[go: up one dir, main page]

CN118961755A - An automatic defect recognition method for LED circuit boards based on machine learning - Google Patents

An automatic defect recognition method for LED circuit boards based on machine learning Download PDF

Info

Publication number
CN118961755A
CN118961755A CN202411450769.3A CN202411450769A CN118961755A CN 118961755 A CN118961755 A CN 118961755A CN 202411450769 A CN202411450769 A CN 202411450769A CN 118961755 A CN118961755 A CN 118961755A
Authority
CN
China
Prior art keywords
welding point
welding
point
led circuit
defect
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202411450769.3A
Other languages
Chinese (zh)
Other versions
CN118961755B (en
Inventor
张红波
刘培培
陈立冬
郑雨停
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Longnan Dingtai Electronic Technology Co ltd
Original Assignee
Longnan Dingtai Electronic Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Longnan Dingtai Electronic Technology Co ltd filed Critical Longnan Dingtai Electronic Technology Co ltd
Priority to CN202411450769.3A priority Critical patent/CN118961755B/en
Publication of CN118961755A publication Critical patent/CN118961755A/en
Application granted granted Critical
Publication of CN118961755B publication Critical patent/CN118961755B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/95—Investigating the presence of flaws or contamination characterised by the material or shape of the object to be examined
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Landscapes

  • Chemical & Material Sciences (AREA)
  • Pathology (AREA)
  • Analytical Chemistry (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biochemistry (AREA)
  • Immunology (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)

Abstract

一种基于机器学习的LED线路板自动缺陷识别方法，涉及缺陷识别技术领域，包括以下步骤：获得可用焊接面图像并获得各个焊接点的焊接点图像，分别获得焊接点中心坐标和焊接点边缘坐标并对各个焊接点的焊接点形态进行评估以判断是否存在形态异常焊接点，获得各个焊接点的焊接点位置信息，构建焊接点分布图并将各个焊接点划分入不同的密集区域，分别获得相邻焊接点之间的邻近距离并判断是否存在距离异常焊接点；对缺陷数据进行采集并构建缺陷预测模型，根据缺陷预测模型对LED线路板进行缺陷预测，生成缺陷预测信号并反馈；通过本发明的技术方案，能够及时地发现形态异常的焊接点和距离异常的焊接点。

A method for automatic defect recognition of an LED circuit board based on machine learning relates to the technical field of defect recognition, and comprises the following steps: obtaining an available welding surface image and obtaining welding point images of each welding point, respectively obtaining welding point center coordinates and welding point edge coordinates and evaluating the welding point morphology of each welding point to determine whether there is a welding point with abnormal morphology, obtaining welding point position information of each welding point, constructing a welding point distribution map and dividing each welding point into different dense areas, respectively obtaining the proximity distance between adjacent welding points and determining whether there is a welding point with abnormal distance; collecting defect data and constructing a defect prediction model, predicting defects of an LED circuit board according to the defect prediction model, generating a defect prediction signal and feeding back the signal; and through the technical solution of the present invention, welding points with abnormal morphology and welding points with abnormal distance can be found in time.

