A furnace mouth image recognition method, device, equipment and storage medium

Abstract

The present application discloses a furnace mouth image recognition method, device, equipment and storage medium, the method comprising: real-time acquisition of furnace mouth images; pre-processing of furnace mouth images to obtain furnace mouth areas; obtaining the area and number of slag flakes in the furnace mouth area, and obtaining pre-splashing recognition results according to the area and number of slag flakes in a first set time period; translating up and down along the height of the furnace mouth area to obtain two splashing detection areas, performing threshold processing on the images of the two splashing detection areas, obtaining the corresponding areas and comparing them with the corresponding set area thresholds respectively, and obtaining splashing recognition results according to the comparison results in a second set time period. In this way, by extracting, analyzing and judging the information and features contained in the furnace mouth image, the state of pre-splashing and splashing can be identified, which can be well used to control the feeding method and oxygen gun adjustment method in the smelting process, thereby achieving the purpose of automated steelmaking, and has the characteristics of accurate prediction and fast processing speed.

Description

Technical Field

[0001] The present invention relates to the field of image recognition, and in particular to a furnace mouth image recognition method, device, equipment and storage medium. Background Art

[0002] BOF steelmaking is the current mainstream steelmaking method. During the BOF steelmaking process, steelmakers generally observe the state of the slag and flame at the furnace mouth with their eyes to control the gun position and the addition of auxiliary materials, so as to blow a furnace of steel to meet the steelmaking standards.

[0003] However, identifying the state of the slag splashing and pre-splashing at the furnace mouth by human eyes relies on the operator's human experience and judgment, which results in operational differences, low prediction accuracy, and slow processing speed.

[0004] Therefore, how to solve the problems of low prediction accuracy and slow processing speed of furnace mouth splashing and pre-splattering identification is a technical problem that needs to be solved urgently by technical personnel in this field. Summary of the invention

[0005] In view of this, the purpose of the present invention is to provide a furnace mouth image recognition method, device, equipment and storage medium, which can achieve the purpose of automated steelmaking, with high prediction accuracy for pre-splashing and splashing recognition and fast processing speed. The specific scheme is as follows:

[0006] A furnace mouth image recognition method, comprising:

[0007] Real-time acquisition of furnace mouth images;

[0008] Preprocessing the furnace mouth image to obtain a furnace mouth area of ​​the furnace mouth image;

[0009] Acquire the area and number of slag pieces in the furnace mouth area, and obtain a pre-splashing identification result according to the area and number of slag pieces in a first set time period;

[0010] Two splash detection areas are obtained by translating up and down along the height of the furnace mouth area. After threshold processing is performed on the images of the two splash detection areas, the corresponding areas are obtained and compared with the corresponding set area thresholds respectively, and the splash recognition result is obtained according to the comparison results within the second set time period.

[0011] Preferably, in the furnace mouth image recognition method provided in the embodiment of the present invention, preprocessing the furnace mouth image to obtain the furnace mouth area of ​​the furnace mouth image specifically includes:

[0012] De-noising the furnace mouth image by using median filtering;

[0013] Performing a scale transformation on the furnace mouth image, and enhancing the brightness of the furnace mouth image by a scale transformation factor;

[0014] Segmenting the furnace mouth image and obtaining a segmentation threshold by using an OTSU automatic threshold algorithm;

[0015] According to the mean value of the furnace mouth image after segmentation, the segmentation threshold is adjusted and the furnace mouth image is binarized to obtain the furnace mouth area of ​​the furnace mouth image.

[0016] Preferably, in the furnace mouth image recognition method provided in the embodiment of the present invention, obtaining the area and number of slag pieces in the furnace mouth area specifically includes:

[0017] Perform edge detection on the furnace mouth area by using the Canny edge operator to obtain the slag area and number of edge detection results;

[0018] The furnace mouth image is transformed by a Gaussian derivative matrix to obtain a derivative graph of the Hessian rule, and features of the derivative graph are extracted and processed to obtain the area and number of slag pieces in the derivative graph;

[0019] The area and number of slag pieces in the edge detection result and the area and number of slag pieces in the derived graph are subjected to a union operation to obtain the area and number of slag pieces in the furnace mouth area.

