An optical element surface micro-nano level target point efficient detection method and system based on optical-AFM fusion

Abstract

The application provides an optical element surface micro-nano level target point efficient detection method and system based on optical-AFM fusion, relates to the micro-nano manufacturing technical field, and aims at solving the problem of low AFM detection efficiency in the prior art, which cannot meet the detection requirements of a large number of micro-nano defects on the surface of a large aperture optical element. The method comprises the following steps: step one, obtaining a full aperture image of the optical element through an optical microscope; step two, obtaining position information, morphology information and shape information of the target point, classifying defect types, whether the target point is a contaminant and a size range of the target point; step three, constructing a target point classification model based on a convolutional neural network, realizing classification of defect types in different size ranges and identification of contaminants; step four, screening target points needing AFM ultra-precision detection according to classification accuracy and AFM detection time consumption; and step five, planning an AFM scanning path. The application realizes the improvement of the ultra-precision detection efficiency of micro-nano level target points.

Description

Technical Field

[0001] This invention relates to the field of micro-nano manufacturing technology, and more specifically, to a method and system for efficient detection of micro-nano-level target points on the surface of optical components based on optical-AFM fusion. Background Technology

[0002] Fusion energy is currently recognized as the most ideal clean energy source and a key pathway to solving the future energy crisis. Inertial confinement fusion technology, driven by high-powered lasers, is the mainstream fusion ignition method at present. Currently, giant laser devices are equipped with numerous high-quality, large-aperture optical components to achieve the transmission, focusing, and conversion of high-powered laser energy, such as wedge lenses and plasma-optical switches. The raw materials for these optical components mainly include fused silica and potassium dihydrogen phosphate crystals, which are hard-brittle or soft-brittle and difficult to process. Due to current manufacturing limitations, micro-nano-level surface damage remains on the surfaces of ultra-precision machined optical components, including scratches, pits, and bumps. Furthermore, the surfaces of optical components may adsorb a large amount of contaminants during production and use. When subjected to high-powered laser irradiation, these defects will violently absorb laser energy, causing laser damage. The size of these damage points will then rapidly increase with the number of laser irradiations until the component is unusable. Currently, laser-induced damage on the surface of optical components has become a bottleneck limiting the output capacity of giant laser devices. However, the micro-nano-scale target points (defects, contaminants, etc.) on the surface of large-aperture optical components are not only complex in shape but also numerous, posing a huge challenge to their efficient and rapid ultra-precision detection and repair.

[0003] The Lawrence Livermore National Laboratory (LLNL) in the United States has significantly improved the laser damage resistance of optical components by employing a "circular strategy" to monitor defects larger than 50 μm on the surface of optical components online and perform timely micro-repair treatments. This has enabled the National Inertial Confinement (NIF) device to stably output 1.8 MJ, 500 TW of laser energy, and even output 2.05 MJ of ultraviolet laser energy to generate 3.15 MJ of fusion energy for laboratory fusion ignition, representing the highest achievement in laser-driven inertial confinement fusion technology to date. However, to achieve a stable and controllable fusion reaction, LLNL has proposed an ignition target of "3.0 MJ," far exceeding the current output capacity of NIF. At this level, micro- and nano-defects with dimensions of 0.5–50 μm on the surface of optical components will also severely affect their laser damage resistance. However, due to the limitations of the optical diffraction limit, efficient detection and feature identification of micro- and nano-defects with dimensions of 0.5–50 μm using only high-resolution optical microscopy presents a significant challenge.

[0004] The detection accuracy of optical components based on scanning and imaging is limited by the optical diffraction limit. Atomic force microscopy (AFM) exhibits extremely high detection accuracy at nanoscale targets, accurately obtaining detailed feature information of nanoscale defects on the surface of optical components, providing a technical means for ultra-precision detection. However, AFM has extremely low detection efficiency, and the detection of a single target point is time-consuming. In addition, large-aperture optical components are large in size, with a large number of micro- and nano-scale target points on their surface, which are randomly distributed. It is difficult to accurately locate the target point using the field of view of the AFM lens. Therefore, if all micro- and nano-scale target points on the surface of optical components are detected by AFM, the detection efficiency will be extremely low, making it unsuitable for efficient and ultra-precision detection of micro- and nano-scale defects on the surface of large-aperture optical components. Summary of the Invention

[0005] The technical problem to be solved by this invention is:

[0006] The existing detection technology, AFM, has too low detection efficiency, making it difficult to meet the detection needs of the large number of micro-nano defects on the surface of large-aperture optical components.

