Flexible circuit board quality detection system and method based on data analysis
By calculating the position offset vector of electronic components of the flexible circuit board and establishing a failure rate model, the problem of change detection before and after the bending test of the flexible circuit board is solved, high-precision automated detection is achieved, and manual detection is reduced.
Patent Information
- Application Number
- CN202411238104.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-05
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-09-05
AI Technical Summary
The prior art lacks effective methods to detect changes in flexible circuit boards before and after bending tests, and cannot meet the detection needs of high precision and high density.
By obtaining the reference image of the flexible circuit board, calculating the offset vector of the position of the electronic component, establishing a relationship model between the offset vector and the failure rate, combining machine vision algorithms for data analysis, identifying feature polygons and bending shafts, and performing image alignment and comparison Yes, generate the quality evaluation results of the flexible circuit board.
The accuracy of image positioning and comparison of flexible circuit boards is improved, the amount of manual detection is reduced, the quality model is constructed through historical detection data, the failure rate evaluation value is generated, and automated detection is realized.
Smart Images

Figure CN118762001B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of flexible circuit board detection, and in particular to a flexible circuit board quality detection system and method based on data analysis. Background Art
[0002] Flexible printed circuit (FPC), also known as soft circuit board, flexible circuit board, has the characteristics of light weight, thin thickness, free bending and folding, which makes full use of equipment space and has a wide range of application scenarios in mobile electronic devices, vehicle-mounted electronic devices, etc. With the rapid development of the electronics industry, circuit board design is becoming more and more high-precision and high-density. Traditional manual inspection methods can no longer meet production needs, and quality inspection needs to be carried out through automated means.
[0003] With the development of artificial intelligence and machine vision technology, quality inspection through image recognition and comparison has been widely used for general rigid circuit boards, such as PCBs. However, since flexible circuit boards need to undergo bending and folding tests compared to rigid circuit boards, it is necessary to compare the flexible circuit boards before and after the bending tests. The prior art lacks a method for comparing the changes of flexible circuit boards before and after the bending tests. Summary of the invention
[0004] The object of the present invention is to provide a flexible circuit board quality detection system and method based on data analysis to solve the problems raised in the above background technology.
[0005] In order to solve the above technical problems, the present invention provides the following technical solution: a flexible circuit board quality detection method based on data analysis, the method comprising:
[0006] Step S100: obtaining a reference image of the flexible circuit board, and obtaining a relative positional relationship between a bending axis of the flexible circuit board and positions of electronic components in the flexible circuit board from the reference image;
[0007] Step S200: comparing the images of the flexible circuit board before and after the bending test, and calculating the offset vector of the position of the electronic components in the image;
[0008] Step S300: obtaining the detection record of the flexible circuit board after the bending test, and establishing a relationship model between the offset vector and the failure rate;
[0009] Step S400: taking a certain circuit board among the flexible circuit boards as a target circuit board, acquiring images of the target circuit board before and after the bending test, and obtaining a quality evaluation result of the target circuit board.
[0010] Furthermore, step S100 includes:
[0011] Step S101: obtaining a design record of a certain flexible circuit board, obtaining a region where the certain flexible circuit board is bent in a usage scenario, obtaining an axis around which the bending region bends, recording the axis as a bending axis of the certain flexible circuit board, and making the bending axis correspond to the bending region;
[0012] Step S102: Selecting images of several feature positions in the reference image to form a reference position set, wherein the reference position set includes at least three feature positions that are not collinear in the reference image, connecting the feature positions in the reference position set to obtain a feature polygon, and obtaining a feature vector of the feature polygon, wherein the feature vector represents the direction or position of the feature polygon in the plane;
[0013] Step S103: establishing a plane coordinate system in the plane where the reference image is located, obtaining the relative position of the feature vector and the bending axis, and using the relative position as the target position of the bending axis relative to the feature vector;
[0014] Step S104: collecting a number of reference images of a certain flexible circuit board, marking the feature positions in the reference images, and performing data training through a machine vision algorithm to obtain a feature position recognition model for identifying the feature positions in the images;
[0015] According to the application scenario of the flexible circuit board, the part of the flexible circuit board that needs to be bent is marked to obtain the bending area of a certain flexible circuit board. During the bending process of the bending area in the application scenario, the axis around which the bending area is folded is recorded as the bending axis;
[0016] Establishing a connection between the bending region and the bending axis is equivalent to establishing a method for retrieving the bending region through the bending axis. When the bending axis can be obtained, the bending region corresponding to the bending axis can be obtained through the corresponding relationship.