[0020] Preferably, in the furnace mouth image recognition method provided in the embodiment of the present invention, edge detection is performed on the furnace mouth area by using a Canny edge operator to obtain the slag sheet area and number of edge detection results, specifically including:

[0021] Extracting the contour of the slag in the furnace mouth area by using the Canny edge operator;

[0022] Perform roundness screening on the extracted contour lines;

[0023] Convert the filtered contours into regions and fill them;

[0024] The filled area is screened, and an area with an area within a set interval is selected, and the area is marked in the furnace mouth image to obtain the identified slag pieces, and the area and number of the slag pieces are obtained.

[0025] Preferably, in the furnace mouth image recognition method provided in the embodiment of the present invention, the roundness screening of the extracted contour line specifically includes:

[0026] Remove the non-arc edges from the extracted contours;

[0027] The remaining contour lines are screened for closed contours to obtain contour lines whose distance between the starting point and the ending point is less than the set distance threshold.

[0028] Preferably, in the furnace mouth image recognition method provided in the embodiment of the present invention, the features of the derived graph are extracted and processed to obtain the area and number of slag pieces in the derived graph, which specifically includes:

[0029] Perform background removal on the derived image;

[0030] Filter out areas where the pixel value is greater than the set pixel threshold;

[0031] The furnace mouth border and area screening are performed on the screened area to obtain the identified slag pieces, and the area and number of the slag pieces are obtained.

[0032] Preferably, in the furnace mouth image recognition method provided in the embodiment of the present invention, obtaining the splash recognition result according to the comparison result within the second set time period specifically includes:

[0033] When the area of ​​the splash detection area obtained by moving downward after threshold processing is greater than the first set area threshold, and the area of ​​the splash detection area obtained by moving upward after threshold processing is less than the second set area threshold, and the number of the furnace mouth images that meet this condition within the second set time period is greater than the set number threshold, it is determined that the furnace mouth image contains splashing; or,

[0034] When the area of ​​the splash detection area obtained by moving downward after threshold processing is greater than half of the total area of ​​the furnace mouth area, and the area of ​​the splash detection area obtained by moving upward after threshold processing is less than the second set area threshold, it is determined that the furnace mouth image contains splashing phenomenon.

[0035] The embodiment of the present invention further provides a furnace mouth image recognition device, comprising:

[0036] Image acquisition module, used for real-time acquisition of furnace mouth images;

[0037] An image processing module, used for preprocessing the furnace mouth image to obtain a furnace mouth area of ​​the furnace mouth image;

[0038] A pre-splashing identification module, used to obtain the area and number of slag pieces in the furnace mouth area, and obtain a pre-splashing identification result according to the area and number of slag pieces in a first set time period;

[0039] The splash recognition module is used to translate up and down along the height of the furnace mouth area to obtain two splash detection areas, perform threshold processing on the images of the two splash detection areas, obtain the corresponding areas and compare them with the corresponding area thresholds respectively, and obtain the splash recognition result according to the comparison results within the second set time period.

[0040] An embodiment of the present invention further provides a furnace mouth image recognition device, comprising a processor and a memory, wherein the processor implements the above-mentioned furnace mouth image recognition method provided in the embodiment of the present invention when executing the computer program stored in the memory.

[0041] An embodiment of the present invention further provides a computer-readable storage medium for storing a computer program, wherein when the computer program is executed by a processor, the furnace mouth image recognition method provided in the embodiment of the present invention is implemented.

