Micro through hole quality detection method and system based on data model feedback

Abstract

The invention discloses a micro-through-hole quality detection method and system based on data model feedback, and relates to the technical field of semiconductor detection, and the method comprises the steps: obtaining micro-through-hole image data, carrying out the gray processing of the micro-through-hole image data, carrying out the binary conversion of a processed gray image, obtaining micro-through-hole binary image data, and carrying out the detection of the micro-through-hole quality. And obtaining two-dimensional binary image data of the micro-through hole according to the binary image data of the micro-through hole. According to the method, the micro-through hole is roughly detected through the shortest diameter of the micro-through hole, the monitoring efficiency is improved, resource waste is avoided, the point cloud spacing information is obtained through the minimum defect size information and the shortest diameter of the micro-through hole, dynamic adjustment of point cloud model construction is achieved, the accuracy of the model is improved, and the accuracy of the model is improved. The micro through hole is further analyzed through the contour curve function of the micro through hole instead of directly comparing the aperture of the micro through hole with the production standard, so that misjudgment of the micro through hole is avoided, and the quality detection efficiency of the micro through hole is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of semiconductor detection, and specifically relates to a method and system for detecting the quality of micro-vias based on data model feedback. Background Art

[0002] In modern manufacturing industries, especially in the fields of semiconductors, electronics, aerospace, etc., the processing of micro-vias is extremely crucial. Taking semiconductor manufacturing as an example, the micro-vias in a chip are responsible for connecting different circuit layers, and their quality affects signal transmission and chip performance. In the electronic devices in the aerospace field, the quality of micro-vias is related to the reliability and safety of the devices. With the development of the manufacturing industry, the requirements for the accuracy, efficiency, and automation of micro-via quality detection are constantly increasing. The quality of micro-vias is directly related to the electrical performance and signal transmission quality after chip packaging. Therefore, the detection of micro-vias in advanced packaging has become a crucial part of the semiconductor detection process.

[0003] Currently, there are still problems in the quality detection of micro-vias, such as being unable to accurately evaluate the quality of micro-vias based on the scanned images of micro-vias. When detecting a large number of micro-vias, it is impossible to ensure the accuracy of each micro-via detection. If each micro-via is detected, it will result in low detection efficiency. For example, when using a probe contact method to detect micro-vias, the detection efficiency is low and the micro-vias may be damaged. It is impossible to further analyze the detection data and dynamically adjust the detection settings according to the actual detection situation. Summary of the Invention

[0004] To solve the above technical problems, a method and system for detecting the quality of micro-vias based on data model feedback are provided. The technical solution solves the problems proposed in the above background art, such as being unable to accurately evaluate the quality of micro-vias based on the scanned images of micro-vias. When detecting a large number of micro-vias, it is impossible to ensure the accuracy of each micro-via detection. If each micro-via is detected, it will result in low detection efficiency. For example, when using a probe contact method to detect micro-vias, the detection efficiency is low and the micro-vias may be damaged. It is impossible to further analyze the detection data and dynamically adjust the detection settings according to the actual detection situation.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A method for detecting the quality of micro-vias based on data model feedback, comprising:

[0007] Obtaining micro-via image data, where the micro-via image data includes micro-via two-dimensional image data and micro-via three-dimensional image data;

[0008] Performing gray-scale processing on the micro-via image data, and performing binary conversion on the processed gray-scale image to obtain micro-via binary image data;

[0009] Based on the binary image data of the micro-vias, obtain the two-dimensional binary image data of the micro-vias. The two-dimensional binary image data of the micro-vias represents the planar data of the micro-vias, including the position information and the opening shape information of the micro-vias.

[0010] Based on the two-dimensional binary image data of the micro-vias and the construction of a point cloud model, obtain the two-dimensional model of the micro-vias.

[0011] Based on the two-dimensional model of the micro-vias, determine whether the aperture of the micro-via meets the production standard. If not, the micro-via is unqualified. If so, based on the two-dimensional binary image data of the micro-vias, obtain the three-dimensional binary image data of the micro-vias. The three-dimensional binary image data of the micro-vias includes the sidewall image of the micro-via.

[0012] Based on the three-dimensional binary image data of the micro-vias, detect the depth and the wall of the micro-via, and determine whether the micro-via meets the production standard. If not, the micro-via is unqualified. If so, the micro-via is qualified.

[0013] Preferably, the step of based on the two-dimensional binary image data of the micro-vias and the construction of a point cloud model to obtain the two-dimensional model of the micro-vias specifically includes:

[0014] Based on the two-dimensional binary image data of the micro-vias and contour recognition, obtain the contour information of the micro-vias.

[0015] According to the contour information of the micro-vias, connect any two positions on the edge contour of the micro-via to obtain the contour feature line segment of the micro-via.

[0016] Measure the length of the contour feature line segment of the micro-via, and take the contour feature line segment with the maximum length as the longest diameter of the micro-via.

[0017] Taking the longest diameter of the micro-via as the reference, take the contour feature line segment that is perpendicular to the longest diameter of the micro-via and has the minimum length as the shortest diameter of the micro-via.

[0018] Based on the production standard of the micro-vias, obtain the standard diameter of the micro-via and the diameter difference threshold of the micro-via.

[0019] According to the standard diameter of the micro-via and the diameter difference threshold of the micro-via, determine whether the shortest diameter of the micro-via meets the production standard. If not, the micro-via is unqualified and mark it. If so, based on the micro-via process technology, obtain the minimum defect size information. The minimum defect size information represents the minimum size of the non-negligible defect of the micro-via.

[0020] According to the minimum defect size information and the shortest diameter of the micro-via, obtain the point cloud spacing information.

[0021] Based on the point cloud spacing information and the two-dimensional binary image data of the micro-vias, construct a point cloud model and obtain the two-dimensional model of the micro-vias.

[0022] Preferably, obtaining the point cloud spacing information according to the minimum defect size information and the shortest diameter of the microvia specifically includes:

[0023] Obtaining the maximum tolerance diameter and the minimum tolerance diameter of the microvia according to the standard diameter of the microvia and the diameter difference threshold of the microvia;

[0024] Taking the maximum tolerance diameter of the microvia and the standard diameter of the microvia as the first diameter threshold of the microvia, and taking the minimum tolerance diameter of the microvia and the standard diameter of the microvia as the second diameter threshold of the microvia;

[0025] Obtaining the basic point cloud spacing information according to the minimum defect size information based on the Nyquist sampling theorem, where the basic point cloud spacing is half of the minimum defect size;

[0026] Obtaining the point cloud spacing information according to the first diameter threshold of the microvia, the second diameter threshold of the microvia, the shortest diameter of the microvia, and the basic point cloud spacing information;

[0027] Among them, if the shortest diameter of the microvia is within the second diameter threshold of the microvia, then taking the basic point cloud spacing as the point cloud average spacing when constructing the point cloud model;

[0028] If the shortest diameter of the microvia is within the first diameter threshold of the microvia, then taking the difference between the shortest diameter of the microvia and the standard diameter of the microvia as the diameter deviation value;

[0029] Taking the ratio of the diameter deviation value to the standard diameter of the microvia as the diameter deviation coefficient;

[0030] Obtaining the adjusted point cloud spacing according to the diameter deviation coefficient and the basic point cloud spacing based on the dynamic reduction coefficient method;

[0031] Taking the adjusted point cloud spacing as the point cloud average spacing when constructing the point cloud model;

[0032] The adjusted point cloud spacing is specifically:

[0033]

[0034] In the formula, D is the adjusted point cloud spacing, D0 is the basic point cloud spacing, w is the diameter deviation coefficient, d min is the shortest diameter of the microvia, and d0 is the standard diameter of the microvia.

