A method and system for compensating white light interferometry for surface microtopography
By using white light interferometry and Unet++ semantic segmentation network, the problem of distinguishing between tool marks and attachments on the surface of key components was solved, and compensation for distorted information was achieved, thus improving the accuracy of morphology evaluation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-02-17
- Publication Date
- 2026-04-10
AI Technical Summary
Existing technologies struggle to accurately distinguish between tool marks and attachments on the surface of critical components, leading to deviations in morphology evaluation results, and the distortions generated during processing cannot be compensated for.
The white light interferometry method is adopted to identify measurable and non-measurable points by numbering, generate ultra-depth two-dimensional images and point cloud data, use Unet++ semantic segmentation network to segment defects, and combine variable spline interpolation calculation to compensate for the distorted data of non-measurable points.
It enables the identification and compensation of actual tool marks and attachments on the surface of components, thereby improving the accuracy of surface morphology evaluation.
Smart Images
Figure CN116007533B_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the field of interferometry, in particular to a white light interferometry method and system for compensating surface micro-topography. BACKGROUND
[0002] Key components are the core of high-end equipment, bearing cyclic loads, such as bearings, transmission shafts, etc. Key components such as cracks in operation, eventually will have no premonition of sudden fracture failure, causing major accidents. Key components in the machining process of turning, milling, planing, grinding and other machining processes will leave tool marks on the machined surface, and the curvature radius of the tool mark valley is very small, thus greatly amplifying the stress level. Under the condition of cyclic load service, tool marks cause damage to the internal microstructure of the material and cracks. The surface micro-topography of the tool mark morphology determines the fatigue resistance of the key component.
[0003] Currently, the measurement of key components mainly uses general commercial instruments, such as laser confocal microscopes and white light interferometers. When these instruments use general methods for measurement, key information of tool marks is easily lost in the data preprocessing process. In addition, due to the use of coaxial light illumination by general commercial instruments, the reflected light of the steep side of the machining tool mark cannot enter the interference lens to form an image, resulting in distortion of the tool mark measurement information.
[0004] During the machining of key components, micro-chips and grinding wheel debris and other attachments generated during the machining of key components adhere to the surface of the key components under the high-temperature and high-speed machining environment, forming part of the surface topography of the key components. Micro-chips and grinding wheel debris are involved in the calculation of surface roughness as part of the surface topography, but such attachments are not integral with the substrate of the key component surface, and it is difficult to distinguish them with general commercial instruments. Therefore, the surface defects and attachment topography of the component cannot be distinguished from the substrate topography, and they are involved in the evaluation of the component surface together, resulting in a deviation in the evaluation results of the component surface topography and low evaluation accuracy. SUMMARY
[0005] The purpose of the embodiments of the present application is to provide a white light interferometry method and system for compensating surface micro-topography, which can identify the true tool marks and defect information on the surface of the component, compensate for distorted topography measurement information on the surface of the component, and improve the accuracy of the evaluation of the surface topography of the component.
[0006] To achieve the above purpose, the embodiments of the present application provide the following solutions:
[0007] A white light interferometry method for compensating surface micro-topography, the method comprising:
[0008] Obtaining a raw sequence image of a component surface white light interference; numbering original images in the raw sequence image in order; different numbers are used to represent different component surface positions;
[0009] According to a standard step surface, obtaining a determination threshold of measurable points and unmeasurable points in the raw sequence image; the measurable points are used to represent true values in the raw sequence image; the unmeasurable points are used to represent distorted values in the raw sequence image; the determination threshold is used to determine the measurable points and the unmeasurable points in the raw sequence image;
[0010] Using a digital image processing algorithm, the raw sequence image is generated into an ultra-depth-of-field two-dimensional image at the same time, the ultra-depth-of-field two-dimensional image includes color information, and the raw sequence image is generated into height point cloud data; according to the ultra-depth-of-field two-dimensional image and the height point cloud data, the measurable point cloud data with color information is generated;
[0011] After the raw sequence image is generated into an ultra-depth-of-field two-dimensional image, the color information in the ultra-depth-of-field two-dimensional image is segmented and positioned using a Unet++ semantic segmentation network, and position data, quantity data and height data relative to a component substrate of a component surface defect or an adherend are obtained;
[0012] Based on the measurable points adjacent to the unmeasurable points, a variable spline interpolation calculation is performed to obtain an unmeasurable point prediction result; according to the unmeasurable point prediction result and the measurable point cloud data, a component surface three-dimensional morphology is obtained.
