A method and system for evaluating wire bonding quality on line

Abstract

The application belongs to the field of circuit packaging, and specifically discloses a wire bonding welding quality online evaluation method and system. The method is as follows: a standard wire template is created, then a detection image of a workpiece to be measured is collected, the detection image is processed to obtain the center coordinates of each wire to be measured, then the standard wire template is called according to the center coordinates of each wire to be measured to sequentially match the wires to be measured, so as to obtain the profile of each wire to be measured; the minimum surrounding rectangle of each wire to be measured is obtained according to the center coordinates and the profile, and the length, width and area of the minimum surrounding rectangle of the standard wire are compared with those of the minimum surrounding rectangle of each wire to be measured and weighted calculation is performed to output a quality evaluation score, thereby realizing online evaluation of the wire bonding welding quality. The application can accurately determine whether all the wires are detected, and can accurately reflect whether the wires have problems such as being too long, too short, having too much welding, having too little welding, being bent and the like, thereby realizing comprehensive judgment of the overall quality of the wires.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of circuit packaging, and more particularly, to a wire bonding welding quality online evaluation method and system. BACKGROUND

[0002] In the field of integrated circuit packaging with high frequency, high speed, high power and high temperature, wire bonding technology has very important significance, which is mainly used to connect the interfaces of chips and substrates through high-purity wires. The quality of wire bonding welding will directly affect the performance of electronic components, and reasonable detection and evaluation of the quality of wire bonding welding has significant meaning in production practice. The length of the wire is generally 15 μm to 250 μm. Traditionally, the main way to detect the quality of wire bonding is to observe it manually using a microscope. The inspector carefully checks each wire through the microscope to ensure that the position, shape and size of the wire meet the standard. This detection method is low in efficiency and precision, and is disturbed by human subjective factors, and has low reliability. At the same time, the number of wires on electronic components is large and the distribution is complex. The traditional manual detection method is time-consuming and laborious, and it is difficult to reasonably evaluate all the wires. The evaluation method and index are also difficult to have a standardized standard. Therefore, under the background that the application of wire bonding technology is becoming more and more widespread and important, a new wire bonding welding quality detection and evaluation method is needed, which can overcome the related pain points of traditional manual detection, realize high efficiency, high precision, high reliability, and standardized evaluation standard.

[0003] Some new detection methods have appeared, such as a detection method based on point cloud deep learning technology and a detection method based on multi-channel information fusion. CN118396961A discloses a SOP chip pin defect detection method and system based on point cloud deep learning. After preprocessing the point cloud image, the corresponding pin information is extracted, and a pre-constructed defect detection model is used to obtain the defect detection result of the SOP chip, realizing three-dimensional detection of the SOP chip pins. However, this method only focuses on the detection of pin-related defects and does not mention whether the wire as a whole is too long, too short, has multiple welds, or has too few welds, which cannot comprehensively judge the overall quality of the wire. CN117745723A discloses a chip wire bonding quality detection method, system and storage medium, and proposes a multi-channel information fusion image processing method. At the same time, for different quality defects of chip wires, a mathematical model for detecting the quality of chip wire bonding is established. However, this method ignores the detection of the number of wires and whether there are wires, and only focuses on the judgment of the quality of the pins and the collimation degree of the wires, which cannot guarantee that all the wires in the image can be detected. SUMMARY

[0004] In view of the defects of the prior art, the application provides a wire bonding welding quality online evaluation method and system, aiming to solve the problems that the existing wire evaluation method cannot comprehensively judge the overall quality of the wire and cannot guarantee that all wires are detected.

[0005] According to an aspect of the application, a wire bonding welding quality online evaluation method is provided, specifically:

[0006] S1 creating a standard wire template;

[0007] S2 collecting a detection image of a workpiece to be detected, processing the detection image to obtain the center coordinates of each wire to be detected, and then calling the standard wire template to match each wire to be detected in sequence according to the center coordinates of each wire to be detected to obtain the contour of each wire to be detected;

[0008] S3 obtaining the minimum surrounding rectangle of each wire to be detected according to the center coordinates and the contour obtained in step S2, and comparing and weighting the length, width and area of the minimum surrounding rectangle of the standard wire to output a quality evaluation score, thereby realizing online evaluation of the wire bonding welding quality.

