Optical lens polishing quality optimization method and system

Abstract

The present invention relates to the field of optical lens technology, and specifically to a method and system for optimizing the polishing quality of optical lenses. First, the surface quality data of the initial optical lens and the finished lens are collected to construct an initial data set. The samples with serious defects are removed by preprocessing to form a standard data set; and feature extraction is performed on the standard data set, and the features that significantly affect the polishing quality are analyzed using the Pearson correlation coefficient, so as to construct an automatic control model, output polishing parameters, and divide different areas to formulate corresponding polishing strategies. Secondly, the optical lens is polished according to the polishing parameters, and the lens surface is continuously monitored in real time during the polishing process. The real-time quality score is calculated by considering the roughness, number of defects, material uniformity and energy consumption of the optical lens. Finally, an early warning mechanism is established based on the real-time quality score to obtain the risk level of the optical lens during the polishing process to ensure timely response when quality risks occur.

Description

Technical Field

[0001] The present invention relates to the technical field of optical lenses, and in particular to a method and system for optimizing the polishing quality of optical lenses. Background Art

[0002] Optical lenses are widely used in modern optical devices, such as camera lenses, microscopes, telescopes, laser equipment, etc. With the advancement of technology, the surface quality and precision of optical lenses are increasingly required. High-quality optical lenses need to have extremely low surface roughness and high shape accuracy to ensure excellent imaging quality and optical performance in the optical system. However, the polishing process of optical lenses is a complex and delicate process that directly affects the final quality of the lenses.

[0003] Traditional optical lens polishing processes usually rely on the operator's experience and manual adjustments. In addition, traditional polishing methods often result in surface defects, edge cracking and shape deviation when processing these high-demand lenses, leading to increased production costs and higher scrap rates.

[0004] In order to solve these problems, CNC grinding equipment and advanced measurement technology are gradually introduced. However, the existing automated grinding system is difficult to flexibly respond to the needs of different lenses, and lacks a real-time feedback control mechanism, which makes it difficult to further improve the grinding quality and production efficiency. Therefore, how to optimize the grinding quality of optical lenses, reduce surface defects, and improve production efficiency has become a technical problem that needs to be solved urgently in the current optical manufacturing field.

[0005] To this end, a method and system for optimizing the polishing quality of optical lenses are proposed. Summary of the invention

[0006] The purpose of the present invention is to provide a method and system for optimizing the polishing quality of optical lenses. First, the surface quality data of the initial optical lens and the finished lens are collected to construct an initial data set. The samples with serious defects are removed by preprocessing to form a standard data set; and the standard data set is subjected to feature extraction, and the features that significantly affect the polishing quality are analyzed using the Pearson correlation coefficient, so as to construct an automatic control model, output polishing parameters, and divide different areas to formulate corresponding polishing strategies. Secondly, the optical lens is polished according to the polishing parameters, and the lens surface is continuously monitored in real time during the polishing process. The real-time quality score is calculated by considering the roughness, number of defects, material uniformity and energy consumption of the optical lens. Finally, an early warning mechanism is established based on the real-time quality score to obtain the risk level of the optical lens during the polishing process to ensure timely response when quality risks occur.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A method for optimizing the polishing quality of an optical lens, comprising:

[0009] Collect surface quality data of initial optical lenses and surface quality data of finished lenses to construct an initial data set;

[0010] Preprocessing the initial data set to remove initial optical lenses with serious defects and obtain a standard data set;

[0011] Extract features from the standard data set, build an automated control model, obtain polishing parameters, divide different areas and formulate corresponding polishing strategies, and adjust the areas that need to be adjusted during the polishing of the optical lens in real time; the polishing parameters include polishing lens pressure, polishing lens speed, polishing mode and coolant flow rate;

[0012] Grinding the optical lens according to the grinding parameters, and monitoring the optical lens in real time during the grinding process to obtain a real-time quality score of the optical lens;

[0013] An early warning mechanism is established based on the quality score of the real-time optical lens; the early warning mechanism divides the risk level into two levels based on the quality score of the real-time optical lens, specifically including:

[0014] High risk level: ;

[0015] Low risk level: ;

[0016] in, By analyzing the distribution of various risk levels in historical data, .

[0017] Preferably, the data in the initial data set include average roughness, peak-to-valley height, root mean square roughness, diameter, thickness, radius of curvature, material properties, scratches, pits, cracks and coolant flow of the optical lens.

[0018] Preferably, the specific process of identifying the initial optical lens with the serious defect includes:

[0019] Performing a missing value check on the data in the initial data set, and performing a secondary check on the data with missing values ​​to obtain a first initial data set;

[0020] The first initial data set is normalized by The function identifies outliers; the initial optical lenses corresponding to the outliers are removed;

[0021] The remaining data are normalized to obtain the standard data set.