[0017] By identifying several feature positions, locating the plane, connecting the identified feature positions into feature polygons, and then obtaining feature vectors representing the direction or position of the feature polygons in the plane, the flexible circuit board image alignment method is simplified. When collecting different photos of the same flexible circuit board, in order to compare the required sampling positions, the image needs to be positioned to prevent comparison errors caused by different positions or angles of the two images.
[0018] Further, step S200 includes:
[0019] Step S201: taking an electronic component in a bending area of a certain flexible circuit board as a target component, obtaining an image P1 of a certain circuit board in the certain flexible circuit board, obtaining an image d1 of the target component in the image P1 before the certain circuit board is bent, marking the contour of d1 in a plane coordinate system, and obtaining the centroid r1 of the contour of d1 in the plane coordinate system;
[0020] Step S202: after completing k bending tests, obtain an image P2 of a circuit board, obtain an image d2 of a target component in the image P2, mark the contour of d2 in the plane coordinate system, and obtain the centroid r2 of the contour of d2 in the plane coordinate system;
[0021] Step S203: Connect r1 and r2 in the plane coordinate system to obtain an offset vector γ12, wherein the direction of the offset vector γ12 is from r1 to r2;
[0022] The morphological features of the image corresponding to the target component are obtained by marking the contour. The target component is regarded as a geometric body with uniform density. When the morphological features of the image corresponding to the target component change, the centroid position of the image changes. The change of the image corresponding to the target component is described by representing the offset vector of the centroid change in the coordinate system.
[0023] Furthermore, step S300 includes:
[0024] Step S301: setting a first reference direction and a second reference direction in a plane coordinate system, wherein the first reference direction is collinear with the bending axis, and the second reference direction is a direction in the plane coordinate system that is not collinear with the first reference direction;
[0025] Step S302: Obtaining the fault record of the flexible circuit board after k bending tests from the detection record of the flexible circuit board, capturing the fault record corresponding to the offset vector γ12, and obtaining the failure rate m12 of the flexible circuit board when the offset vector γ12 appears;
[0026] Step S303: Decompose the vector γ12 in the first reference direction and the second reference direction.
[0027] γ12=c1·α+c2·β, where α represents the unit vector in the first reference direction, β represents the unit vector in the second reference direction, c1 represents the coefficient of the vector α, and c2 represents the coefficient of the vector β;
[0028] In the process of bending the flexible circuit board, by using the bending axis as the first reference direction for analyzing the deformation of the flexible circuit board, it is helpful to specifically analyze the deformation of the image in the flexible circuit board according to the bending shape of the flexible circuit board, and the second reference direction is used to assist the first reference direction to analyze the deformation amount;
[0029] Step S304: c1, c2 and m12 form a data record g, the coefficient of the first reference direction is used as the first independent variable, the coefficient of the second reference direction is used as the second independent variable, and the failure rate is used as the dependent variable. Several data records are obtained for function fitting to obtain a function model, which is a relationship model between the offset vector and the failure rate.
