Online identification and detection method and system for defects of cold-rolled strip steel
By acquiring and preprocessing the cold-rolled strip images, using deep learning models to identify defects, the problems of inefficient and insufficient accuracy of manual visual inspection are solved, and efficient online identification and detection of cold-rolled strip defects are achieved.
Patent Information
- Application Number
- CN202510202787.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-24
- Publication Date
- 2025-06-27
AI Technical Summary
In the prior art, defect detection of cold-rolled strip steel mainly relies on manual visual inspection, resulting in low detection efficiency, frequent missed detection and missed detection, affecting the accuracy and reliability of the detection results.
A cold-rolled strip defect online identification and detection method is adopted to identify and store defect information by collecting strip images, pre-processing and segmenting them.
It realizes efficient screening and identification of defect characteristics of strip steel, improves the accuracy and reliability of inspection, reduces missed inspections and missed inspections, and improves product quality and production efficiency.
Smart Images
Figure CN120219908A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of image recognition, and relates to an on-line recognition and detection method and system for cold-rolled strip steel defects. Background Art
[0002] Cold-rolled strip steel has been widely used in many high-end manufacturing fields such as automobile manufacturing, household appliance production, and precision instrument manufacturing due to its high precision, excellent surface quality, and good mechanical properties. However, during the cold-rolling production process, due to various complex factors such as the accuracy limitation of rolling equipment, the fluctuation of lubrication conditions, and the presence of impurities in raw materials, various defects are easily generated on the strip steel surface, such as scratches, holes, inclusions, and pitting. Traditionally, the defect detection of cold-rolled strip steel mainly relies on manual visual inspection. However, this method has significant defects: the detection efficiency is low, and due to long-term high-intensity visual observation, the inspectors are prone to visual fatigue, resulting in frequent missed detections and misdetections, seriously affecting the accuracy and reliability of the detection results. Therefore, there is an urgent need for a method that can quickly identify the defects of cold-rolled strip steel. Summary of the Invention
[0003] The purpose of the present invention is to solve the problem that the defect detection of cold-rolled strip steel in the prior art mainly relies on manual visual inspection, resulting in frequent missed detections and misdetections, and to provide an on-line recognition and detection method and system for cold-rolled strip steel defects.
[0004] To achieve the above object, the present invention adopts the following technical solutions:
[0005] An on-line recognition and detection method for cold-rolled strip steel defects, comprising:
[0006] Collecting strip steel images being conveyed on a conveyor belt, and screening the collected strip steel images to obtain strip steel images containing defect features;
[0007] Preprocessing the screened strip steel images, and dividing the preprocessed defect images into pictures of a preset size;
[0008] Inputting the divided images into a pre-trained deep learning model to identify and store defect information.
[0009] A further improvement of the present invention lies in:
[0010] Further, preprocessing the screened strip steel images specifically includes: performing filtering and noise reduction and image enhancement processing on the images to reduce noise interference in the images and improve the clarity and contrast of the images.
[0011] Further, the pre-trained deep learning model is specifically:
[0012] Adopt strip steel pictures containing defect features and preprocess the collected pictures;
[0013] Divide the preprocessed pictures to obtain a training set and a test set;
[0014] Set the hyperparameters of the deep learning model and initialize the model parameters;
[0015] Train the deep learning model based on the training set, and verify and evaluate the output results through the test set until the deep learning model reaches the predetermined number of iterations or the loss function is less than the preset value, and the deep learning module stops training.
[0016] Furthermore, training the deep learning model based on the training set is specifically as follows:
[0017] Randomly select a batch of samples from the training set and input them into the deep learning model; perform forward calculation through each layer of the model, keeping the weights and biases of the model unchanged, and finally obtain the prediction results;
[0018] Compare the prediction results with the true labels, calculate the loss value using the loss function; according to the loss value, calculate the gradient of each parameter in the model through the backpropagation algorithm; use the optimization algorithm to update the parameters of the model according to the gradient;
[0019] Repeat the above steps until the predetermined number of iterations is reached or the loss function is less than the preset value, and the deep learning module stops training.