[0042] It can be seen from the above technical scheme that a furnace mouth image recognition method provided by the present invention includes: real-time acquisition of furnace mouth images; preprocessing of the furnace mouth images to obtain a furnace mouth area of ​​the furnace mouth image; obtaining the area and number of slag pieces in the furnace mouth area, and obtaining a pre-splattering recognition result based on the area and number of slag pieces within a first set time period; translating up and down along the height of the furnace mouth area to obtain two splashing detection areas, performing threshold processing on the images of the two splashing detection areas, obtaining the corresponding areas and comparing them with the corresponding set area thresholds respectively, and obtaining the splashing recognition result based on the comparison results within a second set time period.

[0043] The present invention extracts, analyzes and judges the information and features contained in the furnace mouth image, thereby obtaining the state of pre-splashing and splashing. In this way, the pre-splashing and splashing recognition of the furnace mouth image can be well used to control the feeding method and oxygen gun adjustment method in the smelting process, replacing the operation performed by the operator through visual observation, thereby achieving the purpose of automated steelmaking. The method has the characteristics of accurate prediction and fast processing speed. In addition, the present invention also provides a corresponding device, equipment and computer-readable storage medium for the furnace mouth image recognition method, further making the above method more practical, and the device, equipment and computer-readable storage medium have corresponding advantages. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or related technologies, the drawings required for use in the embodiments or related technical descriptions are briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0045] Figure 1 A flow chart of a furnace mouth image recognition method provided by an embodiment of the present invention;

[0046] Figure 2 A specific flow chart of the furnace mouth image recognition method provided by an embodiment of the present invention;

[0047] Figure 3The furnace mouth image before scale transformation provided by the embodiment of the present invention;

[0048] Figure 4 A furnace mouth image after scale transformation provided by an embodiment of the present invention;

[0049] Figure 5 A furnace mouth image after automatic threshold segmentation provided by an embodiment of the present invention;

[0050] Figure 6 A binarized furnace mouth image provided by an embodiment of the present invention;

[0051] Figure 7 An original furnace mouth image corresponding to the furnace mouth pre-splashing provided in an embodiment of the present invention;

[0052] Figure 8 An image after Canny sub-pixel edge extraction provided by an embodiment of the present invention;

[0053] Fig. 9 An image after roundness screening provided by an embodiment of the present invention;

[0054] Fig.10 An image after closed contours are screened out provided by an embodiment of the present invention;

[0055] Fig.11 The slag image identified by the Canny edge operator processing provided in the embodiment of the present invention;

[0056] Fig.12 An image converted into a Hessian matrix provided by an embodiment of the present invention;

[0057] Fig.13 An image with a protruding area after background removal provided by an embodiment of the present invention;

[0058] Fig.14 An image having a protruding area after area screening provided by an embodiment of the present invention;

[0059] Fig.15 A furnace mouth image after overall pre-splashing recognition provided by an embodiment of the present invention;

[0060] Fig.16 The recognition area map obtained after the furnace mouth area is translated up and down provided by the embodiment of the present invention;

[0061] Fig.17 A basic diagram of the furnace mouth fire spraying provided by the embodiment of the present invention;

[0062] Fig.18 A splashing image under the condition of slag accumulation at the furnace mouth provided by an embodiment of the present invention;

[0063] Fig.19 Acquiring an image of a splash detection area in the case of slag accumulation and splashing provided by an embodiment of the present invention;

[0064] Fig. 20 The pre-splash recognition result under normal conditions provided by the embodiment of the present invention;

[0065] Fig.21 The pre-splashing recognition result under the condition of slag accumulation at the furnace mouth provided by the embodiment of the present invention;

[0066] Fig. 22 The pre-splash recognition result under the condition of smoke interference provided by the embodiment of the present invention;

[0067] Fig.23 The splash recognition result under normal conditions provided by the embodiment of the present invention;

[0068] Fig.24 The splash recognition result under the condition of smoke interference provided by the embodiment of the present invention;

[0069] Fig.25 The splashing recognition result under the condition of slag accumulation at the furnace mouth provided by the embodiment of the present invention;

[0070] Fig.26 A schematic structural diagram of a furnace mouth image recognition device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0071] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0072] The present invention provides a furnace mouth image recognition method, such as Figure 1 As shown, the following steps are included:

[0073] S101, collecting furnace mouth images in real time.

