A method and system for adjusting a silicon-based resonant device based on image recognition

Abstract

An embodiment of the present invention provides a method and system for adjusting a silicon-based resonant device based on image recognition, which reduces the influence of image distortion on the accuracy of image processing through a camera calibration method; calculates the center position of the laser spot by using an adaptive threshold segmentation method and a gray center-of-gravity method; quickly identifies the microstructure in the silicon-based resonant device through a shape matching algorithm; accurately measures the relevant dimensions of the microstructure by using a Hough detection algorithm; the image processing system serves as the feedback for motion control and laser processing, and the frequency detection circuit serves as the feedback for adjusting the performance of the silicon-based resonant device, and cooperatively controls the motion parameters of the four-axis motion platform and the energy parameters of the ultrafast pulsed laser, thereby realizing the automatic and precise trimming of the structural dimensions of the silicon-based resonant device. The present invention has the advantages of controllable trimming method, high adjustment accuracy, high adjustment efficiency, etc., and is applicable to the batch manufacturing of high-performance silicon-based resonant devices.

Description

Technical Field

[0001] The present invention relates to the field of micro-nano manufacturing of silicon-based resonant devices, and particularly to a method and system for adjusting a silicon-based resonant device based on image recognition. Background Art

[0002] Silicon-based resonant devices have the characteristics of small size, light weight, low cost, low power consumption, and suitability for mass production, and can be widely applied in fields such as automotive traction control systems, driving stability systems, camera stabilization systems, aircraft stabilization systems, and military, etc., and have become an important direction for the development of microelectromechanical technology. Due to the large processing errors existing in the existing processing technology for silicon-based resonant devices, it restricts the improvement of the yield of current silicon-based resonant devices. The existing process trimming methods for silicon-based devices include chemical trimming, electrical trimming, and laser trimming, etc., and the laser trimming method is the most widely applied due to its advantages of high trimming control accuracy and long-lasting and effective trimming effect.

[0003] However, there are still the following problems in the laser trimming of silicon-based resonant devices: the microstructures of silicon-based devices are complex, and the trimming is difficult; due to processing errors, the microstructure sizes cannot be determined, which affects the setting of trimming parameters and reduces the trimming accuracy; there are difficulties in the positioning of the laser spot and the control of the etching position during trimming, resulting in the deficiency of low batch trimming efficiency. Summary of the Invention

[0004] In view of the above problems, embodiments of the present invention provide a method and system for adjusting a silicon-based resonant device based on image recognition, so as to overcome the above problems or at least partially solve the above problems.

[0005] In a first aspect of embodiments of the present invention, a method for adjusting a silicon-based resonant device based on image recognition is provided, including:

[0006] Performing camera calibration on the imaging system, solving camera parameters, and using the calibrated imaging system for image acquisition;

[0007] Obtaining a microstructure image in the silicon-based resonant device, and identifying the microstructure type in the microstructure image according to microstructure template matching, wherein the microstructure template is the original structure of different microstructures in different silicon-based resonators;

[0008] Extracting the contour of the microstructure image, and measuring the size of the microstructure by using a detection algorithm corresponding to the microstructure;

[0009] Collecting an image of the laser spot, selecting the area where the spot is located as the region of interest in the image of the laser spot, performing threshold segmentation on the region of interest, and calculating the center coordinates of the laser spot;

[0010] Locate the starting coordinates of the area to be trimmed in the microstructure, and achieve automatic positioning according to the center coordinates of the laser spot. Plan the processing route based on the measured size of the microstructure to achieve automated laser trimming.

[0011] Optionally, the camera calibration of the imaging system includes:

[0012] Select M×N squares with a size of a as the calibration board;

[0013] Calibrate the camera using a calibration method based on geometric invariant moments, radial constraints, and the Gaussian theorem.

[0014] Optionally, obtaining the microstructure image in the silicon-based resonant device and identifying the microstructure type in the microstructure image according to microstructure template matching includes:

[0015] Intercept different types of original structures as templates according to the geometric characteristics of the silicon-based resonant device structure;

[0016] Match the microstructure in the microstructure image with the microstructure template according to the shape matching algorithm to determine the microstructure type in the microstructure image.

[0017] Optionally, extracting the contour of the microstructure image and measuring the size of the microstructure using the detection algorithm corresponding to the microstructure includes:

[0018] Detect and remove the incorrect contours in the microstructure contour using an improved chord-to-point distance accumulation algorithm to obtain the correct microstructure contour;

[0019] Measure the pixel distance of the correct microstructure contour according to the Hough line detection algorithm to obtain the size of the microstructure.

[0020] Optionally, selecting the area where the light spot is located as the region of interest in the image of the laser spot, performing threshold segmentation on the region of interest, and calculating the center coordinates of the laser spot includes:

[0021] Obtain the histogram of the region of interest, and segment the region of interest with the gray value corresponding to the first peak in the histogram to obtain the segmented light spot image;

[0022] Calculate the center coordinates of the segmented light spot image using the gray centroid method.

[0023] Optionally, locating the starting coordinates of the area to be trimmed in the microstructure and achieving automatic positioning according to the center coordinates of the laser spot includes:

[0024] By matching and recognizing the real-time microstructure image collected by the calibrated imaging system with the microstructure template, the coordinates of the origin of the microstructure to be adjusted are obtained;

[0025] Calculate the difference between the starting coordinates of the area to be adjusted of the microstructure and the coordinates of the center of the light spot, and control the motion platform to make the two sets of coordinates coincide to achieve automatic positioning.

[0026] Optionally, planning the processing route according to the measured size of the microstructure to achieve automated laser trimming includes:

[0027] Etching parameters, etching shape, laser processing power, and processing speed set according to the measured microstructure size;

[0028] Turn on the laser and the air pump to achieve automated laser trimming.

[0029] In the second aspect of the embodiments of the present invention, a silicon-based resonant device adjustment system based on image recognition is provided, including:

[0030] A camera calibration module for calibrating the imaging system, solving camera parameters, and using the calibrated imaging system for image acquisition;

[0031] An image recognition module for obtaining the microstructure image in the silicon-based resonant device and matching and recognizing the microstructure type in the microstructure image according to the microstructure template, where the microstructure template is the original structure of different microstructures in different silicon-based resonators;

[0032] A size measurement module for extracting the contour of the microstructure image and measuring the size of the microstructure using the detection algorithm corresponding to the microstructure;

[0033] A light spot positioning module for collecting the image of the laser light spot, selecting the area where the light spot is located as the region of interest in the image of the laser light spot, performing threshold segmentation on the region of interest, and calculating the center coordinates of the laser light spot;

[0034] An automatic trimming module for positioning the starting coordinates of the area to be trimmed in the microstructure, achieving automatic positioning according to the center coordinates of the laser light spot, planning the processing route according to the measured size of the microstructure, and achieving automated laser trimming.

[0035] The embodiments of the present invention have the following advantages:

[0036] In an embodiment of the present invention, camera calibration is performed on an imaging system to solve camera parameters. The calibrated imaging system is used for image acquisition to obtain a microstructure image of the silicon-based resonant device. The microstructure type in the microstructure image is identified based on microstructure template matching. The microstructure template is the original structure of different microstructures in different silicon-based resonators. The contour of the microstructure image is extracted, and the size of the microstructure is measured using a detection algorithm corresponding to the microstructure. An image of a laser spot is acquired, and the area where the spot is located in the image of the laser spot is selected as the region of interest. Threshold segmentation is performed on the region of interest, the center coordinates of the laser spot are calculated, the starting coordinates of the area to be trimmed in the microstructure are located, and automatic positioning is achieved based on the center coordinates of the laser spot. A processing route is planned according to the size of the microstructure measured, and automated laser trimming is realized. In an embodiment of the present invention, the influence of image distortion on the image processing accuracy is reduced by the camera calibration method. By identifying different microstructures in the silicon-based resonant device, the structural size of the area to be trimmed is accurately measured, improving the accuracy of parameter setting during process machining. Automatic positioning is achieved through the center coordinates of the laser spot and the starting coordinates of machining, improving the positioning speed and positioning accuracy, and overall improving the trimming accuracy and efficiency. Description of the Drawings

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for the description of the embodiments of the present invention. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0038] Figure 1 It is a flowchart of the steps of a method for adjusting a silicon-based resonant device based on image recognition provided by an embodiment of the present invention;

[0039] Figure 2 It is a schematic structural diagram of a camera distortion model provided by an embodiment of the present invention;

[0040] Figure 3 It is a schematic diagram of a typical microstructure of a silicon-based resonant device provided by an embodiment of the present invention;

[0041] Figure 4 It is a flowchart of a microstructure shape matching algorithm provided by an embodiment of the present invention;

[0042] Figure 5 It is a flowchart of the size measurement of a microstructure provided by an embodiment of the present invention;

[0043] Figure 6 It is a schematic diagram of a picosecond laser spot provided by an embodiment of the present invention;

[0044] Figure 7 It is a schematic diagram of a light spot after threshold segmentation provided by an embodiment of the present invention;

[0045] Figure 8 It is a flowchart of the automatic trimming of the microstructure of a silicon-based resonant device provided by an embodiment of the present invention;

[0046] Figure 9 It is a schematic structural diagram of a silicon-based resonant device adjustment system based on image recognition provided by an embodiment of the present invention. Specific embodiments

[0047] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings in the embodiments of the present invention. Although the exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present invention can be more thoroughly understood and the scope of the present invention can be fully conveyed to those skilled in the art.

[0048] An embodiment of the present invention provides a method for adjusting a silicon-based resonant device based on image recognition. Refer to Figure 1 , Figure 1 It is a flowchart of the steps of a method for adjusting a silicon-based resonant device based on image recognition provided by an embodiment of the present invention. As shown in Figure 1 , the method includes:

[0049] Step S101: Calibrate the camera of the imaging system, solve the camera parameters, and use the calibrated imaging system for image acquisition.

[0050] Since there is distortion in the camera during the shooting process, which affects the subsequent image processing accuracy. Therefore, in order to correct the distorted pictures during the camera shooting process and ensure the measurement accuracy, it is necessary to calibrate the camera in the imaging system and use the calibrated camera to collect pictures of each microstructure in the silicon-based resonant device, so that the images of each microstructure can correctly reflect the correct shape of the microstructure of the silicon-based resonant device to the greatest extent, and then ensure the measurement accuracy of the subsequent microstructure size.

[0051] Optionally, the calibration of the camera for the imaging system includes:

[0052] Select M×N squares with a size of a as the calibration board;

[0053] Calibrate the camera using a calibration method of geometric invariant moments, radial constraints, and Gauss's theorem.

[0054] In this embodiment, first, select M×N squares with a size of a as the calibration board (for example, 7×7 squares with a size of 0.01 mm), take pictures of the calibration board with a camera, and extract the contours of the obtained calibration board pictures. Obtain the centers of the small squares on the calibration board, use the properties of geometric invariant moments, take the centers as the coordinates of the image coordinate system, and then obtain the world coordinate system coordinates corresponding to the centers according to the selected origin of the world coordinate system. Calculate the calibration parameters of the camera through the calibration method. The specific steps are as follows:

[0055] A1: Analyze the camera distortion model, that is, establish and analyze the conversion relationships between the pixel coordinate system, the image coordinate system, the camera coordinate system, and the world coordinate system, as Figure 2 shown, Figure 2 which is a camera distortion model provided by an embodiment of the present invention.

[0056] A2: Determine the relationship between the actual image coordinates and the ideal image coordinates:

[0057]

[0058] where Dx and Dy represent the radial distortion between the ideal image coordinate point and the actual image coordinate point, D x is in the x-axis direction, D y is in the y-axis direction, x u and y u represent the ideal image coordinate point, and x d and y d represent the actual image coordinate point.

[0059] A3: Solve the parameters: t y t x , R, (u0, v0).

[0060] t y t x are respectively 3×1 translation vectors in the x-direction and y-direction, s x is the scale factor, s x = f x / f y , R is a 3×3 rotation matrix, and (u0, v0) is the image center.

[0061] A4: Introduce the Gaussian imaging theorem to solve the parameters: t z t z , f.

[0062] t z is a 3×1 translation vector in the z-direction, z c is the distance from the camera lens to the calibration board, f is the actual focal length of the camera, and F is the equivalent focal length of the lens in the camera, that is:

[0063]

[0064] A5: Optimize the camera parameters using the LM algorithm (Levenberg - Marquarelt). The specific formula is as follows:

[0065]

[0066] Among them, m ij represents the detected point, m^ represents the mapped point, and A is the matrix of distortion parameters k1 and k2. This formula represents the residual value between two points. The purpose of optimizing the parameters is to reduce the residual value between two points.

[0067] In this embodiment, the influence of image distortion on the image processing accuracy is reduced by the camera calibration method, so that the micro - structure image collected by the calibrated imaging system can correctly reflect the correct shape of the micro - structure of the silicon - based resonant device to the greatest extent, thereby ensuring the measurement accuracy of the subsequent micro - structure size.

[0068] Step S102: Obtain the micro - structure image in the silicon - based resonant device, and identify the micro - structure type in the micro - structure image according to the micro - structure template matching. Among them, the micro - structure template is the original structure of different micro - structures in different silicon - based resonators.

[0069] The silicon - based resonant device contains multiple different micro - structures. The micro - structures of the silicon - based resonant device have obvious geometric characteristics. To measure the sizes of different micro - structures in different silicon - based resonant devices, it is necessary to identify the structure type of the device, and then measure the size of the micro - structure according to the algorithms corresponding to different micro - structures. As Figure 3 shown, Figure 3 is a schematic diagram of the typical micro - structure of the silicon - based resonant device, including an elastic beam (a), a mass block (b), a comb tooth (c), etc.

[0070] Optionally, the obtaining the micro - structure image in the silicon - based resonant device and identifying the micro - structure type in the micro - structure image according to the micro - structure template matching includes:

[0071] Intercept different types of original structures as templates according to the geometric characteristics of the silicon - based resonant device structure;

[0072] Match the micro - structure in the micro - structure image with the micro - structure template according to the shape - matching algorithm, and determine the micro - structure type in the micro - structure image.

[0073] Before image recognition, it is necessary to intercept some original structures in the silicon - based resonant device as templates. The intercepted templates are shown as (a), (b), (c) in Figure 3 . Image recognition is mainly realized by the shape - matching algorithm. The process is as shown in Figure 4As shown, first, preprocess the collected microstructure images, then extract the contours of the microstructures in the microstructure images and the microstructure template images, identify their similarity through a shape matching algorithm, and calculate the 7 Hu moment values of the microstructure images and the microstructure template images respectively. The formulas are as follows:

[0074] H0 = η 20 + η 02

[0075] H1 = (η 20 - η 02 ) 2 + 4η 11 2

[0076] H2 = (η 30 - 3η 12 ) 2 +(3η 21 - η 03 ) 2

[0077] H3 = (η 30 + η 12 ) 2 +(η 21 + η 03 ) 2

[0078] H4 = (η 30 - 3η 12 )(η 30 + η 12 )[(η 30 + 3η 12 ) 2 - 3(η 21 + η 03 ) 2 +(3η 21 - η 03 )[3(η 30 + η 12 ) 2 -(η 21 + η 03 ) 2 )

[0079] H5 = (η 20 - η 02 )[(η 30 + η 12 ) 2 -(η 21 + η 03 ) 2 + 4η 11 (η 30 + η 12 )(η21 +η 03 )]

[0080] H6 = (3η 21 -η 03 )(η 30 +η 12 )[(η 30 +η 12 ) 2 -3(η 21 +η 03 ) 2 +(η 30 -3η 12 )(η 21 +η 03 )[3(η 30 +η 12 ) 2 -(η 21 +η 03 ) 2

[0081] The Hu moment value is used to describe the shape features of the microstructure template image or the acquired microstructure image. The matching degree between the acquired microstructure image and the microstructure template image is calculated using the following formula:

[0082]

[0083] where H i B represents the Hu moment value of the microstructure template image, and H i A represents the Hu moment value of the microstructure in the acquired image.

[0084] Calculate the sum of the differences of the 7 Hu moment values between the microstructure image and the microstructure template image. The smaller the result, the higher the matching degree. When the return value is less than the preset value d (for example, d is 0.1), it is determined that the matching is successful. Otherwise, the shape matching algorithm is used to match with another microstructure template image again until the matching is successful.

[0085] Step S103: Extract the contour of the microstructure image, and measure the size of the microstructure using the detection algorithm corresponding to the microstructure.

[0086] There are different microstructures in the silicon-based resonant device, and the shape and size of each microstructure are different. For different microstructures, it is first necessary to extract the contour of the microstructure in the microstructure image and calculate the size of the microstructure according to the extracted contour of the microstructure.

[0087] Optionally, extracting the contour of the microstructure image and measuring the size of the microstructure using the detection algorithm corresponding to the microstructure includes:

[0088] ​The improved chord-to-point distance accumulation algorithm is used to detect and remove the incorrect contours in the microstructure contour, and the correct microstructure contour is obtained.

[0089] According to the Hough line detection algorithm, the pixel distance of the correct microstructure contour is measured to obtain the size of the microstructure.

[0090] In this embodiment, corresponding measurement modules are provided for different microstructures in the silicon-based resonant device, mainly implemented by the Hough line detection algorithm. As Figure 5 shown in the flowchart of the microstructure size measurement. First, the image to be measured needs to be preprocessed, and then the contour of the image structure is extracted. Since there may be impurities in the image that affect the accuracy of the contour, after extracting the contour of the binary image, the improved chord-to-point distance accumulation algorithm is used to detect the incorrect contour and remove it to obtain the correct chip contour. The specific steps are as follows:

[0091] B1: Define the chord length L. The length of L represents the number of contour points, set as 2k. Pi is the midpoint of L, that is, Pi-k and Pi+k are the endpoints of the chord length L.

[0092] B2: Calculate the distance from Pi to the chord length L, that is, the height h of the triangle. Calculate the distance d1 between Pi and Pi-k, and calculate the distance d2 between Pi and Pi+k.

[0093] B3: Calculate the value of the angle r according to the following formula.

[0094]

[0095] B4: Use the angle r as the judgment condition for the curvature of this point.

[0096] B5: Then select Pi as the point at 1 / 4 and 3 / 4 of L respectively, and perform the above steps B1 to B4 for calculation to prevent some incorrect contour points from not being detected.

[0097] B6: Select three different chord lengths L and perform the above calculations on the contour.

[0098] B7: Respectively perform normalization processing on the three groups of calculated angle sets, that is: angle Ci / max(angle C).

[0099] B8: Multiply the three groups of normalized angles to obtain the angle product, which is used as the final curvature for judging the contour corner points.

[0100] B9: Set a threshold to judge the corner point contour.

[0101] Since the microstructure of the silicon-based resonant device presents a geometric shape and the straight lines at the edges of the microstructure of the silicon-based resonant device are at fixed angles, the contour points are traversed first to determine the straight line angle of the structure, and then the Hough line detection is performed on the structure. The specific steps are as follows:

[0102] C1: Obtain the straight line contour points in sequence.

[0103] C2: Traverse the contour points in sequence, take the first point A(x0, y0) and the last point B(x1, y1), and there are T points between the two points.

[0104] C3: Calculate the angle as θ according to the straight line formed between A and B.

[0105] C4: Add θ to the array m.

[0106] C5: Set a threshold k, count the number n of repeated elements in the array m. If n > k, extract the interval [n - 2, n + 3) where the element n is located.

[0107] C6: Obtain all elements n that meet the conditions, and obtain all the angle intervals as the angles in the Hough space.

[0108] Finally, according to the straight lines detected by the Hough line detection and the geometric properties of the structure, the dimensions of the chip structure are calculated to obtain the length, width, angle, etc. of the structure.

[0109] In this embodiment, for the extracted microstructure contour, the improved chord-to-point distance accumulation algorithm is used to detect and remove the incorrect contours in the microstructure contour, and the Hough detection algorithm is used to measure the microstructure contour after removing the incorrect contours, so as to obtain accurate microstructure dimensions.

[0110] Step S104: Collect the image of the laser spot, select the area where the spot is located as the region of interest in the image of the laser spot, perform threshold segmentation on the region of interest, and calculate the center coordinates of the laser spot.

[0111] Compared with femtosecond lasers and nanosecond lasers, the pulse width of picosecond pulsed lasers is close to the thermal conduction time between electrons and the lattice in single-crystalline silicon, and it has a relatively high average power close to that of femtosecond laser pulses. Using picosecond lasers to process single-crystalline silicon can effectively reduce the thermal influence range during the processing, improve the processing accuracy, and obtain higher processing quality. Therefore, a picosecond laser is used to implement the laser adjustment system.

[0112] Optionally, the step of selecting the area where the spot is located as the region of interest in the image of the laser spot, performing threshold segmentation on the region of interest, and calculating the center coordinates of the laser spot includes:

[0113] Obtain the histogram of the region of interest, and segment the region of interest with the gray value corresponding to the first peak in the histogram to obtain the segmented spot image;

[0114] Calculate the central coordinates of the segmented spot image by using the gray centroid method.

[0115] Since the gray value of the central part of the spot generated by the picosecond pulsed laser is the lowest and concentrated within a certain range, the gray value of the peripheral part is higher, and there is a large amount of random noise. Therefore, the laser spot image can be divided into three parts, as Figure 6 shown. Region A is the background of the image, that is, the region outside the laser beam; Region B is the part where the material is not completely damaged due to insufficient laser energy, and there is a large amount of random noise in this region due to the splash of slag generated by the melting of single crystal silicon in the central region. Region C is the central part of the laser spot, which is the region where single crystal silicon is completely melted, so the average gray value is the lowest.

[0116] Among them, only the size of Region C is relatively fixed and the gray distribution is also relatively uniform. Therefore, the C region of the spot image can be segmented from other parts by setting a threshold, and then the central coordinates of the spot can be located. First, the histogram of the spot image needs to be obtained, and then the image is segmented with the gray value corresponding to the first peak in the histogram. The obtained image is closer to the central part of the spot, and the surrounding noise is effectively removed. The gray distribution of the image is relatively uniform and shows overall symmetry, as Figure 7 shown, and then the gray centroid method is used to calculate the central coordinates of the target image, and more accurate results can be obtained.

[0117] Step S105: Locate the starting coordinates of the area to be trimmed in the microstructure, and achieve automatic positioning according to the central coordinates of the laser spot, and plan the processing route according to the size measured by the microstructure to achieve automated laser trimming.

[0118] Through steps S101 to S104, the size of the microstructure of the silicon-based resonant device and the central coordinates of the laser spot are obtained. In addition, the starting coordinates of the area to be trimmed in the microstructure of the silicon-based resonant device need to be obtained to achieve automatic positioning, so as to achieve automated laser trimming. For example, by controlling the motion parameters of the four-axis motion platform and the energy parameters of the ultrafast pulsed laser, the automated precise trimming of the structural size of the silicon-based resonant device is achieved.

[0119] Optionally, the locating the starting coordinates of the area to be trimmed in the microstructure and achieving automatic positioning according to the central coordinates of the laser spot includes:

[0120] By matching and recognizing the real-time microstructure image collected by the calibrated imaging system with the microstructure template, the coordinates of the origin to be trimmed in the microstructure are obtained;

[0121] Calculate the difference between the starting coordinates of the area to be trimmed of the microstructure and the coordinates of the center of the light spot, and control the moving platform to make the two point coordinates coincide to achieve automatic positioning.

[0122] Optionally, planning the processing route according to the measured size of the microstructure to achieve automated laser trimming, including:

[0123] Etching parameters, etching shape, laser processing power, and processing speed set according to the measured size of the microstructure;

[0124] Turn on the laser and the air pump to achieve automated laser trimming.

[0125] In this embodiment, the method for automatically trimming the microstructure in the silicon-based resonant device is as Figure 8 shown. First, obtain the starting coordinates of the area to be trimmed of the microstructure, that is, the position coordinates of the origin to be trimmed of the folded beam or the mass block. Perform pyramid layering on the collected microstructure image. Pyramid layering is a search method from fine to coarse. By reducing the resolution of the template image and the image to be matched, feature matching is performed on the two images with low resolution. After obtaining the matching position, search for the matching position in the neighborhood of the corresponding position on the image with a higher resolution level until the last level of the image, and find the final correct matching position. When matching low-resolution images, there are fewer pixel points and the speed is fast; when matching high-resolution images, due to the reduced search range, the speed is also greatly improved. Among them, the template is automatically selected according to different objects to be trimmed with different microstructure templates. Secondly, extract the edge features of the microstructure image, smooth-filter the microstructure image with a filter, and process the microstructure image using non-maximum suppression technology. Then, use a thinning algorithm to obtain an edge image with a single-pixel width. Finally, by moving the template image point by point in the microstructure image to be matched and using the Hausdorff distance as the similarity metric, obtain the position of the template image when the similarity is the largest, and this position is the correct matching position. Through the above image processing technology, the position coordinates of the origin to be trimmed of the microstructure can be automatically obtained.

[0126] Subtract the coordinates of the point to be processed of the microstructure from the coordinates of the center of the laser spot to obtain the coordinate difference. Calculate the moving distances in the X and Y directions by calibrating to obtain the scale factor between the pixel coordinates and the physical coordinates of the imaging system. The computer controls the motion platform to move using the obtained moving distances. After the movement is completed, it will be determined again whether the starting coordinates of the area to be adjusted coincide with the center coordinates of the laser spot, that is, whether the coordinate difference between the starting coordinates of the area to be adjusted and the center coordinates of the laser spot meets the preset positioning accuracy requirements. For example, if the preset positioning accuracy is 5um, when the coordinate difference between the starting coordinates of the area to be adjusted and the center of the laser beam is less than 5um, the positioning accuracy requirements are met and the automatic positioning is completed. Otherwise, it is necessary to calculate the distance between the starting coordinates of the area to be adjusted and the center coordinates of the laser spot again, and control the moving platform to perform re-positioning until the positioning accuracy requirements are met. After automatically completing the positioning and reaching the corresponding positioning accuracy, set the etching parameters according to the measured microstructure size, such as the power of the laser and the processing speed, and start the laser to achieve automatic adjustment. After completing one adjustment, detect whether the resonator frequency has changed after the adjustment. If it has changed, the adjustment ends. Otherwise, start the laser again for the second adjustment until it is detected that the resonator frequency has changed after the adjustment.

[0127] An embodiment of the present invention also provides a silicon-based resonator device adjustment system based on image recognition. Refer to Figure 9 , Figure 9 which is a schematic structural diagram of a silicon-based resonator device adjustment system proposed in an embodiment of the present application. The device includes:

[0128] A camera calibration module for calibrating the imaging system, solving the camera parameters, and using the calibrated imaging system for image acquisition;

[0129] An image recognition module for obtaining the microstructure image in the silicon-based resonator device and identifying the microstructure type in the microstructure image according to the microstructure template matching, where the microstructure template is the original structure of different microstructures in different silicon-based resonators;

[0130] A size measurement module for extracting the contour of the microstructure image and measuring the size of the microstructure using the detection algorithm corresponding to the microstructure;

[0131] A spot positioning module for collecting the image of the laser spot, selecting the area where the spot is located as the region of interest in the image of the laser spot, performing threshold segmentation on the region of interest, and calculating the center coordinates of the laser spot;

[0132] An automatic adjustment module for positioning the starting coordinates of the area to be adjusted in the microstructure, achieving automatic positioning according to the center coordinates of the laser spot, planning the processing route according to the size of the microstructure measured, and realizing automatic laser adjustment.

[0133] An embodiment of the present invention also provides a method and system for adjusting a silicon-based resonant device based on image recognition. The method includes: performing camera calibration on an imaging system, solving camera parameters, using the calibrated imaging system for image acquisition to obtain a microstructure image in the silicon-based resonant device, identifying the microstructure type in the microstructure image according to microstructure template matching, where the microstructure template is the original structure of different microstructures in different silicon-based resonators, extracting the contour of the microstructure image, measuring the size of the microstructure using a detection algorithm corresponding to the microstructure, collecting an image of a laser spot, selecting the area where the spot is located as the region of interest in the image of the laser spot, performing threshold segmentation on the region of interest, calculating the center coordinates of the laser spot, positioning the starting coordinates of the area to be trimmed in the microstructure, and achieving automatic positioning according to the center coordinates of the laser spot, planning a processing route according to the measured size of the microstructure, and realizing automated laser trimming. In the embodiment of the present invention, the influence of image distortion on the image processing accuracy is reduced by the camera calibration method. By identifying different microstructures in the silicon-based resonant device, the structural size of the area to be trimmed is accurately measured, improving the accuracy of parameter setting during the process of processing. Automatic positioning is achieved through the center coordinates of the laser spot and the starting coordinates of processing, improving the positioning speed and positioning accuracy, and overall improving the trimming accuracy and efficiency.

[0134] Each embodiment in this specification is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the embodiments can be referred to each other.

[0135] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods and systems according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal devices generate a device for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0136] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing terminal device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured product including an instruction device, and the instruction device implements the functions in Figure 1 one process or multiple processes and / or blocksFigure 1 The functions specified in one or more boxes.

[0137] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device, so that a series of operation steps are executed on the computer or other programmable terminal device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable terminal device provide for implementing the steps of the functions specified in one or more processes and / or boxes Figure 1 One process or more processes and / or boxes Figure 1 The steps of the functions specified in one or more boxes.

[0138] Although the preferred embodiments of the embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they learn the basic creative concepts. Therefore, the appended claims are intended to be construed to include the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention.

[0139] Finally, it should also be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or terminal device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or terminal device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or terminal device comprising the element.

[0140] The above has introduced in detail a method and system for adjusting a silicon-based resonant device based on image recognition provided by the present invention. Specific examples are used in this article to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A method for adjusting a silicon-based resonant device based on image recognition, characterized in that, The method includes: Performing camera calibration on the imaging system, solving for camera parameters, and using the calibrated imaging system for image acquisition; Obtaining a microstructure image in the silicon-based resonant device, and identifying the microstructure type in the microstructure image according to microstructure template matching, where the microstructure template is the original structure of different microstructures in different silicon-based resonators; Extracting the contour of the microstructure image, and measuring the size of the microstructure using the detection algorithm corresponding to the microstructure; Collecting an image of the laser spot, selecting the area where the spot is located as the region of interest in the image of the laser spot, performing threshold segmentation on the region of interest, and calculating the center coordinates of the laser spot; Locating the starting coordinates of the area to be trimmed in the microstructure, achieving automatic positioning according to the center coordinates of the laser spot, planning a processing route according to the measured size of the microstructure, and realizing automated laser trimming.

2. The method according to claim 1, wherein The performing camera calibration on the imaging system includes: Selecting M×N squares with a size of a as the calibration board; Calibrating the camera using a calibration method of geometric invariant moments, radial constraint, and Gauss's theorem.

3. The method according to claim 1, characterized in that, The obtaining a microstructure image in the silicon-based resonant device, and identifying the microstructure type in the microstructure image according to microstructure template matching includes: Intercepting different types of original structures as templates according to the geometric characteristics of the silicon-based resonant device structure; Matching the microstructure in the microstructure image with the microstructure template according to the shape matching algorithm, and determining the microstructure type in the microstructure image.

4. The method according to claim 1, characterized in that, The extracting the contour of the microstructure image, and measuring the size of the microstructure using the detection algorithm corresponding to the microstructure includes: Detecting and removing incorrect contours in the microstructure contour using an improved chord-to-point distance accumulation algorithm to obtain the correct microstructure contour; Measuring the pixel distance of the correct microstructure contour according to the Hough line detection algorithm to obtain the size of the microstructure.

5. The method according to claim 1, characterized in that, The selecting the area where the spot is located as the region of interest in the image of the laser spot, performing threshold segmentation on the region of interest, and calculating the center coordinates of the laser spot includes: Obtaining the histogram of the region of interest, segmenting the region of interest using the gray value corresponding to the first peak in the histogram as the threshold to obtain the segmented spot image; Calculating the center coordinates of the segmented spot image using the gray center of gravity method.

6. The method according to claim 1, characterized in that The locating the starting coordinates of the area to be trimmed in the microstructure, and achieving automatic positioning according to the center coordinates of the laser spot includes: Obtaining the coordinates of the origin to be trimmed of the microstructure by matching and identifying the real-time microstructure image collected by the calibrated imaging system with the microstructure template; Calculating the difference between the starting coordinates of the area to be trimmed of the microstructure and the center coordinates of the spot, and controlling the motion platform to make the two point coordinates coincide to achieve automatic positioning.

7. The method according to claim 1, wherein The planning a processing route according to the measured size of the microstructure, and realizing automated laser trimming includes: Setting the etching parameters, etching shape, laser processing power, and processing speed according to the measured microstructure size; Turning on the laser and the air pump to realize automated laser trimming.

8. A silicon-based resonant device adjustment system based on image recognition, characterized in that, Includes: A camera calibration module, which is used to calibrate the camera of the imaging system, solve the camera parameters, and use the calibrated imaging system for image acquisition; An image recognition module, which is used to obtain the microstructure image in the silicon-based resonant device, and identify the microstructure type in the microstructure image according to the microstructure template matching, wherein the microstructure template is the original structure of different microstructures in different silicon-based resonators; A dimension measurement module, which is used to extract the contour of the microstructure image and measure the dimension of the microstructure by using the detection algorithm corresponding to the microstructure; A spot positioning module, which is used to collect the image of the laser spot, select the area where the spot is located as the region of interest in the image of the laser spot, perform threshold segmentation on the region of interest, and calculate the center coordinates of the laser spot; An automatic trimming module, which is used to locate the starting coordinates of the area to be trimmed in the microstructure, achieve automatic positioning according to the center coordinates of the laser spot, plan the processing route according to the dimensions measured of the microstructure, and realize automatic laser trimming.

Images