[0030] Furthermore, step S400 includes:
[0031] Step S401: obtaining an image e1 of a target circuit board before a bending test, where the target circuit board is a certain flexible circuit board, identifying feature positions in e1, obtaining a feature polygon u formed by the feature positions, and obtaining a feature vector ηu of the feature polygon;
[0032] Step S402: Obtain the target position of the bending axis relative to the feature vector ηu, mark the bending axis in e1, obtain the bending area Q in e1 according to the correspondence between the bending axis and the bending area, obtain the image h1 of the target component in the bending area Q, and obtain the centroid w1 of h1;
[0033] By acquiring the characteristic polygon in the image, determining the direction of the characteristic polygon, obtaining the position of the bending axis in the image e1, obtaining the bending area in e1, and retrieving the images of the components in the bending area, the scope of data comparison is reduced and the efficiency of comparison is improved;
[0034] Step S403: After completing k bending tests, obtain the image e2 of the target circuit board, obtain the characteristic polygon in e2, repeat the method of step S401 and step S402, obtain the centroid w2 of the target component image after completing k bending tests, and obtain the offset vector δ;
[0035] After the bending test, the graphics of the target component in image e2 are located by acquiring the characteristic polygon in image e2, and the image e1 and image e2 are aligned to improve the matching accuracy due to the change of placement position or observation angle;
[0036] Step S404: performing vector decomposition on the offset vector δ in the first reference direction and the second reference direction, δ=cx·α+cy·β, where cx represents the coefficient of the unit vector in the first reference direction, and cy represents the coefficient of the unit vector in the second reference direction;
[0037] Step S405: Substitute cx and cy into the relationship model between the offset vector and the failure rate to obtain a predicted value of the failure rate of the target circuit board.
[0038] In order to better implement the above method, a flexible circuit board quality inspection system based on data analysis is also proposed, the system includes: a bending management module, an offset vector management module, a relationship model management module and a comparison and evaluation module, wherein the bending management module is used to manage the relative position relationship between the bending axis and the position of the electronic components in the flexible circuit board, the offset vector management module is used to manage the offset vector of the position of the electronic components in the image, the relationship model management module is used to manage the relationship model between the offset vector and the failure rate, and the comparison and evaluation module is used to compare the image of the target circuit to calculate the predicted value of the failure rate of the target circuit board.
[0039] Furthermore, the bending management module includes: a bending area management unit, a feature position identification unit, a feature vector acquisition unit and a recognition model management unit, wherein the bending area management unit is used to obtain the bending axis in the flexible circuit board and manage the correspondence between the bending axis and the bending area, the feature position identification unit is used to identify the feature position in the image, the feature vector acquisition unit is used to obtain the feature vector of the feature polygon, and the recognition model management unit is used to manage the feature position recognition model.
[0040] Furthermore, the offset vector management module includes: a contour marking unit, a centroid management unit and an offset vector generation unit, wherein the contour marking unit is used to obtain the contour of the target component image, the centroid management unit is used to obtain the centroid of the contour image, and the offset vector generation unit is used to obtain the offset vector before and after the target component image changes.
[0041] Furthermore, the relationship model management module includes: a coordinate system management unit, a vector decomposition unit, a fault record management unit and a function fitting unit, wherein the coordinate system management unit is used to manage the plane coordinate system and the unit vectors in the plane coordinate system, the vector decomposition unit is used to perform vector decomposition on the offset vector, the fault record management unit is used to obtain the fault records of the flexible circuit board in the flexible circuit board detection results, and the function fitting unit is used to collect data records, perform function fitting, and obtain a relationship model between the offset vector and the failure rate.
[0042] Furthermore, the comparison and evaluation module includes: an image acquisition unit, an area acquisition unit, an offset vector acquisition unit and a failure rate prediction unit, wherein the image acquisition unit is used to acquire images of the target circuit before and after completing the bending test, the area acquisition unit is used to acquire the bending area in the target circuit, the offset vector acquisition unit is used to acquire the offset vector; and the failure rate prediction unit is used to acquire the failure rate of the target circuit based on the relationship model.
