Full-field optical coherence tomography three-dimensional industrial defect detection system and method for large size samples

Abstract

The application relates to the technical field of defect detection, in particular to a full-field optical coherence tomography three-dimensional industrial defect detection system and method suitable for large-size samples, wherein an interference fringe generated by a full-field interference imaging system illuminates a sample; a three-dimensional displacement platform is controlled to move the sample in a transverse direction multiple times according to a preset track, a wavelength tunable laser is synchronously controlled to complete wavelength scanning within the exposure time of a surface array camera after each movement of the sample is completed; the surface array camera transmits three-dimensional data collected at each position of the sample to a controller; and the controller performs image reconstruction, splicing fusion and rendering display on the three-dimensional original data. The full-field OCT technology and laser sweep frequency technology of the application are combined, partial three-dimensional structure information of the large-size sample is obtained according to the principle of the full-field sweep frequency OCT, the partial reconstruction results of adjacent areas are spliced and fused by means of an image splicing algorithm, and complete three-dimensional imaging results are presented, so that the upper and lower surfaces and internal defects of the whole large-size sample can be detected.

Description

Technical Field

[0001] This application relates to the field of defect detection technology, and in particular to a full-field optical coherence tomography three-dimensional industrial defect detection system and method suitable for large-size samples. Background Technology

[0002] In the field of industrial quality inspection, current common optical defect detection methods mainly acquire two-dimensional surface information of samples. However, for the detection of internal defects in large quantities of transparent / semi-transparent samples, differentiation of upper and lower surfaces, and imaging of multilayer film structures, three-dimensional imaging methods are necessary. Optical Coherence Tomography (OCT), as a three-dimensional tomographic imaging technology, achieves three-dimensional "tomographic" imaging of the sample's interior by measuring the scattered / reflected light signals at different depths within the sample. In recent years, it has been widely used in the industrial quality inspection industry. However, the current application scenarios of industrial OCT are limited to the sampling inspection of products with small inspection areas (on the order of square millimeters). For the detection of internal defects in large-sized samples, such as LCD screens, large lenses, mobile phone back panels, and watch crystals in the 3C industry, industrial OCT technology is required to have a large imaging field of view.

[0003] Most existing industrial OCT technologies employ point scanning, utilizing the principle of confocal detection to perform a two-dimensional lateral scan of the sample, point by point (depth scan), and then acquiring the sample's three-dimensional structural information based on tomography principles. To achieve wide-field imaging of large samples, the scanning lens needs to be perfectly focused while scanning at large angles, making lens design difficult, bulky, and costly. While using point scanning OCT to scan large samples with its limited imaging field of view can expand the field of view by moving the sample, implementing a three-dimensional image stitching algorithm for this scanning method is quite challenging. Summary of the Invention

[0004] The technical problem that this invention aims to solve is that existing industrial OCT defect detection technologies face significant challenges in detecting large-sized samples due to the difficulty in hardware implementation and algorithm execution.

[0005] Therefore, the present invention provides a full-field optical coherence tomography three-dimensional industrial defect detection system and method suitable for large-size samples.

[0006] The technical solution adopted by this invention to solve its technical problem is:

[0007] A full-field optical coherence tomography (OCT) three-dimensional industrial defect detection system suitable for large-size samples includes a wavelength-tunable laser, a full-field interferometric imaging system, a three-dimensional displacement platform, and a controller. The full-field interferometric imaging system includes a collimating beam expander system, a polarization beam splitter, a reference arm beam expander lens group, a reference arm reflector, a sample arm beam expander lens group, a sample under test, and an area array camera. The wavelength-tunable laser, the area array camera, and the three-dimensional displacement platform are connected to the controller. The polarization beam splitter is positioned between the reference arm beam expander lens group, the sample arm beam expander lens group, the area array camera, and the collimating beam expander system. The wavelength-tunable laser is positioned on the side of the collimating beam expander system away from the polarization beam splitter. The reference arm reflector is positioned on the side of the reference arm beam expander lens group away from the polarization beam splitter. The sample under test is placed on the side of the sample arm beam expander lens group away from the polarization beam splitter.

[0008] Furthermore, the three-dimensional displacement platform is used to place the sample to be tested, adjust the focal plane position of the sample to be tested in the axial direction, and move the sample to be tested in the lateral direction according to a set trajectory.

