Method for measuring object topography based on single-pixel imaging binocular deflection

Abstract

The application discloses a kind of object topography measurement methods based on single-pixel imaging binocular deflection, comprising: step S1: based on Fourier single-pixel imaging principle and Fourier center slice generates multiple horizontal and vertical Fourier fundamental frequency fringe images;Step S2: the multiple horizontal and vertical Fourier fundamental frequency fringe patterns generated are displayed through display screen, and the Fourier fundamental frequency fringe pattern that is distorted by being reflected on the surface of the object to be measured is collected through camera;Step S3: based on single-pixel imaging principle, the image collected by camera is solved;Step S4: combination is the corresponding point coordinate of display screen;Step S5: based on binocular deflection, the display screen coordinates obtained are used to iteratively reconstruct the three-dimensional point cloud and surface normal on the upper surface of the object to be measured under the framework of binocular deflection, and the three-dimensional topography of the object to be measured is obtained by wavefront reconstruction algorithm.The application has the advantages of simple principle, easy operation, strong anti-interference ability, wide application range, etc.

Description

Technical Field

[0001] The present invention mainly relates to the technical field of object shape measurement, and in particular to an object shape measurement method based on single-pixel imaging binocular deflectometry. Background Art

[0002] In the field of industrial production, there is an increasing demand for the topography measurement of mirror-reflective objects such as wafers, glass, and lenses, or smooth transparent objects with a certain thickness (thick transparent surfaces).

[0003] Traditionally, some practitioners have used the streak reflectance method, which collects light reflected from an object's surface and reconstructs the surface topography based on the camera-display geometry. This method offers advantages such as high precision and flexibility, making it a common method for measuring the topography of mirrored objects. However, due to the presence of light reflected from the bottom surface of thick, transparent surfaces, this superposition of light reflected from the top surface creates "ghost images," presenting significant challenges to subsequent reconstruction and directly impacting accuracy.

[0004] Unlike diffusely reflective materials like plaster and wood, smooth surfaces like mirrors and thick, transparent surfaces are primarily specularly reflective, meaning the reflected light has a distinct directionality. If a projector is used as a modulated light source, the incident light will also carry directional information, resulting in only a portion of the light being captured by the camera after reflection from the object, making subsequent object reconstruction difficult. However, the light emitted by diffusely reflective light sources like LCD displays does not have directional properties, allowing the camera to capture the modulated image as completely as possible for subsequent reconstruction. The method of using LCD displays as a system modulated light source is collectively known as deflectometry imaging, which is suitable for measuring the topography of smooth surfaces and has been widely used.

[0005] In deflectometry imaging, practitioners Wan et al. proposed a frequency-shifting method based on fringe frequency optimization and Fourier transform. This method separates the lower surface reflection component from a frequency domain perspective, enabling the measurement of thick transparent surfaces. However, it places high demands on image acquisition and is susceptible to interference from DC components and ambient light. Another practitioner, Ye et al., proposed a phase decoupling method that directly calculates the upper surface phase by projecting frequency-shifted images, enabling lens morphology measurement. However, this method requires high-quality iterative initial values ​​and is also sensitive to illumination. Both of these methods are based on monocular deflectometry. To resolve depth ambiguity, monocular deflectometry requires obtaining the spatial position of the system, including the object being measured. Therefore, it places high demands on equipment installation accuracy, which is difficult to guarantee in actual industrial environments.

[0006] In addition, the difficulty in solving the "ghost image" problem lies in the fact that traditional cameras only collect light intensity information. Due to the lack of light direction dimension, it is difficult to directly separate the reflected light intensity of the upper and lower surfaces from the images collected by the camera. Summary of the Invention

[0007] The technical problem to be solved by the present application is that: in view of the technical problems existing in the prior art, the present application provides an object topography measurement method based on single-pixel imaging binocular deflection, which is simple in principle, easy to operate, strong in anti-interference ability and wide in application range.

