Converter steelmaking state detection method and computer readable storage medium

Abstract

The application discloses a converter steelmaking state detection method and a computer readable storage medium, and the method comprises the following steps: reading a video frame from a real-time furnace mouth monitoring video stream to obtain a first image; removing character information in the first image to obtain a second image; performing flame detection on the second image through an RGB algorithm to obtain a first flame area; performing intense combustion flame detection on the second image through a preset intense combustion flame condition to obtain a second flame area; performing a set operation on the first flame area and the second flame area to obtain a flame image; calculating a total number of pixel points of the flame area in the flame image to obtain a flame area; and performing abnormal detection on the converter steelmaking state according to the flame area. The application can improve the accuracy of flame detection, thereby improving the accuracy of converter steelmaking state detection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of converter steelmaking, and particularly relates to a converter steelmaking state detection method and a computer readable storage medium. BACKGROUND

[0002] As an important link of steel production, the main goal of converter steelmaking is to smelt qualified molten steel in terms of composition and temperature. In the steelmaking process, the changes of the flame and smoke amount at the furnace mouth and whether there is slag spatter or overflow are important parameters for experts in the related field to judge the steelmaking state. The state discrimination of converter steelmaking has a direct impact on the quality of finished steel. In the past, the state discrimination according to manual experience needs the operator to continuously observe the flame changes at the furnace mouth, which has the problems of strong subjectivity and high cost.

[0003] In the paper entitled An early fire-detection method based on image processing, a method for detecting flame by RGB algorithm is disclosed, which discloses the RGB flame detection formula:

[0004] Rule1: R >= G >= B

[0005] Rule2: R >= R T

[0006] Rule3: S >= ((255-R) x S T / R T )

[0007] Wherein, R, G and B represent color components, R T represents the threshold value of R component, S T represents the threshold value of saturation, and the preferred range of the two threshold values is 115 < R T < 135, 55 < S T < 65.

[0008] However, when using a camera to take pictures, the saturation of the flame will change with the changes of various background illuminations, such as the saturation obtained during the day will be greater than that obtained at night. This is because if there is no other background illumination, the flame becomes the only light source. In this case, the hue of the operating flame will be more displayed as white according to the camera imaging. At this time, if only the RGB detection scheme is adopted, the strongly burning flame cannot be accurately detected. SUMMARY

[0009] The technical problems solved by the present application are to provide a converter steelmaking state detection method and a computer readable storage medium, which can improve the accuracy of flame detection, thereby improving the accuracy of converter steelmaking state detection.

[0010] To solve the above technical problems, the technical scheme adopted by the present application is as follows: a converter steelmaking state detection method, comprising:

[0011] reading a video frame from a real-time furnace mouth monitoring video stream to obtain a first image;

[0012] removing character information in the first image to obtain a second image;

[0013] performing flame detection on the second image through an RGB algorithm to obtain a first flame region;

[0014] performing intense combustion flame detection on the second image through a preset intense combustion flame condition to obtain a second flame region;

[0015] performing set operation on the first flame region and the second flame region to obtain a flame image;

[0016] calculating the total number of pixel points of the flame region in the flame image to obtain a flame area;

[0017] performing abnormal detection on the converter steelmaking state according to the flame area.

[0018] The present application also provides a computer readable storage medium having a computer program stored thereon, wherein the program is executed by a processor to implement the method described above.

