Three-dimensional shape measurement method and system based on continuous vertical scanning structure illumination

Abstract

The invention belongs to the related technical field of three-dimensional shape measurement, and discloses a three-dimensional shape measurement method and system based on continuous vertical scanning structure illumination, and the method comprises the steps: S1, collecting a structured light image of a to-be-measured sample at each scanning position; s2, carrying out background light intensity estimation by using the adjacent structured light images in the time domain, obtaining a uniform illumination image, and correcting the structured light image; s3, performing high-pass filtering on the uniform illumination image to obtain high-frequency in-focus information; performing low-pass filtering on the corrected structured light image to obtain low-frequency in-focus information; fusing the two to obtain an optical tomographic image of the to-be-detected sample; and S4, obtaining the three-dimensional shape of the to-be-detected sample. According to the method, a continuous scanning structured light illumination measurement method is adopted, only one structured illumination phase shift cosine fringe pattern is collected at each scanning position, and high-precision chromatography extraction is carried out, so that high-precision chromatography imaging under the condition of an uninterrupted continuous scanning mode can be realized, and the measurement speed and precision are greatly improved.

Description

Technical Field

[0001] The present invention belongs to the technical field related to three-dimensional shape measurement, and more specifically, relates to a three-dimensional shape measurement method and system based on continuous vertical scanning structured illumination. Background Art

[0002] The surface morphology of microstructures is an important geometric feature of manufactured products and scientific research samples. Its measurement, analysis and evaluation are of great significance to the assurance of surface functional quality characteristics. Structured light illumination microscopy (SIM) is based on the principle of wide-field microscopy imaging. It uses structured light to modulate the measured surface or structural information, removes the interference of defocused signals, and combines vertical scanning to obtain optical tomography images. Then, the peak location algorithm is used to extract the peak of the tomography response curve to achieve three-dimensional measurement. This method has the characteristics of non-contact, high efficiency and high precision. It can be applied not only to reflective surfaces, but also to the morphology measurement of scattering surfaces.

[0003] The existing structured light illumination microscopy measurement method needs to stop after fixing the step position during the scanning process, project multiple sets of cosine fringe patterns onto the surface of the sample to be measured through the spatial light modulator (DMD) in the measurement system, and then continue scanning until the scanning is completed. This process is time-consuming and cannot meet the needs of fast measurement. et al. proposed the flying phase method in the paper “Better three-dimensional inspection with structured illumination: speed”. At a fixed scanning step position, the camera only needs a single exposure. By calculating the axial tomographic response through the “flying phase method”, fast three-dimensional reconstruction can be achieved, greatly shortening the measurement time. However, due to the limited number of images collected, the stability of the calculated tomographic response signal decreases. In addition, factors such as the local slope of the sample surface, material properties, and ambient light fluctuations will significantly interfere with the collected signal, resulting in serious noise in the tomographic response. Under complex noise interference, the maximum tomographic response position in the tomographic response cannot be accurately determined (the maximum tomographic response position can determine the sample surface height), resulting in reduced measurement accuracy. Summary of the invention

[0004] In view of the above limitations of the prior art, the present invention proposes a three-dimensional morphology measurement method and system based on continuous vertical scanning structured illumination, which utilizes the "flying phase method" to collect multiple phase-shifted fringe patterns at each step position without stopping, thereby advantageously improving the measurement speed. At the same time, a tomography extraction algorithm that effectively suppresses the influence of noise is adopted to ensure measurement accuracy, thereby realizing rapid and high-precision measurement of three-dimensional morphology of continuous vertical scanning structured illumination.

[0005] To achieve the above object, according to one aspect of the present invention, a high-precision three-dimensional shape measurement method based on continuous vertical scanning structured illumination is provided, comprising:

[0006] S1, using the continuous vertical scanning structured light illumination measurement method, collects a phase-shifted cosine fringe structured light image of the sample to be tested at each vertical scanning position;

[0007] S2, using the adjacent structured light images in the time domain to estimate the background light intensity, and obtaining the estimated background light intensity as a uniformly illuminated image; at the same time, using the estimated background light intensity to correct the background light intensity of the structured light image, and obtaining the corrected structured light image;

[0008] S3, high-pass filtering the uniformly illuminated image to obtain high-frequency in-focus information of the sample to be tested; low-pass filtering the corrected structured light image to obtain low-frequency in-focus information of the sample to be tested; fusing the high-frequency in-focus information and the low-frequency in-focus information to obtain an optical tomography image of the sample to be tested;

[0009] S4, based on the optical tomography image, obtaining the three-dimensional morphology of the sample to be tested.