[0008] To solve the above technical problems, the present application adopts the following technical solutions:

[0009] An object topography measurement method based on single-pixel imaging binocular deflection, comprising:

[0010] Step S1: generating a plurality of horizontal and vertical Fourier fundamental frequency fringe images based on the Fourier single-pixel imaging principle and Fourier center slicing;

[0011] Step S2: displaying the generated horizontal and vertical Fourier fundamental frequency fringe patterns through a display screen, and collecting the Fourier fundamental frequency fringe patterns distorted by the surface of the object to be measured through a camera;

[0012] Step S3: based on the single-pixel imaging principle, solving the images collected by the camera to obtain the LTC horizontal projection and vertical projection curves corresponding to the pixel points on the imaging planes of the two cameras, respectively;

[0013] Step S4: using a Gaussian function to fit the peak point coordinates of the LTC horizontal projection and vertical projection curves, respectively, and combining them into the display screen corresponding point coordinates;

[0014] Step S5: based on binocular deflection, through the obtained display screen coordinates, iteratively reconstructing the three-dimensional point cloud and surface normal of the upper surface of the object to be measured under the binocular deflection framework, and obtaining the three-dimensional topography of the surface of the object to be measured through a wavefront reconstruction algorithm.

[0015] As a further improvement of the method of the present application: in step S1, the specific generation formula of the Fourier fundamental frequency fringe image is as follows:

[0016]

[0017] wherein, represents the horizontal Fourier fundamental frequency fringe image, represents the vertical Fourier fundamental frequency fringe image, A is the average light intensity, B is the modulation light intensity, f x =x / N s represents the horizontal direction frequency, wherein x is the horizontal direction pixel coordinate of the display screen, N s is the horizontal direction resolution of the display screen, f y =y / M s represents the vertical direction frequency, wherein y is the vertical direction pixel coordinate of the display screen, M s is the vertical direction resolution of the display screen.

[0018] As a further improvement of the method of the present invention, each pixel unit in the camera imaging plane is regarded as an independent individual, and the light intensity value collected in the time series of each pixel unit is extracted.

[0019] As a further improvement of the method of the present invention: the light intensity signal collected by the camera is defined as:

[0020] I i (u,v)=I0+∫∫ Ω P(x,y,u,v)·S i (x,y)dxdy

[0021] Where (u, v) represents the coordinates of a point on the camera imaging plane; (x, y) represents the coordinates of a point on the display screen; I0 represents ambient light; Ω represents the pattern area reflected by the object being measured; P(x, y, u, v) represents the horizontal or vertical projection of the light transmission coefficient LTC of the display screen to a point on the camera imaging plane; S i (x,y) represents the horizontal or vertical Fourier fundamental frequency fringe pattern.

[0022] As a further improvement of the method of the present invention: the pixel coordinate (u, v) of the camera corresponds to a frequency coefficient f x The light intensity values ​​obtained by collecting the four horizontal fundamental frequency fringe patterns are I0(u,v,f x ), I1(u,v,f x ), I2(u,v,f x ), I3(u,v,f x ), the four light intensity values ​​obtained are processed as follows:

[0023]

[0024] Where F{·} represents the positive Fourier transform, P v (x, y, u, v) represents the horizontal projection of the light transmission coefficient (LTC) of the display screen to a point on the camera imaging plane; we get:

[0025] P v (x,y,u,v)∝F -1 {[I0(u,v,f x )-I2(u,v,f x )]+j[I1(u,v,f x )-I3(u,v,f x )]}

[0026] Among them, F -1 {·} represents the inverse Fourier transform; similarly, for the frequency coefficient f x The four vertical fundamental frequency stripe patterns generated are collected to obtain the corresponding vertical projection P of LTCh (x,y,u,v).

[0027] As a further improvement of the method of the present invention: in step S4, the P obtained in step S3 is v (x,y,u,v) and P h Perform Gaussian fitting on the highest peak in (x, y, u, v), and obtain the horizontal coordinates of the peak point after fitting, which can be combined into the coordinates of the corresponding point on the display screen.

[0028] As a further improvement of the method of the present invention, the obtained LTC curve is the superposition of multiple pulse curves. This process is equivalent to the following formula:

[0029]

[0030] Among them, P 1 (x, y, u, v) represents the LTC of the first reflection, that is, the upper surface reflection, P i (x, y, u, v) represents the LTC of the i-th reflection. When light passes through a transparent object, it will be reflected and refracted, and the number of reflections and refractions is the same; λ i To express the attenuation coefficient of reflected light.