[0019] The present application has the advantages that: by removing the monitoring character information on the video frame, the character information is prevented from affecting the subsequent converter steelmaking state detection, and the accuracy of detection is ensured; by detecting the flame regions of ordinary combustion and intense combustion respectively and taking the set of the two as the final flame detection result, the accuracy of flame detection is improved, thereby improving the accuracy of converter steelmaking state detection. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 The figure is a flowchart of a converter steelmaking state detection method of the present application;

[0021] Figure 2 The figure is a method flowchart of the first embodiment of the present application;

[0022] Figure 3 The figure is a schematic diagram of the first image of the first embodiment of the present application;

[0023] Figure 4 The figure is a schematic diagram of the second image of the first embodiment of the present application;

[0024] Figure 5 A schematic diagram of a normal furnace mouth region in the first image of the embodiment one of the present application;

[0025] Figure 6 A schematic diagram of an abnormal furnace mouth region in the first image of the embodiment one of the present application;

[0026] Figure 7 A schematic diagram of an abnormal furnace mouth region in the first image of the embodiment one of the present application;

[0027] Figure 8 A schematic diagram of the splicing of the first image, the first flame region, the second flame region and the flame image in the embodiment one of the present application;

[0028] Figure 9 A schematic diagram of the third image in the embodiment one of the present application;

[0029] Figure 10 A schematic diagram of the smoke image in the embodiment one of the present application;

[0030] Figure 11 A schematic diagram of the fourth image in the embodiment one of the present application;

[0031] Figure 12 A schematic diagram of the slag image in the embodiment one of the present application;

[0032] Figure 13 A schematic diagram of Figure 12 A corresponding first image schematic diagram;

[0033] Figure 14 A schematic diagram of the front-end display of the embodiment one of the present application. DETAILED DESCRIPTION

[0034] To make the technical content, the purposes and effects of the present application clear, the following will be described in detail in combination with the embodiments and the accompanying drawings.

[0035] Please refer to Figure 1 A converter steelmaking state detection method, comprising:

[0036] reading a video frame from a real-time furnace mouth monitoring video stream to obtain a first image;

[0037] removing character information in the first image to obtain a second image;

[0038] detecting a flame in the second image by an RGB algorithm to obtain a first flame region;

[0039] detecting a violently burning flame in the second image by a preset violently burning flame condition to obtain a second flame region;

[0040] Performing a union operation on the first flame area and the second flame area to obtain a flame image;

[0041] Counting the total number of pixel points of the flame area in the flame image to obtain a flame area;

[0042] According to the flame area, the abnormality of the converter steelmaking state is detected.

[0043] From the above description, the beneficial effects of the present application are that the accuracy of flame detection can be improved, thereby improving the accuracy of the converter steelmaking state detection.

[0044] Further, the intense combustion flame condition includes 215≤R≤255, 215≤G≤255 and 10≤B≤255, R, G and B are three color components of the pixel point.

[0045] From the above description, the intense combustion flame can be accurately detected when the flame is the only light source.

[0046] Further, before the flame detection on the second image by the RGB algorithm to obtain the first flame area, further comprising:

[0047] When the gun position value and the oxygen flow meet the preset furnace mouth area selection condition, the furnace mouth area selection is performed on the second image, and the preset furnace mouth area selection condition includes that the gun position value is located in a preset first interval and the oxygen flow is located in a preset second interval.

[0048] Determine whether the selected furnace mouth area meets the preset furnace mouth area size condition, and the furnace mouth area size condition includes that the length of the circumscribed rectangle of the selected furnace mouth area is within a preset length range, the width of the circumscribed rectangle is within a preset width range, and the total number of pixel points of the selected furnace mouth area is within a preset area range.

[0049] If it is satisfied, the flame detection on the second image by the RGB algorithm to obtain the first flame area is performed.

[0050] From the above description, only when the selected furnace mouth area is a normal furnace mouth area, the subsequent detection and analysis are performed, thereby ensuring the accuracy of the subsequent detection.

[0051] Further, the first interval is [2000, 2250], the second interval is [25000, 31000], the length range is [1400, 1950], the width range is [130, 350], and the area range is [130000, 290000].

