Heating furnace slab centering detection method, device, equipment and medium

Abstract

The invention discloses a heating furnace slab centering detection method, device and equipment and a medium, and relates to the technical field of heating furnace detection, and the method comprises the steps: carrying out the shooting through a plurality of cameras in an annular array layout, obtaining the image information of a heating furnace slab, carrying out the matching through a preset database according to the image information, and obtaining the slab number and thickness information, the method comprises the steps of obtaining image information, preprocessing the image information to obtain processed image information, processing the processed image information through a coordinate measurement algorithm based on a slab number and thickness information to obtain image feature parameters, and inputting the image feature parameters into a preset slab recognition model to obtain a detection result. Multiple cameras are used for obtaining heating furnace plate blank images, plate blank numbers and thicknesses are extracted, image feature parameters are obtained through preprocessing and a coordinate measurement algorithm, the image feature parameters are input into a plate blank recognition model to obtain a detection result, the accuracy and efficiency of plate blank centering detection are improved, and automatic detection is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of heating furnace detection, and in particular to a method, device, equipment and medium for detecting centering of a heating furnace slab. Background Art

[0002] Hot rolling lines are a critical link in modern steel production, with production efficiency and product quality directly impacting a company's economic returns and market competitiveness. Hot-rolled slabs must be precisely aligned before entering the reheating furnace to ensure they maintain a stable position and orientation during the subsequent rolling process. Slab centering accuracy not only impacts the quality of the rolled product but also the safe operation of production equipment.

[0003] In traditional production processes, the detection of slab lateral offset and rotation angle (i.e., "centering") primarily relies on manual experience or simple infrared detection technology. However, manual detection is subject to high subjectivity, low accuracy, and poor real-time performance. Infrared detection only provides limited position information and cannot meet the requirements of high-precision, fully automated production. Therefore, a more efficient and accurate method is needed to detect heating furnace centering. Summary of the Invention

[0004] The main purpose of this application is to provide a method, device, equipment and medium for detecting the centering of a heating furnace slab, aiming to solve the technical problem of how to improve the efficiency and accuracy of the centering detection of a heating furnace.

[0005] To achieve the above objectives, the present application proposes a method for detecting the centering of a heating furnace slab, the method comprising:

[0006] Multiple cameras are arranged in a circular array to capture images of the heating furnace slab.

[0007] According to the image information, matching is performed through a preset database to obtain slab number and thickness information;

[0008] Preprocessing the image information to obtain processed image information;

[0009] Processing the processed image information using a coordinate measurement algorithm based on the slab number and the thickness information to obtain image feature parameters;

[0010] The image feature parameters are input into a preset slab recognition model to obtain a detection result.

[0011] In one embodiment, the step of processing the processed image information using a coordinate measurement algorithm based on the slab number and the thickness information to obtain image feature parameters includes:

[0012] Get slab database information;

[0013] Based on the slab number, the slab database information is consulted to obtain calibration parameters corresponding to the slab number, wherein the parameters include distortion parameters, an intrinsic parameter matrix, and an extrinsic parameter matrix of the camera;

[0014] Correcting the processed image information according to the thickness information based on the calibration parameters to obtain corrected image information;

[0015] Processing the corrected image information through a coordinate measurement algorithm to obtain actual coordinate information;

[0016] Based on the actual coordinate information, image feature parameters are obtained, and the image feature parameters include the actual coordinates of the slab corner points and the relative position information with respect to the center line of the roller.

[0017] In one embodiment, the step of processing the corrected image information through a coordinate measurement algorithm to obtain actual coordinate information includes:

[0018] Obtain pixel coordinates of the corrected image information;

[0019] The pixel coordinates are converted based on the camera intrinsic parameter matrix to obtain the image coordinates. The specific formula is:

[0020]

[0021] Among them, u0, v0 represent the coordinates of the center point of the image coordinate system in the pixel coordinate system, u, v represent the coordinates of the image coordinate system corresponding to the pixel coordinate system, and A represents the intrinsic parameter matrix;

[0022] The image coordinates are converted into camera coordinates according to the perspective projection algorithm. The specific formula is:

[0023]

[0024] Among them, x c ,y c ,z c represents the coordinates in the camera coordinate system, and f represents the focal length of the camera;

[0025] The camera coordinates are transformed based on the camera extrinsic matrix to obtain the actual coordinate information. The specific formula is:

[0026]

[0027] Among them, R and T represent the external parameter matrix.

[0028] In one embodiment, before the step of applying the image feature parameters to a preset slab recognition model to obtain a detection result, the step includes:

[0029] Obtain image information sample data and build an initial slab recognition model;

[0030] The initial slab recognition model is trained based on the image information sample data to obtain a preset slab recognition model.

[0031] In one embodiment, the step of training the initial slab recognition model based on the image information sample data to obtain a preset slab recognition model includes:

[0032] Initializing weight and bias parameters of the slab recognition model;

[0033] Inputting the image information sample data into the initial slab recognition model for calculation to obtain a prediction result;

[0034] Comparing the predicted result with the true value according to the loss function to obtain an error value;

[0035] Obtaining the gradients of the weight and bias parameters by back propagation algorithm calculation;

[0036] The weight and bias parameters are iteratively updated through an optimization algorithm according to the gradient until a maximum number of iterations is reached or the error value converges to a preset threshold, thereby obtaining a preset slab recognition model.

