CN118275457A - Curved surface optical element surface defect detection device and method based on composite illumination - Google Patents

Abstract

The present invention discloses a device and method for detecting surface defects of curved optical elements based on composite illumination. The device includes a microscope tube, a camera, an illumination light source, an electrically controlled displacement stage and a computer. The camera and the surface to be measured are conjugate surfaces. The surface to be measured is imaged on the camera chip through the lens tube. The illumination light source includes an off-axis illumination mode and a coaxial illumination mode. The coaxial illumination mode includes a white light illumination module and a structured light illumination module. The off-axis illumination mode is selected to quickly locate the defects; the coaxial white light illumination module is switched to realize automatic focusing of the surface; the coaxial structured light illumination module is switched to realize surface defect detection. By performing longitudinal scanning on the curved surface, images of the same target at different focal planes are obtained, and the images are processed and fused to expand the dynamic range of the device; by performing transverse scanning, three-dimensional splicing and fusion, full-aperture surface defect information is obtained, and finally three-dimensional detection and quantitative evaluation of weak defects on the surface of full-aperture curved optical elements are realized.

Description

A device and method for detecting surface defects of curved optical elements based on composite illumination

Technical Field

The present invention belongs to the field of optical detection, and in particular relates to a device and method for detecting surface defects of curved optical elements based on composite illumination.

Background technique

As the application of ultra-precision curved optical components becomes more and more widespread, higher requirements are placed on the ability to detect surface defects. Since curved optical components have higher surface complexity and more complex manufacturing processes, surface defect detection of curved optical components is crucial for optical system quality control. Improving detection methods and devices to achieve full-aperture three-dimensional measurement of defects is one of the main breakthrough directions at present.

Commonly used surface defect detection methods are mainly divided into contact and non-contact. Contact detection methods, such as scanning probe surface profilers, have a small detection range and are difficult to detect high-steepness surfaces. The visual method among non-contact detection methods is still widely used due to its simple operation and low cost, but the detection results depend on the resolution of the human eye and the results are not reliable. In addition, white light interferometry is also a commonly used detection method among non-contact detection methods. The detection results are highly accurate, but a strict and stable detection environment is required. When detecting curved surfaces with small curvature radii, the efficiency is low. Chinese invention patent application CN113970560A (A three-dimensional defect detection method based on multi-sensor fusion) proposes a three-dimensional defect detection method based on multi-sensor fusion. On the basis of using a microscopic imaging head to perform two-dimensional detection of surface defects of optical components, a white light interferometric measurement head is used to perform high-precision interference detection of the defect position according to the defect positioning mark result, so as to obtain defect depth data, and then complete the three-dimensional detection of surface defects of optical components; Chinese invention patent application CN116718551A (An optical component surface defect measurement device) uses low-magnification dark field scattering combined with high-magnification bright and dark field imaging to quickly find surface defects of optical components and accurately measure defect sizes, thereby ensuring the efficiency and accuracy of defect identification and classification; US invention patent application US202318221043A (Inspection system and method for analyzing defects) provides a defect detection system and detection method, which analyzes defects of a wide range of products in a simple and reliable manner by detecting multi-color light beams reflected on the surface to be tested; US invention patent application US202318124225A (Defect detection The device can conveniently indicate the position in the image by simultaneously displaying one or more types of vibration state images and two or more images in the optical image for the same defect in the detection area on the image display unit. Although the above patent application can realize high-precision detection of surface defects of optical components, the system structure is relatively complex and it is mainly aimed at surface defect detection of planar optical components. Further research is still needed for high-precision, high-resolution and high-efficiency three-dimensional detection of weak defects on the surface of curved optical components.

Summary of the invention

In order to solve the problems of small dynamic range, complex operation and strict environmental requirements of existing curved optical component surface defect detection technology, the present invention proposes a curved optical component surface defect detection device and method based on composite lighting. By integrating bright field/dark field microscopic scattering imaging and microstructured illumination detection, and scanning and fusing the surface, the complete two-dimensional information such as position and size of full-aperture surface defects and three-dimensional depth information are effectively obtained, thereby realizing high-precision, high-resolution and large dynamic range curved surface defect detection, with obvious advantages.