[0031] As a further improvement of the method of the present invention: in step S5, all pixel points to be reconstructed on the camera imaging plane are iteratively obtained to obtain a point cloud image of the target, and during the reconstruction process, the surface normal of the object to be measured is calculated as an indirect quantity, and the surface morphology map of the object to be measured is reconstructed according to the Zernike polynomial.

[0032] Compared with the prior art, the advantages of the present invention are:

[0033] 1. The object shape measurement method based on single-pixel imaging binocular deflectometry of the present invention has a simple principle and is easy to operate. It inherits the advantages of single-pixel imaging and has a strong ability to resist ambient light interference. Since there is no need to obtain the prior position of the object to be measured, it can better adapt to the industrial production environment and meet actual measurement requirements.

[0034] 2. The object shape measurement method based on single-pixel imaging binocular deflectometry of the present invention regards each pixel unit in the camera imaging plane as an independent unit, uses binocular deflectometry as a framework, and utilizes the principle of single-pixel imaging to directly reconstruct the one-dimensional light transmission curve of the light source plane, obtain the coordinates of the light source acting on the upper surface of the object, separate the surface reflection component from the source of the "ghost image", and solve the "ghost image" problem caused by reflection from the lower surface of the object.

[0035] 3. The object topography measurement method based on single-pixel imaging binocular deflection of the present application is a method for calculating display screen coordinates based on single-pixel imaging, which solves the ghost image problem from the source and inherits the advantages of single-pixel imaging, and has strong anti-interference ability. The present application fuses Fourier single-pixel imaging with Fourier center slicing, and only obtains display screen coordinates from one-dimensional information, thereby improving the collection efficiency.

[0036] 4. The object topography measurement method based on single-pixel imaging binocular deflection of the present application combines Fourier single-pixel imaging under the framework of binocular deflection, and uses an LCD display screen to project horizontal and vertical Fourier fundamental frequency fringe images. The one-dimensional light transmission curve of the reflected light on the surface of a transparent or mirror object is directly extracted by obtaining the light transmission coefficient of each pixel on the camera, and then the surface topography of the transparent or mirror object is reconstructed based on the principle of binocular deflection. The method proposed in the present application inherits the advantages of single-pixel imaging and has strong anti-interference ability; based on binocular deflection, the object depth and topography are decoupled, and there is no need to provide the shape prior of the measured object, which is more suitable for industrial scene application. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 is a flowchart of the method of the present application.

[0038] Figure 2 is a schematic diagram of the measurement system built in the specific application example of the present application.

[0039] Figure 3 is a schematic diagram of the single-pixel imaging system built in the specific application example of the present application.

[0040] Figure 4 is a discrete light spot pattern in the specific application example of the present application.

[0041] Figure 5 is a schematic diagram of a partial Fourier fundamental frequency fringe pattern in the specific application example of the present application.

[0042] Figure 6 is a schematic diagram of the light transmission coefficient curve in the specific application example of the present application.

[0043] Figure 7 is a schematic diagram of the binocular deflection reconstruction in the specific application example of the present application.

[0044] Figure 8 is a schematic diagram of the measurement result in the specific application example of the present application; wherein (a) is the measurement result of a plane mirror; (b) is the measurement result of a plano-concave mirror. DETAILED DESCRIPTION

[0045] The present application will be further described in detail below in combination with the drawings and specific embodiments of the present application.

[0046] like Figure 1 and Figure 2 As shown, the object shape measurement method based on single-pixel imaging binocular deflectometry of the present invention comprises the following steps:

[0047] Step S1: Based on the Fourier single pixel imaging principle and the Fourier center slice, multiple horizontal and vertical Fourier fundamental frequency fringe images are generated; the total number of images under full sampling conditions is (M s ×N s )·2·r,M s and N s They are the horizontal and vertical resolutions of the projector respectively.

[0048] Step S2: The computer sends a control command to the LCD display to display the generated multiple horizontal and vertical Fourier fundamental frequency fringe patterns, and uses the camera to capture the Fourier fundamental frequency fringe patterns that are distorted by reflection from the surface of the object to be measured, and then transmits the image to the computer.