Description

Automatic defect identification method for LED circuit board based on machine learning
Technical Field
The invention relates to the technical field of defect identification, in particular to an automatic defect identification method for an LED circuit board based on machine learning.
Background
The defect identification of the LED circuit board is to detect various possible defects on the LED circuit board by utilizing computer vision and machine learning technologies, such as short circuit, open circuit, circuit board damage, poor welding and the like, and the method combines the image processing, feature extraction and machine learning technologies, so that the defect identification of the LED circuit board can be efficiently and accurately carried out, and the identification efficiency and the product quality are improved;
In the prior art, defect identification of an LED circuit board is mostly not fine enough, effective identification cannot be carried out on the form and the distance of each welding point, so that the identification accuracy is low, in the prior art, a means for predicting defects of the current LED circuit board by utilizing the defect data of the scrapped LED circuit board is also lacking, and the invention provides an automatic defect identification method of the LED circuit board based on machine learning.
Disclosure of Invention
The invention aims to provide an automatic defect identification method for an LED circuit board based on machine learning.
The aim of the invention can be achieved by the following technical scheme: an automatic defect identification method of an LED circuit board based on machine learning comprises the following steps:
Step S1: acquiring a welding surface image of the LED circuit board, and preprocessing the acquired welding surface image to obtain a corresponding available welding surface image;
step S2: obtaining a welding point image of each welding point in the available welding surface image, respectively obtaining a welding point center coordinate and a welding point edge coordinate in the obtained welding point image, evaluating the welding point form of each welding point according to the obtained welding point center coordinate and the obtained welding point edge coordinate, and judging whether a form abnormal welding point exists or not;
Step S3: obtaining the welding point position information of each welding point in the available welding surface image, constructing a corresponding welding point distribution diagram according to the obtained welding point position information, dividing each welding point into different dense areas in the constructed welding point distribution diagram, respectively obtaining the adjacent distances between adjacent welding points in the different dense areas, and judging whether the distance abnormal welding points exist or not;
step S4: collecting defect data of the scrapped LED circuit board, constructing a corresponding defect prediction model according to the collected defect data, predicting defects of the current LED circuit board according to the constructed defect prediction model, generating a corresponding defect prediction signal and feeding back the corresponding defect prediction signal.
Further, the process of acquiring the welding surface image of the LED circuit board and preprocessing the acquired welding surface image to obtain a corresponding available welding surface image includes:
And acquiring a welding surface image of the LED circuit board, and preprocessing the acquired welding surface image, wherein the preprocessing comprises denoising, graying and size adjustment, and the welding surface image after the preprocessing is marked as an available welding surface image.
Further, the process of obtaining the welding point image of each welding point in the available welding surface image and obtaining the center coordinates and the edge coordinates of the welding point in the obtained welding point image respectively includes:
and identifying the welding point in the available welding surface image by utilizing an edge detection algorithm, intercepting a partial image of the welding point to obtain a welding point image, obtaining the center of the welding point, constructing a welding point coordinate system, obtaining the center coordinate of the welding point, identifying the welding point edge of the welding point, obtaining a plurality of edge points on the welding point edge, and taking the coordinates of each edge point in the welding point coordinate system as the welding point edge coordinate.
Further, the process of evaluating the welding point form of each welding point according to the obtained central coordinate and the obtained edge coordinate of the welding point and judging whether the welding point with abnormal form exists comprises the following steps:
in a welding point coordinate system, obtaining edge distances between the edge coordinates of each welding point and the center coordinates of the welding point, obtaining edge standards of the welding point according to the edge distances of the same welding point, comparing the edge distances with the edge standards, and evaluating the welding point form of the welding point according to a comparison result to obtain an evaluation coefficient;
Setting an evaluation standard, comparing the evaluation coefficient with the evaluation standard, obtaining a morphological abnormality welding point according to a comparison result, generating a morphological abnormality signal and feeding the morphological abnormality signal back to related personnel.
Further, the process of obtaining the welding point position information of each welding point in the available welding surface image and constructing a corresponding welding point distribution map according to the obtained welding point position information includes:
Constructing a welding surface coordinate system in an available welding surface image, and obtaining welding point position information of each welding point in the welding surface coordinate system, wherein the welding point position information refers to coordinate information of the center of each welding point and corresponding edge points in the welding surface coordinate system;
and constructing a welding point distribution diagram of the LED circuit board according to the welding point position information, wherein the welding point distribution diagram is used for reflecting the distribution situation of the center and the edge points of each welding point on the LED circuit board.
Further, dividing each welding point into different dense areas in the constructed welding point distribution diagram, respectively obtaining adjacent distances between adjacent welding points in the different dense areas, and judging whether the abnormal welding points exist or not, wherein the process comprises the following steps:
In the welding point distribution map, the center and the edge points of each welding point are treated as one point, and clustering treatment is carried out on each point in the welding point distribution map by using a K-Means algorithm so as to divide the welding point distribution map into a plurality of different dense areas;
the coordinate information of each point in the dense area with the same degree of density is incorporated into the same dense coordinate set, the shortest distance between each point in the dense coordinate set is obtained, and the shortest distance between adjacent welding points is used as the adjacent distance according to the shortest distance;
and obtaining the adjacent coefficients between each adjacent welding point according to all adjacent distances in the dense coordinate set, setting an adjacent standard, comparing the adjacent coefficients with the adjacent standard, obtaining the distance abnormal welding point according to the comparison result, generating a distance abnormal signal and feeding the distance abnormal signal back to related personnel.
Further, the process of collecting the defect data of the scrapped LED circuit board and constructing a corresponding defect prediction model according to the collected defect data comprises the following steps:
collecting defect data of the scrapped LED circuit board, wherein the defect data refers to defect proportion of morphologically abnormal welding points and distance abnormal welding points in the LED circuit board to all points and working time of the LED circuit board;
The method comprises the steps of selecting a machine learning model as an initial defect prediction model, training and evaluating the initial defect prediction model by using defect data to obtain a latest defect prediction model, wherein the defect prediction model is used for predicting the working time of the defect prediction model according to the input defect data.
Further, the process of performing defect prediction on the current LED circuit board according to the constructed defect prediction model, generating a corresponding defect prediction signal and feeding back the signal comprises the following steps:
Obtaining defect data of a current LED circuit board, inputting the defect data into a defect identification model, predicting the working time length of the LED circuit board by using the defect identification model, obtaining the working time length of the current LED circuit board, comparing the working time length with the working time length, obtaining the LED circuit board in a potential fault state according to a comparison result, generating a defect prediction signal and feeding back the defect prediction signal to related personnel.
Compared with the prior art, the invention has the beneficial effects that:
1. The method and the device can evaluate the welding point form of each welding point by obtaining the central coordinate of the welding point and the edge coordinate of the welding point, are favorable for finding the abnormal welding point and feeding back in time, and can judge whether the distance between the adjacent welding points is too far or too near by classifying the distribution situation of the welding points to obtain the adjacent distance between the adjacent welding points in dense areas with different densities, so as to be favorable for finding the abnormal welding point and feeding back in time;
2. The defect proportion and the working time length occupied by the morphological abnormal welding points and the distance abnormal welding points in the scrapped LED circuit board are obtained, a corresponding defect identification model is constructed according to the corresponding relation between the obtained defect proportion and the working time length, whether the current LED circuit board has potential fault conditions or not can be judged according to the defect identification model, an effective defect prediction mechanism is formed, and defect identification of the LED circuit board can be better achieved.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
As shown in fig. 1, the automatic defect recognition method for the LED circuit board based on machine learning comprises the following steps:
Step S1: acquiring a welding surface image of the LED circuit board, and preprocessing the acquired welding surface image to obtain a corresponding available welding surface image;
step S2: obtaining a welding point image of each welding point in the available welding surface image, respectively obtaining a welding point center coordinate and a welding point edge coordinate in the obtained welding point image, evaluating the welding point form of each welding point according to the obtained welding point center coordinate and the obtained welding point edge coordinate, and judging whether a form abnormal welding point exists or not;
Step S3: obtaining the welding point position information of each welding point in the available welding surface image, constructing a corresponding welding point distribution diagram according to the obtained welding point position information, dividing each welding point into different dense areas in the constructed welding point distribution diagram, respectively obtaining the adjacent distances between adjacent welding points in the different dense areas, and judging whether the distance abnormal welding points exist or not;