To achieve the above-mentioned purpose, the present invention adopts the following technical scheme: a surface defect detection device for curved optical elements based on composite illumination, the device comprising a microscope tube, a camera, an illumination light source, an electrically controlled displacement stage and a computer; the computer is connected to the illumination light source and the camera; the illumination light source comprises an off-axis illumination light source and a coaxial illumination light source, and the coaxial illumination light source comprises a white light illumination module and a structured light illumination module; the computer controls the light intensity and angle of the off-axis illumination light source according to the morphology of the optical element to be measured, so as to realize the rapid positioning of the surface defects; the computer controls the coaxial illumination light source to switch the white light illumination module and the structured light illumination module, the coaxial illumination light source forms a real image above the measured surface through the microscope tube, and is collected by the camera after being reflected by the surface of the element; the camera observes the measured surface through the microscope tube, and the measured surface forms a real image on the camera chip through the microscope tube; the white light illumination module of the coaxial illumination light source is switched, and the image is displayed by the camera chip by controlling the electrically controlled displacement stage; the camera detects the measured surface through the microscope tube, and the measured surface forms a real image on the camera chip through the microscope tube; the white light illumination module of the coaxial illumination light source is switched, and the image is displayed by controlling the electrically controlled displacement stage. The stage moves the component position to achieve automatic focusing on the component surface; the computer sets parameters according to the pre-stored program to make the structured light illumination module produce sinusoidal stripes orthogonal in the horizontal and vertical directions; the position where the defect exists on the component surface has a microscopic three-dimensional mutation, which causes a mutation in the phase distribution and a decrease in contrast; the computer calculates the collected sinusoidal fringe image containing the phase distribution of the component surface to obtain the surface shape of the component and three-dimensional information of the defect; the computer controls the electric-controlled translation stage to move longitudinally, focuses on the surface at different focal planes in the current field of view and collects the current image, processes and fuses the collected image, and expands the dynamic range of detection; the computer controls the electric-controlled translation stage to move horizontally along an "S"-shaped path, collects the surface image in each field of view, obtains the full-aperture surface defect information through three-dimensional stitching and fusion, and realizes three-dimensional detection and quantitative evaluation of weak surface defects in the full field of view.

Furthermore, the off-axis illumination light sources are symmetrically distributed, the coaxial illumination light source is placed on the side of the microscope tube, and coaxial illumination is achieved by deflection through a half-reflecting half-mirror; the camera is placed above the microscope tube; the electric-controlled translation stage is used to place the optical element to be tested, so that the optical element to be tested can be moved or deflected laterally or longitudinally relative to the microscope tube; the optical element to be tested is placed on the focal plane of the microscope tube.

Furthermore, a color camera is used to acquire a color image, and the color difference feature can be used to automatically focus on the component surface.

Furthermore, under illumination of each light source mode, the imaging field of view is the same field of view.

The present invention also proposes a detection method according to the above-mentioned curved optical element surface defect detection device based on composite illumination, comprising the following steps:

S1. Place the optical component to be tested on an electrically controlled translation stage, connect an off-axis illumination source, adjust the brightness and angle of the light source according to the surface morphology of the component to be tested, quickly locate the defect by adjusting the translation stage, and obtain a machine vision image of the bright defect against the dark background in the current field of view;

S2, switch the coaxial illumination light source, select the white light illumination module, control the movement of the translation stage and automatically focus on the surface of the component, present the effect of dark defects under bright background for the defects in the same field of view of S1, and collect the machine vision image of the current defect;

S3, select the structured light illumination module, project the fringe pattern onto the surface in the same field of view as S1 and collect the fringe image after being reflected by the surface, perform phase calculation and three-dimensional reconstruction on the fringe image, and obtain the modulation degree, surface shape and defect three-dimensional information in the current field of view;

S4, select the white light illumination module, control the longitudinal movement of the translation stage, perform longitudinal scanning on the component surface in the current field of view, repeat S2-S3 at each focal plane, and obtain machine vision images and fringe images of defects at different focal planes of the component;

S5, performing image fusion on the two-dimensional and three-dimensional images at different focal planes obtained in S4 to obtain complete three-dimensional information of surface defects in the current field of view;

S6. Control the lateral movement of the translation stage, repeat S2-S5 in each field of view, perform three-dimensional stitching and fusion on the surface defects in all the fields of view, and obtain three-dimensional information on the surface defects of the full-aperture curved optical element.

