Automatic judgment system for surface quality of hot rolled strip product

Abstract

The invention discloses an automatic judgment system for the surface quality of a hot rolled strip product, and relates to the technical field of industrial automation and machine vision, the automatic judgment system comprises an image acquisition module, an image preprocessing module, a defect detection module, a defect classification and evaluation module and a comprehensive judgment module; the image acquisition module is used for acquiring clear image data of the surface of a hot-rolled strip product in real time; the image preprocessing module is electrically connected with the image acquisition module. According to the automatic judgment system provided by the invention, the automatic, efficient and precise judgment of the surface quality of a hot rolled strip product is realized by integrating the image acquisition module, the image preprocessing module, the defect detection module, the defect classification and evaluation module and the comprehensive judgment module, and the system can acquire and process the image data of the surface of the hot rolled strip in real time; the evaluation result is more visual and easy to understand, and the method has important significance for improving the production quality and market competitiveness of hot rolled strip products.

Description

Technical Field

[0001] The invention relates to the technical field of industrial automation and machine vision, and in particular to an automatic determination system for the surface quality of hot-rolled strip products. Background Art

[0002] In the production process of hot-rolled strip products, surface quality is one of the important factors affecting the final quality of the product. Traditionally, the surface quality inspection of hot-rolled strip products mainly relies on manual visual inspection. This method is not only inefficient, but also easily affected by human subjective factors, making it difficult to ensure the accuracy and consistency of the inspection results. In addition, as the speed of the production line continues to increase, manual inspection is increasingly difficult to meet the real-time and efficient inspection needs of the production line.

[0003] Although the existing automated inspection systems have improved the inspection efficiency to a certain extent, there are still some problems. Some systems have deficiencies in image acquisition and processing, making it difficult to obtain clear image data, which affects the accuracy of defect detection. At the same time, the algorithms of some systems in defect detection and classification are not mature enough, and they are unable to accurately identify various types of defects and scientifically evaluate the severity of defects. In this regard, we propose an automatic determination system for the surface quality of hot-rolled plate and strip products. Summary of the invention

[0004] In order to solve the above technical problems, a system for automatically determining the surface quality of hot-rolled strip products is provided. This technical solution solves the above problems.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] An automatic determination system for the surface quality of a hot-rolled strip product, the automatic determination system comprising:

[0007] Image acquisition module, image preprocessing module, defect detection module, defect classification and evaluation module and comprehensive judgment module;

[0008] The image acquisition module is used to obtain clear image data of the surface of hot-rolled strip products in real time;

[0009] The image preprocessing module is electrically connected to the image acquisition module, and the image preprocessing module is used to perform preliminary optimization processing on the acquired image, and the optimization processing includes: denoising, contrast enhancement and geometric distortion correction;

[0010] The defect detection module is electrically connected to the image preprocessing module, and the defect detection module is used to automatically identify defects on the surface of the hot-rolled strip and distinguish defects from normal surfaces through image processing algorithms and pattern recognition technology;

[0011] The defect classification and evaluation module is electrically connected to the defect detection module, and is used to classify the detected defects and evaluate their severity levels according to preset standards, including the size, shape, location of the defects and the degree of impact on product quality;

[0012] The comprehensive determination module is electrically connected to the defect classification and evaluation module, and the comprehensive determination module is used to comprehensively evaluate the surface quality of the entire hot-rolled strip product based on the results of defect detection, classification and evaluation.

[0013] Preferably, the image acquisition module includes:

[0014] An image capturing unit, the image capturing unit is used to capture image data of the surface of the hot-rolled strip;

[0015] A light source control unit, which is used to adjust the intensity, color and direction of the light source to provide light source conditions for capturing image data;

[0016] Image stabilization unit, which is used to stabilize the image acquisition process on a moving hot strip production line to reduce blur caused by vibration;

[0017] The data transmission unit is used to transmit the captured image data to the image acquisition module.