[0009] A three-dimensional industrial defect detection method based on the above-described full-field optical coherence tomography three-dimensional industrial defect detection system suitable for large-size samples includes,

[0010] S1. The interference fringes generated by the full-field interferometric imaging system illuminate the sample under test;

[0011] S2. The controller controls the three-dimensional displacement platform to move the sample to be tested, and synchronously controls the wavelength tunable laser to complete wavelength scanning within the exposure time of the area array camera.

[0012] S3. The area array camera transmits the three-dimensional raw data acquired at various locations on the sample to be tested to the controller.

[0013] S4. The controller performs image reconstruction and image stitching algorithm processing on the acquired three-dimensional raw data and finally renders and displays it to detect surface and internal defects of the sample to be tested.

[0014] Furthermore, step S4 specifically includes:

[0015] S4-1, Calibration file for linearly sampling N quasi-monochromatic light wavenumbers of a wavelength-tunable laser output along the depth direction;

[0016] S4-2. Perform a fast inverse Fourier transform on N frames of two-dimensional interferometric images acquired under N quasi-monochromatic illumination along the calibrated wavenumber arrangement direction to obtain the reconstructed image.

[0017] S4-3. Perform image enhancement operations on the reconstructed image and map the pixel values ​​of the image to a grayscale range of 0 to 255;

[0018] S4-4. Stack the N frames of reconstructed images after image enhancement and grayscale mapping along the wavenumber arrangement direction to obtain the structural information of the sample in the depth direction.

[0019] S4-5: The controller controls the three-dimensional displacement stage to move the sample to be tested to different set positions. The M sets of three-dimensional raw data collected from different areas of the sample to be tested are processed by S4-2 to S4-4. Then, the image stitching algorithm is used to stitch the M sets of reconstructed images to obtain the complete structure of the sample to be tested.

[0020] S4-6. Render and display the stitched 3D reconstructed image, and perform surface and internal defect detection.

[0021] Furthermore, in step S4-1, at the very beginning of the full-field sweep frequency OCT image reconstruction algorithm of the present invention, it is necessary to measure the quasi-monochromatic light wavenumber output by the tunable laser and linearly sample it along the depth direction to obtain a calibration file, so that the wavenumber of the interference signal has equally spaced sampling points in the depth direction, so as to accelerate the reconstruction speed using the fast inverse Fourier transform algorithm.

[0022] Furthermore, in step S4-2, before the fast inverse Fourier transform, a high-pass filter is used to filter out the DC background signal and low-frequency self-interference signal, and windowing is performed to shape the interference spectrum in order to reduce the sidelobes of the point spread function after the inverse fast inverse Fourier transform.

[0023] The beneficial effects of this invention are that it integrates full-field (FF) OCT technology using surface scanning and swept-source (SS) laser technology. Based on the principle of FF-SS-OCT, it obtains the three-dimensional structural information of large-sized samples. Using an image stitching algorithm, it merges and stitches together the local reconstruction results of adjacent areas to present a three-dimensional image of the entire sample, enabling the detection of defects on the upper and lower surfaces and internal structures of the sample. This effectively solves the limitations of traditional industrial OCT's point-scanning imaging method, which suffers from a small field of view and high complexity in system hardware and algorithm implementation. It can effectively detect defects on the upper and lower surfaces and internal structures of large-sized transparent / semi-transparent samples. Attached Figure Description

[0024] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0025] Figure 1 This is a schematic diagram of the structure of the full-field optical coherence tomography three-dimensional industrial defect detection system applicable to large-size samples in this invention.

[0026] Figure 2 This is a schematic diagram of the full-field interferometric imaging system in this invention.

[0027] Figure 3 This is a flowchart illustrating the implementation of the full-field optical coherence tomography three-dimensional industrial defect detection method applicable to large-size samples in this invention.