Furthermore, in S2, the step of controlling the translation stage to move and automatically focusing on the surface of the component includes:

S2.1. The computer controls the longitudinal movement of the translation stage. The camera collects each frame of the image and calculates the square of the grayscale difference between two adjacent pixels according to the following formula to obtain the image focus value:

Where f(x,y) represents the grayscale value of the pixel (x,y) corresponding to image f, and D(f) is the result of image definition calculation;

S2.2, performing peak search on the acquired image to identify the focus peak;

S2.3, the image position corresponding to the peak value is the focus position, and the blurriness of the current frame image is evaluated;

S2.4, judge the color difference of the collected image, observe the blur degree or obvious color edge around the defect, and verify the focus position of S2.3;

S2.5. If the focus positions of S2.3 and S2.4 are the same, the automatic focus check is completed, otherwise S2.1-S2.4 are repeated.

Furthermore, in S3, the step of using structured light to perform three-dimensional detection of component surface defects includes:

S3.1. Calibrate the camera and system geometric structure position parameters of the curved optical element surface defect measurement device, and establish the relationship between the camera coordinate system and phase, and the phase and world coordinate system;

S3.2. The computer generates n sinusoidal fringe patterns in the x-direction and y-direction, with the phase difference between adjacent phase patterns being 2π/n, and displays them on the OLED display. The fringe function after reflection from the optical surface to be measured Expressed as:

Where a and b are normal quantities, is the initial phase, (x, y) is the position of the pixel, is the fringe period, is the phase value introduced by the optical surface to be measured, and its distribution can be obtained by using the phase shift method. According to the n-step phase shift in the x and y directions, the camera collects n sinusoidal fringe patterns containing phase mutations with a phase difference of 2π/n in the x and y directions respectively;

S3.3. Perform phase unwrapping on the sinusoidal fringe pattern containing phase mutations to obtain the phase value and modulation Mod, through the phase value Obtain the slope change α, slope change α and phase value of the surface morphology of the optical element to be measured The relationship is as follows:

Where, d is the distance from the fringe conjugate plane to the surface to be measured; p is the period of the sinusoidal fringes at the fringe conjugate plane;

S3.4. The local slope α of the surface topography of the optical element to be measured is integrated using the Sauchwell reconstruction model to obtain the surface shape S that characterizes the surface height distribution information of the optical element to be measured. The gradient and height value between adjacent pixels are expressed as follows:

Among them, α _i,j , S _i,j represent the slope and height values represented by the point (i, j), , It is the distance between two adjacent pixels in the vertical direction; the superscript x, y and subscript x, y both represent the x direction and the y direction;

S3.5. The surface microscopic morphology of the optical element to be measured is measured by polynomial fitting calculation to obtain three-dimensional information of the defect.

Furthermore, in S5, the step of fusing images at different focal planes includes:

S5.1. Take the two-dimensional images A and B at different focal planes in S4, perform N-layer wavelet decomposition on A and B respectively, and extract their respective low-frequency components A _1N , B _1N and high-frequency components A _2N , A _3N , A _4N , B _2N , B _3N , B _4N ;

S5.2. The low-frequency components A _1N and B _1N are fused by weighted average method to obtain a fused low-frequency image AB _1N , as shown in the following formula:

Where (i, j) represents the position of the pixel, N represents the number of layers after wavelet decomposition, A _1N and B _1N represent the pixel values corresponding to the low-frequency component coefficients, AB _1N represent the low-frequency component coefficients after fusion, and a ₁ and a ₂ represent the weighting coefficients during fusion (a ₁ +a ₂ =1);

S5.3, the high-frequency components A _2N and B _2N are fused by using the coefficient absolute value larger method to obtain a fused high-frequency image AB _2N ;

S5.4, repeat S5.3 for the high-frequency components A _3N , B _3N and A _4N , B _4N to obtain fused high-frequency images AB _3N , AB _4N ;

S5.5, integrating the fused AB _1N , AB _2N , AB _3N , and AB _4N again through inverse wavelet transform to obtain a fused image AB;

S5.6. According to the wavelet coefficients of each component in S5.5, the three-dimensional images corresponding to A and B are fused to obtain a fused three-dimensional image, thereby obtaining complete three-dimensional information of surface defects in the current field of view.

Furthermore, in S6, the step of three-dimensionally stitching and fusing the surface defects of all viewing fields includes:

S6.1. Control the electric-controlled translation stage to move in an "S" shape to collect fringe patterns in each field of view, while ensuring that the overlapping areas of two adjacent fields of view are consistent;

S6.2, according to S3, obtain the modulation map Mod and the surface shape S of each field of view image;

S6.3. The SIFT algorithm is used to extract feature points from the modulation map Mod of the adjacent images to obtain the number, position, scale and direction of feature points. The modulus m(x, y) and direction θ(x, y) of the gradient are as follows:

Where (x, y) represents the position of the image pixel, and L is the scale space value of the key point;

S6.4. The surface shapes of the corresponding adjacent fields of view are registered and fused according to the feature point matrix to obtain the full-aperture surface shape S.

