Detection and early warning system for optical glass processing

Abstract

The invention discloses a detection and early warning system for optical glass processing, and relates to the technical field of optical glass processing, the system comprises the following components: an image capture module, the image capture module is used as the starting part of the whole detection and early warning system, the performance of the image capture module directly determines the accuracy of subsequent image processing and flaw recognition, and the image capture module is used as the detection and early warning module; through the intelligent learning and optimization module and continuous data learning and algorithm optimization, the system can continuously adapt to new detection requirements, the recognition accuracy and stability of the system are improved, long-term effectiveness and advancement are kept, the recognition accuracy of the system can be gradually improved, the risks of missing detection and misjudgment are reduced, and the detection efficiency is improved. The system can meet the processing requirements of optical glass of different types and complexities, has wide applicability and flexibility, improves the automation, intelligence and accuracy of the detection and early warning system for optical glass processing, and brings remarkable value and benefit to the optical glass processing industry.

Description

Technical Field

[0001] The invention relates to the technical field of optical glass processing, in particular to a detection and early warning system for optical glass processing. Background Art

[0002] With the continuous advancement of science and technology, in the optical glass processing industry, defect detection of glass products is a vital task. The defects of glass products will not only affect the appearance quality of the product, but may also have a serious impact on its performance, which in turn affects the market acceptance and user experience of the product. Ensuring efficient, accurate and reliable defect detection in the optical glass processing process is of great significance to improving product quality, reducing production costs and enhancing market competitiveness.

[0003] Therefore, how to build a highly automated and intelligent optical glass processing detection and early warning system that can record, analyze and identify in real time is an urgent market demand and has broad application prospects. The automated and intelligent optical glass processing detection and early warning system should have the following characteristics: it can automatically capture images, accurately identify defects, and provide real-time feedback results. It can also continuously improve recognition performance through self-learning and optimization, and realize efficient, accurate and reliable defect detection in the optical glass processing process.

[0004] However, the current traditional optical glass defect detection mainly relies on manual visual inspection. However, this detection method has obvious limitations. First, manual visual inspection is inefficient and cannot adapt to large-scale, high-speed production needs. Second, manual inspection is easily affected by fatigue and visual error factors, which makes it difficult to ensure the accuracy and stability of the detection results. In addition, manual inspection requires a lot of training and experience accumulation. For novices, the learning cost is high and it is difficult to reach the detection level of skilled workers. Summary of the invention

[0005] The purpose of the present invention is to make up for the shortcomings of existing optical glass defects and provide a highly automated and intelligent detection and early warning system for optical glass processing. The system not only has efficient image capture and processing capabilities, but also incorporates advanced machine learning technology, so that the invented system has high scalability and flexibility. Users can customize and configure the system according to actual needs, such as adjusting the shooting parameters of the image capture module and optimizing the algorithm parameters of the image processing module, thereby realizing automatic detection and early warning of optical glass defects.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a detection and early warning system for optical glass processing, the system comprising the following components:

[0007] Image capture module: As the starting part of the entire detection and early warning system, the performance of the image capture module directly determines the accuracy of subsequent image processing and defect recognition. This module uses a high-resolution camera to ensure that the captured images are clear and delicate, providing high-quality raw data for subsequent defect recognition;

[0008] Image processing module: automatically processes and analyzes the received images, with a built-in optimized defect recognition algorithm to identify defects in optical glass. To further improve the accuracy and efficiency of recognition, the image processing module uses advanced deep learning algorithms to improve image quality and reduce the impact of interference factors on recognition results;

[0009] Result feedback module: Feedback the recognition results to production personnel in real time, including the location, size and type of defects, so that production personnel can respond quickly. It also provides detailed reports and statistical functions to help production personnel understand the distribution and trend of defects, providing a basis for subsequent process improvement and quality control;

[0010] Intelligent learning and optimization module: Optimize the defects in the result feedback through the defect recognition algorithm, improve the recognition accuracy and stability of the system, so that it can adapt to the processing needs of optical glass of different types and complexity, and then use the advanced gradient descent algorithm to improve the stability and convergence speed of the algorithm.

