CN118781099A - Method and system for detecting microscopic defects on surface of optical element - Google Patents
Method and system for detecting microscopic defects on surface of optical element Download PDFInfo
- Publication number
- CN118781099A CN118781099A CN202411244384.1A CN202411244384A CN118781099A CN 118781099 A CN118781099 A CN 118781099A CN 202411244384 A CN202411244384 A CN 202411244384A CN 118781099 A CN118781099 A CN 118781099A
- Authority
- CN
- China
- Prior art keywords
- optical element
- suspected
- detected
- pixel point
- dark field
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000003287 optical effect Effects 0.000 title claims abstract description 152
- 230000007547 defect Effects 0.000 title claims abstract description 106
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000003708 edge detection Methods 0.000 claims abstract description 45
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 31
- 230000002950 deficient Effects 0.000 claims abstract description 24
- 238000003062 neural network model Methods 0.000 claims abstract description 15
- 238000001514 detection method Methods 0.000 claims abstract description 8
- 238000012549 training Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 6
- 230000004927 fusion Effects 0.000 claims description 4
- 238000012545 processing Methods 0.000 abstract description 13
- 230000008569 process Effects 0.000 description 13
- 238000004364 calculation method Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 238000013527 convolutional neural network Methods 0.000 description 4
- 230000001629 suppression Effects 0.000 description 4
- 238000003384 imaging method Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000002146 bilateral effect Effects 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000000877 morphologic effect Effects 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000008358 core component Substances 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000000227 grinding Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000003754 machining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000005498 polishing Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000007493 shaping process Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Image Analysis (AREA)
Abstract
The application relates to the technical field of image processing, in particular to a method and a system for detecting microscopic defects on the surface of an optical element, wherein the method comprises the following steps: collecting dark field images of the surfaces of the optical element to be detected and the standard optical element; acquiring each suspected defect pixel point of an optical element to be detected; obtaining each connected domain of the surface dark field image of the optical element to be detected based on the distribution of the suspected defective pixel points; obtaining scratch suspected degree of each connected domain; obtaining the weak scratch suspected degree of each suspected defective pixel point; and determining a gradient low threshold in an edge detection algorithm, performing edge detection on the surface dark field image of the optical element to be detected, and combining a neural network model to obtain a microscopic defect result of the surface of the optical element to be detected. Thereby improving the accuracy of microscopic defect detection on the surface of the optical element.
Description
Technical Field
The application relates to the technical field of image processing, in particular to a method and a system for detecting microscopic defects on the surface of an optical element.
Background
The quality of the processing of the optical element as a core component of the optical system directly affects the stability and accuracy of the optical system. However, in the process of precisely machining such as grinding, polishing, shaping and the like of the optical element, various microscopic defects are inevitably generated on the surface of the optical element due to the limitation of the process and the technical conditions, and the service performance and the service life of the optical element are affected.
In the process of surface defect detection of an optical element, an edge detection process is generally required to be performed on an acquired image of the surface of the optical element so as to highlight the characteristics of the defect of the surface of the optical element in the image. However, the microscopic defects on the surface of the optical element are also called weak scratches, and the gray value in the image on the surface of the optical element is close to the background, so that when the edge detection processing is performed on the image on the surface of the optical element, the weak scratches in the image are easily buried in the background and are not detected due to the fact that the gradient low threshold is set too high, background information in the image is also extracted due to the fact that the gradient low threshold is set too low, the accuracy of extracting the weak scratch information in the image on the surface of the optical element cannot be guaranteed, and finally the recognition accuracy of the microscopic defects on the surface of the subsequent optical element is affected.