[0022] Preferably, the feature extraction is used to extract features that have a greater impact on the polishing quality from the standard data set, and the specific process includes:

[0023] The Pearson correlation coefficient method is used to calculate the correlation between the data and the polishing quality of the data in the standard data set;

[0024] The larger the correlation coefficient is, the higher the degree of correlation with the polishing quality is; a standard feature vector group is obtained through the correlation coefficient to establish an automatic control model;

[0025] The automatic control model includes a preprocessing layer, a feature analysis layer and a parameter prediction layer;

[0026] The preprocessing layer includes data cleaning, standardization, normalization and missing value processing;

[0027] The feature analysis layer includes an LSTM network structure, which is used to analyze the standard feature vector group to obtain a first feature vector;

[0028] The parameter prediction layer includes four output neurons, each of which corresponds to a prediction target; the prediction targets include lens polishing pressure, lens polishing speed, polishing mode and coolant flow rate; the parameter prediction layer is used to predict the parameters in the standard feature vector group according to the standard feature vector group to obtain the prediction result;

[0029] By comparing the predicted results with the actual parameters in real time, the grinding parameters can be automatically adjusted.

[0030] Preferably, the automated control model includes region identification:

[0031] Based on empirical rules, the initial optical lens is divided into different areas, and corresponding polishing strategies are formulated for the different areas based on the automated control model;

[0032] Based on the linear adjustment algorithm, the grinding pressure is adjusted according to the surface roughness of different areas. The specific formula is:

[0033] ;

[0034] in, is the adjusted grinding pressure, The pressure is set under standard operating conditions. is the pressure adjustment factor, is the target roughness, The measured roughness.

[0035] Preferably, the real-time monitoring captures the image of the optical lens surface in real time through a high-resolution camera, identifies the scratches, pits and cracks on the optical lens surface during the polishing process through an image recognition method; measures the surface roughness of the optical lens through laser scanning and compares it with the target roughness to calculate the deviation;

[0036] Based on the real-time data of the optical lens during the polishing process and the preset quality standards, the real-time optical lens quality score of each lens is automatically generated; the specific calculation formula of the real-time optical lens quality score is:

[0037] ;

[0038] in, Score the quality of real-time optical lenses, is the weight of the roughness score, To adjust the sensitivity parameters affected by roughness deviation, is the target roughness, is the measured roughness, The weights for scoring defects, To adjust the coefficient of the effect of the number of defects on the score, For the The number of class defects, The weights for scoring other quality dimensions, To score the quality dimensions of material uniformity and optical performance testing, is the weight of the efficiency score, The time required for the grinding process, To adjust the coefficient of energy consumption on efficiency score, The energy consumed during the grinding process.

[0039] Preferably, when the quality score of the real-time optical lens is less than When the quality score of the real-time optical lens is greater than or equal to , the corresponding risk level is low risk level.

[0040] An optical lens grinding quality optimization system, comprising:

[0041] A data collection unit, used to collect surface quality data of the initial optical lens and the surface quality data of the finished lens, and construct an initial data set;

[0042] A data analysis unit, used to pre-process the initial data set, remove initial optical lenses with serious defects, and obtain a standard data set;

[0043] An automated control unit, used to extract features from the standard data set, build an automated control model, obtain polishing parameters, divide different areas into different areas, formulate corresponding polishing strategies, and adjust the areas that need to be adjusted during the polishing process of the optical lens in real time; the polishing parameters include polishing lens pressure, polishing lens speed, polishing mode and coolant flow rate;

[0044] A real-time optical detection unit, used to grind the optical lens according to the grinding parameters, and to monitor the optical lens in real time during the grinding process to obtain a real-time quality score of the optical lens;

[0045] The early warning unit is used to establish an early warning mechanism according to the quality score of the real-time optical lens.

[0046] Preferably, the automated control unit includes area identification:

[0047] Based on empirical rules, the initial optical lens is divided into different areas, and corresponding polishing strategies are formulated for the different areas based on the automated control model;

[0048] Based on the linear adjustment algorithm, the grinding pressure is adjusted according to the surface roughness of different areas. The specific formula is:

[0049] ;

[0050] in, is the adjusted grinding pressure, The pressure is set under standard operating conditions. is the pressure adjustment factor, is the target roughness, The measured roughness.

[0051] Preferably, the real-time monitoring in the real-time optical detection unit captures the image of the surface of the optical lens in real time through a high-resolution camera, identifies the scratches, pits and cracks on the surface of the optical lens during the polishing process through an image recognition method; measures the surface roughness of the optical lens through laser scanning and compares it with the target roughness to calculate the deviation;

[0052] Based on the real-time data of the optical lens during the polishing process and the preset quality standards, a real-time optical lens quality score for each lens is automatically generated.

[0053] Compared with the prior art, the present invention has the following beneficial effects:

[0054] 1. When proposing the automatic control model, the present invention firstly performs normalization processing and applies The function identifies outliers, checks the data for missing values, and removes lens samples with serious defects. By promptly removing lenses that have serious defects that affect the quality of the final product, the scrap rate and rework rate are effectively reduced. At the same time, automated data processing reduces the need for manual inspection, thereby improving production efficiency and yield. Subsequently, the Pearson correlation coefficient method is used to extract key features closely related to the polishing quality, and an automated control model is constructed to determine the polishing parameters. This method achieves accurate prediction and dynamic adjustment of polishing parameters, allowing the polishing process to be adaptively optimized based on real-time monitoring data, significantly improving the surface quality of the lens, reducing the occurrence of defects, improving production efficiency, and reducing the defective rate.