[0052] Further, after the union operation on the first flame area and the second flame area is performed to obtain the flame image, the method further comprises:

[0053] performing a grayscale processing on the second image to obtain a third image;

[0054] performing a zero processing on pixel values of pixel points in the third image whose pixel values are greater than a preset smoke pixel threshold to obtain a smoke image, and calculating a number of pixel points in the smoke image whose pixel values are not zero to obtain a smoke area;

[0055] performing an exclusive or operation on the second image and the flame image to obtain a fourth image;

[0056] performing a zero processing on pixel values of pixel points in the fourth image whose pixel values are less than a preset slag pixel threshold to obtain a slag image, and calculating a number of pixel points in the slag image whose pixel values are not zero to obtain a slag area.

[0057] Further, the abnormal detection on the converter steelmaking state according to the flame area specifically comprises:

[0058] judging whether the converter steelmaking state is abnormal according to the flame area, the smoke area, the slag area and a preset warning threshold;

[0059] if yes, generating an abnormal report according to the flame area, the smoke area, the slag area and the real-time furnace mouth monitoring video stream.

[0060] As can be seen from the above description, the abnormal detection on the converter steelmaking state according to the flame area, the smoke area and the slag area can further improve the accuracy of the detection on the converter steelmaking state.

[0061] Further, after the abnormal detection on the converter steelmaking state according to the flame area, the method further comprises:

[0062] splicing the first image, the flame image, the smoke image and the slag image to obtain a front-end display image.

[0063] As can be seen from the above description, the front-end display image is generated, so that the user can intuitively understand the converter steelmaking state.

[0064] Further, after the union operation on the first flame area and the second flame area is performed to obtain the flame image, the method further comprises:

[0065] calculating an average grayscale value of the flame area in the flame image to obtain a flame brightness;

[0066] performing a difference operation on the flame image corresponding to the video frame and a previous video frame to obtain a frame difference image;

[0067] The average pixel value of the frame difference image is calculated, and the average pixel value is divided by a preset value to obtain the flame frequency corresponding to the video frame, and the preset value is 21.

[0068] As described above, by calculating the flame brightness and the flame frequency, the flame burning condition can be reflected to a certain extent, so that the related industrial parameters can be adjusted based on the flame burning condition.

[0069] Further, the method further comprises:

[0070] According to the flame area, the flame brightness, the flame frequency, the smoke area and the slag area, at least one of the industrial parameters including the gun position value, the oxygen flow, the carbon dioxide concentration and the fan speed is adjusted.

[0071] As described above, according to the flame area, the flame brightness, the flame frequency, the smoke area and the slag area, the flame burning condition and the slag overflow condition can be reflected to a certain extent, when the burning is too intense, the burning is insufficient or the slag is overflowed, by timely adjusting the related industrial parameters, the final product quality can be avoided, and economic losses and personnel casualties can be avoided.

[0072] The application further provides a computer readable storage medium, which stores a computer program, and the program is executed by a processor to realize the method.

[0073] Embodiment one

[0074] Please refer to Figures 2-14 , the embodiment one of the application is: a converter steelmaking state detection method, which can be applied to the analysis and control of the intelligent production process of a steelmaking converter ladle.

[0075] As Figure 2 shown, the method comprises the following steps:

[0076] S1: a video frame is read from a real-time furnace mouth monitoring video stream to obtain a first image. That is, a video frame is read from a real-time furnace mouth monitoring video stream through RTSP (Real Time Streaming Protocol), and the read video frame is recorded as a first image, as Figure 3 shown.

[0077] S2: character information in the first image is removed to obtain a second image.

[0078] Specifically, since the characters of the monitoring information are generally white in the video image, the first image can be screened for white characters according to the RGB three channels, and the first image screened for the monitoring information is recorded as a second image, as shown in Figure 4

[0079] S3: When the gun position value and the oxygen flow meet the preset furnace mouth region selection condition, the second image is selected for the furnace mouth region. When selecting, the largest image connected part is selected.

[0080] The gun position value is the vertical distance between the end face of the oxygen lance nozzle and the calm molten pool surface, that is, the oxygen lance height; and the furnace mouth region selection condition includes that the gun position value is in a preset first interval and the oxygen flow is in a preset second interval. Since the furnace mouth region changes when the gun position value is in the interval [2000, 2250] and the oxygen flow is in the interval [25000, 31000], in the embodiment, the first interval is [2000, 2250] and the second interval is [25000, 31000].

