Lens coating defect identification system and method

By acquiring the lens surface image and optical detection parameters and calculating the defect evaluation coefficient, the misjudgment problem caused by uneven coating thickness in the prior art is solved, the recognition accuracy of lens coating is improved, and waste and cost are reduced.

CN118961709AInactive Publication Date: 2024-11-15ANHUI DACHENG SMART GLASS CO LTD
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202410984068.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-22
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

When the existing lens coating automation identification system deals with uneven coating thickness, it is easy to misjudgment as a defect, resulting in a large number of lenses being wasted and the cost is increased.

Method used

By acquiring the lens surface image, the number of defect areas and the proportion of defects are calculated, and the defect evaluation coefficient is calculated based on optical detection parameters to determine whether the lens coating is available to avoid misjudgment.

Benefits of technology

It improves the recognition accuracy of lens coating, reduces unnecessary lens waste, and reduces costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118961709A_ABST
    Figure CN118961709A_ABST
Patent Text Reader

Abstract

The invention discloses a lens coating defect identification system and method. The system comprises a data acquisition module, an optical detection module and a data analysis module, relates to the technical field of lens coating identification, and solves the technical problems that when few defects exist, the use of a lens coating may not be affected, and a large number of coated lenses are wasted by an existing identification system. The number of defect areas on the surface and the defect ratio are obtained according to the surface image, whether the lens coating has defects is preliminarily judged according to the surface evaluation parameters, and when the judgment result is that the defects exist, whether the coated lens is available or not is specifically judged according to the defect evaluation coefficient and through the defect evaluation coefficient; by means of the mode, whether the coating film on the surface of the lens can cause defects of data such as pictures or not in the actual application process can be more accurately judged, the defects of the coating film on the surface of the lens in the actual application process are judged in the mode, and the cost of the coated lens can be reduced to a certain degree.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of lens coating recognition, and in particular to a lens coating defect recognition system and method. Background Art

[0002] In the production process of optical lenses, coating is one of the key steps to improve lens performance. Coating can not only enhance the light transmittance of the lens and reduce reflection, but also provide multiple functions such as anti-ultraviolet, anti-blue light, and anti-oil. However, during the coating process, due to the influence of various factors, such as equipment accuracy, environmental conditions, and operating processes, various defects may appear on the lens surface, such as scratches, color aberration, uneven thickness, etc.

[0003] With the development of computer vision technology and artificial intelligence algorithms, automated lens defect recognition systems have greatly improved detection efficiency and accuracy. However, existing automated recognition systems have certain limitations when dealing with complex and changeable coating defects. For example, when uneven coating thickness appears on the identified surface, the lens coating is directly marked as defective, resulting in the coated lens being directly marked as unqualified. However, when the defects are rare, they may not affect the use of the lens coating. The current recognition system will cause a large number of coated lenses to be wasted, resulting in a significant increase in the cost of coated lenses. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a lens coating defect recognition system and method for solving the technical problem that when the defects are rare and may not affect the use of the lens coating, the current recognition system will cause a large number of coated lenses to be wasted. The present invention obtains the number and defect proportion of defective areas on the surface according to the surface image, and preliminarily determines whether the lens coating has defects based on the surface evaluation parameters. When the judgment result is that there are defects, the defect evaluation coefficient is used to specifically determine whether the coated lens is usable, thereby solving the above-mentioned problem.

[0005] To achieve the above-mentioned object, the first aspect of the present invention provides a lens coating defect recognition system, comprising: a data acquisition module and an optical detection module and a data analysis module connected thereto;

[0006] The optical detection module is used to detect the lens coating to obtain detection parameters; wherein the detection parameters are the test brightness, reference brightness, and brightness test difference and reference brightness difference of several areas;

[0007] The data acquisition module is used to acquire a surface image of the coated lens; and acquire surface evaluation parameters according to the surface image; wherein the surface evaluation parameters include the number of defective areas and the defect ratio;

[0008] The data analysis module calculates the defect assessment coefficient according to the detection parameters; and

[0009] A preliminary judgement is made based on the surface evaluation parameters as to whether the lens coating has defects; if so, the lens coating is marked as defective; if not, an evaluation result is obtained based on the defect evaluation coefficient; wherein the evaluation results include unqualified and qualified.