[0055] 2. The present invention further improves the accuracy and consistency of optical lens polishing by introducing region recognition and linear adjustment algorithms into the automated control model. The initial optical lens is divided into different regions based on empirical rules, and corresponding polishing strategies are formulated according to the specific characteristics of each region. Through the linear adjustment algorithm, the polishing pressure is dynamically adjusted according to the surface roughness of each region, thereby achieving precise control of the polishing process. This method can effectively optimize the polishing effect of different regions, significantly improve the consistency and quality of the lens surface, while reducing the incidence of human errors and defects in the polishing process, and improving overall production efficiency.

[0056] 3. The optical lens polishing quality optimization method of the present invention effectively improves the monitoring and control accuracy of lens polishing. By real-time monitoring of scratches, pits and cracks on the lens surface through high-resolution cameras and laser scanning technology, and measuring surface roughness, this method can timely detect and evaluate quality deviations in the polishing process. Based on the real-time data collected during the polishing process, the quality score of the optical lens is automatically generated, which comprehensively considers the roughness, number of defects and other quality dimensions, such as material uniformity and optical properties, as well as polishing efficiency and energy consumption. In addition, the set quality score threshold can effectively distinguish between high-risk and low-risk lenses, and realize dynamic management of risk levels. This instant feedback mechanism not only reduces the defective rate, but also improves the efficiency of the overall production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 A schematic diagram of a process flow of an optical lens polishing quality optimization method provided by an embodiment of the present invention;

[0058] Figure 2 A schematic diagram of the structure of an optical lens polishing quality optimization system provided by an embodiment of the present invention;

[0059] Figure 3 A schematic diagram of the structure of an automation control unit provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. 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.

[0061] The present invention provides a method for optimizing the polishing quality of an optical lens. The method is applied to an optical lens polishing quality optimization system. The flowchart of the specific method and system is shown in Figure 1 and Figure 2 .

[0062] Embodiment 1

[0063] As an embodiment of the present invention, refer to Figure 1 S10 in the embodiment is applied to a data collection unit of an optical lens polishing quality optimization system. In this embodiment, surface quality data of initial optical lenses and surface quality data of finished lenses of Company A, which produces high-precision optical lenses, are collected to construct an initial data set.

[0064] Furthermore, the data in the initial data set include average roughness, peak-to-valley height, root mean square roughness, diameter, thickness, radius of curvature, material properties, scratches, pits, cracks and coolant flow of the optical lens.

[0065] Among them, the average roughness reflects the overall smoothness of the lens surface and is an important indicator for evaluating the grinding quality; the peak-to-valley height provides the height difference between the highest and lowest points on the lens surface, which helps to evaluate the irregularity of the surface; the root mean square roughness measures the statistical parameters of the surface roughness and reflects the average height of the surface texture; geometric dimensions such as diameter and thickness have a direct impact on determining the grinding pressure and speed of the lens; the radius of curvature affects the optical properties of the lens and needs to be precisely controlled to meet the design specifications; material properties The thermal expansion coefficient and hardness of different materials will affect the setting of grinding parameters; scratches, pits, cracks: these defect data directly affect the quality scoring system and are key indicators for comparing quality before and after grinding; maintaining an appropriate flow rate of coolant is an important condition for preventing lens overheating and reducing wear.

[0066] In this embodiment, the collection of detailed data allows the system to more accurately locate defects and uneven areas on the lens, so that the polishing process can be more finely adjusted. For example, by measuring the average roughness and peak-to-valley height, the system can accurately calculate the areas that need extra attention, as well as the necessary polishing intensity and time.

[0067] Reference Figure 1S20 in the figure is applied to a data analysis unit of an optical lens polishing quality optimization system, and the data analysis unit is used to preprocess the initial data set, remove the initial optical lenses with serious defects, and obtain a standard data set.

[0068] Furthermore, the specific process of identifying the initial optical lens with the serious defect includes:

[0069] Performing a missing value check on the data in the initial data set, and performing a secondary check on the data with missing values ​​to obtain a first initial data set;

[0070] The first initial data set is normalized by The function identifies outliers; the initial optical lenses corresponding to the outliers are removed;

[0071] in The specific calculation method of the function is:

[0072] ;

[0073] in, For the standard data set data, For the The mean of the data, For the The standard deviation of the data;

[0074] when hour, is an outlier, is the threshold for judging abnormal values; ;

[0075] The remaining data are normalized to obtain the standard data set.

[0076] In the optical lens polishing quality optimization system, the process of identifying initial optical lenses with serious defects significantly improves the efficiency and product quality of the entire manufacturing process by enhancing data quality and optimizing production processes. First, by performing missing value checks and secondary checks on the initial data set, the process ensures the integrity and accuracy of the data, providing a solid foundation for the formulation of efficient polishing strategies. Then, through normalization and application The function identifies outliers. This step standardizes the data, reduces the impact of measurement scale differences, and ensures the accuracy and reliability of subsequent analysis. This method effectively reduces the scrap rate and rework rate by promptly removing lenses that may affect the quality of the final product due to serious defects, thereby improving production efficiency and yield.