[0013] Optionally, the different numbers are also used to calculate a relative height relative to a starting position of the component surface; the height data is the relative height.
[0014] Optionally, the determination threshold of the measurable points and the unmeasurable points in the raw sequence image according to the standard step surface specifically includes:
[0015] Converting the raw sequence image into a sequence gray scale image;
[0016] According to the sequence gray scale image, a maximum value and an average value of any gray scale value are obtained;
[0017] The maximum value and the average value are subtracted to obtain a difference value;
[0018] The difference value is determined as the determination threshold.
[0019] Optionally,
[0020] The raw sequence image is generated into an ultra-depth-of-field two-dimensional image at the same time, and the raw sequence image is generated into height point cloud data specifically includes:
[0021] According to the maximum value of the gray value of the same pixel point in the sequence gray image, the relative height of the pixel point is obtained;
[0022] The relative heights of all pixel points in the sequence gray image are obtained;
[0023] According to the relative heights of all pixel points, the height point cloud data is obtained;
[0024] The generating of the measurable point point cloud data with color information from the hyper-zoom depth two-dimensional image and the height point cloud data specifically includes:
[0025] According to the point cloud reconstruction method, the color information on the hyper-zoom depth two-dimensional image is assigned to the height point cloud data, and the measurable point point cloud data with color information is obtained.
[0026] To achieve the above object, the embodiments of the present application further provide the following schemes:
[0027] A compensation surface micro-topography white light interferometric measurement system, the system comprises:
[0028] An image acquisition module is configured to acquire a raw sequence image of white light interference of a component surface, number the raw images in the raw sequence image in order, and use different numbers to represent different component surface positions.
[0029] A determination threshold acquisition module is configured to acquire a determination threshold of measurable points and non-measurable points in the raw sequence image according to a standard step surface, wherein the measurable points represent real values in the raw sequence image, the non-measurable points represent distorted values in the raw sequence image, and the determination threshold is used to determine the measurable points and the non-measurable points in the raw sequence image.
[0030] A point cloud data generation module is configured to generate a hyper-zoom depth two-dimensional image from the raw sequence image by using a digital image processing algorithm, wherein the hyper-zoom depth two-dimensional image includes color information, and the raw sequence image is also used to generate height point cloud data, and the hyper-zoom depth two-dimensional image and the height point cloud data are used to generate measurable point point cloud data with color information.
[0031] An Unet++ semantic segmentation module is configured to segment and locate the color information in the hyper-zoom depth two-dimensional image after the raw sequence image is generated into the hyper-zoom depth two-dimensional image, so as to obtain position data, quantity data and height data relative to a component base body of a component surface defect or an adherent.
[0032] The three-dimensional topography generation module is configured to perform variational spline interpolation calculation based on the measured points adjacent to the unmeasured points to obtain an unmeasured point prediction result; and obtain a three-dimensional topography of a component surface according to the unmeasured point prediction result and the measured point cloud data.
[0033] Optionally, the different numbers are also used to calculate relative heights relative to a starting position of the component surface; and the height data is the relative heights.
[0034] Optionally, the determination threshold acquisition module is specifically configured to:
[0035] convert the original sequence images into sequence gray scale images;
[0036] obtain a maximum value and an average value of any gray scale value according to the sequence gray scale images;
[0037] perform difference value calculation on the maximum value and the average value to obtain a difference value;
[0038] determine the difference value as the determination threshold.
