Ultra-miniature inductor high-precision alignment device and control method thereof

Abstract

The application relates to the technical field of inductance alignment control, in particular to a super-mini inductance high-precision alignment device and a control method thereof. The method comprises the following steps: images of a substrate and a mask plate are collected from preset angles respectively, and are recorded as collected images; for any collected image and a standard template image thereof, feature coefficients in horizontal and vertical directions of the collected image and the standard template image thereof are obtained, and the collected image is divided into small blocks; adjustment coefficients of filter processing in the small blocks of the collected image are obtained, the size of a filter window in the small blocks of the collected image is obtained, and filter processing is performed on the collected image; the alignment of the substrate and the mask plate is measured, if the alignment does not meet the standard, the relative positions of the substrate and the mask plate are adjusted until the alignment meets the standard. The application aims to improve the accuracy of subsequent alignment mark position matching by accurately preprocessing the substrate and the mask plate, and to improve the high-precision alignment control precision of the super-mini inductance.

Description

Technical Field

[0001] The present application relates to the technical field of inductor alignment control, and in particular to an ultra-micro inductor high-precision alignment device and a control method thereof. Background Art

[0002] High-precision alignment plays a vital role in the production process of ultra-micro inductors. Due to their compact size and excellent performance, ultra-micro inductors have been widely used in cutting-edge electronic fields such as 5G communications, the Internet of Things, and wearable devices. Taking the RF module of a 5G base station as an example, the ultra-micro inductor needs to be precisely matched with the circuit parameters, which places extremely high demands on the alignment accuracy of each layer of pattern during its production. Through high-precision alignment, it is possible to ensure that the key dimensions of the inductor, such as the number of turns, line width, and line spacing, are highly consistent with the design values, thereby ensuring the stability and accuracy of electrical properties such as inductance and Q value, laying the foundation for improving the performance and reliability of the entire electronic equipment.

[0003] However, in the actual operation of high-precision printing alignment of inductors, there are various interference factors, resulting in varying printing accuracy. The complex electromagnetic environment in the production workshop can interfere with the sensors and control systems of the alignment device, resulting in large errors in the judgment of position offset during the printing alignment process. Due to the complex and changeable electromagnetic environment in actual production, the characteristics of the alignment process affected by local electromagnetic interference vary significantly. Traditional pre-processing methods are difficult to avoid the differences in the distribution of noise and stripes caused by electromagnetic interference, which in turn leads to large deviations in the printing alignment control during the preparation of ultra-micro inductors, making it impossible to achieve the goal of precise alignment. Summary of the Invention

[0004] In view of the above, it is necessary to provide a high-precision alignment device for ultra-micro inductors and a control method thereof. Compared with the traditional high-precision alignment device for ultra-micro inductors and a control method thereof, the accuracy of subsequent alignment mark position matching is improved by accurately pre-processing the substrate and the mask, thereby improving the high-precision alignment control accuracy of the ultra-micro inductors:

[0005] In a first aspect, an embodiment of the present application provides a method for controlling an ultra-micro inductor high-precision alignment device, the method comprising the following steps:

[0006] Collect images of the substrate and the mask from preset angles, record them as collected images, and obtain template images of each collected image;

[0007] For any acquired image and its template image, the data distribution of the template image's mapping results in the horizontal and vertical directions is used to obtain dynamic sliding windows that divide the acquired image and its template image's mapping results in the horizontal and vertical directions. By evaluating the data differences within each sliding window of each mapping result between the acquired image and its template image under each dynamic sliding window division result, the characteristic coefficients of the acquired image and its template image in the horizontal and vertical directions are obtained, which are used to divide the acquired image and its template image into small blocks.

[0008] Comparing the data probability distribution within each small block of the acquired image with the data probability distribution within the small block at the same position in the template image, and combining the distribution of the corner points within each small block of the acquired image to obtain the adjustment coefficient of the filtering processing within each small block of the acquired image, and filtering the acquired image in combination with the characteristic coefficient;

[0009] By comparing the filtered substrate images and mask images at the same angle, the alignment of the substrate and the mask is measured. If it does not meet the standard, the relative position of the substrate and the mask is adjusted until it meets the standard.

[0010] In one embodiment, the process of obtaining the dynamic sliding windows divided by the mapping results of the captured image and its template image in the horizontal and vertical directions is as follows:

[0011] The collected image and its template image are mapped horizontally and vertically using the Radon algorithm to obtain a horizontal and vertical mapping sequence;

[0012] The number of data between any two adjacent extreme values ​​in each mapping sequence of the template image of the acquired image is counted and duplicated, and the deduplication result is used as each dynamic sliding window for dividing each mapping sequence of each acquired image and its template image.

