A defect tracking method and system for eyeglass coating

Abstract

The present invention discloses a defect tracking method and system for eyeglass coating. A three-dimensional model of an eyeglass product is constructed by a three-dimensional scanning device, and a plurality of positioning points are set in the model. In a plurality of production cycles, the coating image data is collected by using machine vision technology. The image data is feature recognized by using the recognition model and the positioning point features, and image positioning information and defect recognition information are generated. In combination with the eyeglass model, the defect coordinates of each cycle are located, and the defect coordinate changes are analyzed to generate defect change vectors. The defect change vectors are clustered by DBSCAN, the defect distribution deviation is calculated, and the coating process is controlled and optimized based on the defect recognition information. Through the present invention, the overall defect tracking and evaluation of coating production can be achieved, and the production efficiency can be further improved.

Description

Technical Field

[0001] The present invention relates to the field of eyeglass coating production analysis, and more specifically, to a defect tracking method and system for eyeglass coating. Background Art

[0002] The coating process of glasses is an important part of the manufacturing process of glasses. Through coating, the light transmittance, anti-reflection, wear resistance and other properties of the lens can be improved, and the aesthetics and durability of the frame can be improved. However, various defects often occur during the coating process, such as film shedding, bubbles, scratches, etc. These defects not only affect the appearance quality of the lens, but also affect its optical performance and service life. Therefore, real-time monitoring and tracking of coating defects is the key to ensuring product quality and production efficiency. However, the existing technology lacks effective analysis of coating defects and defect distribution evaluation, which makes it difficult to improve production efficiency. Therefore, there is an urgent need for a defect tracking method and system for glasses coating. Summary of the invention

[0003] The present invention overcomes the defects of the prior art and proposes a defect tracking method and system for eyeglass coating.

[0004] A first aspect of the present invention provides a defect tracking method for eyeglass coating, comprising:

[0005] Using a 3D scanning device, a 3D image scan of the target eyewear product is performed and a 3D eyewear model is constructed;

[0006] In the glasses model, multiple positioning points are set based on the target glasses product, the positioning points are marked, and the image features of the marked points are obtained;

[0007] The target eyewear products are coated in multiple production cycles. Based on machine vision technology, the coating image data of the target eyewear products are collected in each cycle.

[0008] By identifying the model and the features of the mark point image, the coating image data is subjected to feature recognition. The extracted features are first compared with the features of the mark point image to generate image positioning information based on the glasses model, and then compared with the coating defect comparison features to generate defect recognition information.

[0009] According to the image positioning information and defect recognition information, the defects in each production cycle are located in coordinates in combination with the eyewear model, the defect coordinate changes in two adjacent production cycles are analyzed, and multiple defect change vectors are generated. The defect change vectors are clustered based on DBSCAN data, and the defect distribution deviation in two adjacent production cycles is determined by the number of clustering groups;

[0010] The defect distribution deviation between all two adjacent production cycles is calculated and analyzed, and the overall defect tracking and evaluation of coating production is carried out through the defect distribution deviation. Based on the defect identification information, the coating production is regulated and optimized.

[0011] In this solution, the three-dimensional scanning device is used to perform three-dimensional image scanning on the target eyewear product and construct a three-dimensional eyewear model, specifically:

[0012] The target eyewear product is scanned in three dimensions at multiple angles using three-dimensional scanning equipment. The three-dimensional structure is analyzed through the scanned image, and a three-dimensional eyewear model is constructed.

[0013] In this solution, in the glasses model, multiple positioning points are set based on the target glasses product, the positioning points are marked, and the image features of the marked points are obtained, specifically:

[0014] In the glasses model, multiple positioning points are set based on the important production points of the target glasses products;

[0015] In the glasses model, an origin is randomly set and three-dimensional coordinates are constructed;

[0016] Based on the three-dimensional coordinates, analyze the shortest distance between multiple positioning points, and add positioning points to ensure that the shortest distance between adjacent positioning points is within the preset range;

[0017] Acquire standard image feature data of multiple product parts of a target eyewear product;

[0018] Marking is performed based on all the positioning points, and the features of the positioning points are extracted from the standard image feature data to obtain the image features of the marking points.

[0019] In this solution, the target eyewear products are coated in multiple production cycles. Based on machine vision technology, the coating image data of the target eyewear products are collected in each cycle, specifically:

[0020] Set production cycle;

[0021] Perform coating processing on target eyewear products in multiple production cycles. During the processing, collect coating image data of target eyewear products in each cycle based on machine vision technology;

[0022] The coating image data includes image data at multiple angles.

[0023] In this solution, the coating image data is feature recognized by identifying the model and the mark point image features, and the extracted features are first compared with the mark point image features to generate image positioning information based on the glasses model, and then compared with the coating defect comparison features to generate defect recognition information, specifically:

[0024] Perform image noise reduction and standardization preprocessing on the coating image data;

[0025] Based on the Canny edge detection operator, the texture and contour features of the coating image data are extracted to obtain real-time features;

[0026] Import the real-time features and the image features of the marked points into the recognition model for the first feature comparison, calculate the similarity of the feature data through the cosine distance, determine the location of the real-time features based on the similarity, and generate the image location information corresponding to the real-time features in combination with the glasses model;

[0027] Obtain coating defect contrast features from the image database;

[0028] Import the real-time features and the coating defect comparison features into the recognition model for a second feature comparison, perform feature learning and defect recognition analysis on the two imported features, and generate corresponding defect recognition information;

[0029] Defect analysis is performed on the coating image data of multiple production cycles to obtain image positioning information and defect identification information corresponding to each production cycle.