[0009] Compared with the prior art, the above technical scheme conceived by the application can effectively determine whether all wires are detected by drawing the contour of all wires to be detected, and can comprehensively judge the overall quality of the wire to be detected by comparing and weighting the size of the minimum surrounding rectangle of the wire to be detected and the standard wire.

[0010] As a further preferred, step S1 specifically comprises:

[0011] S11 collecting a standard image of a standard wire, and then pre-processing the standard image to obtain template information;

[0012] S12 based on the edge information and the gray scale information of the pre-processed standard image, drawing the contour of the standard wire to obtain contour information;

[0013] S13 saving the template information and the contour information of the standard wire.

[0014] As a further preferred, in step S12, a planar projection invariant deformable model or an anisotropic scaling shape model is used to draw the contour of the standard wire.

[0015] As a further preferred, in step S2, when the wire to be detected is rotated and twisted or partially blocked in the detection image, a planar projection invariant deformable model is used for template matching; when the wire to be detected is scaled and deformed, an anisotropic scaling shape model is used for template matching.

[0016] As a further preferred, in step S3, the calculation method of the quality evaluation score is:

[0017] Score [i] = 0.7 x M L[i] + 0.2 x M W[i] + 0.1 x M A[i]

[0018] In the formula, Score [i] is the quality evaluation score of the i-th lead, M L[i] is the length score of the minimum surrounding rectangle of the i-th lead under test, M W[i] is the width score of the minimum surrounding rectangle of the i-th lead under test, and M A[i] is the area score of the minimum surrounding rectangle of the i-th lead under test. The length score, the width score and the area score are calculated by using the following formulae, respectively:

[0019] M [i] = min{S, S [i]} / max{S, S [i]}

[0020] In the formula, M [i] is the size score, S is the size of the minimum surrounding rectangle of the standard lead, and S [i] is the size of the minimum surrounding rectangle of the i-th lead under test.

[0021] According to another aspect of the present application, there is provided a lead bonding welding quality online evaluation system, which comprises a standard module, a lead detection module and a quality evaluation module, wherein:

[0022] The standard module is configured to create a standard lead template;

[0023] The lead detection module is configured to collect detection images of leads under test, process the detection images to obtain the center coordinates of the leads under test, and then call the standard lead template to match the leads under test one by one according to the center coordinates of the leads under test to obtain the contours of the leads under test; and the quality evaluation module is configured to obtain the minimum surrounding rectangle of each of the leads under test according to the center coordinates and the contours, compare the length, the width and the area of the minimum surrounding rectangle of each of the leads under test with those of the standard lead, and perform weighted calculation to output a quality evaluation score, thereby realizing online evaluation of the lead bonding welding quality.

[0024] As a further preferred, the specific process of the standard module to create the standard lead template is as follows: collecting a standard image of a standard lead, then pre-processing the standard image to obtain template information, then drawing the contour of the standard lead based on the edge information and the grayscale information of the pre-processed standard image to obtain contour information, and finally saving the template information and the contour information of the standard lead.

[0025] As further preferred, the standard module draws the contour of the standard lead using a planar projection invariant deformable model or an anisotropic scaling shape model.

[0026] As further preferred, when the to-be-tested lead has a rotational distortion or a partial occlusion in the detection image, the lead detection module uses a planar projection invariant deformable model for template matching; when the to-be-tested lead has a scaling deformation, the lead detection module uses an anisotropic scaling shape model for template matching.

[0027] As further preferred, the calculation method of the quality evaluation score in the quality evaluation module is:

[0028] Score [i] = 0.7 x M L[i] + 0.2 x M W[i] + 0.1 x M A[i]

[0029] In the formula, Score [i] is the quality evaluation score of the i-th lead, M L[i] is the length fraction of the minimum surrounding rectangle of the i-th to-be-tested lead, M W[i] is the width fraction of the minimum surrounding rectangle of the i-th to-be-tested lead, and M A[i] is the area fraction of the minimum surrounding rectangle of the i-th to-be-tested lead. The length fraction, the width fraction and the area fraction are respectively calculated using the following formula:

[0030] M [i] = min{S, S [i]} / max{S, S [i]}

[0031] In the formula, M [i] is the size fraction, S is the size of the minimum surrounding rectangle of the standard lead, and S [i] is the size of the minimum surrounding rectangle of the i-th to-be-tested lead.