[0018] Preferably, the image preprocessing module includes:

[0019] A denoising processing unit, which is responsible for denoising the collected image;

[0020] Contrast enhancement unit, the contrast enhancement unit is used to enhance the contrast of the image and make the details in the image clear;

[0021] The geometric distortion correction unit is used to correct the geometric distortion in the image, and the geometric distortion includes perspective distortion and lens distortion.

[0022] Preferably, the defect detection module comprises:

[0023] An image analysis unit, which is responsible for receiving pre-processed image data and applying an image processing algorithm to extract feature information in the image;

[0024] The defect recognition unit is used to recognize defect areas in the image based on the feature information extracted by the image analysis unit by using pattern recognition technology to distinguish defects from normal surfaces.

[0025] Preferably, the defect classification and assessment module includes:

[0026] A defect classification unit, the defect classification unit is used to classify the detected defects according to a predetermined classification standard;

[0027] Defect assessment unit, which is used to assess the severity of each defect according to preset assessment criteria after classification;

[0028] The result summary and output unit is responsible for summarizing the results of classification and evaluation and outputting them in the form of a report.

[0029] Preferably, the comprehensive determination module includes:

[0030] Result summary unit: The result summary unit is used to collect all defect detection results, classification information and severity level evaluation from the defect classification and evaluation module and summarize them into a comprehensive data set;

[0031] The quality assessment standard unit is used to store and manage preset standards or rules for surface quality assessment;

[0032] The comprehensive evaluation algorithm unit is used to quantitatively evaluate the surface quality of the entire hot-rolled plate and strip product through an algorithm based on the data set provided by the result summary unit and the preset standards in the quality evaluation standard unit.

[0033] Preferably, the method for the defect detection module to automatically identify the surface defects of the hot-rolled strip is:

[0034] The image analysis unit receives the optimized image data from the image preprocessing module;

[0035] Extract features from images and use gray-level co-occurrence matrix to identify abnormalities in surface texture;

[0036] Select the most valuable feature set for defect identification from the extracted features and normalize the feature set;

[0037] Build a defect detection model based on CNN for pattern recognition, and train the model using image data of known defects and normal surfaces;

[0038] The image features to be detected are input into the trained model, and the model outputs the result of whether the image area contains defects;

[0039] The defect area identified by the model is marked on the original image.

[0040] Preferably, the calculation method for identifying abnormal conditions of surface texture using the gray level co-occurrence matrix is:

[0041] Assume that the size of image I is M×N and the gray level is G, then the gray level co-occurrence matrix P is a G×G matrix, where P(i,j) represents the number of occurrences of all pixels with gray level i and gray level j in the image in a specific direction and distance d;

[0042] Among them, the calculation formula of P(i,j) is:

[0043]

[0044] Where δ represents the Kronecker function.

[0045] Preferably, the defect classification and assessment module classifies the detected defects and assesses their severity levels according to preset standards as follows:

[0046] The feature data of the identified defect area received from the defect detection module is further extracted to obtain the specific features of the defect and output a detailed feature description set of the defect;

[0047] The defect assessment unit assesses the severity level of each defect according to the preset assessment criteria;

[0048] Each defect is comprehensively evaluated based on defect size, defect shape and defect location factors to determine its severity level, and the evaluation results of each defect are recorded and output.

[0049] Preferably, the comprehensive determination module performs a method of comprehensively evaluating the surface quality of the entire hot-rolled strip product based on the results of defect detection, classification and evaluation as follows:

[0050] Receive the inspection results of all defects from the defect classification and assessment module, including the type, location, size, shape of the defect and the severity level assessment of each defect;

[0051] Clean the collected data and convert the defect information into quantifiable indicators;

[0052] Define the surface quality assessment standards for hot rolled strip products in the quality assessment standard unit;

[0053] Based on the collected data and the preset evaluation criteria, a comprehensive evaluation algorithm is applied to calculate the overall surface quality score of the hot-rolled strip products;

[0054] Output the evaluation results in the form of a report;

[0055] Among them, the calculation method of the overall surface quality score is:

[0056]

[0057] Where Q represents the overall surface quality score, C brepresents the severity score C of the bth defect, A represents the total number of defects detected, and W b represents the weight w of the bth defect, E represents the maximum number of defects allowed, and F max Indicates the maximum total defect area allowed.