step S4: collecting defect data of the scrapped LED circuit board, constructing a corresponding defect prediction model according to the collected defect data, predicting defects of the current LED circuit board according to the constructed defect prediction model, generating a corresponding defect prediction signal and feeding back the corresponding defect prediction signal.
It should be further noted that, in the implementation process, the process of collecting the soldering surface image of the LED circuit board and preprocessing the collected soldering surface image to obtain the corresponding available soldering surface image includes:
Collecting a welding surface image of an LED circuit board, wherein the welding surface generally refers to a copper foil area used for welding an LED element or other electronic elements, and the welding surface image refers to an original image of one side of the welding surface of the welded LED circuit board;
The acquired welding surface image is preprocessed, the preprocessing is used for optimizing the welding surface image so as to facilitate subsequent analysis and processing, the preprocessing comprises denoising, graying and size adjustment, and the welding surface image after the preprocessing is marked as an available welding surface image.
It should be further noted that, in the specific implementation process, a welding point image of each welding point is obtained from the available welding surface image, and the process of respectively obtaining the center coordinates of the welding point and the edge coordinates of the welding point in the obtained welding point image includes:
In the embodiment of the invention, the form of the welding point on the LED circuit board needs to be analyzed, wherein the welding point generally refers to the physical connection point of the LED element or other electronic elements and the LED circuit board;
In the obtained available welding surface image, each welding point is identified by utilizing an edge detection algorithm, partial images of each welding point and the peripheral area of each welding point are intercepted to obtain corresponding welding point images, each welding point is bound with the corresponding welding point image, and the welding point images only comprise single and complete welding points;
Taking a welding point image of any welding point as an example, marking the position of a welding hole corresponding to the welding point as a welding point center, constructing a welding point coordinate system of the welding point by taking the welding point center as a coordinate origin, marking the welding point center coordinate as (0, 0), identifying the welding point edge of the welding point by utilizing an edge detection algorithm, sequentially obtaining a plurality of edge points on the obtained welding point edge according to a fixed distance, marking the coordinate of each edge point in the welding point coordinate system as the welding point edge coordinate, and marking the coordinates as (x i,yi) respectively, wherein i=1, 2, … … and n, wherein n represents the number of the edge points.
It should be further noted that, in the specific implementation process, the process of evaluating the welding point form of each welding point according to the obtained center coordinates and the obtained edge coordinates of the welding point and judging whether the welding point with abnormal form exists includes:
The welding point shape is a shape surrounded by the welding point edges of each welding point, and most of the welding point shape is approximate to a circle under normal conditions, so in the embodiment of the invention, the welding point shape is defaulted to be a circle for evaluation, taking any welding point as an example, in a welding point coordinate system of the welding point, the edge distance between the edge coordinates of each welding point and the center coordinates of the welding point is obtained, and the obtained edge distance is marked as B i;
Taking the average value of the edge distances of the same welding point as the edge standard of the welding point, marking the average value as B 0, comparing the obtained edge distances with the edge standard, marking the obtained edge distances as a normal state if B i≤B0, marking the obtained edge distances as an abnormal state if B i＞B0, evaluating the welding point form of the welding point according to the comparison result, taking the ratio of the number of the edge distances marked as the normal state in the welding point to the number of all the edge distances as an evaluation coefficient of the welding point, and marking the obtained evaluation coefficient as P i;
Setting an evaluation standard P 0, comparing the obtained evaluation coefficient with the evaluation standard, if P i≥P0, marking the welding point as a morphological normal welding point, if P i＜P0, marking the welding point as a morphological abnormal welding point, generating a corresponding morphological abnormal signal, and feeding back the generated morphological abnormal signal to related personnel.
It should be further noted that, in the implementation process, the process of obtaining the welding point position information of each welding point in the available welding surface image and constructing the corresponding welding point distribution diagram according to the obtained welding point position information includes:
in the embodiment of the invention, the adjacent distance between the adjacent welding points on the LED circuit board is required to be analyzed, and whether the adjacent distance between the adjacent welding points is too far or too close is judged to realize another defect identification mode of the welding points on the LED circuit board;
