace top burden distribution stage is a precondition and basis for realizing the monitoring of the blast furnace burden distribution process. The invention provides a method for realizing the identification of the material distribution stage of the blast furnace top by utilizing the infrared image data of the blast furnace top based on a target detection algorithm, which converts the identification problem of the material distribution stage of the blast furnace top into the target detection problem of a marker, thereby realizing the automatic identification of the material distribution stage of the blast furnace top. In the production and operation process of the blast furnace, the furnace top distribution stage identification method provided by the invention can realize high-precision effective dynamic identification of the furnace top distribution stage, and provides important furnace top distribution information for ensuring the normal operation of the blast furnace.
Drawings
FIG. 1 is a flow chart of a method for identifying a blast furnace roof burden distribution stage based on marker target detection in an embodiment of the invention.
In fig. 2, (a) and (b) are respectively infrared images of the furnace roof of the blast furnace in the non-burden distribution stage and the burden distribution stage in the embodiment of the present invention, and (c) is a marker image selected as a marker for discriminating between different burden distribution stages.
Fig. 3 is a network structure diagram of an object detection model in an embodiment of the present invention.
In fig. 4, (a), (b) and (c) are three types of result diagrams of detection of a marker object in the embodiment of the present invention, which correspond to the case that two temperature measurement bars can detect both, only one temperature measurement bar and both temperature measurement bars cannot detect.
Fig. 5 is a flowchart showing a judgment condition for a top cloth stage in the embodiment of the present invention.
Fig. 6 is a block diagram of a blast furnace roof burden distribution stage recognition device based on marker target detection in an embodiment of the present invention.
Detailed Description
The invention will be further described with reference to the accompanying drawings and specific examples.
FIG. 1 shows the overall flow of a blast furnace top burden distribution stage identification method based on marker target detection in an embodiment of the invention. Firstly, carrying out image pretreatment on an infrared image of the blast furnace top by adopting a graying and median filtering method. Marking the marker by using an image marking tool to establish a marker infrared image data set, and training a target detection model by using the marker infrared image data set to establish a furnace top marker target detection model. Finally, a blast furnace top material distribution stage judgment condition is provided by combining a blast furnace material distribution rule and a marker target detection result, and automatic identification of the blast furnace top material distribution stage is realized according to the blast furnace top material distribution stage judgment condition.
The blast furnace top material distribution stage identification method based on marker target detection in the embodiment of the invention comprises the following steps:
s1: acquiring an infrared image of the top of the blast furnace, and performing image preprocessing operation on the infrared image according to the characteristics of the infrared image of the top of the blast furnace, wherein the specific steps comprise:
s1-1, screenshot sampling is carried out on the infrared video of the blast furnace top at intervals of a certain number of frames, and a sampled infrared image dataset of the blast furnace top is obtained.
S1-2, redundant color information is contained in the sampled infrared image of the blast furnace top, the sampled infrared image of the blast furnace top is subjected to gray-scale treatment according to the following formula, and the color image is converted into a gray-scale image so as to reduce the complexity of subsequent calculation;
Gray(i,j)＝αR(i,j)+βG(i,j)+γB(i,j)
wherein Gray (i, i) represents a Gray value of the pixel having coordinates (i, j) after graying; r (i, j), G (i, j) and B (i, i) represent luminance values of R channel, G channel and B channel before graying of the pixel having coordinates (i, j), respectively; alpha, beta and gamma are the proportion of the three-channel brightness value respectively; in this example, α, β and γ are 0.30,0.59,0.11, respectively.
S1-3, the internal environment of the blast furnace is bad, so that the imaging effect of the furnace top infrared thermal imager is bad, the obtained infrared image of the furnace top of the blast furnace contains a lot of noise information, wherein the most dominant noise is salt and pepper noise, the noise in the infrared image of the furnace top of the blast furnace is filtered by adopting a median filtering method with good filtering effect on the salt and pepper noise, and the expression is as follows:
wherein F is i,j Is the gray value of the pixel with coordinates (i, j) after median filtering; a is a two-dimensional filter window with a size of 3 x 3 centered on a pixel with coordinates (i, j);in order to order the gray values of the pixels in the filter window from small to large.