[0043] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: the present invention improves the accuracy of image positioning and comparison by extracting features from flexible circuit board images, compares the flexible circuit board images before and after the bending test, describes the changes in the images of electronic components in the images, and constructs a model of image changes related to the quality of flexible circuit boards in combination with historical detection data. Based on image recognition, an evaluation value of the failure rate of the flexible circuit board is generated, thereby reducing the amount of manual inspection of flexible circuit boards. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:
[0045] Figure 1 It is a structural schematic diagram of a flexible circuit board quality detection system based on data analysis of the present invention;
[0046] Figure 2 It is a flow chart of a method for quality inspection of a flexible circuit board based on data analysis of the present invention; DETAILED DESCRIPTION
[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0048] See also Figure 1 and Figure 2 , the present invention provides a technical solution:
[0049] Step S100: obtaining a reference image of the flexible circuit board, and obtaining a relative positional relationship between a bending axis of the flexible circuit board and positions of electronic components in the flexible circuit board from the reference image;
[0050] Wherein, step S100 includes:
[0051] Step S101: obtaining a design record of a certain flexible circuit board, obtaining a region where the certain flexible circuit board is bent in a usage scenario, obtaining an axis around which the bending region bends, recording the axis as a bending axis of the certain flexible circuit board, and making the bending axis correspond to the bending region;
[0052] Step S102: Selecting images of several feature positions in the reference image to form a reference position set, wherein the reference position set includes at least three feature positions that are not collinear in the reference image, connecting the feature positions in the reference position set to obtain a feature polygon, and obtaining a feature vector of the feature polygon, wherein the feature vector represents the direction or position of the feature polygon in the plane;
[0053] In an embodiment, a method for obtaining a feature vector of a feature polygon includes:
[0054] A1: Vertex representation conversion: convert the vertex representation of the polygon into vector form;
[0055] A2: Eigenvector calculation: Calculate the characteristics of these vectors, usually including: the length of the vector or the direction of the vector;
[0056] A3: Describe the direction of the characteristic polygon through the characteristics of the vector;
[0057] Step S103: establishing a plane coordinate system in the plane where the reference image is located, obtaining the relative position of the feature vector and the bending axis, and using the relative position as the target position of the bending axis relative to the feature vector;
[0058] Step S104: Collect several reference images of a certain flexible circuit board, mark the feature positions in the reference images, perform data training through a machine vision algorithm, and obtain a feature position recognition model for identifying the feature positions in the images.
[0059] Step S200: comparing the images of the flexible circuit board before and after the bending test, and calculating the offset vector of the position of the electronic components in the image;
[0060] Wherein, step S200 includes:
[0061] Step S201: taking an electronic component in a bending area of a certain flexible circuit board as a target component, obtaining an image P1 of a certain circuit board in the certain flexible circuit board, obtaining an image d1 of the target component in the image P1 before the certain circuit board is bent, marking the contour of d1 in a plane coordinate system, and obtaining the centroid r1 of the contour of d1 in the plane coordinate system;
[0062] Step S202: after completing k bending tests, obtain an image P2 of a circuit board, obtain an image d2 of a target component in the image P2, mark the contour of d2 in the plane coordinate system, and obtain the centroid r2 of the contour of d2 in the plane coordinate system;
[0063] Step S203: Connect r1 and r2 in the plane coordinate system to obtain an offset vector γ12, wherein the direction of the offset vector γ12 is from r1 to r2.
[0064] Step S300: obtaining the detection record of the flexible circuit board after the bending test, and establishing a relationship model between the offset vector and the failure rate;
[0065] Wherein, step S300 includes:
[0066] Step S301: setting a first reference direction and a second reference direction in a plane coordinate system, wherein the first reference direction is collinear with the bending axis, and the second reference direction is a direction in the plane coordinate system that is not collinear with the first reference direction;
[0067] Step S302: Obtaining the fault record of the flexible circuit board after k bending tests from the detection record of the flexible circuit board, capturing the fault record corresponding to the offset vector γ12, and obtaining the failure rate m12 of the flexible circuit board when the offset vector γ12 appears;
[0068] Step S303: Decompose the vector γ12 in the first reference direction and the second reference direction.
[0069] γ12=c1·α+c2·β, where α represents the unit vector in the first reference direction, β represents the unit vector in the second reference direction, c1 represents the coefficient of the vector α, and c2 represents the coefficient of the vector β;
[0070] Step S304: c1, c2 and m12 constitute a data record g, the coefficient of the first reference direction is used as the first independent variable, the coefficient of the second reference direction is used as the second independent variable, and the failure rate is used as the dependent variable, and several data records are obtained for function fitting to obtain a function model, which is a relationship model between the offset vector and the failure rate;
[0071] In the embodiment, an offset vector, that is, a vector coefficient pair, is taken as a statistical category, the number of all statistical cases in this category is collected and recorded as case0, the number of cases with faults is recorded as case1, and the fault rate λ when a certain offset vector appears is calculated, λ= case1 / case0;
[0072] Different vector coefficients and failure rates are collected, such as (3, 6, 0.007), (4, 2, 0.003), (-2, 1, 0.005), and (1.5, 3.6, 0.0002);
[0073] A three-dimensional coordinate system is established, with the coefficient of α as the corresponding value of the x-axis, the coefficient of β as the corresponding value of the y-axis, and the fault green as the corresponding value of the z-axis. Each data is represented in the three-dimensional coordinate system, and surface fitting is performed according to the electricity in the three-dimensional coordinate system to obtain the relationship model between the offset vector and the failure rate.
