A strip plate shape on-line detection method and related equipment

Abstract

The application belongs to the technical field of strip plate shape detection, and discloses a strip plate shape online detection method and related equipment. The method comprises the following steps: acquiring real-time working condition parameters of a to-be-detected strip in a detection area; the real-time working condition parameters at least comprise a motion speed parameter, a surface temperature parameter and a surface state parameter; collecting a reflection image of a preset optical pattern projected on the surface of the to-be-detected strip; determining an image processing strategy matched with a current working condition according to the real-time working condition parameters, so as to perform adaptive processing on the reflection image; the image processing strategy comprises at least one of a motion blur processing strategy, a contrast enhancement processing strategy and a contaminant artifact filtering processing strategy; based on the processed reflection image, plate shape defect features of the to-be-detected strip are extracted and converted into plate shape quantitative results; thereby, high-precision and high-reliability online detection of strip plate shape defects can be realized in a complex industrial environment.

Description

Technical Field

[0001] This application relates to the field of strip shape detection technology, and more specifically, to an online strip shape detection method and related equipment. Background Technology

[0002] Oriented silicon steel is usually produced in the form of steel strip. In order to ensure the shape quality of the product, the shape is usually inspected at multiple stages of the steel strip production process. This allows the process parameters to be adjusted based on the inspection results (for example, inspections are carried out upstream and downstream of the continuous annealing unit after cold rolling to adjust the continuous annealing process parameters based on the inspection results) to avoid the product's shape quality being unqualified.

[0003] Currently, online inspection of steel strip shape primarily utilizes non-contact optical inspection technology. This technology leverages the smooth surface of the steel strip, projecting specific optical patterns and capturing the deformation of the reflected patterns to determine shape defects. This method is particularly suitable for the continuous annealing process after cold rolling (continuous annealing units are mainly used for decarburization, nitriding, and stretching of steel strips) because it avoids damage to the steel strip and provides quantitative data on shape defects. However, in actual industrial production environments, optical inspection systems face numerous challenges.

[0004] First, fluctuations in the steel strip speed during production cause motion blur in the image, making the edges of the grid lines in the reflected pattern unclear and affecting the accurate extraction of plate shape defect features. Second, contaminants such as residual rolling emulsion and dust on the steel strip surface are stretched into stripe-like artifacts under the motion blur effect, mixing with the line deformation caused by actual plate shape defects, severely interfering with image analysis. Finally, when plate shaping is performed upstream of the continuous annealing unit, the detection area is close to the inlet of the continuous annealing unit. The surface temperature of the steel strip in this area will rise, which will change the physical properties of the residual emulsion, forming an oil film and reducing the contrast between the projected optical pattern and the steel strip background. When low-contrast images are superimposed with motion blur and contaminant artifacts, the image quality is severely degraded, and may even lead to measurement failure.

[0005] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0006] The purpose of this application is to provide an online detection method and related equipment for strip shape, which can achieve high-precision and high-reliability online detection of strip shape defects in complex industrial environments.

[0007] In a first aspect, this application provides an online strip shape detection method for online strip shape detection at the inlet of a continuous annealing unit. The method includes the following steps: A1. Obtain the real-time operating parameters of the strip steel to be tested in the detection area; the real-time operating parameters include at least the motion speed parameters, surface temperature parameters, and surface condition parameters. A2. Acquire the reflected image of the preset optical pattern projected onto the surface of the strip steel to be tested; A3. Based on the real-time operating parameters, determine an image processing strategy that matches the current operating conditions to adaptively process the reflected image; the image processing strategy includes at least one of the processing strategies for motion blur processing, contrast enhancement processing, and contaminant artifact filtering processing; A4. Based on the processed reflection image, extract the plate shape defect features of the strip steel to be tested, and convert them into plate shape quantification results.

[0008] Secondly, this application provides an online strip shape detection system for online strip shape detection at the inlet of a continuous annealing unit. The system includes: The working condition parameter acquisition module is used to acquire the real-time working condition parameters of the strip steel under test in the detection area; the real-time working condition parameters include at least the motion speed parameter, surface temperature parameter, and surface condition parameter. The reflection image acquisition module is used to acquire the reflection image of a preset optical pattern projected onto the surface of the strip steel to be tested; The image processing module is used to determine an image processing strategy that matches the current operating conditions based on the real-time operating parameters, so as to adaptively process the reflected image; the image processing strategy includes at least one of the processing strategies for motion blur processing, contrast enhancement processing, and pollutant artifact filtering processing. The strip shape quantization module is used to extract the strip shape defect features of the strip under test based on the processed reflection image and convert them into strip shape quantization results.

[0009] Thirdly, this application provides an electronic device including a processor and a memory, the memory storing a computer program executable by the processor, wherein when the processor executes the computer program, it performs the steps of the strip shape online detection method described above.

[0010] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the steps of the online strip shape detection method described above. Beneficial Effects: This application provides an online strip shape detection method and related equipment. By acquiring real-time operating parameters of the strip to be tested and adaptively determining image processing strategies based on these parameters, the acquired reflective images are processed to ultimately extract and quantify strip shape defect features. This method effectively solves problems in existing technologies such as motion blur caused by high-speed strip movement, artifacts formed by surface contaminants, and contrast reduction caused by increased surface temperature. By introducing the perception of real-time operating parameters and an adaptive image processing strategy, this application can specifically optimize image degradation under different operating conditions, such as compensating for motion blur, enhancing low-contrast areas, and filtering contaminant artifacts, thereby significantly improving the clarity and signal-to-noise ratio of the reflective images. Accordingly, this application can more accurately extract strip shape defect features and convert them into reliable strip shape quantification results, overcoming the limitation of existing technologies where detection accuracy is greatly affected by environmental factors, and achieving high-precision and high-reliability online detection of strip shape defects in complex industrial environments. Attached Figure Description

[0011] Figure 1 A flowchart of an online strip shape detection method provided in this application.