[0010] According to the three-dimensional shape measurement method based on continuous vertical scanning structured illumination provided by the present invention, S1 specifically includes:

[0011] A plurality of cosine fringe patterns with fixed phase shifts are preset, and one of the cosine fringe patterns is projected onto the surface of the sample to be tested at each vertical scanning position. During the scanning process of the sample to be tested, the plurality of cosine fringe patterns are projected in a cycle in sequence.

[0012] According to the three-dimensional shape measurement method based on continuous vertical scanning structured illumination provided by the present invention, in S2, the background light intensity is estimated by using the structured light images adjacent in the time domain, and the estimated background light intensity is obtained as a uniform illumination image, which specifically includes:

[0013] For any vertical scanning position, taking a plurality of adjacent structured light images around the vertical scanning position as the center;

[0014] averaging a plurality of adjacent structured light images and filtering the average to obtain an estimated background light intensity at any scanning position as a uniformly illuminated image;

[0015] The number of the plurality of adjacent structured light images is consistent with the number of the preset cosine fringe patterns.

[0016] According to the three-dimensional shape measurement method based on continuous vertical scanning structured illumination provided by the present invention, the uniform illumination image at any vertical scanning position is specifically shown as follows:

[0017]

[0018] Among them, I b Indicates background light intensity; I est represents uniform illumination image; H represents filter; z i represents the i-th scanning position; I represents the structured light image; and N represents the number of the preset cosine fringe patterns.

[0019] According to the three-dimensional shape measurement method based on continuous vertical scanning structured illumination provided by the present invention, in S2, the estimated background light intensity is used to correct the background light intensity of the structured light image to obtain the corrected structured light image, which specifically includes:

[0020] The collected structured light image is subtracted from the uniformly illuminated image to obtain the corrected structured light image.

[0021] According to the three-dimensional shape measurement method based on continuous vertical scanning structured illumination provided by the present invention, high-pass filtering is performed on the uniform illumination image in S3 to obtain high-frequency in-focus information of the sample to be measured, which specifically includes:

[0022] The contrast of the modified structured light image is calculated using a Laplace operator, and contrast-weighted high-pass filtering is performed on the uniformly illuminated image to obtain high-frequency in-focus information of the sample to be tested.

[0023] According to the three-dimensional shape measurement method based on continuous vertical scanning structured illumination provided by the present invention, the corrected structured light image is low-pass filtered to obtain the low-frequency in-focus information of the sample to be measured, which is specifically shown in the following formula:

[0024] I in,Lo (x,y)=LP[I in (x,y)]≈LP[D(x,y)];

[0025] Among them, I in,Lo represents the low-frequency in-focus information in the tomographic response image; LP[.] represents Gaussian low-pass filtering; I in (x, y) represents the in-focus part of the image; D(x, y) represents the corrected structured light image; (x, y) is the pixel coordinate.

[0026] According to the three-dimensional shape measurement method based on continuous vertical scanning structured illumination provided by the present invention, the uniform illumination image is subjected to high-pass filtering to obtain high-frequency in-focus information of the sample to be measured, which is specifically shown in the following formula:

[0027] I in,Hi (x,y)=HP[I b (x,y)]≈HP[I est (x,y)]·Nm (x,y);

[0028] Among them, I in,Hi Represents the high-frequency in-focus information in the tomographic response image; HP[.]=1-LP[.]; I b Indicates background light intensity; I est represents a uniformly illuminated image; (x, y) is the pixel coordinate; N m (x, y) represents the contrast calculated by using the Laplace operator on the corrected structured light image.