[0039] Optionally, the point cloud data generation module is specifically configured to:
[0040] obtain a relative height of a same pixel point in the sequence gray scale images according to a maximum value of a gray scale value of the pixel point;
[0041] obtain relative heights of all pixel points in the sequence gray scale images;
[0042] obtain the height point cloud data according to the relative heights of all the pixel points;
[0043] assign color information on the hyper-zoom depth two-dimensional image to the height point cloud data according to a point cloud reconstruction method to obtain the measured point cloud data with color information.
[0044] In the embodiment of the application, original sequence images of component surface white light interference are obtained; original images in the original sequence images are numbered in sequence; different numbers represent different component surface positions. A determination threshold of measured points and unmeasured points in the original sequence images is obtained according to a standard step surface, threshold setting is realized, and measured points and unmeasured points of the component surface are determined.
[0045] The original sequence image is generated into the super depth of field two-dimensional image and the height point cloud data by using the digital image processing algorithm; the measurable point point cloud data with color information is generated according to the super depth of field two-dimensional image and the height point cloud data. The prediction result of the unmeasurable point is calculated by the variable spline interpolation based on the measurable point adjacent to the unmeasurable point, and the distortion data generated by the unmeasurable point is removed and compensated; and the three-dimensional topography of the component surface is obtained according to the prediction result of the unmeasurable point and the measurable point point cloud data. In the realization of the identification of the real topography information of the component surface, the measurement of the measurable point is realized, the compensation of the height information of the distortion point on the component surface, the prediction of the unmeasurable point is realized, and the accuracy of the evaluation of the component surface topography is improved. BRIEF DESCRIPTION OF DRAWINGS
[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.
[0047] Figure 1 The flowchart of the compensation surface microtopography white light interferometric measurement method provided by the embodiment of the present application is shown in the figure.
[0048] Figure 2 The schematic diagram of the original sequence image provided by the embodiment of the present application is shown in the figure.
[0049] Figure 3 The curve schematic diagram of the gray value of the measurable point, the gray value of the unmeasurable point and the average gray value is shown in the figure.
[0050] Figure 4 The schematic diagram of the super depth of field two-dimensional image provided by the embodiment of the present application is shown in the figure.
[0051] Figure 5 The schematic diagram of the height point cloud data provided by the embodiment of the present application is shown in the figure.
[0052] Figure 6 The schematic diagram of the measurable point point cloud data with color information provided by the embodiment of the present application is shown in the figure.
[0053] Figure 7 The schematic diagram of the segmentation and positioning of the super depth of field two-dimensional image by using the Unet++ semantic segmentation network provided by the embodiment of the present application is shown in the figure.
[0054] Figure 8 The specific schematic diagram of the compensation surface microtopography white light interferometric measurement method provided by the embodiment of the present application is shown in the figure.
[0055] Figure 9A structure schematic diagram of a surface micro-topography white light interferometric measurement system provided by the embodiment of the present application is provided.
[0056] Symbol explanation:
[0057] An image acquisition module-1, a determination threshold acquisition module-2, a point cloud data generation module-3, a Unet++ semantic segmentation module-4, and a three-dimensional topography generation module-5. DETAILED DESCRIPTION
[0058] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts fall within the scope of the present application.
[0059] The purpose of the present application is to provide a surface micro-topography white light interferometric measurement method and system to solve the problem that the existing method cannot compensate for the distortion topography information of a component surface, and to distinguish defects such as the base body and the adhering matter of the component surface, and to improve the accuracy of the component surface topography evaluation.
[0060] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.
[0061] Figure 1 An exemplary flow of the above-mentioned surface micro-topography white light interferometric measurement method is shown. Each step will be described in detail below.
[0062] Step 1: Obtain the original sequence images of the white light interference of the component surface; number the original images in the original sequence images in order; and different numbers are used to represent different component surface positions.