[0077] In addition, automated data processing reduces the need for manual inspection, improving the speed and efficiency of the production process while maintaining consistency and predictability in processing time. This strategy of identifying and eliminating defective lenses in advance not only saves labor and material costs, but also reduces machine operating time, effectively reducing production costs.

[0078] Reference Figure 1 S30 in the figure is applied to an automatic control unit of an optical lens polishing quality optimization system, and the automatic control unit is used to extract features from the standard data set, build an automatic control model, obtain polishing parameters, divide different areas into corresponding polishing strategies, and adjust the areas that need to be adjusted during the optical lens polishing process in real time; the polishing parameters include polishing lens pressure, polishing lens speed, polishing mode and coolant flow rate.

[0079] Furthermore, the feature extraction is used to extract features that have a greater impact on the polishing quality from the standard data set, and the specific process includes:

[0080] The Pearson correlation coefficient method is used to calculate the correlation between the data and the polishing quality of the data in the standard data set. The specific formula is:

[0081] ;

[0082] in, The standard dataset The correlation coefficient between the column eigenvector and the polishing quality, For the The first column of the eigenvector Row data, For the The mean of the column eigenvectors, For the The first column of the eigenvector Row data, For the The mean of the column eigenvectors;

[0083] The larger the correlation coefficient is, the higher the degree of correlation with the polishing quality is; a standard feature vector group is obtained through the correlation coefficient to establish an automatic control model;

[0084] The automatic control model includes a preprocessing layer, a feature analysis layer and a parameter prediction layer;

[0085] The preprocessing layer includes data cleaning, standardization, normalization and missing value processing;

[0086] The feature analysis layer includes an LSTM network structure, which is used to analyze the standard feature vector group to obtain a first feature vector;

[0087] The parameter prediction layer includes four output neurons, each of which corresponds to a prediction target; the prediction targets include lens polishing pressure, lens polishing speed, polishing mode and coolant flow rate; the parameter prediction layer is used to predict the parameters in the standard feature vector group according to the standard feature vector group to obtain the prediction result;

[0088] By comparing the predicted results with the actual parameters in real time, the grinding parameters can be automatically adjusted.

[0089] In this embodiment, the Pearson correlation coefficient is used to accurately identify features closely related to the polishing quality, thereby achieving more accurate polishing parameter settings, such as polishing pressure, speed, and coolant flow. The adjustment of parameters directly improves the surface treatment quality of the optical lens and ensures the consistency and reliability of the product.

[0090] In addition, by building an automated control model, the system can compare the predicted results with the actual parameters in real time and make adjustments automatically, greatly improving the efficiency and automation of the polishing process. It not only reduces the need for manual operation and reduces production costs, but also shortens the production cycle and improves the efficiency of the overall production line. In particular, the introduction of the LSTM network optimizes the prediction accuracy of the feature analysis layer and further improves the accuracy of parameter prediction by learning the patterns of long-term data. The automated prediction and adjustment process reduces the product scrap rate caused by parameter errors, saves raw materials, and reduces unnecessary energy consumption. The continuous operation of this system also helps to collect important production data, provide data support for future operation optimization, and enable enterprises to continuously adjust and optimize production processes to cope with changing market demands and technical standards.

[0091] Furthermore, the automation control model includes region identification:

[0092] Based on empirical rules, the initial optical lens is divided into different areas, and corresponding polishing strategies are formulated for the different areas based on the automated control model;

[0093] Based on the linear adjustment algorithm, the grinding pressure is adjusted according to the surface roughness of different areas. The specific formula is:

[0094] ;

[0095] in, is the adjusted grinding pressure, The pressure is set under standard operating conditions. is the pressure adjustment factor, is the target roughness, The measured roughness.

[0096] In this embodiment, the system divides the initial optical lens into different areas through regional recognition based on empirical rules. This division enables the system to formulate a special polishing strategy for each area and optimize the processing process of each area in a targeted manner. This method significantly improves the accuracy of the polishing process, ensuring that each area can obtain the best processing effect, thereby improving the overall product quality. Then, the system automatically adjusts the polishing pressure through a linear adjustment algorithm, which calculates the required pressure adjustment based on the difference between the measured surface roughness of each area and the target roughness. Pressure adjustment not only ensures the surface quality of the lens, but also helps to prevent material loss caused by excessive polishing, effectively extending the service life of the lens. In addition, by adjusting the polishing parameters in real time to match the needs of specific areas, the system greatly improves production efficiency and resource utilization. Accurate parameter control reduces errors and scrap rates in the production process, thereby reducing production costs.

[0097] Reference Figure 1 S40 in the figure is applied to a real-time optical detection unit of an optical lens polishing quality optimization system, and the real-time optical detection unit is used to polish the optical lens according to the polishing parameters, and to perform real-time monitoring of the optical lens during the polishing process to obtain a real-time quality score of the optical lens.