[0032] Overall, compared with the prior art, the above technical solutions conceived by the present application mainly have the following technical advantages:

[0033] 1. The present application can accurately determine whether all leads are detected by calling the standard lead template to match the to-be-tested lead and obtain the contour thereof. Meanwhile, the present application can accurately reflect whether the leads have problems such as being too long, too short, having multiple welds, having few welds, being bent and the like by using the length, the width and the area of the minimum surrounding rectangle of the to-be-tested lead and the standard lead for comparison and weighted calculation to output the quality evaluation score, thereby comprehensively judging the overall quality of the leads.

[0034] 2. In particular, the application uses planar projection invariant deformable model or anisotropic scaling shape model when creating a standard lead template and matching the lead to be tested, which can effectively improve the lead detection accuracy;

[0035] 3. In addition, the application also optimizes the calculation method of the quality evaluation score, sets the score calculation formula, and performs weighted calculation on the length, width and area scores, which can further improve the detection accuracy of the overall quality of the lead. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 is a flowchart of the lead bonding and welding quality online evaluation method provided by the embodiment of the application;

[0037] Figure 2 is the contour of the standard lead manually outlined and drawn after selecting the standard lead in the embodiment of the application;

[0038] Figure 3 is the detection image of the workpiece to be tested in the embodiment of the application;

[0039] Figure 4 is the contour and quality evaluation score of all leads to be tested in the workpiece to be tested in the embodiment of the application;

[0040] Figure 5 is an application schematic diagram of the lead bonding and welding quality online evaluation system provided by the embodiment of the application.

[0041] In all the drawings, the same reference signs are used to represent the same elements or structures, wherein:

[0042] 1-movable workbench, 2-workpiece to be tested, 3-split light source, 4-industrial camera, 5-microscope, 6-lead bonding and welding quality online evaluation system. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical scheme and advantages of the application more clear and obvious, the application will be further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the application and not to limit the application.

[0044] As shown in Figure 1 , the application provides a lead bonding and welding quality online evaluation method, specifically:

[0045] S1 create a standard lead template;

[0046] S2 collects a detection image of the workpiece to be detected, processes the detection image to obtain the center coordinates of each lead to be detected, and then calls a standard lead template according to the center coordinates of each lead to be detected to sequentially match the leads to obtain the center coordinates and the contour of all the leads to be detected. The contour can be drawn to intuitively determine whether all the leads to be detected in the detection image are detected, thereby ensuring that all the leads to be detected are detected. When the template is called, parameters including but not limited to a scaling factor can be adjusted to match different circuit boards. After the parameters are adjusted, a preliminary detection is performed. The detection accuracy can be determined by a test picture or real-time monitoring (whether all the leads are detected and how the detection effect is). If the detection accuracy meets the requirements, it indicates that the parameters are appropriate, and the detection of all the leads is continued. If the detection accuracy does not meet the requirements, the parameter adjustment is continued until the detection accuracy meets the requirements. In actual application, the detection accuracy can be adjusted as needed.

[0047] S3 obtains the minimum surrounding rectangle of each lead to be detected according to the center coordinates and the contour obtained in step S2, and compares and weightedly calculates the length, the width and the area of the minimum surrounding rectangle of the standard lead to output a quality evaluation score. The leads that are too long or too short, curved, missing, overwelded or underwelded will all reduce the quality evaluation score, so as to realize the online evaluation of the quality of lead bonding welding.

[0048] The present application can effectively determine whether all the leads are detected by drawing the contour of all the leads to be detected. Meanwhile, the overall quality of the leads to be detected can be comprehensively determined by comparing and weightedly calculating the dimensions of the minimum surrounding rectangle of the leads to be detected and the standard lead.