[0058] Compared with the prior art, the present invention has the following beneficial effects:

[0059] The automatic determination system proposed in the present invention realizes the automated, efficient and precise determination of the surface quality of hot-rolled plate and strip products by integrating multiple modules including image acquisition, image preprocessing, defect detection, defect classification and evaluation, and comprehensive determination. The system can acquire and process image data of the hot-rolled plate and strip surface in real time, automatically identify and evaluate defects, greatly shorten the detection time, improve the detection efficiency, avoid misjudgment caused by subjectivity and fatigue in manual visual inspection, and improve the accuracy and reliability of detection. Through the comprehensive evaluation algorithm, the system can calculate the overall surface quality score of the hot-rolled plate and strip products and output it in the form of a report, making the evaluation result more intuitive and easy to understand, which is of great significance for improving the production quality and market competitiveness of hot-rolled plate and strip products. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 This is the structural framework diagram of the automatic determination system for the surface quality of hot-rolled strip products;

[0061] Figure 2 This is the framework diagram of the image acquisition module;

[0062] Figure 3 This is the framework diagram of the image preprocessing module;

[0063] Figure 4 This is the framework diagram of the defect detection module;

[0064] Figure 5 This is the framework diagram of the defect classification and assessment module;

[0065] Figure 6 This is the framework diagram of the comprehensive judgment module. DETAILED DESCRIPTION

[0066] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0067] Reference Figure 1 As shown, an automatic determination system for the surface quality of hot-rolled strip products, the automatic determination system comprises:

[0068] Image acquisition module, image preprocessing module, defect detection module, defect classification and evaluation module and comprehensive judgment module;

[0069] As the starting point of the entire system, the image acquisition module is responsible for obtaining clear image data of the surface of hot-rolled strip products in real time. This module uses a high-resolution industrial camera, combined with light sources and lenses, to ensure that high-quality, blur-free, and shadow-free images can be captured in various production environments, providing a solid foundation for subsequent processing.

[0070] The image preprocessing module is electrically connected to the image acquisition module. It implements denoising of the acquired images through algorithms, effectively eliminating noise points in the images and improving the purity of the images. At the same time, it also uses contrast enhancement technology to make the details in the images clearer, facilitating subsequent defect detection. Correction of geometric distortion is also an important part of the preprocessing process, which ensures that the shape of the objects in the image is consistent with the actual shape and avoids misjudgment caused by lens distortion.

[0071] The defect detection module is closely connected with the image preprocessing module. Its task is to automatically identify various defects on the surface of hot-rolled strips using advanced image processing algorithms and pattern recognition technology. This module can accurately distinguish defect areas from normal surfaces. Whether the defects are small or significant, they can be quickly captured and marked, greatly improving the accuracy and efficiency of detection.

[0072] The defect classification and evaluation module is electrically connected to the defect detection module, and is used to classify the detected defects and evaluate their severity levels according to preset standards, including the size, shape, location of the defects and the degree of impact on product quality;

[0073] The comprehensive determination module is electrically connected to the defect classification and evaluation module, and the comprehensive determination module is used to comprehensively evaluate the surface quality of the entire hot-rolled strip product based on the results of defect detection, classification and evaluation.

[0074] Reference Figure 2 As shown, the image acquisition module includes:

[0075] The image capture unit, as the source of image acquisition, is responsible for capturing the image data of the hot-rolled strip surface. This unit is equipped with a high-precision, high-sensitivity image sensor that can accurately record the texture and color of the hot-rolled strip surface in real time. At the same time, the image capture unit also has the ability to respond quickly, ensuring that images can be captured stably and continuously even when the production line is running at high speed, providing sufficient data support for subsequent processing;

[0076] The light source control unit is the key to providing ideal light source conditions for image capture. This unit can flexibly adjust the intensity, color and direction of the light source to meet the shooting requirements of different production environments and hot-rolled strips of different materials. The uniformity and stability of the light source are crucial to improving image quality. Therefore, the light source control unit is fully considered in the design to ensure that the light source conditions are always kept in the best state during the shooting process, so as to obtain clear and bright images.