[0063] Step 1 can be performed by the image acquisition module 1. The image acquisition module 1 takes a picture of the component surface to obtain the original sequence images of the white light interference of the component surface, and the surface topography includes defects such as knife marks and adhering matter. Please refer to Figure 2 The original sequence images include n original images, and those skilled in the art can flexibly design the value of n, such as 100, 200, 300, etc., which will not be described here. Then, the original images in the original sequence images are numbered in order from 1 to k. Different numbers are used to represent different component surface positions, i.e., different numbers represent different relative heights of the original images relative to the vertical scanning starting position, i.e., the relative height mentioned later.
[0064] In one example, the image acquisition module 1 can specifically be a white light interferometer. The white light interferometer specifically includes a piezoelectric ceramic device, a CCD industrial camera, and a white light interference lens. The white light interference acquisition component acquires a color sequence original image of the surface white light interference of the component.
[0065] The illumination beam of the white light interferometer is split into two beams by a half-mirror, which are projected onto the surface of the component and the reference mirror surface respectively. The two beams reflected from the two surfaces are combined into one beam again after passing through the beam splitter, and two superimposed images are formed on the photosensitive surface of the CCD camera by the imaging system. Due to the interference of the two beams, interference fringes with light and dark bands are observed on the photosensitive surface of the CCD camera. The brightness of the interference fringes depends on the optical path difference of the two beams. According to the light and dark degree of the white light interference fringes and the position of the interference fringes, the relative height of the measured component is analyzed.
[0066] Step 2: According to the standard step surface, obtain the determination threshold of the measurable points and the non-measurable points in the original sequence image; the measurable points are used to represent the true values in the original sequence image; the non-measurable points are used to represent the distorted values in the original sequence image; the determination threshold is used to determine the measurable points and the non-measurable points in the original sequence image.
[0067] Step 2 can be performed by the determination threshold acquisition module 2. According to the standard step surface, the determination threshold acquisition module 2 pre-acquires and determines the determination threshold of the measurable points and the non-measurable points in each original image of the original sequence image; the measurable points are used to represent the true values in the original sequence image; the non-measurable points are used to represent the distorted values in the original sequence image; the determination threshold is used to determine the measurable points and the non-measurable points in the original sequence image.
[0068] Step 3: using a digital image processing algorithm, generating an ultra-depth-of-field two-dimensional image from the original sequence image, the ultra-depth-of-field two-dimensional image including color information, generating height point cloud data from the original sequence image; generating measurable point cloud data with color information according to the ultra-depth-of-field two-dimensional image and the height point cloud data.
[0069] Step 3 can be performed by the point cloud data generation module 3. Please refer to Figure 4 , the measurable point cloud data generation module 3 uses a digital image processing algorithm to generate an ultra-depth-of-field two-dimensional image from the original sequence image, please refer to Figure 5 , generating height point cloud data from the original sequence image; please refer to Figure 6 , generating measurable point cloud data with color information according to the ultra-depth-of-field two-dimensional image and the height point cloud data.
[0070] In one example, please refer to Figure 8(x, y) represents the coordinates of the pixel point in the original sequence image. Each original image is divided into 3 sequence RGB digital images. The fusion array is the original image numbered k. r represents the color information of the R digital image in the 3 sequence RGB digital images, g represents the color information of the G digital image in the 3 sequence RGB digital images, and b represents the color information of the B digital image in the 3 sequence RGB digital images. All sequence RGB digital images constitute a tomographic image, that is, the original sequence image in the foregoing.
[0071] Step 4: After generating the hyper-zoom two-dimensional image from the original sequence image, the color information in the hyper-zoom two-dimensional image is segmented and located by using the Unet++ semantic segmentation network, to obtain the position data, the quantity data and the height data relative to the component base body of the component surface defect or the adherend.
