A precise stitching method for a small number of discrete defect images

Through the synchronization and in-situ imaging of scattering imaging and fluorescence imaging dual-channel technology, the large field of full-aperture image scattering is used to splice images with large field of view, which solves the problem of poor splicing accuracy of small discrete defects of optical components after hydrofluoric acid treatment, and realizes the Fast and high-precision subsurface defect detection.

CN113720850BActive Publication Date: 2024-11-05LASER FUSION RES CENT CHINA ACAD OF ENG PHYSICS
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202110822801.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-07-21
Publication Date
2024-11-05
Estimated Expiration
2041-07-21

AI Technical Summary

Technical Problem

The prior art is difficult to accurately splice a small number of discrete subsurface defect images of optical components after hydrofluoric acid treatment, resulting in poor positioning accuracy and large detection errors, especially in the process of large-area scanning imaging.

Method used

Using dual-channel technology for scattering imaging and fluorescence imaging, through synchronous and in-situ imaging, the feature points of the scattered image are scattered for large field of view full-aperture images, combining feature point matching and overlapping area styling methods to reduce errors and record splicing rules, assisting Fluorescent image stitching.

Benefits of technology

It realizes rapid and high-precision screening of subsurface defects of optical components, reduces splicing errors, improves detection efficiency and accuracy, and is especially suitable for the detection of subsurface defects of large-size optical components.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113720850B_ABST
    Figure CN113720850B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for accurately stitching a small number of discrete defect images, including constructing a defect detection device; selecting a sample of an optical element, finding a point that can excite fluorescence as a marking point; mounting the sample on a sample stage, dividing the surface into several imaging areas, using a defect detection device to scan and image the sample to obtain a sub-aperture scattering image and a sub-aperture fluorescence image of each imaging area; and then stitching the images. The present invention uses a dual-channel technology of scattering imaging and fluorescence imaging to achieve synchronous and in-situ imaging of scattering signals and fluorescence signals, making full use of the characteristics of the scattering image with more feature points to perform large-field full-aperture image stitching, solving the problem of inaccurate positioning, and improving detection efficiency, especially for sub-surface defect detection of large-size optical elements, which can synchronize image acquisition and image processing, further saving time and cost, and realizing rapid and high-precision screening and characterization of surface and sub-surface defects of fused quartz components.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to an image splicing method, in particular to an accurate splicing method for a small number of discrete defective images. Background Art

[0002] In order to obtain maximum output, large high-power / high-energy laser devices are operated at a flux close to the damage threshold of optical components. Therefore, the damage performance of optical components is particularly important and is the key to determining the output capacity of such laser devices. Subsurface defects of fused quartz components, which are distributed at a depth of submicron to tens of microns below the polished redeposition layer and cannot be detected by conventional detection methods, are the root cause of UV damage to fused quartz components. In order to improve the laser damage resistance of optical components, a common method is to reprocess the surface of optical components, that is, to use hydrofluoric acid to remove the defect-rich subsurface layer. The damage performance of the treated optical components is greatly improved, but there may still be a small number of discrete subsurface defects that cause laser damage.

[0003] Nondestructive testing technology based on the principle of fluorescence imaging can achieve large-area, high-resolution, and fast imaging detection of fused quartz sub-surface defects. However, there are fewer sub-surface defects in optical components after hydrofluoric acid treatment, so there are fewer defects in the images obtained by fluorescence imaging, and there are not enough feature points to splice into a full-aperture image. Large-area scanning imaging requires repeated movement of the electric translation stage, and there will be return errors in the entire process, resulting in poor defect positioning accuracy when splicing the overlapping areas set according to the scan. Summary of the invention

[0004] The purpose of the present invention is to provide a solution to the above-mentioned problem, which adopts the dual-channel technology of scattering imaging and fluorescence imaging to realize the synchronous and in-situ imaging of scattering signals and fluorescence signals, and can make full use of the characteristics of the scattering image with more feature points to perform large-field-of-view full-aperture image stitching, a method for accurately stitching a small number of discrete defective images.