[0020] Furthermore, using the optimization algorithm to update the parameters of the model according to the gradient is specifically as follows: Update the model parameters based on the gradient descent method as:
[0021]
[0022] where θ is the model parameter, is the gradient, and α is the learning rate.
[0023] Furthermore, verifying and evaluating the output results through the test set is specifically as follows: Calculate various evaluation metrics according to the output of the model and the true labels of the test set; The evaluation metrics include: accuracy, the proportion of the number of correctly predicted samples to the total number of samples; precision, the proportion of truly positive samples among all samples predicted as positive; recall, the proportion of samples correctly predicted as positive among all truly positive samples; F1 score, the harmonic mean of precision and recall, comprehensively evaluating the performance of the model.
[0024] Further, the deep learning model is a convolutional neural network model, and the convolutional neural network model includes: three convolutional layers, three pooling layers and four fully connected layers, specifically: Convolutional layer 1 → Pooling layer 1 → Convolutional layer 2 → Pooling layer 2 → Convolutional layer 3 → Pooling layer 3 → Fully connected layer 1 → Fully connected layer 2 → Fully connected layer 3 → Fully connected layer 4.
[0025] An on-line identification and detection system for cold-rolled strip steel defects, comprising:
[0026] A screening module, which collects strip steel images being conveyed on a conveyor belt and screens the collected strip steel images to obtain strip steel images containing defect features;
[0027] A preprocessing module, which preprocesses the screened strip steel images and divides the preprocessed defect images into pictures of a preset size;
[0028] An identification module, which inputs the segmented images into a pre-trained deep learning model to identify and store defect information.
[0029] Compared with the prior art, the present invention has the following beneficial effects:
[0030] The present invention can efficiently screen and identify the defect features of strip steel on a conveyor belt, ensure the image quality through preprocessing and segmentation, and improve the recognition accuracy of the deep learning model; the detection speed is accelerated through the deep learning model; in addition, the stored defect information provides data support for quality traceability and improvement, which helps to improve product quality and production efficiency. The present invention realizes the intelligentization and high efficiency of strip steel defect detection. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.
[0032] Figure 1 It is a schematic diagram of the cold-rolled strip steel image acquisition device of the present invention;
[0033] Figure 2 It is a schematic flow chart of the on-line identification and detection system for cold-rolled strip steel defects of the present invention;
[0034] Figure 3 It is a schematic structural diagram of the on-line identification and detection system for cold-rolled strip steel defects of the present invention.
[0035] Among them, 1 - light source board; 2 - strip steel; 3 - conveyor roller; 4 - power protection rod; 5 - support rod; 6 - industrial high - speed camera. Specific embodiments
[0036] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some but not all of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations.
[0037] Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.
[0038] It should be noted that: Similar reference numerals and letters denote similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.
[0039] In the description of the embodiments of the present invention, it should be noted that if terms such as "upper", "lower", "horizontal", "inner", etc. are used to indicate the orientation or positional relationship, it is based on the orientation or positional relationship shown in the drawings or the orientation or positional relationship when the product of the invention is normally placed. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation of the present invention. In addition, terms such as "first", "second", etc. are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.
[0040] In addition, if the term "horizontal" appears, it does not mean that the component is required to be absolutely horizontal, but it can be slightly inclined. For example, "horizontal" only means that its direction is more horizontal relative to "vertical", and does not mean that the structure must be completely horizontal, but it can be slightly inclined.
[0041] In the description of the embodiments of the present invention, it should also be noted that unless otherwise clearly specified and limited, if the terms "set", "installed", "connected", and "linked" are used, they should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.
[0042] The following further describes the present invention in detail with reference to the accompanying drawings:
[0043] See Figure 1 , the present invention discloses a cold-rolled strip steel image acquisition device, including: a light source plate 1, a strip steel 2, a conveyor roller 3, a power protection rod 4, a support rod 5, and an industrial high-speed camera 6;
[0044] The industrial high-speed camera 6 is arranged on the power protection rod 4. Both ends of the light source plate 1 are provided with support rods 5. The strip steel 2 is placed on the conveyor roller 3. The conveyor roller 3 is externally connected to a motor. The conveyor roller 3 is arranged on a frame. The conveyor roller 3 is arranged below the space of the industrial high-speed camera 6. The support rods 5 are arranged on both sides of the frame. The light source plate 1 is arranged above the space of the conveyor roller 3. The industrial high-speed camera 6 is externally connected to a computer.