[0035] Preferably, taking the adjusted point cloud spacing as the point cloud average spacing when constructing the point cloud model further includes:

[0036] Obtaining the image resolution information according to the microvia image data based on the device parameters;

[0037] According to the image resolution information, compare the adjusted point cloud spacing with the image resolution to determine whether the adjusted point cloud spacing meets the point cloud model spacing setting standard;

[0038] Among them, if the adjusted point cloud spacing is greater than the image resolution, the adjusted point cloud spacing meets the point cloud model spacing setting standard, and the adjusted point cloud spacing is used as the average point cloud spacing when constructing the point cloud model;

[0039] If the adjusted point cloud spacing is less than the image resolution, the adjusted point cloud spacing does not meet the point cloud model spacing setting standard, and the image resolution is used as the average point cloud spacing when constructing the point cloud model.

[0040] Preferably, the judgment of whether the microvia hole diameter meets the production standard according to the microvia hole two-dimensional model specifically includes:

[0041] According to the microvia hole two-dimensional model, based on the construction of the Cartesian coordinate system, obtain the microvia hole coordinate data, and the microvia hole coordinate data represents the coordinate data of the points located on the edge contour of the microvia hole in the microvia hole two-dimensional model;

[0042] According to the microvia hole coordinate data, based on data fitting, obtain the microvia hole contour curve function;

[0043] Among them, if the longest diameter of the microvia hole is equal to the shortest diameter of the microvia hole, based on the shortest diameter of the microvia hole and the circle equation, obtain the microvia hole contour curve function;

[0044] If the longest diameter of the microvia hole is not equal to the shortest diameter of the microvia hole, based on the microvia hole coordinate data, perform ellipse equation fitting based on the random Hough transform and the least squares method to obtain the microvia hole contour curve function;

[0045] According to the microvia hole contour curve function, obtain the major axis information and minor axis information of the microvia hole contour;

[0046] Based on the microvia hole production standard, obtain the microvia hole ellipticity coefficient and the edge roughness threshold, and the microvia hole ellipticity coefficient represents the maximum major axis / minor axis ratio allowed for microvia hole production;

[0047] According to the major axis information of the microvia hole contour, the minor axis information of the microvia hole contour and the microvia hole ellipticity coefficient, judge whether the microvia hole diameter meets the production standard. If the ratio of the major axis and minor axis of the microvia hole contour is greater than the microvia hole ellipticity coefficient, the microvia hole is unqualified and marked. If the ratio of the major axis and minor axis of the microvia hole contour is less than the microvia hole ellipticity coefficient, based on the microvia hole edge roughness standard, obtain the sampling length information;

[0048] Obtain the image resolution information, and based on the image resolution, obtain the horizontal step distance information;

[0049] Obtain the sampling point number information according to the sampling length information and the lateral step distance information;

[0050] Filter the microvia coordinate data according to the sampling point number information to obtain the sampling point coordinate information;

[0051] Based on the polar coordinate system, convert the sampling point coordinates and the microvia contour curve function to obtain the sampling point polar coordinate information and the microvia contour curve polar coordinate function;

[0052] Taking the polar angle in the sampling point polar coordinates as a reference, obtain the contour point polar coordinate information corresponding to the sampling point in the microvia contour curve polar coordinate function, where the polar angles of the sampling point and the contour point are the same;

[0053] Take the difference between the sampling point coordinates and the contour point polar coordinates as the radial deviation value to obtain the radial deviation data;

[0054] Obtain the microvia edge roughness according to the radial deviation data;

[0055] Judge whether the microvia aperture meets the production standard according to the microvia edge roughness and the edge roughness threshold. If the microvia edge roughness is greater than the edge roughness threshold, the microvia is unqualified and marked. If the microvia edge roughness is less than the edge roughness threshold, the microvia aperture is qualified.

[0056] Preferably, the detection of the microvia hole depth and the hole wall according to the microvia three-dimensional binary image data specifically includes:

[0057] Obtain the substrate thickness information and the substrate roughness information, where the substrate thickness information includes the thickness information of each layer in the substrate;

[0058] Based on the microvia application requirements, obtain the calibrated hole depth information corresponding to each microvia, where the calibrated hole depth is the substrate thickness that the microvia needs to penetrate;

[0059] Obtain the microvia hole depth threshold according to the calibrated hole depth information and the substrate roughness;

[0060] Obtain the microvia hole depth information based on the microvia three-dimensional binary image data;

[0061] Judge whether the microvia hole depth meets the production standard according to the microvia hole depth information and the microvia hole depth threshold. If the microvia hole depth exceeds the microvia hole depth threshold, the microvia is unqualified and marked. If the microvia hole depth does not exceed the microvia hole depth threshold, the microvia hole depth meets the production standard;

[0062] Obtain the microvia sidewall cone angle information according to the microvia three-dimensional binary image data;

[0063] Obtain the tolerance information of the sidewall taper angle of the microvia based on the microvia production standard;

[0064] According to the sidewall taper angle information of the microvia and the tolerance information of the sidewall taper angle of the microvia, determine whether the sidewall of the microvia meets the production standard. If the sidewall taper angle of the microvia exceeds the tolerance of the sidewall taper angle of the microvia, the microvia is unqualified and marked. If the sidewall taper angle of the microvia does not exceed the tolerance of the sidewall taper angle of the microvia, the depth of the microvia meets the production standard.

[0065] Furthermore, a microvia quality detection system based on data model feedback is proposed to implement the above detection method, including:

[0066] The main control module is used to judge whether the shortest diameter of the microvia meets the production standard according to the standard diameter of the microvia and the diameter difference threshold of the microvia, compare the adjusted point cloud spacing with the image resolution according to the image resolution information, and judge whether the adjusted point cloud spacing meets the point cloud model spacing setting standard. According to the major axis information of the microvia contour, the minor axis information of the microvia contour and the ellipticity coefficient of the microvia, judge whether the aperture of the microvia meets the production standard. According to the edge roughness of the microvia and the edge roughness threshold, judge whether the aperture of the microvia meets the production standard. Detect the depth and the wall of the microvia according to the three-dimensional binary image data of the microvia, and judge whether the microvia meets the production standard. According to the coordinate data of the microvia, obtain the contour curve function of the microvia based on data fitting. According to the contour curve function of the microvia, obtain the major axis information of the microvia contour and the minor axis information of the microvia contour. According to the sampling point number information, screen the coordinate data of the microvia to obtain the sampling point coordinate information. Based on the polar coordinate system, convert the sampling point coordinates and the microvia contour curve function to obtain the sampling point polar coordinate information and the microvia contour curve polar coordinate function. Taking the polar angle in the sampling point polar coordinates as the reference, obtain the contour point polar coordinate information corresponding to the sampling point in the microvia contour curve polar coordinate function. Take the difference between the sampling point coordinates and the contour point polar coordinates as the radial deviation value, obtain the radial deviation data, and obtain the edge roughness of the microvia according to the radial deviation data;

[0067] The information acquisition module is used to acquire the microvia image data, the microvia two-dimensional image data and the microvia three-dimensional image data, perform gray processing on the microvia image data, and perform binary conversion on the processed gray image to obtain the microvia binary image data. Based on contour recognition, obtain the microvia contour information according to the microvia two-dimensional binary image data. Based on the device parameters, obtain the image resolution information according to the microvia image data, and obtain the horizontal step distance information based on the image resolution;

[0068] An image processing module, which is used to connect any two positions on the edge contour of the micro via according to the micro via contour information to obtain the contour feature line segment of the micro via, obtain the longest diameter and the shortest diameter of the micro via according to the contour feature line segment of the micro via, obtain the point cloud spacing information according to the minimum defect size information and the shortest diameter of the micro via, construct a point cloud model according to the point cloud spacing information and the two-dimensional binary image data of the micro via, and obtain the two-dimensional model of the micro via;

[0069] A display module, which interacts with the main control module and is used to display the two-dimensional binary image data of the micro via, the two-dimensional model of the micro via, the contour curve function of the micro via, and the marking information of the micro via.