[0074] S102, preprocessing the furnace mouth image to obtain a furnace mouth area of ​​the furnace mouth image.

[0075] It should be noted that due to the relatively harsh blowing environment, the collected furnace mouth images will be subject to more interference factors. In order to accurately obtain the furnace mouth area, some preprocessing of the image is required.

[0076] S103, obtaining the area and number of slag pieces in the furnace mouth area, and obtaining a pre-splashing identification result according to the area and number of slag pieces in a first set time period.

[0077] In practical applications, pre-splashing is actually a situation where slag flakes are thrown out of the furnace mouth during converter smelting. In order to more accurately extract the slag flakes thrown out of the furnace mouth, the present invention can perform comprehensive feature extraction and judgment on the processing results of the canny edge detection operator and Gaussian derivative matrix processing to identify the pre-splashing phenomenon. The first set time period here can be set to within the current 2 seconds.

[0078] S104, translate up and down along the height of the furnace mouth area to obtain two splash detection areas, perform threshold processing on the images of the two splash detection areas, obtain the corresponding areas and compare them with the corresponding set area thresholds respectively, and obtain the splash recognition result according to the comparison results within the second set time period. The second set time period here can be set to within the current 1 second.

[0079] In the above-mentioned furnace mouth image recognition method provided in an embodiment of the present invention, the information and features contained in the furnace mouth image are extracted, analyzed and judged to obtain the status of pre-spattering and splashing. In this way, the pre-spattering and splashing recognition of the furnace mouth image can be well used to control the feeding method and oxygen gun adjustment method in the smelting process, replacing the operation performed by the operator through visual observation, thereby achieving the purpose of automated steelmaking. This method has the characteristics of accurate prediction and fast processing speed.

[0080] Furthermore, in the specific implementation, in the above-mentioned furnace mouth image recognition method provided by the embodiment of the present invention, if Figure 2 As shown, step S102 preprocesses the furnace mouth image to obtain the furnace mouth area of ​​the furnace mouth image, which may specifically include: first, using a median filter to remove noise from the furnace mouth image; then, scaling the furnace mouth image, and enhancing the brightness of the furnace mouth image by the scaling factor; then, segmenting the furnace mouth image by the OTSU automatic threshold algorithm and obtaining the segmentation threshold; finally, according to the mean of the segmented furnace mouth image, adjusting the segmentation threshold and binarizing the furnace mouth image to obtain the furnace mouth area of ​​the furnace mouth image. In order to achieve a better processing effect, after obtaining the furnace mouth area of ​​the furnace mouth image, the furnace mouth image may also be equalized according to the mean and variance of the furnace mouth area to accurately obtain basic features such as the width and height of the furnace mouth.

[0081] Specifically, Figures 3 to 6 For example, first Figure 3 The I_i image before transformation is shown by Get Figure 4 The transformed image I_o shown in the figure has the brightness enhanced by the scale factor num_scale. Then the OTSU automatic threshold algorithm is used to obtain Figure 5 The segmented image and segmentation threshold are shown according to Figure 5The segmentation threshold is adjusted by the mean value of the image and the furnace mouth is binarized to obtain Figure 6 The furnace mouth area of ​​the furnace mouth image is shown.

[0082] Furthermore, in the specific implementation, in the furnace mouth image recognition method provided in the embodiment of the present invention, step S103 of acquiring the area and number of slag pieces in the furnace mouth area may specifically include the following steps:

[0083] Step 1: Use the Canny edge operator to perform edge detection on the furnace mouth area and obtain the slag area and number of edge detection results.

[0084] Specifically, the contour lines of the slag pieces in the furnace mouth area can be first extracted by the Canny edge operator; then the extracted contour lines can be screened for roundness; the screened contour lines can be converted into regions and filled; finally, the filled regions can be screened for area, and the regions with areas within a set interval can be selected, and the regions can be marked in the furnace mouth image to obtain the identified slag pieces, and the area and number of the slag pieces can be obtained.