[0049] Further, step S1 specifically includes:

[0050] S11 collects a standard image of a workpiece with standard leads through a microscope and an industrial camera, and then pre-processes the standard image to obtain template information. The template drawing method can adopt a gray matching model based on NCC, a feature-based matching model or a shape-based matching model. The template information includes de-graying and edge information.

[0051] S12 draws the contour of the standard lead based on the edge information and the gray information of the pre-processed standard image to obtain contour information. The contour is manually outlined by artificial manual method or automatically outlined by a program.

[0052] S13 saves the template information and the contour information of the standard lead. The contour information is mainly saved for subsequent detection of the leads to be detected, used for visual display of the final result, and used for determining whether all the leads to be detected are detected.

[0053] Further, in step S12 and step S2, in order to achieve the purpose of good detection effect on the lead on the circuit board of different types, the planar projection invariant deformable model or anisotropic scaling shape model is used to draw the contour of the standard lead. The above two models are existing mathematical methods, and the application of the above two models to the lead detection can detect all the leads in the image and output the center coordinates of each lead, effectively solving the problem of poor detection effect of the traditional template matching algorithm.

[0054] Template matching, i.e. searching for a certain block or several blocks in the whole image corresponding to the template image, considering that the to-be-detected lead shows the characteristics of light-dark-light as a whole in the image, the template matching algorithm can achieve good detection results.

[0055] The planar projection invariant deformable model algorithm is mainly used for detecting the projection image of a planar object under different viewing angles. It uses the property of planar object projection transformation, i.e. the projection of the object on the image plane still maintains certain geometric relationship even if the object is subjected to rotation, scaling, translation and other transformations. Meanwhile, the model mainly focuses on the geometric shape of the object and is less sensitive to light changes. The anisotropic scaling shape model algorithm is mainly used for detecting objects with specific shape and direction characteristics, and can capture the shape and direction information of the object. It establishes a statistical model by learning the shape and direction information of the object, which can describe the shape, direction and variation range of the object, and has certain robustness to scaling, rotation, translation and other transformations of the object.

[0056] Among them, the key parameters of the above two template matching algorithms include image pyramid level, model scaling factor, minimum score threshold, maximum overlap degree, etc. The image pyramid is a series of images generated by using the original image. The original image is the first layer of the image pyramid, which has the highest resolution. The higher the level, the smaller the image, and the lower the image resolution. In the image processing algorithm, the multi-resolution operation of the image pyramid can avoid falling into local points and enhance the robustness of the model. Selecting a suitable image pyramid level can greatly improve the search efficiency and search accuracy, and obtain the best matching result. The model scaling factor is used to set the upper limit and lower limit of the maximum size of the detected lead. Because the size of the template lead and the actual detected lead is not completely consistent, the size of the lead in the actual lead bonding and welding process is large or small. The minimum score threshold is used to exclude those matching results whose algorithm scores are lower than the threshold. Therefore, reasonable setting of the model scaling factor and the minimum score threshold can maximize the avoidance of false detection and missed detection of the lead, and improve the detection accuracy.

[0057] In step S2, the standard lead profile is transformed to each detected lead by affine transformation to show all the detection results. Specifically, the center coordinates of the detected lead are outputted, and then the affine transformation, which is composed of rotation and translation, is calculated according to the point correspondence and the two corresponding angles. The transformation matrix HomMat2D is composed of a rotation matrix R and a translation vector t:

[0058]

[0059] The standard lead template profile is transformed from the upper left corner to the center coordinates of each detected lead to show the detected lead results. The standard lead template profile is transformed to the center coordinates and angle of each detected lead after the following matrix calculation, and the transformed coordinates and angle of the standard lead template are outputted as (Row2, Column2, Angle2).

[0060]

[0061] Preferably, in step S2, the detection mode can be flexibly adjusted according to the lead type or actual working condition during the detection. When the lead to be detected is rotated and distorted in the detection image, a planar projection invariant deformable model is used for template matching; when the lead to be detected is scaled and deformed, an anisotropic scaling shape model is used for template matching.