[0074] Step S400: taking a certain circuit board among the flexible circuit boards as a target circuit board, acquiring images of the target circuit board before and after the bending test, and obtaining a quality evaluation result of the target circuit board;
[0075] Wherein, step S400 includes:
[0076] Step S401: obtaining an image e1 of a target circuit board before a bending test, where the target circuit board is a certain flexible circuit board, identifying feature positions in e1, obtaining a feature polygon u formed by the feature positions, and obtaining a feature vector ηu of the feature polygon;
[0077] Step S402: Obtain the target position of the bending axis relative to the feature vector ηu, mark the bending axis in e1, obtain the bending area Q in e1 according to the correspondence between the bending axis and the bending area, obtain the image h1 of the target component in the bending area Q, and obtain the centroid w1 of h1;
[0078] Step S403: After completing k bending tests, obtain the image e2 of the target circuit board, obtain the characteristic polygon in e2, repeat the method of step S401 and step S402, obtain the centroid w2 of the target component image after completing k bending tests, and obtain the offset vector δ;
[0079] Step S404: performing vector decomposition on the offset vector δ in the first reference direction and the second reference direction, δ=cx·α+cy·β, where cx represents the coefficient of the unit vector in the first reference direction, and cy represents the coefficient of the unit vector in the second reference direction;
[0080] Step S405: Substituting cx and cy into the relationship model between the offset vector and the failure rate to obtain a predicted value of the failure rate of the target circuit board;
[0081] After receiving the predicted value of the failure rate, the relevant staff will further determine whether further fault detection of the flexible circuit board is needed through historical data analysis or experience data.
[0082] The system includes: a bending management module, an offset vector management module, a relational model management module and a comparison and evaluation module;
[0083] The bending management module is used to manage the relative position relationship between the bending axis and the position of the electronic components in the flexible circuit board, wherein the bending management module includes: a bending area management unit, a feature position recognition unit, a feature vector acquisition unit and a recognition model management unit, wherein the bending area management unit is used to obtain the bending axis in the flexible circuit board and manage the corresponding relationship between the bending axis and the bending area, the feature position recognition unit is used to recognize the feature position in the image, the feature vector acquisition unit is used to obtain the feature vector of the feature polygon, and the recognition model management unit is used to manage the feature position recognition model;
[0084] The offset vector management module is used to manage the offset vector of the position of the electronic component in the image, wherein the offset vector management module includes: a contour marking unit, a centroid management unit and an offset vector generation unit, wherein the contour marking unit is used to obtain the contour of the target component image, the centroid management unit is used to obtain the centroid of the contour image, and the offset vector generation unit is used to obtain the offset vector before and after the target component image changes;
[0085] The relationship model management module is used to manage the relationship model between the offset vector and the failure rate, wherein the relationship model management module includes: a coordinate system management unit, a vector decomposition unit, a fault record management unit and a function fitting unit, wherein the coordinate system management unit is used to manage the plane coordinate system and the unit vector in the plane coordinate system, the vector decomposition unit is used to perform vector decomposition on the offset vector, the fault record management unit is used to obtain the fault record of the flexible circuit board in the flexible circuit board detection result, and the function fitting unit is used to collect data records, perform function fitting, and obtain the relationship model between the offset vector and the failure rate;
[0086] Among them, the comparison and evaluation module is used to compare the image of the target circuit and calculate the failure rate prediction value of the target circuit board. Among them, the comparison and evaluation module includes: an image acquisition unit, an area acquisition unit, an offset vector acquisition unit and a failure rate prediction unit. Among them, the image acquisition unit is used to acquire the image of the target circuit before and after completing the bending test, the area acquisition unit is used to acquire the bending area in the target circuit, the offset vector acquisition unit is used to acquire the offset vector. The failure rate prediction unit is used to acquire the failure rate of the target circuit according to the relationship model.