[0037] In one embodiment, the step of inputting the image feature parameters into a preset slab recognition model to obtain a detection result includes:

[0038] The image feature parameters are input into the preset slab recognition model to obtain the slab centerline. The specific formula is:

[0039] y=k1x+b1

[0040] Where k1 represents the slope of the slab centerline, b1 represents the slab intercept, which is obtained by fitting the actual coordinates of the preset corner points of the slab using the least squares method;

[0041] Calculating the relative position information of the slab centerline and the roller centerline to obtain the lateral offset and offset angle;

[0042] A detection result is generated according to the lateral offset and the offset angle.

[0043] In one embodiment, after the step of inputting the image feature parameters into a preset slab recognition model to obtain a detection result, the following steps are included:

[0044] Evaluating the detection result to obtain an evaluation result;

[0045] When the evaluation result does not meet the preset offset requirement, an alarm report is generated, which includes the alarm time, offset amount, offset angle and slab number.

[0046] In addition, to achieve the above-mentioned purpose, the present application also proposes a heating furnace slab centering detection device, the heating furnace slab centering detection device comprising:

[0047] An acquisition module is used to obtain image information of the heating furnace slab by shooting with multiple cameras in a circular array layout;

[0048] A query module, configured to obtain slab number and thickness information by matching the image information with a preset database;

[0049] A processing module, configured to pre-process the image information to obtain processed image information;

[0050] a conversion module, configured to process the processed image information through a coordinate measurement algorithm based on the slab number and the thickness information to obtain image feature parameters;

[0051] The result module is used to input the image feature parameters into a preset slab recognition model to obtain a detection result.

[0052] In addition, to achieve the above-mentioned purpose, the present application also proposes a medium, which is a computer-readable medium and stores a computer program. When the computer program is executed by a processor, the steps of the heating furnace slab centering detection method as described above are implemented.

[0053] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the steps of the heating furnace slab centering detection method as described above.

[0054] This application uses multiple cameras in a circular array layout to capture images of heating furnace slabs. The image information is then matched against a preset database to obtain the slab number and thickness information. The image information is preprocessed to obtain processed image information. The processed image information is then processed using a coordinate measurement algorithm based on the slab number and thickness information to obtain image feature parameters. These image feature parameters are then input into a preset slab recognition model to obtain detection results. By using multiple cameras to capture images of heating furnace slabs, extracting the slab number and thickness, and then preprocessing and using a coordinate measurement algorithm to obtain image feature parameters, the image feature parameters are input into the slab recognition model to obtain detection results. This improves the accuracy and efficiency of slab alignment detection and enables automated detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0056] Figure 1 This is a flow chart of a first embodiment of a method for detecting centering of a heating furnace slab according to the present application;

[0057] Figure 2 This is a schematic diagram of the slab offset of the first embodiment of the heating furnace slab centering detection method of the present application;

[0058] Figure 3 This is a schematic diagram of the slab deviation angle of the first embodiment of the heating furnace slab centering detection method of the present application;

[0059] Figure 4 This is a flow chart of a second embodiment of the method for detecting centering of a heating furnace slab according to the present application;

[0060] Figure 5 This is a flow chart of a third embodiment of a method for detecting centering of a heating furnace slab according to the present application;

[0061] Figure 6 This is a schematic diagram of the module structure of the heating furnace slab centering detection device according to an embodiment of the present application;

[0062] Figure 7 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the heating furnace slab alignment detection method in the embodiment of the present application.

[0063] The purpose, features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0064] It should be understood that the specific embodiments described herein are merely used to explain the technical solutions of the present application and are not intended to limit the present application.

[0065] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0066] In traditional production processes, slab alignment detection relies primarily on manual experience or simple infrared detection technology. However, manual detection is subject to high subjectivity, low accuracy, and poor real-time performance. Infrared detection only provides limited position information and cannot meet the requirements of high-precision, fully automated production.

[0067] The main solution of the embodiment of the present application is: to obtain image information of the heating furnace slab by taking pictures with multiple cameras in a circular array layout, and then to obtain the slab number and thickness information by matching the image information through a preset database, and then to pre-process the image information to obtain processed image information, and to process the processed image information based on the slab number and the thickness information through a coordinate measurement algorithm to obtain image feature parameters, and to input the image feature parameters into a preset slab recognition model to obtain detection results.

[0068] Based on this, the embodiment of the present application provides a method for detecting the centering of a heating furnace slab, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the method for detecting the centering of a heating furnace slab according to the present application.

[0069] In this embodiment, the heating furnace slab centering detection method includes steps S10 to S50:

[0070] Step S10: Using multiple cameras in a circular array layout to take photos, image information of the heating furnace slab is obtained.