[0011] Furthermore, the specific steps of the image capture module are:

[0012] (1) Camera selection and calibration: According to the characteristics of optical glass and testing requirements, select the appropriate camera model and parameters, calibrate the camera, and ensure that the shooting angle and focal length parameters are set correctly to ensure the accuracy and consistency of image acquisition;

[0013] (2) Automatic positioning and focusing: Through the visual positioning algorithm, the camera can automatically position the optical glass to be inspected. According to the thickness and curvature characteristics of the glass, the camera's focus distance is automatically adjusted to ensure a clear and blur-free image.

[0014] (3) Image acquisition and transmission: After positioning and focusing are completed, the camera begins to acquire images of the optical glass to be inspected. The acquired image data is compressed and encoded, and then transmitted to the image processing module in real time through the data transmission interface Ethernet and USB.

[0015] Furthermore, the specific steps of the image processing module are:

[0016] (1) Image preprocessing: De-noising and contrast enhancement preprocessing operations are performed on the received images to improve image quality and reduce the impact of noise and interference on subsequent recognition;

[0017] (2) Feature extraction: Use deep learning algorithms to extract key features in the image, such as edges, textures, and colors. These features will serve as the basis for subsequent defect identification.

[0018] (3) Defect recognition: Based on the extracted features, the deep learning algorithm is used to identify defects in the image. The algorithm learns the characteristics and patterns of defects by training a large amount of sample data, so that defects in the glass can be accurately identified.

[0019] (4) Defect location and classification: For the identified defects, the algorithm further determines their location, size and type. This information will be used for subsequent result feedback and early warning processing.

[0020] Furthermore, the result feedback module comprises the following specific steps:

[0021] (1) Visual display: The identified defects are marked and displayed on the original image, so that production personnel can intuitively understand the location, size and type of the defects;

[0022] (2) Data reporting: The threshold judgment algorithm for defects generates detailed defect data reports, including the number, distribution, and type of defects. These data reports can be used to monitor and analyze the production process, helping production personnel better understand the product quality status;

[0023] (3) Real-time alarm: When a serious defect is identified or the number of defects exceeds the preset threshold, the system triggers a real-time alarm mechanism, reminding production personnel to handle it in a timely manner through sound and light.

[0024] Furthermore, the intelligent learning and optimization module has the following specific steps:

[0025] (1) Data learning: Utilize a large amount of historical data and real-time data to continuously train and optimize the defect recognition algorithm. Through continuous learning and updating, the recognition accuracy and generalization ability of the algorithm are improved.

[0026] (2) Algorithm optimization: According to the actual application scenarios and requirements, the algorithm is optimized and improved. For different types of optical glass or defects, the algorithm parameters are adjusted and new feature extraction methods are introduced to improve the recognition effect.

[0027] (3) Adaptive adjustment: The system can automatically adjust image processing parameters and recognition thresholds based on real-time feedback data and results using a gradient descent optimization algorithm to adapt to detection requirements under different environments and conditions.

[0028] Furthermore, the image capture module uses the SIFT algorithm in the visual positioning algorithm to detect and match feature points, extracts key points and their descriptors in the image through the Gaussian kernel function algorithm, and matches between different images to determine the relative position between the camera and the glass to be detected. The Gaussian filter formula is as follows: Among them, G(x,y) is the Gaussian kernel function, (x,y) is the pixel coordinate, σ is the standard deviation of the Gaussian kernel, and a multi-scale image is generated by convolving the original image with the Gaussian kernel to capture features of different sizes and thus construct a scale space. Local extreme points are searched in the scale space and used as potential feature points. The pixel gradients in the neighborhood of the key points are calculated, and the main direction of each key point is determined to achieve rotation invariance. Based on the position and main direction of the key points, the gradient information in the neighborhood is calculated and encoded into a high-dimensional vector to form a descriptor with robustness to illumination, viewing angle, and scale. The auxiliary image capture module realizes the precise positioning of the glass to be inspected, providing high-quality image data for subsequent image processing and defect recognition.