[0029] According to the three-dimensional shape measurement method based on continuous vertical scanning structured illumination provided by the present invention, high-frequency in-focus information and low-frequency in-focus information are fused to obtain an optical tomography image of the sample to be measured, which is specifically shown in the following formula:

[0030] m HiLo (x,y)=ηI in,Hi (x,y)+I in,Lo (x,y)

[0031] =ηHP[I est (x,y)]·N m (x,y)+LP[|I(x,y)-I est (x,y)|];

[0032] Among them, m HiLo (x, y) represents the optical tomography image; η represents the adjustment factor.

[0033] According to another aspect of the present invention, a three-dimensional shape measurement system based on continuous vertical scanning structured illumination is provided, the system comprising a memory and a processor, the memory storing a computer program, and the processor executing any one of the above-mentioned three-dimensional shape measurement methods based on continuous vertical scanning structured illumination when executing the computer program.

[0034] In general, compared with the prior art, the above technical solutions conceived by the present invention provide a three-dimensional shape measurement method and system based on continuous vertical scanning structured illumination:

[0035] 1. It is proposed to use adjacent structured light images in the time domain to estimate the background light intensity, which can avoid collecting multiple structured light images at the same scanning position and can accurately estimate the background light intensity while realizing continuous scanning; by using the estimated background light intensity to correct the structured light image, it is helpful to eliminate the camera acquisition intensity fluctuation caused by ambient stray light and imaging noise during the scanning process, and realize accurate correction of the structured light image;

[0036] Taking into account that the background light intensity image contains high-frequency information of the tomographic image, and the corrected structured light image contains low-frequency information of the tomographic image, the idea of ​​fusion is proposed. By low-pass filtering the corrected structured light image, the low-frequency in-focus information is extracted to obtain the low-frequency information in the tomographic image. By high-pass filtering the uniformly illuminated image, the high-frequency noise is suppressed while the high-frequency part of the in-focus information in the image is extracted to obtain the high-frequency information in the tomographic image. The high-frequency information and the low-frequency information are fused to obtain a complete tomographic image with a high signal-to-noise ratio in the entire frequency domain. Therefore, the tomographic image can be calculated through a single structured illumination fringe pattern, and the fast and accurate optical tomographic measurement of the three-dimensional morphology is realized, thereby improving the measurement speed.

[0037] The continuous scanning structured light illumination measurement method can be used to collect only one structured illumination phase-shifted cosine fringe pattern at each scanning position, thus achieving an uninterrupted continuous scanning mode and greatly improving the measurement speed;

[0038] 2. The method of estimating the background light intensity by averaging the adjacent structured light images in the time domain and filtering is helpful to reduce the interference of imaging noise; considering that the defocus conditions of the images collected at each scanning position are not consistent, there are residual stripes on the background light intensity image obtained by the mean calculation, which will affect the quality of background intensity estimation. Therefore, it is proposed to use a filter to filter the axial background intensity response to suppress the influence of residual stripes;

[0039] 3. The background-corrected structured light image eliminates the influence of background light intensity fluctuations. The image retains the in-focus stripe information, but due to the existence of the stripe information, it cannot be used directly. It needs to be low-pass filtered to extract the low-frequency in-focus information. Low-pass filtering can effectively remove noise, and only the low-frequency part of the tomographic image is retained, that is, the low-frequency information in the tomographic image is obtained;

[0040] 4. Considering that the uniformly illuminated image contains in-focus information and out-of-focus information, if the high-frequency part of the in-focus information of the uniformly illuminated image is extracted and fused with the low-frequency information in the tomographic image, a complete tomographic image can be obtained. At the same time, considering that the out-of-focus information in the uniformly illuminated image is generally low-frequency, and the in-focus information is the clear texture details on the sample, which is high-frequency information, therefore, a high-pass filter is used to extract only the high-frequency information of the in-focus content and eliminate the out-of-focus content, that is, to obtain the high-frequency information in the tomographic image. At the same time, the Laplace operator is used to calculate the contrast of the corrected structured light image, and the high-pass filtered image is contrast-weighted, which can further suppress the high-frequency noise in the image and improve the image signal-to-noise ratio. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1It is a schematic diagram of the measurement principle of the three-dimensional shape measurement method based on continuous vertical scanning structure illumination provided by the present invention;

[0042] Figure 2 It is a schematic diagram of the measurement process of the three-dimensional shape measurement method based on continuous vertical scanning structured illumination provided by the present invention. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solutions 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 embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0044] See also Figure 1 and Figure 2 This embodiment provides a three-dimensional shape measurement method based on continuous vertical scanning structure illumination, and the three-dimensional shape measurement method includes:

[0045] S1, using the continuous vertical scanning structured light illumination measurement method, collects a phase-shifted cosine fringe structured light image of the sample to be tested at each vertical scanning position;

[0046] S2, using the adjacent structured light images in the time domain to estimate the background light intensity, and obtaining the estimated background light intensity as a uniformly illuminated image; at the same time, using the estimated background light intensity to correct the background light intensity of the structured light image, and obtaining the corrected structured light image;

[0047] S3, high-pass filtering the uniformly illuminated image to obtain the high-frequency part of the in-focus information of the sample to be tested; low-pass filtering the corrected structured light image to obtain the low-frequency part of the in-focus information of the sample to be tested; fusing the high-frequency in-focus information and the low-frequency in-focus information to obtain an optical tomography image of the sample to be tested;

[0048] S4, based on the optical tomography image, obtaining three-dimensional morphological parameters of the sample to be tested.

[0049] This embodiment calculates a tomographic image through a single structured illumination stripe pattern, i.e., a structured light image, thereby realizing fast optical tomographic measurement and improving measurement speed. The existing method requires the acquisition of multiple phase-shifted patterns at each scanning position to calculate the optical tomographic image, so the scanning needs to be intermittent, i.e., a "stop-go-stop" mode. However, the method of this embodiment only needs to acquire one structured illumination stripe pattern at each scanning position, so it can realize an uninterrupted continuous scanning mode, greatly improving the measurement speed, while suppressing the problem of excessive noise introduced by continuous scanning, and can ensure a high signal-to-noise ratio while improving the measurement speed. In addition, the method of this embodiment does not require the addition of additional projection patterns and hardware equipment.

[0050] In some specific embodiments, a three-dimensional shape measurement method based on continuous vertical scanning structured illumination is a continuous scanning structured illumination optical tomography measurement method based on sliding time window background estimation and fusion, and the specific steps of the method are as follows:

[0051] First, N cosine fringe patterns with a fixed phase shift of 2π / N are generated and input into the DMD for projection during the scanning process, that is, S1 specifically includes: presetting multiple cosine fringe patterns with a fixed phase shift, projecting one cosine fringe pattern onto the surface of the sample to be tested at each scanning position, and during the scanning process of the sample to be tested, the multiple cosine fringe patterns are projected in sequence. That is, multiple cosine fringe patterns are projected one at each scanning position in order to collect structured light images, and after the projection from the first to the last is completed, the projection is cyclically started from the first. Specifically, the fringe pattern I DMD It can be expressed as:

[0052]

[0053] Where I0 represents the basic light intensity, (x, y) represents the image pixel coordinates, ω represents the frequency of the fringe pattern, n represents the nth cosine fringe pattern, and N represents the number of preset phase-shifted cosine fringe patterns. In actual measurement, the collected signal is easily disturbed by the fluctuation of background light intensity. Any point I(z) of the structured light image collected by the camera of the continuous scanning structured illumination microscope can be expressed as:

[0054]

[0055] Where z represents the scanning position, m(z) represents the tomographic response of the measurement system (i.e., the degree of modulation of the cosine fringes), and Δz represents the scanning step. b Indicates the background light intensity interfered by ambient stray light and imaging noise during scanning.

[0056] Then the background light intensity needs to be estimated. According to the expression of the collected structured light image signal, the cosine function of the cosine term is periodic. Therefore, the sum of the points with a fixed phase shift of 2π / N within the cosine function period is 0. By using this point, the mean of the adjacent N images can be calculated to eliminate the cosine term in formula 1.2. At the same time, during the scanning process, the background light intensity I b (z) changes slowly compared to the cosine term, so it can be considered that in the N adjacent images in the time domain, I b (z) can be considered unchanged, so by calculating the mean of N images, we can calculate I b (z).

[0057] The mean of adjacent frame images in the time domain window is calculated to reduce the interference of ambient stray light and imaging noise during the scanning process, and the uniform illumination image of each scanning position can be obtained by sliding calculation. Since the defocus of the collected images at each scanning position is not consistent, there are residual stripes on the uniform illumination image obtained by calculation, which affects the quality of background intensity estimation. Therefore, a band-stop filter is used to filter the axial background intensity response to suppress the influence of residual stripes.

[0058] That is, in S2, the background light intensity is estimated by using the structured light images adjacent in the time domain, and the estimated background light intensity is obtained as a uniform illumination image, which specifically includes:

[0059] For any vertical scanning position, taking a plurality of adjacent structured light images around the vertical scanning position as the center;

[0060] averaging a plurality of adjacent structured light images and filtering the average to obtain an estimated background light intensity at any scanning position as a uniformly illuminated image;

[0061] The number of the plurality of adjacent structured light images is consistent with the number of the preset cosine fringe patterns.

[0062] Specifically, the uniformly illuminated image at any vertical scanning position is shown in the following formula:

[0063]

[0064] Among them, I b Indicates background light intensity; I est represents the estimated background light intensity, i.e., uniformly illuminated image; H represents the filter; z i represents the i-th scanning position; I represents the structured light image; and N represents the number of the preset cosine fringe patterns.

[0065] Furthermore, in S2, the estimated background light intensity is used to correct the background light intensity of the structured light image to obtain the corrected structured light image, which specifically includes:

[0066] The collected structured light image is subtracted from the uniformly illuminated image to obtain the corrected structured light image.

[0067] Furthermore, in S3, high-pass filtering is performed on the uniformly illuminated image to obtain high-frequency in-focus information of the sample to be tested, which specifically includes:

[0068] The modulation contrast of the collected structured light image is calculated, and the uniform illumination image is subjected to modulation contrast weighted high-pass filtering to obtain high-frequency in-focus information of the sample to be tested.

[0069] In the case of low signal-to-noise ratio or complex sample texture, imaging noise and high-frequency artifacts will inevitably be introduced, thereby reducing the accuracy of surface topography measurement. From the perspective of defocus, the wide-field image, i.e., the background light intensity distribution I b and the structured light image I can be expressed as:

[0070] I b (x,y)=I in (x,y)+I out (x,y) (1.4)

[0071] I(x,y)=I in (x,y)·cos(ωx)+I b (x,y) (1.5)

[0072] Among them, I in (x,y) and I out (x, y) are the in-focus and out-of-focus parts of the image, x is the pixel coordinate, and ω represents the frequency of the fringe pattern. in The tomographic response image (i.e., in-focus image) can be obtained, thereby determining the focal position for reconstruction. The background intensity estimation can be determined by formula 1.3, by subtracting the estimated background intensity image from the acquired structured light image to eliminate the background intensity fluctuation caused by ambient stray light and imaging noise during the scanning process, thereby achieving correction and establishing the acquired stripe structured light image D(x, y) corrected by the estimated background intensity as follows:

[0073] D(x,y)=|I(x,y)-I est (x,y)|=I in (x,y)|cos(ωx)|(1.6)

[0074] According to formula 1.6, it can be seen that the corrected structured light image can be approximated as an in-focus image, but due to the existence of fringe information, it cannot be used directly. The fringe information is high-frequency information, so low-pass filtering D(x, y) to eliminate the fringe information can approximately obtain the low-frequency information in the in-focus image, that is, low-pass filtering is performed on the corrected structured light image to obtain the low-frequency in-focus information of the sample to be tested, as shown in the following formula:

[0075] I in,Lo (x,y)=LP[I in (x,y)]≈LP[D(x,y)](1.7)

[0076] Among them, I in,Lo Indicates the low-frequency information of the in-focus image; LP[.] indicates Gaussian low-pass filtering; I in (x) represents the in-focus part of the image; D(x, y) represents the corrected structured light image; (x, y) is the pixel coordinate.

[0077] The tomographic image contains high-frequency information and low-frequency information. The background light intensity image contains the high-frequency information of the tomographic image. The corrected structured light image contains the low-frequency information of the tomographic image. This embodiment adopts the idea of ​​fusion to fuse the high-frequency information of the tomographic image and the low-frequency information of the tomographic image to obtain a complete tomographic image of the full frequency domain. The background correction in D(x, y) eliminates the influence of the background light intensity fluctuation. What is retained in the image is the stripe information. After the background correction, the stripe information retained in D(x, y) is in focus, but due to the existence of the stripe information, it cannot be used directly. It needs to be low-pass filtered to extract the low-frequency in-focus information (i.e., the low-frequency information in the tomographic image). I obtained by low-pass filtering in,Lo The noise is effectively removed while only the low-frequency information in the tomographic image is retained.

[0078] In addition, according to formula 1.4, it can be seen that the background light intensity calculated in formula 1.3 contains the in-focus information and out-of-focus information, and there is no fringe interference. Therefore, if I is extracted b The high frequency part of the in-focus information is in,Lo By fusion, a complete tomographic image (complete in-focus image) can be obtained. b The defocus information is generally low-frequency, while the in-focus information is the clear texture details on the sample, which is high-frequency information. Therefore, a high-pass filter can be used to remove the out-of-focus part of the image and extract the high-frequency information of the in-focus part of the image (i.e., the high-frequency information in the tomographic image). In order to restore I in The high-frequency components in the low-pass filter LP are applied to the high-pass filter HP complementary to the low-pass filter LP. b , considering that the information of the out-of-focus part is expressed as low-frequency information, that is, HP[Iout (x,y)]≈0, and the contrast of the corrected structured light image is calculated by using the Laplace operator. The noisy wide-field image is subjected to Gaussian high-pass filtering based on contrast weighting, which can further suppress the high-frequency noise in the image. That is, the high-frequency in-focus information of the sample to be tested is specifically shown in the following formula:

[0079] I in,Hi (x,y)=HP[I b (x,y)]≈HP[I est (x,y)]·N m (x,y)(1.8)

[0080] Among them, I in,Hi Represents the high-frequency in-focus information in the tomographic response image; HP[.]=1-LP[.]; I b Indicates background light intensity; I est represents a uniformly illuminated image; (x, y) is the pixel coordinate; N m (x, y) represents the contrast of the structured light image after calculation using the Laplacian operator. m The modulation contrast calculated for D(x,y) is obtained by using the contrast image N m Weighted processing is performed to effectively suppress the influence of high-frequency noise.

[0081] The signal-to-noise ratio levels of different contrast areas calculated for the corrected structured light image are often different. Low-contrast pixels often have a lower signal-to-noise ratio than high-contrast pixels. When the contrast is close to zero, its high-frequency information will be completely submerged by the noise. Therefore, it is necessary to use a contrast-weighted method to suppress the high-frequency noise in different areas of the tomographic image in order to suppress the influence of noise and retain most of the effective high-frequency information.

[0082] Therefore, the above formula suppresses the influence of high-frequency noise and retains most of the high-frequency information components I of the tomographic image. in,Hi Combining equations 1.7 and 1.8, the tomographic distribution m expressed in the full frequency domain can be obtained HiLo (x, y), that is, the high-frequency in-focus information and the low-frequency in-focus information are fused to obtain the optical tomography image of the sample to be tested, as shown in the following formula:

[0083] m HiLo (x,y)=ηI in,Hi (x,y)+I in,Lo (x,y)

[0084] =ηHP[I est (x,y)]·N m (x,y)+LP[|I(x,y)-I est (x,y)|](1.9)

[0085] Among them, m HiLo (x, y) represents the optical tomography image; η represents the adjustment factor used to ensure that m HiLo (x) The continuity in the frequency domain can be adjusted between 0 and 1. This suppresses the imaging noise in the entire frequency domain and obtains a tomographic response image m with a high signal-to-noise ratio. HiLo (x, y). By extracting the pixel value of the tomographic image stack along the axial direction at each pixel position, the axial tomographic response curve m(z) at each pixel position can be obtained.

[0086] Furthermore, according to the study "Properties of a Defocused Optical System" by PAStokseth, the structured illumination tomography response m(z) can be approximated by the Gaussian function g(z), which is expressed as:

[0087]

[0088] Where a represents the maximum value of the axial tomographic response curve, h represents the position of the maximum value of the tomographic response, and σ FWHM represents a parameter related to the full width at half maximum (FWHM) of the tomographic response curve. Therefore, by least squares Gaussian fitting, the peak position of the axial tomographic response curve can be determined from the selected effective interval, thereby reconstructing the sample height, and its expression is:

[0089]

[0090] Where h' represents the calculated sample surface height, m p (z i ) represents z i Scanning position tomographic response intensity, z i ∈FWHM means that data points with a value greater than half of the peak intensity of the axial tomography response curve are selected as valid data points for least squares Gaussian fitting.

[0091] Furthermore, this embodiment provides a three-dimensional morphology measurement system based on continuous vertical scanning structured illumination, the system comprising a memory and a processor, the memory storing a computer program, and the processor executing any one of the above-mentioned three-dimensional morphology measurement methods based on continuous vertical scanning structured illumination when executing the computer program.

[0092] The measurement system used in this embodiment is as follows Figure 1As shown, it includes a spatial light modulator DMD (LightCrafter6500), a camera (acA1920-155um, Basler), a continuous scanning mechanism (a displacement mechanism for realizing continuous scanning), an LED light source (central wavelength of 518nm), a microscope objective lens (20×0.45 numerical aperture, Nikon Inc), and related lenses and fixed brackets.

[0093] First, generate N cosine fringe patterns with a fixed phase shift of 2π / N (the minimum value of N is 3, and 5 is generally selected) and input them into the DMD. Set the total scanning stroke of the axial scanning mechanism and set it at a fixed scanning step. After each scanning step, the DMD projects a pre-input cosine fringe pattern onto the sample surface and collects it through the camera. After the next scanning step is moved, switch to the next phase-shifted fringe pattern projection, and repeat the above steps in sequence until the scan is completed.

[0094] After scanning is completed, a collection image stack can be obtained, and the tomographic response image is calculated using the continuous scanning structured illumination optical tomography method based on sliding time window background estimation and fusion proposed in the present invention. Then, the sample surface morphology is reconstructed according to formula 1.11.

[0095] Aiming at the problem of low measurement efficiency of structured light illumination micro measurement technology, the present invention proposes a continuous scanning structured illumination optical tomography measurement method based on sliding time window background estimation and fusion.

[0096] Firstly, a continuous scanning structured illumination measurement method is used, where only one structured illumination phase-shifted cosine fringe pattern is collected at each scanning position to improve measurement efficiency.

[0097] Then, the background light intensity is estimated by using the adjacent collected images in the time domain to obtain a uniformly illuminated image. At the same time, the estimated background light intensity is used to correct the background light intensity of the collected fringe pattern to suppress the background light intensity fluctuation caused by ambient stray light and imaging noise during the scanning process.

[0098] The contrast of the corrected fringe pattern is calculated, and the uniformly illuminated image is subjected to contrast-weighted high-pass filtering to obtain the high-frequency in-focus information of the sample under test. The collected fringe pattern after background light intensity correction is subjected to low-pass filtering to eliminate the fringe information and obtain the low-frequency in-focus information of the sample under test. The above two are weightedly fused to obtain an optical tomography image, thereby realizing three-dimensional reconstruction of the surface of the sample under test.

[0099] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A three-dimensional shape measurement method based on continuous vertical scanning structured illumination, characterized in that: include: S1, using the continuous vertical scanning structured light illumination measurement method, collects a phase-shifted cosine fringe structured light image of the sample to be tested at each vertical scanning position; S2, using the adjacent structured light images in the time domain to estimate the background light intensity, and obtaining the estimated background light intensity as a uniformly illuminated image; at the same time, using the estimated background light intensity to correct the background light intensity of the structured light image, and obtaining the corrected structured light image; S3, performing high-pass filtering on the uniformly illuminated image to obtain high-frequency in-focus information of the sample to be tested; Performing low-pass filtering on the corrected structured light image to obtain low-frequency in-focus information of the sample to be tested; The high-frequency in-focus information and the low-frequency in-focus information are fused to obtain an optical tomography image of the sample to be tested; S4, based on the optical tomography image, obtaining the three-dimensional morphology of the sample to be tested.

2. The three-dimensional shape measurement method based on continuous vertical scanning structured illumination according to claim 1, characterized in that: S1 specifically includes: A plurality of cosine fringe patterns with fixed phase shifts are preset, and one of the cosine fringe patterns is projected onto the surface of the sample to be tested at each vertical scanning position. During the scanning process of the sample to be tested, the plurality of cosine fringe patterns are projected in a cycle in sequence.

3. The three-dimensional shape measurement method based on continuous vertical scanning structured illumination according to claim 2, characterized in that: In S2, the background light intensity is estimated by using the structured light images adjacent in the time domain, and the estimated background light intensity is obtained as a uniform illumination image, which specifically includes: For any vertical scanning position, taking a plurality of adjacent structured light images around the vertical scanning position as the center; averaging a plurality of adjacent structured light images and filtering the average to obtain an estimated background light intensity at any scanning position as a uniformly illuminated image; The number of the plurality of adjacent structured light images is consistent with the number of the preset cosine fringe patterns.

4. The three-dimensional shape measurement method based on continuous vertical scanning structured illumination according to claim 3, characterized in that: The uniformly illuminated image at any vertical scanning position is specifically expressed as follows: Among them, I b Indicates background light intensity; I est represents uniform illumination image; H represents filter; z i represents the i-th scanning position; I represents the structured light image; and N represents the number of the preset cosine fringe patterns.

5. The three-dimensional shape measurement method based on continuous vertical scanning structured illumination according to any one of claims 1 to 4, characterized in that: In S2, the estimated background light intensity is used to correct the background light intensity of the structured light image to obtain the corrected structured light image, which specifically includes: The collected structured light image is subtracted from the uniformly illuminated image to obtain the corrected structured light image.

6. The three-dimensional shape measurement method based on continuous vertical scanning structured illumination according to any one of claims 1 to 4, characterized in that: S3 performs high-pass filtering on the uniformly illuminated image to obtain high-frequency in-focus information of the sample to be tested, specifically including: The contrast of the modified structured light image is calculated using a Laplace operator, and contrast-weighted high-pass filtering is performed on the uniformly illuminated image to obtain high-frequency in-focus information of the sample to be tested.

7. The three-dimensional shape measurement method based on continuous vertical scanning structured illumination according to any one of claims 1 to 4, characterized in that: The corrected structured light image is low-pass filtered to obtain the low-frequency in-focus information of the sample to be tested, as shown in the following formula: I in,Lo (x,y)=LP[I in (x,y)]≈LP[D(x,y)]; Among them, I in,Lo represents the low-frequency in-focus information in the tomographic response image; LP[.] represents Gaussian low-pass filtering; I in (x) represents the in-focus part of the image; D(x) represents the corrected structured light image; (x, y) are pixel coordinates.

8. The three-dimensional shape measurement method based on continuous vertical scanning structured illumination according to claim 7, characterized in that: The uniformly illuminated image is subjected to high-pass filtering to obtain high-frequency in-focus information of the sample to be tested, as shown in the following formula: I in,Hi (x,y)=HP[I b (x,y)]≈HP[I est (x,y)]·N m (x,y); Among them, I in,Hi Represents the high-frequency in-focus information in the tomographic response image; HP[.]=1-LP[.]; I b Indicates background light intensity; I est represents a uniformly illuminated image; x is the pixel coordinate; N m (x, y) represents the contrast calculated for the corrected structured light image.

9. The three-dimensional shape measurement method based on continuous vertical scanning structured illumination according to claim 8, characterized in that: The high-frequency in-focus information and the low-frequency in-focus information are fused to obtain the optical tomography image of the sample to be tested, as shown in the following formula: m HiLo (x,y)=ηI in,Hi (x,y)+I in,Lo (x,y)=ηHP[I est (x,y)]·N m (x,y)+LP[|I(x,y)-I est (x,y)|]; Among them, m HiLo (x, y) represents the optical tomography image; η represents the adjustment factor.

10. A three-dimensional shape measurement system based on continuous vertical scanning structured illumination, characterized in that: The system includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the three-dimensional shape measurement method based on continuous vertical scanning structured illumination described in any one of claims 1 to 9 is executed.

Images