[0085] Figure 7 An image of a furnace mouth containing actual slag pieces is shown. Figure 8 The figure shows the contour line of canny edge sub-pixel extraction obtained by setting the corresponding parameters. The roundness value of the extracted contour line can be [0.04559, 1], and the roundness num_r is obtained as follows: F is the area of ​​the contour, and m_dis is the pixel distance from the center of the area to each boundary. Fig. 9 and Fig.10 As shown in the figure, roundness screening can remove those non-arc edges in the contour line, and then perform closed contour screening on the remaining contour lines. The so-called closed contour is obtained by calculating the distance between the starting point and the end point of a contour line and filtering out the points whose distance is greater than the set distance threshold, thereby obtaining a relatively closed contour. The contour line is converted into an area and filled, and then area screening is performed to select the area within a certain threshold range, such as Fig.11 As shown, the area is marked in the image to obtain the identified slag pieces.

[0086] Step 2: The furnace mouth image is transformed by Gaussian derivative matrix to obtain the derivative graph of Hessian rule, and the features of the derivative graph are extracted and processed to obtain the area and number of slag pieces in the derivative graph.

[0087] Specifically, the furnace mouth area is firstly cut out from the furnace mouth image, and then Fig.12 As shown, the image is filtered by Gaussian with sigma = 2. By converting its image into a derivative image (i.e., Hessian matrix image), the area with obvious pixel transformation in the derivative image can be more clearly and intuitively displayed. Next, in the process of processing the derivative image, the background of the derivative image can be removed first; Fig.13 As shown, the background pixels of the matrix image are around 0. The more abrupt the place, the larger the pixel value. At this time, the area with a pixel value greater than the set pixel threshold is screened out, such as Fig.14 As shown, the protruding area in the image can be obtained by selecting a threshold value greater than 2; finally, the furnace mouth frame and area screening are performed on the screened area to obtain the identified slag pieces, and the area and number of the slag pieces are obtained.

[0088] Step 3: perform a union operation on the slag sheet area and number obtained from the edge detection result and the slag sheet area and number in the derived graph to obtain the slag sheet area and number in the furnace mouth area.

[0089] Specifically, the result of the canny edge operator processing is combined with the result of the Gaussian derivative matrix processing, and the regions obtained by the two methods are combined to obtain the following: Fig.15 The status and position of the pre-splash identification block shown in the furnace mouth image.

[0090] Further, in the specific implementation, in the furnace mouth image recognition method provided in the embodiment of the present invention, during the execution of step S104, the furnace mouth area obtained by preprocessing is translated up and down to obtain the following Fig.16 The splash detection area 1 and splash detection area 2 shown in the figure are due to the existence of splash detection area 1 and splash detection area 2 in the furnace mouth blowing process. Fig.17 The flame spraying situation shown in the figure will seriously affect the judgment of the furnace mouth and the splashing. However, the biggest characteristic difference between flame spraying and splashing is that the flame has an upward trend. Therefore, the flame spraying can be filtered and screened by the splashing detection area 2 obtained by moving up. In addition, the flame shrinks very quickly, and the splashing liquid cools much more slowly. This point is also used to store the splashing images that meet the conditions in the array for comprehensive judgment, so as to achieve splashing recognition.

[0091] It should be noted that when there is a lot of slag accumulation at the furnace mouth, making the furnace mouth area very small and difficult to identify, the splashing in this case is Fig.18 As shown in FIG. 1 , the black elliptical area in the image is selected by screening the threshold and coordinate features (if there is no obvious furnace mouth and no black elliptical area, the lower 1 / 3 area in the image is taken as the splash detection area 3), and the uppermost row coordinate is used as the starting row coordinate as shown in FIG. Fig.19 The fixed splash detection area 3 with a height of 200 pixels is planned based on the white dot position in the fixed splash detection area 3, and the part with a pixel value greater than the threshold value 30 is obtained from the fixed splash detection area 3, and its area is calculated and stored. The comprehensive judgment can be