[0081] Specifically, the gun position value and the oxygen flow are obtained from the real-time database, and when the gun position value is in the interval [2000, 2250] and the oxygen flow is in the interval [25000, 31000], the furnace mouth region in the second image is subjected to mask processing. The furnace mouth region is the part surrounded by the brush in Figure 5 After mask processing, the image only shows the part surrounded by the mask.

[0082] S4: It is judged whether the selected furnace mouth region meets the preset furnace mouth region size condition, if yes, step S5 is executed, and if not, the first image is discarded, that is, the current video frame is discarded.

[0083] The furnace mouth region size condition includes that the length of the circumscribed rectangle of the selected furnace mouth region is in a preset length range, the width of the circumscribed rectangle is in a preset width range, and the total number of pixel points of the selected furnace mouth region is in a preset area range; in the embodiment, the length range is [1400, 1950], the width range is [130, 350], and the area range is [130000, 290000].

[0084] In the embodiment, the normal furnace mouth region should be as shown in the part surrounded by the frame in Figure 6 If the furnace mouth region is as shown in Figure 7 , it does not meet the condition of the normal furnace mouth region, in order to avoid affecting the accuracy of subsequent detection, the video frame in which the furnace mouth region does not meet the condition is discarded.

[0085] ​S5: performing flame detection on the second image by using an RGB algorithm to obtain a first flame region; that is, a first flame region in the second image is detected.

[0086] Specifically, a pixel point region in the second image satisfying an RGB flame detection formula is obtained, at which time the flame can be preliminarily identified, and is recorded as a first flame region, as shown in the upper left part of FIG. 5. Figure 8 The RGB flame detection formula includes:

[0087] Rule 1: R≥G≥B

[0088] Rule 2: R≥R T

[0089] Rule 3: S≥((255-R)×S T / R T )

[0090] wherein R, G and B represent color components, R T represents a threshold value of the R component, S T represents a threshold value of the saturation, and the preferred ranges of the two threshold values are 115 T < 135 and 55 T < 65.

[0091] S6: performing intense burning flame detection on the second image by using a preset intense burning flame condition to obtain a second flame region.

[0092] Specifically, a pixel point region in the second image satisfying the preset intense burning flame condition is obtained to obtain the second flame region, that is, a second flame region in the second image is detected.

[0093] Since the use environment of the embodiment is relatively dark, the flame can be regarded as the only light source. Once the flame burns intensely, the RGB algorithm will not be able to accurately detect the intensely burning flame region, and therefore the second image needs to be further identified. According to experiments, under the condition that the flame is the only light source, the intensely burning flame should satisfy the following conditions at the same time:

[0094] Rule 1: 215≤R≤255

[0095] Rule 2: 215≤G≤255

[0096] Rule 3: 10≤B≤255

[0097] When the three conditions are satisfied at the same time, the intensely burning flame can be determined, and is recorded as a second flame region, as shown in the upper left part of FIG. 6. Figure 8

[0098] ​S7: performing a union operation on the first flame region and the second flame region to obtain a flame image; that is, taking the union of the first flame region and the second flame region in the second image to obtain the flame image, as shown in the lower right part of FIG. 6. Figure 8

[0099] S8: calculating the total number of pixel points of the flame region in the flame image to obtain a flame area. The flame region is the region after taking the union of the first flame region and the second flame region, and thus the flame area is the total number of pixel points after taking the union of the first flame region and the second flame region.

[0100] Further, an average gray value of the flame region in the flame image is calculated to obtain a flame brightness.