[0062] Further, in step S3, to determine the geometric size of each lead to be detected, the minimum surrounding rectangle with an arbitrary direction of the profile of the standard lead and the profile of each detected lead to be detected is calculated, that is, the smallest rectangle among all the rectangles containing the standard lead or the lead to be detected (the minimum surrounding rectangle can be obtained by using the method in the prior art). The calculation of the rectangle is based on the center coordinates of the region pixels, that is, the center coordinates of the standard lead or the lead to be detected, so the minimum surrounding rectangle needs to be obtained according to the center coordinates and the profile.

[0063] The calculation method of the quality evaluation score is as follows:

[0064] Score [i] = 0.7 x M L[i] + 0.2 x M W[i] + 0.1 x M A[i]

[0065] In the formula, Score [i] is the quality evaluation score of the i-th lead, M L[i] is the length score of the minimum surrounding rectangle of the i-th lead to be detected, M W[i] is the width score of the minimum surrounding rectangle of the i-th lead to be detected, and M A[i]The area fraction, length fraction, width fraction, and area fraction of the smallest surrounding rectangle of the i-th lead to be tested are calculated using the following formulas:

[0066] M [i] =min{S, S [i]} / max{S,S [i]}

[0067] In the formula, M [i] S is a size fraction, where S is the size of the minimum surrounding rectangle of the standard leader. [i] Let be the size of the smallest surrounding rectangle of the i-th lead to be tested, where the dimensions mentioned above are the length, width, or area, respectively.

[0068] According to another aspect of this application, an online evaluation system for wire bonding welding quality is provided. This system includes a standard module, a wire inspection module, and a quality evaluation module, wherein:

[0069] The standard module is used to create standard lead templates, including acquiring a standard image of the standard lead, then preprocessing the standard image to obtain template information. The template drawing method can adopt an NCC-based grayscale matching model, a feature-based matching model, or a shape-based matching model. Then, based on the edge information and grayscale information of the preprocessed standard image, the outline of the standard lead is drawn to obtain the outline information. Finally, the template information and outline information of the standard lead are saved.

[0070] The lead wire detection module is used to acquire the detection image of the workpiece to be tested, and call the standard lead wire template to match the lead wires to be tested in the detection image in turn to obtain the center coordinates and contours of all the lead wires to be tested. Template matching means searching for a certain area or several areas in the whole image that correspond to the template image. Considering that the lead wires to be tested present a bright-dark-bright feature in the image as a whole, the template matching algorithm can achieve better detection results when used to detect the lead wires to be tested.

[0071] The quality evaluation module is used to obtain the minimum surrounding rectangle of each lead wire to be tested based on the center coordinates and contour, and compares and weights it with the length, width and area of ​​the minimum surrounding rectangle of the standard lead wire to output a quality evaluation score, thereby realizing online evaluation of lead wire bonding welding quality.

[0072] Furthermore, such as Figure 5 As shown, both the standard module and the lead wire detection module use image acquisition units to acquire standard images and images under test. Specifically, these units include a microscope 5, an industrial camera 4, and a split-type light source 3. The standard workpiece or the workpiece under test 2 is fixed on a movable worktable 1. The displacement of the standard workpiece or the workpiece under test 2 is controlled by the movable worktable 1, and the lens is aimed at the position where the lead wire is present to acquire standard images and images under test. These images are then sent to the online evaluation system for lead wire bonding welding quality 6 to display the detection results and scores.

[0073] Further, the lead patterns on different types of circuit boards can be different, to achieve the purpose of good detection effect on leads on different types of circuit boards, a planar projection invariant deformable model or an anisotropic scaling shape model is used to draw the contour of the standard lead. Both of the above-mentioned models are existing mathematical methods, and the application of the above-mentioned models in lead detection can detect all leads in the image and output the center coordinates of each lead; the standard lead contour is transformed to each detected lead by using affine transformation, so as to display all detection results.

[0074] After detecting the lead, the center coordinates of the lead are output, and then affine transformation composed of rotation and translation is calculated according to the point correspondence and two corresponding angles. The transformation matrix HomMat2D is composed of two parts, a rotation matrix R and a translation vector t:

[0075]

[0076] The standard lead template contour is transformed from the upper left corner to the center coordinates of each lead, so as to display the detected lead result. When the standard lead template contour is in the upper left corner, the center coordinates and angle are transmitted as (Row1, Column1, Angle1), and after the following matrix calculation, the coordinates and angle of the standard lead template transformed to the center of each lead are output as (Row2, Column2, Angle2).