[0049] Step S3: Based on the single-pixel imaging principle, the computer solves the image captured by the camera and obtains the LTC horizontal and vertical projection curves corresponding to the pixel points on the two camera imaging planes, where is the main camera pixel, is the pixel point of the secondary camera, and the horizontal and vertical resolutions of the camera are M c and N c .

[0050] Step S4: Use Gaussian function to fit the peak point coordinates of the LTC horizontal projection and vertical projection curves respectively, and combine them into the corresponding point coordinates of the display screen. That is, for the P obtained in step S3, v (x,y,u,v) and P h Perform Gaussian fitting on the highest peak in (x, y, u, v), and obtain the horizontal coordinates of the peak point after fitting, which can be combined into the coordinates of the corresponding point on the display screen.

[0051] Step S5: Based on binocular deflectometry, the three-dimensional point cloud and surface normal of the upper surface of the object to be measured are iteratively reconstructed using the obtained display screen coordinates within the binocular deflectometry framework, and then the three-dimensional morphology of the surface of the object to be measured is obtained through a wavefront reconstruction algorithm.

[0052] The present invention regards each pixel unit in the camera imaging plane as an independent individual and uses binocular deflectometry as a framework to complete the topography measurement of transparent and thick transparent surfaces.

[0053] In a specific application example, in step S1, the specific generation formula of the Fourier baseband fringe image is as follows:

[0054]

[0055] in, represents the horizontal Fourier fundamental frequency fringe image, Represents the vertical Fourier fundamental frequency fringe image, A is the average light intensity, B is the modulated light intensity, f x =x / N s Represents the horizontal frequency, where x is the horizontal pixel coordinate of the display, N s is the horizontal resolution of the display, f y =y / M s Represents the vertical frequency, where y is the vertical pixel coordinate of the display, M s It is the vertical resolution of the display. It should be noted that combining Fourier single-pixel imaging with Fourier center slices only projects horizontal and vertical Fourier fundamental frequency fringe images, greatly reducing the number of projected images.

[0056] In a specific application example, in step S3, in one embodiment, each pixel unit in the camera imaging plane is regarded as an independent individual, and the light intensity value collected in the time series of each pixel unit is extracted. Assume that a certain pixel coordinate (u, v) of the camera has a certain frequency coefficient f x The light intensity values ​​obtained by collecting the four horizontal fundamental frequency fringe patterns are I0(u,v,f x ), I1(u,v,f x ), I2(u,v,f x ), I3(u,v,f x ), the four light intensity values ​​obtained are processed as follows:

[0057]

[0058] Where F{·} represents the positive Fourier transform, P v (x, y, u, v) represents the horizontal projection of the light transmission coefficient (LTC) of the display screen to a certain point on the camera imaging plane. Therefore, we can get:

[0059] P v (x,y,u,v)∝F -1 {[I0(u,v,f x )-I2(u,v,f x )]+j[I1(u,v,f x )-I3(u,v,f x )]}

[0060] Among them, F -1 {·} represents the inverse Fourier transform. Similarly, for the frequency coefficient f x The four vertical fundamental frequency fringe patterns generated are collected, and the vertical projection P of the corresponding LTC can be obtained using the same method as above. h(x,y,u,v).

[0061] refer to Figure 2 In order to implement the above method, the present invention has built a measurement system, including a display screen 1, an object to be measured 2, a secondary camera 3, a main camera 4, an optical support 5, a lens clamp 6 and an optical breadboard 7. The above display screen 1 is used to receive the Fourier fundamental frequency pattern transmitted by the computer and display the pattern in full screen. The above object to be measured 2 is used to reflect the pattern displayed on the display screen to the camera. The above secondary camera 3 and the main camera 4 are used to receive the distorted Fourier fundamental frequency pattern reflected by the object to be measured. Based on the principle of single-pixel imaging, the present invention directly reconstructs the one-dimensional light transmission coefficient curve of the light source plane, obtains the coordinates of the light source acting on the upper surface of the object, and separates the surface reflection component from the source of the "ghost image". The present invention can adapt to different detection environments, reduce the equipment installation accuracy requirements, and can adapt to smooth objects of different materials such as mirrors and thick transparent surfaces.