[0101] Further, a difference operation is performed on the flame images corresponding to two adjacent video frames, that is, the flame image corresponding to the one video frame is subtracted from the flame image corresponding to the previous video frame to obtain a frame difference image; then an average pixel value is calculated according to the pixel values of the pixel points in the frame difference image; and finally the average pixel value is divided by a preset value to obtain a flame frequency corresponding to the one video frame. In this embodiment, the preset value is 21, which is a relatively stable value for representing the flame frequency after multiple experiments.

[0102] By calculating the flame brightness and the flame frequency, the flame burning condition can be reflected to a certain extent, so that the related industrial parameters can be adjusted based on the flame burning condition.

[0103] S9: performing a grayscale processing on the second image to obtain a third image; that is, converting the second image into a grayscale image, as shown in FIG. 7. Figure 9

[0104] S10: performing a zeroing processing on the pixel values of the pixel points with a pixel value greater than a preset smoke pixel threshold value in the third image to obtain a smoke image, and calculating the number of pixel points with a pixel value not equal to zero in the smoke image to obtain a smoke area.

[0105] In this embodiment, the smoke pixel threshold value is 62; and the smoke image is as shown in FIG. 8. Figure 10

[0106] S11: performing an exclusive OR operation on the second image and the flame image to obtain a fourth image.

[0107] Specifically, in the second image and the flame image, if the pixel values of the pixel points at the same position are different, the pixel value of the pixel point at the same position in the second image is retained, and if the pixel values of the pixel points at the same position are the same, the pixel value of the pixel point at the same position is set to 0, so as to obtain the fourth image, as shown in FIG. 9. Figure 11 ​​​​

[0108] That is, the exclusive or operation of the images is to filter the same parts in the two images and obtain the differences between the two images. This step is to remove the flame area in the second image.

[0109] S12: The pixel value of the pixel point with a pixel value less than the preset slag pixel threshold value in the fourth image is zeroed to obtain a slag image, and the number of pixel points with a pixel value not equal to zero in the slag image is calculated to obtain a slag area.

[0110] In this embodiment, the slag pixel threshold value is 35.

[0111] Further, according to common sense, if the slag flows out of the furnace mouth, it must flow downward, so the flame and the area above the flame in the fourth image can be shielded, and the slag image obtained at this time is as shown in Figure 12 The first image corresponding to the slag image is as shown in Figure 13 .

[0112] S13: According to the flame area, the smoke area, the slag area, and the preset warning threshold value, the abnormality of the converter steelmaking state is detected.

[0113] The warning threshold value includes a flame area threshold value, a smoke area threshold value, and a slag area threshold value; in this embodiment, the flame area threshold value is 135%, the smoke area threshold value is 30%, and the slag area threshold value is 10000.

[0114] Specifically, it is determined whether the flame area, the smoke area, and the slag area exceed the corresponding warning threshold value, respectively. If the percentage of the ratio of the flame area to the area of the furnace mouth region (i.e., the total number of pixel points in the furnace mouth region) is greater than the flame area threshold value, the percentage of the ratio of the smoke area to the area of the furnace mouth region is greater than the smoke area threshold value, or the slag area is greater than the slag area threshold value, it is determined that the converter steelmaking state is abnormal, the state parameters (including the flame area, the smoke area, and the slag area, and the flame brightness and the flame frequency) at this time are saved, the RTSP video stream is saved, an abnormal report is generated, and subsequent abnormal video playback can be performed according to different abnormal states.

[0115] Further, according to the flame area, the flame brightness, the flame frequency, the smoke area, and the slag area, the related industrial parameters are adjusted, and the industrial parameters include at least one of a gun position value, an oxygen flow, a carbon dioxide concentration, and a fan speed.

[0116] In the converter steelmaking process, the frequency of flame combustion is of vital importance to the quality of the final product, such as the fire is big or small, or the fire is strong or not, or the fire is big but not fierce, small but fierce, etc. By calculating the flame frequency, the generalized fire state is converted into specific numerical values, which can be used to adjust industrial parameters such as oxygen lance, carbon dioxide concentration, fan speed, etc. In addition, when the flame brightness is very bright, it can be explained from some aspects that the flame combustion is very intense at this time, and the relevant operating personnel need to adjust the oxygen flow or other industrial parameters when the combustion is intense.