[0072] Step 4 can be performed by the Unet++ semantic segmentation module 4. After generating the hyper-zoom two-dimensional image from the original sequence image, the color information in the hyper-zoom two-dimensional image is segmented and located by using the Unet++ semantic segmentation module 4, to obtain the position data, the quantity data and the height data relative to the component base body of the component surface defect or the adherend. Please refer to Figure 7 , the circle is the adherend, and the square is the unmeasurable point. The hyper-zoom two-dimensional image is segmented and located by using the Unet++ semantic segmentation network, to obtain the position data of the component surface adherend, that is, the coordinates of the center pixel point of the adherend; the quantity data, that is, the number of connected domains of the adherend and defects; and the height data relative to the component base body, the height data being the relative height h. After the Unet++ semantic segmentation network identifies the defects in the hyper-zoom two-dimensional image, the relative height h of the adherend and defects is extracted from the measurable point cloud data according to the connected domain coordinates.
[0073] Step 5: based on the measurable points adjacent to the unmeasurable points, a variable spline interpolation calculation is performed to obtain the prediction result of the unmeasurable points; and based on the prediction result of the unmeasurable points and the measurable point cloud data, a three-dimensional topography of the component surface is obtained.
[0074] Step 5 can be performed by the three-dimensional topography generation module 5. The three-dimensional topography generation module 5 performs a variable spline interpolation calculation based on the measurable points adjacent to the unmeasurable points, for example, the measurable points adjacent to the unmeasurable points, to obtain the prediction result of the unmeasurable points; and the prediction result of the unmeasurable points is fused with the measurable point cloud data to complete the distortion values in the point cloud data, to obtain the three-dimensional topography of the component surface. For example, the measurable points and the unmeasurable points are distinguished by a determination threshold, then the unmeasurable points are removed, and the measurable points are subjected to a cubic variable spline interpolation calculation to complete the distortion values in the point cloud data.
[0075] In the embodiment of the present application, the original sequence images of the component surface white light interference are obtained, the original images in the original sequence images are numbered in order, and different numbers represent different component surface positions. The determination threshold of the measurable points and the non-measurable points in the original sequence images is obtained according to the standard step surface, and the threshold setting is realized, and the measurable points and the non-measurable points of the component surface are determined.
[0076] The original sequence images are generated into the super-depth two-dimensional images by using the digital image processing algorithm, and the height point cloud data is generated at the same time, the measurable point cloud data with color information is generated according to the super-depth two-dimensional images and the height point cloud data. The non-measurable point prediction result is obtained by the variable spline interpolation calculation based on the adjacent measurable points of the non-measurable points, and the distortion data generated by the non-measurable points is eliminated and compensated. The component surface three-dimensional topography is obtained according to the non-measurable point prediction result and the measurable point cloud data. In the realization of the component surface real topography information recognition, the measurement of the measurable points is realized, the compensation of the component surface distortion topography information, that is, the compensation of the non-measurable points is realized, and the accuracy of the component surface topography evaluation is improved.
[0077] In other embodiments of the present application, the compensation surface micro-topography white light interference measurement method further comprises that the different numbers are also used for calculating the relative height relative to the starting position of the component surface. The height data is the relative height.
[0078] Please refer to Figure 8 , (x, y) represents the coordinates of the pixel points in the original sequence images. The height array is the original image with the number k. The white light interferometer is adjusted once per vertical shooting height, and the height of each adjustment is the same, that is, the step distance is Δh, and the range of Δh is within 200 um. Then the relative height h=k*Δh. Δh can be as small as 0.1 nm and as large as 200 um. The smaller Δh is, the more accurate the measurement accuracy of the relative height h is, but the measurement speed of the relative height h is also relatively reduced.
[0079] In other embodiments of the present application, the determination threshold of the measurable points and the non-measurable points in the original sequence images is obtained according to the standard step surface, and the determination threshold of the measurable points and the non-measurable points in the original sequence images is obtained according to the standard step surface. Specifically, it comprises:
[0080] Step 21: converting the original sequence images into sequence gray scale images.
[0081] Step 22: obtaining the maximum value and the average value of any gray scale value of the sequence gray scale images.
[0082] Step 23: calculating the difference value by subtracting the average value from the maximum value to obtain the difference value.
[0083] Step 24: determining the difference value as the determination threshold.