S2: according to the visibility difference of all parts of the infrared view field inside the furnace throat at the furnace top of the blast furnace in different material distribution stages, selecting objects with different visibility characteristics at the different material distribution stages of the furnace top of the blast furnace as markers, wherein the specific method is as follows:
fig. 2 (a) and (b) are respectively infrared images of the blast furnace at the non-burden distribution stage and the burden distribution stage in the embodiment of the present invention, and fig. 2 (c) is a selected marker image for discriminating between different burden distribution stages.
In the normal condition, in the non-material distribution stage, the characteristics of a temperature measuring rod, a chute, a material hammer, a central gas flow and the like can be clearly seen in an infrared image of the blast furnace top; in the material distribution stage, dust carried on furnace burden moves upwards irregularly in the furnace under the influence of air flow in the furnace, so that the furnace top infrared thermal imager cannot clearly acquire the characteristics. At least one of the above features is selected as a marker, and in this embodiment, two temperature measurement bars are selected as markers.
S3: building a furnace top marker target detection model based on a target detection algorithm, training the model, inputting an infrared image of the furnace top of the blast furnace after model training is completed, and obtaining a marker target detection result, wherein the method specifically comprises the following steps of:
s3-1, marking the markers in the infrared image dataset according to the selected markers and the format requirements of the selected target detection model, and establishing the infrared image dataset of the markers.
S3-2, training the selected target detection model by using the marker infrared image data set, and establishing a furnace top marker target detection model.
S3-3, detecting the infrared image of the furnace top of the real-time blast furnace by using a furnace top marker target detection model to obtain a marker target detection result, wherein the marker target detection result comprises the size and the position of each marker and the corresponding confidence level.
Fig. 3 is a network structure diagram of the object detection model of the present embodiment. The target detection algorithm used in the embodiment of the invention is a YOLOv5s algorithm, and the network structure of the target detection algorithm is divided into four layers.
The first layer is an input layer that performs an adaptive scaling operation on the input picture to a size suitable for processing by the object detection network.
The second layer is a backbone network part, and the backbone network is formed by combining a series of feature extraction modules, and the main function of the second layer is to extract the features of the input picture.
The third layer is a neck part, and the main function of the neck part is to extract deep features from features extracted from a main network through operations such as convolution and the like, and to fuse the features between different layers through operations such as splicing and the like, so that a fused feature map is obtained.
The fourth layer is an output layer, and has the function of predicting the type and the position of an object contained in the picture according to the feature map obtained in the previous layer and obtaining a final target detection result.
In fig. 4, (a), (b) and (c) are three types of result diagrams of detection of a marker object in this embodiment, which correspond to the case where two temperature measurement bars can detect both, but only one temperature measurement bar and two temperature measurement bars cannot detect both. The two temperature measurement bars in this embodiment are named "rod1" and "rod2", respectively, and if the corresponding temperature measurement bar is detected, it will be framed, and its name and corresponding confidence will be displayed in the upper left corner of the target frame.
S4: according to the distribution rule of the blast furnace, as the physical properties of two furnace charges of iron ore and coke are different, the definition difference exists in the selected markers in the infrared images of the two furnace tops, according to the definition difference of the selected markers in the infrared images of the two furnace tops, and in combination with the target detection result of the markers, the judgment conditions of the distribution stage of the furnace top of the blast furnace are respectively formulated for the two furnace charge conditions, and the identification of the distribution stage of the furnace top of the blast furnace is realized according to the judgment conditions of the distribution stage of the furnace top, and the specific steps comprise:
s4-1 blast furnace top burden distribution stage determination conditions are as follows, and when the current time is the Nth frame, the blast furnace top burden distribution stage determination result of the Nth frame image is usedIndicating (I)>Determined as follows:
wherein S is 1 Is the material distribution stage position of the blast furnace top,indicating that the blast furnace is distributing during the N-1 frame,indicating that the blast furnace is not distributing in the N-1 frame; s is S 2 Is the type of furnace burden>Indicating that the burden being charged at frame N-1 is iron ore, +>Indicating that the charge being charged at frame N-1 is coke; s is S 3 Is the target detection result bit,/->The value of (2) is the number of the detected markers in the Nth frame of image; />Is the result of the summation of the target detection result bits of the images from the nth frame to the n+a frame; a. b, c and d are four parameters to be determined in the model, a is the number of frames of infrared images of the blast furnace top which are required to be inspected when judging the blast furnace top distribution stage of the Nth frame image, the larger the value of a is, the more the number of frames which are required to be inspected is, the more accurate the judging result of the corresponding blast furnace top distribution stage is, meanwhile, the larger the value of a means that the model has higher time delay, so that the two factors are required to be comprehensively considered when determining the value of a, and the value of a is as small as possible on the premise of ensuring the accuracy of the model; b is +.A. judging whether the blast furnace top burden distribution stage of the Nth frame image is changed from the non-burden distribution stage to the burden distribution stage>The value of b is determined by the value of a, the number of the selected throat markers and the visibility difference of the selected throat markers in different distribution stages; c and d are +.A. for judging whether the burden distribution stage of the blast furnace top of the Nth frame image is changed from the burden distribution stage to the non-burden distribution stage>The values of c and d are determined by the value of a, the number of selected throat markers and the difference in definition of the selected markers under different furnace charges; />A blast furnace top material distribution stage identification result of the Nth frame image; in this embodiment, the values of a, b, c, d are respectively: 29,0, 60, 45.