[0077] The image stabilization unit is designed for the characteristics of hot-rolled strip production lines, which are highly mobile and vibrate frequently. During the high-speed operation of the production line, vibration often interferes with image acquisition, resulting in blurred and distorted images. The image stabilization unit uses advanced image stabilization technology to effectively reduce image blur caused by vibration, ensuring that the captured images are always clear and stable.

[0078] The data transmission unit is used to transmit the captured image data to the image acquisition module.

[0079] Reference Figure 3 As shown, the image preprocessing module includes:

[0080] As the first step of image preprocessing, the denoising unit focuses on deep denoising of the collected images. This unit uses denoising algorithms to accurately identify and remove various types of noise in the image, thereby significantly improving the purity and clarity of the image and laying a solid foundation for subsequent processing;

[0081] The contrast enhancement unit is used to enhance the contrast of the image, making the details in the image more vivid and easier to identify. The unit intelligently analyzes the grayscale distribution of the image and uses a contrast enhancement algorithm to adjust the brightness and contrast of the image to ensure that the key information in the image is highlighted, providing strong support for subsequent defect detection;

[0082] The geometric distortion correction unit specifically handles the geometric distortion problems in the image. These geometric distortions are caused by the perspective effect and lens distortion of the shooting equipment, which cause the shapes of objects in the image to become distorted or deformed. The geometric distortion correction unit can accurately identify and correct these distortions to ensure that the shapes of objects in the image are consistent with the actual ones.

[0083] Reference Figure 4 As shown, the defect detection module includes:

[0084] The image analysis unit, as the data receiving and processing center of the defect detection module, is responsible for receiving high-quality image data after preprocessing, conducting in-depth analysis of the image, and extracting key feature information from the image. These feature information include but are not limited to the edge contour, texture pattern, and color distribution of the image, which are crucial for subsequent defect identification.

[0085] The defect recognition unit identifies defective areas using recognition technology based on the feature information extracted by the image analysis unit. The unit has a built-in defect recognition model, which is built through a large amount of training data and machine learning algorithms. It can accurately identify defective areas in images. During the recognition process, the defect recognition unit will compare the feature information with the preset defect feature library, and distinguish the defects in the image from the normal surface by calculating the similarity and matching degree, providing an accurate basis for subsequent classification and evaluation.

[0086] Reference Figure 5 As shown, the defect classification and assessment module includes:

[0087] As the primary component of the module, the core task of the defect classification unit is to scientifically and reasonably classify the detected defects according to the predetermined classification standards. This unit has a built-in complete defect classification system, covering various defects on the surface of hot-rolled strip products. During the classification process, the defect classification unit will use image processing and pattern recognition to conduct a detailed analysis of the shape, size and location of the defects, thereby ensuring that each defect can be accurately classified;

[0088] Based on the classification, the defect assessment unit further conducts an objective and fair assessment of the severity level of each defect. Based on the preset assessment criteria, this unit comprehensively considers the type, quantity, distribution and impact of the defect on product quality, and assigns an accurate severity level to each defect through complex algorithms and model calculations. This level not only reflects the actual situation of the defect, but also provides an important basis for subsequent product quality judgment;

[0089] The result summary and output unit is responsible for summarizing the results of classification and evaluation and outputting them in the form of a report.

[0090] Reference Figure 6 As shown, the comprehensive determination module includes:

[0091] Result summary unit: The result summary unit is used to collect all defect detection results, classification information and severity level evaluation from the defect classification and evaluation module and summarize them into a comprehensive data set;

[0092] The quality assessment standard unit is used to store and manage preset standards or rules for surface quality assessment;

[0093] A comprehensive evaluation algorithm unit, which is used to quantitatively evaluate the surface quality of the entire hot-rolled plate and strip product through an algorithm based on the data set provided by the result summary unit and the preset standards in the quality evaluation standard unit;

[0094] The comprehensive judgment module integrates defect detection results, classification and severity level information, and uses advanced algorithms to quantitatively evaluate the surface quality of hot-rolled plate and strip products based on preset quality assessment standards.

