A machine vision quality inspection method for semiconductor equipment component manufacturing

Abstract

This invention discloses a machine vision quality inspection method for semiconductor equipment component manufacturing, relating to the field of visual quality inspection technology. The method includes the following steps: converting component images to grayscale to obtain a component grayscale image; obtaining a background threshold based on images of a first number of normal components; obtaining a component threshold based on images of the first number of normal components; obtaining a region to be inspected based on the background threshold and the component grayscale image; performing binarization processing on the region to be inspected based on the component threshold to obtain a region binarization image; determining whether an abnormal region exists based on the region binarization image; if an abnormal region exists, obtaining the abnormal location, issuing a defect signal, and simultaneously displaying the abnormal location; if no abnormal region exists, issuing a defect-free signal. This invention addresses the problem in existing technologies where methods are cumbersome when dealing with multiple components, resulting in low efficiency in defect detection of semiconductor equipment components.

Description

Technical Field

[0001] This invention relates to the field of visual quality inspection technology, specifically to a machine vision quality inspection method for semiconductor equipment component manufacturing. Background Technology

[0002] As the foundation of modern information technology, the development of the semiconductor industry is highly dependent on the precision and reliability of manufacturing equipment. Semiconductor equipment itself is composed of a large number of high-precision, complex geometric components. The manufacturing quality of these components directly determines the performance, yield, stability and lifespan of the semiconductor equipment. Therefore, it is necessary to conduct quality inspection on the components of semiconductor equipment.

[0003] Quality inspection includes defect detection. However, semiconductor equipment components have complex structures and often contain holes. When inspecting through acquired images, these holes are easily misidentified as defects. Therefore, it is necessary to scan and compare each type of semiconductor equipment component. This method is computationally intensive and complex. For example, patent application CN115272302A discloses a method, equipment and system for component detection in images. This solution compares with the same component. When dealing with components with multiple structures, it is necessary to acquire images of each normal component, which is cumbersome. In other words, the existing technology is cumbersome when dealing with multiple components, resulting in low efficiency in defect detection of semiconductor equipment components. Summary of the Invention

[0004] This invention aims to at least partially solve one of the technical problems in the prior art by performing grayscale processing on component images to obtain component grayscale images; obtaining a background threshold based on images of a first number of normal components; obtaining a component threshold based on images of the first number of normal components; obtaining a region to be detected based on the background threshold and the component grayscale images; performing binarization processing on the region to be detected based on the component threshold to obtain a region binarization image; determining whether an abnormal region appears based on the region binarization image; if an abnormal region appears, obtaining the abnormal location, issuing a defect signal and displaying the abnormal location; if no abnormal region appears, issuing a defect-free signal. This solves the problem that the existing technology is cumbersome when dealing with multiple components, resulting in low efficiency in detecting component defects in semiconductor devices.

[0005] To achieve the above objectives, this application provides a machine vision quality inspection method for semiconductor equipment component manufacturing, comprising the following steps:

[0006] Acquire images of semiconductor equipment components and label them as component images;

[0007] The component images are converted to grayscale to obtain grayscale images of the components;

[0008] Obtain the background threshold based on the images of a first number of normal components;

[0009] Obtain component thresholds based on images of a first number of normal components;

[0010] The area to be detected is obtained based on the background threshold and the grayscale image of the component.

[0011] Binarization of the area to be detected is performed based on the component threshold to obtain a region binarization map;

[0012] Based on the region binarized map, determine whether an abnormal region exists. If it does, obtain the abnormal location, issue a defect signal, and display the abnormal location. If it does not exist, issue a defect-free signal.

[0013] Further, the grayscale processing of the component images to obtain component grayscale images includes the following sub-steps:

[0014] Obtain the R, G, and B channel values ​​of pixels in the component image, calculate the average of the R, G, and B channel values ​​of each pixel, and mark it as the component grayscale value; replace the R, G, and B channel values ​​of each pixel in the component image with the corresponding component grayscale value to obtain the component grayscale image.