Disclosure of Invention
In order to solve the technical problems, the application aims to provide a method and a system for detecting microscopic defects on the surface of an optical element, wherein the adopted technical scheme is as follows:
In a first aspect, an embodiment of the present application provides a method for detecting microscopic defects on a surface of an optical element, the method including the steps of:
Collecting dark field images of the surfaces of the optical element to be detected and the standard optical element;
acquiring a candidate edge pixel point set in the surface dark field image; obtaining each suspected defect pixel point of the optical element to be detected based on the difference of the candidate edge pixel point set of the optical element to be detected and the standard optical element; obtaining each connected domain of the surface dark field image of the optical element to be detected based on the distribution of the suspected defective pixel points;
analyzing the shape characteristics of each connected domain to obtain scratch suspected degree of each connected domain; combining the scratch suspected degree and the gray level difference between the suspected defect pixel point and a background pixel point in the image to obtain the weak scratch suspected degree of each suspected defect pixel point;
and determining a gradient low threshold in an edge detection algorithm based on the weak scratch suspected degree and the gradient of each suspected defective pixel point, carrying out edge detection on the surface dark field image of the optical element to be detected, and combining a neural network model to obtain a microscopic defect result of the surface of the optical element to be detected.
In one embodiment, the candidate edge pixel point set is all the pixel points reserved in the non-maximum value suppression step after edge detection is performed on the surface dark field image.
In one embodiment, the determining of each suspected defective pixel point includes:
Registering the surface dark field images of the optical element to be detected and the standard optical element, acquiring an intersection of the candidate edge pixel point set of the optical element to be detected and the standard optical element, and removing all the pixel points remained after the intersection in the candidate edge pixel point set of the optical element to be detected as all the suspected defect pixel points.
In one embodiment, the determining of each connected domain includes:
And converting the dark field image of the surface of the optical element to be detected into a binarized image based on all the suspected defective pixel points, and extracting the connected domain of the binarized image to obtain each connected domain.
In one embodiment, the determining of the scratch suspected level includes:
Calculating the circularity of each connected domain, calculating the length-width ratio of the minimum circumscribed rectangle of each connected domain, and obtaining the scratch suspected degree based on the length-width ratio and the circularity.
In one embodiment, a difference between the aspect ratio and a preset aspect ratio threshold is calculated, and a ratio of a forward mapping value of the difference to the circularity is used as the scratch suspected degree.
In one embodiment, the determining the weak scratch plausibility includes:
Calculating the gray average value of all background pixel points in a dark field image of the surface of the optical element to be detected, calculating the difference between the gray value of each suspected defect pixel point and the gray average value, marking the difference as a first difference, and taking the ratio of the scratch suspected degree of the connected domain where each suspected defect pixel point is located to the forward mapping value of the first difference as the weak scratch suspected degree.
In one embodiment, the gradient low threshold is a fusion result of gradient magnitudes of all suspected defective pixels and a normalized value of the weak scratch suspected level.
In one embodiment, the combining the neural network model to obtain the microscopic defect result of the surface of the optical element to be detected includes:
Respectively acquiring a preset number of surface dark field images of optical elements containing weak scratch defects and optical elements not containing weak scratch defects, and acquiring edge detection result images of all the surface dark field images by using an edge detection algorithm to serve as training data of a neural network model;
And taking the edge detection result image of the optical element to be detected as the input of the neural network model after training, and outputting the neural network model as the microscopic defect result of the surface of the optical element to be detected.
In a second aspect, an embodiment of the present application further provides an optical element surface micro defect detection system, including a memory, a processor, and a computer program stored in the memory and running on the processor, where the processor implements the steps of any one of the methods described above when executing the computer program.
