Transparent element surface shape detection method

Abstract

The invention discloses a transparent element surface shape detection method. The method mainly solves the technical problem that an existing transparent element surface shape detection method is difficult to comprehensively and efficiently eliminate parasitic reaction to achieve accurate surface shape detection. After a transparent element detection system is constructed, a display is controlled to respectively display a phase-shift fringe pattern and a multi-frequency fringe pattern, and after being reflected by a transparent element, the fringe pattern is correspondingly collected by a camera and a superposed fringe pattern is generated; then, a superposition phase is obtained through corresponding calculation of the superposition fringe pattern and converted into superposition display coordinates; and extracting the intensity value of the superposed fringe pattern, forming a time intensity sequence, and correspondingly separating parasitic reflection through correlation matching and nonlinear optimization to complete the surface shape detection of the front surface of the transparent element. The whole detection process is efficient and convenient and is not limited by types of transparent elements, and the detection accuracy is high.

Description

Technical Field

[0001] The present invention relates to the field of optical detection technology, and in particular to a method for detecting the surface shape of a transparent element. Background Art

[0002] Planar transparent components are widely used in modern industry in optical instruments, electronic equipment, precision manufacturing and other fields. The surface quality of planar transparent components directly affects the performance and reliability of the product. The traditional method of transparent component detection is to use deflectometry detection, which obtains component surface information by measuring the deflection of light on the component surface. However, due to the reflection of the front and back surfaces of the transparent component, parasitic reflection will occur in the deflectometry detection, so that the detection result contains the reflection superposition information of the front and back surfaces of the transparent component, thereby affecting the accuracy of the detection.

[0003] In order to achieve accurate detection of the surface shape of lossless components, the inventor knows a special light source method, which utilizes the characteristic of transparent components that do not refract in the ultraviolet light band to eliminate the adverse effects of the above-mentioned parasitic reflections, so as to achieve detection of the surface shape of transparent components, but the types of detection elements of this detection method are limited. In addition, the inventor also knows a spatial domain method, which determines the area containing only the front surface reflection by displaying line structured light, binary patterns or dot patterns, but this method is difficult to quickly and accurately determine the appropriate threshold to determine the area. There is also a frequency domain method, which regards the parasitic reflection as the superposition of two sinusoidal signals and uses signal analysis methods to separate the parasitic reflection. In order to obtain the power spectrum distribution or envelope zero point of the sequence, the sampling frequency of the time intensity sequence needs to meet the requirements of the Nyquist sampling theorem, which results in the number of multi-frequency fringe images reaching thousands of images, resulting in low efficiency in transparent component surface shape detection.

[0004] The information disclosed in this background technology section is only used to deepen the understanding of the background technology of the present disclosure, and should not be regarded as acknowledging or suggesting in any form that the information constitutes the prior art known to those skilled in the art. Summary of the invention

[0005] In view of at least one of the above technical problems, the present disclosure provides a transparent component surface shape detection method, aiming to solve the technical problem that the existing transparent component surface shape detection method is difficult to completely and efficiently eliminate parasitic reactions and achieve accurate surface shape detection.

[0006] According to one aspect of the present disclosure, a method for detecting a transparent element surface shape is provided, which comprises the following steps: (1) Constructing a transparent component detection system including a display, a camera, and a transparent component to be tested; (2) controlling the display to display the phase-shift fringe pattern and the multi-frequency fringe pattern respectively, and the camera to collect and generate a superimposed fringe pattern after being reflected by the transparent element; (3) Obtaining the superimposed phase by corresponding calculation from the superimposed fringe pattern and converting it into the coordinates of the superimposed display; (4) Extracting the intensity values ​​of the superimposed fringe pattern and forming a time intensity sequence, and separating the parasitic reflections through correlation matching and nonlinear optimization to complete the front surface shape detection of the transparent element.

[0007] In some embodiments of the present disclosure, in step (1), the relative positions of the display, the camera, and the transparent element to be tested are adjusted accordingly, so that the transparent element to be tested completely reflects the display information of the display and the camera completely captures the reflection information of the transparent element to be tested.