[0030] In this solution, in the coating image data, each production cycle includes corresponding coating image data.

[0031] In this solution, the defects in each production cycle are located according to the image positioning information and defect recognition information in combination with the glasses model, the defect coordinate changes in two adjacent production cycles are analyzed, and multiple defect change vectors are generated. The defect change vectors are clustered based on DBSCAN data, and the defect distribution deviation in two adjacent production cycles is determined by the number of clustering groups.

[0032] According to the image positioning information and defect recognition information corresponding to a production cycle, the identified defects are located in three dimensions. The coordinate positioning is performed based on the glasses model to generate defect positioning data.

[0033] The defect location data includes a plurality of defect coordinates;

[0034] Obtain defect location data of two adjacent production cycles, marked as first defect location data and second defect location data;

[0035] In the first defect location data, a defect coordinate P1 is selected, and the defect coordinates with the shortest and farthest distances from P1 are selected from the second defect location data through the glasses model, and are marked as P2 and P3 respectively;

[0036] Based on P1 and P2, P2 and P3, two vectors are generated respectively and marked as defect change vectors, defect coordinate change analysis is performed on all defect coordinates in the first defect location data, and multiple defect change vectors are generated;

[0037] Integrate multiple defect change vectors to form a vector data set, cluster the vector data set based on the DBSCAN clustering algorithm, and obtain the clustering results;

[0038] The number of clustering groups is obtained through clustering results, and the number of clustering groups is used as an indicator value for judging defect deviation to calculate the corresponding defect distribution deviation.

[0039] In this solution, the defect distribution deviation of all two adjacent production cycles is calculated and analyzed, and the overall defect tracking and evaluation of the coating production is carried out through the defect distribution deviation, and the production regulation and optimization of the coating production is carried out based on the defect identification information, specifically:

[0040] Obtain coating production equipment and environmental parameters for each production cycle;

[0041] Calculate and analyze the defect distribution deviation between all two adjacent production cycles, evaluate the change of defect distribution deviation based on the time dimension, combine the coating production equipment and environmental parameters, conduct overall defect tracking and equipment production quality evaluation for coating production, and generate an evaluation report;

[0042] Control and optimize coating production based on evaluation reports and defect identification information;

[0043] Set the defect distribution deviation warning value, and use the warning value to judge in real time whether the defect distribution of the production cycle meets expectations, and issue a real-time warning.

[0044] The second aspect of the present invention further provides a defect tracking system for eyeglass coating, the system comprising: a memory, a processor, the memory comprising a defect tracking program for eyeglass coating, the defect tracking program for eyeglass coating implementing the following steps when executed by the processor:

[0045] Using a 3D scanning device, a 3D image scan is performed on the target eyewear product and a 3D eyewear model is constructed;

[0046] In the glasses model, multiple positioning points are set based on the target glasses product, the positioning points are marked, and the image features of the marked points are obtained;

[0047] The target eyewear products are coated in multiple production cycles. Based on machine vision technology, the coating image data of the target eyewear products are collected in each cycle.

[0048] By identifying the model and the features of the mark point image, the coating image data is subjected to feature recognition. The extracted features are first compared with the features of the mark point image to generate image positioning information based on the glasses model, and then compared with the coating defect comparison features to generate defect recognition information.

[0049] According to the image positioning information and defect recognition information, the defects in each production cycle are located in coordinates in combination with the eyewear model, the defect coordinate changes in two adjacent production cycles are analyzed, and multiple defect change vectors are generated. The defect change vectors are clustered based on DBSCAN data, and the defect distribution deviation in two adjacent production cycles is determined by the number of clustering groups;

[0050] The defect distribution deviation between all two adjacent production cycles is calculated and analyzed, and the overall defect tracking and evaluation of coating production is carried out through the defect distribution deviation. Based on the defect identification information, the coating production is regulated and optimized.

[0051] The third aspect of the present invention also provides a computer-readable storage medium, which includes a defect tracking program for eyeglass coating. When the defect tracking program for eyeglass coating is executed by a processor, the steps of the defect tracking method for eyeglass coating as described in any one of the above items are implemented.

[0052] The present invention discloses a defect tracking method and system for eyeglass coating. A three-dimensional model of an eyeglass product is constructed by a three-dimensional scanning device, and a plurality of positioning points are set in the model. In a plurality of production cycles, the coating image data is collected by using machine vision technology. The image data is feature recognized by using the recognition model and the positioning point features, and image positioning information and defect recognition information are generated. In combination with the eyeglass model, the defect coordinates of each cycle are located, and the defect coordinate changes are analyzed to generate defect change vectors. The defect change vectors are clustered by DBSCAN, the defect distribution deviation is calculated, and the coating process is controlled and optimized based on the defect recognition information. Through the present invention, the overall defect tracking and evaluation of coating production can be achieved, and the production efficiency can be further improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A flow chart of a defect tracking method for eyeglass coating according to the present invention is shown;

[0054] Figure 2 A block diagram of a defect tracking system for eyeglass coating according to the present invention is shown. DETAILED DESCRIPTION

[0055] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0056] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0057] Figure 1 A flow chart of a defect tracking method for eyeglass coating according to the present invention is shown.