[0005] In order to achieve the above-mentioned object, the technical solution adopted by the present invention is as follows: a method for accurately stitching a small number of discrete defective images, comprising the following steps;

[0006] (1) constructing a defect detection device, wherein the defect detection device includes a sample stage, a laser, and an image acquisition unit;

[0007] The sample stage is arranged horizontally, and is used to place the sample and drive its three-dimensional movement;

[0008] The laser emits laser light, which is incident on the sample surface and subsurface from obliquely above after passing through the optical path;

[0009] The image acquisition unit is located directly above the sample stage and is used to perform scattering imaging and fluorescence imaging on the sample surface to obtain a scattering image and a fluorescence image;

[0010] (2) Select a sample of an optical element and find a point that can excite fluorescence as a marking point. The surface where the marking point is located is the front side of the sample;

[0011] (3) Mount the sample on the sample stage, adjust the laser and image acquisition unit, excite fluorescence at the marked point, and make the scattering image and fluorescence image at the marked point clear;

[0012] (4) Dividing the sample surface into several imaging areas, with adjacent imaging areas partially overlapping, moving the sample, exciting and imaging each imaging area along an S-shaped path, sequentially obtaining a scattering image and a fluorescence image of each imaging area, and using the scattering image and the fluorescence image of each imaging area as the sub-aperture scattering image and the sub-aperture fluorescence image of the area;

[0013] (5) Sub-aperture scattering image stitching;

[0014] Each sub-aperture scattering image is first stitched using the feature point matching method. If the feature points of the sub-aperture scattering image do not match, the overlapped area image stitching method is used to stitch it, and finally the full-aperture scattering image is obtained, and its stitching rules are saved;

[0015] (6) According to the stitching rules obtained in (5), the sub-aperture fluorescence images are stitched to obtain a full-aperture fluorescence image.

[0016] Preferably, the sample stage is a high-precision three-dimensional electric mobile platform;

[0017] The image acquisition unit includes a microscope imaging lens, a first high-pass filter, a second high-pass filter and a single-photon CCD, which are arranged in sequence from bottom to top. The first high-pass filter is arranged at an angle of 45° to the horizontal plane and a UV-enhanced CCD is arranged on one side. The second high-pass filter is arranged horizontally. The microscope imaging lens is used to collect ultraviolet rays scattered from the sample surface and is divided into two paths. One path is scattered imaging in the UV-enhanced CCD through the first high-pass filter, and the other path is fluorescent imaging in the single-photon CCD through the second high-pass filter.

[0018] Preferably, the oblique incident laser is a 355nm continuous laser, and is incident on the front of the sample after being shaped and focused, and the first high-pass filter and the second high-pass filter are 375nm high-pass filters.

[0019] The sample is fused quartz.

[0020] In the present invention, the excitation light source generated by the laser is a single-wavelength 355nm continuous laser. In fact, other wavelengths can be selected according to the characteristics of material defects. The maximum output power of the laser is 500mw. The sub-surface defects of the component are easily excited by the laser to produce fluorescence. When the laser beam is incident on the component surface at a specific angle, the fluorescence excited by the laser and the ultraviolet light scattered by the component surface are collected and converged by the microscope imaging lens. Among them, the scattered ultraviolet light passes through a first high-pass filter set at an angle of 45° to the horizontal plane and is collected by the ultraviolet enhanced CCD, while the fluorescence first passes through a first high-pass filter at 45°, then passes through a second high-pass filter set horizontally, and is finally collected by the single-photon CCD, and fluorescence imaging and scattering imaging are performed simultaneously.

[0021] The present invention adopts dual-channel fusion detection of scattering image and fluorescence image, which can realize synchronous and in-situ imaging of scattering and fluorescence.

[0022] When performing image stitching, all sub-aperture scattering images are first stitched according to the feature point matching method. If there are many feature points, the match will be successful. If there are too few feature points, there will be a mismatch. At this time, the overlap area image stitching method is used for stitching. Each time stitching is performed, the stitching path is recorded, and finally the stitched full-aperture scattering image and its stitching rules are obtained.