[0029] Furthermore, the image processing module uses a deep neural network (CNN) to build a model in the feature extraction process through a deep learning algorithm, automatically extracts useful features from the original data, and realizes a hierarchical representation of the features. This feature extraction and representation learning capability enables the deep learning model to capture the inherent structure and regularity of the data. In classification and regression tasks, the deep learning model uses a convolutional neural network to learn the complex mapping relationship from input to output. Through a large amount of training data, these models can continuously optimize their own parameters and structures to improve the accuracy of prediction and classification. The convolution operation in the convolution layer of the convolutional neural network (CNN) is used to extract local features of the input data. Mathematically, the convolution operation can be expressed as: S(i,j)=(I*K)(i,j)=∑ m ∑ n I(i+m,j+n)K(m,n), where I is the input data, K is the convolution kernel, and K is the output after convolution. The complex mapping relationship from input to output enables high-precision prediction and classification.

[0030] Furthermore, the result feedback module uses a threshold judgment algorithm to determine whether the severity or number of defects exceeds a preset threshold, thereby triggering an alarm mechanism. Assuming that the preset defect number threshold is (T) and the currently detected number of defects is (N), the logic of the threshold judgment can be expressed as: if N>T, then trigger an alarm. After the result feedback module triggers the alarm, the threshold judgment algorithm organizes the defect information into an easy-to-understand form, and then the visualization algorithm converts this information into graphics or icons and displays them on the user interface. The real-time communication algorithm ensures that this information can be transmitted to production personnel in real time so that they can respond promptly based on this information.

[0031] Furthermore, the intelligent learning and optimization module plays a key role in the detection and early warning system for optical glass processing through the defect recognition algorithm. This algorithm realizes automatic detection and recognition of defects in glass products by using a machine learning algorithm. In the defect recognition task, the logistic regression in the machine learning algorithm predicts whether a glass product contains defects by learning image features. The formula of logistic regression is as follows: P(y=1|x)=sigmoid(w*x+b), where P(y=1|x) represents the probability that the output y is 1 (i.e., there is a defect) given the input feature x. The sigmoid function is used to map the output of the linear regression to a probability value between 0 and 1. w and b are parameters of the model, representing weights and biases, respectively. Through complex calculations and analysis, the algorithm can accurately identify defects in optical glass, and can classify and identify defects whether they are surface defects or internal defects, helping production personnel understand the type and severity of defects.

[0032] Furthermore, the intelligent learning and optimization module uses a gradient descent optimization algorithm, and the formula of this algorithm is: Where t represents the model parameters of the tth step, {t+1} represents the updated model parameters, and L represents the gradient of the loss function at t. These data are then used to update and adjust the algorithm online, and iterate and optimize continuously. In order for the algorithm to gradually adapt to new environments and requirements and improve recognition accuracy and generalization capabilities, a gradient descent algorithm is used to improve stability and convergence speed. At the same time, it also has adaptive adjustment capabilities, and can automatically adjust algorithm parameters and computing resources according to the system's operating status and performance requirements to ensure efficient operation of the system.

[0033] Compared with the existing technology, the optical glass processing detection and early warning system has the following beneficial effects:

[0034] 1. The present invention uses the intelligent learning and optimization module through continuous data learning and algorithm optimization. The system can continuously adapt to new detection needs, improve its own recognition accuracy and stability, maintain long-term effectiveness and advancement, and the recognition accuracy of the system will gradually improve, reducing the risks of missed detection and misjudgment. The system can adapt to the processing needs of optical glass of different types and complexities, and has wide applicability and flexibility. It improves the automation, intelligence and accuracy of the detection and early warning system for optical glass processing, bringing significant value and benefits to the optical glass processing industry.