[0098] Furthermore, the real-time monitoring captures the image of the optical lens surface in real time through a high-resolution camera, identifies the scratches, pits and cracks on the optical lens surface during the polishing process through an image recognition method; measures the surface roughness of the optical lens through laser scanning and compares it with the target roughness to calculate the deviation;

[0099] Based on the real-time data of the optical lens during the polishing process and the preset quality standards, the real-time optical lens quality score of each lens is automatically generated; the specific calculation formula of the real-time optical lens quality score is:

[0100] ;

[0101] in, Score the quality of real-time optical lenses, is the weight of the roughness score, To adjust the sensitivity parameters affected by roughness deviation, is the target roughness, is the measured roughness, The weights for scoring defects, To adjust the coefficient of the effect of the number of defects on the score, For the The number of class defects, The weights for scoring other quality dimensions, To score the quality dimensions of material uniformity and optical performance testing, is the weight of the efficiency score, The time required for the grinding process, To adjust the coefficient of energy consumption on efficiency score, The energy consumed during the grinding process.

[0102] The real-time monitoring system provides in-depth insights and precise control of the optical lens polishing process through high-resolution cameras and laser scanning, significantly improving product quality and production efficiency. These technologies can identify scratches, pits and cracks on the lens surface in real time to ensure that defects are discovered and handled in a timely manner. The introduction of laser scanning makes the roughness measurement more accurate. The system adjusts the polishing parameters by comparing the deviation between the measured roughness and the target roughness. This real-time adjustment ensures that each lens meets the highest quality standards.

[0103] In addition, the system automatically generates a real-time quality score for each lens, taking into account roughness, number of defects and other quality dimensions such as material uniformity and optical performance. This scoring system provides operators with instant feedback, allowing for more precise adjustments during production, greatly reducing scrap rates and improving the efficient use of resources. Real-time data not only optimizes the production process, but also supports continuous quality improvement and production decisions by providing key performance indicators.

[0104] Reference Figure 1 S50 in the figure is applied to an early warning unit of an optical lens polishing quality optimization system, and the early warning unit is used to establish an early warning mechanism based on the real-time quality score of the optical lens.

[0105] Furthermore, an early warning mechanism is established according to the quality score of the real-time optical lens; the early warning mechanism divides the risk level into two levels according to the quality score of the real-time optical lens, specifically including:

[0106] High risk level: ;

[0107] Low risk level: ;

[0108] in, By analyzing the distribution of various risk levels in historical data, ;

[0109] When the quality score of the real-time optical lens is less than When the corresponding risk level is a high risk level, a sound warning is given for the high risk level; when the quality score of the real-time optical lens is greater than or equal to When , the corresponding risk level is low risk level. For specific results, please refer to Table 1.

[0110] Table 1 Scoring and risk warning table

[0111]

[0112] Among them, the sensitivity parameter of adjusting the roughness deviation is , target roughness , adjust the number of defects to score , the unit of grinding time is minutes, and the unit of energy consumption is kilowatt-hour.

[0113] In this embodiment, an early warning mechanism based on real-time quality scoring is introduced, which significantly enhances the quality control and risk management capabilities of the production process by dividing the risk level into high risk and low risk. The system automatically classifies according to the set scoring threshold, identifies problems that may affect product quality in real time, and takes corresponding measures in a timely manner. For example, when the quality score of an optical lens is lower than the high-risk threshold, the system will mark it as high risk, which may trigger a suspension of the production line or additional quality inspections to avoid defective products from entering the market. Conversely, when the score is higher than the low-risk threshold, the product is deemed to meet the quality standards and the production process can continue without obstacles. This early warning system not only reduces the cost loss caused by quality problems, but also improves production efficiency and product pass rate. The real-time monitoring and scoring mechanism ensures that each batch of products can meet consistent quality standards, greatly reducing the possibility of rework and returns.

[0114] The optical lens polishing quality optimization method and system of the present invention integrate advanced data acquisition, real-time monitoring and automatic control technologies, greatly improving the quality and efficiency of optical lens production. By adopting high-resolution cameras and laser scanners, the system can capture defects such as scratches, pits and cracks on the lens surface in real time, and measure the surface roughness, so as to promptly detect quality deviations in production. These real-time data enable the system to dynamically adjust polishing parameters such as pressure, speed and coolant flow to ensure that the lens quality meets the preset standards. In addition, by applying the Pearson correlation coefficient method to identify key features that affect polishing quality from a large amount of data, and using the LSTM network for in-depth analysis, the automated control model can accurately predict and adjust key polishing parameters. The system also establishes an early warning mechanism based on real-time quality scoring, and adjusts the production strategy in real time according to the risk level of the lens to prevent the continued production of unqualified products.

[0115] This comprehensive quality control strategy not only significantly reduces the cost and scrap rate caused by quality problems, but also improves overall production efficiency and resource utilization. Through continuous monitoring and automatic adjustment, the system ensures the consistency and high quality of each batch of products, enhances the market competitiveness of enterprises, and meets the dual needs of modern precision manufacturing for high efficiency and high quality, thus bringing significant economic and technical benefits to the optical manufacturing industry.