[0010] Preferably, obtaining the surface evaluation parameters according to the surface image includes:

[0011] Extracting a surface image, identifying RGB color values ​​of a number of pixels on the surface image, and inputting the surface image into a coating image recognition model to obtain a coating image; wherein the coating image recognition model is established based on a convolutional neural network model;

[0012] Determine whether the RGB color values ​​of several pixels in the coating image are within the standard color value range; if yes, mark the corresponding pixels as standard pixels; if no, mark the corresponding pixels as defective pixels;

[0013] Determine whether the pixels adjacent to the defective pixel are defective pixels; if yes, mark the area occupied by the adjacent pixels as defective pixels as defective areas, count the number of defective areas and the defect ratio of defective areas in the coating image; if no, mark the defective pixels as transition pixels.

[0014] It should be noted that the standard color value range is the specific RGB color value range determined by experts in this field according to the type and purpose of the coating.

[0015] By identifying the RGB color value of the pixel, it can directly reflect whether there is any abnormality in the coating effect on the lens surface. That is, when there are scratches, uneven coating, bubbles, etc. on the surface, the RGB color value at that location will not overlap with the standard color value range. Therefore, this method is not only relatively simple, but also has a higher recognition accuracy.

[0016] Preferably, the coating image recognition model is established based on a convolutional neural network model, including:

[0017] Surface images and coating images are obtained from a database; the surface images and corresponding coating images are integrated into a number of training data groups and test data groups; a convolutional neural network model is trained using the surface images and coating images in the training data group, and the trained convolutional neural network model is then tested using the surface images and coating images in the test data group, and the trained convolutional neural network model is adjusted according to the test results to obtain a coating image recognition model whose input is a surface image and whose output is a coating image.

[0018] Preferably, the preliminarily judging whether the lens coating has defects according to the surface evaluation parameters includes:

[0019] Extract the number of defective areas and the defect proportion in the evaluation parameters, and compare the number of defective areas with the quantity threshold;

[0020] If the number of defective areas is greater than the quantity threshold, the lens coating corresponding to the set of evaluation parameters is marked as defective;

[0021] Otherwise, determine whether the defect ratio of the defective area is greater than the ratio threshold; if yes, mark the lens coating corresponding to the set of evaluation parameters as defective; if no, mark the lens coating corresponding to the set of evaluation parameters as not defective.

[0022] The defect area and defect ratio on the surface can be used to directly and simply quickly evaluate the coating condition.

[0023] Preferably, the detecting of the lens coating to obtain the detection parameters comprises:

[0024] Divide the illumination of the coated lens and the illumination of the uncoated lens into a number of detection areas of equal area, and obtain the test brightness LJi and the reference brightness LZi of the several detection areas respectively;

[0025] The brightness test difference LJc and the reference brightness difference LZc are calculated by the calculation formulas LJc=LJmax-LJmin and LZc=LZmax-LZmin, respectively; wherein LJmax and LZmax are the maximum test brightness in several detection areas when the lens is illuminated by the coated lens and the lens is not illuminated by the coated lens, respectively, LJmin and LZmin are the minimum measured brightness in several detection areas, and i is a positive integer.

[0026] Preferably, the calculating of the defect assessment coefficient according to the detection parameters includes:

[0027] Extract the test brightness, reference brightness, brightness test difference and reference brightness difference of several areas in the detection parameters, and use the formula Calculate the defect assessment coefficient; wherein α and β are preset proportional coefficients, n is a positive integer, and n represents the number of detection areas.

[0028] It should be noted that the higher the defect assessment coefficient is, the greater the defect problem is, and vice versa.

[0029] Preferably, obtaining the evaluation result based on the defect evaluation coefficient includes:

[0030] The defect assessment coefficient is compared with the coefficient threshold. If the defect assessment coefficient is greater than the coefficient threshold, the assessment result of the coated lens is marked as unqualified; otherwise, the assessment result of the coated lens is marked as qualified.

[0031] The present invention also provides a method for identifying lens coating defects, comprising the following steps:

[0032] Step 1: Obtain a surface image of the coated lens, and obtain surface evaluation parameters based on the surface image;

[0033] Step 2: Preliminarily determine whether the lens coating has defects based on the surface evaluation parameters; if yes, mark the lens coating as defective; if no, jump to step 3; the evaluation results include unqualified and qualified;

[0034] Step 3: The coating is inspected to obtain inspection parameters, defect assessment coefficients are calculated according to the inspection parameters, and assessment results are obtained based on the defect assessment coefficients.