[0008] In some embodiments of the present disclosure, in step (3), the phase unwrapping is used to obtain x The superposition phase of the direction; x Overlay fringe pattern in direction I x Expressed as ; Superposition truncated phase for: ; y Overlay fringe pattern in direction I y Expressed as ; Superposition truncated phase for: ; in, To superimpose fringe patterns x Directional background light intensity; Indicates the display coordinates corresponding to the front surface of the transparent element S 1 point x phase information of the direction, and ; Indicates the display coordinates corresponding to the back surface of the transparent element S 2 point x phase information of the direction, and ; M x1 and M x2 are the modulation degrees of the superimposed fringe patterns, respectively; f is the frequency of the superimposed fringe pattern; x S1 and x S2 They are S 1 and S2 point x Display coordinates of the direction; To superimpose fringe patterns y Directional background light intensity; Indicates the display coordinates corresponding to the front surface of the transparent element S 1 point y phase information of the direction, and ; Indicates the display coordinates corresponding to the back surface of the transparent element S 2 point y phase information of the direction, and ; M y1 and M y2 are the modulation degrees of the superimposed fringe patterns, respectively; f is the frequency of the superimposed fringe pattern; y S1 and y S2 They are S 1 and S 2 point y The display coordinates of the direction.

[0009] In some embodiments of the present disclosure, in step (3), x Direction and y The overlay display coordinates of the directions are: , ; After normalization, the corresponding , ; in, W is the width of the display; ( x s , y s ) is the overlay display coordinate, ( x sn , y sn ) are the normalized overlay display coordinates.

[0010] In some embodiments of the present disclosure, in step (4), the overlay display coordinates are used as the center position of the relevant matching.

[0011] In some embodiments of the present disclosure, in step (4), the sampling frequency is F , length is NThe time intensity series: ; and , , , , ; in,( x s1 , y s1 )and( x s2 , y s2 ) are S 1 Point and S 2 The display coordinates of the point, ( x sn1 , y sn1 )and( x sn2 , y sn2 ) are S 1 Point and S 2 The point's normalized display coordinates, τ is the number of cycles of the superimposed fringe pattern.

[0012] In some embodiments of the present disclosure, in step (4), the time series I x The spectrum interval Fs / N is the radius of the correlation matching.

[0013] In some embodiments of the present disclosure, in step (4), the matching interval is determined according to the center position of the relevant matching and the radius of the relevant matching: , ; Combine all values ​​in the two matching intervals to form corresponding frequency pairs ( ), determine the time intensity sequence corresponding to the frequency pair I cal ; and extract the time intensity sequence corresponding to the gray value of the superimposed fringe image I get ; Correlate the two time intensity series: ; in, and The time intensity series Ical and I get The mean of I cal，i and I get，i They are I cal and I get No. i Strength value; Pearson ( I cal ,I get ) is the correlation coefficient.

[0014] A correlation coefficient matrix is ​​constructed using the correlation coefficients corresponding to each frequency pair, and the frequency pair corresponding to the maximum value in the correlation coefficient matrix is ​​extracted as the initial value of the correlation matching.

[0015] In some embodiments of the present disclosure, in step (4), the nonlinear optimization includes the following sub-steps: (41) Extract the frequency pair corresponding to the maximum value in the correlation coefficient matrix ( ), based on the non-negative least squares method, the corresponding superimposed fringe pattern is calculated x Background light intensity in the direction , modulation degree of superimposed fringe pattern M x1 and M x2 After that, build I get = WF T ;in: .

[0016] (42) Constructing time-intensity series I’ : ; calculate I’ and I get Strength difference between diff : ; Based on Nelder-Mead Algorithm optimization and To strength difference diff When it is less than the threshold ε, the output x Display coordinates corresponding to the direction and .

[0017] (43) Repeat steps (41) and (42) to obtain the corresponding calculation y Display coordinates corresponding to the direction and .