[0012] Figure 2 A schematic diagram of an online strip shape detection system provided in this application.

[0013] Figure 3 A schematic diagram of the structure of the electronic device provided in this application.

[0014] Labeling Explanation: 1. Operating Parameter Acquisition Module; 2. Reflection Image Acquisition Module; 3. Image Processing Module; 4. Plate Shape Quantization Module; 301. Processor; 302. Memory; 303. Communication Bus. Detailed Implementation

[0015] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0016] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0017] Please refer to Figure 1 This application discloses an online strip shape detection method in some embodiments, used for online strip shape detection at the inlet of a continuous annealing unit. The method includes the following steps: A1. Obtain the real-time operating parameters of the strip steel to be tested in the detection area; the real-time operating parameters include at least the motion speed parameters, surface temperature parameters, and surface condition parameters. A2. Acquire the reflected image of the preset optical pattern projected onto the surface of the strip steel to be tested; A3. Based on the real-time operating parameters, determine an image processing strategy that matches the current operating conditions to adaptively process the reflected image; the image processing strategy includes at least one of the processing strategies for motion blur processing, contrast enhancement processing, and contaminant artifact filtering processing; A4. Based on the processed reflection image, extract the plate shape defect features of the strip steel to be tested, and convert them into plate shape quantification results.

[0018] This application effectively addresses issues such as motion blur, low contrast, and contaminant artifacts by acquiring real-time operating parameters and adaptively adjusting image processing strategies based on these parameters, thereby significantly improving the accuracy of plate shape defect feature extraction and the reliability of plate shape quantization results.

[0019] This method is mainly applied to the online detection of strip steel at the inlet of the continuous annealing unit. The strip steel in this area is usually in a high-speed state and may have contaminants such as rolling emulsion and dust on its surface, while the temperature is also high.

[0020] "Real-time operating parameters" refer to dynamic data reflecting the current state of the strip during the inspection process. Their acquisition is crucial for the adaptive adjustment of subsequent image processing strategies. These parameters include at least motion speed parameters, surface temperature parameters, and surface state parameters. Motion speed parameters reflect the strip's running speed and directly affect the degree of motion blur in the image; surface temperature parameters reflect the temperature of the strip surface and may affect contaminant characteristics and image contrast; surface state parameters describe the type, distribution, and concentration of contaminants on the strip surface, providing guidance for the identification and filtering of contaminant artifacts.

[0021] "Preset optical pattern" refers to a specific geometric pattern projected onto the surface of the strip to be tested, such as a grid pattern, dot matrix pattern, or stripe pattern. After these patterns deform on the surface of the strip, the degree of deformation of their reflected image can be used to invert the strip's shape defects.

[0022] A "reflected image" refers to an image captured by an image acquisition device (such as a high-speed camera) showing a pre-defined optical pattern reflected from the surface of a steel strip. This image contains key information about the surface morphology of the steel strip, but it may also be affected by various interference factors.

[0023] "Image processing strategy" refers to a combination of algorithms and techniques used to improve the quality of reflective images, aiming to eliminate or mitigate interference such as motion blur, low contrast, and contaminant artifacts. Its adaptability is reflected in its ability to dynamically adjust processing methods and parameters based on real-time operating conditions to achieve optimal image processing results.

[0024] "Strip shape defect features" refer to geometric or physical quantities extracted from the processed reflection image that can characterize the strip shape defect, such as wave height, edge warping degree, etc.

[0025] "Strip shape quantification results" refers to converting the extracted strip shape defect features into quantifiable values ​​through a certain mathematical model or standard, so as to evaluate the strip shape quality.

[0026] The core of the online strip shape detection method of this application lies in the acquisition of real-time working parameters, the acquisition of reflective images, the adaptive image processing based on working parameters, and the extraction and quantification of strip shape defect features.

[0027] In step A1, real-time operating parameters of the strip under test in the detection area are acquired. These parameters include at least motion speed parameters, surface temperature parameters, and surface condition parameters. For example, motion speed parameters can be acquired using a speed sensor installed on the strip's movement path. This sensor can be a laser Doppler effect-based velocimeter, which calculates the real-time speed of the strip by measuring the frequency shift of the laser beam after reflection from the strip surface. Surface temperature parameters can be acquired using a non-contact infrared thermometer, which can measure the radiant energy of the strip surface in real time and convert it into a temperature value. Surface condition parameters can be acquired using multispectral imaging technology, which identifies and quantifies the type and concentration of surface contaminants by analyzing the reflectance spectral characteristics of the strip surface in different wavelength bands.

[0028] In step A2, a reflected image of a preset optical pattern projected onto the surface of the strip to be tested is acquired. This is typically achieved using an optical projection system and an image acquisition system. The optical projection system projects the preset optical pattern (e.g., generated by a laser or LED array) onto the strip surface. The image acquisition system, such as a high-resolution industrial camera, synchronously captures images of these patterns reflected off the strip surface at a certain frame rate. To ensure image quality, the camera is usually equipped with suitable lenses and filters to reduce ambient light interference.