[0070] Optionally, the main control module specifically includes:

[0071] A control unit, which is used to obtain the contour curve function of the micro via based on data fitting according to the micro via coordinate data, obtain the major axis information and the minor axis information of the micro via contour according to the contour curve function of the micro via, screen the micro via coordinate data according to the sampling point number information to obtain the sampling point coordinate information, convert the sampling point coordinates and the micro via contour curve function based on the polar coordinate system to obtain the sampling point polar coordinate information and the micro via contour curve polar coordinate function, take the polar angle in the sampling point polar coordinates as a reference, obtain the contour point polar coordinate information corresponding to the sampling point in the micro via contour curve polar coordinate function, use the difference between the sampling point coordinates and the contour point polar coordinates as the radial deviation value to obtain the radial deviation data, and obtain the edge roughness of the micro via according to the radial deviation data;

[0072] An information receiving unit, which interacts with the information acquisition module and the image processing module and is used to receive data and transmit it to the judgment unit;

[0073] A judgment unit, which is used to judge whether the shortest diameter of the micro via meets the production standard according to the standard diameter of the micro via and the diameter difference threshold of the micro via, judge whether the adjusted point cloud spacing meets the point cloud model spacing setting standard according to the image resolution information, judge whether the aperture of the micro via meets the production standard according to the major axis information of the micro via contour, the minor axis information of the micro via contour, and the ellipticity coefficient of the micro via, judge whether the aperture of the micro via meets the production standard according to the edge roughness of the micro via and the edge roughness threshold, and judge whether the micro via meets the production standard according to the three-dimensional binary image data of the micro via.

[0074] Optionally, the information acquisition module specifically includes:

[0075] A first acquisition unit, which is used to acquire micro-via image data, two-dimensional micro-via image data, and three-dimensional micro-via image data, perform gray-scale processing on the micro-via image data, and perform binary conversion on the processed gray-scale image to obtain binary micro-via image data;

[0076] A second acquisition unit, which is used to obtain micro-via contour information based on contour recognition according to the two-dimensional binary micro-via image data, obtain image resolution information based on the micro-via image data and device parameters, and obtain horizontal step distance information based on the image resolution.

[0077] Optionally, the image processing module specifically includes:

[0078] An image processing unit, which is used to connect any two positions on the micro-via edge contour according to the micro-via contour information to obtain a micro-via contour feature line segment, and obtain the longest diameter and the shortest diameter of the micro-via according to the micro-via contour feature line segment;

[0079] A model construction unit, which is used to obtain point cloud spacing information according to the minimum defect size information and the shortest diameter of the micro-via, construct a point cloud model according to the point cloud spacing information and the two-dimensional binary micro-via image data, and obtain a two-dimensional micro-via model.

[0080] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0081] The present invention provides a micro-via quality detection method and system based on data model feedback. By using the shortest diameter of the micro-via, the micro-via is roughly detected, which improves the monitoring efficiency and avoids waste of resources. By using the minimum defect size information and the shortest diameter of the micro-via, the point cloud spacing information is obtained, which realizes the dynamic adjustment of the point cloud model construction and improves the accuracy of the model. By further analyzing the micro-via through the micro-via contour curve function instead of directly comparing the micro-via aperture with the production standard, misjudgment of the micro-via is avoided and the micro-via quality detection efficiency is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 It is a flowchart of a micro-via quality detection method based on data model feedback proposed by the present invention;

[0083] Figure 2 It is a flowchart for obtaining a two-dimensional micro-via model in the present invention;

[0084] Figure 3 It is a flowchart for obtaining point cloud spacing information in the present invention;

[0085] Figure 4 It is a flowchart for obtaining the edge roughness of the micro-via in the present invention;

[0086] Figure 5 This is a structural block diagram of a microvia quality detection system based on data model feedback proposed by the present invention. Detailed implementation manners

[0087] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments in the following description are only examples, and those skilled in the art can think of other obvious variations.

[0088] Referring to Figure 1 - Figure 4 As shown, a microvia quality detection method based on data model feedback in an embodiment of the present invention includes:

[0089] Obtain microvia image data, where the microvia image data includes microvia two-dimensional image data and microvia three-dimensional image data;

[0090] Perform grayscale processing on the microvia image data, and perform binary conversion on the processed grayscale image to obtain microvia binary image data;

[0091] According to the microvia binary image data, obtain microvia two-dimensional binary image data, where the microvia two-dimensional binary image data represents the plane data of the microvia, including microvia position information and opening shape information;

[0092] According to the microvia two-dimensional binary image data, obtain a microvia two-dimensional model based on the construction of a point cloud model;

[0093] Specifically, according to the microvia two-dimensional binary image data, obtaining a microvia two-dimensional model based on the construction of a point cloud model specifically includes:

[0094] According to the microvia two-dimensional binary image data, obtain microvia contour information based on contour recognition;

[0095] According to the microvia contour information, connect any two positions on the microvia edge contour to obtain a microvia contour feature line segment;

[0096] Measure the length of the microvia contour feature line segment, and take the microvia contour feature line segment with the maximum length as the longest diameter of the microvia;

[0097] Taking the longest diameter of the microvia as a reference, take the microvia contour feature line segment that is perpendicular to the longest diameter of the microvia and has the minimum length as the shortest diameter of the microvia;

[0098] Based on the microvia production standard, obtain the standard diameter of the microvia and the diameter difference threshold of the microvia;

[0099] According to the standard diameter of the microvia and the diameter difference threshold of the microvia, determine whether the shortest diameter of the microvia meets the production standard. If not, the microvia is unqualified and marked. If so, based on the microvia process technology, obtain the minimum defect size information, where the minimum defect size information represents the minimum size of non-negligible defects of the microvia;

[0100] According to the minimum defect size information and the shortest diameter of the microvia, obtain the point cloud spacing information;

[0101] According to the point cloud spacing information and the two-dimensional binary image data of the microvia, construct a point cloud model and obtain the two-dimensional model of the microvia.

[0102] In this solution, through the two-dimensional binary image data of the microvia, based on contour recognition, obtain the microvia contour information. According to the microvia contour information, connect any two positions on the edge contour of the microvia to obtain the microvia contour feature line segment. Take the microvia contour feature line segment with the largest length as the longest diameter of the microvia. Based on the longest diameter of the microvia, take the microvia contour feature line segment that is perpendicular to the longest diameter of the microvia and has the smallest length as the shortest diameter of the microvia. According to the standard diameter of the microvia and the diameter difference threshold of the microvia, determine whether the shortest diameter of the microvia meets the production standard. If not, the microvia is unqualified and marked. If so, based on the microvia process technology, obtain the minimum defect size information. According to the minimum defect size information and the shortest diameter of the microvia, obtain the point cloud spacing information; According to the point cloud spacing information and the two-dimensional binary image data of the microvia, construct a point cloud model and obtain the two-dimensional model of the microvia;

[0103] It can be understood that the aperture of the microvia is often directly related to the electrical performance and signal transmission quality after chip packaging. Therefore, the detection of the microvia aperture is the top priority of microvia quality detection. By comparing the standard diameter of the microvia and the diameter difference threshold of the microvia, rapid judgment of the microvia aperture quality is achieved, avoiding subsequent precise detection steps for the microvia, saving resources and improving detection efficiency.