[0015] Furthermore, obtaining the background threshold based on the images of a first number of normal components includes the following sub-steps:

[0016] The image of a normal component is processed into a grayscale image to obtain a normal grayscale image. The grayscale values ​​of pixels that are not part of the component in the normal grayscale image are then obtained and marked as background grayscale values.

[0017] Obtain the range of background grayscale values, and divide the range of background grayscale values ​​into N equal intervals, marking them as the background division range;

[0018] Calculate the frequency of the background grayscale value within each background segmentation range and mark it as the background grayscale frequency;

[0019] Establish a Cartesian coordinate system with the background grayscale value as the X-axis and the background grayscale frequency as the Y-axis, and label it as the background coordinate system.

[0020] Furthermore, obtaining the background threshold based on the images of the first number of normal components also includes the following sub-steps:

[0021] Obtain the median value of the background segment and mark it as the background median value;

[0022] The background median and the corresponding background grayscale frequency are used as the x and y coordinates of the coordinate points and marked as the background coordinate points;

[0023] Plot all background coordinate points into the background coordinate system to obtain a scatter plot, and label it as the background scatter plot;

[0024] The function obtained by fitting the background scatter plot is marked as the background fitting function.

[0025] Obtain the intersection points of the background fitting function and the X-axis of the background coordinate system, and mark them as background intersection points;

[0026] Obtain the maximum value of the x-coordinate among the background intersection points and mark it as the background threshold.

[0027] Furthermore, obtaining the component threshold based on the images of a first number of normal components includes the following sub-steps:

[0028] Obtain the grayscale values ​​of the pixels of the component parts in the normal grayscale image and mark them as component grayscale values;

[0029] Obtain the range of grayscale values ​​of the component, and divide the range of grayscale values ​​of the component into M equal intervals, marking them as the component division range;

[0030] Calculate the frequency of grayscale values ​​of each component within the defined range, and mark it as the grayscale frequency of the component;

[0031] A Cartesian coordinate system is established using the grayscale values ​​of the components as the X-axis and the grayscale frequencies of the components as the Y-axis. This system is then labeled as the component coordinate system.

[0032] Furthermore, obtaining the component threshold based on the images of the first number of normal components also includes the following sub-steps:

[0033] Obtain the median value of the component's segmentation range and mark it as the component's median value;

[0034] The median value of the component and the corresponding gray frequency of the component are used as the x and y coordinates of the coordinate points, and these are marked as component coordinate points.

[0035] Plot all component coordinate points into the component coordinate system to obtain a scatter plot, and label it as the component scatter plot;

[0036] The function obtained by fitting the scatter plot of the parts is marked as the part fitting function;

[0037] Obtain the intersection points of the component fitting function and the X-axis of the component coordinate system, and mark them as component intersection points;

[0038] Find the minimum x-coordinate of the intersection points of the components and mark it as the component threshold.

[0039] Furthermore, obtaining the region to be detected based on the background threshold and the grayscale image of the component includes the following sub-steps:

[0040] Set the grayscale values ​​in the grayscale image of the component that are less than or equal to the background threshold to 255, and set the grayscale values ​​in the grayscale image of the component that are greater than the background threshold to 0 to obtain the binarized image of the component.

[0041] Establish a Cartesian coordinate system, labeled as the position coordinate system, and place the binarized image of the component in the position coordinate system; obtain the largest area region composed of pixels with a grayscale value of 0 in the binarized image of the component, and label it as the initial region; obtain the position of the initial region in the position coordinate system, and label it as the initial region position;

[0042] Place the grayscale image of the component in the same coordinate system as the binary image of the component, obtain the region at the initial location in the grayscale image of the component, and mark it as the region to be detected.

[0043] Furthermore, the process of binarizing the region to be detected based on the component threshold to obtain a binarized map includes the following sub-steps:

[0044] Set the grayscale value of pixels in the detection area that are greater than or equal to the component threshold to 255, and set the grayscale value of pixels in the detection area that are less than the component threshold to 0 to obtain a region binarized image.