[0035] 2. The present invention adopts a high-resolution camera through an image capture module, which can capture subtle defects on the surface of the optical glass, improve the precision and accuracy of detection, realize automatic operation, reduce manual operation, reduce labor intensity, and avoid omissions or misjudgments caused by human factors.

[0036] Other advantages, objectives and features of the present invention will be set forth in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0038] Figure 1 This is the process operation diagram of the detection and early warning system for optical glass processing.

[0039] Figure 2 This is a flow chart of the detection and early warning system for optical glass processing. DETAILED DESCRIPTION

[0040] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] Example 1

[0042] Detection and early warning system for optical glass processing, the system includes the following components:

[0043] Image capture module: As the starting part of the entire detection and early warning system, the performance of the image capture module directly determines the accuracy of subsequent image processing and defect recognition. This module uses a high-resolution camera to ensure that the captured images are clear and delicate, providing high-quality raw data for subsequent defect recognition;

[0044] Image processing module: It has a built-in optimized defect recognition algorithm to automatically process and analyze the received images to identify defects in optical glass. In order to further improve the accuracy and efficiency of recognition, this module uses advanced deep learning algorithms to improve image quality and reduce the impact of interference factors on recognition results;

[0045] Result feedback module: Feedback the recognition results to production personnel in real time, including the location, size and type of defects, so that production personnel can respond quickly. It also provides detailed reports and statistical functions to help production personnel understand the distribution and trend of defects, providing a basis for subsequent process improvement and quality control;

[0046] Intelligent learning and optimization module: Through continuous data learning and algorithm optimization, the recognition accuracy and stability of the system are improved, so that it can adapt to the processing needs of optical glass of different types and complexities. This module uses an advanced gradient descent algorithm to improve the stability and convergence speed of the algorithm.

[0047] The intelligent learning and optimization module uses the gradient descent optimization algorithm. The formula of this algorithm is: Where t represents the model parameters of the tth step, {t+1} represents the updated model parameters, and L represents the gradient of the loss function at t. These data are used to update and adjust the algorithm online, and iterate and optimize continuously. In order for the algorithm to gradually adapt to new environments and requirements and improve recognition accuracy and generalization capabilities, the gradient descent algorithm is used to improve stability and convergence speed. At the same time, it also has adaptive adjustment capabilities, and can automatically adjust algorithm parameters and computing resources according to the system's operating status and performance requirements to ensure efficient operation of the system.

[0048] Through the intelligent learning and optimization module, data can be continuously collected from the actual detection process, and the data can be analyzed and processed. By learning a large amount of data through the gradient descent algorithm, the module can automatically adjust and optimize the internal algorithm parameters, thereby improving the accuracy and efficiency of defect identification. At the same time, it can also automatically adjust and optimize the detection strategy according to different types of optical glass and processing requirements, making the system more intelligent and adaptive.

[0049] Example 2

[0050] Detection and early warning system for optical glass processing, the system includes the following components:

[0051] Image capture module: As the starting part of the entire detection and early warning system, the performance of the image capture module directly determines the accuracy of subsequent image processing and defect recognition. This module uses a high-resolution camera to ensure that the captured images are clear and delicate, providing high-quality raw data for subsequent defect recognition;

[0052] Image processing module: It has a built-in optimized defect recognition algorithm to automatically process and analyze the received images to identify defects in optical glass. In order to further improve the accuracy and efficiency of recognition, this module uses advanced deep learning algorithms to improve image quality and reduce the impact of interference factors on recognition results;

[0053] Result feedback module: Feedback the recognition results to production personnel in real time, including the location, size and type of defects, so that production personnel can respond quickly. It also provides detailed reports and statistical functions to help production personnel understand the distribution and trend of defects, providing a basis for subsequent process improvement and quality control;

[0054] Intelligent learning and optimization module: Through continuous data learning and algorithm optimization, the recognition accuracy and stability of the system are improved, so that it can adapt to the processing needs of optical glass of different types and complexities. This module uses an advanced gradient descent algorithm to improve the stability and convergence speed of the algorithm.