[0087] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.
[0088] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. Flexible circuit board quality inspection method based on data analysis, It is characterized in that The method comprises the following steps: Step S100: obtaining a reference image of the flexible circuit board, and obtaining a relative positional relationship between a bending axis of the flexible circuit board and positions of electronic components in the flexible circuit board from the reference image; Step S200: comparing the images of the flexible circuit board before and after the bending test, and calculating the offset vector of the position of the electronic components in the image; Step S300: obtaining a detection record of a flexible circuit board after a bending test, and establishing a relationship model between the offset vector and the failure rate; Step S400: taking a certain circuit board among the flexible circuit boards as a target circuit board, acquiring images of the target circuit board before and after the bending test, and obtaining a quality evaluation result of the target circuit board; Step S400 includes: Step S401: obtaining an image e1 of a target circuit board before a bending test, wherein the target circuit board is a certain flexible circuit board, identifying characteristic positions in e1, obtaining a characteristic polygon u formed by the characteristic positions, and obtaining a characteristic vector ηu of the characteristic polygon; Step S402: Obtain the target position of the bending axis relative to the feature vector ηu, mark the bending axis in e1, obtain the bending area Q in e1 according to the correspondence between the bending axis and the bending area, obtain the image h1 of the target component in the bending area Q, and obtain the centroid w1 of h1; Step S403: After completing k bending tests, obtain the image e2 of the target circuit board, obtain the characteristic polygon in e2, repeat the method of step S401 and step S402, obtain the centroid w2 of the target component image after completing k bending tests, and obtain the offset vector δ; Step S404: performing vector decomposition on the offset vector δ in the first reference direction and the second reference direction, δ= cx·α+cy·β, where cx represents the coefficient of the unit vector in the first reference direction, and cy represents the coefficient of the unit vector in the second reference direction; Step S405: Substitute cx and cy into the relationship model between the offset vector and the failure rate to obtain a predicted value of the failure rate of the target circuit board.
2. The method for detecting quality of a flexible circuit board based on data analysis according to claim 1, Features: Step S100 includes: Step S101: obtaining a design record of a certain flexible circuit board, obtaining a region where the certain flexible circuit board is bent in a usage scenario, obtaining an axis around which the bending region is bent, recording the axis as a bending axis of the certain flexible circuit board, and making the bending axis correspond to the bending region; Step S102: Selecting images of several feature positions in the reference image to form a reference position set, wherein the reference position set includes at least three feature positions that are not collinear in the reference image, connecting the feature positions in the reference position set to obtain a feature polygon, and obtaining a feature vector of the feature polygon, wherein the feature vector represents the direction or position of the feature polygon in the plane; Step S103: establishing a plane coordinate system in the plane where the reference image is located, obtaining the relative position of the feature vector and the bending axis, and using the relative position as the target position of the bending axis relative to the feature vector; Step S104: collecting a number of reference images of the certain flexible circuit board, marking the feature positions in the reference images, performing data training through a machine vision algorithm, and obtaining a feature position recognition model for identifying the feature positions in the images.
3. The method for detecting quality of a flexible circuit board based on data analysis according to claim 2, Features: Step S200 includes: Step S201: taking an electronic component in a bending area of the certain flexible circuit board as a target component, obtaining an image P1 of a certain circuit board in the certain flexible circuit board, obtaining an image d1 of the target component in the image P1 before the certain circuit board is bent, marking the contour of d1 in the plane coordinate system, and obtaining the centroid r1 of the contour of d1 in the plane coordinate system; Step S202: after completing k bending tests, obtaining an image P2 of the circuit board, obtaining an image d2 of a target component in the image P2, marking the contour of d2 in the plane coordinate system, and obtaining a centroid r2 of the contour of d2 in the plane coordinate system; Step S203: Connect r1 and r2 in the plane coordinate system to obtain an offset vector γ12, wherein the direction of the offset vector γ12 is from r1 to r2.