[0007] The technical solution adopted by the present invention to solve the above-mentioned technical problems is as follows:

[0008] This invention provides a method for efficient detection of micro-nano-scale target points on the surface of optical components based on optical-AFM fusion, comprising the following steps:

[0009] Step 1: Obtain sub-aperture images of the optical element surface using an optical microscope, and stitch the sub-aperture images together to obtain a full-aperture image;

[0010] Step 2: Extract the minimum bounding rectangle of the target point based on the gradient of the full-aperture image to obtain the position information, morphological information and shape information of the target point in the full-aperture image, and classify the defect type, whether it is a contaminant and the size range of the target point.

[0011] Step 3: Based on the location, morphology, and shape information of the target point, as well as the corresponding defect type, whether it is a contaminant, and size range classification information, construct a target point classification model based on a convolutional neural network to classify defect types of different size ranges and identify contaminants.

[0012] Step 4: Set the size threshold of the AFM detection target point based on the classification accuracy and AFM detection time, and screen the target points that need to be subjected to AFM ultra-precision detection.

[0013] Step 5: Based on the selected AFM ultra-precision detection target point location information, plan the AFM scanning path and perform AFM ultra-precision scanning detection on the corresponding target points.

[0014] Furthermore, the optical microscope described in step one has a resolution of 5120×5120, a zoom lens, and is equipped with a 4.5x eyepiece and a 2x objective lens to acquire sub-aperture images of the surface of the optical elements.

[0015] Furthermore, in step one, a sub-aperture image of the surface of the optical element is obtained using an optical microscope, and overlapping areas are set in the sub-aperture images to ensure effective image stitching.

[0016] Further, in step one, the sub-aperture images are stitched together. An adaptive threshold segmentation method is used to perform threshold segmentation on the sub-aperture images. A threshold is calculated for each pixel in the image to avoid losing feature points due to lighting conditions and background. The SURF algorithm is used to extract feature points from the threshold segmentation results. The feature points are then detected, located, oriented, and described. The FLANN algorithm is used to calculate the matching rate of each feature point and perform feature point matching. Based on this, the feature matching pairs of individual feature points are restricted, and the RANSAC algorithm is used to remove unreliable matching points to improve the feature point matching rate. Finally, the sub-aperture images are stitched together.

[0017] Furthermore, step two, which involves extracting the minimum bounding rectangle of the target point based on the gradient of the full-aperture image, includes four steps: edge extraction, morphological processing, calculation of the minimum bounding rectangle, and contour clustering. In edge extraction, the location with the larger image gradient is determined as the edge location of the target point. Then, morphological processing is performed on the edge breakpoints and false detection phenomena that occur during edge extraction. Opening operations are used to remove false detections, and closing operations are used to connect the edges to obtain the complete contour. Next, the minimum bounding rectangle of the obtained complete contour is extracted. Finally, contour clustering is used to group multiple contours with close distances into one class to obtain the minimum bounding rectangle of the final detected contour.

[0018] Furthermore, the morphological information mentioned in step two refers to the angle between the smallest bounding rectangle of the target point and the horizontal direction, and the shape information refers to the aspect ratio of the bounding rectangle of the target point.

[0019] Furthermore, the target point classification model based on convolutional neural networks described in step three uses EfficientNet-B4 as the backbone network and employs transfer learning to train the model.

[0020] Furthermore, in step four, when detecting micro-nano-scale target points of 0.5–50 μm on the surface of optical components, the size threshold of the AFM detection target points is set to 20 μm, that is, AFM ultra-precision detection is performed on target points with a size range of no more than 20 μm and which are not contaminants.

[0021] Furthermore, in step five, the ant colony algorithm is used to plan the AFM scanning path. To accommodate any AFM scanning starting point, the number of ants selected is greater than the number of target points. Then, the ants traverse all the selected AFM ultra-precise detection target points, iterating step by step until the optimal path length tends to stabilize, thus obtaining the globally optimal AFM scanning path.

[0022] A high-efficiency detection system for micro-nano-level target points on the surface of optical components based on optical-AFM fusion is provided. The system has a program module corresponding to the steps of any of the above-mentioned technical solutions, and executes the steps in the above-mentioned high-efficiency detection method for micro-nano-level target points on the surface of optical components based on optical-AFM fusion when running.