The application has at least the following beneficial effects:
Acquiring a candidate edge pixel point set in a surface dark field image by acquiring the surface dark field images of an optical element to be detected and a standard optical element; obtaining each suspected defect pixel point of the optical element to be detected based on the difference of the candidate edge pixel point set of the optical element to be detected and the standard optical element; based on the distribution of the suspected defect pixel points, each connected domain of the surface dark field image of the optical element to be detected is obtained, the pixel points corresponding to the surface microscopic defects of the optical element to be detected in the surface dark field image are represented, and the reliability of analysis of the surface microscopic defects of the optical element to be detected is improved; further, analyzing the shape characteristics of each connected domain to obtain scratch suspected degree of each connected domain; combining the scratch suspected degree and the gray level difference between the suspected defect pixel point and a background pixel point in the image to obtain the weak scratch suspected degree of each suspected defect pixel point, so that the identification precision of the suspected weak scratch defect pixel point in the candidate edge pixel points in the edge detection algorithm of the surface dark field image of the optical element to be detected is improved; then, based on the weak scratch suspected degree and the gradient of each suspected defective pixel point, determining a gradient low threshold value in an edge detection algorithm, and analyzing the distribution characteristics of the suspected defective pixel points, so that the possibility that the weak scratch pixel points in the surface dark field image of the optical element to be detected are annihilated in the background and are not detected is reduced, and meanwhile, the possibility that the background information in the surface dark field image of the optical element to be detected is also extracted is also reduced, and the accuracy of determining the gradient low threshold value in the edge detection algorithm is improved; and finally, carrying out edge detection on the surface dark field image of the optical element to be detected, and combining a neural network model to obtain a microscopic defect result of the surface of the optical element to be detected, thereby improving the accuracy of detecting the microscopic defect of the surface of the optical element to be detected.
Drawings
In order to more clearly illustrate the embodiments of the application or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the application, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart showing the steps of a method for detecting microscopic defects on a surface of an optical element according to an embodiment of the present application;
FIG. 2 is a schematic illustration of scratch defects in a binarized image of a surface dark field image of an optical element;
fig. 3 is a gradient low threshold acquisition flow chart.
Detailed Description
In order to further describe the technical means and effects adopted by the present application to achieve the preset purpose, the following detailed description refers to the specific implementation, structure, characteristics and effects of a method and a system for detecting microscopic defects on the surface of an optical element according to the present application, which are described in detail below with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
The following specifically describes a specific scheme of the method and system for detecting microscopic defects on the surface of an optical element provided by the application with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of steps of a method for detecting microscopic defects on a surface of an optical element according to an embodiment of the application is shown, the method includes the steps of:
S1, collecting dark field images of surfaces of an optical element to be detected and a standard optical element, and preprocessing.
In one embodiment of the present application, a microscopic scattering dark field imaging system is used to collect a surface dark field image of an optical element to be detected and a surface dark field image of a standard optical element, respectively, because the minimum defect size that can be detected by the microscopic scattering dark field imaging system on the surface of the optical element to be detected is far smaller than the defect size detected under the bright field background, which is favorable for detecting the surface defect of the optical element to be detected, wherein the standard optical element is an optical element which has the same type and specification as the optical element to be detected and has no surface microscopic defect after manual detection, and the microscopic scattering dark field imaging system mainly comprises a CCD camera, a microscope objective, an annular LED illumination source, a detection platform, etc., which are specific to the prior art, and the present application is not described in detail herein.
In order to reduce interference of noise data in the surface dark field image on subsequent image processing, in one embodiment of the present application, the acquired surface dark field image is respectively denoised by using a bilateral filtering algorithm to obtain a surface dark field image a of an optical element to be detected after denoise and a surface dark field image B of a standard optical element after denoise, wherein the bilateral filtering algorithm is a prior known technology, and a specific process is not repeated.
It should be noted that, the implementer may select other available denoising algorithms according to the actual situation, and the present application is not limited herein.
S2, acquiring a candidate edge pixel point set in the surface dark field image; obtaining each suspected defect pixel point of the optical element to be detected based on the difference of the candidate edge pixel point set of the optical element to be detected and the candidate edge pixel point set of the standard optical element; and obtaining each connected domain of the surface dark field image of the optical element to be detected based on the distribution of the suspected defective pixel points.
In this embodiment, taking a Canny edge detection algorithm as an example, a gradient low threshold value in the edge detection processing process of the surface dark field image of the optical element is adaptively adjusted by using the Canny edge detection algorithm.
The Canny edge detection algorithm comprises 4 steps of Gaussian blur, image gradient calculation, non-maximum suppression and double-threshold boundary tracking, wherein a group of candidate edge pixel points are reserved in the non-maximum suppression step for the subsequent double-threshold boundary tracking step.