[0013] In one embodiment, the process of obtaining the characteristic coefficient is as follows:

[0014] Where, represents the feature difference value of the kth mapping sequence between the captured image and its template image under the jth dynamic sliding window; Indicates the length of the k-th mapping sequence of the acquired image; represents the data difference between the captured image and its template image in the i-th sliding window of the k-th mapping sequence under the j-th dynamic sliding window; Represents the discreteness of the data in the i-th sliding window of the k-th mapping sequence of the image collected under the j-th dynamic sliding window; represents the sum of the discrete degrees of the data in all sliding windows of the kth mapping sequence of the image collected under the jth dynamic sliding window;

[0015] The characteristic coefficients are screened from all dynamic sliding windows by collecting the distribution of characteristic difference values ​​of each mapping sequence between the image and its template image under all dynamic sliding windows.

[0016] In one embodiment, the method for screening the characteristic coefficient is: counting the maximum value of the characteristic difference values ​​of any mapping sequence between the acquired image and its template image under all dynamic sliding windows; if any mapping sequence is a horizontal mapping sequence, the dynamic sliding window corresponding to the maximum value is used as the characteristic coefficient of the acquired image and its template image in the horizontal direction; otherwise, it is used as the characteristic coefficient of the acquired image and its template image in the vertical direction.

[0017] In one embodiment, the method of dividing the captured image and its template image into small blocks is:

[0018] By collecting the characteristic coefficients of the image and its template image in the horizontal and vertical directions, the collected image is divided equally in the horizontal and vertical directions to obtain small blocks, and the template image of the collected image is divided equally in the horizontal and vertical directions to obtain small blocks.

[0019] In one embodiment, the process of obtaining the adjustment coefficient is as follows:

[0020] Perform probability statistics on the grayscale values ​​of the pixels in each small block, and fit a probability distribution curve based on the probability of the grayscale values ​​of the pixels in each small block; obtain the KL divergence value of the probability distribution curve between each small block of the acquired image and the small block at the same position in its template image;

[0021] Calculate the average distance between any two corner points in each small block; calculate the proportion of corner points in all pixels in each small block;

[0022] The adjustment coefficient can be further obtained through the KL divergence value, the mean and the proportion.

[0023] In one embodiment, the adjustment coefficient is the product of the KL divergence value, the mean value, and the proportion.

[0024] In one embodiment, the process of filtering the acquired image is:

[0025] The product of the characteristic coefficients in the horizontal and vertical directions of the acquired image and the normalized value of the adjustment coefficient in each small block is rounded up as the size of the filter window in the horizontal and vertical directions in each small block, and the acquired image is filtered.

[0026] In one embodiment, the process of measuring whether the alignment between the substrate and the mask meets the standard is as follows:

[0027] A template matching algorithm is used to identify each alignment mark image in each captured image. For any alignment mark on the substrate, the similarity score between the alignment mark image of the any alignment mark at each angle and the alignment mark image at the corresponding position on the mask is obtained, and recorded as each similarity score of the any alignment mark. The average value of all similarity scores of the any alignment mark is calculated. If the average value of all alignment marks on the substrate is greater than a preset threshold, it is determined that the alignment between the substrate and the mask is qualified; otherwise, it is determined that the alignment is unqualified.

[0028] In a second aspect, an embodiment of the present application further provides an ultra-micro inductor high-precision alignment device, wherein the device comprises: a visual positioning system, an image processing and control system;

[0029] The visual positioning system is used to collect images of the substrate and the mask from preset angles, which are recorded as collected images, and obtain template images of each collected image;

[0030] An image processing and control system, for obtaining, for any acquired image and its template image, dynamic sliding windows that divide the mapping results of the acquired image and its template image in the horizontal and vertical directions based on the data distribution of the mapping results of the template image in the horizontal and vertical directions, and obtaining characteristic coefficients of the acquired image and its template image in the horizontal and vertical directions by evaluating the difference in data within each sliding window of each mapping result between the acquired image and its template image under each dynamic sliding window division result, for use in dividing the acquired image and its template image into small blocks;

[0031] Comparing the data probability distribution within each small block of the acquired image with the data probability distribution within the small block at the same position in the template image, and combining the distribution of the corner points within each small block of the acquired image to obtain the adjustment coefficient of the filtering processing within each small block of the acquired image, and filtering the acquired image in combination with the characteristic coefficient;

[0032] By comparing the filtered substrate images and mask images at the same angle, the alignment of the substrate and the mask is measured. If it does not meet the standard, the relative position of the substrate and the mask is adjusted until it meets the standard.