[0062] like Figure 3 As shown, the single-pixel imaging system includes a single-pixel sensor 8 (e.g., a photodiode) and a light source 9 (e.g., a projector, spatial modulator, display screen, etc.), both directed toward the object 2 to be measured. During the measurement process, the modulation pattern (e.g., random speckle pattern, Fourier image, Hadamard image, etc.) transmitted by the projector interacts with the object image. The single-pixel sensor 8, which has no spatial resolution, collects the resulting one-dimensional light intensity signal. After multiple projections, the object image can be reconstructed based on the correlation between the modulation pattern sequence and the corresponding light intensity signal sequence.

[0063] In the present invention, each pixel unit in the camera imaging plane is regarded as a single pixel sensor 8, and the entire imaging plane can constitute a group of single pixel sensor arrays.

[0064] During the image acquisition process, for a pixel point at a certain position on the imaging plane, the light intensity value obtained can be expressed as:

[0065] I i (u,v)=I0+∫∫ Ω P(x,y,u,v)·S i (x,y)dxdy

[0066] Where (u, v) is the coordinate of a point on the camera imaging plane; (x, y) is the coordinate of a point on the display screen; I0 represents the ambient light; Ω is the pattern area reflected by the object being measured; P(x, y, u, v) represents the horizontal or vertical projection of the light transmission coefficient (LTC) of the display screen to a point on the camera imaging plane; S i (x, y) represents the horizontal or vertical Fourier baseband fringe pattern. It should be noted that, in this embodiment, if S i(x, y) and P(x, y, u, v) do not specify horizontal or vertical, which means they are applicable to both directions. For traditional single-pixel imaging, P(x, y, u, v) appears as multiple discrete light spots (such as Figure 4 shown).

[0067] However, single-pixel imaging often requires thousands of illuminations. In particular, in deflectometry, treating the camera imaging plane as a single-pixel sensor array requires single-pixel reconstruction for each pixel, posing significant challenges for signal acquisition and data storage and readout. To address this, the present invention incorporates Fourier center slicing technology to project only the horizontal and vertical Fourier fundamental frequency fringe images.

[0068] The specific generation formula of the image is as follows:

[0069]

[0070] in, represents the horizontal Fourier fundamental frequency fringe image, Represents the vertical Fourier fundamental frequency fringe image, A is the average light intensity, B is the modulated light intensity, f x =x / N s Represents the horizontal frequency, where x is the horizontal pixel coordinate of the display, N s is the horizontal resolution of the display, f y =y / M s Represents the vertical frequency, where y is the vertical pixel coordinate of the display, M s It is the vertical resolution of the display. It should be noted that combining Fourier single-pixel imaging with Fourier center slices, only projecting horizontal and vertical Fourier fundamental frequency fringe images, greatly reduces the number of projected images. Figure 5 A portion of the Fourier fundamental frequency fringe pattern is shown.

[0071] In a specific application embodiment, each pixel unit in the camera imaging plane is regarded as an independent individual, and the light intensity value collected in the time series of each pixel unit is extracted. Assume that a certain pixel coordinate (u, v) of the camera has a certain frequency coefficient f x The light intensity values ​​obtained by collecting the four horizontal fundamental frequency fringe patterns are I0(u,v,f x ), I1(u,v,f x ), I2(u,v,f x ), I3(u,v,f x ). The four light intensity values ​​obtained are processed as follows:

[0072]

[0073] Where F{·} represents the positive Fourier transform, Pv (x, y, u, v) represents the horizontal projection of the light transmission coefficient function (LTC) of the display screen to a certain point on the camera imaging plane. Therefore, we can get:

[0074] P v (x,y,u,v)=F -1 {[I0(u,v,f x )-I2(u,v,f x )]+j[I1(u,v,f x )-I3(u,v,f x )]}

[0075] Among them, F -1 {·} represents the inverse Fourier transform. From the analytical formula, we can see that by using a display to display the four-step phase-shifted Fourier images of different spatial frequencies, we can obtain the Fourier coefficients of LTC corresponding to different spatial frequencies. In addition, the DC component is eliminated during the calculation process, which makes the algorithm have a certain anti-interference ability. Similarly, for the frequency coefficient f x The four vertical fundamental frequency fringe patterns generated are collected, and the vertical projection P of the corresponding light transmission coefficient function (LTC) can be obtained using the same method as above. h (x, y, u, v). To this end, the present invention combines Fourier single-pixel imaging with Fourier center slice technology, which greatly reduces the number of required projections.