[0023] Compared with the prior art, the beneficial effects of the present invention are:

[0024] This invention discloses a highly efficient method and system for detecting micro-nano-level target points on the surface of optical components based on optical-AFM fusion. It integrates optical detection and AFM detection through a dual-station system. A classification strategy is designed based on the characteristics of micro-nano-level target points on the optical component surface to categorize these points, eliminating target points that can be precisely detected optically and contaminants, thereby improving the efficiency of ultra-precision detection of micro-nano-level target points. Simultaneously, it provides the location information of micro-nano-level target points on the optical component surface and plans the path for AFM ultra-precision detection, further improving AFM detection efficiency. Attached Figure Description

[0025] Figure 1 This is a flowchart of an efficient method for detecting micro-nano-level target points on the surface of optical elements based on optical-AFM fusion in an embodiment of the present invention.

[0026] Figure 2 This is a schematic diagram of the optical microscope scanning and imaging trajectory in an embodiment of the present invention;

[0027] Figure 3 This is a 9×9 full-aperture image stitching result diagram in an embodiment of the present invention;

[0028] Figure 4 This is a diagram illustrating the process of extracting the minimum area bounding rectangle of the target point in an embodiment of the present invention.

[0029] Figure 5 This is an image showing the result of extracting multiple feature information of the target point in an embodiment of the present invention;

[0030] Figure 6 This is a diagram illustrating the full-aperture image target point detection, feature information extraction, and storage in an embodiment of the present invention.

[0031] Figure 7 This is a diagram showing the AFM ultra-precision detection results in an embodiment of the present invention;

[0032] Figure 8 This is a comparison chart of the AFM detection path optimization results in the embodiments of the present invention. Detailed Implementation

[0033] In the description of this invention, it should be noted that the terms "first," "second," and "third" mentioned in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first," "second," and "third" may explicitly or implicitly include one or more of that feature.

[0034] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0035] Specific Implementation Plan 1: (e.g.) Figure 1 As shown, this invention provides a highly efficient method for detecting micro-nano-level target points on the surface of optical components based on optical-AFM fusion, comprising the following steps:

[0036] Step 1: Obtain sub-aperture images of the optical element surface using an optical microscope, and stitch the sub-aperture images together to obtain a full-aperture image;

[0037] Step 2: Extract the minimum bounding rectangle of the target point based on the gradient of the full-aperture image to obtain the position information, morphological information and shape information of the target point in the full-aperture image, and classify the defect type, whether it is a contaminant and the size range of the target point.

[0038] Step 3: Based on the location, morphology, and shape information of the target point, as well as the corresponding defect type, whether it is a contaminant, and size range classification information, construct a target point classification model based on a convolutional neural network to classify defect types of different size ranges and identify contaminants.

[0039] Step 4: Set the size threshold of the AFM detection target point based on the classification accuracy and AFM detection time, and screen the target points that need to be subjected to AFM ultra-precision detection.

[0040] Step 5: Based on the selected AFM ultra-precision detection target point location information, plan the AFM scanning path and perform AFM ultra-precision scanning detection on the corresponding target points.

[0041] Specific Implementation Scheme Two: In Step One, the high-resolution optical microscope in the dual-station inspection device (application number: 202310603247.1) that uses KDP crystal surface micro / nano defect optical imaging for rough inspection and AFM scanning for fine inspection moves along the surface of the optical element according to a set trajectory and takes pictures at designated positions. The optical microscope has a resolution of 5120×5120, a zoom lens, and is equipped with a 4.5x eyepiece and a 2x objective lens to acquire sub-aperture images of the optical element surface. All other aspects of this implementation scheme are the same as in Specific Implementation Scheme One.

[0042] Specific Implementation Scheme 3: As described in Step 1, a sub-aperture image of the optical element surface is obtained using an optical microscope. An overlapping area is set in the sub-aperture image to ensure effective image stitching, and to guarantee image quality, the overlapping area of ​​the sub-aperture images is set to 20%. All other aspects of this implementation scheme are the same as in Specific Implementation Scheme 1.