And carrying out image registration processing on the surface dark field image A and the surface dark field image B by using an SURF-based image registration algorithm to obtain pixel points corresponding to each pixel point in the surface dark field image A in the surface dark field image B, wherein the SURF-based image registration algorithm is a prior known technology, the specific process is not repeated, and an implementer can select other prior feasible image registration algorithms according to actual conditions, so that the application is not limited.
Taking the surface dark field image A as an example, a Canny edge detection algorithm is used for obtaining a group of candidate edge pixel points which are reserved in a non-maximum value suppression step of the Canny edge detection algorithm and form a candidate edge pixel point set A1, wherein the set is used for representing the surface profile of an optical element to be detected and the pixel point set which is formed by the edge of the surface microscopic defect and corresponds to the surface dark field image A, the set is marked as the candidate edge pixel point set A1 of the surface dark field image A, and meanwhile, the set which is formed by the pixel points which do not belong to the candidate edge pixel point set A1 in the surface dark field image A is marked as the background pixel point set AM of the surface dark field image A, wherein the Canny edge detection algorithm is a known technology, and the specific process is not repeated.
And acquiring a candidate pixel point set B1 of the surface dark field image B by using the same acquisition method as the candidate edge pixel point set A1, wherein the candidate pixel point set B1 is used for representing a set formed by pixels corresponding to the surface dark field image B on the edge of the surface profile of the standard optical element.
Calculating an intersection AB between the candidate edge pixel point set A1 and the candidate edge pixel point set B1, calculating a pixel point set A2 obtained by subtracting the intersection AB from the candidate edge pixel point set A1, and marking a set formed by pixel points corresponding to the surface dark field image A of the edge suspected to be the surface microscopic defect of the optical element to be detected in the candidate edge pixel point set A1, and marking each pixel point in the pixel point set A2 as each suspected defect pixel point.
In this embodiment, the pixels belonging to the pixel set A2 in the surface dark-field image a are assigned 255, and the remaining pixels are assigned 0, so as to obtain a binarized image C after reassigning the pixels in the surface dark-field image a, which is used to characterize the distribution of the pixels corresponding to the edges of the surface microscopic defects of the optical element to be detected in the surface dark-field image a. Wherein, the practitioner can adopt other modes to carry out assignment according to actual conditions to obtain a binary image, and the application is not limited to this.
The binarized image C is processed by using morphological open operation to obtain a binarized image C1, so as to remove small spots or small noise points in the surface dark field image a from corresponding pixels in the binarized image C, wherein the morphological open operation is a prior art, the detailed process is not described herein, and an implementer can select other algorithms according to actual situations, which is not limited by the present application.
And extracting connected domains of the binarized image C1 to obtain all connected domains of the binarized image C1, wherein each connected domain respectively represents a pixel point corresponding to one surface microscopic defect of the optical element to be detected in the surface dark field image A.
S3, analyzing the shape characteristics of each connected domain to obtain scratch suspected degree of each connected domain; and combining the scratch suspected degree and the gray level difference between the suspected defect pixel point and the background pixel point in the image to obtain the weak scratch suspected degree of each suspected defect pixel point.
Since scratches on the surface of the optical element generally take a long-strip shape, the shape characteristics of each connected domain are analyzed. A schematic diagram of scratch defects in a binary image of a surface dark field image of an optical element is shown in fig. 2.
Taking the ith connected domain Di of the binarized image C1 as an example, calculating the circularity Si of the connected domain Di, the closer the edge shape of the microscopic defect on the surface of the optical element to be detected corresponding to the connected domain Di is to be circular, the larger the circularity Si of the connected domain Di is, and the lower the possibility of the microscopic scratch defect is correspondingly.
Obtaining the minimum circumscribed rectangle of the connected domain Di, calculating the length-width ratio ki of the connected domain Di, and determining the scratch suspected degree of the connected domain Di based on the circularity Si and the length-width ratio ki of the connected domain Di.
The method comprises the following steps: calculating the difference between the length-width ratio ki of the connected domain Di and a preset length-width ratio threshold k, and determining the ratio of the forward mapping value of the difference to the circularity Si as the scratch suspected degree of the connected domain Di.