[0028] In the figure: 1. Wavelength-tunable laser; 2. Full-field interferometric imaging system; 2-1. Collimation and beam expansion system; 2-2. Polarization beam splitter; 2-3. Reference arm beam expander lens group; 2-4. Reference arm reflector; 2-5. Sample arm beam expander lens group; 2-6. Sample under test; 2-7. Area array camera; 3. Three-dimensional displacement platform; 4. Controller. Detailed Implementation

[0029] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0030] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, features defined with "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0031] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0032] Full-field (FF) OCT technology using an area scanning method employs an area array camera to capture lateral images of the sample 2-6 under test, acquiring its two-dimensional information. Depth information can be obtained in the time domain by moving a reference arm to perform a longitudinal scan of the coherent signal within the sample 2-6. However, this method is limited in imaging speed due to the mechanical movement of the reference arm. Swept Source (SS) method, on the other hand, uses a wavelength-tunable laser for wavelength scanning. This involves quasi-monochromatic light of different wavelengths incident on the system at different times. Based on the principle of SS OCT, the depth of the sample 2-6 is scanned, ultimately obtaining its three-dimensional structural information. The imaging speed is not limited by the wavelength-tunable laser but only by the industrial area array camera.

[0033] Due to the limitations of the size of industrial area array cameras, the field of view of a single FF-SS-OCT image is still limited and insufficient to cover the entire large-sized sample. Therefore, it is necessary to capture the original three-dimensional data of different regions of the sample under test 2-6 for local reconstruction, and then use an image stitching algorithm to stitch and fuse the local reconstruction results of adjacent regions to present the three-dimensional imaging result of the entire sample under test 2-6, so as to detect the defects on the upper and lower surfaces and internal surfaces of the entire sample under test 2-6.

[0034] In summary, we used full-field frequency sweep optical coherence tomography (FF-SS-OCT) imaging technology combined with image stitching algorithms to perform three-dimensional detection of surface / internal defects in large-size samples.

[0035] Accordingly, this application provides a full-field optical coherence tomography three-dimensional industrial defect detection system suitable for large-size samples, including a wavelength-tunable laser 1, a full-field interferometric imaging system 2, a three-dimensional displacement platform 3, and a controller 4. The full-field interferometric imaging system 2 includes a collimating beam expander system 2-1, a polarization beam splitter 2-2, a reference arm beam expander lens group 2-3, a reference arm reflector 2-4, a sample arm beam expander lens group 2-5, a sample to be tested 2-6, and an area array camera 2-7. The wavelength-tunable laser 1, the area array camera 2-7, the three-dimensional displacement platform 3, and the controller 4 are connected.

[0036] Specifically, the polarization beam splitter 2-2 is positioned between the reference arm beam expander lens group 2-3, the sample arm beam expander lens group 2-5, the area array camera 2-7, and the collimation beam expander system 2-1. The wavelength-tunable laser 1 is positioned on the side of the collimation beam expander system 2-1 away from the polarization beam splitter 2-2. The reference arm reflector 2-4 is positioned on the side of the reference arm beam expander lens group 2-3 away from the polarization beam splitter 2-2. The sample to be tested 2-6 is placed on the side of the sample arm beam expander lens group 2-5 away from the polarization beam splitter 2-2 via a three-dimensional displacement platform 3. The three-dimensional displacement platform 3 can adjust the focal plane position of the sample to be tested axially and move the sample to be tested 2-6 laterally along a set trajectory.

[0037] The laser emitted from the wavelength-tunable laser 1 enters the collimation and beam-expanding system 2-1 through an optical fiber. After collimation and expansion, the laser beam is split into two paths at the polarization beam splitter 2-2: the reference arm containing the transmitted light path and the sample arm containing the reflected light path. The transmitted light is expanded by the reference arm beam-expanding lens group 2-3 and then illuminates the reference arm reflector 2-4. Symmetrically, the reflected light is expanded by the sample arm beam-expanding lens group 2-5 and then illuminates the sample under test 2-6.

[0038] The light reflected from the reference arm mirror 2-4 passes through the reference arm beam expander lens group 2-3 again. Similarly, the light reflected / scattered from the sample 2-6 passes through the sample arm beam expander lens group 2-5 again. The two beams meet and interfere at the polarization beam splitter 2-2, and the resulting interference fringes are captured by the area array camera 2-7. The area array camera 2-7 has a resolution of 4096×3072 and a pixel size of 5.5μm. The reference arm beam expander lens group 2-3 and the sample arm beam expander lens group 2-5 have the same magnification, both being 1.62x. Finally, the single imaging field of view of the full-field interferometric imaging system 2 is approximately 36mm×27mm.