[0023] The sub-aperture fluorescence images are then stitched together. When stitching together, the stitching rules of the scattering images are adopted. Since each imaging area corresponds to a sub-aperture scattering image and a sub-aperture fluorescence image, the sub-aperture scattering image and the sub-aperture fluorescence image are also one-to-one corresponding in each imaging area. In this way, if the sub-aperture scattering images are stitched together using the feature point matching method, the corresponding sub-aperture fluorescence images are also stitched together using the feature point matching method. If the sub-aperture scattering images are stitched together using the overlapped area image stitching method, the corresponding sub-aperture fluorescence images are also stitched together using the overlapped area image stitching method. Finally, the full-aperture fluorescence image is obtained according to the stitching rules of the full-aperture scattering image.

[0024] This is because: when the sample stage is performing three-dimensional displacement, there will be certain errors, and the accumulated error after scanning is relatively large. Therefore, directly using the overlapping area rule stitching will result in relatively large errors; therefore, for sub-aperture scattering images with more defect points, that is, more feature points, feature point stitching is used, and for sub-aperture scattering images with fewer feature points or no feature points, they are stitched according to the scanned overlapping area image stitching method; in this way, the stitching position of each sub-aperture scattering image is determined, which can effectively reduce errors.

[0025] Compared with the prior art, the advantages of the present invention are:

[0026] The present invention adopts the dual-channel technology of scattering imaging and fluorescence imaging to realize synchronous and in-situ imaging of scattering signals and fluorescence signals. It can make full use of the characteristics of the scattering image with more feature points to perform large-field-of-view full-aperture image stitching, reduce the errors occurring in the stitching process, and record its stitching rules, paths, etc., and use the stitching rules to assist fluorescence image stitching, thereby obtaining a precisely stitched fluorescence image.

[0027] This method not only solves the problem of inaccurate positioning, but also greatly improves detection, especially the detection of sub-surface defects of large-size optical components. It can synchronize image acquisition and image processing, further save time costs, and realize rapid and high-precision screening and characterization of surface and sub-surface defects of fused quartz components.

[0028] The present invention can quickly and accurately stitch sub-surface defect images with discrete signals, which is extremely important for high-precision positioning, rapid, high-resolution and non-destructive detection of sub-surface defects of optical components. The present invention can be used not only for high-precision stitching of sub-surface defect images of large-caliber fused quartz components after hydrofluoric acid treatment, but also for stitching similar images with a small amount of signals. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a structural diagram of the defect detection device in the present invention;

[0030] Figure 2 The sub-aperture scattering image obtained by the UV-enhanced CCD obtained in step (4);

[0031] Figure 3 The sub-aperture fluorescence image obtained by the single-photon CCD obtained in step (4);

[0032] Figure 4 is the full aperture scattering image;

[0033] Figure 5 is a full-aperture fluorescence image;

[0034] Figure 6 This is the flow chart for stitching the full-aperture scattering image and the full-aperture fluorescence image;

[0035] Figure 7 is a full-aperture fluorescence image obtained by the prior art method;

[0036] Figure 8 This is a full-aperture fluorescence image obtained by the method of the present invention.

[0037] In the figure: 1. Single-photon CCD; 2. Second high-pass filter; 3. First high-pass filter; 4. UV-enhanced CCD; 5. Microscope imaging lens; 6. Oblique incident laser; 7. Scattered light; 8. Sample stage; 9. Translation stage. DETAILED DESCRIPTION

[0038] The present invention will be further described below in conjunction with the accompanying drawings.

[0039] Example 1: See Figure 1-Figure 6 ,A method for accurately stitching a small number of discrete ,defective images, including the following steps;

[0040] (1) constructing a defect detection device, wherein the defect detection device includes a sample stage 8, a laser, and an image acquisition unit;

[0041] The sample stage 8 is arranged horizontally, and is used to place the sample and drive its three-dimensional movement;

[0042] The laser emits laser light, which is incident on the sample surface and subsurface from obliquely above after passing through the optical path. We call the laser light obliquely incident through the optical path obliquely incident laser light 6;

[0043] The image acquisition unit is located directly above the sample stage 8 and is used to perform scattering imaging and fluorescence imaging on the sample surface to obtain a scattering image and a fluorescence image;

[0044] (2) Select a sample of an optical element and find a point that can excite fluorescence as a marking point. The surface where the marking point is located is the front side of the sample;