[0018] (44) According to the display coordinates corresponding to the front surface of the transparent element ( ) are calculated separately x Direction and y The slopes corresponding to the directions are: ; in,( x c , y c )and( x m1 , y m1 ) are the coordinates of the entrance pupil center point C of the camera and the front surface of the transparent element M 1 Point coordinates; z m2c and z m2s The distance between the front surface of the transparent element and the center of the camera entrance pupil and the display are z Distance of direction; d m2c and d m2s are the distances from the front surface of the transparent element to the center of the camera entrance pupil and the display, respectively.

[0019] (45) Adoption Zernike The slope of the polynomial fitting is used to reconstruct the front surface shape of the transparent element.

[0020] One or more technical solutions provided in the embodiments of the present application have at least any of the following technical effects or advantages: The transparent element detection system has a simple structure and a small number of components, which allows for rapid construction of the detection system. At the same time, by separately displaying the phase-shift fringe pattern and the multi-frequency fringe pattern on the display, a superimposed fringe pattern is generated after being reflected by the transparent element to be tested and captured by the camera. Parasitic reflection separation and transparent element front surface shape detection can be achieved through correlation matching calculation and nonlinear optimization. The entire process is efficient and convenient, the detection results are reliable, it is not limited by the type of transparent element, and the detection efficiency is high. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 Schematic diagram of the structural principle of a transparent element surface detection system in one embodiment of the present application.

[0022] Figure 2 This is a superimposed phase-shift fringe pattern in an embodiment of the present application.

[0023] Figure 3 This is a superimposed multi-frequency fringe pattern in one embodiment of the present application.

[0024] In the above figures, 1 is a display, 2 is a camera, 3 is a transparent element, 31 is a front surface of the transparent element, and 32 is a rear surface of the transparent element. DETAILED DESCRIPTION

[0025] The programs involved or relied upon in the following embodiments are all conventional programs or simple programs in the technical field, and those skilled in the art can make conventional selections or adaptive adjustments according to specific application scenarios. In order to better understand the technical solution of this application, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0026] In order to solve the problem that the existing transparent component surface shape detection method cannot efficiently and reliably eliminate the adverse effects of parasitic reactions on the detection results, this example discloses a transparent component surface shape detection method, which includes the following steps: (1) Construct a transparent component inspection system including a display, a camera, and a transparent component to be inspected.

[0027] In this embodiment, see Figure 1 The transparent element surface shape detection system includes a display 1 and a camera 2, wherein the display 1 is used to display a phase-shifted fringe pattern or a multi-frequency fringe pattern, and the camera 2 is used to capture the reflection images of the fringe pattern displayed on the display by the front surface 31 and the rear surface 32 of the transparent element. Accordingly, in this example, the relative positions of the transparent element 3, the display 1 and the camera 2 are adjusted accordingly to ensure that the transparent element 3 can effectively and comprehensively reflect the fringe pattern displayed on the display 1, and the reflection information can be completely and clearly captured by the camera 2, thereby avoiding the problem that the transparent element cannot completely reflect all the fringe pattern information displayed on the display 1, and the camera 2 cannot completely capture all the fringe pattern images reflected by the transparent element, resulting in detection distortion and failure to accurately reflect the surface shape structure of the transparent element.

[0028] (2) The display is controlled to display the phase-shift fringe pattern and the multi-frequency fringe pattern respectively. After being reflected by the transparent element, the camera collects the corresponding fringe patterns and generates a superimposed fringe pattern.

[0029] In this embodiment, the display 1 is specifically connected to the computer for communication, and the computer program controls the display 1 to display the phase-shift fringe pattern and the multi-frequency fringe pattern respectively. After the fringe pattern is reflected by the front surface 31 of the transparent element and the rear surface 32 of the transparent element, the corresponding reflected image is captured by the camera. Then, the superimposed fringe pattern is generated based on the image captured by the camera 2. Specifically in this example, the camera 2 is connected to the computer for communication, and the computer generates the superimposed phase-shift fringe pattern correspondingly after acquiring the image captured by the camera (see Figure 2) and superimposed multi-frequency fringe patterns (see Figure 3 ).

[0030] (3) The superimposed phase is obtained by calculating the superimposed phase-shift fringe pattern and converted into the superimposed display coordinates.