[0092]

[0093] Among them, Area1, Area2 and Area3 are the threshold areas that meet the conditions in splash detection area 1, splash detection area 2 and splash detection area 3 respectively, TH_area1, TH_area2 and TH_area3 are the first set area threshold corresponding to splash detection area 1, the second set area threshold corresponding to splash detection area 2 and the third area threshold corresponding to splash detection area 3 respectively, num is the number of images that meet the conditions of (Area1>TH_area1)&&(Area2<TH_area2) or (Area3>TH_area3)&&(Area2<TH_area2) in the current 1s, TH_num is the set number threshold, and TH_area_all is the total area of ​​the furnace mouth area. When one of the above three conditions is met, it is considered to be a splash phenomenon.

[0094] That is to say, step S104 obtains the splash recognition result according to the comparison result within the second set time period, which may specifically include: when the area Area1 of the splash detection area 1 obtained by moving down after threshold processing is greater than the first set area threshold TH_area1, and the area Area2 of the splash detection area 2 obtained by moving up after threshold processing is less than the second set area threshold TH_area2, and the number of furnace mouth images that meet this situation within the second set time period is greater than the set number threshold TH_num, then it is determined that the furnace mouth image contains splashing; or, when the area Area3 of the splash detection area 3 obtained after threshold processing is greater than the third set area threshold TH_area3, and when the area Area2 obtained by moving up after threshold processing of the splash detection area 2 is less than the second set area threshold TH_area2, and the number of furnace mouth images that meet this situation within the second set time period is greater than the set number threshold TH_num, then it is determined that the furnace mouth image contains splashing; or, when the area Area1 obtained by moving down after threshold processing of the splash detection area 1 is greater than half of the total area TH_area_all of the furnace mouth area, and the area Area2 obtained by moving up after threshold processing of the splash detection area 2 is less than the second set area threshold TH_area2, then it is determined that the furnace mouth image contains splashing.

[0095] After the test, no matter it is the pre-splashing under normal conditions, or the pre-splashing under the condition of a certain amount of slag accumulation at the furnace mouth, or the pre-splashing under the condition of large smoke interference, it can be accurately identified. The identification results are as follows: Fig. 20 , Fig.21 and Fig. 22Similarly, whether it is splashing in normal conditions, splashing under the interference of smoke, or splashing under the condition of slag accumulation at the furnace mouth, it can be accurately identified, and the identification results are as follows: Fig.23 , Fig.24 and Fig.25 Therefore, the furnace mouth image recognition method provided by the embodiment can accurately recognize the results of pre-splashing or splashing in various situations.

[0096] Based on the same inventive concept, an embodiment of the present invention also provides a furnace mouth image recognition device. Since the principle of solving the problem by the device is similar to that of the aforementioned furnace mouth image recognition method, the implementation of the device can refer to the implementation of the furnace mouth image recognition method, and the repeated parts will not be repeated.

[0097] In a specific implementation, the furnace mouth image recognition device provided by the embodiment of the present invention is as follows: Fig.26 As shown, specifically including:

[0098] An image acquisition module 11 is used to acquire furnace mouth images in real time;

[0099] An image processing module 12 is used to pre-process the furnace mouth image to obtain a furnace mouth area of ​​the furnace mouth image;

[0100] A pre-splashing identification module 13 is used to obtain the area and number of slag pieces in the furnace mouth area, and obtain a pre-splashing identification result according to the area and number of slag pieces in a first set time period;

[0101] The splash recognition module 14 is used to translate up and down along the height of the furnace mouth area to obtain two splash detection areas, perform threshold processing on the images of the two splash detection areas, obtain the corresponding areas and compare them with the corresponding area thresholds respectively, and obtain the splash recognition result according to the comparison results within the second set time period.