[0045] The working method of the cold-rolled strip steel image acquisition device is as follows: start the motor, the conveyor roller 3 moves under the drive of the motor, and then drives the strip steel 2 to move. The industrial high-speed camera 6 takes pictures of the strip steel 2 and sends them to the computer for processing to identify and store the defect information of the strip steel 2.
[0046] See Figure 2 , the present invention discloses an on-line identification and detection method for cold-rolled strip steel defects, including:
[0047] S101, collect the strip steel images being conveyed on the conveyor belt, and screen the collected strip steel images to obtain the strip steel images containing defect features;
[0048] S102, preprocess the screened strip steel images, and segment the preprocessed defect images into pictures of a preset size;
[0049] Perform filtering and noise reduction and image enhancement processing on the images to reduce noise interference in the images and improve the clarity and contrast of the images.
[0050] S103, input the segmented images into a pre-trained deep learning model to identify and store the defect information.
[0051] The pre-trained deep learning model is specifically:
[0052] S103.1, adopt strip steel pictures containing defect features, and preprocess the collected pictures;
[0053] S103.2, divide the preprocessed pictures to obtain a training set and a test set;
[0054] S103.3, set the hyperparameters of the deep learning model and initialize the model parameters;
[0055] S103.4, train the deep learning model based on the training set, and verify and evaluate the output results through the test set until the deep learning model reaches a predetermined number of iterations or the loss function is less than a preset value, and the deep learning module stops training.
[0056] Training the deep learning model based on the training set specifically includes:
[0057] Randomly select a batch of samples from the training set and input them into the deep learning model; perform forward calculation through each layer of the model, keeping the weights and biases of the model unchanged, and finally obtain a prediction result;
[0058] Compare the prediction result with the true label, calculate the loss value using the loss function; calculate the gradient of each parameter in the model according to the loss value through the backpropagation algorithm; update the parameters of the model according to the gradient using the optimization algorithm;
[0059] Repeat the above steps until a predetermined number of iterations is reached or the loss function is less than a preset value, and the deep learning module stops training.
[0060] Updating the parameters of the model according to the gradient using the optimization algorithm specifically includes: updating the model parameters based on the gradient descent method as:
[0061]
[0062] where θ is the model parameter, is the gradient, and α is the learning rate.
[0063] Verifying and evaluating the output results through the test set specifically includes: calculating various evaluation metrics according to the output of the model and the true labels of the test set; the evaluation metrics include: accuracy rate, the proportion of the number of correctly predicted samples to the total number of samples; precision rate, the proportion of the truly positive samples among all the samples predicted as positive; recall rate, the proportion of the samples correctly predicted as positive among all the truly positive samples; F1 score, the harmonic mean of the precision rate and the recall rate, comprehensively evaluating the performance of the model.
[0064] The deep learning model is a convolutional neural network model, and the convolutional neural network model includes: three convolutional layers, three pooling layers and four fully connected layers, specifically: Convolutional layer 1 → Pooling layer 1 → Convolutional layer 2 → Pooling layer 2 → Convolutional layer 3 → Pooling layer 3 → Fully connected layer 1 → Fully connected layer 2 → Fully connected layer 3 → Fully connected layer 4.
[0065] See Figure 3 , the present invention discloses an on-line recognition and detection system for cold-rolled strip steel defects, including:
[0066] A screening module, which collects strip steel images being conveyed on a conveyor belt and screens the collected strip steel images to obtain strip steel images containing defect features;
[0067] A preprocessing module, which preprocesses the screened strip steel images and divides the preprocessed defect images into pictures of a preset size;
[0068] An identification module, which inputs the segmented images into a pre-trained deep learning model to identify and store defect information.
[0069] The terminal device provided by the embodiment of the present invention. The terminal device of this embodiment includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps in the above-mentioned various method embodiments are implemented. Alternatively, when the processor executes the computer program, the functions of each module / unit in the above-mentioned various device embodiments are implemented.