[0095] The method by which the defect detection module automatically identifies surface defects of hot-rolled strips is as follows:

[0096] As the starting point of the defect detection process, the image analysis unit is responsible for receiving the image data that has been optimized and processed from the image preprocessing module. These image data have been processed by noise reduction, contrast enhancement and geometric distortion correction to ensure the high quality and accuracy of the image, providing a solid foundation for subsequent feature extraction and defect identification. The image analysis unit will extract features from the received images. In this step, the grayscale co-occurrence matrix, an advanced image processing technology, is used to conduct an in-depth analysis of the texture features of the surface of the hot-rolled strip to identify abnormal textures. The grayscale co-occurrence matrix can reflect the spatial relationship and grayscale relationship between pixels in the image, and has a significant effect on identifying surface defects such as cracks and inclusions.

[0097] After feature extraction is completed, it is necessary to select the most valuable feature sets for defect identification from a large number of features. These feature sets can accurately reflect the essential characteristics of surface defects of hot-rolled strips and provide strong support for subsequent pattern recognition. At the same time, in order to eliminate the dimensional differences and numerical range differences between different features, it is also necessary to normalize the feature sets to ensure that the weights of each feature in the pattern recognition process are equal;

[0098] Subsequently, a defect detection model is built based on CNN for pattern recognition. CNN is a deep learning algorithm with feature extraction and pattern recognition capabilities. During the model building process, the model is trained using image data of known defects and normal surfaces so that it can learn the feature differences between defects and normal surfaces. The trained model will serve as the core tool for defect detection and is used for pattern recognition of images to be detected.

[0099] In the process of pattern recognition, the image features to be detected are input into the trained CNN model. The model will judge the image area based on the learned feature differences and output the result of whether the image area contains defects. This result is not only accurate and reliable, but also efficient, and can meet the real-time requirements of automatic determination of the surface quality of hot-rolled strip products.

[0100] The defect areas identified by the model are marked on the original image. This step helps production personnel to intuitively understand the location and shape of the defects, providing strong support for subsequent product quality improvement and defect repair. At the same time, the marked defect areas can also be used as part of the training data to continuously optimize and improve the defect detection model and improve its recognition accuracy and generalization ability.

[0101] The calculation method for identifying abnormal conditions of surface texture using gray-level co-occurrence matrix is:

[0102] Assume that the size of image I is M×N and the gray level is G, then the gray level co-occurrence matrix P is a G×G matrix, where P(i,j) represents the number of occurrences of all pixels with gray level i and gray level j in the image in a specific direction and distance d;

[0103] Among them, the calculation formula of P(i,j) is:

[0104]

[0105] Where δ represents the Kronecker function.

[0106] The defect classification and assessment module classifies the detected defects and assesses their severity levels according to the preset standards as follows:

[0107] The feature data of the identified defect area received from the defect detection module is further extracted to obtain the specific features of the defect and output a detailed feature description set of the defect;

[0108] The defect assessment unit assesses the severity level of each defect according to the preset assessment criteria;

[0109] Each defect is comprehensively evaluated based on defect size, defect shape and defect location factors to determine its severity level, and the evaluation results of each defect are recorded and output.