[0076] In this embodiment, if the object to be measured is a thick and transparent surface, the light will be reflected multiple times when passing through the surface, such as Figure 6 As shown, the LTC curve obtained by the present invention is actually the superposition of multiple pulse curves, and this process is equivalent to the following formula:

[0077]

[0078] Among them, P 1 (x, y, u, v) represents the LTC of the first reflection, that is, the upper surface reflection, P i (x, y, u, v) represents the LTC of the i-th reflection. When light passes through a transparent object, it will be reflected and refracted, and the number of reflections and refractions is the same. In addition, as the number of refractions increases, the intensity of the reflected light will gradually decay. The present invention uses λ i To express the attenuation coefficient of reflected light. Assuming that the curvature of the object surface does not change much, the intensity of light reflected from different positions on the transparent object surface can be considered the same. After the light is reflected by the subsurface, the intensity of the reflected light is less than the intensity of the surface reflected light due to the increase in the number of refractions. The curve is represented by pulse curves of different heights, among which the curve with the highest peak corresponds to the surface reflection. Therefore, the horizontal coordinates of the peak point of the highest pulse curve are the coordinates of the corresponding point to be obtained. For the P obtained in step S3, v(x,y,u,v) and P h Perform Gaussian fitting on the highest peak in (x, y, u, v), and obtain the horizontal coordinates of the peak point after fitting, which can be combined into the coordinates of the corresponding point on the display screen.

[0079] Next, the present invention performs three-dimensional reconstruction of the surface of the object to be measured based on the principle of binocular deflectometry. Figure 7 As shown, assume that the light emitted from the display screen S1 reaches C1 of camera A after being reflected by an object. Let the incident light at point C1 be i1. This ray passes through the optical center O1 and C1 of camera A, and the mirror point to be determined is located on this ray. In the previous step, the coordinates of point S1 on the display screen corresponding to point C1 have been obtained. Then, the position of the target point is searched through an iterative algorithm. The known conditions are point C1, point S1, ray i1, and the target point is located on ray i1. Assuming that the target point is at P1, the internal and external parameters between the cameras can be used to obtain the mapping point C2 from P1 to camera 2. Similarly, the coordinates of point C2 on the display screen S2 are also known. According to the vector and Normals can be calculated separately and if and If they are equal, then point P1 is the real target point. However, in the actual process, due to the existence of noise and error, and It is difficult to be completely equal, by setting the threshold τ, when The target point position is considered to have been obtained when . The above iteration is repeated for all pixels to be reconstructed on the camera imaging plane to obtain the target point cloud image. In addition, during the reconstruction process, the surface normal of the object to be measured is calculated as an indirect quantity, and the surface topography of the object to be measured can also be reconstructed based on the Zernike polynomial. Figure 8 These are the measurement results of the morphology of flat lenses and concave lenses, where (a) is a flat lens and (b) is a flat concave lens.

[0080] Traditional deflectometry measurement systems use diffuse light sources, which solves the imaging problem but also introduces depth ambiguity. The two variables, object depth and topography, are coupled to determine the reflected fringe pattern captured by the camera. Monocular deflectometry uses object depth as a priori, eliminating this coupling and completing object topography measurement. However, in some scenarios where it is not possible to obtain object depth priors, this method is no longer applicable. The rationale behind the present invention's use of binocular deflectometry lies in its ability to decouple object depth and topography, based on binocular analysis of reflected light, without requiring a priori knowledge of object depth, making it more suitable for practical use cases.