[0058] like Figure 1 As shown, the first aspect of the present invention provides a defect tracking method for eyeglass coating, comprising:

[0059] S102, performing a three-dimensional image scan of the target eyewear product and constructing a three-dimensional eyewear model by using a three-dimensional scanning device;

[0060] S104, in the glasses model, multiple positioning points are set based on the target glasses product, the positioning points are marked, and image features of the marked points are obtained;

[0061] S106, performing coating processing on the target eyewear product in multiple production cycles, and collecting coating image data of the target eyewear product in each cycle based on machine vision technology;

[0062] S108, performing feature recognition on the coating image data by identifying the model and the mark point image features, comparing the extracted features with the mark point image features for the first time and generating image positioning information based on the glasses model, and comparing the extracted features with the coating defect comparison features for the second time and generating defect recognition information;

[0063] S110, based on the image positioning information and the defect recognition information, coordinate positioning of the defects in each production cycle is performed in combination with the eyewear model, the defect coordinate changes in two adjacent production cycles are analyzed, and multiple defect change vectors are generated, and data clustering based on DBSCAN is performed on the defect change vectors, and the defect distribution deviation in two adjacent production cycles is determined by the number of clustering groups;

[0064] S112, calculate and analyze the defect distribution deviation of all two adjacent production cycles, conduct overall defect tracking and evaluation of coating production through the defect distribution deviation, and conduct production regulation and optimization of coating production based on defect identification information.

[0065] According to an embodiment of the present invention, the three-dimensional scanning device is used to perform three-dimensional image scanning on the target eyewear product and construct a three-dimensional eyewear model, specifically:

[0066] The target eyewear product is scanned in three dimensions at multiple angles using three-dimensional scanning equipment. The three-dimensional structure is analyzed through the scanned image, and a three-dimensional eyewear model is constructed.

[0067] It should be noted that the target eyewear product is generally an unprocessed product, which is used to construct an original three-dimensional model, and the model is used for subsequent defect location and deviation analysis. The three-dimensional scanning device includes a multi-angle camera device and a scanning device.

[0068] According to an embodiment of the present invention, in the glasses model, multiple positioning points are set based on the target glasses product, the positioning points are marked, and the image features of the marked points are obtained, specifically:

[0069] In the glasses model, multiple positioning points are set based on the important production points of the target glasses products;

[0070] In the glasses model, an origin is randomly set and three-dimensional coordinates are constructed;

[0071] Based on the three-dimensional coordinates, analyze the shortest distance between multiple positioning points, and add positioning points to ensure that the shortest distance between adjacent positioning points is within the preset range;

[0072] Acquire standard image feature data of multiple product parts of a target eyewear product;

[0073] Marking is performed based on all the positioning points, and the features of the positioning points are extracted from the standard image feature data to obtain the image features of the marking points.

[0074] It should be noted that the important production points are generally important points in the production of glasses or defect points that appear frequently during the production process, and can be set by the user.

[0075] The shortest distance between adjacent positioning points is the distance between a positioning point and the nearest positioning point, and the two positioning points are regarded as adjacent positioning points. The image feature of the marking point is the image feature of the positioning point.

[0076] According to an embodiment of the present invention, the target eyewear product is subjected to coating processing in multiple production cycles, and based on machine vision technology, the coating image data of the target eyewear product is collected in each cycle, specifically:

[0077] Set production cycle;

[0078] Perform coating processing on target eyewear products in multiple production cycles. During the processing, collect coating image data of target eyewear products in each cycle based on machine vision technology;

[0079] The coating image data includes image data at multiple angles.

[0080] It should be noted that the coating image data generally includes image data before and after coating processing, which is used for subsequent defect analysis.

[0081] According to an embodiment of the present invention, the feature recognition of the coating image data is performed by identifying the model and the mark point image features, and the extracted features are first compared with the mark point image features to generate image positioning information based on the glasses model, and then compared with the coating defect comparison features to generate defect recognition information, specifically:

[0082] Perform image noise reduction and standardization preprocessing on the coating image data;

[0083] Based on the Canny edge detection operator, the texture and contour features of the coating image data are extracted to obtain real-time features;

[0084] Import the real-time features and the image features of the marked points into the recognition model for the first feature comparison, calculate the similarity of the feature data through the cosine distance, determine the location of the real-time features based on the similarity, and generate the image location information corresponding to the real-time features in combination with the glasses model;

[0085] Obtain coating defect contrast features from the image database;

[0086] Import the real-time features and the coating defect comparison features into the recognition model for a second feature comparison, perform feature learning and defect recognition analysis on the two imported features, and generate corresponding defect recognition information;

[0087] Defect analysis is performed on the coating image data of multiple production cycles to obtain image positioning information and defect identification information corresponding to each production cycle.