[0104] It should be noted that in this embodiment, the microvia aperture threshold is set according to the application scenario of the microvia, and the standard diameter of the microvia takes the middle value of the threshold range. For example: for TSV (Through-Silicon Via), the front-end-of-line (FEOL) TSV: the diameter is usually 2 - 4μm, mainly used for high-density interconnect scenarios (such as 3DIC integration), and the back-end-of-line (BEOL) TSV: the diameter range is 5 - 20μm, while for TGV (Through-Glass Via) the standard diameter is 10 - 100μm.

[0105] Specifically, according to the minimum defect size information and the shortest diameter of the microvia, obtaining the point cloud spacing information specifically includes:

[0106] Obtain the maximum tolerance diameter and the minimum tolerance diameter of the micro-vias according to the standard diameter of the micro-vias and the diameter difference threshold of the micro-vias;

[0107] Take the maximum tolerance diameter and the standard diameter of the micro-vias as the first diameter threshold of the micro-vias, and take the minimum tolerance diameter and the standard diameter of the micro-vias as the second diameter threshold of the micro-vias;

[0108] According to the minimum defect size information, obtain the basic point cloud spacing information based on the Nyquist sampling theorem, where the basic point cloud spacing is half of the minimum defect size;

[0109] Obtain the point cloud spacing information according to the first diameter threshold of the micro-vias, the second diameter threshold of the micro-vias, the shortest diameter of the micro-vias, and the basic point cloud spacing information;

[0110] Among them, if the shortest diameter of the micro-via is within the second diameter threshold of the micro-via, then take the basic point cloud spacing as the average point cloud spacing when constructing the point cloud model;

[0111] If the shortest diameter of the micro-via is within the first diameter threshold of the micro-via, then take the difference between the shortest diameter of the micro-via and the standard diameter of the micro-via as the diameter deviation value;

[0112] Take the ratio of the diameter deviation value to the standard diameter of the micro-via as the diameter deviation coefficient;

[0113] Obtain the adjusted point cloud spacing based on the dynamic reduction coefficient method according to the diameter deviation coefficient and the basic point cloud spacing;

[0114] Take the adjusted point cloud spacing as the average point cloud spacing when constructing the point cloud model;

[0115] The adjusted point cloud spacing is specifically:

[0116]

[0117] In the formula, D is the adjusted point cloud spacing, D0 is the basic point cloud spacing, w is the diameter deviation coefficient, d min is the shortest diameter of the micro-via, and d0 is the standard diameter of the micro-via.

[0118] In this solution, by taking the maximum tolerance diameter and the standard diameter of the micro-vias as the first diameter threshold of the micro-vias, and taking the minimum tolerance diameter and the standard diameter of the micro-vias as the second diameter threshold of the micro-vias, according to the minimum defect size information, obtain the basic point cloud spacing information based on the Nyquist sampling theorem, and obtain the point cloud spacing information according to the first diameter threshold of the micro-vias, the second diameter threshold of the micro-vias, the shortest diameter of the micro-vias, and the basic point cloud spacing information;

[0119] It is understandable that in advanced packaging technologies, the standard diameter of micro-vias (such as TSVs and TGvs) and the threshold of their diameter difference (i.e., the tolerance range allowed by the process) are key parameters affecting the interconnect density and reliability. In this solution, the minimum tolerance diameter of the micro-via and the standard diameter of the micro-via are used as the second diameter threshold of the micro-via, that is, the allowable diameter range below the standard diameter is used as the second diameter threshold of the micro-via (negative tolerance range), and the maximum tolerance diameter of the micro-via and the standard diameter of the micro-via are used as the first diameter threshold of the micro-via, that is, the allowable diameter range above the standard diameter is used as the first diameter threshold of the micro-via (positive tolerance range). If the diameter of the micro-via is within the negative tolerance range, defects such as burrs may occur. Therefore, half of the minimum defect size (such as the minimum burr height that cannot be ignored, etc.) and the basic point cloud spacing are used as the average point cloud spacing when constructing the point cloud model, ensuring that the model can accurately reflect the actual situation of the micro-via. If the diameter of the micro-via is within the positive tolerance range, insufficient solder filling (void risk), stress concentration on the hole wall (crack tendency), and local expansion deformation (abnormal ellipticity) may occur. At this time, to ensure the accuracy of the model, it is necessary to further adjust the point cloud spacing. Therefore, in this solution, through the dynamic reduction coefficient method, the point cloud spacing is adjusted according to the deviation between the shortest diameter of the micro-via and the standard diameter of the micro-via.

[0120] Specifically, using the adjusted point cloud spacing as the average point cloud spacing when constructing the point cloud model further includes:

[0121] Based on the device parameters, obtain the image resolution information according to the micro-via image data;

[0122] Compare the adjusted point cloud spacing with the image resolution according to the image resolution information, and determine whether the adjusted point cloud spacing meets the point cloud model spacing setting standard;

[0123] Among them, if the adjusted point cloud spacing is greater than the image resolution, the adjusted point cloud spacing meets the point cloud model spacing setting standard, and the adjusted point cloud spacing is used as the average point cloud spacing when constructing the point cloud model;

[0124] If the adjusted point cloud spacing is less than the image resolution, the adjusted point cloud spacing does not meet the point cloud model spacing setting standard, and the image resolution is used as the average point cloud spacing when constructing the point cloud model.

[0125] In this solution, by comparing the adjusted point cloud spacing with the image resolution to determine whether the adjusted point cloud spacing meets the point cloud model spacing setting standard, it is understandable that the setting of the point cloud spacing affects the accuracy of the point cloud model. However, if the adjusted point cloud spacing is less than the image resolution, at this time, it is impossible to obtain the accurate position relationship between adjacent points in the point cloud model according to the image, and it is impossible to determine the accurate position of each point in the point cloud model. Therefore, it is necessary to detect the point cloud spacing to determine whether it meets the actual requirements.

[0126] According to the two-dimensional model of the microvia, determine whether the aperture of the microvia meets the production standard. If not, the microvia is unqualified. If so, based on the binary image data of the microvia, obtain the three-dimensional binary image data of the microvia. The three-dimensional binary image data of the microvia includes the sidewall image of the microvia;

[0127] Specifically, determining whether the aperture of the microvia meets the production standard according to the two-dimensional model of the microvia specifically includes:

[0128] According to the two-dimensional model of the microvia, based on the construction of the Cartesian coordinate system, obtain the coordinate data of the microvia. The coordinate data of the microvia represents the coordinate data of the points located on the edge contour of the two-dimensional model of the microvia;

[0129] According to the coordinate data of the microvia, based on data fitting, obtain the contour curve function of the microvia;

[0130] Among them, if the longest diameter of the microvia is equal to the shortest diameter of the microvia, then according to the shortest diameter of the microvia, based on the circle equation, obtain the contour curve function of the microvia;

[0131] If the longest diameter of the microvia is not equal to the shortest diameter of the microvia, then based on the coordinate data of the microvia, perform ellipse equation fitting based on the random Hough transform and the least squares method to obtain the contour curve function of the microvia;

[0132] According to the contour curve function of the microvia, obtain the major axis information of the microvia contour and the minor axis information of the microvia contour;

[0133] Based on the microvia production standard, obtain the microvia ellipticity coefficient and the edge roughness threshold. The microvia ellipticity coefficient represents the maximum major axis / minor axis ratio allowed for microvia production;

[0134] According to the major axis information of the microvia contour, the minor axis information of the microvia contour, and the microvia ellipticity coefficient, determine whether the aperture of the microvia meets the production standard. If the ratio of the major axis of the microvia contour to the minor axis of the microvia contour is greater than the microvia ellipticity coefficient, the microvia is unqualified and is marked. If the ratio of the major axis of the microvia contour to the minor axis of the microvia contour is less than the microvia ellipticity coefficient, then based on the microvia edge roughness standard, obtain the sampling length information;