[0029] In step A3, an image processing strategy matching the current operating conditions is determined based on real-time operating parameters to adaptively process the reflected image. The image processing strategy includes at least one of the following: motion blur processing, contrast enhancement processing, and contaminant artifact filtering processing. For example, when the motion speed parameter is high, the system can prioritize and enhance the motion blur processing strategy, such as using deconvolution-based algorithms or motion compensation algorithms to restore image sharpness. When the surface temperature parameter is high or the surface condition parameters indicate the presence of an oil film, the system can focus on contrast enhancement processing strategies, such as using histogram equalization, gamma correction, or local contrast enhancement algorithms to improve image visibility. When the surface condition parameters indicate the presence of dust or emulsion, the system can apply contaminant artifact filtering processing strategies, such as using morphological filtering, wavelet transform, or deep learning-based denoising algorithms to remove artifacts. These processing strategies can be dynamically selected and adjusted according to the combination of real-time operating parameters to achieve the best image processing effect.

[0030] In step A4, based on the processed reflection image, the shape defect features of the strip steel under test are extracted and converted into shape quantification results. For example, feature points of a preset optical pattern (such as the intersection of grid lines or the edges of stripes) can be identified first from the processed reflection image. Then, based on the spatial location information of these feature points, the three-dimensional topography data of the strip steel surface under test is reconstructed using the principles of triangulation or stereo vision technology. Next, according to the preset shape defect feature decomposition rules (e.g., based on Fourier transform, wavelet analysis, or eigenmode analysis), the three-dimensional topography data is decomposed into spatial frequencies to obtain shape components corresponding to different shape defect types (such as edge waves, center waves, and undulations). Finally, for each shape component, combined with the preset shape defect quantification standard, the quantification results of each shape defect type are calculated, such as the height, length, or curvature of the wave, thereby comprehensively evaluating the shape quality of the strip steel.

[0031] The online strip shape detection method proposed in this application has significant advantages and innovations compared to existing technologies. Traditional optical detection methods often suffer from decreased detection accuracy or even failure when faced with complex conditions such as high-speed strip movement at the inlet of continuous annealing units, surface contaminants, and temperature changes, due to image quality degradation. For example, motion blur makes the edges of the grid lines in the reflection pattern unclear, affecting the accurate extraction of strip shape defect features; stripe-like artifacts formed by contaminants remaining on the strip surface are mixed with real strip shape defects, severely interfering with image analysis; and when low-contrast images are superimposed with motion blur and contaminant artifacts, image quality is severely degraded, potentially leading to measurement failure.

[0032] The core innovation of this application lies in introducing the concept of "real-time operating parameters" and implementing "adaptive image processing strategies" based on these parameters. Specifically, by acquiring the motion speed parameters, surface temperature parameters, and surface state parameters of the strip under test, the system can comprehensively perceive the current detection environment. For example, when the strip speed increases, the system can sense the increased risk of motion blur and automatically adjust the image processing strategy, such as enhancing the intensity of motion blur processing or switching to a more efficient deblurring algorithm. When specific contaminants are detected on the strip surface, the system can select and optimize the contaminant artifact filtering strategy according to the type and concentration of the contaminants, thereby more accurately removing artifacts and avoiding confusion with real strip shape defects. In addition, when the surface temperature rises and the contrast decreases, the system can adaptively adjust the contrast enhancement processing strategy to ensure that effective image information is not lost. This adaptive processing mechanism is not available in existing technologies. Traditional non-contact optical inspection systems typically use fixed image processing algorithms or manually adjust parameters, making it difficult to cope with dynamic changes in operating conditions during production. Once the operating conditions deviate from the preset range, their detection performance will drop sharply. This application effectively solves problems such as motion blur, low contrast, and contaminant artifacts by sensing the working conditions in real time and dynamically adjusting the processing strategy, significantly improving the accuracy of strip shape defect feature extraction and the reliability of strip shape quantification results. Therefore, the method of this application can ensure high-quality detection images and accurate strip shape quantification results in various complex and changing industrial environments, thus providing strong technical support for quality control in the strip steel production process.

[0033] In some implementations, step A1 includes: A101. Motion speed parameters are obtained using a laser velocimeter based on the Doppler effect; A102. Surface temperature parameters are obtained using an infrared thermometer; A103. Obtain multispectral reflectance images of the surface of the strip under test in multiple preset wavebands; A104. Analyze the light intensity and spectral characteristics of the multispectral reflectance images under different spectral bands to identify the type, distribution information and concentration of pollutants, as surface state parameters.

[0034] Specifically, motion velocity parameters can be obtained using a laser velocimeter based on the Doppler effect. The laser velocimeter emits a laser beam and receives the scattered light reflected from the surface of a moving object, calculating the object's velocity using the Doppler frequency shift principle. Its purpose is to provide high-precision, non-contact strip motion velocity data, providing an accurate basis for subsequent motion blur processing. Surface temperature parameters can be obtained using an infrared thermometer. Infrared thermometers determine the temperature of an object by measuring the infrared energy radiated from its surface, offering advantages such as non-contact operation and fast response. Its purpose is to monitor the strip surface temperature in real time, providing input for temperature-related image processing (such as contrast adjustment) and assisting in judging the surface condition.