[0088] It should be noted that the recognition model is an image recognition model based on machine learning, which has the function of feature learning and recognition. The image positioning information corresponding to the real-time feature is also based on the positioning information of the corresponding coating image data, and the positioning information includes the coordinate information in the mirror model. The image positioning information is used to determine the three-dimensional model positioning corresponding to each part of the feature in the coating image data, and then perform defect positioning analysis. Based on the similarity comparison between the real-time feature and the mark point image feature, the consistency of the real-time feature and the mark point image feature is determined, and the positioning information of the implementation feature is determined.

[0089] The coating defect contrast features include various types of defect features for contrast identification, such as film shedding, bubbles, scratches, uneven coating, etc.

[0090] According to an embodiment of the present invention, in the coating image data, each production cycle includes corresponding coating image data.

[0091] According to an embodiment of the present invention, the coordinates of defects in each production cycle are located in combination with the glasses model based on the image positioning information and the defect identification information, and the defect coordinate changes in two adjacent production cycles are analyzed, and multiple defect change vectors are generated. The defect change vectors are clustered based on DBSCAN data, and the defect distribution deviation in two adjacent production cycles is determined by the number of clustering groups, which has the following characteristics:

[0092] According to the image positioning information and defect recognition information corresponding to a production cycle, the identified defects are located in three dimensions. The coordinate positioning is performed based on the glasses model to generate defect positioning data.

[0093] The defect location data includes a plurality of defect coordinates;

[0094] Obtain defect location data of two adjacent production cycles, marked as first defect location data and second defect location data;

[0095] In the first defect location data, a defect coordinate P1 is selected, and the defect coordinates with the shortest and farthest distances from P1 are selected from the second defect location data through the glasses model, and are marked as P2 and P3 respectively;

[0096] Based on P1 and P2, P2 and P3, two vectors are generated respectively and marked as defect change vectors, defect coordinate change analysis is performed on all defect coordinates in the first defect location data, and multiple defect change vectors are generated;

[0097] Integrate multiple defect change vectors to form a vector data set, cluster the vector data set based on the DBSCAN clustering algorithm, and obtain the clustering results;

[0098] The number of cluster groups is obtained through clustering results, and the number of cluster groups is used as an indicator value for judging defect deviation to calculate the corresponding defect distribution deviation.

[0099] It should be noted that the defect identification information includes defect characteristics, defect type, and defect position marking in the coating image data.

[0100] It can be understood here that in the defect coordinate analysis, by analyzing the defect change vector, the defect distribution and distribution deviation within two adjacent cycles can be effectively determined. The DBSCAN clustering algorithm is introduced here to cluster the vector data set (i.e., defect change vector), and perform deviation analysis through the number of clustering groups. The larger the number of clustering groups, the higher the degree of discreteness of the defect change vector, that is, the greater the deviation of the defect distribution, and the range of defects that occur in processing tends to expand, and the distribution deviation can be further comprehensively calculated.

[0101] The defect distribution deviation calculation formula is as follows:

[0102] D = K × N,

[0103] Among them, D is the defect distribution deviation, K is the correction coefficient, and N is the number of clustering groups.

[0104] According to an embodiment of the present invention, the defect distribution deviations of all two adjacent production cycles are calculated and analyzed, and the overall defect tracking and evaluation of the coating production is performed through the defect distribution deviations, and the production regulation and optimization of the coating production is performed based on the defect identification information, specifically:

[0105] Obtain coating production equipment and environmental parameters for each production cycle;

[0106] Calculate and analyze the defect distribution deviation between all two adjacent production cycles, evaluate the change of defect distribution deviation based on the time dimension, combine the coating production equipment and environmental parameters, conduct overall defect tracking and equipment production quality evaluation for coating production, and generate an evaluation report;

[0107] Control and optimize coating production based on evaluation reports and defect identification information;

[0108] Set the defect distribution deviation warning value, and use the warning value to judge in real time whether the defect distribution of the production cycle meets expectations, and issue a real-time warning.

[0109] It should be noted that the defect distribution deviation warning value is used to compare with the defect distribution deviation degree and realize warning analysis.

[0110] It is worth mentioning here that in the prior art, the defect analysis of coating is often based on worker inspection, coating equipment and process analysis, and the optimization scheme is poor, and there is no overall evaluation of the distribution characteristics of the defects, which makes it difficult to improve production efficiency. Based on this, the present invention models and locates the target eyewear products in the form of three-dimensional coordinates, obtains the corresponding processed coating images based on different production cycles, and locates and analyzes the processing defects, generates corresponding defect coordinates, and performs multi-cycle defect distribution feature analysis and defect distribution deviation analysis through defect coordinates, effectively analyzes the distribution changes of defects and evaluates the processing from a holistic perspective, realizes accurate processing defect status distribution analysis and process evaluation, and further optimizes the coating equipment and processing technology.

[0111] The coating production equipment and environmental parameters include temperature, humidity, vacuum, cleanliness, coating deposition rate, spraying equipment power, coating processing time and other parameters.