[0043] Specific Implementation Plan Four: As described in Step One, the sub-aperture images are stitched together. An adaptive threshold segmentation method is used to perform threshold segmentation on the sub-aperture images. A threshold is calculated for each pixel in the image to avoid losing feature points due to lighting conditions and background. This method has good adaptability to the image. The SURF algorithm is used to extract feature points from the threshold segmentation results. These feature points are then detected, located, oriented, and described to provide a foundation for reliable feature point matching. The FLANN algorithm is used to calculate the matching rate of each feature point and perform feature point matching. Based on this, feature matching pairs limiting individual feature points are used, along with the RANSAC algorithm to eliminate unreliable matches, improving the feature point matching rate. Finally, the sub-aperture images are stitched together. The stitched result of the 9×9 full-aperture sub-aperture images is shown below. Figure 3 As shown in the figure, the splicing effect is quite good. This implementation plan is otherwise the same as specific implementation plan three.

[0044] Specific Implementation Plan 5: Step 2, which describes extracting the minimum bounding rectangle of the target point based on the gradient of the full-aperture image, specifically involves extracting the minimum bounding rectangle of the target point based on the gradient of the grayscale change rate of the full-aperture image, including four steps: edge extraction, morphological processing, minimum bounding rectangle calculation, and contour clustering. Specifically, during edge extraction, the location with the larger image gradient is identified as the edge position of the target point. Then, morphological processing is performed to address edge breaks and false detections. Opening operations are used to remove false detections, and closing operations are used to connect the edges to obtain a complete contour. Next, the minimum bounding rectangle of the obtained complete contour is extracted. Finally, to address the problem of repeated detection of the target point contour, contour clustering is used to group multiple contours with close distances into one class to obtain the final minimum bounding rectangle of the detected contour. The specific process is as follows: Figure 4 As shown. This implementation plan is otherwise the same as specific implementation plan one.

[0045] Specific Implementation Plan Six: (e.g.) Figure 5 As shown, the morphological information mentioned in step two refers to the angle between the smallest bounding rectangle of the target point and the horizontal direction, and the shape information refers to the aspect ratio of the bounding rectangle of the target point. This implementation scheme is otherwise the same as specific implementation scheme five.

[0046] In this implementation scheme, for target points with a size range of 0.5–50 μm, such as Figure 6 As shown, the extracted target point feature information is saved using the xlwt function library on the Python platform. Target points are classified according to defect type, whether they are contaminants, and size range. Due to the detection limitations of optical microscopes, target points with a size range of 0.5–2 μm are not identified or classified; their location information is directly sent to the AFM for ultra-precision detection. For target points with a size range of 2–50 μm, separate datasets are created based on their size: 2–50 μm, 10–50 μm, 20–50 μm, and 30–50 μm.

[0047] Specific Implementation Scheme Seven: The target point classification model based on convolutional neural networks described in Step Three uses EfficientNet-B4 as the backbone network and employs transfer learning to train the model to improve its classification accuracy. All other aspects of this implementation scheme are the same as Specific Implementation Scheme One.

[0048] In this implementation scheme, the obtained datasets of different size ranges from optical inspection are used as the training set for the classification network. During defect and contaminant classification, high-pressure clean gas is used to pre-treat the surface of the optical components to remove contaminants, achieving accurate differentiation between defects and contaminants. The classification accuracies obtained for different size range training sets are as follows: 87.92% for 2–50 μm, 96.30% for 10–50 μm, 99.04% for 20–50 μm, and 99.41% for 30–50 μm. High classification accuracy indicates that the optical microscope images contain sufficient target point feature information, eliminating the need for AFM ultra-precision inspection. Then, with a probability threshold of 0.601, the classification model achieves a defect classification accuracy of 99.74% and a contaminant classification accuracy of 97.91% for target points with a size range of 20–50 μm, meeting the high confidence requirement of this invention.

[0049] Specific Implementation Scheme Eight: In step four, when detecting micro-nano-scale target points of 0.5–50 μm on the surface of optical components, the size threshold for AFM detection target points is set to 20 μm. That is, AFM ultra-precision detection is performed on target points with a size range of no more than 20 μm that are not contaminants. The rest of this implementation scheme is the same as Specific Implementation Scheme One.

[0050] Specific Implementation Scheme Nine: In step five, the ant colony algorithm is used to plan the AFM scanning path. To accommodate any AFM scanning starting point, the number of ants selected must be greater than the number of target points. Then, the ants traverse all the selected AFM ultra-precise detection target points, iterating step by step until the optimal path length tends to stabilize, thus obtaining the globally optimal AFM scanning path. The rest of this implementation scheme is the same as Specific Implementation Scheme One.