[0135] Obtain the image resolution information, and based on the image resolution, obtain the horizontal step information;

[0136] According to the sampling length information and the horizontal step information, obtain the sampling point number information;

[0137] According to the sampling point number information, screen the coordinate data of the microvia to obtain the sampling point coordinate information;

[0138] Based on the polar coordinate system, the coordinates of the sampling points and the micro via hole profile curve function are transformed to obtain the polar coordinate information of the sampling points and the polar coordinate function of the micro via hole profile curve;

[0139] Taking the polar angle in the polar coordinates of the sampling points as a reference, the polar coordinate information of the profile points corresponding to the sampling points in the polar coordinate function of the micro via hole profile curve is obtained, and the polar angles of the sampling points and the profile points are the same;

[0140] The difference between the coordinates of the sampling points and the polar coordinates of the profile points is used as the radial deviation value to obtain the radial deviation data;

[0141] Based on the radial deviation data, the edge roughness of the micro via hole is obtained;

[0142] According to the edge roughness of the micro via hole and the edge roughness threshold, it is judged whether the aperture of the micro via hole meets the production standard. If the edge roughness of the micro via hole is greater than the edge roughness threshold, the micro via hole is unqualified and marked. If the edge roughness of the micro via hole is less than the edge roughness threshold, the aperture of the micro via hole is qualified.

[0143] In this solution, based on the micro via hole coordinate data and data fitting, the micro via hole profile curve function is obtained. According to the major axis information, minor axis information and ellipticity coefficient of the micro via hole profile, it is judged whether the aperture of the micro via hole meets the production standard. If not, the micro via hole is marked as unqualified. If so, the edge roughness of the micro via hole is further detected.

[0144] It can be understood that in the design of the micro via hole opening, the standard opening is usually defined as circular, which is based on the following process requirements:

[0145] Electroplating uniformity: The circular opening can ensure uniform flow of the electroplating solution and avoid voids or stress concentration during copper filling (for example, in the TSV redundancy design of AMD graphics cards, the via hole shape directly affects the electroplating reliability)

[0146] Signal integrity: Regular openings reduce electromagnetic field distortion. Especially in high-speed signal transmission (such as HBM memory), an aperture deviation exceeding ±2μm may cause impedance mismatch

[0147] However, in the actual production process, the micro via hole opening will inevitably have certain changes, such as oval shape, etc. Therefore, in the actual micro via hole quality inspection, the micro via holes with non-circular openings are not directly marked as unqualified. For example, in the TSMC CoWoS package, the interposer TSV allows a slight oval shape, but it needs to be dynamically monitored by an optical profiler + SPC statistical control. Therefore, in this embodiment, according to the application requirements of the micro via hole, the ellipticity coefficient and edge roughness threshold of the micro via hole are set. For example:

[0148]

[0149] It should be noted that in this embodiment, based on the recommended parameters of the ISO 1997 standard, 0.08 mm is selected as the sampling length. For the image acquisition device, the horizontal step represents the minimum distance at which features can be recognized in the images collected by the device, that is, the definition of resolution. Distances less than this will not allow image features to be recognized. Therefore, the ratio of the sampling length to the horizontal step is used as the number of sampling points, which complies with the national production standard ISO 1997.

[0150] In this embodiment, if the longest diameter of the micro via hole is equal to the shortest diameter of the micro via hole, the micro via hole is circular, and the data is fitted according to the conventional circular function expression to obtain the contour curve function of the micro via hole. If the longest diameter of the micro via hole is not equal to the shortest diameter of the micro via hole, the micro via hole is elliptical, and the data is fitted based on the elliptical equation to obtain the contour curve function of the micro via hole.

[0151] Based on the three-dimensional binary image data of the micro via hole, the depth and the wall of the micro via hole are detected to determine whether the micro via hole meets the production standard. If not, the micro via hole is unqualified; if so, the micro via hole is qualified.

[0152] Specifically, based on the three-dimensional binary image data of the micro via hole, the depth and the wall of the micro via hole are detected, which specifically includes:

[0153] Obtain the substrate thickness information and the substrate roughness information, where the substrate thickness information includes the thickness information of each layer in the substrate;

[0154] Based on the application requirements of the micro via hole, obtain the calibrated hole depth information corresponding to each micro via hole, where the calibrated hole depth is the substrate thickness that the micro via hole needs to penetrate;

[0155] According to the calibrated hole depth information and the substrate roughness, obtain the micro via hole depth threshold;

[0156] Based on the three-dimensional binary image data of the micro via hole, obtain the micro via hole depth information;

[0157] According to the micro via hole depth information and the micro via hole depth threshold, determine whether the micro via hole depth meets the production standard. If the micro via hole depth exceeds the micro via hole depth threshold, the micro via hole is unqualified and is marked. If the micro via hole depth does not exceed the micro via hole depth threshold, the micro via hole depth meets the production standard;

[0158] Based on the three-dimensional binary image data of the micro via hole, obtain the side wall cone angle information of the micro via hole;

[0159] Based on the micro via hole production standard, obtain the side wall cone angle tolerance information of the micro via hole;

[0160] According to the microvia sidewall taper angle information and the microvia sidewall taper angle tolerance information, determine whether the microvia sidewall meets the production standard. If the microvia sidewall taper angle exceeds the microvia sidewall taper angle tolerance, the microvia is unqualified and marked. If the microvia sidewall taper angle does not exceed the microvia sidewall taper angle tolerance, the microvia hole depth meets the production standard.

[0161] In this solution, based on the microvia three-dimensional binary image data, obtain the microvia hole depth information. According to the microvia hole depth information and the microvia hole depth threshold, determine whether the microvia hole depth meets the production standard. According to the microvia three-dimensional binary image data, obtain the microvia sidewall taper angle information. According to the microvia sidewall taper angle information and the microvia sidewall taper angle tolerance information, determine whether the microvia sidewall meets the production standard.

[0162] It can be understood that the microvia aperture directly affects the interconnection density, signal transmission performance, and mechanical strength. For example, in TSMC's CoWoS technology, the aperture accuracy of micro-bumps (μBmps) and through-silicon vias (TSVs) determines the interconnection density and electrical performance of chip stacking1. If the aperture deviation is too large, it may lead to signal delay, short circuit, or mechanical stress concentration6. Therefore, detailed detection of the aperture is a core link to ensure packaging reliability. For hole depth and hole wall detection, in mature processes, the depth control of laser drilling (such as CO2 laser or UV / YAG laser) is usually relatively stable10, the probability of hole depth deviation is low, and the hole wall roughness may affect the electroplating filling quality, but the unqualified rate can be significantly reduced through process optimization (such as chemical cleaning and electroplating solution parameter control)14. Under stable processes, simple sampling inspection can cover the risks. The aperture accuracy usually reflects the stability of the process equipment (such as the focusing accuracy and parameter control of the laser drilling machine). If the aperture meets the standard, it indicates that the equipment is operating stably, indirectly indicating a high consistency of the hole depth and hole wall10.

[0163] For example, in Intel's glass substrate technology, the strict control of the TGV (through-glass via) aperture is achieved at a 75μm pitch, and its process stability also ensures the qualification rates of the hole depth and hole wall2. Therefore, in this solution, when the microvia aperture detection is qualified, only the hole depth and hole wall taper angle of the microvia are detected, which improves the detection efficiency and ensures the accuracy of the detection.

[0164] In this embodiment, the product of the calibrated hole depth and the substrate roughness is used as the allowable hole depth deviation amount, and the microvia hole depth threshold is set according to the hole depth deviation amount, specifically:

[0165]

[0166] In the formula, T1 and T2 are the microvia hole depth thresholds, T0 is the calibrated hole depth, and Q is the roughness;

[0167] If the depth T of the micro via hole belongs to [T1, T2], the depth of the micro via hole meets the production standard.