[0112] According to an embodiment of the present invention, it also includes:

[0113] Obtain defect identification information and defect change vector for each production cycle;

[0114] Obtain defect characteristics through defect identification information;

[0115] Based on the three-dimensional space coordinates, the defect change vectors are sorted to form a distribution feature set;

[0116] Associating the distribution feature set with the defect feature to form an associated feature set;

[0117] Based on the time dimension, the associated feature set of each production cycle is serialized to form feature serialization data;

[0118] Build a LSTM-based prediction model and initialize the model;

[0119] Import the feature serialized data into the prediction model for sequence prediction, set the prediction sequence length, and obtain the prediction output sequence;

[0120] Perform data analysis on the predicted output sequence to generate predicted defect distribution information;

[0121] Carry out production early warning assessment and production control based on predicted defect distribution information.

[0122] It should be noted that the defect change vector of the first production cycle is the defect change vector obtained by analyzing it and the second production cycle (adjacent cycle). In the prediction analysis, the last cycle is not analyzed. The defect change vectors are sorted based on the three-dimensional space coordinates. The user can set the sorting rules, such as based on the length or direction of the vector. The sorting is used to form regular serialized data to improve the accuracy of subsequent data prediction.

[0123] It is worth mentioning that traditional defect prediction is often based on the number of defects, with limited prediction accuracy and effect. The present invention serializes dual features and performs prediction based on the LSTM model. The dual features include coordinate-based distribution features and image-based defect features (i.e., distribution feature sets and defect features). This can effectively evaluate the distribution and predict the distribution characteristics of coating production defects, thereby improving the prediction effect based on holistic analysis. Compared with traditional technologies, the present invention has a better guiding role in production regulation and effectively improves production efficiency.

[0124] Figure 2 A block diagram of a defect tracking system for eyeglass coating according to the present invention is shown.

[0125] The second aspect of the present invention further provides a defect tracking system 2 for eyeglass coating, the system comprising: a memory 21, a processor 22, the memory 21 comprising a defect tracking program for eyeglass coating, the defect tracking program for eyeglass coating when executed by the processor 22 implements the following steps:

[0126] Using a 3D scanning device, a 3D image scan of the target eyewear product is performed and a 3D eyewear model is constructed;

[0127] In the glasses model, multiple positioning points are set based on the target glasses product, the positioning points are marked, and the image features of the marked points are obtained;

[0128] The target eyewear products are coated in multiple production cycles. Based on machine vision technology, the coating image data of the target eyewear products are collected in each cycle.

[0129] By identifying the model and the features of the mark point image, the coating image data is subjected to feature recognition. The extracted features are first compared with the features of the mark point image to generate image positioning information based on the glasses model, and then compared with the coating defect comparison features to generate defect recognition information.

[0130] According to the image positioning information and defect recognition information, the defects in each production cycle are located in coordinates in combination with the eyewear model, the defect coordinate changes in two adjacent production cycles are analyzed, and multiple defect change vectors are generated. The defect change vectors are clustered based on DBSCAN data, and the defect distribution deviation in two adjacent production cycles is determined by the number of clustering groups;

[0131] The defect distribution deviation between all two adjacent production cycles is calculated and analyzed, and the overall defect tracking and evaluation of coating production is carried out through the defect distribution deviation. Based on the defect identification information, the coating production is regulated and optimized.

[0132] According to an embodiment of the present invention, the three-dimensional scanning device is used to perform three-dimensional image scanning on the target eyewear product and construct a three-dimensional eyewear model, specifically:

[0133] The target eyewear products are scanned in three dimensions at multiple angles using three-dimensional scanning equipment. The three-dimensional structure is analyzed through the scanned images, and a three-dimensional eyewear model is constructed.

[0134] It should be noted that the target eyewear product is generally an unprocessed product, which is used to construct an original three-dimensional model, and the model is used for subsequent defect location and deviation analysis. The three-dimensional scanning device includes a multi-angle camera device and a scanning device.

[0135] According to an embodiment of the present invention, in the glasses model, multiple positioning points are set based on the target glasses product, the positioning points are marked, and the image features of the marked points are obtained, specifically:

[0136] In the glasses model, multiple positioning points are set based on the important production locations of the target glasses products;

[0137] In the glasses model, an origin is randomly set and three-dimensional coordinates are constructed;

[0138] Based on the three-dimensional coordinates, analyze the shortest distance between multiple positioning points, and add positioning points to ensure that the shortest distance between adjacent positioning points is within the preset range;

[0139] Acquire standard image feature data of multiple product parts of a target eyewear product;

[0140] Marking is performed based on all the positioning points, and the features of the positioning points are extracted from the standard image feature data to obtain the image features of the marking points.

[0141] It should be noted that the important production points are generally important points in the production of glasses or defect points that appear frequently during the production process, and can be set by the user.

[0142] The shortest distance between adjacent positioning points is the distance between a positioning point and the nearest positioning point, and the two positioning points are regarded as adjacent positioning points. The image feature of the marking point is the image feature of the positioning point.