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

a. The present invention combines the bright field/dark field microscopic scattering imaging method and the microstructure illumination detection method, realizes the cross-integration of multiple methods, has a compact system structure and simple operation, and can quickly realize the surface defect detection of curved optical components;

b. The present invention combines physical and image judgment methods to achieve automatic focusing on the surface of the component;

c. The present invention uses the wavelet transform method to achieve the fusion of multi-focus images and expand the dynamic range of the system;

d. The present invention combines the method of image feature point recognition to perform three-dimensional stitching and fusion, thereby improving the accuracy of fusion;

e. The present invention switches the light source under the premise of keeping the field of view unchanged, omitting the complicated positioning and calibration process between different methods, and improving the accuracy of defect information extraction and fusion;

f. The present invention uses a color camera, which is more conducive to distinguishing dust and pollutants, thereby improving the accuracy of the detection results.

BRIEF DESCRIPTION OF THE DRAWINGS

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

FIG1 is a schematic structural diagram of a device for detecting surface defects of a curved optical element based on composite illumination according to the present invention;

FIG2 is a flow chart of a method for detecting surface defects of a curved optical element based on composite illumination according to the present invention;

FIG3a and FIG3b are surface defect images captured by a camera under off-axis white light source illumination and on-axis white light source illumination in the same field of view in a specific embodiment of the device of the present invention;

4a and 4b are sinusoidal fringe diagrams containing surface defect information in the x-direction and the y-direction in a specific embodiment of the device of the present invention; FIG. 4c is a schematic diagram of three-dimensional defect information;

FIG5 is a schematic diagram of a longitudinal fusion result of an image in a specific embodiment of the device of the present invention;

Fig. 6a is a lateral scanning path of the device of the present invention, Fig. 6b is a schematic diagram of lateral image fusion in a specific embodiment of the device of the present invention, and Fig. 6c is a schematic diagram of defect quantity, positioning, and area detection results in a specific embodiment of the present invention.

In the attached drawings: 1—microscope tube, 2—camera, 3—coaxial illumination source, 4—off-axis illumination source, 5—electrically controlled translation stage, 6—computer.

Detailed ways

In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and do not constitute a limitation of the present invention.

As shown in FIG1 , the surface defect detection device of the curved optical element based on composite illumination of the present invention comprises a microscope tube 1, a camera 2, a coaxial illumination light source 3, an off-axis illumination light source 4, an electrically controlled displacement stage 5 and a computer 6; the computer 6 is connected to the light source and the camera 2; the coaxial illumination light source 3 further comprises a white light illumination module and a structured light illumination module; the computer 6 controls the light intensity and angle of the off-axis illumination light source according to the morphology of the optical element to be measured, so as to realize the rapid positioning of the surface defects; the computer 6 controls the coaxial illumination light source 3 to switch between the white light illumination module and the structured light illumination module, the coaxial illumination light source 3 forms a real image above the measured surface through the microscope tube 1, and after being reflected by the surface of the element, it is collected by the camera 2; the camera 2 observes the measured surface through the microscope tube 1, and the measured surface forms a real image on the camera chip through the microscope tube 1; the coaxial white light illumination mode is switched, and the position of the element is moved by controlling the electrically controlled displacement stage 5 , realizing automatic focusing on the component surface; the computer 6 sets parameters according to the pre-stored program so that the structured light illumination module generates sinusoidal stripes orthogonal to the horizontal and vertical directions; the position where the defect exists on the component surface has a microscopic three-dimensional mutation, resulting in a mutation of the phase distribution and a decrease in contrast; the computer 6 calculates the collected sinusoidal fringe diagram containing the phase distribution to obtain the surface shape and defect information of the component; the computer 6 controls the electric-controlled displacement stage 5 to move longitudinally, focuses on the surface at different focal planes in the current field of view and collects the current image, processes and fuses the collected image, and expands the dynamic range of detection; the computer 6 controls the electric-controlled displacement stage 5 to move horizontally along an "S"-shaped path, collects the surface image in each field of view, obtains the full-aperture surface defect information through three-dimensional stitching and fusion methods, and realizes the three-dimensional detection and quantitative evaluation of the surface weak defects in the full field of view.