4. The method for detecting quality of a flexible circuit board based on data analysis according to claim 3, Features: Step S300 includes: Step S301: setting a first reference direction and a second reference direction in the plane coordinate system, wherein the first reference direction is collinear with the bending axis, and the second reference direction is a direction in the plane coordinate system that is not collinear with the first reference direction; Step S302: Obtaining the fault record of the flexible circuit board after k bending tests from the detection record of the flexible circuit board, capturing the fault record corresponding to the offset vector γ12, and obtaining the failure rate m12 of the flexible circuit board when the offset vector γ12 appears; Step S303: Decompose the vector γ12 in the first reference direction and the second reference direction. γ12=c1·α+c2·β, where α represents the unit vector in the first reference direction, β represents the unit vector in the second reference direction, c1 represents the coefficient of the vector α, and c2 represents the coefficient of the vector β; Step S304: c1, c2 and m12 form a data record g, the coefficient of the first reference direction is used as the first independent variable, the coefficient of the second reference direction is used as the second independent variable, and the failure rate is used as the dependent variable, and several data records are obtained for function fitting to obtain a function model, which is a relationship model between the offset vector and the failure rate.
5. A flexible circuit board quality detection system for use in the flexible circuit board quality detection method based on data analysis as claimed in any one of claims 1 to 4, It is characterized in that The system includes the following modules: a bending management module, an offset vector management module, a relationship model management module and a comparison and evaluation module, wherein the bending management module is used to manage the relative position relationship between the bending axis and the position of the electronic components in the flexible circuit board, the offset vector management module is used to manage the offset vector of the position of the electronic components in the image, the relationship model management module is used to manage the relationship model between the offset vector and the failure rate, and the comparison and evaluation module is used to compare the image of the target circuit to calculate the predicted value of the failure rate of the target circuit board.
6. The flexible circuit board quality inspection system according to claim 5, Features: The bending management module includes: a bending area management unit, a feature position identification unit, a feature vector acquisition unit and a recognition model management unit, wherein the bending area management unit is used to obtain the bending axis in the flexible circuit board and manage the correspondence between the bending axis and the bending area, the feature position identification unit is used to identify the feature position in the image, the feature vector acquisition unit is used to obtain the feature vector of the feature polygon, and the recognition model management unit is used to manage the feature position recognition model.
7. The flexible circuit board quality inspection system according to claim 5, Features: The offset vector management module includes: a contour marking unit, a centroid management unit and an offset vector generation unit, wherein the contour marking unit is used to obtain the contour of the target component image, the centroid management unit is used to obtain the centroid of the contour image, and the offset vector generation unit is used to obtain the offset vector before and after the target component image changes.
8. The flexible circuit board quality inspection system according to claim 5, Features: The relational model management module includes: a coordinate system management unit, a vector decomposition unit, a fault record management unit and a function fitting unit, wherein the coordinate system management unit is used to manage the plane coordinate system and the unit vectors in the plane coordinate system, the vector decomposition unit is used to perform vector decomposition on the offset vector, the fault record management unit is used to obtain the fault records of the flexible circuit board in the flexible circuit board detection results, and the function fitting unit is used to collect data records, perform function fitting, and obtain a relationship model between the offset vector and the failure rate.
9. The flexible circuit board quality inspection system according to claim 5, Features: The comparison and evaluation module includes: an image acquisition unit, an area acquisition unit, an offset vector acquisition unit and a failure rate prediction unit, wherein the image acquisition unit is used to acquire images of the target circuit before and after completing the bending test, the area acquisition unit is used to acquire the bending area in the target circuit, the offset vector acquisition unit is used to acquire the offset vector, and the failure rate prediction unit is used to acquire the failure rate of the target circuit based on the relationship model.
Citation Information
Cited By
Identification and evaluation method and system based on multi-modal data analysis
CN119397486A
Flexible circuit board production quality monitoring method and system based on data feedback
CN120704266A
A data feedback-based flexible circuit board production quality monitoring method and system
CN120704266B