[0168] The taper angle tolerance of the side wall of the micro via hole needs to be set according to the production process of the micro via hole. The taper angle tolerances of the side walls of the micro via holes in different application scenarios are also different. For example:

[0169]

[0170] Refer to Figure 5 As shown, further, in combination with the above-mentioned method for detecting the quality of micro via holes based on data model feedback, a system for detecting the quality of micro via holes based on data model feedback is proposed, including:

[0171] A main control module, which is used to judge whether the shortest diameter of the micro via hole meets the production standard according to the standard diameter of the micro via hole and the diameter difference threshold of the micro via hole, compare the adjusted point cloud spacing with the image resolution according to the image resolution information, and judge whether the adjusted point cloud spacing meets the point cloud model spacing setting standard. According to the major axis information of the micro via hole contour, the minor axis information of the micro via hole contour, and the ellipticity coefficient of the micro via hole, judge whether the aperture of the micro via hole meets the production standard. According to the edge roughness of the micro via hole and the edge roughness threshold, judge whether the aperture of the micro via hole meets the production standard. Detect the depth and hole wall of the micro via hole according to the three-dimensional binary image data of the micro via hole, and judge whether the micro via hole meets the production standard. According to the coordinate data of the micro via hole, based on data fitting, obtain the contour curve function of the micro via hole. According to the contour curve function of the micro via hole, obtain the major axis information of the micro via hole contour and the minor axis information of the micro via hole contour. According to the sampling point number information, screen the coordinate data of the micro via hole to obtain the sampling point coordinate information. Based on the polar coordinate system, convert the sampling point coordinates and the micro via hole contour curve function to obtain the sampling point polar coordinate information and the micro via hole contour curve polar coordinate function. Taking the polar angle in the sampling point polar coordinates as the reference, obtain the contour point polar coordinate information corresponding to the sampling point in the micro via hole contour curve polar coordinate function. Take the difference between the sampling point coordinates and the contour point polar coordinates as the radial deviation value, obtain the radial deviation data, and obtain the edge roughness of the micro via hole according to the radial deviation data;

[0172] An information acquisition module, which is used to acquire the micro via hole image data, the two-dimensional image data of the micro via hole, and the three-dimensional image data of the micro via hole, perform gray processing on the micro via hole image data, and perform binary conversion on the processed gray image to obtain the binary image data of the micro via hole. Based on contour recognition, obtain the micro via hole contour information according to the two-dimensional binary image data of the micro via hole. Based on the device parameters, obtain the image resolution information according to the micro via hole image data, and obtain the horizontal step distance information based on the image resolution;

[0173] An image processing module, which is used to connect any two positions on the edge contour of the micro via according to the micro via contour information to obtain the contour feature line segment of the micro via, obtain the longest diameter and the shortest diameter of the micro via according to the contour feature line segment of the micro via, obtain the point cloud spacing information according to the minimum defect size information and the shortest diameter of the micro via, construct a point cloud model according to the point cloud spacing information and the two-dimensional binary image data of the micro via, and obtain the two-dimensional model of the micro via;

[0174] A display module, which interacts with the main control module and is used to display the two-dimensional binary image data of the micro via, the two-dimensional model of the micro via, the contour curve function of the micro via, and the marking information of the micro via.

[0175] The main control module specifically includes:

[0176] A control unit, which is used to obtain the contour curve function of the micro via based on data fitting according to the micro via coordinate data, obtain the major axis information and the minor axis information of the micro via contour according to the contour curve function of the micro via, screen the micro via coordinate data according to the sampling point number information to obtain the sampling point coordinate information, convert the sampling point coordinates and the contour curve function of the micro via based on the polar coordinate system to obtain the sampling point polar coordinate information and the contour curve polar coordinate function of the micro via, take the polar angle in the sampling point polar coordinates as the reference, obtain the contour point polar coordinate information corresponding to the sampling point in the contour curve polar coordinate function of the micro via, use the difference between the sampling point coordinates and the contour point polar coordinates as the radial deviation value to obtain the radial deviation data, and obtain the edge roughness of the micro via according to the radial deviation data;

[0177] An information receiving unit, which interacts with the information acquisition module and the image processing module and is used to receive data and transmit it to the judgment unit;

[0178] A judgment unit, which is used to judge whether the shortest diameter of the micro via meets the production standard according to the standard diameter of the micro via and the diameter difference threshold of the micro via, judge whether the adjusted point cloud spacing meets the point cloud model spacing setting standard according to the image resolution information, judge whether the aperture of the micro via meets the production standard according to the major axis information of the micro via contour, the minor axis information of the micro via contour, and the ellipticity coefficient of the micro via, judge whether the aperture of the micro via meets the production standard according to the edge roughness of the micro via and the edge roughness threshold, and judge whether the micro via meets the production standard according to the three-dimensional binary image data of the micro via.

[0179] The information acquisition module specifically includes:

[0180] A first acquisition unit, which is used to acquire micro via hole image data, two-dimensional micro via hole image data and three-dimensional micro via hole image data, perform gray processing on the micro via hole image data, and perform binary conversion on the processed gray image to obtain binary micro via hole image data;

[0181] A second acquisition unit, which is used to obtain micro via hole contour information based on contour recognition according to the two-dimensional binary micro via hole image data, obtain image resolution information based on the device parameters according to the micro via hole image data, and obtain lateral step distance information based on the image resolution.

[0182] An image processing module, specifically including:

[0183] An image processing unit, which is used to connect any two positions on the edge contour of the micro via hole according to the micro via hole contour information to obtain a micro via hole contour feature line segment, and obtain the longest diameter and the shortest diameter of the micro via hole according to the micro via hole contour feature line segment;

[0184] A model construction unit, which is used to obtain point cloud spacing information according to the minimum defect size information and the shortest diameter of the micro via hole, construct a point cloud model according to the point cloud spacing information and the two-dimensional binary micro via hole image data, and obtain a two-dimensional micro via hole model.

[0185] In summary, the advantages of the present invention are as follows: By using the shortest diameter of the micro via hole, the micro via hole is roughly detected, which improves the monitoring efficiency and avoids waste of resources. By using the minimum defect size information and the shortest diameter of the micro via hole, the point cloud spacing information is obtained, realizing the dynamic adjustment of the point cloud model construction and improving the accuracy of the model. By fitting the micro via hole contour, the micro via hole contour curve function is obtained, and the micro via hole is further analyzed through the micro via hole contour curve function, rather than directly comparing the micro via hole aperture with the production standard, avoiding misjudgment of the micro via hole and improving the micro via hole quality detection efficiency.

[0186] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art of this industry should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification is only the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A micro-through hole quality detection method based on data model feedback, characterized in that: include: Acquire micro-through-hole image data, wherein the micro-through-hole image data includes micro-through-hole two-dimensional image data and micro-through-hole three-dimensional image data; Performing grayscale processing on the micro-through-hole image data, and performing binary conversion on the processed grayscale image to obtain micro-through-hole binary image data; According to the micro-through-hole binary image data, two-dimensional binary image data of the micro-through-hole is obtained, wherein the two-dimensional binary image data of the micro-through-hole represents the plane data of the micro-through-hole, including the micro-through-hole position information and the opening shape information; According to the two-dimensional binary image data of the micro-through hole, a two-dimensional model of the micro-through hole is obtained based on the point cloud model construction; According to the two-dimensional model of the micro-through hole, it is judged whether the aperture of the micro-through hole meets the production standard. If not, the micro-through hole is unqualified. If so, three-dimensional binary image data of the micro-through hole is obtained according to the binary image data of the micro-through hole, and the three-dimensional binary image data of the micro-through hole includes a side wall image of the micro-through hole; According to the three-dimensional binary image data of the micro-through hole, the hole depth and hole wall of the micro-through hole are detected to determine whether the micro-through hole meets the production standard. If not, the micro-through hole is unqualified, and if so, the micro-through hole is qualified.