[0071] It should be noted that four 4-megapixel industrial cameras are deployed in the furnace inlet area of the heating furnace. They are arranged in a circular array, with multiple cameras evenly arranged in a circle around the center of the heating furnace. This layout can capture image information of the slab in all directions and from multiple angles, ensuring accurate detection of the slab from different directions. Each camera shoots continuously at a frame rate of 30fps and transmits image data to the image processing server in real time via Gigabit Ethernet. To adapt to high-temperature industrial environments, the camera is equipped with a high-temperature resistant protective cover (can withstand an ambient temperature of 80°C) and an automatic cleaning device to prevent dust and heat radiation from affecting the imaging quality. During the image acquisition process, a hardware trigger synchronization mechanism is used: when the PLC detects the slab transport arrival signal (changes from 0 to 1), all cameras are started simultaneously through the IO trigger signal to ensure the time synchronization of multi-view images. At the same time, the timestamp and slab number information of each frame of the image will be recorded to achieve data traceability. To optimize image quality, an adaptive exposure control algorithm is integrated to dynamically adjust camera exposure parameters according to on-site lighting conditions, and a high-brightness LED fill light (6000K color temperature, 120W power) is used to eliminate shadow and reflection interference.

[0072] At the same time, image compression technology (such as H.264 encoding) is used in the data transmission process to reduce network load, and CRC checksums are used to ensure the integrity of data transmission. After receiving the image, the server will first perform cache management and establish an image database based on slab number and time sequence, providing complete raw data support for subsequent processing.

[0073] Step S20: Matching is performed through a preset database according to the image information to obtain the slab number and thickness information.

[0074] It should be noted that after the image acquisition is completed, a real-time data connection is established with the first-level automation (L1) of the production line through the OPC UA protocol. When the slab arrival signal is detected, a query request containing a timestamp is immediately sent to L1 to obtain the complete information package of the current slab. The information package is encapsulated in JSON format and contains key data such as slab number, specification parameters (thickness, width, length), material brand, etc. The slab thickness information is accurate to 0.1mm, and the latest process parameters provided by process automation (L2) through the MES interface ensure accuracy. The preset database is updated and constructed by performing detailed measurements and records of each slab during the production process, including key parameters such as slab number, thickness, length, width, etc.

[0075] In order to cope with network delays or data loss, a three-level data fault tolerance mechanism has been designed: first try to read from the L1 cache, if it fails, turn to the L2 database query, and finally call the historical data backup. All extracted data will undergo triple verification: format verification (such as regular expression matching numbering rules), range verification (such as thickness value in the range of 50-300mm) and logical verification (such as aspect ratio meets process specifications). The data that passes the verification will be marked with a timestamp and acquisition device ID, stored in the Redis real-time database, and written to SQL Server for persistent storage. At the same time, a data association module has been specially developed to automatically bind the image file (stored as a binary file named in time series) with the process parameters through the slab number to establish a complete inspection data chain. The process uses multi-threaded parallel processing to ensure that all data extraction and association operations are completed within 500ms, providing complete context information for subsequent image processing.

[0076] Step S30: pre-process the image information to obtain processed image information.

[0077] It should be noted that image preprocessing involves optimizing the original image to enhance image features and improve the accuracy and efficiency of subsequent detection and recognition. First, the original image needs to be corrected to eliminate the effects of camera distortion and uneven ambient lighting. This is typically done using a calibration plate to obtain the camera's intrinsic and extrinsic parameters, which are then used to perform geometric correction on the image. Furthermore, the heating furnace environment is complex, and various noise levels are present in the image. Filtering algorithms (such as Gaussian filtering and median filtering) can effectively remove random noise from the image and improve image quality. To enhance the clarity of image features, techniques such as histogram equalization or adaptive threshold segmentation can be used to enhance image contrast, making the slab number and thickness information more clearly visible. Edge detection algorithms (such as the Canny operator and the Sobel operator) can then accurately extract edge information from the image, which is critical for identifying slab outlines and details. Finally, the image is segmented into distinct regions, facilitating more detailed analysis and processing of specific areas (such as the slab number and thickness identification areas).

[0078] Step S40 : processing the processed image information through a coordinate measurement algorithm based on the slab number and thickness information to obtain image feature parameters.

[0079] It should be noted that the slab number is used to query the slab database, which stores detailed information on all slabs, including but not limited to production batches, material properties, and processing parameters. In particular, parameters related to camera calibration, such as distortion parameters, intrinsic parameter matrices, and extrinsic parameter matrices, are also stored in this database. Based on the slab number, the slab database is consulted to retrieve the calibration parameters corresponding to the slab. These parameters are crucial for correcting the images of heating furnace slabs captured by multiple cameras. Specifically, the camera's distortion parameters are used to correct image distortion caused by the lens; the intrinsic parameter matrix contains internal camera parameters (such as focal length and principal point coordinates) to facilitate the conversion of pixel coordinates to the physical coordinate system; and the extrinsic parameter matrix describes the camera's position and attitude relative to the world coordinate system and is crucial for accurately calculating the object's position in real space. Using these calibration parameters, combined with slab thickness information, the preprocessed image can be further corrected to eliminate errors caused by factors such as shooting angle and equipment characteristics, resulting in more accurate corrected image information. This process ensures the accuracy of the subsequent coordinate measurement algorithm.

[0080] Based on the corrected image, by applying the coordinate measurement algorithm, the image feature parameters such as the actual coordinates of the slab corner points and their relative position information with the center line of the roller can be accurately calculated.

[0081] Step S50: input the image feature parameters into a preset slab recognition model to obtain a detection result.