[0045] Furthermore, based on the region binarized map, it is determined whether an abnormal region exists. If it does, the abnormal location is obtained, a defect signal is issued, and the abnormal location is displayed simultaneously. If it does not exist, a defect-free signal is issued, including the following sub-steps:

[0046] The independent regions formed by pixels with a gray value of 0 within the region to be detected in the region binarized image are marked as the initial defect regions.

[0047] Obtain the minimum area of ​​the permissible defect portion of the semiconductor and mark it as the minimum area.

[0048] Find the pixel in the component image that corresponds to the minimum area and mark it as the threshold for the number of pixels;

[0049] Obtain the number of pixels in the initial defect region. If the number of pixels in the initial defect region is greater than the pixel count threshold, then mark the initial defect region as an abnormal region.

[0050] If the number of pixels in the initial defect area is less than or equal to the pixel count threshold, it indicates that no abnormal area has appeared.

[0051] Furthermore, obtaining the anomaly location includes the following sub-steps:

[0052] Obtain the maximum and minimum values ​​of the horizontal coordinate of the abnormal region, and mark them as the maximum horizontal value and the minimum horizontal value of the region, respectively;

[0053] Obtain the maximum and minimum values ​​of the vertical coordinate of the abnormal region, and mark them as the maximum and minimum values ​​of the region, respectively.

[0054] Calculate the difference between the horizontal maximum and horizontal minimum values ​​of the region, and mark it as the horizontal difference value of the region; calculate the difference between the vertical maximum and vertical minimum values ​​of the region, and mark it as the vertical difference value of the region.

[0055] Calculate the product of the longitudinal difference and the transverse difference of the region, and label it as the difference product;

[0056] Rotate the abnormal region 360° around any point within the region. When the product of the differences is minimized, mark the first coordinate point as the x-coordinate and y-coordinate of the region's smallest horizontal and vertical values ​​as a single coordinate point. Mark the second coordinate point as the x-coordinate and y-coordinate of the region's largest horizontal and vertical values ​​as a single coordinate point. Mark the third coordinate point as the x-coordinate and y-coordinate of the region's largest horizontal and vertical values ​​as a single coordinate point. Mark the fourth coordinate point as the x-coordinate and y-coordinate of the region's smallest horizontal and vertical values ​​as a single coordinate point. The rectangle formed by connecting the first, second, third, and fourth coordinate points in sequence is marked as the abnormal location.

[0057] The beneficial effects of this invention are as follows: This invention obtains a grayscale image of a component by converting the component image to grayscale; obtains a background threshold based on images of a first number of normal components; obtains a component threshold based on images of the first number of normal components; obtains the region to be detected based on the background threshold and the component grayscale image; performs binarization processing on the region to be detected based on the component threshold to obtain a region binarization image; and determines whether an abnormal region exists based on the region binarization image. If an abnormal region exists, the abnormal location is obtained, a defect signal is issued, and the abnormal location is displayed simultaneously; if no abnormal region exists, a defect signal is issued. The advantage is that it can handle defect detection of various components and improve the efficiency of component defect detection in semiconductor equipment.

[0058] This invention obtains the area to be detected based on a background threshold and a grayscale image of the component. Its advantage is that it can not only filter the parts in the image, but also eliminate the interference of holes on the threshold defect identification of the component. This not only improves the accuracy of component defect detection in semiconductor equipment, but also improves the efficiency of component defect detection in semiconductor equipment. Attached Figure Description

[0059] Figure 1 This is a flowchart of the steps of the method of the present invention;

[0060] Figure 2 This is a schematic diagram of the background fitting function of the present invention;

[0061] Figure 3 This is a schematic diagram of the component fitting function of the present invention. Detailed Implementation

[0062] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] Example 1, please refer to Figure 1 As shown, this application provides a machine vision quality inspection method for semiconductor equipment component manufacturing, comprising the following steps:

[0064] Step S1: Obtain images of semiconductor device components and label them as component images. When obtaining component images, it is necessary to obtain a specific scene, i.e., a black background, to facilitate subsequent image analysis.