[0117] At the same time, according to the experience of converter steelmaking and daily life, when the flame area is large, it means that the internal combustion of the converter is intense, which will affect the quality of the final product. At this time, the system user needs to pay attention at all times, and manual operation can prevent problems from happening. Similarly, when the smoke area is large, it means that there is an insufficient combustion phenomenon inside, which needs to be adjusted in a timely manner. When there is slag spatter or a large amount of slag overflow, if not stopped in time, it will endanger the production environment and cause economic losses, and even more serious, it may cause casualties. Therefore, by monitoring the flame area, flame brightness, flame frequency, smoke area and slag area, the production safety management can be strengthened and unnecessary economic losses of enterprises can be reduced.

[0118] Further, the first image, the flame image, the smoke image and the slag image are spliced to obtain a front-end display image, and displayed on the front-end interface, as shown in Figure 14

[0119] Further, the front-end display image, the flame area, the smoke area and the slag area and other results can be transmitted to the client through network communication.

[0120] The embodiment reads real-time video stream to process and identify key frame images, uses computer vision and graphics operations to detect and analyze the flame, smoke and slag of the converter mouth area, realizes automatic discrimination of the converter steelmaking state, and saves labor cost.

[0121] By detecting the flame area of ordinary combustion and intense combustion respectively, and taking the union of the two as the final flame detection result, the accuracy of flame detection is improved; by setting the threshold values of smoke and slag detection as variable parameters, industry experts can dynamically adjust to the best segmentation interval according to their experience, greatly improving the accuracy and robustness of system detection.

[0122] ​In addition, a production early warning system is designed according to the analysis of production factors, which improves the security while ensuring the accuracy. Meanwhile, by connecting the real-time database, important parameters can be freely added on the front-end page and monitored, and the abnormal rules of production factors are set according to the expert experience, and the abnormal data is written into the real-time database to form an abnormal information report, which facilitates the process backtracking of the system user, thereby greatly improving the robustness of the system.

[0123] Embodiment two

[0124] The embodiment is a computer readable storage medium corresponding to the above embodiment, which stores a computer program, the program is executed by a processor to realize the steps of the converter steelmaking state detection method in the above embodiment, and the same technical effect can be achieved, which is not repeated here.

[0125] In summary, the converter steelmaking state detection method and the computer readable storage medium provided by the application can remove the monitoring character information on the video frame, avoid the influence of character information on subsequent converter steelmaking state detection, and ensure the accuracy of detection. The accuracy of flame detection can be improved by detecting the flame area of ordinary combustion and intense combustion respectively and taking the union of the two as the final flame detection result, thereby improving the accuracy of converter steelmaking state detection. The accuracy and robustness of detection are improved by setting the threshold values of smoke and slag detection as variable parameters, so that industry experts can dynamically adjust to the best segmentation interval according to their own experience. The accuracy of converter steelmaking state detection can be further improved by detecting the abnormality of the converter steelmaking state according to the flame area, smoke area and slag area. By adjusting the relevant industrial parameters according to the flame area, flame brightness, flame frequency, smoke area and slag area, the relevant industrial parameters can be adjusted in time when there is excessive intense combustion, insufficient combustion or a large amount of slag overflow, thereby avoiding affecting the quality of the final product and also avoiding economic losses and personnel casualties.

[0126] The application realizes the automatic discrimination of the converter steelmaking state, saves the labor cost, and improves the accuracy of converter steelmaking state detection.

[0127] The above description is only an embodiment of the application, and does not limit the patent range of the application, and any equivalent transformation or direct or indirect application in the related technical field according to the content of the specification and drawings of the application is also included in the patent protection range of the application.