[0055] The image capture module detects and matches feature points in the visual positioning algorithm, extracts key points and their descriptors in the image, and matches them between different images to determine the relative position between the camera and the glass to be detected. The formula of the visual positioning algorithm is as follows: Among them, G(x,y) is the Gaussian kernel function, (x,y) is the pixel coordinate, σ is the standard deviation of the Gaussian kernel, and Gaussian filtering achieves image smoothing and denoising by convolving the image with the Gaussian kernel.

[0056] The visual positioning algorithm in the image capture module is used to detect and match feature points, so that the system can automatically adjust the position and angle of the camera to ensure that the captured image can fully and accurately reflect the surface condition of the optical glass. In addition, the image capture module also has the characteristics of strong real-time performance and can quickly capture and transmit image data to meet the real-time requirements of the detection and early warning system, which is of great significance for timely detection and processing of defects in the optical glass processing process.

[0057] Example 3

[0058] The steps of the system based on Example 1 and Example 2 are as follows:

[0059] (1) Initialize the system

[0060] Launching the detection and early warning system for optical glass processing;

[0061] Initialize the parameters of each module, including camera settings and algorithm parameters.

[0062] (2) Image Capture

[0063] The image capture module starts a high-resolution camera to capture an image of the optical glass to be inspected;

[0064] The camera obtains a clear image of the glass surface according to the preset position and angle;

[0065] The captured images are transmitted to the image processing module in real time.

[0066] (3) Image processing and defect recognition

[0067] After receiving the image, the image processing module first performs preprocessing, such as denoising and enhancement, to improve the image quality;

[0068] The behavior recording module starts working, capturing the experimenter's hand movements and eye movement data in real time. Then, the built-in defect recognition algorithm automatically processes and analyzes the pre-processed images.

[0069] The algorithm uses feature extraction, classification and regression methods to identify defects in optical glass and determine their location, size and type.

[0070] (4) Result feedback and alarm

[0071] The result feedback module displays the recognition results to production personnel in real time, including detailed information of defects;

[0072] If the number or severity of identified defects exceeds the preset threshold, the result feedback module will trigger the alarm mechanism to remind production personnel through sound and light;

[0073] At the same time, detailed reports and statistical functions are provided to help production personnel understand the distribution and trends of defects.

[0074] (5) Intelligent learning and optimization

[0075] The intelligent learning and optimization module continuously collects data during the detection process, including images and recognition results;

[0076] Using this data, the module conducts model training and optimization through deep learning algorithms to improve the accuracy and stability of defect recognition;

[0077] With the accumulation of data and optimization of algorithms, the performance of the system will continue to improve to adapt to the processing needs of optical glass of different types and complexities.

[0078] (6) Cycle detection and monitoring

[0079] The system continuously performs a cycle of image capture, processing, feedback and learning to achieve real-time detection and monitoring of the optical glass processing process;

[0080] Production personnel can adjust processing parameters or take other necessary measures in a timely manner according to the feedback from the system to ensure product quality and production efficiency.

[0081] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A detection and early warning system for optical glass processing, characterized in that: The system consists of the following components: Image capture module: As the starting part of the entire detection and early warning system, the performance of the image capture module directly determines the accuracy of subsequent image processing and defect recognition. This module uses a high-resolution camera to ensure that the captured images are clear and delicate, providing high-quality raw data for subsequent defect recognition; Image processing module: automatically processes and analyzes the received images, with a built-in optimized defect recognition algorithm to identify defects in optical glass. In order to further improve the accuracy and efficiency of recognition, this module uses advanced deep learning algorithms to improve image quality and reduce the impact of interference factors on recognition results; Result feedback module: Feedback the recognition results to production personnel in real time, including the location, size and type of defects, so that production personnel can respond quickly. It also provides detailed reports and statistical functions to help production personnel understand the distribution and trend of defects, providing a basis for subsequent process improvement and quality control; Intelligent learning and optimization module: Optimize the defects in the result feedback through the defect recognition algorithm, improve the recognition accuracy and stability of the system, so that it can adapt to the processing needs of optical glass of different types and complexity, and then use the advanced gradient descent algorithm to improve the stability and convergence speed of the algorithm.