In the obtained available welding surface image, constructing a welding surface coordinate system of the welding surface by taking any vertex as a coordinate origin, and obtaining welding point position information of each welding point in the constructed welding surface coordinate system, wherein the welding point position information refers to coordinate information of the center of each welding point and corresponding edge points in the welding surface coordinate system;
And constructing a welding point distribution diagram of the LED circuit board according to the obtained welding point position information, wherein the constructed welding point distribution diagram only comprises the welding point center and the edge point of each welding point and is used for reflecting the distribution condition of each welding point center and the edge point on the LED circuit board.
It should be further noted that, in the implementation process, dividing each welding point into different dense areas in the constructed welding point distribution diagram, and respectively obtaining the adjacent distances between adjacent welding points in the different dense areas and judging whether the distance abnormal welding points exist includes:
in the constructed welding point distribution diagram, the center and the edge points of each welding point are treated as one point, each point in the welding point distribution diagram is clustered by using a K-Means algorithm to divide the welding point distribution diagram into a plurality of different dense areas, the dense degree of each dense area is different, and each dense area is ordered according to the dense degree from high to low;
The coordinate information of each point in the dense area with the same dense degree is brought into the same dense coordinate set, any point in any dense coordinate set is taken as an example, the shortest distance between the point and other points in the dense coordinate set is obtained, the point corresponding to the smallest shortest distance is taken as the nearest welding point of the point, the nearest welding point of the point and the point are marked as adjacent welding points, and the shortest distance between the adjacent welding points is marked as an adjacent distance;
Obtaining all adjacent distances in the dense coordinate set by adopting the same method, numbering the obtained adjacent distances, and marking the adjacent distances as L j, wherein j=1, 2, … … and m, wherein m represents the number of the adjacent distances in the dense coordinate set, and obtaining the adjacent coefficient of each adjacent welding point according to the obtained adjacent distances, and marking the adjacent coefficient as R j;
Setting a proximity standard R 0, comparing the obtained proximity coefficient with the proximity standard, marking the proximity standard as a distance normal welding point if R j≤R0, marking the proximity standard as a distance abnormal welding point if R j＞R0, generating a corresponding distance abnormal signal, and feeding back the generated distance abnormal signal to related personnel.
It should be further noted that, in the implementation process, the process of collecting the defect data of the scrapped LED circuit board and constructing the corresponding defect prediction model according to the collected defect data includes:
Collecting defect data of the scrapped LED circuit board, wherein the defect data refers to the proportion of morphological abnormal welding points and distance abnormal welding points in the LED circuit board to all points, and the defect data is marked as defect proportion, and the defect data also comprises the working time of the LED circuit board, namely the time interval from delivery to scrapping of the LED circuit board;
Selecting a machine learning model as an initial defect prediction model, dividing the obtained defect data into a training set and a testing set, training the initial defect prediction model by using the training set, and obtaining a trained defect prediction model by learning the corresponding relation between the defect proportion and the working time;
and evaluating the trained defect prediction model by using a test set, tuning the trained defect prediction model according to an evaluation result to obtain a tuned defect prediction model, and taking the latest tuned defect prediction model as the latest defect prediction model, wherein the defect prediction model is used for predicting the working time of the defect prediction model according to the input defect data.
It should be further noted that, in the implementation process, the process of performing defect prediction on the current LED circuit board according to the constructed defect prediction model, generating a corresponding defect prediction signal and feeding back the signal includes:
obtaining defect data of a current LED circuit board, inputting the obtained defect data into a constructed defect identification model, and predicting the working time of the LED circuit board according to the input defect data by using the defect identification model;
The method comprises the steps of obtaining the current working time length of the LED circuit board, comparing the obtained working time length with the corresponding working time length, if the working time length is smaller than or equal to the working time length, not performing any other operation on the LED circuit board, if the working time length is larger than the working time length, marking the LED circuit board as a potential fault state, generating a corresponding defect prediction signal, and feeding the defect prediction signal back to related personnel.
The above embodiments are only for illustrating the technical method of the present invention and not for limiting the same, and it should be understood by those skilled in the art that the technical method of the present invention may be modified or substituted without departing from the spirit and scope of the technical method of the present invention.