[0065] Step S2 involves converting the component image to grayscale to obtain a grayscale image of the component. Step S2 includes the following sub-steps:

[0066] Step S201: Obtain the R, G, and B channel values ​​of pixels in the component image, calculate the mean of the R, G, and B channel values ​​of each pixel, and mark it as the component grayscale value; replace the R, G, and B channel values ​​of each pixel in the component image with the corresponding component grayscale value to obtain the component grayscale image; since the image only contains the black background and the color of the component, the mean method is the most convenient.

[0067] In practical applications, for example, if the R, G, and B channel values ​​of a pixel in a component image are 23, 22, and 24 respectively, then the grayscale value of this pixel is 23.

[0068] Step S3: Obtain a background threshold based on a first number of images of normal components; the first number is needed to summarize the distribution of gray values ​​in the images, so multiple images are required to make the acquired data more accurate, for example, the first number is 10. Step S3 includes the following sub-steps:

[0069] Step S301: Perform grayscale processing on the image of normal parts to obtain a normal grayscale image, and obtain the grayscale values ​​of pixels that are not parts in the normal grayscale image, and mark them as background grayscale values; the grayscale processing method here is the same as in step 201.

[0070] Step S302: Obtain the range of background grayscale values, and divide the range of background grayscale values ​​into equal intervals of N, marking them as the background division range; in order to facilitate the calculation and observation of the distribution of background grayscale values, N should not be too large or too small, for example, N is set to 7;

[0071] Step S303: Calculate the frequency of the background grayscale value within each background division range and mark it as the background grayscale frequency;

[0072] Step S304: Establish a Cartesian coordinate system with the background grayscale value as the X-axis data and the background grayscale frequency as the Y-axis data, and mark it as the background coordinate system;

[0073] Step S305: Obtain the median of the background segmentation range and mark it as the background median;

[0074] Step S306: Use the background median and the corresponding background grayscale frequency as the x and y coordinates of the coordinate points and mark them as background coordinate points;

[0075] Step S307: Plot all background coordinate points into the background coordinate system to obtain a scatter plot, and mark it as a background scatter plot;

[0076] Step S308: Fit the background scatter plot to obtain a function, which is then marked as the background fitting function. The function fitting method can not only reduce the influence of abnormal background gray values ​​that are too large or too small, but also make the data obtained based on the background fitting function more accurate.

[0077] Step S309: Obtain the intersection point of the background fitting function and the X-axis of the background coordinate system, and mark it as the background intersection point;

[0078] Step S310: Obtain the maximum value of the x-coordinate among the background intersection points and mark it as the background threshold. This method is used to obtain the background threshold, making the obtained background grayscale range more accurate. Here, the maximum value is obtained because the background is black, and the range of background grayscale values ​​is close to 0; therefore, values ​​less than the maximum value can separate the background image portion. For practical applications, please refer to... Figure 2 As shown, if the maximum value of the horizontal coordinate among the background intersection points is 40, then the background threshold is 40.

[0079] Step S4: Obtain the component threshold based on the images of a first number of normal components; Step S4 includes the following sub-steps:

[0080] Step S401: Obtain the grayscale values ​​of the pixels of the component part in the normal grayscale image and mark them as component grayscale values;

[0081] Step S402: Obtain the range of grayscale values ​​of the component, and divide the range of grayscale values ​​of the component into equal intervals of M, and mark them as the component division range; in order to facilitate the calculation and observation of the distribution of grayscale values ​​of the component, M should not be too large or too small, for example, M is set to 7;

[0082] Step S403: Calculate the frequency of grayscale values ​​of each component within the defined range and mark it as the grayscale frequency of the component.

[0083] Step S404: Establish a Cartesian coordinate system with the grayscale values ​​of the parts as the X-axis data and the grayscale frequencies of the parts as the Y-axis data, and mark it as the parts coordinate system;

[0084] Step S405: Obtain the median value of the component division range and mark it as the component median value;

[0085] Step S406: Use the median value of the component and the corresponding grayscale frequency of the component as the x-coordinate and y-coordinate of the coordinate point, and mark it as the component coordinate point;

[0086] Step S407: Plot all component coordinate points into the component coordinate system to obtain a scatter plot, and mark it as a component scatter plot;

[0087] Step S408: Fit the scatter plot of the parts to obtain a function, which is marked as the part fitting function; the function fitting method can not only reduce the influence of abnormal gray values ​​of parts that are too large or too small, but also make the data obtained based on the part fitting function more accurate.