[0045] (3) Mount the sample on the sample stage 8, adjust the laser and the image acquisition unit, excite fluorescence at the marked point, and make the scattering image and the fluorescence image at the marked point clear;

[0046] (4) Dividing the sample surface into several imaging areas, with adjacent imaging areas partially overlapping, moving the sample, exciting and imaging each imaging area along an S-shaped path, sequentially obtaining a scattering image and a fluorescence image of each imaging area, and using the scattering image and the fluorescence image of each imaging area as the sub-aperture scattering image and the sub-aperture fluorescence image of the area;

[0047] (5) Sub-aperture scattering image stitching;

[0048] Each sub-aperture scattering image is first stitched using the feature point matching method. If the feature points of the sub-aperture scattering image do not match, the overlapped area image stitching method is used to stitch it, and finally the full-aperture scattering image is obtained, and its stitching rules are saved;

[0049] (6) According to the stitching rules obtained in (5), the sub-aperture fluorescence images are stitched to obtain a full-aperture fluorescence image.

[0050] In the defect detection device, the laser emits a laser, which is reflected and refracted by the optical path to the sample surface. The surface and sub-surface defects of the sample are easily excited by the laser to produce fluorescence, and ultraviolet light is also scattered on the sample surface. We use fluorescence and ultraviolet light to simultaneously form scattering images and fluorescence images respectively. Taking advantage of the fact that there are many feature points in the scattering image, we perform large-field full-aperture image stitching and obtain the stitching rules for stitching the scattering image into the full-aperture image. This stitching rule is used to assist the stitching of the fluorescence image, and then a precisely stitched fluorescence image is obtained.

[0051] Example 2: See Figure 1-Figure 6 , this embodiment 2 is further improved on the basis of embodiment 1, wherein the sample stage 8 is a high-precision three-dimensional electric mobile platform;

[0052] The image acquisition unit includes a microscope imaging lens 5, a first high-pass filter 3, a second high-pass filter 2 and a single-photon CCD1, which are arranged in sequence from bottom to top. The first high-pass filter 3 is arranged at a 45° angle to the horizontal plane and a UV-enhanced CCD4 is arranged on one side. The second high-pass filter 2 is arranged horizontally. The microscope imaging lens 5 is used to collect ultraviolet rays scattered from the sample surface and is divided into two paths. One path is scattered and imaged in the UV-enhanced CCD4 through the first high-pass filter 3, and the other path is fluorescently imaged in the single-photon CCD1 through the second high-pass filter 2.

[0053] The oblique incident laser 6 is a 355nm continuous laser, and is incident on the front of the sample after being shaped and focused. The first high-pass filter 3 and the second high-pass filter 2 are 375nm high-pass filters.

[0054] In the defect detection device of the present invention, when the laser beam is incident obliquely on the sample surface at a specific angle, scattered ultraviolet light and fluorescence are generated, which we collectively call scattered light 7, wherein the ultraviolet light is imaged in the ultraviolet enhanced CCD4 to form a scattered image, which is used to form a sub-aperture scattered image and subsequently spliced, and the fluorescence is imaged in the single-photon CCD1 to form a fluorescence image, which is used to form a sub-aperture fluorescence image and subsequently spliced. In this embodiment, the sample of the optical element is fused quartz.

[0055] Specifically: the laser generates a single-wavelength 355nm continuous laser, which is incident on the component surface at an angle. The fluorescence excited by the laser and the ultraviolet light scattered by the component surface are collected and converged by the microscope imaging lens 5, wherein the scattered ultraviolet light passes through a first high-pass filter 3 set at an angle of 45° to the horizontal plane, and is collected by the ultraviolet enhanced CCD4 for scattered imaging, while the fluorescence first passes through a first high-pass filter 3 at 45°, and then passes through a second high-pass filter 2 set horizontally, and is finally collected by a single-photon CCD1 for fluorescence imaging.

[0056] In this embodiment, the excitation light source generated by the laser is a single-wavelength 355nm continuous laser. In fact, other wavelengths can be selected according to the characteristics of material defects. The maximum output power of the laser is 500mw.

[0057] Example 3: See Figures 1 to 8 In order to better illustrate the effect of the present invention, we further improved on the basis of Example 1.