S4-2 asWhen N frame image is in "non-cloth phase", when +.>When the Nth frame image is in a cloth stage; after the recognition result of the nth frame is obtained, according to +.>And->And the subsequent marker target detection result realizes the identification of the blast furnace top material distribution stage of the subsequent frame according to the blast furnace top material distribution stage judgment condition.
In this embodiment, a schematic diagram of the determination conditions in the top distribution stage is shown in fig. 5. The distribution stage of the Nth frame image is determined by the recognition result of the top distribution stage of the N-1 th frame and the detection result of the marker targets of the N th frame to the (n+29) th frame.
In this embodiment, a certain 2650m of China is selected 3 And obtaining infrared image data sets of the blast furnace top for 2 days respectively by sampling the infrared video of the blast furnace top for 2 days in a certain month of a certain year in the blast furnace database. The furnace top marker target detection model is obtained through model training and testing by using the blast furnace top infrared image data set on the 1 st day. Blast furnace roof infrared image dataset pair acquisition using day 2And verifying the obtained furnace top marker target detection model. And selecting 3 groups of sample data, and performing experimental verification on 400 pictures in each group.
Evaluating the estimation effect of the model by adopting an identification accuracy ACC, wherein the identification accuracy ACC is equal to the number M of images which are identified correctly in the cloth stage r Divided by the total number of pictures M t The following formula is shown:
the evaluation index results are as follows:
the table shows that the identification accuracy of the blast furnace top distribution stage identification method based on the marker target detection exceeds 95% on 3 groups of verification data sets, and the effectiveness and the reliability of the blast furnace top distribution stage identification method based on the marker target detection provided by the invention are intuitively proved.
Corresponding to the embodiment of the blast furnace top burden distribution stage identification method based on the marker target detection, the invention also provides an embodiment of the blast furnace top burden distribution stage identification device based on the marker target detection.
Referring to fig. 6, the blast furnace top burden distribution stage identifying device based on the detection of the marker targets provided by the embodiment of the invention comprises a memory and one or more processors, wherein executable codes are stored in the memory, and the processors are used for realizing the blast furnace top burden distribution stage identifying method based on the detection of the marker targets in the embodiment when executing the executable codes.
The embodiment of the blast furnace top distribution stage identification device based on the marker target detection can be applied to any equipment with data processing capability, wherein the equipment with data processing capability can be equipment or a device such as a computer. The apparatus embodiments may be implemented by software, or may be implemented by hardware or a combination of hardware and software. Taking software implementation as an example, the device in a logic sense is formed by reading corresponding computer program instructions in a nonvolatile memory into a memory by a processor of any device with data processing capability. In terms of hardware, as shown in fig. 6, a hardware structure diagram of an apparatus with optional data processing capability where the identification device for detecting a blast furnace top burden distribution stage based on a marker object according to the present invention is located is shown in fig. 6, and in addition to a processor, a memory, a network interface, and a nonvolatile memory shown in fig. 6, the optional apparatus with data processing capability in the embodiment generally includes other hardware according to an actual function of the optional apparatus with data processing capability, which is not described herein again.
The implementation process of the functions and roles of each unit in the above device is specifically shown in the implementation process of the corresponding steps in the above method, and will not be described herein again.