[0082] It should be noted that after image feature parameters are extracted, they are input into a pre-configured slab recognition model for analysis and processing. This model utilizes a modified YOLOv8 architecture, specifically tuned for slab detection. Trained on a large-scale slab image dataset using transfer learning techniques, it achieves a recognition accuracy of 98.7%. The model input layer receives a standardized feature tensor, including 32-dimensional feature parameters such as the coordinates of slab contour feature points, grayscale distribution histograms, and texture feature vectors. In terms of network architecture, the model uses depthwise separable convolution to reduce computational complexity and incorporates a CBAM (Central Attention Mechanism) to enhance key feature extraction. During the forward propagation process, the model outputs slab position, pose, and bounding box information through five detection heads, while also predicting the OCR recognition result for the slab number. In the post-processing stage, a non-maximum suppression (NMS) algorithm is used to eliminate duplicate detections, and a trajectory smoothing algorithm based on historical frame information is applied to improve detection stability. Structured data such as the slab center coordinates, rotation angle, and centering deviation are output in real time, and quality assessment metrics including confidence scores are generated.

[0083] Furthermore, the image feature parameters are input into the preset slab recognition model to obtain the slab centerline, calculate the relative position information of the slab centerline and the roller centerline, obtain the lateral offset and offset angle, and generate the detection result based on the lateral offset and offset angle. Specifically, after obtaining the equation of the slab centerline, it is necessary to calculate the relative position information between it and the roller centerline, which includes the lateral offset and offset angle. The lateral offset refers to the difference in distance between the slab centerline and the roller centerline in the horizontal direction; the offset angle reflects the size of the angle between the two, and is used to describe the degree of deviation of the slab from the ideal travel path. In order to accurately calculate these values, it is first necessary to clarify the specific mathematical expressions of the two centerlines,

[0084] The specific formula for the slab centerline is:

[0085] y=k1x+b1

[0086] Among them, k1 represents the slope of the slab centerline, b1 represents the slab intercept, which is obtained by fitting the actual coordinates of the preset corner points of the slab using the least squares method. The specific formula of the roller centerline is:

[0087] y=k2x+b2

[0088] Among them, k2 represents the slope of the slab centerline, and b2 represents the slab intercept, which are determined according to the roller geometry parameters in the actual production scenario.

[0089] By comparing the equations of the two lines, we can use geometric methods or knowledge of analytic geometry to solve the relative position relationship between them. For example, using vector analysis methods can effectively calculate the shortest distance between the two center lines (i.e., the lateral offset) and the angle between them (offset angle). Figure 2 The diagram of slab offset is shown in the figure. The solid line in the figure represents the center line of the slab roller, the dotted line in the figure represents the center line of the slab in the heating furnace, and the blue arrow represents the lateral offset of the slab center line relative to the center line of the roller. +10mm and -10mm are marked in the figure, indicating the maximum allowable offset in the positive and negative directions. Based on the calculated lateral offset, if the roller center line is closer to the heating furnace than the slab center line, then this distance is defined as a positive value; conversely, if the roller center line is farther away from the heating furnace than the slab center line, then this distance is defined as a negative value. Figure 3 The diagram below shows a schematic diagram of slab offset angles. +1° indicates a 1-degree rotation of the slab centerline to the right, and -1° indicates a 1-degree rotation to the left, with a maximum offset angle of 1 degree. Calculations can generate detailed inspection results. These results not only reveal whether the slabs maintain proper alignment during transportation but also provide the necessary data for adjusting related mechanical equipment on the production line. For example, detecting a large offset or angle indicates a calibration issue with the conveyor system or other potential operational anomalies. In this case, timely adjustments can be made based on the inspection results to ensure smooth subsequent production.

[0090] Furthermore, after step S50, it also includes: evaluating the test results to obtain an evaluation result; when the evaluation result does not meet the preset offset requirement, generating an alarm report, the alarm report includes the alarm time, offset, offset angle and slab number. Specifically, the test results are compared with the preset process standards. In this embodiment, when the lateral offset exceeds ±10mm or the rotation angle exceeds ±1°, it is determined to be an alignment abnormality. The evaluation process adopts a sliding window algorithm to comprehensively analyze the results of the last three consecutive tests to avoid misjudgment. For abnormal situations, a structured alarm report is generated, including core data such as the alarm timestamp (format: YYYY-MM-DD HH:MM:SS.fff), the measured offset (accuracy 0.01mm), the offset angle (accuracy 0.01°), the slab number, and the associated camera view screenshot and original measurement data. When an abnormal situation is detected, the re-inspection process is automatically triggered or the operator is notified to intervene to ensure the continuity and safety of the production process.

[0091] These inspection results are transmitted to the PLC control system in real time via the OPC UA protocol, triggering the automatic correction device to operate. The data is then persistently stored in an SQL database for quality traceability. The model also features online learning capabilities, continuously collecting actual production line data for regular incremental training, gradually improving its adaptability to new slab specifications.

[0092] This embodiment provides a method for detecting the centering of heating furnace slabs. Multiple cameras are arranged in a circular array to capture images of the heating furnace slabs. The image information is then matched against a preset database to obtain the slab number and thickness information. The image information is preprocessed to obtain processed image information. The processed image information is then processed using a coordinate measurement algorithm based on the slab number and thickness information to obtain image feature parameters. These image feature parameters are then input into a preset slab recognition model to obtain a detection result. By using multiple cameras to capture images of the heating furnace slabs, extracting the slab number and thickness, and then preprocessing and using a coordinate measurement algorithm to obtain image feature parameters, the image feature parameters are input into the slab recognition model to obtain a detection result. This method improves the accuracy and efficiency of slab centering detection and enables automated detection.