Claims

1. A method for detecting the condition of steelmaking in a converter, characterized in that, include: Read a video frame from the real-time furnace monitoring video stream to obtain the first image; Remove the character information from the first image to obtain the second image; The second image is subjected to flame detection using the RGB algorithm to obtain the first flame region; By using preset intense burning flame conditions, the second image is subjected to intense burning flame detection to obtain the second flame region; the intense burning flame conditions include 215≤R≤255, 215≤G≤255 and 10≤B≤255, where R, G and B are the three color components of the pixel. Perform a union operation on the first flame region and the second flame region to obtain a flame image; Calculate the total number of pixels in the flame region of the flame image to obtain the flame area; Based on the flame area, anomaly detection is performed on the converter steelmaking status; After performing a union operation on the first flame region and the second flame region to obtain a flame image, the process further includes: The second image is converted to grayscale to obtain the third image; The pixel values ​​of pixels in the third image that are greater than a preset smoke pixel threshold are zeroed to obtain a smoke image, and the number of pixels in the smoke image with non-zero pixel values ​​is calculated to obtain the smoke area. Perform an XOR operation on the second image and the flame image to obtain the fourth image; The pixel values ​​of pixels in the fourth image whose pixel values ​​are less than a preset slag pixel threshold are zeroed to obtain a slag image, and the number of pixels in the slag image whose pixel values ​​are not zero is calculated to obtain the slag area. The specific steps for detecting abnormalities in the converter steelmaking state based on the flame area are as follows: Based on the flame area, smoke area, and slag area, as well as the preset warning thresholds, it is determined whether the converter steelmaking status is abnormal; wherein, the warning thresholds include the flame area threshold, the smoke area threshold, and the slag area threshold. If so, an anomaly report is generated based on the flame area, smoke area, slag area, and the real-time furnace opening monitoring video stream.

2. The method for detecting the state of converter steelmaking according to claim 1, characterized in that, Before performing flame detection on the second image using the RGB algorithm to obtain the first flame region, the method further includes: When the gun position value and oxygen flow rate are detected to meet the preset furnace opening area selection conditions, the furnace opening area is selected for the second image. The preset furnace opening area selection conditions include the gun position value being in a preset first interval and the oxygen flow rate being in a preset second interval. Determine whether the selected furnace opening area meets the preset furnace opening area size conditions. The furnace opening area size conditions include that the length of the outer rectangle of the selected furnace opening area is within a preset length range, the width of the outer rectangle is within a preset width range, and the total number of pixels in the selected furnace opening area is within a preset area range. If the conditions are met, then the step of performing flame detection on the second image using the RGB algorithm to obtain the first flame region is executed.

3. The method for detecting the state of converter steelmaking according to claim 2, characterized in that, The first interval is [2000, 2250], the second interval is [25000, 31000]; the length range is [1400, 1950], the width range is [130, 350], and the area range is [130000, 290000].

4. The method for detecting the state of converter steelmaking according to claim 1, characterized in that, After performing anomaly detection on the converter steelmaking status based on the flame area, the process further includes: The first image, flame image, smoke image, and slag image are stitched together to obtain the front-end display image.

5. The method for detecting the state of converter steelmaking according to claim 1, characterized in that, After performing a union operation on the first flame region and the second flame region to obtain a flame image, the process further includes: Calculate the average grayscale value of the flame region in the flame image to obtain the flame brightness; Perform a difference operation on the flame image corresponding to the video frame and the previous video frame to obtain a frame difference image; Calculate the average pixel value of the frame difference image, and divide the average pixel value by a preset value to obtain the flame frequency corresponding to the video frame. The preset value is 21.

6. The method for detecting the state of converter steelmaking according to claim 5, characterized in that, The method further includes: Based on the flame area, flame brightness, flame frequency, smoke area, and slag area, relevant industrial parameters are adjusted, including at least one of the following: lance position value, oxygen flow rate, carbon dioxide concentration, and fan speed.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-6.

Images