For the device embodiments, reference is made to the description of the method embodiments for the relevant points, since they essentially correspond to the method embodiments. The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purposes of the present invention. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
The embodiment of the invention also provides a computer readable storage medium, wherein a program is stored on the computer readable storage medium, and when the program is executed by a processor, the method for identifying the blast furnace top burden distribution stage based on the detection of the marker targets in the embodiment is realized.
The computer readable storage medium may be an internal storage unit, such as a hard disk or a memory, of any of the data processing enabled devices described in any of the previous embodiments. The computer readable storage medium may be any external storage device that has data processing capability, such as a plug-in hard disk, a Smart Media Card (SMC), an SD Card, a Flash memory Card (Flash Card), or the like, which are provided on the device. Further, the computer readable storage medium may include both internal storage units and external storage devices of any data processing device. The computer readable storage medium is used for storing the computer program and other programs and data required by the arbitrary data processing apparatus, and may also be used for temporarily storing data that has been output or is to be output.
It should be noted that the above embodiments are only for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention may be modified or substituted without departing from the spirit and scope of the technical solution of the present invention, which is intended to be covered in the scope of the claims of the present invention.
The foregoing detailed description of the preferred embodiments of the present invention has been presented in detail, and it should be understood that the foregoing description is of the most preferred embodiment of the invention and is not intended to limit the invention. Any modifications, additions, and equivalent substitutions made within the principle of the invention should be included in the protection scope of the invention.
Claims (7)
1. The blast furnace top distribution stage identification method based on marker target detection is characterized by comprising the following steps of:
s1: acquiring an infrared image of the top of the blast furnace, and performing image preprocessing operation according to the characteristics of the infrared image of the top of the blast furnace;
s2: according to the visibility difference of all parts of the infrared view field inner furnace throat at the top of the blast furnace in different material distribution stages, selecting objects with different visibility characteristics at the top of the blast furnace in different material distribution stages as markers;
s3: building a furnace top marker target detection model based on a target detection algorithm, training the model, and inputting an infrared image of the furnace top of the blast furnace after model training is completed to obtain a marker target detection result;
s4: according to a blast furnace burden distribution rule, as the physical properties of iron ore and coke are different, the definition difference exists in the selected markers in the infrared images of the two furnace tops, according to the definition difference of the selected markers in the infrared images of the two furnace tops, and the target detection results of the markers are combined, the blast furnace top burden distribution stage judgment conditions are respectively formulated for the two furnace burden conditions, and the blast furnace top burden distribution stage identification is realized according to the furnace top burden distribution stage judgment conditions; the method specifically comprises the following substeps:
s4-1 blast furnace top burden distribution stage determination conditions are as follows, and when the current time is the Nth frame, the blast furnace top burden distribution stage determination result of the Nth frame image is usedIndicating (I)>Determined as follows:
wherein S is 1 Is the material distribution stage position of the blast furnace top,indicating that the blast furnace is distributing during the N-1 th frame,/->Indicating that the blast furnace is not distributing in the N-1 frame; s is S 2 Is the type of furnace burden>Indicating that the burden being charged at frame N-1 is iron ore, +>Indicating that the charge being charged at frame N-1 is coke; s is S 3 Is the target detection result bit,/->The value of (2) is the number of the detected markers in the Nth frame of image; />Is the result of the summation of the target detection result bits of the images from the nth frame to the n+a frame; a is the frame number of infrared images of the blast furnace top which are required to be inspected when judging the blast furnace top distribution stage of the Nth frame image, b is the +.>C and is the +.f for judging whether the blast furnace top burden distribution stage of the Nth frame image is changed from the "burden distribution stage" to the "non-burden distribution stage->A threshold value of (2); />A blast furnace top material distribution stage identification result of the Nth frame image;
s4-2 asWhen N frame image is in "non-cloth phase", when +.>When the Nth frame image is in a cloth stage; after the recognition result of the nth frame is obtained, according to +.>And->And the subsequent marker target detection result realizes the identification of the blast furnace top material distribution stage of the subsequent frame according to the blast furnace top material distribution stage judgment condition.