Among them, the off-axis illumination light source 4 is symmetrically distributed, the coaxial illumination light source 3 is placed on the side of the microscope tube 1, and coaxial illumination is realized by deflection through a semi-reflective and semi-mirror lens; the camera 2 is placed above the microscope tube 1; the electric-controlled displacement stage 5 is used to place the optical element to be measured, so that the optical element to be measured can be moved or deflected horizontally and vertically relative to the microscope tube 1; the optical element to be measured is placed on the focal plane of the microscope tube.

Based on the above-mentioned curved optical element surface defect detection device based on composite illumination, the present invention is further described through a specific embodiment. The curved optical element surface defect detection method based on composite illumination of the present invention comprises the following steps, as shown in FIG2:

S1, place the optical component to be tested on the electric-controlled translation stage 5, connect the off-axis illumination light source 4, adjust the brightness and angle of the light source according to the surface morphology of the component to be tested, quickly locate the defect by adjusting the electric-controlled translation stage 5, and obtain the machine vision image of the bright defect under the dark background in the current field of view, as shown in FIG3a;

S2, switch the coaxial illumination light source 3, select the white light module, control the electric control translation stage 5 to move and automatically focus on the surface of the component, present the effect of dark defects under bright background for the defects in the same field of view of S1, and collect the machine vision image of the current defect, as shown in FIG3b;

S3, select the structured light module, project the fringe pattern onto the surface in the same field of view as S1 and collect the fringe image after the surface reflection, as shown in Figures 4a and 4b, perform phase solution and three-dimensional reconstruction on the fringe image, and obtain the surface modulation, surface shape and three-dimensional defect information in the current field of view. As shown in Figure 4c, the length of the defect is 551.73um, the width is 14.87um, and the depth is 198.52nm;

S4, select the white light module, control the electric-controlled translation stage 5 to move longitudinally, perform longitudinal scanning on the component surface in the current field of view, repeat S2-S3 at each focal plane, and obtain machine vision images and fringe images of defects at different focal planes of the component;

S5, performing image fusion on the two-dimensional and three-dimensional images at different focal planes obtained in S4 to obtain complete three-dimensional information of surface defects in the current field of view, as shown in FIG5 ;

S6, control the electric-controlled translation stage 5 to move laterally, and the movement path is shown in Figure 6a, and S2-S5 are repeated in each field of view, and the surface defects in all fields of view are three-dimensionally stitched and fused to obtain the surface defect information of the full-aperture curved optical element, as shown in Figures 6b and 6c.

Furthermore, in S2, the step of controlling the electric-controlled translation stage 5 to move and automatically focusing the surface of the component includes:

S2.1. The computer 6 controls the electric-controlled displacement stage 5 to move longitudinally. The camera 2 collects each frame of the image and calculates the square of the grayscale difference between two adjacent pixels according to the following formula to obtain the image focus value:

Where f(x,y) represents the grayscale value of the pixel (x,y) corresponding to image f, and D(f) is the result of image definition calculation;

S2.2, performing peak search on the acquired image to identify the focus peak;

S2.3, the image position corresponding to the peak value is the focus position, and the blurriness of the current frame image is evaluated;

S2.4, judge the color difference of the collected image, observe the blur degree or obvious color edge around the defect, and verify the focus position of S2.3;

S2.5. If the focus positions of S2.3 and S2.4 are the same, the automatic focus check is completed, otherwise S2.1-S2.4 are repeated.

Furthermore, in S3, the step of using a structured light module to perform three-dimensional detection of component surface defects includes:

S3.1. Calibrate the camera and system geometric structure position parameters of the curved optical element surface defect measurement device, and establish the relationship between the camera coordinate system and phase, and between the phase and the world coordinate system;

S3.2. The computer generates n sinusoidal fringe patterns in the x-direction and y-direction, with a phase difference of 2 π/n between adjacent phase patterns, and displays them on the OLED display. The fringe function after reflection from the optical surface to be measured Expressed as:

Where a and b are normal quantities, is the initial phase, (x, y) is the position of the pixel, is the fringe period, The phase value introduced by the optical surface to be measured, the distribution of which can be obtained by using the phase shifting technology. According to the n-step phase shifting in the x and y directions, the camera collects n sinusoidal fringe patterns containing phase mutations with a phase difference of 2π/n in the x and y directions respectively;