[0143] According to an embodiment of the present invention, the target eyewear product is subjected to coating processing in multiple production cycles, and based on machine vision technology, the coating image data of the target eyewear product is collected in each cycle, specifically:

[0144] Set production cycle;

[0145] Perform coating processing on target eyewear products in multiple production cycles. During the processing, collect coating image data of target eyewear products in each cycle based on machine vision technology;

[0146] The coating image data includes image data at multiple angles.

[0147] It should be noted that the coating image data generally includes image data before and after coating processing, which is used for subsequent defect analysis.

[0148] According to an embodiment of the present invention, the feature recognition of the coating image data is performed by identifying the model and the mark point image features, and the extracted features are first compared with the mark point image features to generate image positioning information based on the glasses model, and then compared with the coating defect comparison features to generate defect recognition information, specifically:

[0149] Perform image noise reduction and standardization preprocessing on the coating image data;

[0150] Based on the Canny edge detection operator, the texture and contour features of the coating image data are extracted to obtain real-time features;

[0151] Import the real-time features and the image features of the marked points into the recognition model for the first feature comparison, calculate the similarity of the feature data through the cosine distance, determine the location of the real-time features based on the similarity, and generate the image location information corresponding to the real-time features in combination with the glasses model;

[0152] Obtain coating defect contrast features from the image database;

[0153] Import the real-time features and the coating defect comparison features into the recognition model for a second feature comparison, perform feature learning and defect recognition analysis on the two imported features, and generate corresponding defect recognition information;

[0154] Defect analysis is performed on the coating image data of multiple production cycles to obtain image positioning information and defect identification information corresponding to each production cycle.

[0155] It should be noted that the recognition model is an image recognition model based on machine learning, which has the function of feature learning and recognition. The image positioning information corresponding to the real-time feature is also based on the positioning information of the corresponding coating image data, and the positioning information includes the coordinate information in the mirror model. The image positioning information is used to determine the three-dimensional model positioning corresponding to each part of the feature in the coating image data, and then perform defect positioning analysis. Based on the similarity comparison between the real-time feature and the mark point image feature, the consistency of the real-time feature and the mark point image feature is determined, and the positioning information of the implementation feature is determined.

[0156] The coating defect contrast features include various types of defect features for contrast identification, such as film shedding, bubbles, scratches, uneven coating, etc.

[0157] According to an embodiment of the present invention, in the coating image data, each production cycle includes corresponding coating image data.

[0158] According to an embodiment of the present invention, the coordinates of defects in each production cycle are located in combination with the glasses model based on the image positioning information and the defect identification information, and the defect coordinate changes in two adjacent production cycles are analyzed, and multiple defect change vectors are generated. The defect change vectors are clustered based on DBSCAN data, and the defect distribution deviation in two adjacent production cycles is determined by the number of clustering groups, which has the following characteristics:

[0159] According to the image positioning information and defect recognition information corresponding to a production cycle, the identified defects are located in three dimensions. The coordinate positioning is performed based on the glasses model to generate defect positioning data.

[0160] The defect location data includes a plurality of defect coordinates;

[0161] Obtain defect location data of two adjacent production cycles, marked as first defect location data and second defect location data;

[0162] In the first defect location data, a defect coordinate P1 is selected, and the defect coordinates with the shortest and farthest distances from P1 are selected from the second defect location data through the glasses model, and are marked as P2 and P3 respectively;

[0163] Based on P1 and P2, P2 and P3, two vectors are generated respectively and marked as defect change vectors, defect coordinate change analysis is performed on all defect coordinates in the first defect location data, and multiple defect change vectors are generated;

[0164] Integrate multiple defect change vectors to form a vector data set, cluster the vector data set based on the DBSCAN clustering algorithm, and obtain the clustering results;

[0165] The number of cluster groups is obtained through clustering results, and the number of cluster groups is used as an indicator value for judging defect deviation to calculate the corresponding defect distribution deviation.

[0166] It should be noted that the defect identification information includes defect characteristics, defect type, and defect position marking in the coating image data.

[0167] It can be understood here that in the defect coordinate analysis, by analyzing the defect change vector, the defect distribution and distribution deviation within two adjacent cycles can be effectively determined. The DBSCAN clustering algorithm is introduced here to cluster the vector data set (i.e., defect change vector), and perform deviation analysis through the number of clustering groups. The larger the number of clustering groups, the higher the degree of discreteness of the defect change vector, that is, the greater the deviation of the defect distribution, and the range of defects that occur in processing tends to expand, and the distribution deviation can be further comprehensively calculated.

[0168] The defect distribution deviation calculation formula is as follows:

[0169] D = K × N,

[0170] Among them, D is the defect distribution deviation, K is the correction coefficient, and N is the number of clustering groups.

[0171] According to an embodiment of the present invention, the defect distribution deviations of all two adjacent production cycles are calculated and analyzed, and the overall defect tracking and evaluation of the coating production is performed through the defect distribution deviations, and the production regulation and optimization of the coating production is performed based on the defect identification information, specifically:

[0172] Obtain coating production equipment and environmental parameters for each production cycle;

[0173] Calculate and analyze the defect distribution deviation between all two adjacent production cycles, evaluate the change of defect distribution deviation based on the time dimension, combine the coating production equipment and environmental parameters, conduct overall defect tracking and equipment production quality evaluation for coating production, and generate an evaluation report;

[0174] Control and optimize coating production based on evaluation reports and defect identification information;

[0175] Set the defect distribution deviation warning value, and use the warning value to judge in real time whether the defect distribution of the production cycle meets expectations, and issue a real-time warning.