[0039] The controller 4 synchronously controls the wavelength tunable laser 1, the area array camera 2-7 in the full-field interferometric imaging system 2, and the three-dimensional displacement platform 3. When the three-dimensional displacement platform 3 completes the lateral movement of the sample 2-6 to be tested, the wavelength tunable laser 1 completes the scanning of the depth direction of the sample 2-6 to be tested within the exposure time of the area array camera 2-7. The three-dimensional raw data acquired by the area array camera 2-6 is transmitted to the controller for subsequent data processing, including the reconstruction of each set of raw data, the stitching of the reconstructed images, and the three-dimensional display of the stitched images.

[0040] A full-field optical coherence tomography three-dimensional industrial defect detection method suitable for large-size samples includes:

[0041] S1, The interference fringes generated by the full-field interferometric imaging system 2 illuminate the samples under test 2-6.

[0042] S2, Controller 4 controls the three-dimensional displacement platform 3 to move the sample to be tested 2-6, and synchronously controls the wavelength tunable laser 1 to complete wavelength scanning within the exposure time of the area array camera 2-7.

[0043] S3. The area array camera 2-7 transmits the three-dimensional raw data collected at various locations on the sample to be tested 2-6 to the controller 4.

[0044] S4. The controller 4 performs image reconstruction and image stitching algorithm processing on the acquired three-dimensional raw data and finally renders and displays it to detect surface and internal defects of the test samples 2-6.

[0045] In step S4, the process of image reconstruction, image stitching algorithm processing, and final rendering and display of the 3D raw data is as follows:

[0046] S4-1, Wavenumber Linear Sampling:

[0047] According to the SS-OCT imaging principle, to obtain information about the depth direction of the sample under test (2-6), it is necessary to perform an inverse Fourier transform on the three-dimensional raw data along the scanning direction of the wavelength (assuming the number of sampling points is N, N≥0 and is an integer power of 2, such as 1024). This invention chooses to use the fast inverse Fourier transform algorithm to speed up the processing. The algorithm requires that the wavenumber of the interference signal (i.e., the N frames of two-dimensional interference images acquired under the illumination of N quasi-monochromatic wavelengths) has an equally spaced feature in the depth direction of the sampling points. Therefore, at the beginning of the FF-SS-OCT image reconstruction algorithm of this invention, it is necessary to measure the calibration file of the N quasi-monochromatic light wavenumbers output by the tunable laser linearly sampled along the depth direction.

[0048] S4-2. Perform fast inverse Fourier transform on N frames of two-dimensional interferometric images along the calibrated wavenumber arrangement direction. In order to obtain a clean background signal in the reconstructed image, use a high-pass filter to filter out the DC background signal and low-frequency self-interference signal before fast inverse Fourier transform, and perform windowing to shape the interference spectrum to reduce the sidelobes of the point spread function after inverse fast inverse Fourier transform.

[0049] S4-3. Perform image enhancement operations on the N-frame reconstructed images after inverse fast Fourier transform and filtering, and map the pixel values ​​of the images to a grayscale range of 0 to 255 to facilitate subsequent image processing and display.

[0050] S4-4. Stack the N frames of reconstructed images after image enhancement and grayscale mapping along the wavenumber arrangement direction to obtain the structural information of the sample under test in the depth direction of 2-6.

[0051] S4-5, Controller 4 controls the three-dimensional displacement stage 3 to move the sample 2-6 to different set positions. The above operation is performed on the M sets of three-dimensional raw data collected from different areas of the sample 2-6. Then, the image stitching algorithm is used to stitch the M sets of reconstructed images to obtain the complete structure of the sample 2-6.

[0052] S4-6. Render and display the stitched 3D reconstructed image, and perform surface and internal defect detection.

[0053] Based on the above-described preferred embodiments of the present invention, and through the foregoing description, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification, but must be determined by the scope of the claims.