2. The method for identifying a blast furnace roof burden distribution stage based on marker object detection according to claim 1, wherein the step S1 specifically comprises the following sub-steps:
the method comprises the steps of S1, carrying out screenshot sampling on an infrared video of the blast furnace top at intervals of a certain frame number to obtain a sampled infrared image dataset of the blast furnace top;
s1-2, carrying out gray processing on the sampled infrared image of the blast furnace top, wherein the formula is as follows:
Gray(i，j)＝αR(i，j)+βG(i，j)+γB(i，j)
wherein Gray (i, j) represents a Gray value of the pixel having coordinates (i, j) after graying; r (i, j), G (i, j) and B (i, j) represent luminance values of R channel, G channel and B channel before graying of the pixel having coordinates (i, j), respectively; alpha, beta and gamma are the proportion of the three-channel brightness value respectively;
s1-3, filtering noise in the infrared image of the blast furnace top by adopting a median filtering method, wherein the expression is as follows:
wherein F is i，j Is the gray value of the pixel with coordinates (i, j) after median filtering; a is a two-dimensional filter window with a size of m x m centered on a pixel with coordinates (i, j);to order the gray values of pixels in the filter window from small to large。
3. The method for identifying a blast furnace top burden-distribution stage based on marker object detection according to claim 1, wherein in step S2, in the non-burden-distribution stage, the following features can be clearly seen in the blast furnace top infrared image: temperature measuring rod, chute, material hammer and central gas flow; in the material distribution stage, dust carried on furnace burden moves upwards irregularly in the furnace under the influence of air flow in the furnace, so that the furnace top infrared thermal imager cannot clearly acquire the characteristics; at least one of the characteristics of the temperature measuring rod, the chute, the material hammer and the central gas flow is selected as a marker.
4. The method for identifying a blast furnace roof burden distribution stage based on marker object detection according to claim 1, wherein the step S3 specifically comprises the following sub-steps:
s3-1, marking the markers in the infrared image dataset according to the format requirements of the selected target detection model and establishing the marker infrared image dataset;
s3-2, training the selected target detection model by using a marker infrared image data set, and establishing a furnace top marker target detection model;
s3-3, detecting the infrared image of the furnace top of the real-time blast furnace by using a furnace top marker target detection model to obtain a marker target detection result, wherein the marker target detection result comprises the size and the position of each marker and the corresponding confidence level.
5. The method for identifying the blast furnace top distribution stage based on the marker target detection according to claim 1 is characterized in that in the blast furnace top distribution stage judging condition, the larger the value of a parameter is, the more the number of frames to be inspected is, the more accurate the judging result of the corresponding blast furnace top distribution stage is, meanwhile, the larger the value of a means that a model has higher time delay, so that the two factors need to be comprehensively considered when determining the value of a, and the value of a is as small as possible on the premise of ensuring the model precision; the value of the parameter b is determined by the value of a, the number of the selected throat markers and the visibility difference of the selected throat markers in different distribution stages; the values of parameters c and d are determined by the value of a, the number of throat markers selected and the difference in definition of the markers selected under different charges.
6. A blast furnace roof burden distribution stage identification device based on marker target detection, comprising a memory and one or more processors, wherein the memory stores executable code, and wherein the processor is configured to implement the blast furnace roof burden distribution stage identification method based on marker target detection as claimed in any one of claims 1 to 5 when executing the executable code.