It should be noted that, the difference represents the difference degree of the two variables, and specifically, the difference, the absolute value of the difference, the ratio and other modes can be adopted for calculation; the forward map indicates that the dependent variable will increase with increasing independent variable and the dependent variable will decrease with decreasing independent variable.
In this embodiment, the expression of the scratch suspicion Hi of the connected domain Di may be: Wherein ki represents the aspect ratio of the connected domain Di, k represents a preset aspect ratio threshold, in this embodiment, k=4, which can be set by an implementer according to the actual situation, and the application is not limited to this, and is used for representing the minimum aspect ratio of the scratch defect edge on the surface of the optical element; exp () represents an exponential function with a natural constant as a base, in order to avoid the value of (ki-k) being negative, affecting the calculation; si represents the circularity of the connected domain Di.
The larger the aspect ratio of the edge of the micro defect on the surface of the optical element to be detected is compared with the minimum aspect ratio of the edge of the scratch defect on the surface of the optical element, (ki-k) is larger, and the less the edge shape of the micro defect on the surface of the optical element to be detected corresponding to the connected domain Di is close to a circle, the smaller the circularity Si is, the more suspected the connected domain Di is a pixel point corresponding to the edge of the scratch defect on the surface of the optical element to be detected, and the greater the scratch suspected degree Hi of the connected domain Di is.
And obtaining the scratch suspected degree of each connected domain by adopting a calculation method which is the same as the scratch suspected degree of the connected domain Di.
Because the color of the weak scratch defect in the scratch defect of the surface of the optical element is relatively close to that of the normal area of the surface of the optical element, the gray value of the pixel point corresponding to the surface dark field image A at the edge of the weak scratch defect of the surface of the optical element to be detected is relatively close to that of the pixel point corresponding to the surface dark field image A at the normal area of the surface of the optical element to be detected.
Taking the j-th suspected defect pixel point A1 (j) as an example, calculating the weak scratch suspected degree R (j) of the pixel point A1 (j), specifically: calculating the gray average value of all pixel points in the background pixel point set AM, calculating the difference between the gray value of the suspected defective pixel point A1 (j) and the gray average value, marking the difference as a first difference, and taking the ratio of the scratch suspected degree of the connected domain where the suspected defective pixel point A1 (j) is positioned and the forward mapping value of the first difference as the weak scratch suspected degree of the pixel point A1 (j).
In the present embodiment, the expression of the weak scratch suspected level R (j) of the pixel point A1 (j) may be: Wherein Hi (j) represents a scratch suspected degree of the connected domain where the pixel point A1 (j) is located in the binarized image C1, and if the pixel point A1 (j) is not located in the connected domain, the Hi (j) is assigned to 0; a (j) represents a gradation value of the pixel point A1 (j); a represents the average value of gray values of all pixel points in a background pixel point set AM; exp () represents an exponential function based on a natural constant, in order to prevent the value of (a (j) -a) from affecting the calculation as a negative number.
The more suspected the pixel point A1 (j) is the pixel point corresponding to the surface dark field image A of the scratch defect of the surface of the optical element to be detected, the larger the scratch suspected degree Hi (j) of the pixel point A1 (j) in the connected domain in the binarized image C1 is, and the closer the gray value between the pixel point A1 (j) and the pixel point corresponding to the surface dark field image A of the normal area of the surface of the optical element to be detected is, (the smaller the (a (j) -a) is, the more suspected the pixel point A1 (j) is the pixel point corresponding to the surface dark field image A of the weak scratch defect in the surface scratch defect of the optical element to be detected is, and the larger the weak scratch suspected degree R (j) of the pixel point A1 (j) is.
And obtaining the weak scratch suspected degree of each suspected defective pixel point by adopting the same calculation method as the weak scratch suspected degree of the pixel point A1 (j).
And S4, determining a gradient low threshold in an edge detection algorithm based on the weak scratch suspected degree and the gradient of each suspected defective pixel point, carrying out edge detection on the surface dark field image of the optical element to be detected, and combining a neural network model to obtain a microscopic defect result of the surface of the optical element to be detected.