[0051] Specific Implementation Scheme Ten: A high-efficiency detection system for micro-nano-level target points on the surface of optical components based on optical-AFM fusion, comprising:

[0052] The image acquisition module is used to acquire sub-aperture images of the surface of optical elements using an optical microscope, and to stitch the sub-aperture images together to obtain a full-aperture image.

[0053] The dataset construction module is used to extract the minimum bounding rectangle of the target point based on the gradient of the full-aperture image, obtain the position information, morphological information and shape information of the target point in the full-aperture image, and classify the defect type, whether it is a contaminant and the size range of the target point.

[0054] The model building module is used to construct a target point classification model based on a convolutional neural network, based on the target point's location, morphology, and shape information, as well as the corresponding defect type, whether it is a contaminant, and size range classification information. This enables the classification of defect types within different size ranges and the identification of contaminants.

[0055] The target point filtering module is used to set the size threshold of AFM detection target points based on classification accuracy and AFM detection time, and to filter target points that need to be subjected to AFM ultra-precision detection.

[0056] The path planning module is used to plan the AFM scanning path based on the selected AFM ultra-precision detection target point location information, and to perform AFM ultra-precision scanning detection on the corresponding target points.

[0057] To verify the effectiveness of this invention, a specific embodiment of the method is described using a 50mm × 50mm aperture KDP crystal optical element. The implementation process employed a Mercury ME2P-2621-4GM / CP camera from Daheng Imaging Technology Co., Ltd., with a resolution of 5120 × 5120. The lens was a POMEAS VP-LZH-7505 motorized zoom lens, set with a 4.5x eyepiece and a 2x objective lens. The sub-aperture image size was approximately 1.423mm × 1.423mm. The scanning and imaging trajectory of the optical microscope is shown below. Figure 2As shown, the motion platform in the X and Y directions is a Beijing Micro-Nano Optics Technology WN202WA100×100. The optical microscope and motion platform are installed in a dual-station inspection device for coarse optical imaging and fine AFM scanning of micro-nano defects on the surface of KDP crystals (application number: 202310603247.1). The sub-aperture image overlap area is set to 20%.

[0058] Statistics show that a 50mm × 50mm diameter KDP crystal optical element has an average of approximately 250 target points smaller than 20μm on its surface, including 70 defect points and 180 contaminant points. The time required for a single target point to be scanned and inspected at the AFM precision inspection station (20*20μm) is approximately 8.5 minutes. Without pre-inspection and classification using an optical microscope, AFM ultra-precision inspection of 250 target points on one optical element would take approximately 35.4 hours. However, after classification and screening using the method of this invention, the AFM ultra-precision inspection time is reduced to 10.48 hours, improving inspection efficiency by approximately 70.3%. The greater the number of contaminants and the more target points with a size range of 20–50μm, the more significant the improvement in AFM ultra-precision inspection efficiency brought about by this invention. Figure 7 As shown, the AFM ultra-precision detection method can effectively obtain detailed feature information of micro-nano target points below 20μm, which confirms the effectiveness of the optical-AFM fusion method proposed in this invention for detecting micro-nano target points on the surface of optical components.

[0059] The ant colony algorithm is used to plan the AFM detection path for the target points on the sample surface. The AFM scanning paths before and after optimization are as follows: Figure 8 As shown, the total path length before optimization was 1253.71 mm, and the total path length after optimization was 327.83 mm. The AFM probe movement efficiency was improved by 73.7% after optimization, further confirming that the present invention can achieve efficient and ultra-precise detection of micro-nano-level target points of 0.5-50 μm on the surface of optical elements, thereby improving detection efficiency.

[0060] While the present invention has been disclosed above, its scope of protection is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and all such changes and modifications will fall within the scope of protection of the present invention.

Claims

1. A highly efficient method for detecting micro-nano-scale target points on the surface of optical components based on optical-AFM fusion, characterized in that, Includes the following steps: Step 1: Obtain sub-aperture images of the optical element surface using an optical microscope, and stitch the sub-aperture images together to obtain a full-aperture image; Step 2: Extract the minimum bounding rectangle of the target point based on the gradient of the full-aperture image to obtain the position information, morphological information and shape information of the target point in the full-aperture image, and classify the defect type, whether it is a contaminant and the size range of the target point. Step 3: Based on the location, morphology, and shape information of the target point, as well as the corresponding defect type, whether it is a contaminant, and size range classification information, construct a target point classification model based on a convolutional neural network to classify defect types of different size ranges and identify contaminants. Step 4: Set the size threshold of the AFM detection target point based on the classification accuracy and AFM detection time, and screen the target points that need to be subjected to AFM ultra-precision detection. Step 5: Based on the selected AFM ultra-precision detection target point location information, plan the AFM scanning path and perform AFM ultra-precision scanning detection on the corresponding target points.