[0088] Step S409: Obtain the intersection point of the component fitting function and the X-axis of the component coordinate system, and mark it as the component intersection point;

[0089] Step S410: Obtain the minimum x-coordinate of the intersection points of the components and mark it as the component threshold. This method is used to obtain the component threshold, making the obtained component grayscale range more accurate. Here, the minimum value is obtained because the component is silver-white, and its grayscale distribution is different from the background. Therefore, the minimum value can distinguish the components from the background. For practical applications, please refer to [link to relevant documentation]. Figure 2 As shown, if the maximum value of the horizontal coordinate among the background intersection points is 80, then the background threshold is 80.

[0090] Step S5: Obtain the area to be detected based on the background threshold and the grayscale image of the component; Step S5 includes the following sub-steps:

[0091] Step S501: Set the gray values ​​in the grayscale image of the component that are less than or equal to the background threshold to 255, and set the gray values ​​in the grayscale image of the component that are greater than the background threshold to 0, to obtain a binarized image of the component.

[0092] In practical applications, gray values ​​less than or equal to 40 in the grayscale image of a component are set to 255, and gray values ​​greater than 40 in the grayscale image of a component are set to 0 to obtain a binarized image of the component.

[0093] Step S502: Establish a Cartesian coordinate system and mark it as the position coordinate system; place the binarized image of the component in the position coordinate system; obtain the largest area region composed of pixels with a gray value of 0 in the binarized image of the component and mark it as the initial region; obtain the position of the initial region in the position coordinate system and mark it as the initial region position.

[0094] Step S503: Place the grayscale image of the component in the same coordinate system as the binary image of the component, obtain the region at the initial region position in the grayscale image of the component, and mark it as the region to be detected; this method obtains the region to be detected in the grayscale image of the component based on the binary image of the component; the region to be detected is the region of the component that needs to be defect-judged.

[0095] Step S6: Binarize the area to be detected based on the component threshold to obtain a region binarized map; Step S6 includes the following sub-steps:

[0096] Step S601: Set the grayscale value of pixels in the detection area that are greater than or equal to the component threshold to 255, and set the grayscale value of pixels in the detection area that are less than the component threshold to 0 to obtain a region binarized image; if a defect occurs, under the action of light, the grayscale value of the defect part will be lower than the grayscale value of the component, and the method can obtain the defect part.

[0097] In practical applications, the grayscale value of pixels with a value greater than or equal to 80 in the region to be detected is set to 255, and the grayscale value of pixels with a value less than 40 in the region to be detected is set to 0, thus obtaining a binarized image of the region.

[0098] Step S7: Based on the region binarized map, determine whether an abnormal region exists. If it does, obtain the abnormal location, issue a defect signal, and simultaneously display the abnormal location; if it does not exist, issue a defect-free signal. Step S7 includes the following sub-steps:

[0099] Step S701: Mark the independent region consisting of pixels with a gray value of 0 in the region to be detected in the region binarized image as the initial defect region.

[0100] Step S702: Obtain the minimum area of ​​the permissible defect portion of the semiconductor and mark it as the minimum area.

[0101] Step S703: Obtain the pixel corresponding to the minimum area in the component image and mark it as the pixel number threshold. Here, the components for manufacturing semiconductors have high precision, so they cannot contain defects. However, in order to prevent interference from independent pixels, the pixel number threshold is set to be greater than or equal to 2.

[0102] Step S704: Obtain the number of pixels in the initial defect area. If the number of pixels in the initial defect area is greater than the pixel count threshold, then mark the initial defect area as an abnormal area.

[0103] Step S705: If the number of pixels in the initial defect area is less than or equal to the pixel count threshold, it indicates that no abnormal area has appeared.

[0104] In practical applications, if the minimum area of ​​the semiconductor defect portion is obtained and marked as the minimum area of ​​0, and the pixel count threshold is set to be greater than or equal to 2, then the pixel count threshold is 2. If an initial defect region exists and the number of pixels in the initial defect region is 122, and if the number of pixels in the initial defect region 122 is greater than the pixel count threshold 3, then the initial defect region is marked as an abnormal region.