[0116] Embodiment 2

[0117] As an embodiment of the present invention, refer to Figure 1 S10 in the embodiment is applied to a data collection unit of an optical lens polishing quality optimization system. In this embodiment, surface quality data of initial optical lenses and surface quality data of finished lenses of Company B, which produces high-precision optical lenses, are collected to construct an initial data set.

[0118] Furthermore, the data in the initial data set include average roughness, peak-to-valley height, root mean square roughness, diameter, thickness, radius of curvature, material properties, scratches, pits, cracks and coolant flow of the optical lens.

[0119] Reference Figure 1 S20 in the figure is applied to a data analysis unit of an optical lens polishing quality optimization system, and the data analysis unit is used to preprocess the initial data set, remove the initial optical lenses with serious defects, and obtain a standard data set.

[0120] Furthermore, the specific process of identifying the initial optical lens with the serious defect includes:

[0121] Performing a missing value check on the data in the initial data set, and performing a secondary check on the data with missing values ​​to obtain a first initial data set;

[0122] The first initial data set is normalized by The function identifies outliers; the initial optical lenses corresponding to the outliers are removed;

[0123] in The specific calculation method of the function is:

[0124] ;

[0125] in, For the standard data set data, For the The mean of the data, For the The standard deviation of the data;

[0126] when hour, is an outlier, is the threshold for judging abnormal values; ;

[0127] The remaining data are normalized to obtain the standard data set.

[0128] Reference Figure 1 S30 in the figure is applied to an automatic control unit of an optical lens polishing quality optimization system, and the automatic control unit is used to extract features from the standard data set, build an automatic control model, obtain polishing parameters, divide different areas into corresponding polishing strategies, and adjust the areas that need to be adjusted during the optical lens polishing process in real time; the polishing parameters include polishing lens pressure, polishing lens speed, polishing mode and coolant flow rate.

[0129] Furthermore, the feature extraction is used to extract features that have a greater impact on the polishing quality from the standard data set, and the specific process includes:

[0130] The Pearson correlation coefficient method is used to calculate the correlation between the data and the polishing quality of the data in the standard data set. The specific formula is:

[0131] ;

[0132] in, The standard dataset The correlation coefficient between the column eigenvector and the polishing quality, For the The first column of the eigenvector Row data, For the The mean of the column eigenvectors, For the The first column of the eigenvector Row data, For the The mean of the column eigenvectors;

[0133] The larger the correlation coefficient is, the higher the degree of correlation with the polishing quality is; a standard feature vector group is obtained through the correlation coefficient to establish an automatic control model;

[0134] The automatic control model includes a preprocessing layer, a feature analysis layer and a parameter prediction layer;

[0135] The preprocessing layer includes data cleaning, standardization, normalization and missing value processing;

[0136] The feature analysis layer includes an LSTM network structure, which is used to analyze the standard feature vector group to obtain a first feature vector;

[0137] The parameter prediction layer includes four output neurons, each of which corresponds to a prediction target; the prediction targets include lens polishing pressure, lens polishing speed, polishing mode and coolant flow rate; the parameter prediction layer is used to predict the parameters in the standard feature vector group according to the standard feature vector group to obtain the prediction result;

[0138] By comparing the predicted results with the actual parameters in real time, the grinding parameters can be automatically adjusted.

[0139] Furthermore, the automation control model includes region identification:

[0140] Based on empirical rules, the initial optical lens is divided into different areas, and corresponding polishing strategies are formulated for the different areas based on the automated control model;

[0141] Based on the linear adjustment algorithm, the grinding pressure is adjusted according to the surface roughness of different areas. The specific formula is:

[0142] ;

[0143] in, is the adjusted grinding pressure, The pressure is set under standard operating conditions. is the pressure adjustment factor, is the target roughness, The measured roughness.

[0144] Reference Figure 1 S40 in the figure is applied to a real-time optical detection unit of an optical lens polishing quality optimization system, and the real-time optical detection unit is used to polish the optical lens according to the polishing parameters, and to perform real-time monitoring of the optical lens during the polishing process to obtain a real-time quality score of the optical lens.

[0145] Furthermore, the real-time monitoring captures the image of the optical lens surface in real time through a high-resolution camera, identifies the scratches, pits and cracks on the optical lens surface during the polishing process through an image recognition method; measures the surface roughness of the optical lens through laser scanning and compares it with the target roughness to calculate the deviation;

[0146] Based on the real-time data of the optical lens during the polishing process and the preset quality standards, the real-time optical lens quality score of each lens is automatically generated; the specific calculation formula of the real-time optical lens quality score is:

[0147] ;

[0148] in, Score the quality of real-time optical lenses, is the weight of the roughness score, To adjust the sensitivity parameters affected by roughness deviation, is the target roughness, is the measured roughness, The weights for scoring defects, To adjust the coefficient of the effect of the number of defects on the score, For the The number of class defects, The weights for scoring other quality dimensions, To score the quality dimensions of material uniformity and optical performance testing, is the weight of the efficiency score, The time required for the grinding process, To adjust the coefficient of energy consumption on efficiency score, is the energy consumed during the grinding process. Please refer to Table 2 for details.