2. The efficient detection method for micro-nano-level target points on the surface of optical elements based on optical-AFM fusion according to claim 1, characterized in that, The optical microscope described in step one has a resolution of 5120×5120, a zoom lens, and is equipped with a 4.5x eyepiece and a 2x objective lens to acquire sub-aperture images of the surface of the optical elements.

3. The efficient detection method for micro-nano-level target points on the surface of optical components based on optical-AFM fusion according to claim 1, characterized in that, In step one, a sub-aperture image of the surface of the optical element is obtained using an optical microscope. An overlapping area is set in the sub-aperture image to ensure effective image stitching.

4. The efficient detection method for micro-nano-level target points on the surface of optical elements based on optical-AFM fusion according to claim 3, characterized in that, Step one involves stitching the sub-aperture images. An adaptive thresholding method is used to perform thresholding on the sub-aperture images, calculating a threshold for each pixel in the image to avoid losing feature points due to lighting conditions and background. The SURF algorithm is used to extract feature points from the thresholding results, and these feature points are then detected, located, oriented, and described. The FLANN algorithm is used to calculate the matching rate of each feature point, and feature point matching is performed. Based on this, a feature matching pair limiting individual feature points is used, along with the RANSAC algorithm to remove unreliable matching points, thereby improving the feature point matching rate. Finally, the sub-aperture images are stitched together.

5. The efficient detection method for micro-nano-level target points on the surface of optical elements based on optical-AFM fusion according to claim 1, characterized in that, Step two, which involves extracting the minimum bounding rectangle of the target point based on the gradient of the full-aperture image, includes four steps: edge extraction, morphological processing, calculation of the minimum bounding rectangle, and contour clustering. During edge extraction, the location with the larger image gradient is determined as the edge location of the target point. Then, morphological processing is performed on the edge breakpoints and false detections that occur during edge extraction. Opening operations are used to remove false detections, and closing operations are used to connect the edges to obtain the complete contour. Next, the minimum bounding rectangle of the obtained complete contour is extracted. Finally, contour clustering is used to group multiple contours with close distances into one class to obtain the minimum bounding rectangle of the final detected contour.

6. The efficient detection method for micro-nano-level target points on the surface of optical components based on optical-AFM fusion according to claim 5, characterized in that, The morphological information mentioned in step two refers to the angle between the smallest bounding rectangle of the target point and the horizontal direction, and the shape information refers to the aspect ratio of the bounding rectangle of the target point.

7. The efficient detection method for micro-nano-level target points on the surface of optical components based on optical-AFM fusion according to claim 1, characterized in that, The target point classification model based on convolutional neural networks described in step three uses EfficientNet-B4 as the backbone network and employs transfer learning to train the model.

8. The efficient detection method for micro-nano-level target points on the surface of optical components based on optical-AFM fusion according to claim 1, characterized in that, In step four, when detecting micro-nano-scale target points of 0.5–50 μm on the surface of optical components, the size threshold of the AFM detection target points is set to 20 μm, that is, AFM ultra-precision detection is performed on target points with a size range of no more than 20 μm and which are not contaminants.

9. The efficient detection method for micro-nano-level target points on the surface of optical elements based on optical-AFM fusion according to claim 1, characterized in that, In step five, the ant colony algorithm is used to plan the AFM scanning path. To accommodate any AFM scanning starting point, the number of ants selected is greater than the number of target points. Then, the ants traverse all the selected AFM ultra-precise detection target points, iterating step by step until the optimal path length tends to stabilize, thus obtaining the globally optimal AFM scanning path.

10. A high-efficiency detection system for micro-nano-level target points on the surface of optical components based on optical-AFM fusion, characterized in that, The system has a program module corresponding to the steps of any one of the claims 1 to 9 above, and executes the steps in the above-described efficient detection method for micro-nano-level target points on the surface of optical elements based on optical-AFM fusion when it is run.

Images