[0084] In one example, please refer to Figure 3, the ordinate is the gray value, and the abscissa is the sequence number of the original sequence image, that is, the number. The pixel point data with a difference between the maximum value and the average value of the gray value less than a determination threshold is removed, that is, the unmeasurable points on the surface of the component that are too steep and cannot enter the white light interference lens are removed. In step 22, the maximum value of the gray value of any gray image pixel point in the sequence gray image is obtained, for example, the gray values of any gray image pixel point (500, 500) in the sequence gray image are 1, 5, 6, 7, and 8, and the maximum value of the gray value of the pixel point is 8, at this time, the average value of the gray value of the sequence gray image is 5, that is, the difference is 8-5=3.
[0085] In other embodiments of the present application, a digital image processing algorithm is used to generate the original sequence image into the hyper-depth two-dimensional image and the height point cloud data, specifically including:
[0086] Step 31: According to the number corresponding to the maximum value of the gray value of the same pixel point in the sequence gray image, the relative height of the pixel point is obtained.
[0087] Step 32: Obtain the relative height of all pixel points in the sequence gray image.
[0088] Step 33: According to the relative height of all pixel points, the height point cloud data is obtained.
[0089] In one example, first, the original sequence image is converted into a sequence gray image, the number corresponding to the maximum value of the gray value of the same pixel point in any one of the sequence gray images is extracted, and the relative height of the pixel point is obtained according to the relative height h=k*Δh. In this way, the relative height of all pixel points in the sequence gray image is obtained. Then, according to the relative height of all pixel points, the height point cloud data is obtained.
[0090] In other embodiments of the present application, the hyper-depth two-dimensional image and the height point cloud data are used to generate measurable point point cloud data with color information, specifically including:
[0091] Step 34: According to the point cloud reconstruction method, the color information on the hyper-depth two-dimensional image is assigned to the height point cloud data to obtain the measurable point point cloud data with color information.
[0092] In one example, according to the point cloud reconstruction method, the color information of the R digital image in the above r represents the color information of the R digital image in the three sequence RGB digital images, the color information of the G digital image in the three sequence RGB digital images, and the color information of the B digital image in the three sequence RGB digital images, which is assigned to the height point cloud data to obtain the measurable point point cloud data with color information.
[0093] In other embodiments of the present application, the method for compensating surface micro-topography white light interferometry further comprises, after generating the super-depth-of-field two-dimensional image from the original sequence image:
[0094] Figure 9 An exemplary structure of the above-mentioned white light interferometry system for compensating surface micro-topography is shown. The system at least comprises: an image acquisition module 1, a determination threshold acquisition module 2, a point cloud data generation module 3, an Unet++ semantic segmentation module 4, and a three-dimensional topography generation module 5. Each module will be described in detail below.
[0095] The image acquisition module 1 is used to acquire the original sequence image of the white light interference of the component surface; the original images in the original sequence image are numbered in order; different numbers are used to represent different component surface positions.
[0096] In one example, the specific description of the image acquisition module 1 is as described above and will not be repeated here.
[0097] The determination threshold acquisition module 2 is used to acquire the determination threshold of the measurable point and the non-measurable point in the original sequence image according to the standard step surface; the measurable point is used to represent the true value in the original sequence image; the non-measurable point is used to represent the distorted value in the original sequence image; the determination threshold is used to determine the measurable point and the non-measurable point in the original sequence image.
[0098] In one example, the specific description of the determination threshold acquisition module 2 is as described above and will not be repeated here.
[0099] The point cloud data generation module 3 is used to generate a super-depth-of-field two-dimensional image from the original sequence image by using a digital image processing algorithm, and the super-depth-of-field two-dimensional image includes color information, and the original sequence image generates height point cloud data; the measurable point cloud data with color information is generated according to the super-depth-of-field two-dimensional image and the height point cloud data.
[0100] In one example, the specific description of the point cloud data generation module 3 is as described above and will not be repeated here.
[0101] The Unet++ semantic segmentation module 4 is used to segment and locate the color information in the super-depth-of-field two-dimensional image after the original sequence image generates the super-depth-of-field two-dimensional image, to obtain the position data, the quantity data and the height data relative to the component base body of the component surface defects or the adherend.