2. A micro-through hole quality detection method based on data model feedback according to claim 1, characterized in that: The method of obtaining a two-dimensional model of the micro-through hole based on the two-dimensional binary image data of the micro-through hole and based on the point cloud model construction specifically includes: According to the two-dimensional binary image data of the micro-through hole, based on contour recognition, the contour information of the micro-through hole is obtained; According to the micro-through hole contour information, any two positions in the micro-through hole edge contour are connected to obtain the micro-through hole contour feature line segment; The length of the characteristic line segment of the micro-through hole contour is measured, and the characteristic line segment of the micro-through hole contour with the longest length is taken as the longest diameter of the micro-through hole; Taking the longest diameter of the micro-through hole as a reference, the micro-through hole contour characteristic line segment perpendicular to the longest diameter of the micro-through hole and with the shortest length is taken as the shortest diameter of the micro-through hole; Based on the micro-via production standard, the micro-via standard diameter and the micro-via diameter difference threshold are obtained; According to the micro-through hole standard diameter and the micro-through hole diameter difference threshold, it is judged whether the shortest diameter of the micro-through hole meets the production standard. If not, the micro-through hole is unqualified and marked. If yes, the minimum defect size information is obtained based on the micro-through hole process technology. The minimum defect size information indicates the minimum size of the micro-through hole that cannot be ignored; According to the minimum defect size information and the shortest diameter of the micro-through hole, the point cloud spacing information is obtained; According to the point cloud spacing information and the two-dimensional binary image data of the micro-through hole, a point cloud model is constructed to obtain a two-dimensional model of the micro-through hole.

3. A micro-through hole quality detection method based on data model feedback according to claim 2, characterized in that: The step of obtaining the point cloud spacing information based on the minimum defect size information and the shortest diameter of the micro-through hole specifically includes: According to the micro-via standard diameter and the micro-via diameter difference threshold, the maximum tolerance diameter of the micro-via and the minimum tolerance diameter of the micro-via are obtained; The maximum tolerance diameter of the micro-via and the standard diameter of the micro-via are used as the first diameter threshold of the micro-via, and the minimum tolerance diameter of the micro-via and the standard diameter of the micro-via are used as the second diameter threshold of the micro-via; According to the minimum defect size information, based on the Nyquist sampling theorem, basic point cloud spacing information is obtained, where the basic point cloud spacing is half of the minimum defect size; Obtaining point cloud spacing information according to the first micro-through hole diameter threshold, the second micro-through hole diameter threshold, the shortest diameter of the micro-through hole, and basic point cloud spacing information; If the shortest diameter of the micro-through hole is within the second diameter threshold of the micro-through hole, the basic point cloud spacing is used as the average point cloud spacing when constructing the point cloud model; If the shortest diameter of the micro-through hole is within the first diameter threshold of the micro-through hole, the difference between the shortest diameter of the micro-through hole and the standard diameter of the micro-through hole is used as the diameter deviation value; The ratio of the diameter deviation value to the standard diameter of the micro-through hole is taken as the diameter deviation coefficient; According to the diameter deviation coefficient and the basic point cloud spacing, the adjustment point cloud spacing is obtained based on the dynamic reduction coefficient method; The adjusted point cloud spacing is used as the average point cloud spacing when constructing the point cloud model; The adjustment of the point cloud spacing is specifically as follows: Where D is the adjusted point cloud spacing, D0 is the basic point cloud spacing, w is the diameter deviation coefficient, d min is the shortest diameter of the micro-through hole, and d0 is the standard diameter of the micro-through hole.

4. A micro-through hole quality detection method based on data model feedback according to claim 3, characterized in that: The adjusting the point cloud spacing as the average point cloud spacing when constructing the point cloud model also includes: According to the micro-through hole image data, image resolution information is obtained based on the equipment parameters; According to the image resolution information, the adjusted point cloud spacing is compared with the image resolution to determine whether the adjusted point cloud spacing meets the point cloud model spacing setting standard; If the adjusted point cloud spacing is greater than the image resolution, the adjusted point cloud spacing complies with the point cloud model spacing setting standard, and the adjusted point cloud spacing is used as the average point cloud spacing when constructing the point cloud model; If the adjusted point cloud spacing is smaller than the image resolution, the adjusted point cloud spacing does not meet the point cloud model spacing setting standard, and the image resolution is used as the average point cloud spacing when constructing the point cloud model.

5. The micro-through hole quality detection method based on data model feedback according to claim 1 is characterized in that: The step of judging whether the micro-through hole diameter meets the production standard according to the micro-through hole two-dimensional model specifically includes: According to the micro-through hole two-dimensional model, based on the Cartesian coordinate system, the micro-through hole coordinate data is obtained, wherein the micro-through hole coordinate data represents the coordinate data of the point located at the edge contour of the micro-through hole in the micro-through hole two-dimensional model; According to the micro-through hole coordinate data, based on data fitting, a micro-through hole contour curve function is obtained; Wherein, if the longest diameter of the micro-through hole is equal to the shortest diameter of the micro-through hole, then according to the shortest diameter of the micro-through hole and based on the circle equation, the micro-through hole contour curve function is obtained; If the longest diameter of the micro-through hole is not equal to the shortest diameter of the micro-through hole, an ellipse equation is fitted based on random Hough transform and least square method according to the micro-through hole coordinate data to obtain the micro-through hole contour curve function; According to the micro-through hole profile curve function, the micro-through hole profile major axis information and the micro-through hole profile minor axis information are obtained; Based on the micro-through hole production standard, the micro-through hole ellipticity coefficient and the edge roughness threshold are obtained, wherein the micro-through hole ellipticity coefficient represents the maximum major axis / minor axis ratio allowed in the micro-through hole production; According to the micro-through hole profile major axis information, micro-through hole profile minor axis information and micro-through hole ellipticity coefficient, determine whether the micro-through hole aperture meets the production standard; if the ratio of the micro-through hole profile major axis to the micro-through hole profile minor axis is greater than the micro-through hole ellipticity coefficient, the micro-through hole is unqualified and is marked; if the ratio of the micro-through hole profile major axis to the micro-through hole profile minor axis is less than the micro-through hole ellipticity coefficient, obtain the sampling length information based on the micro-through hole edge roughness standard; Obtain image resolution information, and based on the image resolution, obtain lateral step information; According to the sampling length information and the lateral step information, the sampling point number information is obtained; According to the sampling point number information, the micro-through hole coordinate data is screened to obtain the sampling point coordinate information; Based on the polar coordinate system, the sampling point coordinates and the micro-through hole profile curve function are converted to obtain the sampling point polar coordinate information and the micro-through hole profile curve polar coordinate function; Taking the polar angle in the polar coordinates of the sampling point as a reference, obtaining the polar coordinate information of the contour point corresponding to the sampling point in the polar coordinate function of the micro-through hole contour curve, wherein the polar angle of the sampling point is the same as that of the contour point; The difference between the sampling point coordinates and the contour point polar coordinates is used as a radial deviation value to obtain radial deviation data; According to the radial deviation data, the edge roughness of the micro-through hole is obtained; According to the edge roughness of the micro-through hole and the edge roughness threshold, it is judged whether the aperture of the micro-through hole meets the production standard. If the edge roughness of the micro-through hole is greater than the edge roughness threshold, the micro-through hole is unqualified and marked. If the edge roughness of the micro-through hole is less than the edge roughness threshold, the aperture of the micro-through hole is qualified.