[0081] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions based on the principles of the present invention are within the scope of protection of the present invention. It should be noted that for those skilled in the art, various improvements and modifications that do not depart from the principles of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A method for measuring object shape based on single-pixel imaging binocular deflectometry, characterized in that: include: Step S1: generating multiple horizontal and vertical Fourier fundamental frequency fringe images based on the Fourier single pixel imaging principle and the Fourier center slice; Step S2: Displaying the generated multiple horizontal and vertical Fourier fundamental frequency fringe patterns on a display screen, and capturing the Fourier fundamental frequency fringe patterns distorted by reflection from the surface of the object to be measured by a camera; Step S3: Based on the single-pixel imaging principle, the image captured by the camera is solved to obtain the horizontal and vertical projection curves of the light transmission coefficient LTC corresponding to the pixel points on the two camera imaging planes respectively; Step S4: Use Gaussian functions to fit the peak point coordinates of the LTC horizontal projection and vertical projection curves respectively, and combine them into the corresponding point coordinates of the display screen; Step S5: Based on binocular deflectometry, the three-dimensional point cloud and surface normal of the upper surface of the object to be measured are iteratively reconstructed using the obtained display screen coordinates within the binocular deflectometry framework, and the three-dimensional morphology of the surface of the object to be measured is obtained through a wavefront reconstruction algorithm.

2. The object shape measurement method based on single-pixel imaging binocular deflectometry according to claim 1, characterized in that: In step S1, the specific generation formula of the Fourier baseband fringe image is as follows: in, represents the horizontal Fourier fundamental frequency fringe image, represents the vertical Fourier fundamental frequency fringe image, is the average light intensity, is the modulated light intensity, represents the horizontal frequency, where is the horizontal pixel coordinate of the display screen, is the horizontal resolution of the display. represents the vertical frequency, where is the pixel coordinate in the vertical direction of the display screen, It is the vertical resolution of the display.

3. The object shape measurement method based on single-pixel imaging binocular deflectometry according to claim 1, characterized in that: Each pixel unit in the camera imaging plane is regarded as an independent individual, and the light intensity value collected in the time series of each pixel unit is extracted.

4. The object shape measurement method based on single-pixel imaging binocular deflectometry according to claim 3, characterized in that: The light intensity signal collected by the camera is defined as: in, Indicates the coordinates of a point on the camera imaging plane; Indicates the coordinates of a point on the display screen; Represented as ambient light; Indicates the pattern area reflected by the object being measured; Indicates the horizontal or vertical projection of the light transmission coefficient LTC of the display screen to a certain point on the camera imaging plane; Represents a horizontal or vertical Fourier fundamental fringe pattern.

5. The object shape measurement method based on single-pixel imaging binocular deflectometry according to claim 4, characterized in that: A pixel coordinate of the camera For a certain frequency coefficient The light intensity values ​​obtained by collecting the four horizontal fundamental frequency fringe patterns are , , , , the four light intensity values ​​obtained are processed as follows: in, represents the forward Fourier transform, Represents the LTC horizontal projection of the display screen to a point on the camera imaging plane; we get: in, Represents the inverse Fourier transform; similarly, for the frequency coefficient The four vertical fundamental frequency fringe patterns generated are collected to obtain the corresponding vertical projection of LTC .

6. The object shape measurement method based on single-pixel imaging binocular deflectometry according to claim 5, characterized in that: In step S4, the and Perform Gaussian fitting on the highest peak in the wave, and obtain the horizontal coordinates of the peak point after fitting, which can be combined into the coordinates of the corresponding point on the display screen.

7. The object shape measurement method based on single-pixel imaging binocular deflectometry according to claim 6, characterized in that: The obtained LTC curve is the superposition of multiple pulse curves. This process is equivalent to the following formula: in, represents the LTC of the first reflection, i.e. the upper surface reflection, represents the LTC of the i-th reflection. When light passes through a transparent object, it will be reflected and refracted, and the number of reflections and refractions is the same; To express the attenuation coefficient of reflected light.

8. The object shape measurement method based on single-pixel imaging binocular deflectometry according to any one of claims 1 to 5, characterized in that: In step S5, all pixel points to be reconstructed on the camera imaging plane are iterated to obtain a point cloud image of the target, and during the reconstruction process, the surface normal of the object to be measured is calculated as an indirect quantity, and the surface topography of the object to be measured is reconstructed according to the Zernike polynomial.

Images