[0176] It should be noted that the defect distribution deviation warning value is used to compare with the defect distribution deviation degree and realize warning analysis.

[0177] It is worth mentioning here that in the prior art, the defect analysis of coating is often based on worker inspection, coating equipment and process analysis, and the optimization scheme is poor, and there is no overall evaluation of the distribution characteristics of the defects, which makes it difficult to improve production efficiency. Based on this, the present invention models and locates the target eyewear products in the form of three-dimensional coordinates, obtains the corresponding processed coating images based on different production cycles, and locates and analyzes the processing defects, generates corresponding defect coordinates, and performs multi-cycle defect distribution feature analysis and defect distribution deviation analysis through defect coordinates, effectively analyzes the distribution changes of defects and evaluates the processing from a holistic perspective, realizes accurate processing defect status distribution analysis and process evaluation, and further optimizes the coating equipment and processing technology.

[0178] The coating production equipment and environmental parameters include temperature, humidity, vacuum, cleanliness, coating deposition rate, spraying equipment power, coating processing time and other parameters.

[0179] The third aspect of the present invention also provides a computer-readable storage medium, which includes a defect tracking program for eyeglass coating. When the defect tracking program for eyeglass coating is executed by a processor, the steps of the defect tracking method for eyeglass coating as described in any one of the above items are implemented.

[0180] The present invention discloses a defect tracking method and system for eyeglass coating. A three-dimensional model of an eyeglass product is constructed by a three-dimensional scanning device, and a plurality of positioning points are set in the model. In a plurality of production cycles, the coating image data is collected by using machine vision technology. The image data is feature recognized by using the recognition model and the positioning point features, and image positioning information and defect recognition information are generated. In combination with the eyeglass model, the defect coordinates of each cycle are located, and the defect coordinate changes are analyzed to generate defect change vectors. The defect change vectors are clustered by DBSCAN, the defect distribution deviation is calculated, and the coating process is controlled and optimized based on the defect recognition information. Through the present invention, the overall defect tracking and evaluation of coating production can be achieved, and the production efficiency can be further improved.

[0181] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0182] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0183] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0184] A person of ordinary skill in the art can understand that: all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0185] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0186] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A defect tracking method for eyeglass coating, characterized in that: include: Using a 3D scanning device, a 3D image scan of the target eyewear product is performed and a 3D eyewear model is constructed; In the glasses model, multiple positioning points are set based on the target glasses product, the positioning points are marked, and the image features of the marked points are obtained; The target eyewear products are coated in multiple production cycles. Based on machine vision technology, the coating image data of the target eyewear products are collected in each cycle. By identifying the model and the features of the mark point image, the coating image data is subjected to feature recognition. The extracted features are first compared with the features of the mark point image to generate image positioning information based on the glasses model, and then compared with the coating defect comparison features to generate defect recognition information. According to the image positioning information and defect recognition information, the defects in each production cycle are located in coordinates in combination with the eyewear model, the defect coordinate changes in two adjacent production cycles are analyzed, and multiple defect change vectors are generated. The defect change vectors are clustered based on DBSCAN data, and the defect distribution deviation in two adjacent production cycles is determined by the number of clustering groups; The defect distribution deviation between all two adjacent production cycles is calculated and analyzed, and the overall defect tracking and evaluation of coating production is carried out through the defect distribution deviation. Based on the defect identification information, the coating production is regulated and optimized.

2. A defect tracking method for eyeglass coating according to claim 1, characterized in that: The method of performing a three-dimensional image scanning on the target eyewear product and constructing a three-dimensional eyewear model by using a three-dimensional scanning device is specifically as follows: The target eyewear product is scanned in three dimensions at multiple angles using three-dimensional scanning equipment. The three-dimensional structure is analyzed through the scanned image, and a three-dimensional eyewear model is constructed.

3. A defect tracking method for eyeglass coating according to claim 1, characterized in that: In the glasses model, multiple positioning points are set based on the target glasses product, the positioning points are marked, and the image features of the marked points are obtained, specifically: In the glasses model, multiple positioning points are set based on the important production points of the target glasses products; In the glasses model, an origin is randomly set and three-dimensional coordinates are constructed; Based on the three-dimensional coordinates, analyze the shortest distance between multiple positioning points, and add positioning points to ensure that the shortest distance between adjacent positioning points is within the preset range; Acquire standard image feature data of multiple product parts of a target eyewear product; Marking is performed based on all the positioning points, and the features of the positioning points are extracted from the standard image feature data to obtain the image features of the marking points.

4. A defect tracking method for eyeglass coating according to claim 3, characterized in that: In the above, the target eyewear products are subjected to coating processing in multiple production cycles. Based on machine vision technology, the coating image data of the target eyewear products are collected in each cycle, specifically: Set production cycle; Perform coating processing on target eyewear products in multiple production cycles. During the processing, collect coating image data of target eyewear products in each cycle based on machine vision technology; The coating image data includes image data at multiple angles.