6. A micro-through hole quality detection method based on data model feedback according to claim 5, characterized in that: The detecting of the micro-through hole depth and hole wall according to the micro-through hole three-dimensional binary image data specifically includes: Acquire substrate thickness information and substrate roughness information, wherein the substrate thickness information includes thickness information of each layer in the substrate; Based on the application requirements of the micro-through hole, the calibrated hole depth information corresponding to each micro-through hole is obtained, where the calibrated hole depth is the thickness of the substrate that the micro-through hole needs to penetrate; According to the calibrated hole depth information and the substrate roughness, a micro-through hole depth threshold is obtained; Based on the three-dimensional binary image data of the micro-through hole, the hole depth information of the micro-through hole is obtained; According to the micro-via hole depth information and the micro-via hole depth threshold, it is judged whether the micro-via hole depth meets the production standard. If the micro-via hole depth exceeds the micro-via hole depth threshold, the micro-via hole is unqualified and marked. If the micro-via hole depth does not exceed the micro-via hole depth threshold, the micro-via hole depth meets the production standard. According to the three-dimensional binary image data of the micro-through hole, the taper angle information of the micro-through hole side wall is obtained; Based on the micro-via production standards, obtain the micro-via sidewall taper angle tolerance information; According to the micro-via sidewall taper angle information and the micro-via sidewall taper angle tolerance information, it is judged whether the micro-via sidewall meets the production standard. If the micro-via sidewall taper angle exceeds the micro-via sidewall taper angle tolerance, the micro-via is unqualified and is marked. If the micro-via sidewall taper angle does not exceed the micro-via sidewall taper angle tolerance, the micro-via hole depth meets the production standard.

7. A micro-through hole quality detection system based on data model feedback, used to implement the detection method according to any one of claims 1 to 6, characterized in that: include: The main control module is used to determine whether the shortest diameter of the micro-through hole meets the production standard according to the standard diameter of the micro-through hole and the threshold value of the difference in diameter of the micro-through hole, and to compare the adjusted point cloud spacing with the image resolution according to the image resolution information to determine whether the adjusted point cloud spacing meets the point cloud model spacing setting standard, and to determine whether the aperture of the micro-through hole meets the production standard according to the major axis information of the micro-through hole contour, the minor axis information of the micro-through hole contour and the ellipticity coefficient of the micro-through hole, and to determine whether the aperture of the micro-through hole meets the production standard according to the edge roughness of the micro-through hole and the edge roughness threshold value, and to detect the hole depth and hole wall of the micro-through hole according to the three-dimensional binary image data of the micro-through hole to determine whether the micro-through hole meets the production standard, and to determine whether the micro-through hole meets the production standard according to the coordinate data of the micro-through hole. , based on data fitting, obtain the micro-through-hole contour curve function, obtain the micro-through-hole contour major axis information and the micro-through-hole contour minor axis information according to the micro-through-hole contour curve function, screen the micro-through-hole coordinate data according to the sampling point number information, obtain the sampling point coordinate information, based on the polar coordinate system, convert the sampling point coordinates and the micro-through-hole contour curve function, obtain the sampling point polar coordinate information and the micro-through-hole contour curve polar coordinate function, take the polar angle in the sampling point polar coordinate as the reference, obtain the contour point polar coordinate information corresponding to the sampling point in the micro-through-hole contour curve polar coordinate function, take the difference between the sampling point coordinate and the contour point polar coordinate as the radial deviation value, obtain the radial deviation data, and obtain the micro-through-hole edge roughness according to the radial deviation data; An information acquisition module, the information acquisition module is used to acquire micro-through-hole image data, micro-through-hole two-dimensional image data and micro-through-hole three-dimensional image data, grayscale processing is performed on the micro-through-hole image data, and the processed grayscale image is binarized to acquire micro-through-hole binary image data, and based on the micro-through-hole two-dimensional binary image data and contour recognition, obtain micro-through-hole contour information, and based on the micro-through-hole image data and equipment parameters, obtain image resolution information, and based on the image resolution, obtain lateral step information; An image processing module, wherein the image processing module is used to connect any two positions in the edge contour of the micro-through hole according to the micro-through hole contour information, obtain the micro-through hole contour feature line segment, obtain the longest diameter of the micro-through hole and the shortest diameter of the micro-through hole according to the micro-through hole contour feature line segment, obtain the point cloud spacing information according to the minimum defect size information and the shortest diameter of the micro-through hole, construct a point cloud model according to the point cloud spacing information and the two-dimensional binary image data of the micro-through hole, and obtain the two-dimensional model of the micro-through hole; The display module interacts with the main control module and is used to display the micro-through hole binary image data, the micro-through hole two-dimensional model, the micro-through hole contour curve function and the marking information of the micro-through hole.

8. A micro-through hole quality detection system based on data model feedback according to claim 7, characterized in that: The main control module specifically includes: A control unit, the control unit is used to obtain a micro-through-hole contour curve function based on the micro-through-hole coordinate data and data fitting, obtain micro-through-hole contour major axis information and micro-through-hole contour minor axis information based on the micro-through-hole contour curve function, screen the micro-through-hole coordinate data according to the sampling point number information, obtain sampling point coordinate information, convert the sampling point coordinates and the micro-through-hole contour curve function based on the polar coordinate system, obtain the sampling point polar coordinate information and the micro-through-hole contour curve polar coordinate function, obtain the contour point polar coordinate information corresponding to the sampling point in the micro-through-hole contour curve polar coordinate function based on the polar angle in the sampling point polar coordinate, use the difference between the sampling point coordinates and the contour point polar coordinates as the radial deviation value, obtain radial deviation data, and obtain the micro-through-hole edge roughness according to the radial deviation data; An information receiving unit, which interacts with the information acquisition module and the image processing module to receive data and transmit it to the judgment unit; A judgment unit, wherein the judgment unit is used to judge whether the shortest diameter of the micro-through hole meets the production standard based on the standard diameter of the micro-through hole and the threshold value of the difference in diameter of the micro-through hole, judge whether the adjusted point cloud spacing meets the point cloud model spacing setting standard based on the image resolution information, judge whether the aperture of the micro-through hole meets the production standard based on the major axis information of the micro-through hole contour, the minor axis information of the micro-through hole contour and the ellipticity coefficient of the micro-through hole, judge whether the aperture of the micro-through hole meets the production standard based on the edge roughness of the micro-through hole and the edge roughness threshold, and judge whether the micro-through hole meets the production standard based on the three-dimensional binary image data of the micro-through hole.

9. The micro-through hole quality detection system based on data model feedback according to claim 7, characterized in that: The information acquisition module specifically includes: A first acquisition unit, the first acquisition unit is used to acquire micro-through-hole image data, micro-through-hole two-dimensional image data and micro-through-hole three-dimensional image data, grayscale processing is performed on the micro-through-hole image data, and the processed grayscale image is binarized to acquire micro-through-hole binary image data; The second acquisition unit is used to obtain micro-through hole contour information based on contour recognition according to the micro-through hole two-dimensional binary image data, obtain image resolution information based on the micro-through hole image data and device parameters, and obtain lateral step information based on the image resolution.

10. The micro-through hole quality detection system based on data model feedback according to claim 7, characterized in that: The image processing module specifically includes: An image processing unit, wherein the image processing unit is used to connect any two positions in the edge contour of the micro-through hole according to the micro-through hole contour information, obtain a characteristic line segment of the micro-through hole contour, and obtain the longest diameter and the shortest diameter of the micro-through hole according to the characteristic line segment of the micro-through hole contour; The model building unit is used to obtain point cloud spacing information according to the minimum defect size information and the shortest diameter of the micro-through hole, build a point cloud model according to the point cloud spacing information and the two-dimensional binary image data of the micro-through hole, and obtain a two-dimensional model of the micro-through hole.

Images