Claims (8)

1.一种基于机器学习的LED线路板自动缺陷识别方法，其特征在于，包括以下步骤：1. A method for automatic defect recognition of LED circuit boards based on machine learning, characterized in that it comprises the following steps: 步骤S1：对LED线路板的焊接面图像进行采集，对所采集的焊接面图像进行预处理以获得相应的可用焊接面图像；Step S1: collecting the welding surface image of the LED circuit board, and preprocessing the collected welding surface image to obtain a corresponding usable welding surface image; 步骤S2：在可用焊接面图像中获得各个焊接点的焊接点图像，在所获得的焊接点图像中分别获得焊接点中心坐标和焊接点边缘坐标，根据所获得的焊接点中心坐标和焊接点边缘坐标对各个焊接点的焊接点形态进行评估并判断是否存在形态异常焊接点；Step S2: obtaining a welding point image of each welding point in the available welding surface image, obtaining the welding point center coordinates and welding point edge coordinates in the obtained welding point image, respectively, evaluating the welding point morphology of each welding point according to the obtained welding point center coordinates and welding point edge coordinates, and determining whether there is an abnormal welding point in morphology; 步骤S3：在可用焊接面图像中获得各个焊接点的焊接点位置信息，根据所获得的焊接点位置信息构建相应的焊接点分布图，在所构建的焊接点分布图中将各个焊接点划分入不同的密集区域，分别获得不同密集区域内相邻焊接点之间的邻近距离并判断是否存在距离异常焊接点；Step S3: obtaining welding point location information of each welding point in the available welding surface image, constructing a corresponding welding point distribution map according to the obtained welding point location information, dividing each welding point into different dense areas in the constructed welding point distribution map, respectively obtaining the proximity distances between adjacent welding points in different dense areas, and determining whether there are welding points with abnormal distances; 步骤S4：对已经报废的LED线路板的缺陷数据进行采集，根据所采集的缺陷数据构建相应的缺陷预测模型，根据所构建的缺陷预测模型对当前的LED线路板进行缺陷预测，生成相应的缺陷预测信号并反馈。Step S4: Collect defect data of scrapped LED circuit boards, build a corresponding defect prediction model based on the collected defect data, predict defects of the current LED circuit boards based on the built defect prediction model, generate corresponding defect prediction signals and feedback. 2.根据权利要求1所述的一种基于机器学习的LED线路板自动缺陷识别方法，其特征在于，对LED线路板的焊接面图像进行采集，对所采集的焊接面图像进行预处理以获得相应的可用焊接面图像的过程包括：2. The automatic defect recognition method for LED circuit boards based on machine learning according to claim 1 is characterized in that the process of collecting the welding surface image of the LED circuit board and preprocessing the collected welding surface image to obtain the corresponding usable welding surface image comprises: 对LED线路板的焊接面图像进行采集，对所采集的焊接面图像进行预处理，所述预处理包括去噪、灰度化、尺寸调整，将经过上述预处理后的焊接面图像标记为可用焊接面图像。The welding surface image of the LED circuit board is collected, and the collected welding surface image is preprocessed, wherein the preprocessing includes denoising, graying, and resizing, and the welding surface image after the above preprocessing is marked as a usable welding surface image. 3.根据权利要求2所述的一种基于机器学习的LED线路板自动缺陷识别方法，其特征在于，在可用焊接面图像中获得各个焊接点的焊接点图像，在所获得的焊接点图像中分别获得焊接点中心坐标和焊接点边缘坐标的过程包括：3. The automatic defect recognition method for LED circuit boards based on machine learning according to claim 2 is characterized in that the welding point image of each welding point is obtained in the available welding surface image, and the process of respectively obtaining the welding point center coordinates and the welding point edge coordinates in the obtained welding point image comprises: 在可用焊接面图像中利用边缘检测算法对其中的焊接点进行识别，对焊接点的局部图像进行截取以获得焊接点图像，获得焊接点的焊接点中心并构建焊接点坐标系，获得焊接点中心坐标，对焊接点的焊接点边缘进行识别，在焊接点边缘上获得若干个边缘点，将各个边缘点在焊接点坐标系中的坐标作为焊接点边缘坐标。An edge detection algorithm is used to identify the welding points in the available welding surface image, a local image of the welding point is intercepted to obtain a welding point image, the welding point center of the welding point is obtained and a welding point coordinate system is constructed to obtain the welding point center coordinates, the welding point edge of the welding point is identified, a number of edge points are obtained on the welding point edge, and the coordinates of each edge point in the welding point coordinate system are used as the welding point edge coordinates. 4.根据权利要求3所述的一种基于机器学习的LED线路板自动缺陷识别方法，其特征在于，根据所获得的焊接点中心坐标和焊接点边缘坐标对各个焊接点的焊接点形态进行评估并判断是否存在形态异常焊接点的过程包括：4. The automatic defect recognition method for LED circuit boards based on machine learning according to claim 3 is characterized in that the process of evaluating the soldering point morphology of each soldering point and judging whether there is a soldering point with abnormal morphology according to the obtained soldering point center coordinates and soldering point edge coordinates comprises: 在焊接点坐标系中，获得各个焊接点边缘坐标与焊接点中心坐标之间的边缘距离，根据同一焊接点的各个边缘距离获得该焊接点的边缘标准，将边缘距离与边缘标准进行比较，根据比较结果对焊接点的焊接点形态进行评估并获得评估系数；In the welding point coordinate system, the edge distance between the edge coordinates of each welding point and the center coordinates of the welding point is obtained, the edge standard of the welding point is obtained according to the edge distances of the same welding point, the edge distance is compared with the edge standard, and the welding point shape of the welding point is evaluated according to the comparison result to obtain the evaluation coefficient; 设置评估标准，将评估系数与评估标准进行比较，根据比较结果获得形态异常焊接点，生成形态异常信号并将其反馈至相关人员处。An evaluation standard is set, the evaluation coefficient is compared with the evaluation standard, the abnormal morphology welding points are obtained according to the comparison result, and a morphology abnormality signal is generated and fed back to the relevant personnel. 5.