The method comprises the steps of obtaining gradient amplitude values obtained by calculating each suspected defective pixel point in an image gradient calculation step of a Canny edge detection algorithm, respectively carrying out normalization processing on weak scratch suspected degrees of each suspected defective pixel point by using a softmax function, and calculating a gradient low threshold value T of a surface dark field image A in a double-threshold value boundary tracking step of the Canny edge detection algorithm, wherein the gradient low threshold value T is specifically as follows: and taking a fusion result of the gradient amplitude values of all the suspected defect pixel points and the normalized value of the weak scratch suspected degree as a gradient low threshold T.
The fusion expression combines a plurality of variables, and specifically, may be calculated by adding, multiplying, averaging, or the like.
In the present embodiment, the expression of the gradient low threshold T may beWherein d (j) represents the gradient amplitude of the j-th suspected defect pixel point; v (j) represents the normalization result of the weak scratch suspected degree of the jth suspected defective pixel point; n represents the number of suspected defective pixels. A gradient low threshold acquisition flow chart is shown in fig. 3.
According to the obtained gradient low threshold T of the surface dark field image A in the double-threshold boundary tracking step of the Canny edge detection algorithm, the Canny edge detection algorithm is used for finishing the edge detection processing of the surface dark field image A, and an edge detection image F after the edge detection processing of the surface dark field image A is obtained, wherein the Canny edge detection algorithm is a prior known technology, and the specific process is not repeated.
In one embodiment of the present application, m=100, and the practitioner can set itself according to the actual situation, which is not limited in this aspect of the present application. And respectively carrying out edge detection processing on all acquired surface dark field images by using the same method as the surface dark field image A, taking all images obtained after the edge detection processing as training data of a convolutional neural network model, setting the image containing the weak scratch defect in the training data as 1, setting the image not containing the weak scratch defect as 0, and respectively taking a random gradient descent method and a cross entropy loss function as an optimization algorithm and a loss function of the convolutional neural network model in the training process, wherein the training of the convolutional neural network model is the prior known technology, and the specific process is not repeated.
It should be noted that, the implementer may also select other available neural network models according to the actual situation, which is not limited by the present application.
And taking the edge detection image F as the input of the trained convolutional neural network model, outputting a label of the edge detection image F, judging that the optical element to be detected has the weak scratch defect if the label is 1, and judging that the optical element to be detected does not have the weak scratch defect if the label is 0.
Based on the same inventive concept as the above method, the embodiment of the present application further provides an optical element surface micro defect detection system, which includes a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of any one of the above optical element surface micro defect detection methods.
In summary, the application acquires the candidate edge pixel point set in the surface dark field image by acquiring the surface dark field images of the optical element to be detected and the standard optical element; obtaining each suspected defect pixel point of the optical element to be detected based on the difference of the candidate edge pixel point set of the optical element to be detected and the standard optical element; based on the distribution of the suspected defect pixel points, each connected domain of the surface dark field image of the optical element to be detected is obtained, the pixel points corresponding to the surface microscopic defects of the optical element to be detected in the surface dark field image are represented, and the reliability of analysis of the surface microscopic defects of the optical element to be detected is improved; further, analyzing the shape characteristics of each connected domain to obtain scratch suspected degree of each connected domain; combining the scratch suspected degree and the gray level difference between the suspected defect pixel point and a background pixel point in the image to obtain the weak scratch suspected degree of each suspected defect pixel point, so that the identification precision of the suspected weak scratch defect pixel point in the candidate edge pixel points in the edge detection algorithm of the surface dark field image of the optical element to be detected is improved; then, based on the weak scratch suspected degree and the gradient of each suspected defective pixel point, determining a gradient low threshold value in an edge detection algorithm, and analyzing the distribution characteristics of the suspected defective pixel points, so that the possibility that the weak scratch pixel points in the surface dark field image of the optical element to be detected are annihilated in the background and are not detected is reduced, and meanwhile, the possibility that the background information in the surface dark field image of the optical element to be detected is also extracted is also reduced, and the accuracy of determining the gradient low threshold value in the edge detection algorithm is improved; and finally, carrying out edge detection on the surface dark field image of the optical element to be detected, and combining a neural network model to obtain a microscopic defect result of the surface of the optical element to be detected, thereby improving the accuracy of detecting the microscopic defect of the surface of the optical element to be detected.