[0033] This application has at least the following beneficial effects:

[0034] The present application takes into account that during the high-precision printing alignment process of the ultra-micro inductor preparation process, due to the differences in the influence of electromagnetic interference, large calculation errors of the alignment deviation may be caused. Therefore, first, images of the installed substrate and mask are captured in different directions, and the local data distribution differences of the captured images caused by electromagnetic interference are compared and analyzed based on the data distribution characteristics of the template image. The data distribution differences in the horizontal and vertical directions are comprehensively analyzed to accurately determine the local area range affected by electromagnetic interference in the captured image, and based on the data change characteristics caused by noise and stripes in the local area, the local influence characteristics of electromagnetic interference are analyzed. Based on the analysis results, the preprocessing is optimized and adjusted to accurately retain the characteristics of the alignment mark position. The beneficial effect is that the local characteristics of electromagnetic interference are fully considered, the substrate and mask are accurately preprocessed, and the accuracy of subsequent matching of the alignment mark position is improved, thereby improving the high-precision alignment control accuracy of the ultra-micro inductor. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0036] Figure 1 A block diagram of an ultra-micro inductor high-precision alignment device provided in one embodiment of the present application;

[0037] Figure 2 A flowchart of the steps of a control method for an ultra-micro inductor high-precision alignment device provided in one embodiment of the present application;

[0038] Figure 3 Schematic diagram of the alignment adjustment process between the substrate and the mask. DETAILED DESCRIPTION

[0039] In the description of the embodiments of this application, words such as "exemplary," "or," and "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary," "or," and "for example" is intended to present the relevant concepts in a concrete manner.

[0040] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application relates. The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. It should be understood that, unless otherwise indicated, " / " represents or.

[0041] It should also be noted that the terms "first" and "second" in this application are used to distinguish similar objects, rather than to describe a specific order or sequence.

[0042] The components of the ultra-micro inductor high-precision alignment device provided by the present application are described in detail below with reference to the accompanying drawings.

[0043] See also Figure 1 , which shows a block diagram of an ultra-micro inductor high-precision alignment device provided by an embodiment of the present application, the ultra-micro inductor high-precision alignment device includes: a visual positioning system, a high-precision motion platform, an image processing and control system, and a lighting system;

[0044] The visual positioning system is used to capture images of the substrate and mask from preset angles, referred to as captured images, and to obtain a template image of each captured image. Specifically, a high-resolution industrial camera is used to clearly capture the alignment marks on the substrate and mask. The high-resolution industrial camera has a resolution of up to 5000 × 5000 pixels and a frame rate of 60 fps, enabling rapid acquisition of image information.

[0045] The high-precision motion platform consists of three-axis linear modules (X, Y, and Z), with positioning accuracy of up to ±0.5μm for each axis. Ball screw drive and linear guide rails ensure smooth and high-precision motion. The drive motor is a servo motor with high response speed and precise torque control. The high-precision motion platform is responsible for carrying the substrate and mask, and accurately adjusts the relative position of the substrate and mask based on information fed back by the image processing and control system to achieve high-precision alignment.

[0046] An image processing and control system, for obtaining, for any acquired image and its template image, dynamic sliding windows that divide the mapping results of the acquired image and its template image in the horizontal and vertical directions based on the data distribution of the mapping results of the template image in the horizontal and vertical directions, and obtaining characteristic coefficients of the acquired image and its template image in the horizontal and vertical directions by evaluating the difference in data within each sliding window of each mapping result between the acquired image and its template image under each dynamic sliding window division result, for use in dividing the acquired image and its template image into small blocks;

[0047] Comparing the data probability distribution within each small block of the acquired image with the data probability distribution within the small block at the same position in the template image, and combining the distribution of the corner points within each small block of the acquired image to obtain the adjustment coefficient of the filtering processing within each small block of the acquired image, and filtering the acquired image in combination with the characteristic coefficient;

[0048] By comparing the filtered substrate images and mask images at the same angle, the alignment of the substrate and the mask is measured. If the alignment does not meet the standard, the alignment deviation of the substrate and the mask is obtained and used to adjust the relative position of the substrate and the mask until the alignment meets the standard. The alignment deviation of the substrate and the mask is calculated through image recognition algorithm and control algorithm, and control instructions are sent to the high-precision motion platform based on the deviation to realize automatic alignment operation. Specifically, the alignment deviation of the substrate and the mask can be calculated through template matching algorithm and PID control (proportional-integral-derivative control) algorithm.

[0049] The lighting system uses a ring-shaped LED light source with luminous intensity adjustable within the range of 500~5000 lux, ensuring sufficient and uniform light in the shooting area, improving the quality of images collected by the visual positioning system, and facilitating accurate identification of alignment marks.

[0050] See also Figure 2 , which shows a flowchart of a control method for a high-precision alignment device for an ultra-micro inductor according to an embodiment of the present application, the method comprises the following steps:

[0051] Step 1: Capture images of the substrate and the mask from preset angles, record them as captured images, and obtain template images of each captured image.

[0052] A substrate with a clean surface and meeting required flatness is placed on the carrier platform of the high-precision motion platform of the ultra-micro inductor high-precision alignment device. Vacuum suction ensures that the substrate will not move during subsequent operations. Simultaneously, a mask with a designed pattern is mounted on the corresponding fixture. At this point, the device's lighting system is activated, adjusting the luminous intensity to 2000 lux to provide suitable light for the device's visual positioning system to capture images. Specifically, the high-resolution industrial camera in the visual positioning system captures images of the substrate and mask from various preset angles. The image processing and control system in the ultra-micro inductor high-precision alignment device evaluates the alignment between the substrate and mask based on the captured images. However, given that the captured images may contain noise and fringes during the actual acquisition process, which can affect image quality, reduce the accuracy of subsequent matching and recognition of the alignment marks on the substrate and mask, and thus affect the accuracy of the subsequent alignment assessment, filtering is required.

[0053] In this embodiment, the number of preset angles is 4, and the preset angles are 、 、 、 , where at each preset angle, the substrate and the mask each correspond to a template image.

[0054] Step 2: By comparing the acquired substrate image with its template image, and the mask image with its template image, the acquired substrate image and mask image are filtered, and the alignment of the substrate and mask is measured by the filtered substrate image and mask image. If the alignment does not meet the standard, the relative position of the substrate and mask is adjusted until the alignment meets the standard.

[0055] Since different batches of materials are subject to different environmental interference factors during the alignment and matching process during the actual acquisition process, the actual acquired images are affected by different electromagnetic interference characteristics in different directions. As a result, it is difficult to avoid the impact caused by the difference in electromagnetic interference during the preprocessing process using traditional image preprocessing methods. Therefore, in this application, based on the differences in the impact of electromagnetic interference during the actual production of RF inductors, the interference features such as noise and stripes in the alignment and matching process of the substrate and mask are accurately analyzed, the preprocessing process of the acquired images is optimized and adjusted, the accuracy of the alignment and matching recognition is improved, and high-precision alignment of the substrate and mask is achieved; the specific analysis and processing process is as follows:

[0056] (1) For any acquisition image and its template image, the data distribution of the mapping results of the template image in the horizontal and vertical directions is used to obtain the dynamic sliding windows that divide the mapping results of the acquisition image and its template image in the horizontal and vertical directions. By evaluating the differences in the data in each sliding window under the division results of each dynamic sliding window, the characteristic coefficients of the acquisition image and its template image in the horizontal and vertical directions are obtained.

[0057] During the matching and recognition process of radio frequency inductive printing alignment, due to the influence of electromagnetic interference, different degrees of noise and stripe interference appear in the collected image, resulting in blurred edges of the alignment mark, affecting the accuracy of printed alignment mark recognition; therefore, the noise characteristics caused by noise and stripes in the collected image under the influence of electromagnetic interference are considered, among which, noise is a randomly isolated bright or dark spot, which will destroy the edge line of the alignment mark and cause gaps or burrs in the originally continuous contour; stripes are periodic and directional light and dark stripes, covering the entire image or a large area, reducing the signal-to-noise ratio, making the grayscale of the alignment mark area no longer uniform, thereby blurring the true boundary of the alignment mark; in short, both noise and stripes will make the edge of the alignment mark blurred, making it difficult to lock the position of the alignment mark, and ultimately reducing the accuracy of printing alignment.

[0058] Based on the above analysis, taking the Lth captured image of the substrate and its template image as an example, where the template image of the Lth captured image is in the same direction as the shooting direction of the Lth captured image and has a standard alignment area, the Lth captured image and its template image are respectively used as input, and the Radon algorithm is used to perform horizontal and vertical mapping. The mapping results are used as the horizontal mapping sequence and vertical mapping sequence of the Lth captured image, and the horizontal mapping sequence and vertical mapping sequence of the template image of the Lth captured image. The Radon algorithm is a well-known technology and will not be described in detail in this application.

[0059] Furthermore, the local distribution characteristics of the data in each direction between the Lth acquisition image and its template image are compared. Specifically, taking the Lth acquisition image and the kth mapping sequence of its template image as an example, where the kth mapping sequence refers to the mapping sequence in the horizontal direction, the number of data between all two adjacent extreme values ​​in the kth mapping sequence of the template image of the Lth acquisition image is counted, and duplicates are removed. The deduplication results are used as the dynamic sliding windows that divide the Lth acquisition image and the kth mapping sequence of its template image. Taking the jth dynamic sliding window as an example, by evaluating the difference in data within each sliding window of the kth mapping sequence between the Lth acquisition image and its template image under the jth dynamic sliding window division result, the characteristic difference value of the kth mapping sequence between the Lth acquisition image and its template image under the jth dynamic sliding window is obtained. The expression is:

[0060] Where, represents the feature difference value of the kth mapping sequence between the captured image and its template image under the jth dynamic sliding window; Indicates the length of the k-th mapping sequence of the acquired image; represents the data difference between the captured image and its template image in the i-th sliding window of the k-th mapping sequence under the j-th dynamic sliding window; Represents the discreteness of the data in the i-th sliding window of the k-th mapping sequence of the image collected under the j-th dynamic sliding window; Represents the sum of the discrete degrees of the data within all sliding windows of the kth mapping sequence of the image captured under the jth dynamic sliding window. If the dynamic sliding window exceeds the range of the mapping sequence, a data filling method is used to fill the missing data. This embodiment uses the mean filling method, which is a well-known technique and will not be described in detail in this application. Implementers may choose other existing feasible data filling methods. The sliding step size of the sliding window is 1, and implementers may set the sliding step size of the sliding window based on actual conditions.

[0061] In this embodiment, the method for calculating the data difference amount between two sliding windows is: arrange the data in each sliding window according to the order in the mapping sequence to form a data sequence of each sliding window, and use the Manhattan distance between the data sequences of the two sliding windows as the data difference amount within the two sliding windows. Manhattan distance is a well-known technology and will not be described in detail in this application. As other implementation methods, on the basis of being able to measure the difference amount of data in the two sliding windows, the implementer may adopt other existing technologies, such as the Euclidean distance between the data sequences of the two sliding windows, etc., and this application does not impose any special restrictions.

[0062] In this embodiment, the discreteness is the variance. As other implementation methods, on the basis of being able to measure the degree of uneven distribution of data in each sliding window, the implementer may adopt other existing technologies, such as standard deviation, coefficient of variation, etc., and this application does not impose any special restrictions.

[0063] It should be noted that: The larger the calculated result of , the more likely it is that the data in the ith sliding window of the kth mapping sequence of the Lth acquisition image under the jth dynamic sliding window, within the data range determined by the template image, has abnormal fluctuations that exceed the overall level due to electromagnetic interference. The larger the calculated feature difference value, the more complete and significant the electromagnetic interference features are captured under the jth dynamic sliding window, the higher the discrimination and credibility of the comparative analysis, and the more accurate the comparative analysis results.

[0064] To accurately identify the impact of noise and streaks caused by electromagnetic interference on local data distribution in the acquired image, the maximum value of the feature difference values ​​in the kth mapping sequence between the Lth acquired image and its template image under all dynamic sliding windows is calculated. Since the kth mapping sequence is a horizontal mapping sequence, the dynamic sliding window corresponding to the maximum value is used as the horizontal feature coefficient of the Lth acquired image and its template image. According to the method for obtaining the horizontal feature coefficient of the Lth acquired image and its template image, the horizontal and vertical feature coefficients of each acquired image and its template image are obtained.

[0065] (2) Divide the acquired image and its template image into small blocks; compare the data probability distribution within each small block of the acquired image with the small block at the same position in its template image, and obtain the adjustment coefficient of the filtering processing within each small block of the acquired image in combination with the distribution of the corner points within each small block of the acquired image.

[0066] The L-th acquired image is divided equally in the horizontal and vertical directions according to the characteristic coefficients of the L-th acquired image, respectively, to obtain small blocks; the template image of the L-th acquired image is divided equally in the horizontal and vertical directions according to the characteristic coefficients of the L-th acquired image, respectively, to obtain small blocks.

[0067] The process of calculating the feature difference value and then dividing the Lth acquired image and its template image into small blocks takes into account the different impacts of electromagnetic interference on different areas during the actual acquisition process. Therefore, based on the premise that the data distribution characteristics of the Lth acquired image and its template image are similar, a comparative analysis of the local data distribution differences of the Lth acquired image is performed according to the distribution intervals of the local areas of the data in the template image, and then the horizontal and vertical area ranges that can reflect the significant data distribution differences are determined, which facilitates the subsequent accurate analysis of the local electromagnetic interference characteristics of the Lth acquired image.

[0068] Furthermore, probability statistics are performed on the grayscale values ​​of the pixels within each small block, and a probability distribution curve is fitted based on the probability of the grayscale values ​​of the pixels within each small block, where the horizontal axis represents the grayscale value of the pixel and the vertical axis represents the probability of the grayscale value of the pixel. The KL divergence value of the probability distribution curve between each small block of the Lth acquired image and the small block at the same position in its template image is used as the first eigenvalue of each small block of the Lth acquired image; the larger the first eigenvalue, the greater the deviation of the data distribution within each small block due to electromagnetic interference from the data distribution at the same position in the template image after the distribution characteristics of the mapped data are divided based on the horizontal and vertical directions. The calculation of the KL divergence value is a well-known technique and will not be described in detail in this application.

[0069] In this embodiment, the least squares method is used to obtain the probability distribution curve. The least squares method is a well-known technology and will not be described in detail in this application. Based on the obtainable probability distribution curve, the implementer can select other existing feasible algorithms.

[0070] Furthermore, considering the interference effects of possible noise distribution within each small block, the grayscale values ​​of all pixels within each small block are used as input, and a corner detection algorithm is used to obtain the corner points within each small block. The mean distance between any two corner points within each small block is calculated. A larger mean indicates a more dispersed distribution of corner points and a larger range of local interference. The proportion of corner points among all pixels within each small block is calculated. A larger proportion indicates a greater distribution density of corner points within the local area and a larger range of local electromagnetic interference. The product of the mean and the proportion within each small block is used as the second eigenvalue of each small block. A larger second eigenvalue indicates a more significant electromagnetic interference feature within each small block.

[0071] Based on the above analysis, the adjustment coefficient for the filtering process within each small block of the Lth acquired image is obtained using the first eigenvalue and the second eigenvalue of each small block of the Lth acquired image. Specifically, the product of the first eigenvalue and the second eigenvalue of each small block of the Lth acquired image is used as the adjustment coefficient for the filtering process within each small block of the Lth acquired image. The larger the calculated adjustment coefficient, the greater the deviation of the data distribution within each small block from the data distribution at the same position in the template image, and the larger the range of electromagnetic interference. Therefore, a larger filtering window needs to be set to reduce the impact of electromagnetic interference.

[0072] (3) Filter the captured image; by comparing the filtered substrate image and mask image at the same angle, measure the alignment of the substrate and mask. If it does not meet the standard, adjust the relative position of the substrate and mask until it meets the standard.

[0073] Based on the above analysis, the calculated adjustment coefficients of all small blocks of all acquired images are used as input, and a normalization method is used to obtain the normalized value of the adjustment coefficient of each small block of each acquired image, which is used to optimize the filtering process of each small block. Specifically, in order to better preserve the edge features of the alignment marks on the substrate and the mask, the acquired images are filtered using a median filtering algorithm. The size of the filtering window in each small block of each acquired image is obtained by combining the adjustment coefficient of the filtering process in each small block of each acquired image with the characteristic coefficients in the horizontal and vertical directions of each acquired image. The expression is:

[0074] , Where, 、 They represent the horizontal and vertical sizes of the filter window in the vth small block of the Lth acquired image respectively; Indicates rounding up operation; 、 Respectively represent the characteristic coefficients in the horizontal and vertical directions of the L-th acquired image; represents the normalized value of the adjustment coefficient of the filtering process in the vth small block of the Lth acquired image. The size of the filtering window in the vth small block of the Lth acquired image is , set the minimum filter window size to , if exists or In the case of , the size of the filter window in the vth small block of the Lth captured image is set to The minimum filter window size is preset manually and can be set by the implementer. This application does not impose any special restrictions. The median filter algorithm is a well-known technology and the specific implementation steps will not be described in detail in this application.

[0075] In this embodiment, the Min-Max normalization method is used to obtain the normalized value of the adjustment coefficient. The Min-Max normalization method is a well-known technology and will not be described in detail in this application.

[0076] Each image is filtered based on a median filtering algorithm, wherein the filtering window in the vth small block of the Lth acquired image is used as the filtering window when filtering the data in the vth small block of the Lth acquired image using the median filtering algorithm.

[0077] Furthermore, a template matching algorithm is used to identify each alignment mark image in each image. For any alignment mark on the substrate, the similarity score between the alignment mark image of any alignment mark at each angle and the alignment mark image of the corresponding position on the mask is obtained, and recorded as each similarity score of any alignment mark. The average value of all similarity scores of any alignment mark is calculated. If the average values ​​of all alignment marks on the substrate are greater than the preset threshold, it is determined that the alignment between the substrate and the mask meets the standard; otherwise, it is determined that the alignment does not meet the standard. Among them, the template matching algorithm is a well-known technology and will not be described in detail in this application. The similarity score between two alignment mark images is calculated as follows: the normalized value of the reciprocal of the sum of the absolute differences between the two alignment mark images. Among them, the Sigmoid function is used to obtain the normalized value of the reciprocal of the sum of the absolute differences. The Sigmoid function is a well-known technology and will not be described in detail in this application.

[0078] In this embodiment, the value of the preset threshold is 0.8. The value of the preset threshold is preset manually and the implementer can set it according to actual conditions. This application does not impose any special restrictions.

[0079] If the alignment does not meet the standards, the alignment deviation of the substrate and the mask is calculated through the image processing and control system, and a fine-tuning instruction is sent to the high-precision motion platform. Specifically, for example, if there is a deviation of 1μm in the X-axis direction, the image processing and control system controls the X-axis servo motor to drive the ball screw to rotate, so that the high-precision motion platform moves 1μm in the X-axis direction, thereby adjusting the relative position of the substrate and the mask. During the fine-tuning process, the visual positioning system continuously collects images and performs matching recognition analysis until the alignment of the substrate and the mask meets the standards. In the above process, the moving speed of the high-precision motion platform is controlled within 10μm / s to ensure the accuracy of the fine-tuning. The schematic diagram of the alignment adjustment process of the substrate and the mask is as follows Figure 3 shown.

[0080] Step 3: After the substrate and the mask are precisely aligned, exposure, curing and inspection operations are performed.

[0081] Next, the exposure operation is performed. Specifically, after precise alignment, the high-precision motion platform remains stable. The substrate coated with photosensitive slurry is placed firmly against the mask, and then exposed using an exposure machine. The exposure machine uses an intensity of 100mJ / cm² and an exposure time of 5s. This causes the photosensitive slurry to undergo a photochemical reaction under light, transferring the pattern on the mask to the substrate.

[0082] Furthermore, the curing and inspection operations are completed, specifically: the exposed substrate is placed into the curing equipment and cured at a temperature of 150°C for 10 minutes to solidify the photosensitive paste into shape. Subsequently, the printed pattern is inspected by automatic optical inspection equipment and compared with the design pattern to check whether there are problems such as pattern missing, deformation, and excessive alignment deviation. If a problem is found, the relevant data is recorded and the parameters of the printing equipment are adjusted accordingly to optimize the printing alignment effect in the next batch of production. For example, if it is detected that the pattern has an obvious problem of excessive alignment deviation, it may be that the positioning system of the printing equipment has a slight offset. At this time, the X / Y axis alignment compensation parameters of the printing machine can be adjusted. For example, the original X-axis compensation value is 0.02mm. According to the direction and size of the deviation, it can be adjusted to 0.03mm or 0.01mm. By accurately correcting the compensation amount of the mechanical positioning, the offset of the pattern in the horizontal and vertical directions can be reduced.

[0083] To sum up, the present application takes into account that in the high-precision printing alignment process of the ultra-micro inductor preparation process, due to the difference in the influence of electromagnetic interference, a large calculation error of the alignment deviation may be caused. Therefore, the installed substrate and mask are firstly captured in different directions, and the local data distribution differences of the captured image caused by the influence of electromagnetic interference are compared and analyzed based on the data distribution characteristics of the template image. The data distribution differences in the horizontal and vertical directions are comprehensively analyzed to accurately determine the local area range affected by electromagnetic interference in the captured image, and based on the data change characteristics caused by the noise and stripes in the local area, the local influence characteristics of the electromagnetic interference are analyzed. Based on the analysis results, the pre-processing is optimized and adjusted to accurately retain the characteristics of the alignment mark position. The beneficial effect is to fully consider the local characteristics of the electromagnetic interference, accurately pre-process the substrate and the mask, improve the accuracy of the subsequent matching of the alignment mark position, and thereby improve the high-precision alignment control accuracy of the ultra-micro inductor.

[0084] The flowcharts and block diagrams in the accompanying drawings show the possible implementation architectures, functions and operations of the systems, methods and computer program products according to the embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, program segment or part of the code, and the module, program segment or part of the code contains one or more executable instructions for implementing the specified logical functions. In some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, or they can sometimes be executed in the opposite order, which can depend on the functions involved. In the descriptions corresponding to the flowcharts and block diagrams in the accompanying drawings, the operations or steps corresponding to different boxes can also occur in an order different from that disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps can actually be executed substantially in parallel, or they can sometimes be executed in the opposite order, which can depend on the functions involved. Each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented by a dedicated hardware-based system that performs the specified function or action, or may be implemented by a combination of dedicated hardware and computer instructions.

[0085] It is obvious to those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and that the present application can be implemented in other specific forms without departing from the basic characteristics of the present application. Therefore, from all perspectives, the above embodiments of the present application should be regarded as exemplary and non-restrictive.

Claims

1. A control method for a high-precision alignment device for ultra-micro inductors, characterized in that: The method comprises the following steps: Collect images of the substrate and the mask from preset angles, record them as collected images, and obtain template images of each collected image; For any acquired image and its template image, the data distribution of the template image's mapping results in the horizontal and vertical directions is used to obtain dynamic sliding windows that divide the acquired image and its template image's mapping results in the horizontal and vertical directions. By evaluating the data differences within each sliding window of each mapping result between the acquired image and its template image under each dynamic sliding window division result, the characteristic coefficients of the acquired image and its template image in the horizontal and vertical directions are obtained, which are used to divide the acquired image and its template image into small blocks. Comparing the data probability distribution within each small block of the acquired image with the data probability distribution within the small block at the same position in the template image, and combining the distribution of the corner points within each small block of the acquired image to obtain the adjustment coefficient of the filtering processing within each small block of the acquired image, and filtering the acquired image in combination with the characteristic coefficient; By comparing the filtered substrate images and mask images at the same angle, the alignment between the substrate and the mask is measured. If the alignment does not meet the standard, the relative position of the substrate and the mask is adjusted until it meets the standard. The process of obtaining the adjustment coefficient is as follows: Perform probability statistics on the grayscale values ​​of the pixels in each small block, and fit a probability distribution curve based on the probability of the grayscale values ​​of the pixels in each small block; obtain the KL divergence value of the probability distribution curve between each small block of the acquired image and the small block at the same position in its template image; Calculate the average distance between any two corner points in each small block; calculate the proportion of corner points in all pixels in each small block; The adjustment coefficient can be further obtained through the KL divergence value, the mean and the proportion; The adjustment coefficient is the product of the KL divergence value, the mean value, and the proportion.

2. The control method of the ultra-micro inductor high-precision alignment device according to claim 1, wherein: The process of obtaining the dynamic sliding windows divided by the mapping results of the acquisition image and its template image in the horizontal and vertical directions is as follows: The collected image and its template image are mapped horizontally and vertically using the Radon algorithm to obtain a horizontal and vertical mapping sequence; The number of data between any two adjacent extreme values ​​in each mapping sequence of the template image of the acquired image is counted and duplicated, and the deduplication result is used as each dynamic sliding window for dividing each mapping sequence of each acquired image and its template image.

3. The control method of the ultra-micro inductor high-precision alignment device according to claim 2, wherein: The process of obtaining the characteristic coefficient is as follows: Where, represents the feature difference value of the kth mapping sequence between the captured image and its template image under the jth dynamic sliding window; Indicates the length of the k-th mapping sequence of the acquired image; represents the data difference between the captured image and its template image in the i-th sliding window of the k-th mapping sequence under the j-th dynamic sliding window; Represents the discreteness of the data in the i-th sliding window of the k-th mapping sequence of the image collected under the j-th dynamic sliding window; represents the sum of the discrete degrees of the data in all sliding windows of the kth mapping sequence of the image collected under the jth dynamic sliding window; The characteristic coefficients are screened from all dynamic sliding windows by collecting the distribution of characteristic difference values ​​of each mapping sequence between the image and its template image under all dynamic sliding windows.

4. The control method of the ultra-micro inductor high-precision alignment device according to claim 3, wherein: The method for screening the characteristic coefficients is as follows: the maximum value of the characteristic difference values ​​of any mapping sequence between the acquired image and its template image under all dynamic sliding windows is counted; if any mapping sequence is a horizontal mapping sequence, the dynamic sliding window corresponding to the maximum value is used as the characteristic coefficient of the acquired image and its template image in the horizontal direction; otherwise, it is used as the characteristic coefficient of the acquired image and its template image in the vertical direction.

5. The control method of the ultra-micro inductor high-precision alignment device according to claim 1, wherein: The method for dividing the collected image and its template image into small blocks is: By collecting the characteristic coefficients of the image and its template image in the horizontal and vertical directions, the collected image is divided equally in the horizontal and vertical directions to obtain small blocks, and the template image of the collected image is divided equally in the horizontal and vertical directions to obtain small blocks.

6. The control method of the ultra-micro inductor high-precision alignment device according to claim 1, wherein: The process of filtering the collected image is as follows: The product of the characteristic coefficients in the horizontal and vertical directions of the acquired image and the normalized value of the adjustment coefficient in each small block is rounded up as the size of the filter window in the horizontal and vertical directions in each small block, and the acquired image is filtered.

7. The control method of the ultra-micro inductor high-precision alignment device according to claim 1, wherein: The process of measuring the alignment between the substrate and the mask is as follows: A template matching algorithm is used to identify each alignment mark image in each captured image. For any alignment mark on the substrate, the similarity score between the alignment mark image of the any alignment mark at each angle and the alignment mark image at the corresponding position on the mask is obtained, and recorded as each similarity score of the any alignment mark. The average value of all similarity scores of the any alignment mark is calculated. If the average value of all alignment marks on the substrate is greater than a preset threshold, it is determined that the alignment between the substrate and the mask is qualified; otherwise, it is determined that the alignment is unqualified.

8. An ultra-micro inductor high-precision alignment device, using the control method of the ultra-micro inductor high-precision alignment device as claimed in claim 1, characterized in that: The device includes: a visual positioning system, an image processing and control system; The visual positioning system is used to collect images of the substrate and the mask from preset angles, which are recorded as collected images, and obtain template images of each collected image; An image processing and control system, for obtaining, for any acquired image and its template image, dynamic sliding windows that divide the mapping results of the acquired image and its template image in the horizontal and vertical directions based on the data distribution of the mapping results of the template image in the horizontal and vertical directions, and obtaining characteristic coefficients of the acquired image and its template image in the horizontal and vertical directions by evaluating the difference in data within each sliding window of each mapping result between the acquired image and its template image under each dynamic sliding window division result, for use in dividing the acquired image and its template image into small blocks; Comparing the data probability distribution within each small block of the acquired image with the data probability distribution within the small block at the same position in the template image, and combining the distribution of the corner points within each small block of the acquired image to obtain the adjustment coefficient of the filtering processing within each small block of the acquired image, and filtering the acquired image in combination with the characteristic coefficient; By comparing the filtered substrate images and mask images at the same angle, the alignment of the substrate and the mask is measured. If it does not meet the standard, the relative position of the substrate and the mask is adjusted until it meets the standard.

Images