[0102] In the above-mentioned furnace mouth image recognition device provided in the embodiment of the present invention, the states of pre-spattering and splashing can be obtained through the interaction of the above-mentioned four modules, and then can be well used to control the feeding method and oxygen gun adjustment method in the smelting process, to replace the operator's operation through visual observation, so as to achieve the purpose of automated steelmaking, and has the characteristics of accurate prediction and fast processing speed.

[0103] For more specific working processes of the above modules, please refer to the corresponding contents disclosed in the above embodiments, which will not be repeated here.

[0104] Correspondingly, an embodiment of the present invention further discloses a furnace mouth image recognition device, comprising a processor and a memory; wherein the furnace mouth image recognition method disclosed in the aforementioned embodiment is implemented when the processor executes a computer program stored in the memory.

[0105] For more specific processes of the above method, please refer to the corresponding contents disclosed in the aforementioned embodiments, which will not be repeated here.

[0106] Furthermore, the present invention also discloses a computer-readable storage medium for storing a computer program; when the computer program is executed by a processor, the furnace mouth image recognition method disclosed above is implemented.

[0107] For more specific processes of the above method, please refer to the corresponding contents disclosed in the aforementioned embodiments, which will not be repeated here.

[0108] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the devices, equipment, and storage media disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method part.

[0109] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0110] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0111] A furnace mouth image recognition method provided by an embodiment of the present invention comprises: real-time acquisition of furnace mouth images; preprocessing of the furnace mouth images to obtain the furnace mouth area of ​​the furnace mouth images; obtaining the area and number of slag flakes in the furnace mouth area, and obtaining a pre-splashing recognition result according to the area and number of slag flakes in a first set time period; translating up and down along the height of the furnace mouth area to obtain two splashing detection areas, performing threshold processing on the images of the two splashing detection areas, obtaining the corresponding areas and comparing them with the corresponding set area thresholds respectively, and obtaining a splashing recognition result according to the comparison results in a second set time period. In this way, by extracting, analyzing and judging the information and features contained in the furnace mouth image, the state of pre-splashing and splashing can be identified, which can be well used to control the feeding method and oxygen gun adjustment method in the smelting process, replacing the operation performed by the operator through visual observation, thereby achieving the purpose of automated steelmaking, and the method has the characteristics of accurate prediction and fast processing speed. In addition, the present invention also provides a corresponding device, equipment and computer-readable storage medium for the furnace mouth image recognition method, which further makes the above method more practical, and the device, equipment and computer-readable storage medium have corresponding advantages.

[0112] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0113] The furnace mouth image recognition method, device, equipment and storage medium provided by the present invention are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for general technical personnel in this field, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A furnace mouth image recognition method, characterized in that: include: Real-time acquisition of furnace mouth images; Preprocessing the furnace mouth image to obtain a furnace mouth area of ​​the furnace mouth image; Perform edge detection on the furnace mouth area by using the Canny edge operator to obtain the slag area and number of edge detection results; The furnace mouth image is transformed by a Gaussian derivative matrix to obtain a derivative graph of the Hessian rule, and features of the derivative graph are extracted and processed to obtain the area and number of slag pieces in the derivative graph; The obtained slag sheet area and number in the edge detection result and the slag sheet area and number in the derived graph are subjected to a union operation to obtain the slag sheet area and number in the furnace mouth area, and the pre-splashing recognition result is obtained according to the slag sheet area and number in the first set time period; Two splash detection areas are obtained by translating up and down along the height of the furnace mouth area. After threshold processing is performed on the images of the two splash detection areas, the corresponding areas are obtained and compared with the corresponding set area thresholds respectively, and the splash recognition result is obtained according to the comparison results within the second set time period.

2. The furnace mouth image recognition method according to claim 1, characterized in that: Preprocessing the furnace mouth image to obtain a furnace mouth area of ​​the furnace mouth image specifically includes: De-noising the furnace mouth image by using median filtering; Performing a scale transformation on the furnace mouth image, and enhancing the brightness of the furnace mouth image by a scale transformation factor; Segmenting the furnace mouth image and obtaining a segmentation threshold by using an OTSU automatic threshold algorithm; According to the mean value of the furnace mouth image after segmentation, the segmentation threshold is adjusted and the furnace mouth image is binarized to obtain the furnace mouth area of ​​the furnace mouth image.