It should be noted that: the sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. And the foregoing description has been directed to specific embodiments of this specification. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
The foregoing description of the preferred embodiments of the present application is not intended to be limiting, but rather, any modifications, equivalents, improvements, etc. that fall within the principles of the present application are intended to be included within the scope of the present application.
Claims (10)
1. A method for detecting microscopic defects on a surface of an optical element, the method comprising the steps of:
Collecting dark field images of the surfaces of the optical element to be detected and the standard optical element;
acquiring a candidate edge pixel point set in the surface dark field image; obtaining each suspected defect pixel point of the optical element to be detected based on the difference of the candidate edge pixel point set of the optical element to be detected and the standard optical element; obtaining each connected domain of the surface dark field image of the optical element to be detected based on the distribution of the suspected defective pixel points;
analyzing the shape characteristics of each connected domain to obtain scratch suspected degree of each connected domain; combining the scratch suspected degree and the gray level difference between the suspected defect pixel point and a background pixel point in the image to obtain the weak scratch suspected degree of each suspected defect pixel point;
and determining a gradient low threshold in an edge detection algorithm based on the weak scratch suspected degree and the gradient of each suspected defective pixel point, carrying out edge detection on the surface dark field image of the optical element to be detected, and combining a neural network model to obtain a microscopic defect result of the surface of the optical element to be detected.
2. A method for detecting microscopic defects on a surface of an optical element according to claim 1, wherein the set of candidate edge pixels is all pixels reserved in the non-maximum suppressing step for edge detection of a surface dark field image.
3. The method for detecting microscopic defects on a surface of an optical element according to claim 1, wherein the determining of each pixel of suspected defects comprises:
Registering the surface dark field images of the optical element to be detected and the standard optical element, acquiring an intersection of the candidate edge pixel point set of the optical element to be detected and the standard optical element, and removing all the pixel points remained after the intersection in the candidate edge pixel point set of the optical element to be detected as all the suspected defect pixel points.
4. A method for detecting microscopic defects on a surface of an optical element according to claim 1, wherein the determining of each of the communicating regions includes:
And converting the dark field image of the surface of the optical element to be detected into a binarized image based on all the suspected defective pixel points, and extracting the connected domain of the binarized image to obtain each connected domain.
5. A method for detecting microscopic defects on a surface of an optical element as recited in claim 1, wherein the determining of the plausibility of scratches includes:
Calculating the circularity of each connected domain, calculating the length-width ratio of the minimum circumscribed rectangle of each connected domain, and obtaining the scratch suspected degree based on the length-width ratio and the circularity.
6. The method according to claim 5, wherein a difference between the aspect ratio and a predetermined aspect ratio threshold is calculated, and a ratio of a forward mapping value of the difference to the circularity is used as the scratch suspected level.
7. A method for detecting microscopic defects on a surface of an optical element as recited in claim 1, wherein the determining of the plausibility of the weak scratches includes:
Calculating the gray average value of all background pixel points in a dark field image of the surface of the optical element to be detected, calculating the difference between the gray value of each suspected defect pixel point and the gray average value, marking the difference as a first difference, and taking the ratio of the scratch suspected degree of the connected domain where each suspected defect pixel point is located to the forward mapping value of the first difference as the weak scratch suspected degree.
8. The method for detecting microscopic defects on a surface of an optical element according to claim 1, wherein the gradient low threshold is a fusion result of gradient magnitudes of all suspected defective pixels and a normalized value of the weak scratch suspected level.
9. The method for detecting microscopic defects on a surface of an optical element according to claim 1, wherein the step of combining the neural network model to obtain microscopic defect results on the surface of the optical element to be detected comprises the steps of:
Respectively acquiring a preset number of surface dark field images of optical elements containing weak scratch defects and optical elements not containing weak scratch defects, and acquiring edge detection result images of all the surface dark field images by using an edge detection algorithm to serve as training data of a neural network model;
And taking the edge detection result image of the optical element to be detected as the input of the neural network model after training, and outputting the neural network model as the microscopic defect result of the surface of the optical element to be detected.