[0035] In practical applications, to obtain more comprehensive surface condition information, multispectral reflectance images of the strip surface under test can be acquired in multiple preset wavelength bands. Multispectral imaging technology can capture the reflectance characteristics of objects in different wavelength ranges, thereby revealing surface details and material composition information invisible to the naked eye. Its purpose is to provide rich spectral data for identifying contaminants on the strip surface. Furthermore, by analyzing the light intensity and spectral characteristics of the multispectral reflectance images in different wavelength bands, the type, distribution information, and concentration of contaminants can be identified as surface condition parameters. For example, different types of oil, oxides, or dust have unique absorption or reflectance spectral characteristics in specific wavelength bands; by comparing with a preset spectral database, these contaminants can be accurately identified. Its purpose is to provide detailed surface contaminant information so that targeted contaminant artifact filtering can be performed during image processing, improving the accuracy of strip shape defect identification.

[0036] This application's solution employs a Doppler-effect-based laser velocimeter to obtain precise motion velocity parameters, providing accurate compensation for subsequent motion blur processing and effectively reducing image blur caused by high-speed strip movement. Simultaneously, real-time surface temperature parameters are acquired using an infrared thermometer, providing temperature compensation for image contrast enhancement and ensuring consistent image quality under different temperature conditions. More importantly, by acquiring multispectral reflectance images and analyzing their light intensity and spectral characteristics, the type, distribution information, and concentration of contaminants are identified, enabling the system to accurately grasp the contamination status of the strip surface. Therefore, when determining subsequent image processing strategies, contaminant artifact filtering strategies can be selectively chosen and adjusted. For example, based on the specific type and concentration of contaminants, filter parameters can be dynamically adjusted or specific decontamination algorithms can be selected, effectively removing artifacts caused by contaminants and avoiding their interference with the extraction of strip shape defect features.

[0037] In some implementations, step A3 includes: A301. Based on the real-time operating parameters, determine the image processing strategy and its initial parameter set that match the current operating conditions; A302. The image reflection image is preliminarily processed using the image processing strategy and its initial parameter set; A303. Evaluate the image processing effect of the pre-processed reflected image based on the preset evaluation model; A304. If the image processing effect does not meet the preset defect recognition reliability requirements, the parameters of the image processing strategy are adjusted, and the adjusted parameters are repeatedly used to process the reflected image, and the image processing effect is evaluated until the image processing effect meets the defect recognition reliability requirements, and the processed reflected image is obtained.

[0038] Specifically, in step A301, the system selects one or more basic strategies from a preset image processing strategy library based on the acquired real-time operating parameters, such as motion speed parameters, surface temperature parameters, and surface state parameters, and determines a set of initial parameters. These initial parameter sets can be pre-trained based on historical data, empirical rules, or machine learning models. The image processing strategies can include at least one of motion blur processing, contrast enhancement processing, and contaminant artifact filtering processing. For example, multiple different motion blur processing strategies, multiple different contrast enhancement processing strategies, and multiple contaminant artifact filtering processing strategies can be pre-stored in the image processing strategy library. Each motion blur processing strategy has its optimal motion speed range pre-calibrated, each contrast enhancement processing strategy has its optimal temperature range pre-calibrated, and each contaminant artifact filtering processing strategy has its optimal surface state parameter range pre-calibrated. Thus, the optimal combination of processing strategies can be matched based on the ranges to which the real-time operating parameters fall.

[0039] Further, in step A302, the reflected image of the preset optical pattern projected onto the surface of the strip to be tested is first pre-processed using the image processing strategy and initial parameter set determined in step A301. This pre-processing aims to perform preliminary denoising, enhancement, or correction on the image.

[0040] Subsequently, in step A303, the image processing effect of the pre-processed reflective image is evaluated according to a preset evaluation model. The evaluation model may include various image quality evaluation indicators, such as image sharpness, signal-to-noise ratio, edge sharpness, contrast of defect features, and recognizability. These indicators are used to quantify the quality of the processed image and compare it with preset defect recognition reliability requirements. For example, a weighted sum of these image quality evaluation indicators can be calculated to obtain a comprehensive image quality score, and these image quality evaluation indicators and the comprehensive image quality score can be used together as quantitative data of the image processing effect.

[0041] In a preferred embodiment, in step A304, if the evaluation result shows that the image processing effect does not meet the preset defect recognition reliability requirements, a parameter adjustment mechanism is triggered. This mechanism intelligently adjusts the parameters of the image processing strategy based on the evaluation results. For example, if the image is still blurry, the intensity of motion blur processing is increased; if the contrast is insufficient, the contrast enhancement parameters are adjusted. The adjusted parameters are then applied to the reflected image again for processing, and its effect is re-evaluated. This iterative process continues until the image processing effect meets the preset defect recognition reliability requirements, thereby obtaining the final processed reflected image. For example, the defect recognition reliability requirements may be: each image quality evaluation index is not less than its respective preset index threshold, and the overall image quality score is not less than the preset overall score threshold.

[0042] This application's solution addresses the limitations of determining the image processing strategy solely based on real-time operating parameters by introducing a closed-loop adaptive adjustment mechanism. Specifically, in step A301, the system preliminarily determines the image processing strategy and its initial parameter set based on real-time operating parameters, providing a starting point for subsequent processing. Subsequently, in step A302, the reflected image undergoes preliminary processing. The key steps are A303 and A304, where a preset evaluation model quantitatively evaluates the effect of the preliminarily processed image and compares it with the reliability requirements for defect identification. If the effect is unsatisfactory, the parameters of the image processing strategy are iteratively adjusted in step A304. This iterative adjustment and evaluation feedback mechanism allows the image processing strategy to be dynamically optimized based on the actual processing effect, thereby overcoming the impact of potentially inaccurate initial parameters or changes in operating parameters, ensuring that the final processed reflected image meets the requirements for high-precision plate-shaped defect identification.

[0043] In some implementations, step A4 includes: A401. Identify feature points of the preset optical pattern from the processed reflection image; A402. Based on the feature points, construct the three-dimensional morphology data of the surface of the strip steel to be tested; A403. Perform spatial frequency decomposition on the three-dimensional topography data, and combine it with the preset plate shape defect feature decomposition rules to obtain topography components corresponding to different plate shape defect types, which are used as plate shape defect features for each plate shape defect type. A404. For each shape component, calculate the quantification results of each type of plate shape defect based on the preset plate shape defect quantification standard.

[0044] In step A401, the preset optical pattern typically refers to a structured light pattern projected onto the surface of the steel strip, such as parallel stripes, grids, or dot patterns. Feature points are specific points in the reflected image that can be precisely identified and located within these optical patterns (e.g., corners or vertices of polygonal patterns, centers of circular or dotted patterns, intersections of straight lines or grids, etc.). Their positional information in the image is fundamental to subsequent 3D reconstruction. In practical applications, these feature points can be identified with high precision using image processing algorithms such as sub-pixel edge detection and phase unwrapping techniques.

[0045] Further, in step A402, based on the identified feature points and combined with the known geometric parameters of the projection and acquisition devices (e.g., camera calibration parameters, projector calibration parameters, and their relative positional relationships), the two-dimensional image coordinates of these feature points are converted into three-dimensional spatial coordinates of the strip surface under test using the principles of triangulation or structured light 3D reconstruction algorithms, thereby constructing the three-dimensional topographic data of the strip surface. This three-dimensional topographic data accurately reflects the undulations and deformation of the strip surface.

[0046] In step A403, spatial frequency decomposition is a technique that decomposes complex morphology into components of different spatial scales or modes. For example, methods such as Fourier transform, wavelet analysis, or eigenmode analysis can be used. Different types of plate defects, such as edge waves, center waves, and warps, exhibit different wavelengths, amplitudes, and distribution characteristics in space, i.e., they have different spatial frequency features. Through spatial frequency decomposition, these morphological components with specific spatial frequency ranges can be separated. Each morphological component corresponds to one type or class of plate defects, thus serving as the plate defect feature for each type of plate defect.

[0047] Finally, in step A404, for each separated topographic component, the quantification result for each type of plate shape defect is calculated based on a preset plate shape defect quantification standard. The plate shape defect quantification standard is a pre-defined evaluation criterion used to convert the geometric characteristics of the topographic components (such as amplitude, wavelength, curvature, gradient, etc.) into specific quantified values ​​or levels. For example, the severity of the defect can be determined by comparing the maximum amplitude of the topographic component with a preset threshold, and the corresponding quantification result can be output.

[0048] The proposed solution, by refining step A4 into a series of specific operations, effectively addresses the potential issues of insufficient accuracy and low discriminative power in the extraction and quantification of strip defect features in the basic approach. Specifically, firstly, by identifying feature points of a preset optical pattern and constructing three-dimensional topographic data of the strip surface to be tested, the true geometric morphology of the strip surface can be accurately obtained, providing high-fidelity raw data for subsequent defect analysis. Secondly, spatial frequency decomposition is performed on the three-dimensional topographic data. This process decomposes the complex strip surface morphology into independent topographic components corresponding to different strip defect types. For example, different types of strip defects such as edge waves, center waves, and warps have different characteristic frequencies and distribution patterns in space. Through spatial frequency decomposition, the unique features of these defects can be effectively separated and identified, overcoming the limitations of traditional methods in accurately distinguishing and quantifying various strip defects. Finally, for each topographic component, the quantification results are calculated based on preset strip defect quantification standards, ensuring that each defect type can be objectively and accurately evaluated, providing a reliable basis for precise control of the production process.

[0049] Preferably, step A403 may include: The three-dimensional topography data is decomposed using the characteristic modality analysis method to obtain topography modes with spatial distribution characteristics; Based on the preset plate shape defect feature decomposition rules, the morphological modes related to each of the plate shape defect types are identified as corresponding morphological components, which serve as plate shape defect features for each plate shape defect type.

[0050] Specifically, the characteristic modal analysis method refers to a method that decomposes complex morphology data into a series of basis functions (i.e., morphology modes) with specific spatial distribution patterns. These morphology modes are mutually orthogonal and can effectively capture the main deformation characteristics of the strip surface morphology. The morphology modes with spatial distribution characteristics can be understood as the inherent shapes or patterns obtained through the characteristic modal analysis method that reflect specific deformation patterns on the strip surface. For example, some modes may represent central wave defects in the strip, while others may represent edge waves or half-wave defects. Its purpose is to decompose the complex overall morphology into basic components that are easier to understand and quantify.

[0051] In practical applications, identifying the morphological modes associated with each type of plate defect as the corresponding morphological components specifically refers to matching and associating the decomposed morphological modes with known plate defect types (such as center waves, edge waves, half waves, etc.) according to preset plate defect feature decomposition rules. For example, the defect type represented by a morphological mode can be determined by comparing its shape, amplitude, and spatial location with standard defect patterns, and then used as the morphological component of that defect type.

[0052] The proposed solution introduces a characteristic modal analysis method to decompose the three-dimensional morphology data of the strip surface under test into a series of morphology modes with clear physical meaning and spatial distribution characteristics. These modes can more accurately capture and characterize different types of plate shape defects, overcoming the limitations of traditional spatial frequency decomposition in distinguishing complex defect modes. Because these morphology modes have a stronger correlation with specific plate shape defect types, subsequent methods, based on preset plate shape defect feature decomposition rules, can more accurately identify the morphology modes related to each plate shape defect type and use them as corresponding morphology components, thereby improving the accuracy and reliability of plate shape defect feature extraction.

[0053] The above technical solution enables the decomposition of complex three-dimensional strip morphology data into morphology modes with clear physical meaning and spatial distribution characteristics, thereby more accurately and effectively identifying and separating different types of strip shape defects. Compared with general spatial frequency decomposition, this method can provide more interpretable morphology components, significantly improving the accuracy and reliability of strip shape defect feature extraction, and providing a more solid foundation for subsequent strip shape quantization.

[0054] In some preferred embodiments, after acquiring the three-dimensional morphology data of the strip surface to be tested, characteristic mode analysis can be performed using methods such as Principal Component Analysis (PCA), Independent Component Analysis (ICA), or Empirical Mode Decomposition (EMD). For example, using the PCA method, the three-dimensional morphology data can be projected onto a set of orthogonal principal components, which are morphology modes with spatial distribution characteristics. Among them, the principal component with the largest contribution may correspond to the main strip shape defects, such as center waves or edge waves. Subsequently, according to the preset strip shape defect characteristic decomposition rules, such as by analyzing the shape characteristics, amplitude distribution, and correspondence with the strip width direction of each principal component, the morphology modes related to strip shape defect types such as center waves, edge waves, and half waves can be identified and used as the corresponding morphology components. Thus, the characteristics of each defect can be accurately extracted, providing accurate input for subsequent quantitative analysis.

[0055] Preferably, step A404 may include: For each morphological component, its geometric feature parameters are extracted; the geometric feature parameters include at least one of amplitude, wavelength, curvature and gradient; By comparing the geometric feature parameters with preset thresholds and combining them with preset plate shape defect quantification standards, the quantification results of each plate shape defect type are determined.

[0056] Among them, geometric characteristic parameters refer to numerical attributes that can quantitatively describe the physical morphology and severity of plate-shaped defects. Specifically, amplitude can be understood as the peak and valley height or maximum deviation of the defect, reflecting the severity of the defect; wavelength can be understood as the periodic length of the defect, reflecting the distribution characteristics of the defect; curvature can be understood as the degree of bending of the defect surface, reflecting the severity of local deformation; and gradient can be understood as the rate of change of the defect surface height, reflecting the steepness of the defect. These parameters can be accurately extracted from the morphological components obtained by spatial frequency decomposition, with the aim of providing an objective and measurable basis for the quantification of plate-shaped defects.

[0057] Preset thresholds refer to critical values ​​set in advance based on industry standards, product specifications, historical data, or expert experience. These thresholds are used to determine whether the extracted geometric feature parameters exceed acceptable ranges, thereby distinguishing between normal fluctuations and actual defects, and making a preliminary classification of the severity of defects. Plate shape defect quantification standards refer to a complete set of rules used to convert the comparison results of geometric feature parameters and thresholds into specific quantitative values ​​or levels. For example, defects can be divided into different levels such as minor, moderate, and severe, or specific quantitative indices can be provided.

[0058] The proposed solution extracts specific geometric feature parameters from each morphological component, transforming abstract morphological information into quantifiable physical properties. By comparing these geometric feature parameters with preset thresholds, it is possible to objectively determine whether a defect exceeds the standard and the degree of exceeding the standard. Furthermore, by combining this with preset plate shape defect quantification standards, the comparison results can be transformed into meaningful quantitative values ​​or levels. This method ensures that the quantification of plate shape defects no longer relies solely on the overall perception of morphological components but is based on their inherent, measurable physical characteristics, thereby significantly improving the accuracy and reliability of the quantification results.

[0059] Through the above technical solutions, the quantification process of plate shape defects becomes more refined and objective. By extracting geometric feature parameters such as amplitude, wavelength, curvature, and gradient, the characteristics of different types of plate shape defects can be described more comprehensively and accurately. Comparing these parameters with preset thresholds and combining them with quantification standards, the severity of defects can be accurately assessed, avoiding the bias of subjective judgment. This improves the reliability and consistency of plate shape detection, providing more solid data support for subsequent production process control and product quality assessment.

[0060] In some preferred embodiments, it is assumed that a typical edge waviness defect is identified during the detection process, which manifests as a specific shape component in the three-dimensional topographic data. The maximum amplitude and average wavelength can then be extracted from this shape component. For example, if the extracted amplitude is 0.8 mm and the wavelength is 150 mm, the amplitude is compared with a preset amplitude threshold (e.g., 0.5 mm), and it is found to exceed the threshold. Simultaneously, the wavelength is compared with a preset wavelength range (e.g., 100-200 mm), and it is found to be within that range. Based on a preset plate shape defect quantification standard, and considering both the amplitude exceeding the threshold and the wavelength being within the specific range, the edge waviness defect may be quantified as a "moderate defect," and a corresponding quantification index, such as 70 points (out of 100), is given. This quantification method based on specific geometric parameters makes the defect evaluation results more specific and operable.

[0061] Preferably, the step of comparing the geometric feature parameters with a preset threshold and combining them with a preset plate shape defect quantification standard to determine the quantification result of each plate shape defect type may include: Obtain the product specification information of the strip steel to be tested; the product specification information includes at least one of material, thickness and width. Based on the product specification information, the preset geometric feature parameter thresholds and plate shape defect quantification standards are adjusted to obtain dynamic thresholds and dynamic quantification standards that match the current product specifications. By comparing the geometric feature parameters with the dynamic threshold and combining them with the dynamic quantification standard, the quantification results of each plate shape defect type are determined.