[0149] Table 2 Real-time quality score table

[0150]

[0151] Among them, the sensitivity parameter of adjusting the roughness deviation is , target roughness , adjust the number of defects to score , the unit of grinding time is minutes, and the unit of energy consumption is kilowatt-hour.

[0152] Among them, the specific calculation formula for the scores of material uniformity and optical performance test quality dimensions is:

[0153] ;

[0154] in, For the The weight of each quality indicator, which reflects the importance of each indicator to the quality of the final product; For the The actual measured value of the quality indicator, For the The target value of a quality indicator.

[0155] Reference Figure 1 S50 in the figure is applied to an early warning unit of an optical lens polishing quality optimization system, and the early warning unit is used to establish an early warning mechanism based on the real-time quality score of the optical lens.

[0156] Furthermore, an early warning mechanism is established according to the quality score of the real-time optical lens; the early warning mechanism divides the risk level into two levels according to the quality score of the real-time optical lens, specifically including:

[0157] High risk level: ;

[0158] Low risk level: ;

[0159] in, By analyzing the distribution of various risk levels in historical data, ;

[0160] When the quality score of the real-time optical lens is less than When the quality score of the real-time optical lens is greater than or equal to , the corresponding risk level is low risk level.

[0161] The optical lens polishing quality optimization method and system proposed in the present invention significantly improves the quality and efficiency of optical lens production by integrating advanced data acquisition, real-time monitoring, and automated control technologies. The system captures the surface state of the lens in real time through a high-resolution camera and laser scanning equipment, accurately identifies defects such as scratches, pits, cracks, and measures the surface roughness. These real-time acquired data enable the system to dynamically adjust polishing parameters, such as polishing pressure, speed, and coolant flow, to ensure that each lens meets the set quality standards. In addition, the system uses the Pearson correlation coefficient method to extract key features that have a significant impact on the polishing quality from the collected data, and uses the LSTM network to deeply analyze these features, thereby optimizing the prediction and adjustment of polishing parameters. At the same time, through the real-time quality scoring mechanism, the system can perform risk assessment in real time according to the quality performance of the lens, establish an early warning mechanism, distinguish high-risk and low-risk products, and adjust or stop the production of unqualified products in time.

[0162] The system not only ensures product quality, reduces scrap and rework, but also improves the automation and efficiency of the production process. Through continuous monitoring and precise control, the invention effectively optimizes resource utilization, reduces production costs, and enhances the market competitiveness of enterprises.

[0163] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for optimizing the polishing quality of an optical lens, characterized in that: include: Collect surface quality data of initial optical lenses and surface quality data of finished lenses to construct an initial data set; Preprocessing the initial data set to remove initial optical lenses with serious defects and obtain a standard data set; Extract features from the standard data set, build an automated control model, obtain polishing parameters, divide different areas and formulate corresponding polishing strategies, and adjust the areas that need to be adjusted during the polishing of the optical lens in real time; the polishing parameters include polishing lens pressure, polishing lens speed, polishing mode and coolant flow rate; Grinding the optical lens according to the grinding parameters, and monitoring the optical lens in real time during the grinding process to obtain a real-time quality score of the optical lens; The real-time monitoring captures the image of the optical lens surface in real time through a high-resolution camera, identifies the scratches, pits and cracks on the optical lens surface during the polishing process through an image recognition method; measures the surface roughness of the optical lens through laser scanning and compares it with the target roughness to calculate the deviation; Based on the real-time data of the optical lens during the polishing process and the preset quality standards, the real-time optical lens quality score of each lens is automatically generated; the specific calculation formula of the real-time optical lens quality score is: ； in, Score the quality of real-time optical lenses, is the weight of the roughness score, To adjust the sensitivity parameters affected by roughness deviation, is the target roughness, is the measured roughness, The weights for scoring defects, To adjust the coefficient of the effect of the number of defects on the score, For the The number of class defects, The weights used to score the other quality dimensions, To score the quality dimensions of material uniformity and optical performance testing, is the weight of the efficiency score, The time required for the grinding process, To adjust the coefficient of energy consumption on efficiency score, The energy consumed during the grinding process; An early warning mechanism is established based on the quality score of the real-time optical lens; the early warning mechanism divides the risk level into two levels based on the quality score of the real-time optical lens, specifically including: High risk level: ; Low risk level: ; in, is the quality score threshold, .

2. The method for optimizing the polishing quality of an optical lens according to claim 1, characterized in that: The data in the initial data set include average roughness, peak-to-valley height, root mean square roughness, diameter, thickness, radius of curvature, material properties, scratches, pits, cracks and coolant flow of the optical lens.

3. The method for optimizing the polishing quality of an optical lens according to claim 1, characterized in that: The specific process of identifying the initial optical lens with the serious defects includes: Performing a missing value check on the data in the initial data set, and performing a secondary check on the data with missing values ​​to obtain a first initial data set; The first initial data set is normalized by The function identifies outliers; the initial optical lenses corresponding to the outliers are removed; The remaining data are normalized to obtain the standard data set.