[0035] Compared with the prior art, the beneficial effects of the present invention are as follows: first, a surface image of a coated lens is obtained, and the number and defect ratio of defective areas on the surface are obtained based on the surface image, and a preliminary judgment is made based on the surface evaluation parameters whether the lens coating has defects; when the judgment result is that defects exist, the lens coating can be tested again, and a defect evaluation coefficient can be obtained based on the test brightness, reference brightness, brightness test difference, and reference brightness difference of several obtained areas, and then a specific judgment is made based on the defect evaluation coefficient whether the coated lens is usable; when the above method is used, it can be more accurately judged whether the coating on the lens surface will cause defects in data such as images when actually used; and by judging the defects in the actual application of the lens coating in this way, the cost of the coated lens can be reduced to a certain extent. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 A schematic diagram of the process of identifying coating defects of lenses of the present invention;

[0038] Figure 2 A schematic diagram of a process for obtaining surface evaluation parameters of the present invention;

[0039] Figure 3 It is a connection block diagram of the system of the present invention. DETAILED DESCRIPTION

[0040] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than 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.

[0041] See also Figure 1-Figure 3 , a first aspect of the present invention provides a lens coating defect recognition system, comprising: a data acquisition module and an optical detection module and a data analysis module connected thereto;

[0042] The optical detection module is used to detect the lens coating to obtain detection parameters; wherein the detection parameters are the test brightness, reference brightness, and brightness test difference and reference brightness difference of several areas;

[0043] The data acquisition module is used to acquire a surface image of the coated lens; and acquire surface evaluation parameters according to the surface image; wherein the surface evaluation parameters include the number of defective areas and the defect ratio;

[0044] The data analysis module calculates the defect assessment coefficient according to the detection parameters; and

[0045] A preliminary judgement is made based on the surface evaluation parameters as to whether the lens coating has defects; if so, the lens coating is marked as defective; if not, an evaluation result is obtained based on the defect evaluation coefficient; wherein the evaluation results include unqualified and qualified.

[0046] As a first step of the embodiment of the present invention: obtaining surface evaluation parameters according to the surface image includes:

[0047] Extracting a surface image, identifying RGB color values ​​of a number of pixels on the surface image, and inputting the surface image into a coating image recognition model to obtain a coating image; wherein the coating image recognition model is established based on a convolutional neural network model;

[0048] Determine whether the RGB color values ​​of several pixels in the coating image are within the standard color value range; if yes, mark the corresponding pixels as standard pixels; if no, mark the corresponding pixels as defective pixels;

[0049] Determine whether the pixels adjacent to the defective pixel are defective pixels; if yes, mark the area occupied by the adjacent pixels as defective pixels as defective areas, count the number of defective areas and the defect ratio of defective areas in the coating image; if no, mark the defective pixels as transition pixels.

[0050] Exemplarily, first, a high-resolution camera is used to photograph the lens to obtain its surface image, and then the image is converted into an RGB color mode to facilitate subsequent pixel color recognition.

[0051] Identify the RGB color value of each pixel, and input the surface image into a pre-trained coating image recognition model to obtain a coating image reflecting the coating effect of the lens.

[0052] Experts in the field of optics have defined the standard color value ranges of R: 100-150, G: 80-120, and B: 70-110 based on the coating characteristics of this polarized lens.

[0053] Each pixel on the coating image will be checked one by one to determine whether its RGB color value is within the above standard color value range. Pixels within the standard range will be marked as standard pixels; pixels outside the range will be marked as defective pixels;

[0054] Then check all the points marked as "defective pixels" to determine whether their adjacent pixels are also marked as defective pixels. If so, these adjacent defective pixels will be regarded as a whole to form a "defective area". And count the number of all defective areas and calculate the proportion of defective areas in the entire coating image, that is, the defect ratio.

[0055] Based on the number of defective areas and the defect ratio, the system will make a preliminary judgment on whether the lens coating has quality problems. If there are too many defective areas or the defect ratio is too high, the lens is defective.

[0056] Furthermore, the coating image recognition model is established based on the convolutional neural network model, including:

[0057] Surface images and coating images are obtained from a database; the surface images and corresponding coating images are integrated into a number of training data groups and test data groups; a convolutional neural network model is trained using the surface images and coating images in the training data group, and the trained convolutional neural network model is then tested using the surface images and coating images in the test data group, and the trained convolutional neural network model is adjusted according to the test results to obtain a coating image recognition model whose input is a surface image and whose output is a coating image.