[0058] Steps (1) to (6) are the same as in Example 1, and a full-aperture scattering image and a full-aperture fluorescence image are obtained, such as Figure 4 , Figure 5 shown.

[0059] (7) A computer is used to extract defect information from the full-aperture scattering image and the full-aperture fluorescence image, respectively. The defect information includes the location, size, intensity, etc. of the defect. Since the full-aperture image is a grayscale image, the intensity refers to the grayscale value, that is, a value of 0-255. The larger the value, the greater the brightness, the higher the intensity, and the more obvious the defect at that location.

[0060] (8) After step (7), the defect blocks in the two full-aperture images can be marked. The defect blocks in the full-aperture scattering image and the full-aperture fluorescence image are found respectively. Then, all the defect blocks in the two full-aperture images are compared. We use the center position of the defect block as the target position to determine whether the positions of the scattering defect and the fluorescence defect in the two full-aperture images are consistent.

[0061] (9) If there are defective blocks at the same position in the two full-aperture images, they are treated as surface fluorescence defects and sub-surface fluorescence defects of the sample, screened out, marked separately, and the information is stored and output.

[0062] In order to illustrate the effect of the method of the present invention, we refer to Figure 7 and Figure 8 , Figure 7 This is a full-aperture fluorescence image obtained by the prior art method; it can be seen that in the middle right side of the image, there should be one scratch, but after stitching, it appears as two. Figure 8 This is a full-aperture fluorescence image obtained by the method of the present invention. It can be seen that in the middle right side of the image, a complete scratch is connected.

[0063] 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 method for accurately stitching a small number of discrete defect images, characterized by: The steps include: (1) constructing a defect detection device, wherein the defect detection device includes a sample stage, a laser, and an image acquisition unit; The sample stage is horizontally arranged and used to place the sample and drive it to move in three dimensions, and the sample is fused quartz; The laser emits laser light, which is incident on the sample surface and subsurface from obliquely above after passing through the optical path; The image acquisition unit is located directly above the sample stage and is used to perform scattering imaging and fluorescence imaging on the sample surface to obtain a scattering image and a fluorescence image; (2) Select a sample of an optical element and find a point that can excite fluorescence as a marking point. The surface where the marking point is located is the front side of the sample; (3) Mount the sample on the sample stage, adjust the laser and image acquisition unit, excite fluorescence at the marked point, and make the scattering image and fluorescence image at the marked point clear; (4) Dividing the sample surface into several imaging areas, with adjacent imaging areas partially overlapping, moving the sample, exciting and imaging each imaging area along an S-shaped path, sequentially obtaining a scattering image and a fluorescence image of each imaging area, and using the scattering image and the fluorescence image of each imaging area as the sub-aperture scattering image and the sub-aperture fluorescence image of the area; (5) Sub-aperture scattering image stitching; Each sub-aperture scattering image is first stitched using the feature point matching method. If the feature points of the sub-aperture scattering image do not match, the overlapped area image stitching method is used to stitch it, and finally the full-aperture scattering image is obtained, and its stitching rules are saved; (6) According to the stitching rules obtained in (5), the sub-aperture fluorescence images are stitched to obtain a full-aperture fluorescence image.

2. The method for accurately stitching a small number of discrete defective images according to claim 1, characterized in that: The sample stage is a high-precision three-dimensional electric mobile platform; The image acquisition unit includes a microscope imaging lens, a first high-pass filter, a second high-pass filter and a single-photon CCD, which are arranged in sequence from bottom to top. The first high-pass filter is arranged at an angle of 45° to the horizontal plane and a UV-enhanced CCD is arranged on one side. The second high-pass filter is arranged horizontally. The microscope imaging lens is used to collect ultraviolet rays scattered from the sample surface and divide it into two paths. One path is scattered imaging in the UV-enhanced CCD through the first high-pass filter, and the other path is fluorescent imaging in the single-photon CCD through the second high-pass filter.

3. The method for accurately stitching a small number of discrete defective images according to claim 2, characterized in that: The oblique incident laser is a 355nm continuous laser, and is incident on the front of the sample after being shaped and focused. The first high-pass filter and the second high-pass filter are 375nm high-pass filters.