Claims

1. A full-field optical coherence tomography three-dimensional industrial defect detection system suitable for large-size samples, characterized in that, The system includes a wavelength-tunable laser (1), a full-field interferometric imaging system (2), a three-dimensional displacement platform (3), and a controller (4). The full-field interferometric imaging system (2) includes a collimating beam expander (2-1), a polarization beam splitter (2-2), a reference arm beam expander lens group (2-3), a reference arm reflector (2-4), a sample arm beam expander lens group (2-5), a sample to be tested (2-6), and an area array camera (2-7). The polarization beam splitter is located in the reference arm beam expander lens group (2-7). 3) Between the sample arm beam expander lens group (2-5), the area array camera (2-7), and the collimating beam expander system (2-1), the wavelength tunable laser (1) is located on the side of the collimating beam expander system (2-1) away from the polarization beam splitter (2-2), the reference arm reflector (2-4) is located on the side of the reference arm beam expander lens group (2-3) away from the polarization beam splitter, and the sample to be tested (2-6) is placed on the side of the sample arm beam expander lens group (2-5) away from the polarization beam splitter.

2. The full-field optical coherence tomography three-dimensional industrial defect detection system suitable for large-size samples according to claim 1, characterized in that, The wavelength-tunable laser (1), the area array camera (2-7), the three-dimensional displacement platform (3) are connected to the controller (4). The three-dimensional displacement platform (3) is used to place the sample to be tested (2-6), and to adjust the focal plane position of the sample to be tested in the axial direction and move the sample to be tested (2-6) in the lateral direction according to the set trajectory.

3. A three-dimensional industrial defect detection method based on the full-field optical coherence tomography three-dimensional industrial defect detection system suitable for large-size samples as described in claim 1 or 2, characterized in that, include, S1. The interference fringes generated by the full-field interferometric imaging system (2) illuminate the sample under test (2-6); S2, the controller (4) controls the three-dimensional displacement platform (3) to move the sample to be tested (2-6), and synchronously controls the wavelength tunable laser (1) to complete the wavelength scan within the exposure time of the area array camera (2-7); S3. The area array camera (2-7) transmits the three-dimensional raw data collected at various positions of the sample to be tested (2-6) to the controller (4); S4. The controller (4) performs image reconstruction and image stitching algorithm processing on the acquired three-dimensional raw data and finally renders and displays it to detect surface and internal defects of the sample to be tested (2-6).

4. The three-dimensional industrial defect detection method according to claim 3, characterized in that, Step S4 specifically includes: S4-1, Calibration file for linearly sampling N quasi-monochromatic light wavenumbers output by a wavelength-tunable laser (1) along the depth direction; S4-2. Perform a fast inverse Fourier transform on N frames of two-dimensional interferometric images acquired under N quasi-monochromatic illumination along the calibrated wavenumber arrangement direction to obtain the reconstructed image. S4-3. Perform image enhancement operations on the N frames of reconstructed images and map the pixel values ​​of the images to a grayscale range of 0 to 255. S4-4. Stack the N frames of reconstructed images after image enhancement and grayscale mapping along the wavenumber arrangement direction to obtain the structural information of the sample (2-6) in the depth direction of the test sample. S4-5, Controller (4) controls the three-dimensional displacement stage (3) to move the sample to be tested (2-6) to different set positions. The M sets of three-dimensional raw data collected from different areas of the sample to be tested (2-6) are all operated from S4-2 to S4-4. Then, the image stitching algorithm is used to stitch the M sets of reconstructed images to obtain the complete structure of the sample to be tested (2-6). S4-6. Render and display the stitched 3D reconstructed image, and perform surface and internal defect detection.

5. The three-dimensional industrial defect detection method according to claim 3, characterized in that, In step S4-1, at the very beginning of the full-field sweep frequency OCT image reconstruction algorithm of the present invention, it is necessary to measure the calibration file of the quasi-monochromatic light wavenumber output by the tunable laser linearly sampled along the depth direction, so that the wavenumber of the interference signal has equally spaced sampling points in the depth direction, so as to accelerate the reconstruction speed using the fast inverse Fourier transform algorithm.

6. The three-dimensional industrial defect detection method according to claim 3, characterized in that, In step S4-2, before the fast inverse Fourier transform, a high-pass filter is used to filter out the DC background signal and low-frequency self-interference signal, and windowing is performed to shape the interference spectrum to reduce the side lobes of the point spread function after the inverse fast inverse Fourier transform.

Images