[0058] As the second step of this embodiment, a preliminary judgment is made as to whether the lens coating has defects based on the surface evaluation parameters, including:

[0059] Extract the number of defective areas and the defect proportion in the evaluation parameters, and compare the number of defective areas with the quantity threshold;

[0060] If the number of defective areas is greater than the quantity threshold, the lens coating corresponding to the set of evaluation parameters is marked as defective;

[0061] Otherwise, determine whether the defect ratio of the defective area is greater than the ratio threshold; if yes, mark the lens coating corresponding to the set of evaluation parameters as defective; if no, mark the lens coating corresponding to the set of evaluation parameters as not defective.

[0062] The defect area and defect ratio on the surface can be used to directly and simply quickly evaluate the coating condition.

[0063] For example, the factory has set two key thresholds based on past experience and product requirements - the quantity threshold and the percentage threshold. The quantity threshold is set to 5, which means that the maximum number of defective areas allowed on a single lens is 5; the percentage threshold is set to 0.1%, that is, the total area of ​​defective areas on a single lens cannot exceed 0.1% of the total area of ​​the lens. If 7 defective areas are found and the defect percentage reaches 0.2%, the lens coating will be marked as defective.

[0064] As the third step in this embodiment, the lens coating is tested to obtain test parameters, including:

[0065] Divide the illumination of the coated lens and the illumination of the uncoated lens into a number of detection areas of equal area, and obtain the test brightness LJi and the reference brightness LZi of the several detection areas respectively;

[0066] The brightness test difference LJc and the reference brightness difference LZc are calculated by the calculation formulas LJc=LJmax-LJmin and LZc=LZmax-LZmin, respectively; wherein LJmax and LZmax are the maximum test brightness in several detection areas when the lens is illuminated by the coated lens and the lens is not illuminated by the coated lens, respectively, LJmin and LZmin are the minimum measured brightness in several detection areas, and i is a positive integer.

[0067] Exemplarily, the brightness value of the brightest area in the test sample is LJmax=800lux, and the brightness value of the darkest area is LJmin=600lux; while the brightness value of the brightest area in the control group is LZmax=1000lux, and the brightness value of the darkest area is LZmin=800lux.

[0068] Furthermore, the defect assessment coefficient is calculated based on the detection parameters, including:

[0069] Extract the test brightness, reference brightness, brightness test difference and reference brightness difference of several areas in the detection parameters, and substitute the measured test brightness and reference brightness of several areas and the calculated brightness test difference and reference brightness difference into the formula Calculate the defect assessment coefficient; where α=0.6, β=0.4, and n is the number of inspection areas.

[0070] It should be noted that the higher the defect assessment coefficient is, the greater the defect problem is, and vice versa.

[0071] Specifically, the evaluation results are obtained based on the defect evaluation coefficient, including:

[0072] The defect assessment coefficient is compared with the coefficient threshold. If the defect assessment coefficient is greater than the coefficient threshold, the assessment result of the coated lens is marked as unqualified; otherwise, the assessment result of the coated lens is marked as qualified.

[0073] The present invention also provides a method for identifying lens coating defects, comprising the following steps:

[0074] Step 1: Obtain a surface image of the coated lens, and obtain surface evaluation parameters based on the surface image;

[0075] Step 2: Preliminarily determine whether the lens coating has defects based on the surface evaluation parameters; if yes, mark the lens coating as defective; if no, jump to step 3; the evaluation results include unqualified and qualified;

[0076] Step 3: The coating is inspected to obtain inspection parameters, defect assessment coefficients are calculated according to the inspection parameters, and assessment results are obtained based on the defect assessment coefficients.

[0077] Part of the data in the above formula is calculated by removing the dimension and taking its numerical value. The formula is a formula closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0078] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A lens coating defect recognition system, characterized in that: include: Data acquisition module, optical detection module and data analysis module; The optical detection module is used to detect the lens coating to obtain detection parameters; wherein the detection parameters are the test brightness, reference brightness, brightness test difference and reference brightness difference of several areas; The data acquisition module is used to acquire a surface image of the coated lens; and acquire surface evaluation parameters according to the surface image; wherein the surface evaluation parameters include the number of defective areas and the defect ratio; The data analysis module calculates the defect assessment coefficient according to the detection parameters; and A preliminary judgement is made based on the surface evaluation parameters as to whether the lens coating has defects; if so, the lens coating is marked as defective; if not, an evaluation result is obtained based on the defect evaluation coefficient; wherein the evaluation results include unqualified and qualified.