4. The method for optimizing the polishing quality of an optical lens according to claim 1, characterized in that: The feature extraction is used to extract features that have a greater impact on the polishing quality from the standard data set, and the specific process includes: The Pearson correlation coefficient method is used to calculate the correlation between the data and the polishing quality of the data in the standard data set; The larger the correlation coefficient is, the higher the degree of correlation with the polishing quality is; a standard feature vector group is obtained through the correlation coefficient to establish an automatic control model; The automatic control model includes a preprocessing layer, a feature analysis layer and a parameter prediction layer; The preprocessing layer includes data cleaning, standardization, normalization and missing value processing; The feature analysis layer includes an LSTM network structure, which is used to analyze the standard feature vector group to obtain a first feature vector; The parameter prediction layer includes four output neurons, each of which corresponds to a prediction target; the prediction targets include lens polishing pressure, lens polishing speed, polishing mode and coolant flow rate; the parameter prediction layer is used to predict the parameters in the standard feature vector group according to the standard feature vector group to obtain the prediction result; By comparing the predicted results with the actual parameters in real time, the grinding parameters can be automatically adjusted.

5. The method for optimizing the polishing quality of an optical lens according to claim 4, characterized in that: The automated control model includes region identification: Based on empirical rules, the initial optical lens is divided into different areas, and corresponding polishing strategies are formulated for the different areas based on the automated control model; Based on the linear adjustment algorithm, the grinding pressure is adjusted according to the surface roughness of different areas. The specific formula is: ； in, is the adjusted grinding pressure, The pressure is set under standard operating conditions. is the pressure adjustment factor, is the target roughness, The measured roughness.

6. The method for optimizing the polishing quality of an optical lens according to claim 1, characterized in that: When the quality score of the real-time optical lens is less than When , the corresponding risk level is high risk level; When the quality score of the real-time optical lens is greater than or equal to , the corresponding risk level is low risk level.

7. An optical lens polishing quality optimization system, characterized in that: include: A data collection unit, used to collect surface quality data of the initial optical lens and the surface quality data of the finished lens, and construct an initial data set; A data analysis unit, used to pre-process the initial data set, remove initial optical lenses with serious defects, and obtain a standard data set; An automated control unit, used to extract features from the standard data set, build an automated control model, obtain polishing parameters, divide different areas into different areas, formulate corresponding polishing strategies, and adjust the areas that need to be adjusted during the polishing process of the optical lens in real time; the polishing parameters include polishing lens pressure, polishing lens speed, polishing mode and coolant flow rate; A real-time optical detection unit, used to grind the optical lens according to the grinding parameters, and to monitor the optical lens in real time during the grinding process to obtain a real-time quality score of the optical lens; The real-time monitoring captures the image of the optical lens surface in real time through a high-resolution camera, identifies the scratches, pits and cracks on the optical lens surface during the polishing process through an image recognition method; measures the surface roughness of the optical lens through laser scanning and compares it with the target roughness to calculate the deviation; Based on the real-time data of the optical lens during the polishing process and the preset quality standards, the real-time optical lens quality score of each lens is automatically generated; the specific calculation formula of the real-time optical lens quality score is: ； in, Score the quality of real-time optical lenses, is the weight of the roughness score, To adjust the sensitivity parameters affected by roughness deviation, is the target roughness, is the measured roughness, The weights for scoring defects, To adjust the coefficient of the effect of the number of defects on the score, For the The number of class defects, The weights used to score the other quality dimensions, To score the quality dimensions of material uniformity and optical performance testing, is the weight of the efficiency score, The time required for the grinding process, To adjust the coefficient of energy consumption on efficiency score, The energy consumed during the grinding process; The early warning unit is used to establish an early warning mechanism according to the quality score of the real-time optical lens; the early warning mechanism divides the risk level into two levels according to the quality score of the real-time optical lens, specifically including: High risk level: ; Low risk level: ; in, is the quality score threshold, .

8. The optical lens polishing quality optimization system according to claim 7, characterized in that: The automated control unit includes zone identification: Based on empirical rules, the initial optical lens is divided into different areas, and corresponding polishing strategies are formulated for the different areas based on the automated control model; Based on the linear adjustment algorithm, the grinding pressure is adjusted according to the surface roughness of different areas. The specific formula is: ； in, is the adjusted grinding pressure, The pressure is set under standard operating conditions. is the pressure adjustment factor, is the target roughness, The measured roughness.

9. The optical lens polishing quality optimization system according to claim 7, characterized in that: The real-time monitoring in the real-time optical detection unit captures the image of the optical lens surface in real time through a high-resolution camera, identifies the scratches, pits and cracks on the optical lens surface during the polishing process through an image recognition method; measures the surface roughness of the optical lens through laser scanning and compares it with the target roughness to calculate the deviation; Based on the real-time data of the optical lens during the polishing process and the preset quality standards, a real-time optical lens quality score for each lens is automatically generated.

Images