7. A computer-readable storage medium, on which a program is stored, which program, when being executed by a processor, implements a method for identifying a blast furnace roof burden distribution stage based on detection of a marker object according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211241185.6A CN115578612B (en) | 2022-10-11 | 2022-10-11 | Blast furnace top distribution stage identification method and device based on marker target detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211241185.6A CN115578612B (en) | 2022-10-11 | 2022-10-11 | Blast furnace top distribution stage identification method and device based on marker target detection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115578612A CN115578612A (en) | 2023-01-06 |
| CN115578612B true CN115578612B (en) | 2023-08-04 |
Family
ID=84585091
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211241185.6A Active CN115578612B (en) | 2022-10-11 | 2022-10-11 | Blast furnace top distribution stage identification method and device based on marker target detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115578612B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114757935A (en) * | 2022-05-05 | 2022-07-15 | 河钢数字技术股份有限公司 | Multi-target automatic temperature measurement and early warning method for components of steel cold-rolling annealing furnace |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2574873A1 (en) * | 2011-09-28 | 2013-04-03 | Siemens VAI Metals Technologies GmbH | Method and device for decreasing the temperature of blast furnace gas temperature peaks |
| CN102796835B (en) * | 2012-07-25 | 2013-09-11 | 北京科技大学 | Charge level monitoring method based on data fusion of measured data and cloth model and charge level monitoring system |
| CN105695652B (en) * | 2016-04-21 | 2017-10-20 | 上海中冶横天自动化工程有限公司 | Contourgraph and high furnace control system for detecting material top surface in top of blast furnace contour shape |
| CN107609261A (en) * | 2017-09-07 | 2018-01-19 | 北京金恒博远科技股份有限公司 | The emulation mode of blast furnace material distribution process, apparatus and system |
| CN109852748B (en) * | 2019-02-27 | 2020-05-08 | 内蒙古科技大学 | Method for monitoring development process of gas flow in distribution period of blast furnace and predicting gas utilization rate |
| CN111666683B (en) * | 2020-06-05 | 2022-11-11 | 中冶赛迪信息技术(重庆)有限公司 | Method and system for predicting distribution state of furnace burden in blast furnace |
| CN112270237B (en) * | 2020-10-22 | 2023-08-29 | 中冶南方工程技术有限公司 | Blast furnace gas flow state identification method, system and storage medium |
| CN113139275B (en) * | 2021-03-22 | 2022-08-19 | 浙江大学 | Blast furnace throat temperature estimation method based on multilayer ore-coke ratio distribution model |
| CN113887308A (en) * | 2021-09-02 | 2022-01-04 | 中南大学 | Blast furnace charge identification method, device and system based on image multivariate features |
| CN114743152A (en) * | 2022-05-12 | 2022-07-12 | 中南大学 | Automatic extraction method and system for video key frames of blast furnace burden surface |
-
2022
- 2022-10-11 CN CN202211241185.6A patent/CN115578612B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114757935A (en) * | 2022-05-05 | 2022-07-15 | 河钢数字技术股份有限公司 | Multi-target automatic temperature measurement and early warning method for components of steel cold-rolling annealing furnace |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115578612A (en) | 2023-01-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | Autonomous damage segmentation and measurement of glazed tiles in historic buildings via deep learning | |
| WO2021047232A1 (en) | Interaction behavior recognition method, apparatus, computer device, and storage medium | |
| CN110032992B (en) | Examination cheating detection method based on gestures | |
| EP3655923B1 (en) | Weakly supervised anomaly detection and segmentation in images | |
| CN103439342B (en) | Based on the Infrared Non-destructive Testing method of thermal map temporal aspect | |
| CN103020632B (en) | The method for quickly identifying of localization for Mobile Robot monumented point in a kind of indoor environment | |
| Wang et al. | Investigation into recognition algorithm of helmet violation based on YOLOv5-CBAM-DCN | |
| CN105678737B (en) | A kind of digital picture angular-point detection method based on Radon transformation | |
| CN102509109B (en) | Method for distinguishing Thangka image from non-Thangka image | |
| CN107862675A (en) | A kind of real-time vision detection method for electric bucket tooth missing | |
| CN111563896A (en) | Image processing method for catenary anomaly detection | |
| CN115995056A (en) | Automatic bridge disease identification method based on deep learning | |
| CN112989901A (en) | Deep learning-based liquid level meter reading identification method | |
| CN115578612B (en) | Blast furnace top distribution stage identification method and device based on marker target detection | |
| CN116721104B (en) | Live three-dimensional model defect detection method and device, electronic equipment and storage medium | |
| CN113902939A (en) | Industrial product large defect detection method based on twin network | |
| CN113487570A (en) | High-temperature continuous casting billet surface defect detection method based on improved yolov5x network model | |
| CN112651315A (en) | Information extraction method and device of line graph, computer equipment and storage medium | |
| CN116002480A (en) | Automatic detection method and system for accidental falling of passengers in elevator car | |
| CN115330743A (en) | Method for detecting defects based on double lights and corresponding system | |
| You et al. | A mechanical part sorting system based on computer vision | |
| CN113536895A (en) | Disc pointer meter identification method | |
| CN109360289B (en) | Power meter detection method fusing inspection robot positioning information | |
| CN111488785A (en) | Rotary kiln working condition detection method based on in-kiln image | |
| CN114821044B (en) | Square pointer instrument indication recognition method based on gradient transformation |
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 |