[0105] Step S706: Obtain the maximum and minimum values ​​of the horizontal coordinate of the abnormal region, and mark them as the maximum horizontal value and the minimum horizontal value of the region, respectively.

[0106] Step S707: Obtain the maximum and minimum values ​​of the vertical coordinates of the abnormal region, and mark them as the maximum vertical value and the minimum vertical value of the region, respectively.

[0107] Step S708: Calculate the difference between the horizontal maximum value and the horizontal minimum value of the region, and mark it as the horizontal difference value of the region; calculate the difference between the vertical maximum value and the vertical minimum value of the region, and mark it as the vertical difference value of the region.

[0108] Step S709: Calculate the product of the vertical difference value and the horizontal difference value of the region, and mark it as the difference product; use the size of the difference product to obtain the minimum outer rectangle;

[0109] Step S710: Rotate the abnormal area 360° around any point within the abnormal area. When the product of the differences is minimized, mark the first coordinate point as the x-coordinate and y-coordinate of the area's smallest horizontal and vertical values ​​as a single coordinate point. Mark the second coordinate point as the x-coordinate and y-coordinate of the area's largest horizontal and vertical values ​​as a single coordinate point. Mark the third coordinate point as the x-coordinate and y-coordinate of the area's largest horizontal and vertical values ​​as a single coordinate point. Mark the fourth coordinate point as the x-coordinate and y-coordinate of the area's smallest horizontal and vertical values ​​as a single coordinate point. Connect the first, second, third, and fourth coordinate points in sequence to form a rectangle, which is then marked as the abnormal location. The obtained abnormal location is obtained using a minimum outer rectangle method. Displaying the abnormal location facilitates personnel's observation and further judgment of the defective parts of the components.

[0110] Example 2: This application also provides an electronic device, which may include: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus. The memory stores computer-readable instructions, and the processor can call the instructions in the memory. When the computer-readable instructions are executed by the processor, the steps of a machine vision quality inspection method for semiconductor equipment component production are performed to achieve the following functions: acquiring an image of a semiconductor equipment component and marking it as a component image; performing grayscale processing on the component image to obtain a component grayscale image; acquiring a background threshold based on a first number of normal component images; acquiring a component threshold based on the first number of normal component images; acquiring a region to be detected based on the background threshold and the component grayscale image; performing binarization processing on the region to be detected based on the component threshold to obtain a region binarization image; determining whether an abnormal region appears based on the region binarization image; if an abnormal region appears, acquiring the abnormal location, issuing a defect signal, and displaying the abnormal location; if no abnormal region appears, issuing a defect-free signal.

[0111] Furthermore, when the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0112] Example 3: This application also provides a computer program product, which includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute a machine vision quality inspection method for semiconductor equipment component production provided by the above methods. The method includes: acquiring an image of a semiconductor equipment component and marking it as a component image; performing grayscale processing on the component image to obtain a component grayscale image; acquiring a background threshold based on a first number of normal component images; acquiring a component threshold based on the first number of normal component images; acquiring a region to be detected based on the background threshold and the component grayscale image; performing binarization processing on the region to be detected based on the component threshold to obtain a region binarization image; determining whether an abnormal region appears based on the region binarization image; if an abnormal region appears, acquiring the abnormal location, issuing a defect signal and displaying the abnormal location; if no abnormal region appears, issuing a defect-free signal.

[0113] Example 4: This application also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it performs the steps of the machine vision quality inspection method for semiconductor equipment component production described above to achieve the following functions: acquiring images of semiconductor equipment components and marking them as component images; performing grayscale processing on the component images to obtain a component grayscale image; acquiring a background threshold based on images of a first number of normal components; acquiring a component threshold based on images of the first number of normal components; acquiring a region to be detected based on the background threshold and the component grayscale image; performing binarization processing on the region to be detected based on the component threshold to obtain a region binarization image; determining whether an abnormal region appears based on the region binarization image; if an abnormal region appears, acquiring the abnormal location, issuing a defect signal and displaying the abnormal location; if no abnormal region appears, issuing a defect-free signal.