[0062] Specifically, obtaining the product specification information of the strip steel to be tested refers to automatically acquiring the product specification information of the strip steel currently in production by interacting with the production management system (MES) or production planning system before performing strip shape inspection. This product specification information may include, but is not limited to, the material type, thickness, and width of the strip steel. This information forms the basis for subsequent dynamic adjustment of thresholds and quantification standards.

[0063] Specifically, based on the product specification information, preset geometric feature parameter thresholds and strip shape defect quantification standards are adjusted to obtain dynamic thresholds and dynamic quantification standards that match the current product specifications. This can be understood as the system maintaining a database or rule base containing mapping relationships between various product specifications and corresponding strip shape defect evaluation parameters. Upon obtaining the current strip steel product specification information, the system queries this database or rule base and intelligently selects or calculates a set of geometric feature parameter thresholds and strip shape defect quantification standards most suitable for the current product evaluation based on the strip steel's material, thickness, width, and other characteristics. For example, for thinner strip steel, its sensitivity to strip shape defects is higher, therefore the thresholds for corresponding geometric feature parameters such as amplitude and curvature will be set more strictly; while for strip steel of a specific material, the allowable wavelength range may differ.

[0064] In practical applications, comparing the geometric feature parameters with the dynamic threshold, and combining this with the dynamic quantification standard, to determine the quantification result of each plate defect type refers to comparing the geometric feature parameters (such as amplitude, wavelength, curvature, gradient, etc.) of each plate defect type extracted from the processed reflection image with the threshold dynamically adjusted according to the current product specifications. For example, if the amplitude of a plate defect exceeds the dynamic amplitude threshold set for the current product specifications, the defect is judged to be out of specification. Simultaneously, the severity of the defect is quantified using the dynamic quantification standard, for example, by classifying it into different levels such as minor, moderate, and severe, or assigning it a specific quantification score.

[0065] This application's solution achieves dynamic adaptive adjustment of the strip shape defect assessment standard by incorporating product specification information. Traditional methods, with their fixed thresholds and quantification standards, struggle to adapt to the diverse production needs of strip steel products, potentially leading to deviations in the assessment of strip shape defects for different specifications. For example, for thin strip steel with high precision requirements, even minute shape fluctuations may be considered serious defects; while for thick strip steel, the same fluctuations might be within acceptable limits. By acquiring the product specification information of the strip steel currently being tested, the system can intelligently adjust the thresholds of geometric feature parameters and the quantification standard for strip shape defects based on the strip steel's material, thickness, width, and other characteristics. This dynamic adjustment mechanism ensures that the assessment of strip shape defects always matches the actual quality requirements of the current product, thus avoiding misjudgments or omissions due to standard mismatches. Consequently, the accuracy and reliability of the strip shape quantification results are significantly improved, enabling the strip shape detection system to more accurately guide production and optimize process parameters.

[0066] Through the above technical solution, this application can dynamically adjust the evaluation threshold and quantification standard of strip shape defects according to different product specifications, thereby significantly improving the accuracy and adaptability of strip shape defect detection. Compared with the traditional method using fixed thresholds, this application can avoid misjudgments or omissions caused by differences in product specifications, making the strip shape quantification results more in line with actual production needs and product quality standards. This not only helps to improve the precision of product quality control and reduce unnecessary rework or scrap, but also provides more accurate strip shape data feedback for the production line, supports the optimization and adjustment of process parameters, and thus improves production efficiency and economic benefits.

[0067] refer to Figure 2 This application provides an online strip shape detection system for online strip shape detection at the inlet of a continuous annealing unit. The system includes: The working condition parameter acquisition module 1 is used to acquire the real-time working condition parameters of the strip steel under test in the detection area; the real-time working condition parameters include at least the motion speed parameter, surface temperature parameter and surface state parameter (the specific process can be referred to step A1 above). The reflection image acquisition module 2 is used to acquire the reflection image of the preset optical pattern projected onto the surface of the strip steel to be tested (the specific process can be referred to step A2 above). Image processing module 3 is used to determine an image processing strategy that matches the current operating conditions based on the real-time operating parameters, so as to adaptively process the reflected image; the image processing strategy includes at least one of the processing strategies for motion blur processing, contrast enhancement processing and pollutant artifact filtering processing (the specific process can be referred to step A3 above). The strip shape quantification module 4 is used to extract the strip shape defect features of the strip under test based on the processed reflection image and convert them into strip shape quantification results (for details, please refer to step A4 above).

[0068] Please refer to Figure 3This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device includes a processor 301 and a memory 302. The processor 301 and the memory 302 are interconnected and communicate with each other via a communication bus 303 and / or other connection mechanisms (not shown). The memory 302 stores a computer program executable by the processor 301. When the electronic device is running, the processor 301 executes the computer program to perform online strip shape detection in any optional implementation of the above embodiments, to achieve the following functions: acquiring real-time operating parameters of the strip under test in the detection area; the real-time operating parameters include at least motion speed parameters, surface temperature parameters, and surface state parameters; acquiring a reflected image of a preset optical pattern projected onto the surface of the strip under test; determining an image processing strategy matching the current operating conditions based on the real-time operating parameters to adaptively process the reflected image; the image processing strategy includes at least one of motion blur processing, contrast enhancement processing, and contaminant artifact filtering processing; and extracting the strip shape defect features of the strip under test based on the processed reflected image, and converting them into a strip shape quantification result.