[0110] The comprehensive judgment module conducts a comprehensive evaluation of the surface quality of the entire hot-rolled strip product based on the results of defect detection, classification and evaluation:

[0111] Receive the inspection results of all defects from the defect classification and assessment module, including the type, location, size, shape of the defect and the severity level assessment of each defect;

[0112] Clean the collected data and convert the defect information into quantifiable indicators;

[0113] Define the surface quality assessment standards for hot rolled strip products in the quality assessment standard unit;

[0114] Based on the collected data and the preset evaluation criteria, a comprehensive evaluation algorithm is applied to calculate the overall surface quality score of the hot-rolled strip products;

[0115] Output the evaluation results in the form of a report;

[0116] Among them, the calculation method of the overall surface quality score is:

[0117]

[0118] Where Q represents the overall surface quality score, C b represents the severity score C of the bth defect, A represents the total number of defects detected, and W b represents the weight w of the bth defect, E represents the maximum number of defects allowed, and F max Indicates the maximum total defect area allowed.

[0119] The use process of the present invention is as follows: the image acquisition module acquires clear image data of the surface of the hot-rolled plate and strip product in real time; the image preprocessing module performs preliminary optimization processing on the acquired image by denoising, enhancing contrast and correcting geometric distortion; the defect detection module automatically identifies defects on the surface of the hot-rolled plate and strip, and distinguishes defects from normal surfaces through image processing algorithms and pattern recognition technology; the defect classification and evaluation module classifies the detected defects and evaluates their severity levels according to preset standards, including the size, shape, location of the defects and the degree of impact on product quality; the comprehensive judgment module comprehensively evaluates the surface quality of the entire hot-rolled plate and strip product based on the results of defect detection, classification and evaluation, and outputs an evaluation report, and outputs the evaluation results in the form of a report for use in production management and quality control.

[0120] In summary, the advantages of the present invention are: by integrating image acquisition, preprocessing, defect detection, classification and evaluation, and comprehensive judgment modules, automated, efficient, and precise detection is achieved. The system significantly improves detection efficiency, reduces human errors, comprehensively evaluates surface quality and quantifies evaluation results. At the same time, it has strong adaptability and can be adjusted according to different standards and needs. This innovative system is of great significance to improving the production quality and market competitiveness of hot-rolled plate and strip products.

[0121] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. An automatic determination system for the surface quality of hot-rolled strip products, characterized in that: The automatic determination system includes: Image acquisition module, image preprocessing module, defect detection module, defect classification and evaluation module and comprehensive judgment module; The image acquisition module is used to obtain clear image data of the surface of hot-rolled strip products in real time; The image preprocessing module is electrically connected to the image acquisition module, and the image preprocessing module is used to perform preliminary optimization processing on the acquired image, and the optimization processing includes: denoising, contrast enhancement and geometric distortion correction; The defect detection module is electrically connected to the image preprocessing module, and the defect detection module is used to automatically identify defects on the surface of the hot-rolled strip and distinguish defects from normal surfaces through image processing algorithms and pattern recognition technology; The defect classification and evaluation module is electrically connected to the defect detection module, and is used to classify the detected defects and evaluate their severity levels according to preset standards, including the size, shape, location of the defects and the degree of impact on product quality; The comprehensive determination module is electrically connected to the defect classification and evaluation module, and the comprehensive determination module is used to comprehensively evaluate the surface quality of the entire hot-rolled strip product based on the results of defect detection, classification and evaluation.

2. The automatic determination system for the surface quality of hot-rolled strip products according to claim 1 is characterized in that: The image acquisition module includes: An image capturing unit, the image capturing unit is used to capture image data of the surface of the hot-rolled strip; A light source control unit, which is used to adjust the intensity, color and direction of the light source to provide light source conditions for capturing image data; Image stabilization unit, which is used to stabilize the image acquisition process on a moving hot strip production line to reduce blur caused by vibration; The data transmission unit is used to transmit the captured image data to the image acquisition module.

3. The automatic determination system for the surface quality of hot-rolled strip products according to claim 1 is characterized in that: The image preprocessing module includes: A denoising processing unit, which is responsible for denoising the collected image; Contrast enhancement unit, the contrast enhancement unit is used to enhance the contrast of the image and make the details in the image clear; The geometric distortion correction unit is used to correct the geometric distortion in the image, and the geometric distortion includes perspective distortion and lens distortion.