[0114] Based on the above description of the embodiments, the embodiments of the present invention can be provided as methods, systems, or computer program products. Based on this understanding, the above technical solutions, in essence or in terms of their contribution to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or certain parts of the embodiments.

[0115] In the embodiments provided in this application, it should be understood that the disclosed system or method can be implemented in other ways. The embodiments described above are merely illustrative. For example, the division of modules or units is only a logical functional division, and there may be other division methods in actual implementation. Furthermore, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the coupling or direct coupling or communication connection shown or discussed may be through some communication interfaces. The indirect coupling or communication connection between systems, modules, and units may be electrical, mechanical, or other forms.

[0116] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A machine vision quality inspection method for semiconductor equipment component manufacturing, characterized in that, Includes the following steps: Acquire images of semiconductor equipment components and label them as component images; The component images are converted to grayscale to obtain grayscale images of the components; Obtain the background threshold based on the images of a first number of normal components; Obtain component thresholds based on images of a first number of normal components; The area to be detected is obtained based on the background threshold and the grayscale image of the component. Binarization of the area to be detected is performed based on the component threshold to obtain a region binarization map; Based on the region binarization map, determine whether an abnormal region has appeared. If it has appeared, obtain the abnormal location, issue a defect signal and display the abnormal location at the same time. If it does not occur, a defect no-defect signal will be issued; Obtaining a background threshold based on images of a first number of normal components includes the following sub-steps: performing grayscale processing on the images of normal components to obtain a normal grayscale image; obtaining the grayscale values ​​of pixels in the normal grayscale image that are not part of the component, and marking them as background grayscale values; obtaining the range of background grayscale values; dividing the range of background grayscale values ​​into N equal intervals, and marking them as background division ranges; calculating the frequency of background grayscale values ​​in each background division range, and marking them as background grayscale frequencies. Establish a Cartesian coordinate system with background grayscale values ​​as the X-axis and background grayscale frequency as the Y-axis, and mark it as the background coordinate system; obtain the median of the background division range and mark it as the background median; Use the background median and the corresponding background grayscale frequency as the x and y coordinates of the coordinate points, and mark them as background coordinate points; plot all background coordinate points into the background coordinate system to obtain a scatter plot, and mark it as a background scatter plot; fit the background scatter plot to obtain a function, and mark it as a background fitting function; obtain the intersection point of the background fitting function and the X-axis of the background coordinate system, and mark it as a background intersection point; Obtain the maximum value of the x-coordinate among the background intersection points and mark it as the background threshold; Obtaining the location of an anomaly includes the following sub-steps: obtaining the maximum and minimum values ​​of the horizontal coordinate of the anomaly region, and marking them as the region's horizontal maximum value and the region's horizontal minimum value, respectively; obtaining the maximum and minimum values ​​of the vertical coordinate of the anomaly region, and marking them as the region's vertical maximum value and the region's vertical minimum value, respectively. Calculate the difference between the largest and smallest values ​​in the region and label it as the region's horizontal difference value; Calculate the difference between the maximum and minimum vertical values ​​of the region and mark it as the regional vertical difference value; Calculate the product of the vertical difference and the horizontal difference of the region, and mark it as the difference product. Rotate the abnormal region around any point within the abnormal region. When the difference product is minimized, use the minimum horizontal and vertical values ​​of the region as the x and y coordinates of a single coordinate point, and mark it as the first coordinate point. Use the maximum horizontal and vertical values ​​of the region as the x and y coordinates of a single coordinate point, and mark it as the second coordinate point. Use the maximum horizontal and vertical values ​​of the region as the x and y coordinates of a single coordinate point, and mark it as the third coordinate point. Use the minimum horizontal and vertical values ​​of the region as the x and y coordinates of a single coordinate point, and mark it as the fourth coordinate point. Connect the first, second, third, and fourth coordinate points in sequence and mark the rectangle formed by connecting them end to end as the abnormal location.