S3.3. Perform phase unwrapping on the sinusoidal fringe pattern containing phase mutations to obtain the phase value and modulation Mod, through the phase value Obtain the slope change of the surface morphology of the optical element to be measured , the slope changes and phase value The relationship is as follows:

in, is the distance from the fringe conjugate plane to the surface to be measured; is the period of the sinusoidal fringes at the conjugate plane of the fringes;

S3.4. The local slope α of the surface topography of the optical element to be measured is integrated using the Sauchwell reconstruction model to obtain the surface shape S that characterizes the surface height distribution information of the optical element to be measured. The gradient and height value between adjacent pixels are expressed as follows:

Among them, α _i,j , S _i,j represent the slope and height values represented by the point (i, j), , It is the distance between two adjacent pixels in the vertical direction; the superscript x, y and subscript x, y both represent the x direction and the y direction;

S3.5. The surface microscopic morphology of the optical element to be measured is measured by polynomial fitting calculation to obtain three-dimensional information of the defect.

Furthermore, in S5, the step of fusing images at different focal planes includes:

S5.1. Take the two-dimensional images A and B at different focal planes in S4, perform N-layer wavelet decomposition on A and B respectively, and extract their respective low-frequency components A _1N , B _1N and high-frequency components A _2N , A _3N , A _4N , B _2N , B _3N , B _4N ;

S5.2. The low-frequency components A _1N and B _1N are fused by weighted average method to obtain a fused low-frequency image AB _1N , as shown in the following formula:

Where (i, j) represents the position of the pixel, N represents the number of layers after wavelet decomposition, A _1N and B _1N represent the pixel values corresponding to the low-frequency component coefficients, AB _1N represent the low-frequency component coefficients after fusion, and a ₁ and a ₂ represent the weighting coefficients during fusion (a ₁ +a ₂ =1);

S5.3, the high-frequency components A _2N and B _2N are fused by using the coefficient absolute value larger method to obtain a fused high-frequency image AB _2N ;

S5.4, repeat S5.3 for the high-frequency components A _3N , B _3N and A _4N , B _4N to obtain fused high-frequency images AB _3N , AB _4N ;

S5.5, integrating the fused AB _1N , AB _2N , AB _3N , and AB _4N again through inverse wavelet transform to obtain a fused image AB;

S5.6. According to the wavelet coefficients of each component in S5.5, the three-dimensional images corresponding to A and B are fused to obtain a fused three-dimensional image, thereby obtaining complete three-dimensional information of surface defects in the current field of view.

Furthermore, in S6, the step of three-dimensionally stitching and fusing the surface defects of all viewing fields includes:

S6.1. Control the electric-controlled translation stage to move in an "S" shape to collect fringe patterns in each field of view, while ensuring that the overlapping areas of two adjacent fields of view are consistent;

S6.2, according to S3, obtain the modulation map Mod and the surface shape S of each field of view image;

S6.3. The SIFT algorithm is used to extract feature points from the modulation map Mod of the adjacent images to obtain the number, position, scale and direction of feature points. The modulus m(x, y) and direction θ(x, y) of the gradient are as follows:

Where (x, y) represents the position of the image pixel, and L is the scale space value of the key point;

S6.4. The surface shapes of the corresponding adjacent fields of view are registered and fused according to the feature point matrix to obtain the full-aperture surface shape S.

Although the preferred embodiments of the present invention have been disclosed for illustrative purposes, those skilled in the art will appreciate that various modifications, additions and substitutions are possible, without departing from the scope and spirit of the invention as disclosed in the accompanying claims.

Claims (9)

1. The device is characterized by comprising a microscope tube, a camera, an illumination light source, an electric control displacement table and a computer; the computer is connected with the illumination light source and the camera; the illumination light source comprises an off-axis illumination light source and an on-axis illumination light source, and the on-axis illumination light source comprises a white light illumination module and a structural light illumination module; the computer controls the light intensity and the angle of the off-axis illumination light source according to the appearance of the optical element to be detected, so as to realize the rapid positioning of the surface defect; the computer controls the coaxial illumination light source to switch the white light illumination module and the structural light illumination module, the coaxial illumination light source forms a real image above the surface to be measured through the microscope tube, and the coaxial illumination light source is collected by the camera after being reflected by the element surface; the camera observes the surface to be measured through the microscope tube, and the surface to be measured forms a real image on the camera chip through the microscope tube; the white light illumination module of the coaxial illumination light source is switched, and the automatic focusing on the surface of the element is realized by controlling the position of the element of the electric control displacement table; the computer sets parameters according to a pre-stored program to enable the structured light illumination module to generate sine stripes orthogonal in the horizontal direction and the vertical direction; the position of the defect on the surface of the element has microscopic three-dimensional mutation, which causes mutation of phase distribution and reduction of contrast; the computer calculates the acquired sinusoidal fringe pattern containing the phase distribution of the element surface to obtain the three-dimensional information of the surface shape and the defect of the element surface; the computer controls the electric control displacement table to longitudinally move, focuses the surfaces at different focal planes under the current view field and collects the current image, processes and fuses the collected images, and expands the dynamic range of detection; the computer controls the electric control displacement table to transversely move according to the S-shaped path, surface images under each view field are collected, and the full-caliber surface defect information is obtained through three-dimensional splicing and fusion, so that the three-dimensional detection and quantitative evaluation of the surface weak defects of the full view field are realized.