2. A lens coating defect recognition system according to claim 1, characterized in that: The step of acquiring the surface evaluation parameter according to the surface image comprises: Extracting a surface image, identifying RGB color values ​​of a number of pixels on the surface image, and inputting the surface image into a coating image recognition model to obtain a coating image; wherein the coating image recognition model is established based on a convolutional neural network model; Determine whether the RGB color values ​​of several pixels in the coating image are within the standard color value range; if yes, mark the corresponding pixels as standard pixels; if no, mark the corresponding pixels as defective pixels; Determine whether the pixels adjacent to the defective pixel are defective pixels; if yes, mark the area occupied by the adjacent pixels as defective pixels as defective areas, count the number of defective areas and the defect ratio of defective areas in the coating image; if no, mark the defective pixels as transition pixels.

3. The lens coating defect recognition system according to claim 1, characterized in that: The preliminary determination of whether the lens coating has defects based on the surface evaluation parameters includes: Extract the number of defective areas and the defect proportion in the evaluation parameters, and compare the number of defective areas with the quantity threshold; If the number of defective areas is greater than the quantity threshold, the lens coating corresponding to the set of evaluation parameters is marked as defective; Otherwise, determine whether the defect ratio of the defective area is greater than the ratio threshold; if yes, mark the lens coating corresponding to the set of evaluation parameters as defective; if no, mark the lens coating corresponding to the set of evaluation parameters as not defective.

4. The lens coating defect recognition system according to claim 1, characterized in that: The method of detecting the lens coating to obtain detection parameters includes: Divide the illumination of the coated lens and the illumination of the uncoated lens into a number of detection areas of equal area, and obtain the test brightness LJi and the reference brightness LZi of the several detection areas respectively; The brightness test difference LJc and the reference brightness difference LZc are calculated by the calculation formulas LJc=LJmax-LJmin and LZc=LZmax-LZmin, respectively; wherein LJmax and LZmax are the maximum test brightness in several detection areas when the lens is illuminated by the coated lens and the lens is not illuminated by the coated lens, respectively, LJmin and LZmin are the minimum measured brightness in several detection areas, and i is a positive integer.

5. The lens coating defect recognition system according to claim 2, characterized in that: The coating image recognition model is established based on a convolutional neural network model, including: Acquire surface images and coating images from a database; integrate the surface images and corresponding coating images into a plurality of training data sets and test data sets; The convolutional neural network model is trained using the surface images and coating images in the training data set, and then the trained convolutional neural network model is tested using the surface images and coating images in the test data set. The trained convolutional neural network model is adjusted according to the test results to obtain a coating image recognition model whose input is the surface image and output is the coating image.

6. The lens coating defect recognition system according to claim 1, characterized in that: The defect assessment coefficient is calculated according to the detection parameters, including: Extract the test brightness, reference brightness, brightness test difference and reference brightness difference of several areas in the detection parameters, and use the formula Calculate the defect assessment coefficient; wherein α and β are preset proportional coefficients, n is a positive integer, and n represents the number of detection areas.

7. The lens coating defect recognition system according to claim 1, characterized in that: The obtaining of the evaluation result based on the defect evaluation coefficient includes: The defect assessment coefficient is compared with the coefficient threshold. If the defect assessment coefficient is greater than the coefficient threshold, the assessment result of the coated lens is marked as unqualified; otherwise, the assessment result of the coated lens is marked as qualified.

8. A lens coating defect recognition method according to claim 1, based on the lens coating defect recognition system according to any one of claims 1 to 7, characterized in that: The following steps are involved: Step 1: Obtain a surface image of the coated lens, and obtain surface evaluation parameters based on the surface image; Step 2: Preliminarily determine whether the lens coating has defects based on the surface evaluation parameters; if yes, mark the lens coating as defective; if no, jump to step 3; the evaluation results include unqualified and qualified; Step 3: The coating is inspected to obtain inspection parameters, defect assessment coefficients are calculated according to the inspection parameters, and assessment results are obtained based on the defect assessment coefficients.

Citation Information

Cited By

  • Defect tracking method and system for glasses coating

    CN119359706A

  • Lens coating defect intelligent identification classification and detection system

    CN122694848A