2. The machine vision quality inspection method for semiconductor equipment component manufacturing according to claim 1, characterized in that, The process of converting component images to grayscale to obtain component grayscale images includes the following sub-steps: Obtain the R, G, and B channel values ​​of pixels in the component image, calculate the average of the R, G, and B channel values ​​of each pixel, and mark it as the component grayscale value; replace the R, G, and B channel values ​​of each pixel in the component image with the corresponding component grayscale value to obtain the component grayscale image.

3. The machine vision quality inspection method for semiconductor equipment component manufacturing according to claim 2, characterized in that, Obtaining the component threshold based on images of a first number of normal components includes the following sub-steps: Obtain the grayscale values ​​of the pixels of the component parts in the normal grayscale image and mark them as component grayscale values; Obtain the range of grayscale values ​​of the component, and divide the range of grayscale values ​​of the component into M equal intervals, marking them as the component division range; Calculate the frequency of grayscale values ​​of each component within the defined range, and mark it as the grayscale frequency of the component; A Cartesian coordinate system is established using the grayscale values ​​of the components as the X-axis and the grayscale frequencies of the components as the Y-axis. This system is then labeled as the component coordinate system.

4. The machine vision quality inspection method for semiconductor equipment component manufacturing according to claim 3, characterized in that, Obtaining the component threshold based on images of a first number of normal components also includes the following sub-steps: Obtain the median value of the component's segmentation range and mark it as the component's median value; The median value of the component and the corresponding gray frequency of the component are used as the x and y coordinates of the coordinate points, and these are marked as component coordinate points. Plot all component coordinate points into the component coordinate system to obtain a scatter plot, and label it as the component scatter plot; The function obtained by fitting the scatter plot of the parts is marked as the part fitting function; Obtain the intersection points of the component fitting function and the X-axis of the component coordinate system, and mark them as component intersection points; Find the minimum x-coordinate of the intersection points of the components and mark it as the component threshold.

5. The machine vision quality inspection method for semiconductor equipment component manufacturing according to claim 4, characterized in that, The process of obtaining the region to be detected based on the background threshold and component grayscale image includes the following sub-steps: Set the grayscale values ​​in the grayscale image of the component that are less than or equal to the background threshold to 255, and set the grayscale values ​​in the grayscale image of the component that are greater than the background threshold to 0 to obtain the binarized image of the component. Establish a Cartesian coordinate system, labeled as the position coordinate system, and place the binary image of the component in the position coordinate system; Obtain the largest area region composed of pixels with a grayscale value of 0 in the binarized image of the component, and mark it as the initial region; Obtain the position of the initial region in the position coordinate system and mark it as the initial region position; Place the grayscale image of the component in the same coordinate system as the binary image of the component, obtain the region at the initial location in the grayscale image of the component, and mark it as the region to be detected.

6. The machine vision quality inspection method for semiconductor equipment component manufacturing according to claim 5, characterized in that, Binarizing the region to be detected based on component thresholds to obtain a binarized map includes the following sub-steps: Set the grayscale value of pixels in the detection area that are greater than or equal to the component threshold to 255, and set the grayscale value of pixels in the detection area that are less than the component threshold to 0 to obtain a region binarized image.

7. The machine vision quality inspection method for semiconductor equipment component manufacturing according to claim 6, characterized in that, Based on the region binarized map, determine whether an abnormal region exists. If it does, obtain the abnormal location, issue a defect signal, and simultaneously display the abnormal location. If it does not exist, issue a defect absence signal, including the following sub-steps: The independent regions formed by pixels with a gray value of 0 within the region to be detected in the region binarized image are marked as the initial defect regions. Obtain the minimum area of ​​the permissible defect portion of the semiconductor and mark it as the minimum area. Find the pixel in the component image that corresponds to the minimum area and mark it as the threshold for the number of pixels; Obtain the number of pixels in the initial defect region. If the number of pixels in the initial defect region is greater than the pixel count threshold, then mark the initial defect region as an abnormal region. If the number of pixels in the initial defect area is less than or equal to the pixel count threshold, it indicates that no abnormal area has appeared.

Images