[0093] Based on the first embodiment of the present application, in the third embodiment of the present application, the same or similar contents as those in the first embodiment can be referred to the above introduction and will not be described in detail later. Figure 4 The heating furnace slab centering detection method step S40 further includes steps S201 to S205:

[0094] Step S201, obtaining slab database information.

[0095] It's important to note that sensors, surveillance cameras, and other equipment on the production line collect information about the slabs. This includes, but is not limited to, basic information such as slab number, size, and thickness. This collected data is processed and entered into a database. Therefore, this database not only stores basic information about each slab, such as key parameters like slab number, size, and thickness, but also records various dynamic data during the production process, including, but not limited to, positional offset and rotation angle before entering the heating furnace.

[0096] Step S202 : Based on the slab number, the slab database information is searched to obtain calibration parameters corresponding to the slab number.

[0097] It's important to note that once a slab is assigned a unique slab number, it becomes its identifier throughout the entire production process. The slab database not only records basic information about each slab (such as size and thickness), but also stores its associated calibration parameters, acquired through a precise camera calibration process.

[0098] Specifically, when precise position and angular offset measurements of a specific slab are required, the corresponding calibration parameters are first retrieved from the database based on the slab number. These parameters, including camera distortion parameters and intrinsic and extrinsic parameters, are crucial for converting image coordinates into actual physical coordinates. For example, when employing the Zhang Zhengyou calibration method, detailed camera calibration data is obtained by analyzing a series of images captured using a calibration plate. This data is then used to correct the slab image captured by the camera, eliminating any possible geometric distortion and improving measurement accuracy.

[0099] Step S203 , correcting the processed image information according to the thickness information based on the calibration parameters to obtain corrected image information.

[0100] It should be noted that the standard billet is photographed with a high-resolution camera, and the original image is preliminarily corrected using the previously determined camera calibration parameters (including distortion parameters and internal and external parameter parameters) to eliminate the geometric distortion caused by the camera hardware characteristics. This process ensures that the dimensional information in the image can be accurately mapped to the actual physical size. However, billets of different thicknesses will have different effects on the imaging effect. This is because light changes when passing through or reflecting from materials of different thicknesses, resulting in image distortion. Therefore, after completing the preliminary correction, the image needs to be further fine-tuned according to the specific thickness information of the billet. This step usually involves adjusting specific parameters in the image processing algorithm, such as introducing a thickness-related compensation factor to correct the projection deformation or perspective error caused by the difference in billet thickness.

[0101] Step S204: Process the corrected image information through a coordinate measurement algorithm to obtain actual coordinate information.

[0102] It should be noted that the above-mentioned corrected image information is input into the coordinate measurement algorithm based on the imaging principle to obtain the precise coordinate information of the slab in the real space. The core of the coordinate measurement algorithm is to convert the two-dimensional image coordinates into three-dimensional physical coordinates. This step first involves using the internal and external parameters of the camera obtained by Zhang Zhengyou's calibration method to map the points in the image to the actual physical space through the coordinate transformation formula. Specifically, for each key point identified from the image (such as the edge of the slab or a feature point), the algorithm calculates its three-dimensional coordinates in the camera coordinate system based on its pixel coordinates and the internal parameters of the camera (such as focal length, principal point, etc.). Subsequently, these coordinates are further converted to the real world coordinate system in combination with the position and posture of the camera relative to the world coordinate system (i.e., external parameters).

[0103] Furthermore, the pixel coordinates of the corrected image information are obtained, and the pixel coordinates are converted based on the camera intrinsic parameter matrix to obtain the image coordinates. The specific formula is:

[0104]

[0105] Among them, u0, v0 represent the coordinates of the center point of the image coordinate system in the pixel coordinate system, u, v represent the coordinates of the image coordinate system corresponding to the pixel coordinate system, and A represents the intrinsic parameter matrix;

[0106] According to the perspective projection algorithm, the image coordinates are converted into camera coordinates. The specific formula is:

[0107]

[0108] Among them, x c ,y c ,z c represents the coordinates in the camera coordinate system, and f represents the focal length of the camera;

[0109] The camera coordinates are transformed based on the camera extrinsic matrix to obtain the actual coordinate information. The specific formula is:

[0110]

[0111] Where R and T represent extrinsic parameter matrices. In this embodiment, camera calibration is performed using the Zhang Zhengyou calibration method. By fixing the camera or a marker and taking multiple images of the marker from different angles, the pixel coordinates of each corner point are accurately extracted, and their physical coordinates in the world coordinate system are calculated. The maximum likelihood estimation method is then used to optimize and solve the camera's intrinsic and extrinsic parameters and distortion coefficients. The specific calibration process uses a checkerboard as a calibration plate with a known size. By taking multiple photos from different perspectives to obtain sufficient information, feature points in each photo are detected to obtain pixel coordinate values. The physical coordinate values are calculated based on the known world coordinate system information. The camera extrinsic parameter matrix corresponding to each image is gradually solved, completing the analysis of the entire calibration process. Finally, radial and tangential distortion parameters are calculated to correct errors caused by lens distortion.