[0031] In this embodiment, a phase shift algorithm is used to calculate the superimposed phase shift fringe pattern, from which the superimposed phase is obtained and then converted into the superimposed display coordinates.

[0032] Specifically, the phase shift fringe pattern first displayed by the display 1 is reflected by the front surface and the rear surface of the transparent element, and then recorded by the camera to generate a superimposed phase shift fringe pattern. Then: x Overlay fringe pattern in direction I x Expressed as ; y Overlay fringe pattern in direction I y Expressed as .

[0033] Use the phase unwrapping algorithm to calculate the superposition truncated phase separately φ x and φ y : ; .

[0034] in, To superimpose fringe patterns x Directional background light intensity; Indicates the display coordinates corresponding to the front surface of the transparent element S 1 point x phase information of the direction, and ; Indicates the display coordinates corresponding to the back surface of the transparent element S 2 point x phase information of the direction, and ; M x1 and M x2 are the modulation degrees of the superimposed fringe patterns, respectively; f is the frequency of the superimposed fringe pattern; x S1 and x S2 They are S 1 and S 2 pointx Display coordinates of the direction; To superimpose fringe patterns y Directional background light intensity; Indicates the display coordinates corresponding to the front surface of the transparent element S 1 point y phase information of the direction, and ; Indicates the display coordinates corresponding to the back surface of the transparent element S 2 point y phase information of the direction, and ; M y1 and M y2 are the modulation degrees of the superimposed fringe patterns, respectively; f is the frequency of the superimposed fringe pattern; y S1 and y S2 They are S 1 and S 2 point y The display coordinates of the direction.

[0035] Therefore, the superimposed truncated phase φ x and φ y Expand and get x Superposition phase of direction 𝛷 x and y Superposition phase of direction 𝛷 y .

[0036] Furthermore, we can get x Direction and y Overlay display coordinates of direction: , .

[0037] Normalize the overlay display coordinates: , ; in, W is the width of the display; ( x s , y s ) is the overlay display coordinate, ( x sn ,y sn ) are the normalized overlay display coordinates.

[0038] (4) The intensity values ​​of the superimposed fringe patterns are extracted and formed into a time intensity sequence. The parasitic reflections are separated by correlation matching and nonlinear optimization to complete the surface shape detection of the transparent element front surface.

[0039] In order to separate the parasitic reflection and improve the accuracy of detection, the parasitic reflection is separated by correlation matching and nonlinear optimization in this example. Considering that the superimposed display coordinates contain the reflection information of the transparent element front surface 31 and the transparent element rear surface 32, in this embodiment, the superimposed display coordinates are used as the center position of the correlation matching.

[0040] To determine the relevant matching radius, in this embodiment, after displaying the phase-shifted fringe pattern in step (3), the display 1 then displays fringe patterns of different frequencies. Similarly, after being reflected by the front and rear surfaces of the transparent element, the reflection information is captured by the camera 2 to generate a superimposed multi-frequency fringe pattern.

[0041] After obtaining the superimposed multi-frequency fringe pattern, the intensity value of the superimposed multi-frequency fringe pattern is extracted to form a sampling frequency of F , length is N The time intensity series of: ; and , , , , ; in,( x s1 , y s1 )and( x s2 , y s2 ) are S 1 Point and S 2 The display coordinates of the point, ( x sn1 , y sn1 )and( x sn2 , y sn2 ) are S 1 Point and S 2 The point's normalized display coordinates, τ is the number of cycles of the superimposed fringe pattern.

[0042] In order to effectively record the possible frequency changes in the time intensity series without using Fourier transform, in this embodiment, the time series is selected I x The spectrum interval Fs / N Serves as the radius for the relevant matches.

[0043] Furthermore, the matching interval can be determined according to the center position of the relevant matching and the radius of the relevant matching: , .