[0102] In one example, the specific description of the Unet++ semantic segmentation module 4 is as described above and will not be repeated here.
[0103] The three-dimensional topography generation module 5 is configured to perform variable spline interpolation calculation based on the measured points adjacent to the unmeasured points to obtain a prediction result of the unmeasured points, and obtain the three-dimensional topography of the component surface according to the prediction result of the unmeasured points and the point cloud data of the measured points.
[0104] In one example, the three-dimensional topography generation module 5 is specifically described above, and will not be described here.
[0105] In other embodiments of the present application, the compensation surface micro-topography white light interferometric measurement system further comprises different numbers for calculating the relative height relative to the starting position of the component surface. The height data is the relative height.
[0106] In one example, the calculation process of the relative height is described above, and will not be described here.
[0107] In other embodiments of the present application, the determination threshold acquisition module 2 is specifically configured to:
[0108] Convert the original sequence image into a sequence gray scale image.
[0109] According to the sequence gray scale image, obtain the maximum value and the average value of any gray scale value.
[0110] Subtract the maximum value from the average value to obtain a difference value.
[0111] The difference value is determined as the determination threshold.
[0112] In one example, the determination threshold acquisition process is described above, and will not be described here.
[0113] In other embodiments of the present application, the point cloud data generation module 3 is specifically configured to:
[0114] According to the maximum value of the gray scale value corresponding to the same pixel point in the sequence gray scale image, the relative height of the pixel point is obtained.
[0115] Obtain the relative height of all pixel points in the sequence gray scale image.
[0116] According to the relative height of all pixel points, height point cloud data is obtained.
[0117] According to the point cloud reconstruction method, the color information on the super-depth-of-field two-dimensional image is assigned to the height point cloud data to obtain the measured point cloud data with color information.
[0118] In one example, the point cloud data acquisition process is described above, and will not be described here. The Unet++ semantic segmentation module 4 can be used to identify and segment the attachments and other defects according to the color difference between the attachments and other defects and the component surface in the super-depth-of-field two-dimensional image.
[0119] After the Unet++ semantic segmentation module 4 identifies the defects in the super-depth-of-field two-dimensional image, the relative height h of the defects such as the attachments is extracted from the measurable point cloud data according to the connected domain coordinates.
[0120] The various embodiments are described in a progressive manner in the specification, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referred to each other. For the system disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method part.
[0121] The principles and implementation manners of the embodiments of the present application are described by using specific examples in the present application, and the above embodiment description is only used to help understand the method and core idea of the embodiments of the present application; meanwhile, for the general technical personnel in the art, the specific implementation manners and application ranges will be changed according to the idea of the embodiments of the present application. In conclusion, the content of the specification should not be understood as the limitation of the embodiments of the present application.
Claims
1. A white light interferometry method for compensating surface micro-topography, characterized in that, the method comprises: acquiring a raw sequence image of white light interference of a component surface; numbering original images in the raw sequence image in order; different numbers are used to represent different positions of the component surface; acquiring a judgment threshold of measurable points and unmeasurable points in the raw sequence image according to a standard step surface; the measurable points are used to represent true values in the raw sequence image; the unmeasurable points are used to represent distorted values in the raw sequence image; the judgment threshold is used to judge the measurable points and the unmeasurable points in the raw sequence image; generating an ultra-depth-of-field two-dimensional image while generating height point cloud data of the raw sequence image by using a digital image processing algorithm, the ultra-depth-of-field two-dimensional image including color information; generating measurable point cloud data with color information according to the ultra-depth-of-field two-dimensional image and the height point cloud data; after generating the ultra-depth-of-field two-dimensional image from the raw sequence image, segmenting and locating the color information in the ultra-depth-of-field two-dimensional image by using a Unet++ semantic segmentation network to obtain position data, quantity data and height data of a component surface defect or an adherent relative to a component substrate; performing variable spline interpolation calculation based on the measurable points adjacent to the unmeasurable points to obtain a prediction result of the unmeasurable points; obtaining a three-dimensional topography of the component surface according to the prediction result of the unmeasurable points and the measurable point cloud data.