根据权利要求4所述的一种基于机器学习的LED线路板自动缺陷识别方法，其特征在于，在可用焊接面图像中获得各个焊接点的焊接点位置信息，根据所获得的焊接点位置信息构建相应的焊接点分布图的过程包括：5. The automatic defect recognition method for LED circuit boards based on machine learning according to claim 4 is characterized in that the process of obtaining the welding point position information of each welding point in the available welding surface image and constructing the corresponding welding point distribution map according to the obtained welding point position information comprises: 在可用焊接面图像中构建焊接面坐标系，在焊接面坐标系中获得各个焊接点的焊接点位置信息，所述焊接点位置信息是指各个焊接点中心以及相应的边缘点在该焊接面坐标系中的坐标信息；Constructing a welding surface coordinate system in the available welding surface image, and obtaining welding point position information of each welding point in the welding surface coordinate system, wherein the welding point position information refers to the coordinate information of the center of each welding point and the corresponding edge point in the welding surface coordinate system; 根据焊接点位置信息构建LED线路板的焊接点分布图，所述焊接点分布图用于反映各个焊接点中心以及边缘点在LED线路板上的分布情况。A welding point distribution map of the LED circuit board is constructed according to the welding point position information, and the welding point distribution map is used to reflect the distribution of the center and edge points of each welding point on the LED circuit board. 6.根据权利要求5所述的一种基于机器学习的LED线路板自动缺陷识别方法，其特征在于，在所构建的焊接点分布图中将各个焊接点划分入不同的密集区域，分别获得不同密集区域内相邻焊接点之间的邻近距离并判断是否存在距离异常焊接点的过程包括：6. The automatic defect recognition method for LED circuit boards based on machine learning according to claim 5 is characterized in that the process of dividing each welding point into different dense areas in the constructed welding point distribution map, obtaining the proximity distances between adjacent welding points in different dense areas and judging whether there are welding points with abnormal distances comprises: 在焊接点分布图中，将其中的各个焊接点中心以及边缘点均作为一个点进行处理，利用K-Means算法对焊接点分布图中的各个点进行聚类处理以将其划分为若干个不同的密集区域；In the welding point distribution map, the center and edge points of each welding point are treated as one point, and the K-Means algorithm is used to cluster the points in the welding point distribution map to divide them into several different dense areas; 将同一密集程度的密集区域内的各个点的坐标信息纳入同一个密集坐标集合，获得密集坐标集合中各点之间的最短距离，根据最短距离获得各点的相邻焊接点，将相邻焊接点之间的最短距离作为邻近距离；The coordinate information of each point in the dense area of the same density is included in the same dense coordinate set, the shortest distance between each point in the dense coordinate set is obtained, the adjacent welding points of each point are obtained according to the shortest distance, and the shortest distance between adjacent welding points is used as the proximity distance; 根据密集坐标集合中所有的邻近距离获得各个相邻焊接点之间的邻近系数，设置邻近标准，将邻近系数与邻近标准进行比较，根据比较结果获得距离异常焊接点，生成距离异常信号并将其反馈至相关人员处。According to all the proximity distances in the dense coordinate set, the proximity coefficient between each adjacent welding point is obtained, the proximity standard is set, the proximity coefficient is compared with the proximity standard, the distance abnormal welding point is obtained according to the comparison result, the distance abnormality signal is generated and fed back to the relevant personnel. 7.根据权利要求6所述的一种基于机器学习的LED线路板自动缺陷识别方法，其特征在于，对已经报废的LED线路板的缺陷数据进行采集，根据所采集的缺陷数据构建相应的缺陷预测模型的过程包括：7. The method for automatic defect recognition of LED circuit boards based on machine learning according to claim 6 is characterized in that the defect data of scrapped LED circuit boards are collected, and the process of constructing a corresponding defect prediction model according to the collected defect data comprises: 对已经报废的LED线路板的缺陷数据进行采集，所述缺陷数据是指该LED线路板中形态异常焊接点和距离异常焊接点占所有点的缺陷比例以及该LED线路板的可工作时长；Collect defect data of scrapped LED circuit boards, wherein the defect data refers to the defect ratio of abnormal soldering points in shape and abnormal soldering points in distance to all points in the LED circuit board and the working time of the LED circuit board; 选择机器学习模型作为初始的缺陷预测模型，利用缺陷数据对初始的缺陷预测模型进行训练和评估以获得最新的缺陷预测模型，所述缺陷预测模型用于根据所输入的缺陷数据对其可工作时长进行预测。A machine learning model is selected as the initial defect prediction model, and the initial defect prediction model is trained and evaluated using defect data to obtain the latest defect prediction model, which is used to predict the working time of the defect according to the input defect data. 8.根据权利要求7所述的一种基于机器学习的LED线路板自动缺陷识别方法，其特征在于，根据所构建的缺陷预测模型对当前的LED线路板进行缺陷预测，生成相应的缺陷预测信号并反馈的过程包括：8. The automatic defect recognition method for LED circuit boards based on machine learning according to claim 7 is characterized in that the process of performing defect prediction on the current LED circuit board according to the constructed defect prediction model, generating corresponding defect prediction signals and feeding back comprises: 获得当前的LED线路板的缺陷数据，将缺陷数据输入至缺陷识别模型中，利用缺陷识别模型对该LED线路板的可工作时长进行预测，获得当前的LED线路板的已工作时长，将已工作时长与可工作时长进行比较，根据比较结果获得潜在故障状态的LED线路板，生成缺陷预测信号并将其反馈至相关人员处。Obtain defect data of the current LED circuit board, input the defect data into a defect recognition model, use the defect recognition model to predict the working time of the LED circuit board, obtain the actual working time of the current LED circuit board, compare the actual working time with the working time, obtain the LED circuit board in a potential fault state based on the comparison result, generate a defect prediction signal and feed it back to relevant personnel.
CN202411450769.3A 2024-10-17 2024-10-17 An automatic defect recognition method for LED circuit boards based on machine learning Active CN118961755B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202411450769.3A CN118961755B (en) 2024-10-17 2024-10-17 An automatic defect recognition method for LED circuit boards based on machine learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202411450769.3A CN118961755B (en) 2024-10-17 2024-10-17 An automatic defect recognition method for LED circuit boards based on machine learning