2. The composite illumination-based curved surface optical element surface defect detection device according to claim 1, wherein the off-axis illumination light sources are symmetrically distributed, the coaxial illumination light sources are arranged on the side surface of the microscope tube, and the coaxial illumination is realized by deflection through the semi-reflection semi-transparent mirror; the camera is arranged above the microscope tube; the electric control displacement table is used for placing the optical element to be measured, so that the optical element to be measured can move or deflect transversely and longitudinally relative to the microscope tube; the optical element to be measured is placed on the focal plane of the microscope tube.

3. The device for detecting surface defects of a curved optical element based on compound illumination according to claim 1, wherein a color camera is used to obtain a color image, and the color difference feature is used to automatically focus the surface of the element.

4. The composite illumination-based surface defect detection device for a curved optical element according to claim 1, wherein under illumination of each light source mode, the imaging fields of view are the same field of view.

5. A method for detecting a surface defect of a curved optical element based on composite illumination according to any one of claims 1 to 4, comprising the steps of:

S1, placing an optical element to be detected on an electric control displacement table, connecting an off-axis illumination light source, adjusting the brightness and the angle of the light source according to the surface morphology of the element to be detected, rapidly positioning the defect by adjusting the displacement table, and acquiring a machine vision image of a bright defect under a dark background in a current field of view;

s2, switching the coaxial illumination light source, selecting a white light illumination module, controlling the displacement table to move and automatically focusing the surface of the element, presenting the effect of a dark defect under a bright background to the defect in the same view field of S1, and collecting a machine vision image of the current defect;

S3, selecting a structured light illumination module, projecting a fringe image to the surface in the same view field of the S1, collecting the fringe image after surface reflection, carrying out phase resolving and three-dimensional reconstruction on the fringe image, and obtaining modulation degree, surface shape and defect three-dimensional information under the current view field;

S4, selecting a white light illumination module, controlling the displacement table to longitudinally move, longitudinally scanning the surface of the element under the current view field, repeating S2-S3 at each focal plane, and obtaining machine vision images and stripe images of defects at different focal planes of the element;

s5, performing image fusion on the two-dimensional and three-dimensional images at different focal planes acquired in the S4 to obtain complete three-dimensional information of the surface defects under the current view field;

S6, controlling the displacement table to transversely move, repeating the steps S2-S5 under each view field, and performing three-dimensional splicing and fusion on the obtained surface defects under all view fields to obtain the three-dimensional information of the surface defects of the full-caliber curved surface optical element.

6. The method for detecting surface defects of a curved optical element based on compound illumination according to claim 5, wherein in S2, the step of controlling the displacement stage to move and automatically focus the surface of the element comprises:

s2.1, controlling the displacement platform to longitudinally move by a computer, collecting each frame of image by a camera, and calculating the square of the gray level difference of two adjacent pixels according to the following formula to obtain an image focusing value:

,

wherein f (x, y) represents the gray value of the corresponding pixel point (x, y) of the image f, and D (f) is the image definition calculation result;

s2.2, carrying out peak search on the acquired image, and judging a focusing peak value;

s2.3, evaluating the current frame image ambiguity by taking the image position corresponding to the peak value as a focusing position;

S2.4, performing color difference judgment on the acquired image, observing the blurring degree or obvious color edges around the defect, and verifying the focusing position of S2.3;

and S2.5, if the focusing positions of the S2.3 and the S2.4 are the same, completing automatic focusing, otherwise repeating the steps of S2.1-S2.4.