5. A defect tracking method for eyeglass coating according to claim 4, characterized in that: The feature recognition of the coating image data is performed by identifying the model and the mark point image features, and the extracted features are compared with the mark point image features for the first time to generate image positioning information based on the glasses model, and the second time with the coating defect comparison features to generate defect recognition information, specifically: Perform image noise reduction and standardization preprocessing on the coating image data; Based on the Canny edge detection operator, the texture and contour features of the coating image data are extracted to obtain real-time features; Import the real-time features and the image features of the marked points into the recognition model for the first feature comparison, calculate the similarity of the feature data through the cosine distance, determine the location of the real-time features based on the similarity, and generate the image location information corresponding to the real-time features in combination with the glasses model; Obtain coating defect contrast features from the image database; Import the real-time features and the coating defect comparison features into the recognition model for a second feature comparison, perform feature learning and defect recognition analysis on the two imported features, and generate corresponding defect recognition information; Defect analysis is performed on the coating image data of multiple production cycles to obtain image positioning information and defect identification information corresponding to each production cycle.

6. A defect tracking method for eyeglass coating according to claim 5, characterized in that: In the coating image data, each production cycle includes corresponding coating image data.

7. A defect tracking method for eyeglass coating according to claim 6, characterized in that: According to the image positioning information and defect recognition information, the defects in each production cycle are located in coordinates in combination with the glasses model, the defect coordinate changes in two adjacent production cycles are analyzed, and multiple defect change vectors are generated. The defect change vectors are clustered based on DBSCAN data, and the defect distribution deviation in two adjacent production cycles is determined by the number of clustering groups. According to the image positioning information and defect recognition information corresponding to a production cycle, the identified defects are located in three dimensions. The coordinate positioning is performed based on the glasses model to generate defect positioning data. The defect location data includes a plurality of defect coordinates; Obtain defect location data of two adjacent production cycles, marked as first defect location data and second defect location data; In the first defect location data, a defect coordinate P1 is selected, and the defect coordinates with the shortest and farthest distances from P1 are selected from the second defect location data through the glasses model, and are marked as P2 and P3 respectively; Based on P1 and P2, P2 and P3, two vectors are generated respectively and marked as defect change vectors, defect coordinate change analysis is performed on all defect coordinates in the first defect location data, and multiple defect change vectors are generated; Integrate multiple defect change vectors to form a vector data set, cluster the vector data set based on the DBSCAN clustering algorithm, and obtain the clustering results; The number of cluster groups is obtained through clustering results, and the number of cluster groups is used as an indicator value for judging defect deviation to calculate the corresponding defect distribution deviation.

8. A defect tracking method for eyeglass coating according to claim 7, characterized in that: The calculation and analysis of the defect distribution deviation of all two adjacent production cycles is used to track and evaluate the overall defects of the coating production through the defect distribution deviation, and the production control and optimization of the coating production are carried out based on the defect identification information, specifically: Obtain coating production equipment and environmental parameters for each production cycle; Calculate and analyze the defect distribution deviation between all two adjacent production cycles, evaluate the change of defect distribution deviation based on the time dimension, combine the coating production equipment and environmental parameters, conduct overall defect tracking and equipment production quality evaluation for coating production, and generate an evaluation report; Control and optimize coating production based on evaluation reports and defect identification information; Set the defect distribution deviation warning value, and use the warning value to judge in real time whether the defect distribution of the production cycle meets expectations, and issue a real-time warning.

9. A defect tracking system for eyeglass coating, characterized in that: The system includes: a memory and a processor. The memory includes a defect tracking program for eyeglass coating. When the defect tracking program for eyeglass coating is executed by the processor, the following steps are implemented: Using a 3D scanning device, a 3D image scan of the target eyewear product is performed and a 3D eyewear model is constructed; In the glasses model, multiple positioning points are set based on the target glasses product, the positioning points are marked, and the image features of the marked points are obtained; The target eyewear products are coated in multiple production cycles. Based on machine vision technology, the coating image data of the target eyewear products are collected in each cycle. By identifying the model and the features of the mark point image, the coating image data is subjected to feature recognition. The extracted features are first compared with the features of the mark point image to generate image positioning information based on the glasses model, and then compared with the coating defect comparison features to generate defect recognition information. According to the image positioning information and defect recognition information, the defects in each production cycle are located in coordinates in combination with the eyewear model, the defect coordinate changes in two adjacent production cycles are analyzed, and multiple defect change vectors are generated. The defect change vectors are clustered based on DBSCAN data, and the defect distribution deviation in two adjacent production cycles is determined by the number of clustering groups; The defect distribution deviation between all two adjacent production cycles is calculated and analyzed, and the overall defect tracking and evaluation of coating production is carried out through the defect distribution deviation. Based on the defect identification information, the coating production is regulated and optimized.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a defect tracking program for eyeglass coating, and when the defect tracking program for eyeglass coating is executed by a processor, the steps of the defect tracking method for eyeglass coating as described in any one of claims 1 to 8 are implemented.

Images