10. An optical element surface micro defect detection system comprising a memory, a processor and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the method according to any of claims 1-9 when the computer program is executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411244384.1A CN118781099A (en) | 2024-09-06 | 2024-09-06 | Method and system for detecting microscopic defects on surface of optical element |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411244384.1A CN118781099A (en) | 2024-09-06 | 2024-09-06 | Method and system for detecting microscopic defects on surface of optical element |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118781099A true CN118781099A (en) | 2024-10-15 |
Family
ID=92981125
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411244384.1A Pending CN118781099A (en) | 2024-09-06 | 2024-09-06 | Method and system for detecting microscopic defects on surface of optical element |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118781099A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105631857A (en) * | 2015-12-17 | 2016-06-01 | 中国科学院自动化研究所 | Method and device for detecting scratch on surface of optical element |
| CN115908415A (en) * | 2023-01-06 | 2023-04-04 | 深圳新视智科技术有限公司 | Defect detection method, device and equipment based on edge and storage medium |
-
2024
- 2024-09-06 CN CN202411244384.1A patent/CN118781099A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105631857A (en) * | 2015-12-17 | 2016-06-01 | 中国科学院自动化研究所 | Method and device for detecting scratch on surface of optical element |
| CN115908415A (en) * | 2023-01-06 | 2023-04-04 | 深圳新视智科技术有限公司 | Defect detection method, device and equipment based on edge and storage medium |
Non-Patent Citations (2)
| Title |
|---|
| 冯凯萍;吕笑文;张丽文;杨耀雨;陈良奇;江晓亮;: "一种新的光学元件表面划痕检测算法", 电脑知识与技术, no. 01, 5 January 2018 (2018-01-05) * |
| 李克斌;余厚云;周申江;: "基于形态学特征的机械零件表面划痕检测", 光学学报, no. 08, 27 June 2018 (2018-06-27) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110490914B (en) | Image fusion method based on brightness self-adaption and significance detection | |
| CN109490316B (en) | Surface defect detection algorithm based on machine vision | |
| CN113592861B (en) | Bridge crack detection method based on dynamic threshold | |
| CN108765402B (en) | Non-woven fabric defect detection and classification method | |
| JP2010534432A (en) | System and method for enhancing digital image processing with noise removal function | |
| CN112614062A (en) | Bacterial colony counting method and device and computer storage medium | |
| CN113393426B (en) | Steel rolling plate surface defect detection method | |
| CN116012291A (en) | Industrial part image defect detection method and system, electronic equipment and storage medium | |
| CN113177924A (en) | Industrial production line product flaw detection method | |
| CN113034474A (en) | Test method for wafer map of OLED display | |
| CN115018797A (en) | Screen defect detection method, screen defect detection device and computer-readable storage medium | |
| CN116664457B (en) | Image processing method for enhancing denoising | |
| CN110021012B (en) | Mobile phone lens window glass defect detection method based on machine vision technology | |
| CN115311289A (en) | Method for detecting oil stain defects of plain-color cloth | |
| CN112308843A (en) | Camera overexposure abnormity adjusting method and system for photovoltaic cell panel image acquisition | |
| CN113673396B (en) | Spore germination rate calculation method, device and storage medium | |
| CN115272362A (en) | Method and device for segmenting effective area of digital pathology full-field image | |
| CN114022539A (en) | Microscopic image cell position judgment method | |
| CN112381084B (en) | Automatic contour recognition method for tomographic image | |
| CN113888503A (en) | Product appearance detection method and device and storage medium | |
| CN116977335B (en) | Intelligent detection method for pitting defects on surface of mechanical part | |
| CN118781099A (en) | Method and system for detecting microscopic defects on surface of optical element | |
| CN111563869B (en) | Stain test method for quality inspection of camera module | |
| CN112184619A (en) | Metal part surface defect detection method based on deep learning | |
| CN115254674A (en) | Bearing defect sorting method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination |