A method and system for measuring complex surface topography

Abstract

The application belongs to the field of surface topography measurement, and particularly discloses a complex surface topography measurement method and a measurement system, which comprises the following steps: collecting a structured light image of a sample to be measured, obtaining an optical tomography image, a wide-field image and a modulation contrast image; processing the optical tomography image, and reconstructing the surface topography of the sample according to the processed optical tomography image; the processing of the optical tomography image comprises the following steps: performing low-pass filtering on the optical tomography image through a Gaussian low-pass filter to obtain a low-frequency component; performing high-pass filtering on the wide-field image through a Gaussian high-pass filter which is complementary to the Gaussian low-pass filter to obtain a high-frequency component with noise; performing weighted processing on the high-frequency component with noise by using the modulation contrast image to obtain a high-frequency component; and performing weighted fusion on the low-frequency component and the high-frequency component to obtain the processed optical tomography image. The application can reduce the noise influence of the optical tomography image and improve the complex surface topography measurement effect.

Description

Technical Field

[0001] This invention belongs to the field of surface morphology measurement, and more specifically, relates to a method and system for measuring complex surface morphology. Background Technology

[0002] In advanced manufacturing fields such as integrated circuits (ICs), microelectromechanical systems (MEMS), display panels, and electronics manufacturing, complex composite surfaces composed of regions with varying reflectivity and roughness formed by different materials and processes are widely present. Examples include 3D integrated circuit packaging bonding interfaces, MEMS structural surfaces, printed circuit board surfaces, electronic packaging interfaces, microfluidic chip structural surfaces, and textured surfaces. The quality of these wide dynamic range composite surfaces, rich in regions with different optical properties, significantly impacts the functional characteristics of manufactured products, necessitating rapid and high-precision measurement instruments adapted for their quality inspection and analysis.

[0003] Existing non-contact optical measurement methods for surface morphology, such as white light interferometers, confocal microscopes, and zoom microscopes, struggle to achieve rapid and high-precision measurements of surfaces with wide dynamic ranges. Solutions integrating multiple measurement and sensing principles not only suffer from low measurement efficiency but also face challenges in unifying and fusing coordinates from multiple sensors. Structured illumination microscopy is particularly attractive due to its incoherent, surface-scanning imaging characteristics and adaptability to surfaces with diverse features (e.g., rough or smooth).

[0004] However, current methods and devices for measuring surface topography under structured lighting have the following problems:

[0005] 1) When dealing with surfaces with complex features (such as smooth, high-curvature, or transparent multilayer structures) and materials with significant differences in reflectivity, the fringe contrast in structured illumination images captured by a camera is easily affected and rapidly decreases. This decrease in contrast leads to a reduction in the signal-to-noise ratio of the axial response curve, resulting in numerous outliers, missing data, and significant measurement noise in existing surface reconstruction algorithms when measuring complex surfaces. The noise inherent in optical tomography images also contributes to a poor signal-to-noise ratio in the axial response signal, thus affecting the effectiveness of surface topography measurements.

[0006] 2) Existing structural lighting surface topography measurement devices mostly use a structure of a fully internal reflection prism combined with a spatial light modulation device. However, due to the relatively complex design and high processing cost of the fully internal reflection prism, it will significantly increase the early development cost of the structural lighting surface topography measurement device and its prototype, and is not suitable for the development of high-precision surface topography measurement instruments with relatively small production batches.

[0007] 3) To ensure the uniformity of illumination on the DMD surface, existing structured illumination micro-surface measurement devices mostly use Kohler illumination structures. Because this illumination structure requires multiple lens groups to ensure uniform illumination of the object surface, the measurement device is often quite large.

[0008] 4) Because highly coherent lasers are prone to speckle, affecting the accuracy of surface measurements under structured illumination, and broadband light sources such as xenon and halogen lamps are relatively expensive, existing illumination systems for surface measurement using structured illumination microscopes mostly use high-brightness, low-cost LED light sources. However, high-power LED light sources generate significant heat, requiring passive cooling structures to address their heat dissipation issues. This results in poor heat dissipation and increases the complexity of the measurement device. Summary of the Invention

[0009] In view of the above-mentioned defects or improvement needs of the existing technology, the present invention provides a method and system for measuring complex surface morphology, the purpose of which is to reduce the noise influence of optical tomography images and improve the measurement effect of complex surface morphology.

[0010] To achieve the above objectives, according to one aspect of the present invention, a method for measuring the morphology of complex surfaces is proposed, comprising the following steps:

[0011] Structured light images of the sample to be tested are acquired, and corresponding optical tomographic images, wide-field images, and modulation contrast images are obtained. The optical tomographic images are processed, and the surface morphology of the sample is reconstructed based on the processed optical tomographic images.

[0012] The processing of the optical tomography image includes the following steps:

[0013] The optical tomography image is low-pass filtered by a Gaussian low-pass filter to obtain the low-frequency component I. LOS ;

[0014] High-pass filtering of a wide-field image is performed using a Gaussian high-pass filter that is complementary to a Gaussian low-pass filter to obtain the noisy high-frequency component I. HOS ;

[0015] Using modulated contrast image I MC For the noisy high-frequency component I HOS Weighted noise reduction is performed to obtain the high-frequency component I. WHOS ;

[0016] Low-frequency component I LOS With high frequency component I WHOS Weighted fusion is performed to obtain the processed optical tomography image I. HLOS .

[0017] As a further preferred option, a modulated contrast image I is used.MC For the noisy high-frequency component I HOS Weighted noise reduction is performed, specifically as follows:

[0018] I WHOS (x)=Λ(1-I MC (x))·I HOS (x)

[0019]

[0020] Where x represents the pixel position on the image, Λ(·) is the weight function, and y is the intermediate variable.

[0021] As a further preferred option, the optical tomography image is first subjected to noise reduction processing before low-pass filtering with a Gaussian low-pass filter.

[0022] As a further preferred option, when performing noise reduction processing on optical tomography images, image noise reduction is performed in the sparse domain.

[0023] According to another aspect of the present invention, a measurement system for implementing the above-described complex surface morphology measurement method is provided, comprising a measurement device and a processor, wherein:

[0024] The measuring device is used to acquire structured light images of the sample to be tested;

[0025] The processor is used to obtain corresponding optical tomographic images, wide-field images, and modulated contrast images based on structured light images; then, it processes the optical tomographic images and reconstructs the surface morphology of the sample based on the processed optical tomographic images.

[0026] As a further preferred embodiment, the measuring device includes a core optical measuring module, which comprises a liquid optical waveguide, a beam collector, a beam expander, a plane mirror, a DMD, a cubic beam splitter, a sleeve lens, a microscope objective, a right-angle mirror, and a CMOS camera, wherein:

[0027] The light emitted from the far end of the liquid optical waveguide is collected and collimated by a beam collector, and then expanded by a beam expander before being reflected by a plane mirror to the DMD. The DMD adjusts the incident light into a striped pattern with varying intensity and brightness. This striped pattern is reflected by a cubic beam splitter and then projected onto the sample surface through a sleeve lens and a microscope objective.

[0028] The stripe pattern is reflected after being highly modulated by the sample surface. After passing through a microscope objective, a sleeve lens, and a cubic beam splitter, it is finally imaged in a CMOS camera by a right-angle mirror to obtain a structured light image.

[0029] As a further preferred embodiment, the measuring device also includes a piezoelectric ceramic objective actuator for driving the movement of the microscope objective.

[0030] As a further preferred embodiment, the measuring device also includes a displacement stage and a fiber optic interferometer. The displacement stage is used to drive the core optical measuring module to move, and the fiber optic interferometer is coaxially mounted above the microscope objective to measure the displacement when the core optical measuring module moves.

[0031] As a further preferred embodiment, the liquid waveguide is provided with a light source by an externally mounted LED, and the external LED is equipped with an active water cooling system.

[0032] As a further preferred embodiment, the core optical measurement module also includes a dual polarizer, which is installed between the light collector and the beam expander to control the intensity of light incident on the DMD.

[0033] In summary, compared with the prior art, the above-described technical solutions conceived by this invention mainly possess the following technical advantages:

[0034] 1. This invention divides a complex noise reduction task into two concise sub-steps by processing high-frequency and low-frequency components separately. In the high-frequency information processing sub-step, weighted noise reduction of noisy high-frequency components is performed based on the modulated contrast image, thereby suppressing high-frequency noise in the tomographic image. Compared with traditional structured illumination surface reconstruction methods, the method of this invention exhibits higher measurement robustness and can achieve excellent measurement results on many complex surfaces.

[0035] 2. The low-frequency information processing sub-step of this invention performs image denoising in the sparse domain before low-pass filtering of the optical tomographic image, ignoring the influence of high-frequency information and achieving rapid denoising. Thus, the method of this invention can achieve excellent denoising effects in both high-frequency and low-frequency processing steps, thereby greatly enhancing the reconstruction quality of the optical tomographic image and the signal-to-noise ratio of the axial response signal, and significantly improving the reconstruction effect of complex surfaces.

[0036] 3. This invention proposes a compact, low-cost microscopic surface topography measurement system based on structured illumination to achieve rapid and high-precision measurement of wide dynamic combination surface structures composed of regions with different reflectivities and roughnesses formed by different materials and processes. This is of great significance for solving quality and functional assurance problems in advanced manufacturing fields such as IC, MEMS, display panels and electronic manufacturing.

[0037] 4. To ensure a compact design, a plane mirror was chosen instead of a fully internal reflection prism, thus avoiding a complex design process and significantly reducing manufacturing costs. Furthermore, the use of a plane mirror simplifies system installation and commissioning. Simultaneously, to improve efficiency and reduce costs, the illumination and imaging optical paths were integrated, avoiding the cumbersome designs common in traditional technologies where illumination and imaging were separated. This common optical path design simplifies the structure, reduces aberrations caused by redundant optical components, and further simplifies the commissioning process.

[0038] 5. Traditional structured illumination microscopes often use high-power collimated LED light sources to ensure uniformity and high brightness. However, such LED light sources generate a large amount of heat, which, if not properly managed, can affect measurement stability. Therefore, a bulky passive cooling system is required, significantly increasing the overall size of the device. To address this issue, this invention proposes a solution that separates the light source from the measurement system, effectively isolating the thermal effects of the light source from interfering with the system. This independent design not only allows for the use of higher-power LED light sources but also enables the integration of a highly efficient active water cooling system, eliminating the LED's thermal problems and avoiding vibration issues that might arise from active cooling. Furthermore, by employing high-quality liquid waveguide technology, the large-angle scattered light emitted by the LED is homogenized and collimated, successfully constructing a high-quality, high-brightness, and compact collimated illumination system.

[0039] 6. This invention employs a piezoelectric ceramic objective lens actuator. Benefiting from the nanoscale positioning accuracy and excellent repeatability exhibited by piezoelectric ceramics within a measurement range of tens of micrometers, scanning using this actuator enables measurements of microscale samples requiring high precision. Furthermore, by setting up a displacement stage to drive the core optical measurement module for scanning, combined with a traceable high-precision fiber optic interferometer as a metrology device, the measurement range is effectively extended to tens of millimeters, enabling accurate measurements of large-scale sample surfaces. This ensures that the measurement system can accurately measure samples of various sizes. Attached Figure Description

[0040] Figure 1 This is an axonometric view of the overall structure of the complex surface morphology measuring device according to an embodiment of the present invention;

[0041] Figure 2 This is an exploded view of the overall structure of the complex surface morphology measuring device according to an embodiment of the present invention;

[0042] Figure 3 This is a flowchart of the complex surface morphology measurement method according to an embodiment of the present invention;

[0043] Figure 4 This is a flowchart of the optical tomography image processing method according to an embodiment of the present invention;

[0044] Figure 5 This is a flowchart of a first preferred method for optical tomography image processing according to an embodiment of the present invention;

[0045] Figure 6 This is a flowchart of a second preferred method for optical tomography image processing according to an embodiment of the present invention.

[0046] In all the accompanying drawings, the same reference numerals are used to denote the same elements or structures, wherein: 1-base, 2-three-axis displacement stage, 3-column, 4-rapid displacement mechanism, 5-mounting bracket, 6-displacement stage, 7-adapter plate, 8-measurement module body, 9-light intensity detector, 10-camera mount, 11-camera C-mount extension tube, 12-CMOS camera, 13-fiber interferometer, 14-resonant motor kit, 15-light source mount, 16-light collector, 17-liquid waveguide distal mount, 18-liquid waveguide, 19-... - Resonant motor mounting base, 20- Polarizer mounting base, 21- Beam expander, 22- DMD mounting adapter plate, 23- DMD main board, 24- Right-angle mirror pressure plate, 25- Plane mirror, 26- DMD extension board, 27- DMD extension board mounting adapter plate, 28- Cubic beam splitter pressure plate, 29- Right-angle mirror, 30- Cubic beam splitter, 31- Sleeve lens, 32- Mounting base, 33- Objective turret adapter ring, 34- Objective turret, 35- Driver adapter ring, 36- Piezoelectric ceramic objective driver, 37- Microscope objective. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0048] This invention provides a complex surface morphology measurement system, which includes a measuring device and a processor, wherein:

[0049] The measuring device is used to acquire structured light images of the sample to be tested;

[0050] The processor is used to obtain corresponding optical tomographic images, wide-field images, and modulated contrast images based on structured light images; then, it processes the optical tomographic images and reconstructs the surface morphology of the sample based on the processed optical tomographic images.

[0051] Specifically, such as Figure 1 and Figure 2As shown, the measuring device includes a core optical measuring module, which includes a measuring module body 8, and a liquid optical waveguide 18, a light-collecting mirror 16, a beam expander 21, a plane mirror 25, a DMD, a cubic beam splitter 30, a sleeve lens 31, a microscope objective 37, a right-angle mirror 29, and a CMOS camera 12 mounted on the measuring module body 8.

[0052] The liquid waveguide 18 is fixed to the light source mounting base 15 via the liquid waveguide distal mounting base 17, and then illuminated by an externally mounted LED. This separation of the light source from the measurement device isolates the thermal effect of the light source. Combined with a highly efficient active water cooling system, the liquid waveguide technology is then used for light homogenization and collimation. The light emitted from the distal end of the liquid waveguide 18 is collected and collimated by a high numerical aperture focusing lens 16. Subsequently, after passing through a high-quality beam expander 21, the collimated beam is further expanded to ensure that it has uniform light intensity and can completely cover the DMD chip when reflected by the plane mirror 25 to the DMD mainboard 23. Then, the field-programmable gate array (FPGA) on the DMD expansion board 26 controls millions of micromirrors on the DMD chip to modulate the uniform incident light into a striped pattern with varying light intensity. This pattern is reflected by a cubic beam splitter prism 30, then through a sleeve lens 31, and finally projected onto the sample surface by a microscope objective 37. When the sample surface is at the focal plane, the fringe pattern will be very clear with high contrast; conversely, when the sample surface area is out of focus, the fringe pattern in that area will appear blurry with lower modulation contrast. After high modulation of the sample surface, the fringe pattern with varying brightness is reflected and, through the same microscope objective 37 and sleeve lens 31, then through the cubic beam splitter prism 30, and finally imaged onto the target surface of the CMOS camera 12 by the right-angle mirror 29, to obtain a structured light image. To obtain accurate measurement results, the system uses a DMD to phase-shift the projected fringe pattern and simultaneously triggers the CMOS camera 12 to acquire a series of phase-shifted fringe images; these images use a specific contrast decoding algorithm to calculate the contrast value at each pixel position in real time, thereby obtaining an optical slice image similar to that of a confocal microscope.

[0053] In some embodiments, the core optical measurement module employs dual polarizers to adjust the light intensity, ensuring stable and reliable light intensity, thereby reducing the requirements for the accuracy of constant current source current control. These dual polarizers are mounted between the beam collector 16 and the beam expander 21. One polarizer is placed on the resonant motor assembly 14 within the resonant motor mounting base 19, and the other is placed on the polarizer mounting base 20. By controlling the rotation of the upper polarizer via the resonant motor, the light intensity incident on the DMD can be precisely controlled.

[0054] In some embodiments, a light intensity detector 8 is provided on the measurement module body 8; a CMOS camera 12 is mounted on a camera mounting base 10 via a camera C-mount extension tube 11, and the camera mounting base 10 is fixed to the measurement module body 8. A DMD motherboard 23 is fixed to the side of the measurement module body 8 via a DMD mounting adapter 22, while a DMD expansion board 26 is fixed to the front of the measurement module body 8 via a DMD expansion board mounting adapter 27. The DMD motherboard 23 and the DMD expansion board 26 are connected by a flexible flat cable to transmit control signals. A right-angle reflector 29 and a cubic beam splitter prism 30 are fixed inside the measurement module body 8 via a right-angle reflector plate 24 and a cubic beam splitter plate 28, respectively.

[0055] The aforementioned core optical measurement module not only integrates the basic functions of imaging and measurement, but also features a dustproof design specifically for various measurement environments, forming a compact and functionally independent measurement module. This independent design allows the measurement module to be perfectly integrated into a column-type three-axis worktable, and can also be easily integrated into other gantry-type supports or robotic arm ends with suitable mounting holes, demonstrating strong adaptability and flexibility.

[0056] Specifically, the measuring device also includes a scanning component for adjusting the distance between the objective lens and the sample to achieve scanning.

[0057] In some embodiments, the scanning assembly includes an objective converter 34 and a piezoelectric ceramic objective driver 36. The objective converter 34 is mounted on the mounting base 32 via an objective converter adapter ring 33. Multiple microscope objectives 37 can be mounted under the objective converter 34, and the microscope objectives 37 can be quickly switched as needed. Alternatively, the piezoelectric ceramic objective driver 36 can be mounted under the objective converter 34 first, and then the microscope objectives 37 can be further mounted on it.

[0058] In some embodiments, the scanning assembly includes a displacement stage 6 and a fiber optic interferometer 13. The displacement stage 6 is connected to the core optical measurement module via an adapter plate 7 to drive the core optical measurement module to move. The fiber optic interferometer 13 is coaxially mounted above the microscope objective 37 via a mounting bracket 5 to measure the displacement when the core optical measurement module moves.

[0059] The above design provides two independent vertical scanning methods: a high-precision piezoelectric ceramic objective scanning scheme and a large-scale traceable motor-driven continuous scanning scheme. For the piezoelectric ceramic objective scanning scheme, the microscope objective 37 is driven by a piezoelectric ceramic objective driver 36 connected to a driver adapter ring 35 via an objective converter 34, achieving precise measurement. Thanks to the nanometer-level positioning accuracy and excellent repeatability exhibited by piezoelectric ceramics within a measurement range of tens of micrometers, this scanning scheme is particularly suitable for measuring small-scale samples with high precision requirements. The large-scale motor-driven continuous vertical scanning scheme uses a high-precision motorized displacement stage 6 to drive the core optical measurement module for vertical scanning, combined with a traceable high-precision fiber optic interferometer 13 as a metrology device, and achieves accurate measurement of large-scale sample surfaces based on the Abbe measurement principle.

[0060] By using any of the aforementioned vertical scanning schemes—that is, scanning the objective lens or the entire core optical module along the z-axis (perpendicular to the sample plane)—a tomographic image stack with intensity variations along the z-axis can be obtained. In this image stack, each pixel corresponds to a Gaussian-shaped axial response curve, and the peak position of this curve represents the sample's height at that point. By employing a high-precision Gaussian peak localization algorithm, the peak positions of the axial response curves at all pixels can be calculated, thereby ultimately achieving a complete reconstruction of the sample's three-dimensional surface morphology.

[0061] In some embodiments, the measuring device further includes a column-type triaxial stage, which comprises a base 1, a triaxial displacement stage 2, a column 3, and a rapid displacement mechanism 4. The base 1 and column 3 are preferably made of marble, ensuring excellent thermal stability and superior shock absorption. The high-precision triaxial displacement stage 2, mounted on the marble base 1, allows for convenient movement and positioning of the sample. The rapid displacement mechanism 4, mounted on the column 3, is used to quickly adjust and fix the position of the core optical measurement module, thereby adapting to samples of different heights and sizes.

[0062] This invention provides a method for measuring complex surface morphology, such as... Figure 3 As shown, it includes the following steps:

[0063] S1: Prepare a set of structured lighting patterns with equal step phase shifts (in this embodiment, the number of patterns n=5) and load them into the memory of the DMD; and set the core optical measurement module to the initial position.

[0064] S2: The core optical measurement module is adjusted to a specific axial position by the scanning component. Then, the DMD switches the loaded structured illumination patterns according to a predetermined sequence. Simultaneously, to ensure synchronization, the CMOS camera is triggered to acquire the corresponding structured light image each time the DMD switches patterns. Based on the acquired set of structured light images, a contrast decoding algorithm can be used to calculate and obtain the optical tomographic image at that axial position. Other intermediate calculation results, including wide-field images and modulated contrast images, can also be obtained.

[0065] S3: Based on wide-field image I WF and modulated contrast image I MC For optical tomography image I OS The image is then processed to obtain a denoised optical tomographic image.

[0066] S4: Adjust the core optical measurement module to the next axial position, and repeat steps S2 and S3 until the entire vertical scanning process is completed.

[0067] S5: After all scans are completed, the obtained optical tomography image stack is used to perform high-precision Gaussian peak localization, thereby accurately reconstructing the surface morphology of the sample.

[0068] S6: End the measurement process and input the obtained data into the 3D surface morphology display and evaluation software for subsequent data processing and analysis.

[0069] Specifically, such as Figure 4 As shown, step S3 includes the following steps:

[0070] S31: Optical tomography image I is processed through a Gaussian low-pass filter (LP). OS Low-pass filtering is performed to obtain the low-frequency component I of the optical tomography image. LOS .

[0071] S32: Because S31 can only obtain optical tomographic images I OS The low-frequency components after noise reduction are used to recover the optical tomography image I. OS The missing high-frequency components are addressed by using a Gaussian high-pass filter (HP = 1 - LP) complementary to LP for the wide-field image I. WF High-pass filtering is performed to obtain the noisy high-frequency component I. HOS .

[0072] Specifically, low-pass filtering and high-pass filtering are relative concepts. The cutoff frequency of the Gaussian filter is selected according to different applications. Filters with frequencies below the cutoff frequency are low-frequency (Gaussian low-pass filter), and filters with frequencies above the cutoff frequency are high-frequency (Gaussian high-pass filter).

[0073] S33: In order to further suppress high-frequency components I in tomographic images HOS High-frequency noise in the image, based on the modulation contrast image I MC Construct a weighting function for the noisy high-frequency component I. HOS Processing is performed to obtain the high-frequency component I. WHOS Specifically:

[0074] I WHOS (x)=Λ(1-I MC (x))·I HOS (x)

[0075]

[0076] Where x represents the pixel position in the image, y is an intermediate variable, and Λ(·) is the weighting function, which is a function similar to the at-focus optical transfer function (OTF). Since pixels with low modulation contrast typically have more noise than pixels with high modulation contrast, after weighting, all potentially low signal-to-noise ratio (SNR) high-frequency signals are further suppressed, resulting in the tomographic image I. OS Low-noise high-frequency component I WHOS .

[0077] S34: To obtain a full-resolution optical tomographic image after noise reduction I HLOS The low-frequency component I LOS With high frequency component I WHOS Weighted fusion is performed to obtain the processed optical tomography image I. HLOS Specifically:

[0078] I HLOS (x)=η·I MC (x)·I WHOS (x)+I LOS (x)

[0079] Where η is a scaling factor used to ensure the low-frequency component I LOS With high frequency component I WHOS It can be seamlessly integrated in the frequency domain.

[0080] In some embodiments, the optical tomography image I is filtered by a Gaussian low-pass filter in step S31. OS Before performing low-pass filtering, the optical tomography image is first subjected to noise reduction processing.

[0081] Furthermore, such as Figure 5 As shown, for optical tomography image I OS During noise reduction, a sparse domain noise reduction method is used to obtain image I. DOS Then, a Gaussian low-pass filter is used to filter image I. DOSLow-pass filtering is performed to eliminate high-frequency artifacts that may be introduced during sparse noise reduction, resulting in a cleaner low-frequency component I of the optical tomography image. LOS Furthermore, it is easy to understand that adopting, for example Figure 6 The state-of-the-art (SOTA) image denoising methods shown (i.e., other various image denoising methods) can also achieve denoising.

[0082] Furthermore, the use of a plug-and-play (PnP) post-processing noise reduction design not only enhances the robustness of traditional multi-step phase-shifting algorithms but also makes it applicable to other methods such as defocus background estimation algorithms.

[0083] In some embodiments, image denoising is performed using a sparse domain denoising method, including the following steps:

[0084] 1) Use any noise estimation algorithm to analyze the optical tomography image I. OS Noise estimation is performed to obtain the noise level function (NLF).

[0085] 2) Because the obtained NLF generally has the same characteristics as optical tomography image I OS To facilitate subsequent processing, the optical tomography image I is processed using variance-stabilized transform (VST) based on the intensity value correlation characteristics. OS , making I OS The transformed image's NLF is decoupled from the optical tomography intensity value, becoming Gaussian white noise with a variance of 1.

[0086] 3) Use the BM3D algorithm to match and stack similar sub-image blocks in the image after the above VST transformation.

[0087] 4) Perform a three-dimensional discrete cosine transform (DCT) on each stack containing similar sub-image blocks obtained above, and set an appropriate threshold coefficient for filtering, followed by performing an inverse DCT. This step fully utilizes the sparsity of the similar sub-image block signals in the discrete cosine transform domain, thereby effectively separating them from noise and obtaining denoised image blocks.

[0088] 5) Recombine these denoised image blocks according to the position of each sub-image block in the original image to obtain the final denoised optical tomographic image I in the sparse domain. DOS .

[0089] Those skilled in the art will readily understand that the above description is merely 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 within the scope of protection of the present invention.

Claims

1. A method for measuring the morphology of complex surfaces, characterized in that, Includes the following steps: Structured light images of the sample to be tested are acquired, and corresponding optical tomographic images, wide-field images, and modulation contrast images are obtained. The optical tomographic images are processed, and the surface morphology of the sample is reconstructed based on the processed optical tomographic images. The processing of the optical tomography image includes the following steps: Low-frequency components are obtained by low-pass filtering the optical tomography image using a Gaussian low-pass filter. ; High-pass filtering of a wide-field image is performed using a Gaussian high-pass filter that is complementary to a Gaussian low-pass filter to obtain noisy high-frequency components. ; Using modulated contrast images For noisy high-frequency components Weighted noise reduction is performed to obtain the high-frequency components. Specifically: in, x Indicates the pixel position on the image. For the weight function, y As an intermediate variable; Low frequency components With high frequency components Weighted fusion is performed to obtain the processed optical tomographic image. .

2. The method for measuring complex surface morphology as described in claim 1, characterized in that, Before applying a Gaussian low-pass filter to the optical tomography image, noise reduction processing is performed on the optical tomography image.

3. The method for measuring complex surface morphology as described in claim 2, characterized in that, When performing noise reduction on optical tomography images, image noise reduction is performed in the sparse domain.

4. A measurement system for implementing the complex surface morphology measurement method as described in any one of claims 1-3, characterized in that, Includes a measuring device and a processor, wherein: The measuring device is used to acquire structured light images of the sample to be tested; The processor is used to obtain corresponding optical tomographic images, wide-field images, and modulated contrast images based on structured light images; then, it processes the optical tomographic images and reconstructs the surface morphology of the sample based on the processed optical tomographic images.

5. The measurement system as described in claim 4, characterized in that, The measuring device includes a core optical measuring module, which comprises a liquid optical waveguide (18), a beam collector (16), a beam expander (21), a plane mirror (25), a DMD, a cubic beam splitter (30), a sleeve lens (31), a microscope objective (37), a right-angle mirror (29), and a CMOS camera (12), wherein: The light emitted from the far end of the liquid optical waveguide (18) is collected and collimated by the beam collector (16). The collimated light is expanded by the beam expander (21) and then reflected by the plane mirror (25) to the DMD. The DMD adjusts the incident light into a striped pattern with varying light intensity. The striped pattern is reflected by the cubic beam splitter (30) and then projected onto the sample surface through the sleeve lens (31) and the microscope objective (37). The stripe pattern is reflected after being highly modulated on the sample surface. After passing through the microscope objective (37), the sleeve lens (31), and the cubic beam splitter (30), it is finally imaged in the CMOS camera (12) by the right-angle mirror (29) to obtain a structured light image.

6. The measurement system as described in claim 5, characterized in that, The measuring device also includes a piezoelectric ceramic objective drive (36) for driving the movement of the microscope objective (37).

7. The measurement system as described in claim 5, characterized in that, The measuring device also includes a displacement stage (6) and a fiber optic interferometer (13). The displacement stage (6) is used to drive the core optical measurement module to move. The fiber optic interferometer (13) is coaxially mounted above the microscope objective (37) and is used to measure the displacement when the core optical measurement module moves.

8. The measurement system as described in claim 5, characterized in that, The liquid waveguide (18) is provided with a light source by externally mounted LEDs, which are equipped with an active water cooling system.

9. The measurement system according to any one of claims 5-8, characterized in that, The core optical measurement module also includes a dual polarizer, which is installed between the light collector (16) and the beam expander (21) to control the intensity of light incident on the DMD.

Images