7. The method for detecting surface defects of a curved optical element based on compound illumination according to claim 5, wherein in S3, the step of three-dimensionally detecting the surface defects of the element using structured light comprises:

s3.1, calibrating position parameters of a camera and a system geometric structure of the surface defect measuring device of the curved optical element, and establishing a relation between a camera coordinate system and a phase, a phase and a world coordinate system;

s3.2, generating n sine fringe patterns in the x direction and the y direction respectively by a computer, wherein the phase difference between adjacent phase patterns is 2 pi/n, and displaying the phase difference on an OLED display screen; fringe function after reflection by optical surface to be measured Expressed as:

,

Wherein a and b are normal amounts respectively, For the initial phase, (x, y) is the position of the pixel,In the form of a fringe period,The phase value introduced by the optical surface to be measured can be obtained by a phase shift method, and according to n steps of phase shift in the x direction and the y direction, the camera respectively acquires n sine fringe patterns containing phase mutation, wherein the n phase difference is 2 pi/n, in the x direction and the y direction;

S3.3, performing phase unwrapping on the sinusoidal fringe pattern containing the phase mutation to obtain a phase value And a modulation Mod by a phase valueObtaining the slope change of the surface morphology of the optical element to be measuredSlope changeAnd phase valueThe relation of (2) is as follows:

,

Wherein d is the distance from the conjugate surface of the stripe to the surface to be measured; p is the period of a sine stripe at the conjugate plane of the stripe;

S3.4, integrating the local slope alpha of the surface morphology of the optical element to be detected by adopting a Shaoxing-Wilker reconstruction model to obtain the surface shape S of the surface-height distribution information of the optical element to be detected, wherein the gradient and the height value between adjacent pixels are expressed as the following formula:

,

Where α _i,j, S_i,j represents the slope and height values represented by the (i, j) point, and d _x、d_y is the distance between two vertically adjacent pixels; the superscript x, y and the subscript x, y both represent the x direction and the y direction;

And S3.5, completing measurement of the surface microscopic morphology of the optical element to be measured through polynomial fitting calculation, and obtaining the three-dimensional information of the defect.

8. The method for detecting surface defects of a curved optical element based on compound illumination according to claim 5, wherein in S5, the step of performing image fusion on images at different focal planes includes:

S5.1, taking two-dimensional images A and B at different focal planes in S4, respectively carrying out N-layer wavelet decomposition on the images A and B, and extracting respective low-frequency components A _1N、B_1N and high-frequency components A _2N、A_3N、A_4N、B_2N、B_3N、B_4N;

S5.2, fusing the low-frequency components A _1N、B_1N by adopting a weighted average method to obtain a fused low-frequency image AB _1N, wherein the following formula is adopted:

,

Wherein, (i, j) represents the position of the pixel point, N represents the number of layers subjected to wavelet decomposition, a _1N、B_1N represents the pixel value corresponding to the low-frequency component coefficient, AB _1N represents the low-frequency component coefficient after fusion, and a ₁、a₂ represents the weighting coefficient (a ₁+a₂ =1) at fusion;

S5.3, fusing the high-frequency components A _2N、B_2N by adopting a method with larger absolute coefficient value to obtain a fused high-frequency image AB _2N;

S5.4, repeating the step S5.3 on the high-frequency components A _3N、B_3N and A _4N、B_4N to obtain a fused high-frequency image AB _3N、AB_4N;

S5.5, integrating the fused AB _1N、AB_2N、AB_3N、AB_4N again through wavelet inverse transformation to obtain a fused image AB;

And S5.6, fusing the three-dimensional images corresponding to the A and the B according to the wavelet coefficients of the components in the S5.5 to obtain the fused three-dimensional image, thereby obtaining complete three-dimensional information of the surface defect under the current view field.

9. The method for detecting surface defects of a curved optical element based on compound illumination according to claim 5, wherein in S6, the step of three-dimensionally stitching and fusing the surface defects of all fields of view comprises:

S6.1, controlling the electric control displacement table to move along the S shape, collecting fringe patterns of each view field, and simultaneously ensuring that overlapping areas of two adjacent view fields are consistent;

S6.2, obtaining a modulation degree map Mod and a surface shape S of each view field image according to the S3;

S6.3, extracting feature points from a modulation degree map Mod of the adjacent images by adopting a SIFT algorithm to obtain the number, the positions, the scales and the directions of the feature points, and the modulus value m (x, y) and the direction theta (x, y) of the gradient are as follows:

,

wherein (x, y) represents the position of the image pixel, and L is the scale space value of the key point;

and S6.4, registering and fusing the surface shapes of the corresponding adjacent fields according to the characteristic point matrix to obtain a full-caliber surface shape S.