2. The compensated surface microtopography white light interferometry method of claim 1, wherein, the different numbers are also used to calculate relative heights relative to a starting position of the component surface; the height data is the relative height.
3. The compensated surface microtopography white light interferometry method of claim 1, wherein, the acquiring of the judgment threshold of the measurable points and the unmeasurable points in the raw sequence image according to the standard step surface specifically comprises: converting the raw sequence image into a sequence gray scale image; obtaining a maximum value and an average value of a gray scale value of any gray scale image according to the sequence gray scale image; calculating a difference value by subtracting the average value from the maximum value to obtain the difference value; determining the difference value as the judgment threshold.
4. The white light interferometry method for compensating surface micro-topography according to claim 3, characterized in that, the generating of the ultra-depth-of-field two-dimensional image while generating the height point cloud data of the raw sequence image by using the digital image processing algorithm specifically comprises: obtaining a relative height of a pixel point according to a number corresponding to a maximum value of a gray scale value of the pixel point in the same pixel point in the sequence gray scale image; acquiring relative heights of all pixel points in the sequence gray scale image; obtaining the height point cloud data according to the relative heights of all the pixel points; the generating of the measurable point cloud data with color information according to the ultra-depth-of-field two-dimensional image and the height point cloud data specifically comprises: assigning the color information on the ultra-depth-of-field two-dimensional image to the height point cloud data to obtain the measurable point cloud data with color information according to a point cloud reconstruction method.
5. A white light interferometry system for compensating surface micro-topography, characterized in that, the system comprises: An image acquisition module is configured to acquire original sequence images of white light interference of a component surface, number original images in the original sequence images in sequence, and use different numbers to represent different component surface positions. A determination threshold acquisition module is configured to acquire determination thresholds of measurable points and unmeasurable points in the original sequence images according to a standard step surface, use the measurable points to represent true values in the original sequence images, use the unmeasurable points to represent distorted values in the original sequence images, and use the determination thresholds to determine the measurable points and the unmeasurable points in the original sequence images. A point cloud data generation module is configured to generate, by using a digital image processing algorithm, an ultra-deep-field two-dimensional image including color information from the original sequence images and height point cloud data from the original sequence images, and generate measurable point cloud data with color information from the ultra-deep-field two-dimensional image and the height point cloud data. An Unet++ semantic segmentation module is configured to segment and locate color information in the ultra-deep-field two-dimensional image to obtain position data, quantity data and height data relative to a component base body of a component surface defect or an adherent. A three-dimensional topography generation module is configured to perform a variable spline interpolation calculation based on measurable points adjacent to unmeasurable points to obtain an unmeasurable point prediction result, and obtain a component surface three-dimensional topography from the unmeasurable point prediction result and the measurable point cloud data.
6. The compensated surface microtopography white light interferometry system of claim 5, wherein, The different numbers are also used to calculate relative heights relative to a starting position of the component surface, and the height data is the relative height.
7. The compensated surface microtopography white light interferometry system of claim 5, wherein, The determination threshold acquisition module is specifically configured to: convert the original sequence images into sequence gray scale images, obtain a maximum value and an average value of any gray scale value of the sequence gray scale images, calculate a difference value by subtracting the average value from the maximum value to obtain the difference value, and determine the difference value as the determination threshold.
8. The compensated surface microtopography white light interferometry system of claim 7, wherein, The point cloud data generation module is specifically configured to: obtain a relative height of a pixel point in the sequence gray scale images according to a number corresponding to a maximum value of a gray scale value of the pixel point in the sequence gray scale images, obtain relative heights of all pixel points in the sequence gray scale images, obtain the height point cloud data according to the relative heights of the all pixel points, and assign color information on the ultra-deep-field two-dimensional image to the height point cloud data according to a point cloud reconstruction method to obtain the measurable point cloud data with color information.