[0044] Combine all values ​​in the two matching intervals to form corresponding frequency pairs ( ), then the frequency pair is substituted into the above time intensity sequence formula to determine the time intensity sequence corresponding to the frequency pair. I cal Then extract the time intensity sequence obtained by superimposing the grayscale values ​​of the multi-frequency fringe image I get ; Correlate the two time intensity series: ; in, and The time intensity series I cal and I get The mean of I cal，i and I get，i They are I cal and I get No. i Strength value; Pearson ( I cal ,I get ) is the correlation coefficient. Therefore, through Pearson It can quickly evaluate the linear relationship between two sequences. Pearson When the coefficient is large, it indicates that there is a strong consistency between the two time intensity series. In addition, since each frequency pair corresponds to a correlation coefficient, this example constructs a correlation coefficient matrix with the correlation coefficients corresponding to each frequency pair, and extracts the frequency pair corresponding to the maximum value in the correlation coefficient matrix as the initial value of the correlation matching.

[0045] Furthermore, in this embodiment, obtaining and Then, the display coordinates are obtained by nonlinear iteration, and the nonlinear iteration has the following sub-steps: (41) Extract the frequency pair corresponding to the maximum value in the correlation coefficient matrix ( ), based on the non-negative least squares method, the corresponding superimposed fringe pattern is calculated x Background light intensity in the direction , modulation degree of superimposed fringe pattern M x1 and M x2 After that, build I get =WF T ;in: .

[0046] (42) Constructing time-intensity series I’ : ; calculate I’ and I get Strength difference between diff : ; Based on Nelder-Mead Algorithm optimization and To strength difference diff In this example, the threshold ε is set so that when the intensity difference diff When it is less than the threshold ε, the output x Normalized coordinates corresponding to the direction and , multiply the normalized coordinates by the display width to get the display coordinates and .

[0047] (43) Repeat steps (41) and (42) and calculate in the same way to obtain y Display coordinates corresponding to the direction and .

[0048] (44) After obtaining the display coordinates, according to the display coordinates corresponding to the front surface of the transparent element ( ) and system structure parameters, respectively calculate x Direction and y The slopes corresponding to the directions are: ; in,( x c , y c )and( xm1 , y m1 ) are the coordinates of the entrance pupil center point C of the camera and the front surface of the transparent element M 1 Point coordinates; z m2c and z m2s The distance between the front surface of the transparent element and the center of the camera entrance pupil and the display are z Distance of direction; d m2c and d m2s are the distances from the front surface of the transparent element to the center of the camera entrance pupil and the display, respectively.

[0049] (45) In obtaining x Direction and y After the direction slope, use Zernike The slope of the polynomial fitting is used to complete the surface reconstruction of the front surface of the transparent element.

[0050] Although some preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application.

[0051] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the inventive concept. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for detecting the surface shape of a transparent element, characterized in that: The steps include: (1) Constructing a transparent component detection system including a display, a camera, and a transparent component to be tested; (2) controlling the display to display the phase-shift fringe pattern and the multi-frequency fringe pattern respectively, and the camera to collect and generate a superimposed fringe pattern after being reflected by the transparent element; (3) Obtaining the superimposed phase by corresponding calculation from the superimposed fringe pattern and converting it into the coordinates of the superimposed display; (4) Extracting the intensity values ​​of the superimposed fringe pattern and forming a time intensity sequence, and separating the parasitic reflections through correlation matching and nonlinear optimization to complete the front surface shape detection of the transparent element.

2. The transparent element surface shape detection method according to claim 1, characterized in that: In the step (1), the relative positions of the display, the camera and the transparent element to be tested are adjusted accordingly, so that the transparent element to be tested completely reflects the display information of the display and the camera completely captures the reflection information of the transparent element to be tested.