[0070] The computer program can be divided into one or more modules / units, and the one or more modules / units are stored in the memory and executed by the processor to complete the present invention.
[0071] The terminal device can be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The terminal device may include, but is not limited to, a processor and a memory.
[0072] The processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
[0073] The memory can be used to store the computer program and / or module. By running or executing the computer program and / or module stored in the memory, and invoking the data stored in the memory, the processor implements various functions of the terminal device.
[0074] If the modules / units integrated in the terminal device are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such understanding, to implement all or part of the processes in the above-described embodiment methods of the present invention, it can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, it can implement the steps of the above various method embodiments. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.
[0075] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.
Claims
1. A method for online identification and detection of cold-rolled strip defects, characterized in that: include: Collecting images of the steel strip being transported on the conveyor belt, and screening the collected steel strip images to obtain steel strip images containing defect features; Preprocessing the screened strip steel images, and segmenting the preprocessed defective images into images of preset sizes; The segmented images are input into a pre-trained deep learning model to identify and store defect information.
2. The method for online identification and detection of cold-rolled strip defects according to claim 1, characterized in that: The preprocessing of the screened strip steel images is specifically: filtering and denoising the images and performing image enhancement processing to reduce noise interference in the images and improve the clarity and contrast of the images.
3. The method for online identification and detection of cold-rolled strip defects according to claim 2, characterized in that: The pre-trained deep learning model is specifically: Using strip steel images containing defect features and preprocessing the collected images; Divide the preprocessed images into training sets and test sets; Set the hyperparameters of the deep learning model and initialize the model parameters; The deep learning model is trained based on the training set, and the output results are verified and evaluated through the test set until the deep learning model reaches a predetermined number of iterations or the loss function is less than a preset value, and the deep learning module stops training.
4. The method for online identification and detection of cold-rolled strip defects according to claim 3, characterized in that: The deep learning model is trained based on the training set, specifically: A batch of samples are randomly selected from the training set and input into the deep learning model; forward calculations are performed through each layer of the model, and the weights and biases of the model remain unchanged, and finally the prediction results are obtained; Compare the predicted results with the true labels and calculate the loss value using the loss function; Based on the loss value, the gradient of each parameter in the model is calculated through the back-propagation algorithm; the optimization algorithm is used to update the parameters of the model according to the gradient; Repeat the above steps until the predetermined number of iterations is reached or the loss function is less than the preset value, and the deep learning module stops training.
5. The method for online identification and detection of cold-rolled strip defects according to claim 4, characterized in that: The optimization algorithm is used to update the parameters of the model according to the gradient, specifically: the model parameters are updated based on the gradient descent method as follows: Among them, θ is the model parameter, is the gradient and α is the learning rate.
6. The method for online identification and detection of cold-rolled strip defects according to claim 5, characterized in that: The output results are verified and evaluated through the test set, specifically: various evaluation indicators are calculated according to the output of the model and the true label of the test set; the evaluation indicators include: accuracy, the ratio of correctly predicted samples to the total number of samples; precision, the ratio of all samples predicted to be positive that are truly positive; recall, the ratio of all samples truly positive that are correctly predicted to be positive; F1 score, the harmonic mean of precision and recall, comprehensively evaluates the performance of the model.
7. The method for online identification and detection of cold-rolled strip defects according to claim 6, characterized in that: The deep learning model is a convolutional neural network model, which includes: three convolutional layers, three pooling layers and four fully connected layers, specifically: convolutional layer 1→pooling layer 1→convolutional layer 2→pooling layer 2→convolutional layer 3→pooling layer 3→fully connected layer 1→fully connected layer 2→fully connected layer 3→fully connected layer 4.
8. An online recognition and detection system for cold-rolled strip defects, characterized in that: include: A screening module, wherein the screening module collects images of the steel strip being conveyed on the conveyor belt, and screens the collected steel strip images to obtain steel strip images containing defect features; A preprocessing module, wherein the preprocessing module preprocesses the screened strip steel image and divides the preprocessed defect image into images of a preset size; The recognition module inputs the segmented image into a pre-trained deep learning model to recognize and store defect information.