2. The optical glass processing detection and warning system according to claim 1, characterized in that: The specific steps of the image capture module are as follows: (1) Camera selection and calibration: According to the characteristics of optical glass and testing requirements, select the appropriate camera model and parameters, calibrate the camera, and ensure that the shooting angle and focal length parameters are set correctly to ensure the accuracy and consistency of image acquisition; (2) Automatic positioning and focusing: Through the visual positioning algorithm, the camera can automatically position the optical glass to be inspected. According to the thickness and curvature characteristics of the glass, the camera's focus distance is automatically adjusted to ensure a clear and blur-free image. (3) Image acquisition and transmission: After positioning and focusing are completed, the camera begins to acquire images of the optical glass to be inspected. The acquired image data is compressed and encoded, and then transmitted to the image processing module in real time through the data transmission interface Ethernet and USB.

3. The optical glass processing detection and warning system according to claim 1, characterized in that: The image processing module has the following specific steps: (1) Image preprocessing: De-noising and contrast enhancement preprocessing operations are performed on the received images to improve image quality and reduce the impact of noise and interference on subsequent recognition; (2) Feature extraction: Use deep learning algorithms to extract key features in the image, such as edges, textures, and colors. These features will serve as the basis for subsequent defect identification. (3) Defect recognition: Based on the extracted features, the defect recognition algorithm is used to identify defects in the image. The algorithm learns the characteristics and patterns of defects by training a large amount of sample data, so that defects in the glass can be accurately identified. (4) Defect location and classification: For the identified defects, the algorithm further determines their location, size and type. This information will be used for subsequent result feedback and early warning processing.

4. The optical glass processing detection and warning system according to claim 1, characterized in that: The specific steps of the result feedback module are as follows: (1) Visual display: The identified defects are marked and displayed on the original image, so that production personnel can intuitively understand the location, size and type of the defects; (2) Data reporting: The threshold judgment algorithm for defects generates detailed defect data reports, including the number, distribution, and type of defects. These data reports can be used to monitor and analyze the production process, helping production personnel better understand the product quality status; (3) Real-time alarm: When a serious defect is identified or the number of defects exceeds the preset threshold, the system triggers a real-time alarm mechanism, reminding production personnel to handle it in a timely manner through sound and light.

5. The optical glass processing detection and warning system according to claim 1, characterized in that: The specific steps of the intelligent learning and optimization module are as follows: (1) Data learning: Utilize a large amount of historical data and real-time data to continuously train and optimize the defect recognition algorithm. Through continuous learning and updating, the recognition accuracy and generalization ability of the algorithm are improved. (2) Algorithm optimization: According to the actual application scenarios and requirements, the algorithm is optimized and improved in a targeted manner. For different types of optical glass defects, the algorithm parameters are adjusted and new feature extraction methods are introduced to improve the recognition effect. (3) Adaptive adjustment: The system can automatically adjust image processing parameters and recognition thresholds based on real-time feedback data and results using a gradient descent optimization algorithm to adapt to detection requirements under different environments and conditions.