[0112] Step S205: obtaining image feature parameters based on the actual coordinate information.

[0113] It's important to note that after completing the coordinate transformation, we obtain the precise real-world position of the slab, including its lateral offset and rotation angle. This real-world coordinate information provides a solid foundation for subsequent image feature extraction. Using this corrected real-world coordinate data, we can precisely locate the slab's edges or specific feature points. This step typically involves applying edge detection algorithms such as Canny edge detection or the Sobel operator. By identifying key points of the slab's contour, we can further calculate its size, shape, and positional deviation relative to the furnace centerline.

[0114] This embodiment obtains calibration parameters through the slab database, corrects the image in combination with thickness information, and uses a coordinate measurement algorithm to calculate the actual coordinates. Finally, the slab corner point position and its relative position to the roller centerline are extracted, thereby improving detection accuracy and production efficiency and realizing intelligent monitoring and early warning.

[0115] Based on the first embodiment of the present application, in the third embodiment of the present application, the same or similar contents as those in the first embodiment can be referred to the above introduction and will not be described in detail later. Figure 5 The heating furnace slab centering detection method further includes steps S301 to S302 before step S50:

[0116] Step S301: Acquire image information sample data and construct an initial slab recognition model.

[0117] It should be noted that a large number of slab images are collected from the production line as sample data. These images should cover slabs of different thicknesses, sizes, and under various lighting conditions to ensure that the model has good generalization capabilities. The collected image samples need to be preprocessed, including cropping, scaling, grayscale conversion and other operations, to facilitate subsequent feature extraction and model training. During the preprocessing process, it is also necessary to apply the previously calibrated camera parameters to eliminate geometric distortion in the image and ensure that each pixel accurately reflects the actual physical position. Next, an initial slab recognition model is constructed based on a deep learning target detection algorithm (such as YOLOv8). This process involves training with a large number of images marked with slab positions and key feature points. Through multiple rounds of iterative optimization, the model can automatically learn the feature representation of the slab and have high-precision positioning capabilities.

[0118] In addition, slab thickness information needs to be introduced as an additional feature input during the model training stage, which helps improve the model's adaptability to slabs of different thicknesses.

[0119] Step S302 : training the initial slab recognition model based on the image information sample data to obtain a preset slab recognition model.

[0120] It should be noted that training the initial slab recognition model involves first initializing the model's weights and bias parameters. Sample image data is then fed into the initial slab recognition model for computation and prediction. The predictions are then compared with the true values using a loss function to obtain an error. The gradients of the weights and bias parameters are then calculated using a backpropagation algorithm. Finally, the weights and bias parameters are iteratively updated using an optimization algorithm based on the gradients until the maximum number of iterations is reached or the error converges to a preset threshold, resulting in the preset slab recognition model. Specifically, the model's weights and bias parameters are initialized using a random or pretrained method to provide a starting point for subsequent training. The preprocessed image data is then fed into the initial slab recognition model, and a forward propagation algorithm is used to obtain predictions. After obtaining the predictions, the predictions are compared with the true labels using an appropriate loss function (such as mean squared error or cross entropy loss) to quantify the accuracy of the model's predictions and calculate the error. This process helps us understand the gap between the current model output and the desired output. Next, the backpropagation algorithm is applied to calculate the gradients of the weights and bias parameters for each layer based on the error values. This is achieved by calculating the influence of each parameter on the total error layer by layer using the chain rule. Based on these gradients, an optimization algorithm (such as stochastic gradient descent SGD, Adam, etc.) is used to iteratively update the weights and bias parameters to gradually reduce the overall error of the model.

[0121] Throughout the training process, parameters are continuously updated iteratively until the maximum number of iterations is reached or the model error converges to a set threshold. The resulting slab recognition model can efficiently and accurately identify the position and contour of the slab, while also precisely calculating its offset and angular deviation relative to the centerline of the heating furnace.

[0122] After obtaining the preset slab recognition model, the method further includes: evaluating the preset slab recognition model to obtain an evaluation result; if the evaluation result does not meet the detection requirements, retraining the preset slab recognition model until the detection requirements are met.

[0123] This embodiment acquires image sample data, constructs and trains an initial slab recognition model, iteratively updates parameters through an optimization algorithm, and ultimately obtains a preset model, thereby improving the accuracy and efficiency of slab detection and laying the foundation for intelligent monitoring and quality control of the production line.

[0124] This application also provides a heating furnace slab centering detection device, please refer to Figure 6 , the device comprises:

[0125] The acquisition module 10 is used to obtain image information of the heating furnace slab by shooting with multiple cameras in a circular array layout.

[0126] The query module 20 is used to obtain the slab number and thickness information by matching the image information through a preset database.

[0127] The processing module 30 is used to pre-process the image information to obtain processed image information.

[0128] The conversion module 40 is used to process the processed image information through a coordinate measurement algorithm based on the slab number and thickness information to obtain image feature parameters.

[0129] The result module 50 is used to input the image feature parameters into a preset slab recognition model to obtain a detection result.