[0077]

[0078] Preferably, in step S2, the detection mode can be flexibly adjusted according to the lead type or actual working condition during the detection, when the lead to be detected is rotated and distorted in the detection image, a planar projection invariant deformable model is used for template matching; when the lead to be detected is scaled and deformed, an anisotropic scaling shape model is used for template matching.

[0079] Further, to determine the geometric size of each lead to be detected, the minimum surrounding rectangle with an arbitrary direction of the contour of the standard lead and the contour of each detected lead to be detected is calculated, that is, the smallest rectangle (the calculation method of the minimum surrounding rectangle is an existing mathematical calculation method) among all rectangles containing the standard lead or the lead to be detected. The calculation of the rectangle is based on the center coordinates of the region pixels, that is, the center coordinates of the standard lead or the lead to be detected, so the minimum surrounding rectangle needs to be obtained according to the center coordinates and the contour.

[0080] The calculation method of the quality evaluation score is:

[0081] Score [i] = 0.7 x M L[i] + 0.2 x MW[i] +0.1 x M A[i]

[0082] Scorei = (L i / L) + (W i / W) + (A i / A) (1) [i] Scorei is the quality evaluation score of the i th lead, M L[i] L i is the length score of the minimum surrounding rectangle of the i th lead to be measured, M W[i] W i is the width score of the minimum surrounding rectangle of the i th lead to be measured, M A[i] A i is the area score of the minimum surrounding rectangle of the i th lead to be measured, the length score, the width score and the area score are calculated by using the following formula respectively:

[0083] M [i] = min{S, S [i]} / max{S, S [i]}

[0084] M [i] is the size score, S is the size of the minimum surrounding rectangle of the standard lead, S [i] is the size of the minimum surrounding rectangle of the i th lead to be measured, the size mentioned above is the length, the width or the area respectively.

[0085] The technical solutions provided by the present application are further described below according to specific embodiments.

[0086] S1 creates a standard lead template, as shown in FIG. 1, wherein the red color is the contour of the standard lead; Figure 2

[0087] S2 collects the detection image of the workpiece to be measured, as shown in FIG. 2, and calls the standard lead template to sequentially match the leads to be measured in the detection image to obtain the center coordinates and the contour of all the leads to be measured; Figure 3 S3 obtains the minimum surrounding rectangle of each of the leads to be measured according to the center coordinates and the contour obtained in step S2, and compares and weightedly calculates the length, the width and the area of the minimum surrounding rectangle with those of the standard lead to output the quality evaluation score, as shown in FIG. 3, wherein the red color is the contour of the lead to be measured, and the numbers above each of the leads to be measured are the label and the quality evaluation score of the lead respectively, for example, 1:78 means that the quality evaluation score of the 1 st lead to be measured is 78, and it is set that the quality evaluation score above 60 is qualified.

[0088] Figure 4

[0089] It is easy for those skilled in the art to understand that the above description is only the preferred embodiments of the present application, and is not used to limit the present application, and any modification, equivalent replacement and improvement made within the spirit and principle of the present application should be included in the protection scope of the present application.​​​