3. The transparent element surface shape detection method according to claim 1, characterized in that: In step (3), the phase unwrapping is used to obtain x The superposition phase of the direction; x Overlay fringe pattern in direction I x Expressed as ; Superposition truncated phase for: ； y Overlay fringe pattern in direction I y Expressed as ; Superposition truncated phase for: ； in, To superimpose fringe patterns x Directional background light intensity; Indicates the display coordinates corresponding to the front surface of the transparent element S 1 point x phase information of the direction, and ; Indicates the display coordinates corresponding to the back surface of the transparent element S 2 point x phase information of the direction, and ; M x1 and M x2 are the modulation degrees of the superimposed fringe patterns, respectively; f is the frequency of the superimposed fringe pattern; x S1 and x S2 They are S 1 and S 2 point x Display coordinates of the direction; To superimpose fringe patterns y Directional background light intensity; Indicates the display coordinates corresponding to the front surface of the transparent element S 1 point y phase information of the direction, and ; Indicates the display coordinates corresponding to the back surface of the transparent element S 2 point y phase information of the direction, and ; M y1 and M y2 are the modulation degrees of the superimposed fringe patterns, respectively; f is the frequency of the superimposed fringe pattern; y S1 and y S2 They are S 1 and S 2 point y The display coordinates of the direction.

4. The transparent element surface shape detection method according to claim 3, characterized in that: In step (3), x Direction and y The overlay display coordinates of the directions are: ， ； After normalization, the corresponding ， ； in, W is the width of the display; ( x s , y s ) is the overlay display coordinate, ( x sn , y sn ) are the normalized overlay display coordinates.

5. The transparent element surface shape detection method according to claim 1 or 4, characterized in that: In the step (4), the overlay display coordinates are used as the center position of the relevant matching.

6. The transparent element surface shape detection method according to claim 1, characterized in that: In step (4), the sampling frequency is F , length is N The time intensity series: ； and , , , , ; in,( x s1 , y s1 )and( x s2 , y s2 ) are S 1 Point and S 2 The display coordinates of the point, ( x sn1 , y sn1 )and( x sn2 , y sn2 ) are S 1 Point and S 2 The point's normalized display coordinates, τ is the number of cycles of the superimposed fringe pattern.

7. The transparent element surface shape detection method according to claim 6, characterized in that: In step (4), the time series I x The spectrum interval Fs / N is the radius of the correlation matching.

8. The transparent element surface shape detection method according to claim 6, characterized in that: In step (4), the matching interval is determined according to the center position of the relevant matching and the radius of the relevant matching: ， ； Combine all values ​​in the two matching intervals to form corresponding frequency pairs ( ), determine the time intensity sequence corresponding to the frequency pair I cal ; and extract the time intensity sequence corresponding to the gray value of the superimposed fringe image I get ; Correlate the two time intensity series: ； in, and The time intensity series I cal and I get The mean of I cal，i and I get，i They are I cal and I get No. i Strength value; Pearson ( I cal ,I get ) is the correlation coefficient; A correlation coefficient matrix is ​​constructed using the correlation coefficients corresponding to each frequency pair, and the frequency pair corresponding to the maximum value in the correlation coefficient matrix is ​​extracted as the initial value of the correlation matching.

9. The transparent element surface shape detection method according to claim 8, characterized in that: In step (4), the nonlinear optimization includes the following sub-steps: (41) Extract the frequency pair corresponding to the maximum value in the correlation coefficient matrix ( ), based on the non-negative least squares method, the corresponding superimposed fringe pattern is calculated x Directional background light intensity , modulation degree of superimposed fringe pattern M x1 and M x2 After that, build I get =WF T ;in: ； (42) Constructing time-intensity series I’ : ； calculate I’ and I get Strength difference between diff : ； Based on Nelder-Mead Algorithm optimization and To strength difference diff When it is less than the threshold ε, the output x Display coordinates corresponding to the direction and ; (43) Repeat steps (41) and (42) to obtain the corresponding calculation y The display coordinates corresponding to the direction and ; (44) According to the display coordinates corresponding to the front surface of the transparent element ( ) are calculated separately x Direction and y The slopes corresponding to the directions are: ； in,( x c , y c )and( x m1 , y m1 ) are the coordinates of the entrance pupil center point C of the camera and the front surface of the transparent element M 1 Point coordinates; z m2c and z m2s The distance between the front surface of the transparent element and the center of the camera entrance pupil and the display are z Distance of direction; d m2c and d m2s are the distances from the front surface of the transparent element to the center of the camera entrance pupil and the display, respectively; (45) Adoption Zernike The slope of the polynomial fitting is used to reconstruct the front surface shape of the transparent element.

Images