[0130] The heating furnace slab centering detection device provided in this application utilizes the heating furnace slab centering detection method described in the aforementioned embodiment, addressing the technical problem of improving the efficiency and accuracy of heating furnace centering detection. Compared to the prior art, the heating furnace slab centering detection device provided in this application achieves the same beneficial effects as the heating furnace slab centering detection method described in the aforementioned embodiment. Other technical features of the heating furnace slab centering detection device are the same as those disclosed in the aforementioned embodiment and are not further elaborated here.

[0131] In one embodiment, the conversion module 40 also obtains slab database information; searches the slab database information based on the slab number to obtain calibration parameters corresponding to the slab number; corrects the processed image information according to the thickness information based on the calibration parameters to obtain corrected image information; processes the corrected image information through a coordinate measurement algorithm to obtain actual coordinate information; and obtains image feature parameters based on the actual coordinate information.

[0132] In one embodiment, the conversion module 40 is further used to obtain pixel coordinates of the corrected image information, convert the pixel coordinates based on the camera intrinsic parameter matrix to obtain image coordinates, convert the image coordinates into camera coordinates according to the perspective projection algorithm, and convert the camera coordinates based on the camera extrinsic parameter matrix to obtain actual coordinate information.

[0133] In one embodiment, the result module 50 is further configured to obtain image information sample data and construct an initial slab recognition model; and train the initial slab recognition model based on the image information sample data to obtain a preset slab recognition model.

[0134] In one embodiment, the result module 50 is also used to initialize the weights and bias parameters of the slab recognition model; input the image information sample data into the initial slab recognition model for calculation to obtain a prediction result; compare the prediction result with the true value according to the loss function to obtain an error value; calculate through the back propagation algorithm to obtain the gradient of the weight and bias parameters; iteratively update the weight and bias parameters through the optimization algorithm according to the gradient until the maximum number of iterations is reached or the error value converges to a preset threshold, thereby obtaining a preset slab recognition model.

[0135] In one embodiment, the result module 50 is also used to input the image feature parameters into a preset slab recognition model to obtain the slab centerline, calculate the relative position information of the slab centerline and the roller centerline, obtain the lateral offset and offset angle, and generate a detection result based on the lateral offset and offset angle.

[0136] In one embodiment, the result module 50 is further configured to evaluate the detection result to obtain an evaluation result; and generate an alarm report when the evaluation result does not meet a preset deviation requirement.

[0137] The present application provides a heating furnace slab centering detection device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the heating furnace slab centering detection method in the above-mentioned embodiment one.

[0138] Reference below Figure 7 , which shows a schematic structural diagram of a heating furnace slab centering detection device suitable for implementing an embodiment of the present application. The heating furnace slab centering detection device in the embodiment of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (e.g., vehicle-mounted navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 7 The heating furnace slab centering detection device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0139] like Figure 7The heating furnace slab centering detection device may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes based on programs stored in a ROM (Read Only Memory) 1002 or programs loaded from a storage device 1003 into a RAM (Random Access Memory) 1004. RAM 1004 also stores various programs and data required for the operation of the heating furnace slab centering detection device. Processing device 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, hard disk, etc.; and communication devices 1009. The communication devices 1009 can allow the heating furnace slab centering detection device to communicate wirelessly or wired with other devices to exchange data. While the figure shows a heating furnace slab centering detection device with various systems, it should be understood that implementation or presence of all the illustrated systems is not required. More or fewer systems may alternatively be implemented or present.

[0140] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes program code for executing the method described in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are performed.

[0141] The heating furnace slab centering detection device provided in this application utilizes the heating furnace slab centering detection method described in the aforementioned embodiment, addressing the technical problem of improving the efficiency and accuracy of heating furnace centering detection. Compared to the prior art, the heating furnace slab centering detection device provided in this application achieves the same beneficial effects as the heating furnace slab centering detection method described in the aforementioned embodiment. Other technical features of the heating furnace slab centering detection device are the same as those disclosed in the aforementioned embodiment and are not further elaborated upon here.

[0142] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0143] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0144] The present application provides a computer-readable medium having computer-readable program instructions (ie, computer program) stored thereon for performing calculations to obtain computer-readable program instructions for executing the heating furnace slab centering detection method in the above-mentioned embodiment.

[0145] The computer-readable medium provided in this application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination thereof. More specific examples of computer-readable media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. Calculations are performed in this embodiment to obtain a machine-readable medium that may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0146] The computer-readable medium may be included in the heating furnace slab centering detection device; or may exist independently without being assembled into the heating furnace slab centering detection device.

[0147] The computer-readable medium carries one or more programs. When the one or more programs are executed by the heating furnace slab centering detection device, the heating furnace slab centering detection device can write computer program codes for performing the operations of the present application in one or more programming languages or a combination thereof. The programming languages include object-oriented programming languages such as Java, Smalltalk, and C++, and also conventional procedural programming languages such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, through the Internet using an Internet service provider).

[0148] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0149] The modules described in the embodiments of the present application may be implemented in software or hardware, wherein the name of a module does not necessarily limit the unit itself.

[0150] The computer-readable medium provided in this application stores computer-readable program instructions (i.e., a computer program) for executing the aforementioned method for detecting heating furnace slab centering. This computer-readable medium addresses the technical problem of improving the efficiency and accuracy of heating furnace centering detection. Compared to the prior art, the beneficial effects of the computer-readable medium provided in this application are similar to those of the method for detecting heating furnace slab centering provided in the aforementioned embodiment, and are not further elaborated here.