Claims

1. A method for evaluating wire bonding quality on line, characterized by, The online evaluation method specifically comprises the following steps: S1, creating a standard lead template, specifically comprising the following steps: S11, collecting a standard image of a standard lead, and then pre-processing the standard image to obtain template information; S12, based on the edge information and grayscale information of the pre-processed standard image, drawing the contour of the standard lead to obtain contour information; S13, saving the template information and contour information of the standard lead; S2, collecting a detection image of a workpiece to be tested, processing the detection image to obtain the center coordinates of each lead to be tested, and then calling the standard lead template to match each lead to be tested in sequence according to the center coordinates of each lead to be tested, including using affine transformation to transform the standard lead contour to each detected lead, to display all detection results, specifically comprising the following steps: After detecting the lead, the lead center coordinates are output, and an affine transformation, which is composed of rotation and translation, is calculated according to the point correspondence and two corresponding angles, and the transformation matrix HomMat 2 D It is composed of a rotation matrix R and a translation vector t: The standard lead contour is transformed to the center coordinates of each lead, to display the detected lead result, the center coordinates and angle of the standard lead contour are input as (Row1, Column1, Angle1), after matrix calculation, the coordinates and angle of the standard lead contour transformed to the center of each lead are output as (Row2, Column2, Angle2), When the lead to be tested is rotated and distorted, or partially occluded in the detection image, a planar projection invariant deformable model is used for template matching; when the lead to be tested is scaled and deformed, an anisotropic scaling shape model is used for template matching, to obtain the contour of each lead to be tested; S3, obtaining the minimum surrounding rectangle of each lead to be tested according to the center coordinates and contour obtained in step S2, and comparing and weighting the length, width and area of the minimum surrounding rectangle with those of the standard lead to output a quality evaluation score, thereby realizing online evaluation of the quality of lead bonding welding.

2. The wire bonding weld quality on-line evaluation method according to claim 1, wherein, In step S3, the calculation method of the quality evaluation score is as follows: wherein is a length score of the minimum surrounding rectangle of the i-th lead under test, is a width score of the minimum surrounding rectangle of the i-th lead under test, is an area score of the minimum surrounding rectangle of the i-th lead under test, The length score, the width score and the area score are calculated using the following equations, respectively: In the formula, is a size fraction, and the sizes are length, width, area, respectively, is the size of the minimum surrounding rectangle of the standard lead, is the size of the minimum surrounding rectangle of the ith lead under test.

3. A wire bonding quality on-line evaluation system characterized by comprising: The online evaluation system comprises a standard module, a lead detection module and a quality evaluation module, wherein: The standard module is used to create a standard lead template, specifically comprising the following steps: Collecting a standard image of a standard lead, and then pre-processing the standard image to obtain template information, then based on the edge information and grayscale information of the pre-processed standard image, drawing the contour of the standard lead to obtain contour information, and finally saving the template information and contour information of the standard lead; The lead detection module is used to collect a detection image of a workpiece to be tested, process the detection image to obtain the center coordinates of each lead to be tested, and then call the standard lead template to match each lead to be tested in sequence according to the center coordinates of each lead to be tested, including using affine transformation to transform the standard lead contour to each detected lead, to display all detection results, specifically comprising the following steps: After detecting the lead, the lead center coordinates are output, and an affine transformation, which is composed of rotation and translation, is calculated according to the point correspondence and two corresponding angles, and the transformation matrix HomMat 2 D It is composed of a rotation matrix R and a translation vector t: The standard lead contour is transformed to the center coordinates of each lead, to display the detected lead result, the center coordinates and angle of the standard lead contour are input as (Row1, Column1, Angle1), after matrix calculation, the coordinates and angle of the standard lead contour transformed to the center of each lead are output as (Row2, Column2, Angle2), When the to-be-tested lead wires are in rotation distortion, partial occlusion in the detection image, a planar projection invariant deformable model is used for template matching; when the to-be-tested lead wires are in scaling deformation, an anisotropic scaling shape model is used for template matching to obtain the profiles of the to-be-tested lead wires; The quality evaluation module is used for obtaining the minimum surrounding rectangle of each to-be-tested lead wire according to the center coordinates and the profile, and comparing and weightedly calculating the length, width and area of the minimum surrounding rectangle of the standard lead wire to output a quality evaluation score, thereby realizing online evaluation of the quality of lead bonding welding.

4. The wire bonding weld quality on-line evaluation system of claim 3, wherein, The calculation method of the quality evaluation score in the quality evaluation module is: wherein is a quality evaluation score of the i-th lead under test, is a length score of the minimum surrounding rectangle of the i-th lead under test, is a width score of the minimum surrounding rectangle of the i-th lead under test, is an area score of the minimum surrounding rectangle of the i-th lead under test, the length score, the width score and the area score are calculated by using the following equations, respectively: In the formula, is the size fraction, and the size is length, width, area, respectively, is the size of the minimum surrounding rectangle of the standard lead, is the size of the minimum surrounding rectangle of the i-th lead under test.

Images