4. The automatic determination system for the surface quality of hot-rolled strip products according to claim 1 is characterized in that: The defect detection module includes: An image analysis unit, which is responsible for receiving pre-processed image data and applying an image processing algorithm to extract feature information in the image; The defect recognition unit is used to recognize defect areas in the image based on the feature information extracted by the image analysis unit by using pattern recognition technology to distinguish defects from normal surfaces.

5. The automatic determination system for the surface quality of hot-rolled strip products according to claim 1 is characterized in that: The defect classification and assessment module includes: A defect classification unit, the defect classification unit is used to classify the detected defects according to a predetermined classification standard; Defect assessment unit, which is used to assess the severity of each defect according to preset assessment criteria after classification; The result summary and output unit is responsible for summarizing the results of classification and evaluation and outputting them in the form of a report.

6. The automatic determination system for the surface quality of hot-rolled strip products according to claim 1, characterized in that: The comprehensive judgment module includes: Result summary unit: The result summary unit is used to collect all defect detection results, classification information and severity level evaluation from the defect classification and evaluation module and summarize them into a comprehensive data set; The quality assessment standard unit is used to store and manage preset standards or rules for surface quality assessment; The comprehensive evaluation algorithm unit is used to quantitatively evaluate the surface quality of the entire hot-rolled plate and strip product through an algorithm based on the data set provided by the result summary unit and the preset standards in the quality evaluation standard unit.

7. The automatic determination system for the surface quality of hot-rolled strip products according to claim 1 is characterized in that: The method by which the defect detection module automatically identifies surface defects of hot-rolled strips is as follows: The image analysis unit receives the optimized image data from the image preprocessing module; Extract features from images and use gray-level co-occurrence matrix to identify abnormalities in surface texture; Select the most valuable feature set for defect identification from the extracted features and normalize the feature set; Build a defect detection model based on CNN for pattern recognition, and train the model using image data of known defects and normal surfaces; The image features to be detected are input into the trained model, and the model outputs the result of whether the image area contains defects; The defect area identified by the model is marked on the original image.

8. The automatic determination system for the surface quality of hot-rolled strip products according to claim 7, characterized in that: The calculation method for identifying abnormal conditions of surface texture using gray-level co-occurrence matrix is: Assume that the size of image I is M×N and the gray level is G, then the gray level co-occurrence matrix P is a G×G matrix, where P(i,j) represents the number of occurrences of all pixels with gray level i and gray level j in the image in a specific direction and distance d; Among them, the calculation formula of P(i,j) is: Where δ represents the Kronecker function.

9. The automatic determination system for the surface quality of hot-rolled strip products according to claim 1, characterized in that: The defect classification and assessment module classifies the detected defects and assesses their severity levels according to the preset standards as follows: The feature data of the identified defect area received from the defect detection module is further extracted to obtain the specific features of the defect and output a detailed feature description set of the defect; The defect assessment unit assesses the severity level of each defect according to the preset assessment criteria; Each defect is comprehensively evaluated based on defect size, defect shape and defect location factors to determine its severity level, and the evaluation results of each defect are recorded and output.

10. The automatic determination system for the surface quality of hot-rolled strip products according to claim 1, characterized in that: The comprehensive judgment module conducts a comprehensive evaluation of the surface quality of the entire hot-rolled strip product based on the results of defect detection, classification and evaluation: Receive the inspection results of all defects from the defect classification and assessment module, including the type, location, size, shape of the defect and the severity level assessment of each defect; Clean the collected data and convert the defect information into quantifiable indicators; Define the surface quality assessment standards for hot rolled strip products in the quality assessment standard unit; Based on the collected data and the preset evaluation criteria, a comprehensive evaluation algorithm is applied to calculate the overall surface quality score of the hot-rolled strip products; Output the evaluation results in the form of a report; Among them, the calculation method of the overall surface quality score is: Where Q represents the overall surface quality score, C b represents the severity score C of the bth defect, A represents the total number of defects detected, and W b represents the weight w of the bth defect, E represents the maximum number of defects allowed, and F max Indicates the maximum total defect area allowed.

Images