[0069] This application provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it performs online strip shape detection in any optional implementation of the above embodiments to achieve the following functions: acquiring real-time operating parameters of the strip under test in the detection area; the real-time operating parameters include at least motion speed parameters, surface temperature parameters, and surface state parameters; acquiring a reflection image of a preset optical pattern projected onto the surface of the strip under test; determining an image processing strategy matching the current operating conditions based on the real-time operating parameters to adaptively process the reflection image; the image processing strategy includes at least one of motion blur processing, contrast enhancement processing, and contaminant artifact filtering processing; and extracting the strip shape defect features of the strip under test based on the processed reflection image and converting them into strip shape quantification results. The computer-readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0070] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for online strip shape detection, used for online strip shape detection at the inlet of a continuous annealing unit, characterized in that, The steps of this method include: A1. Obtain the real-time operating parameters of the strip steel to be tested in the detection area; the real-time operating parameters include at least the motion speed parameters, surface temperature parameters, and surface condition parameters. A2. Acquire the reflected image of the preset optical pattern projected onto the surface of the strip steel to be tested; A3. Based on the real-time operating parameters, determine an image processing strategy that matches the current operating conditions to adaptively process the reflected image; the image processing strategy includes at least one of the processing strategies for motion blur processing, contrast enhancement processing, and contaminant artifact filtering processing; A4. Based on the processed reflection image, extract the plate shape defect features of the strip steel to be tested, and convert them into plate shape quantification results.

2. The method for online detection of strip shape according to claim 1, characterized in that, Step A1 includes: A101. Motion speed parameters are obtained using a laser velocimeter based on the Doppler effect; A102. Surface temperature parameters are obtained using an infrared thermometer; A103. Obtain multispectral reflectance images of the surface of the strip under test in multiple preset wavebands; A104. Analyze the light intensity and spectral characteristics of the multispectral reflectance images under different spectral bands to identify the type, distribution information and concentration of pollutants, as surface state parameters.

3. The method for online detection of strip shape according to claim 1, characterized in that, Step A3 includes: A301. Based on the real-time operating parameters, determine the image processing strategy and its initial parameter set that match the current operating conditions; A302. The image reflection image is preliminarily processed using the image processing strategy and its initial parameter set; A303. Evaluate the image processing effect of the pre-processed reflected image based on the preset evaluation model; A304. If the image processing effect does not meet the preset defect recognition reliability requirements, the parameters of the image processing strategy are adjusted, and the adjusted parameters are repeatedly used to process the reflected image, and the image processing effect is evaluated until the image processing effect meets the defect recognition reliability requirements, and the processed reflected image is obtained.

4. The method for online detection of strip shape according to claim 1, characterized in that, Step A4 includes: A401. Identify feature points of the preset optical pattern from the processed reflection image; A402. Based on the feature points, construct the three-dimensional morphology data of the surface of the strip steel to be tested; A403. Perform spatial frequency decomposition on the three-dimensional topography data, and combine it with the preset plate shape defect feature decomposition rules to obtain topography components corresponding to different plate shape defect types, which are used as plate shape defect features for each plate shape defect type. A404. For each shape component, calculate the quantification results of each type of plate shape defect based on the preset plate shape defect quantification standard.

5. The method for online detection of strip shape according to claim 4, characterized in that, Step A403 includes: The three-dimensional topography data is decomposed using the characteristic modality analysis method to obtain topography modes with spatial distribution characteristics; Based on the preset plate shape defect feature decomposition rules, the morphological modes related to each of the plate shape defect types are identified as corresponding morphological components, which serve as plate shape defect features for each plate shape defect type.

6. The method for online detection of strip shape according to claim 5, characterized in that, Step A404 includes: For each morphological component, its geometric feature parameters are extracted; the geometric feature parameters include at least one of amplitude, wavelength, curvature and gradient; By comparing the geometric feature parameters with preset thresholds and combining them with preset plate shape defect quantification standards, the quantification results of each plate shape defect type are determined.

7. The method for online detection of strip shape according to claim 6, characterized in that, The step of comparing the geometric feature parameters with a preset threshold and combining them with a preset plate shape defect quantification standard to determine the quantification result of each plate shape defect type includes: Obtain the product specification information of the strip steel to be tested; the product specification information includes at least one of material, thickness and width. Based on the product specification information, the preset geometric feature parameter thresholds and plate shape defect quantification standards are adjusted to obtain dynamic thresholds and dynamic quantification standards that match the current product specifications. By comparing the geometric feature parameters with the dynamic threshold and combining them with the dynamic quantification standard, the quantification results of each plate shape defect type are determined.

8. A strip shape online detection system for online strip shape detection at the inlet of a continuous annealing unit, characterized in that, The system includes: The working condition parameter acquisition module is used to acquire the real-time working condition parameters of the strip steel under test in the detection area; the real-time working condition parameters include at least the motion speed parameter, surface temperature parameter, and surface condition parameter. The reflection image acquisition module is used to acquire the reflection image of a preset optical pattern projected onto the surface of the strip steel to be tested; The image processing module is used to determine an image processing strategy that matches the current operating conditions based on the real-time operating parameters, so as to adaptively process the reflected image; the image processing strategy includes at least one of the processing strategies for motion blur processing, contrast enhancement processing, and pollutant artifact filtering processing. The strip shape quantization module is used to extract the strip shape defect features of the strip under test based on the processed reflection image and convert them into strip shape quantization results.

9. An electronic device, characterized in that, It includes a processor and a memory, the memory storing a computer program executable by the processor, which, when executing the computer program, performs the steps of the strip shape online detection method as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it performs the steps of the online strip shape detection method as described in any one of claims 1-7.

Images