6. The optical glass processing detection and warning system according to claim 2, characterized in that The image capture module uses the SIFT algorithm in the visual positioning algorithm to detect and match feature points, extracts key points and their descriptors in the image through the Gaussian kernel function algorithm, and matches between different images to determine the relative position between the camera and the glass to be detected. The Gaussian filter formula is as follows: Among them, G(x,y) is the Gaussian kernel function, (x,y) is the pixel coordinate, σ is the standard deviation of the Gaussian kernel, and a multi-scale image is generated by convolving the original image with the Gaussian kernel to capture features of different sizes and thus construct a scale space. Local extreme points are searched in the scale space and used as potential feature points. The pixel gradients in the neighborhood of the key points are calculated, and the main direction of each key point is determined to achieve rotation invariance. Based on the position and main direction of the key points, the gradient information in the neighborhood is calculated and encoded into a high-dimensional vector to form a descriptor with robustness to illumination, viewing angle, and scale. The auxiliary image capture module realizes the precise positioning of the glass to be inspected, providing high-quality image data for subsequent image processing and defect recognition.

7. The optical glass processing detection and warning system according to claim 3, characterized in that ,The image processing module uses a deep neural network (CNN) to build a model in the feature extraction process through a deep learning algorithm, automatically extracts useful features from the original data, and realizes hierarchical representation of features. This feature extraction and representation learning capability enables the deep learning model to capture the inherent structure and laws of the data. In classification and regression tasks, the deep learning model uses a convolutional neural network to learn the complex mapping relationship from input to output. Through a large amount of training data, these models can continuously optimize their own parameters and structures to improve the accuracy of prediction and classification. The convolution operation in the convolution layer of the convolutional neural network (CNN) is used to extract local features of the input data. Mathematically, the convolution operation can be expressed as: S(i,j)=(I*K)(i,j)=∑ m ∑ n I(i+m,j+n)K(m,n), where I is the input data, K is the convolution kernel, and K is the output after convolution. The complex mapping relationship from input to output enables high-precision prediction and classification.

8. The optical glass processing detection and warning system according to claim 4, characterized in that ,The result feedback module uses a threshold judgment algorithm to judge whether the severity or quantity of ,defects exceeds the preset threshold and thus triggers the alarm mechanism. ,Assuming that the preset threshold of the number of defects is (T) and the number of ,defects currently detected is (N), the logic of the threshold judgment can be expressed as: if N>T, then the alarm is triggered. ,When the result feedback module triggers the alarm, the threshold judgment algorithm ,organizes the defect information into an easy-to-understand form, and then converts the ,information into graphics or icons and displays them on the user interface. ,The real-time communication algorithm ensures that this information can be transmitted to ,production personnel in real time so that they can make timely responses based on ,the information.

9. The optical glass processing detection and warning system according to claim 3, characterized in that The image processing module realizes automatic detection and identification of defects in glass products in the optical glass processing detection and warning system through the defect recognition algorithm. The logistic regression in this algorithm predicts whether an optical glass contains defects through image features in the defect recognition task. The formula of logistic regression is as follows: P(y=1|x)=sigmoid(w*x+b), where P(y=1|x) represents the probability that the output y is 1 (i.e., there is a defect) given the input feature x. The sigmoid function is used to map the output of linear regression to a probability value between 0 and 1. w and b are parameters of the model, representing weights and biases respectively. Through complex calculations and analysis, the algorithm can accurately identify defects in optical glass. Whether it is surface defects or internal defects, it can classify and identify defects, helping production personnel understand the type and severity of defects.

10. The optical glass processing detection and warning system according to claim 1, characterized in that ,The intelligent learning and optimization module uses the gradient descent optimization algorithm, and the formula of this algorithm is: Where t represents the model parameters of the tth step, {t+1} represents the updated model parameters, and L represents the gradient of the loss function at t. These data are then used to update and adjust the algorithm online, and iterate and optimize continuously. In order for the algorithm to gradually adapt to new environments and requirements and improve recognition accuracy and generalization capabilities, a gradient descent algorithm is used to improve stability and convergence speed. At the same time, it also has adaptive adjustment capabilities, and can automatically adjust algorithm parameters and computing resources according to the system's operating status and performance requirements to ensure efficient operation of the system.

Images