[0151] The present application also provides a computer program product, comprising a computer program, which implements the steps of the above-mentioned heating furnace slab centering detection method when executed by a processor.

[0152] The computer program product provided in this application can solve the technical problem of improving the efficiency and accuracy of heating furnace alignment detection. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the heating furnace slab alignment detection method provided in the above-mentioned embodiment, and will not be elaborated here.

[0153] The above description is only part of the embodiments of the present application and does not limit the patent scope of the present application. All equivalent structural transformations made by using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A method for detecting the centering of a heating furnace slab, characterized in that: The method comprises: Multiple cameras are arranged in a circular array to capture images of the heating furnace slab. According to the image information, matching is performed through a preset database to obtain slab number and thickness information; Preprocessing the image information to obtain processed image information; Processing the processed image information using a coordinate measurement algorithm based on the slab number and the thickness information to obtain image feature parameters; The image feature parameters are input into a preset slab recognition model to obtain a detection result.

2. The method according to claim 1, wherein The step of processing the processed image information based on the slab number and the thickness information by a coordinate measurement algorithm to obtain image feature parameters includes: Get slab database information; Based on the slab number, the slab database information is consulted to obtain calibration parameters corresponding to the slab number, wherein the parameters include distortion parameters, an intrinsic parameter matrix, and an extrinsic parameter matrix of the camera; Correcting the processed image information according to the thickness information based on the calibration parameters to obtain corrected image information; Processing the corrected image information through a coordinate measurement algorithm to obtain actual coordinate information; Based on the actual coordinate information, image feature parameters are obtained, and the image feature parameters include the actual coordinates of the slab corner points and the relative position information with respect to the center line of the roller.

3. The method according to claim 2, wherein The step of processing the corrected image information through a coordinate measurement algorithm to obtain actual coordinate information includes: Obtain pixel coordinates of the corrected image information; The pixel coordinates are converted based on the camera intrinsic parameter matrix to obtain the image coordinates. The specific formula is: Among them, u0, v0 represent the coordinates of the center point of the image coordinate system in the pixel coordinate system, u, v represent the coordinates of the image coordinate system corresponding to the pixel coordinate system, and A represents the intrinsic parameter matrix; The image coordinates are converted into camera coordinates according to the perspective projection algorithm. The specific formula is: Among them, x c ,y c ,z c represents the coordinates in the camera coordinate system, and f represents the focal length of the camera; The camera coordinates are transformed based on the camera extrinsic matrix to obtain the actual coordinate information. The specific formula is: Among them, R and T represent the external parameter matrix.

4. The method according to claim 1, wherein Before the step of applying the image feature parameters to a preset slab recognition model to obtain a detection result, the method includes: Obtain image information sample data and build an initial slab recognition model; The initial slab recognition model is trained based on the image information sample data to obtain a preset slab recognition model.

5. The method according to claim 4, wherein The step of training the initial slab recognition model based on the image information sample data to obtain a preset slab recognition model includes: Initializing weight and bias parameters of the slab recognition model; Inputting the image information sample data into the initial slab recognition model for calculation to obtain a prediction result; Comparing the predicted result with the true value according to the loss function to obtain an error value; Obtaining the gradients of the weight and bias parameters by back propagation algorithm calculation; The weight and bias parameters are iteratively updated through an optimization algorithm according to the gradient until a maximum number of iterations is reached or the error value converges to a preset threshold, thereby obtaining a preset slab recognition model.

6. The method according to claim 1, wherein The step of inputting the image feature parameters into a preset slab recognition model to obtain a detection result includes: The image feature parameters are input into the preset slab recognition model to obtain the slab centerline. The specific formula is: y=k1x+b1 Where k1 represents the slope of the slab centerline, b1 represents the slab intercept, which is obtained by fitting the actual coordinates of the preset corner points of the slab using the least squares method; Calculating the relative position information of the slab centerline and the roller centerline to obtain the lateral offset and offset angle; A detection result is generated according to the lateral offset and the offset angle.

7. The method according to claim 1, wherein After the step of inputting the image feature parameters into a preset slab recognition model to obtain a detection result, the method further comprises: Evaluating the detection result to obtain an evaluation result; When the evaluation result does not meet the preset offset requirement, an alarm report is generated, which includes the alarm time, offset amount, offset angle and slab number.

8. A heating furnace slab centering detection device, characterized in that: The device comprises: An acquisition module is used to obtain image information of the heating furnace slab by shooting with multiple cameras in a circular array layout; A query module, configured to obtain slab number and thickness information by matching the image information with a preset database; A processing module, configured to pre-process the image information to obtain processed image information; a conversion module, configured to process the processed image information through a coordinate measurement algorithm based on the slab number and the thickness information to obtain image feature parameters; The result module is used to input the image feature parameters into a preset slab recognition model to obtain a detection result.

9. A heating furnace slab centering detection device, characterized in that: The device includes: a memory, a processor, and a heating furnace slab centering detection program stored in the memory and running on the processor, wherein the heating furnace slab centering detection program is configured to implement the steps of the heating furnace slab centering detection method according to any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium stores a heating furnace slab centering detection program, which, when executed by the processor, implements the steps of the heating furnace slab centering detection method according to any one of claims 1 to 7.

Images