3. The furnace mouth image recognition method according to claim 1, characterized in that: The furnace mouth area is subjected to edge detection by using the Canny edge operator to obtain the slag area and number of edge detection results, specifically including: Extracting the contour of the slag in the furnace mouth area by using the Canny edge operator; Perform roundness screening on the extracted contour lines; Convert the filtered contours into regions and fill them; The filled area is screened, and an area with an area within a set interval is selected, and the area is marked in the furnace mouth image to obtain the identified slag pieces, and the area and number of the slag pieces are obtained.

4. The furnace mouth image recognition method according to claim 3, characterized in that: The extracted contour lines are screened for roundness, including: Remove the non-arc edges from the extracted contours; The remaining contour lines are screened for closed contours to obtain contour lines whose distance between the starting point and the ending point is less than the set distance threshold.

5. The furnace mouth image recognition method according to claim 1, characterized in that: The features of the derived graph are extracted and processed to obtain the area and number of slag pieces in the derived graph, including: Perform background removal on the derived image; Filter out areas where the pixel value is greater than the set pixel threshold; The furnace mouth border and area screening are performed on the screened area to obtain the identified slag pieces, and the area and number of the slag pieces are obtained.

6. The furnace mouth image recognition method according to claim 1, characterized in that: The splash recognition result is obtained according to the comparison result within the second set time period, specifically including: When the area of ​​the splash detection area obtained by moving downward after threshold processing is greater than the first set area threshold, and the area of ​​the splash detection area obtained by moving upward after threshold processing is less than the second set area threshold, and within a second set time period, when the number of the furnace mouth images that satisfies the conditions that the area of ​​the splash detection area obtained by moving downward after threshold processing is greater than the first set area threshold, and the area of ​​the splash detection area obtained by moving upward after threshold processing is less than the second set area threshold is greater than the set number threshold, it is determined that the furnace mouth image contains splashing; or, When the area of ​​the splash detection area obtained by moving downward after threshold processing is greater than half of the total area of ​​the furnace mouth area, and the area of ​​the splash detection area obtained by moving upward after threshold processing is less than the second set area threshold, it is determined that the furnace mouth image contains splashing phenomenon.

7. A furnace mouth image recognition device, characterized in that: include: Image acquisition module, used for real-time acquisition of furnace mouth images; An image processing module, used for preprocessing the furnace mouth image to obtain a furnace mouth area of ​​the furnace mouth image; A pre-splashing identification module is used to perform edge detection on the furnace mouth area by using a Canny edge operator to obtain the area and number of slag pieces of the edge detection result; convert the furnace mouth image by using a Gaussian derivative matrix to obtain a derivative graph of the Hessian rule, extract and process the features of the derivative graph, and obtain the area and number of slag pieces in the derivative graph; perform a union operation on the area and number of slag pieces of the edge detection result obtained and the area and number of slag pieces in the derivative graph to obtain the area and number of slag pieces in the furnace mouth area, and obtain the pre-splashing identification result according to the area and number of slag pieces in the first set time period; The splash recognition module is used to translate up and down along the height of the furnace mouth area to obtain two splash detection areas, perform threshold processing on the images of the two splash detection areas, obtain the corresponding areas and compare them with the corresponding area thresholds respectively, and obtain the splash recognition result according to the comparison results within the second set time period.

8. A furnace mouth image recognition device, characterized in that: The method comprises a processor and a memory, wherein the processor implements the furnace mouth image recognition method according to any one of claims 1 to 6 when executing the computer program stored in the memory.

9. A computer-readable storage medium, characterized in that: Used to store a computer program, wherein when the computer program is executed by a processor, the furnace mouth image recognition method according to any one of claims 1 to 6 is implemented.

Images