Publications (2)

Publication Number Publication Date
CN118961755A true CN118961755A (en) 2024-11-15
CN118961755B CN118961755B (en) 2024-12-20

Family

ID=93389809

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202411450769.3A Active CN118961755B (en) 2024-10-17 2024-10-17 An automatic defect recognition method for LED circuit boards based on machine learning

Country Status (1)

Country Link
CN (1) CN118961755B (en)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101204031B1 (en) * 2011-07-20 2012-11-27 주식회사 쎄크 Nondestructive inspection method for soldering section
CN110660052A (en) * 2019-09-23 2020-01-07 武汉科技大学 A deep learning-based detection method for surface defects of hot-rolled strip steel
US20220163461A1 (en) * 2020-11-24 2022-05-26 Inventec (Pudong) Technology Corporation Defective Soldering Point Intensive Extent Analysis System For Solder Paste Inspection And Method Thereof
KR102590040B1 (en) * 2022-12-09 2023-10-18 크레플 주식회사 Automated system and method for real-time detection of process defects using a machine learning model
CN117405703A (en) * 2023-10-24 2024-01-16 湖北东禾电子科技有限公司 PCBA circuit board welding spot detection system based on image recognition
CN117635620A (en) * 2024-01-26 2024-03-01 深圳市嘉熠精密自动化科技有限公司 Circuit board defect detection method and system based on image processing
WO2024087640A1 (en) * 2022-10-26 2024-05-02 上海第二工业大学 Printed circuit board welding spot defect detection method based on digital image processing
CN118637738A (en) * 2024-07-29 2024-09-13 龙南鼎泰电子科技有限公司 An intelligent monitoring method and monitoring system for the reaction progress of circuit board etching waste liquid for automobiles

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101204031B1 (en) * 2011-07-20 2012-11-27 주식회사 쎄크 Nondestructive inspection method for soldering section
CN110660052A (en) * 2019-09-23 2020-01-07 武汉科技大学 A deep learning-based detection method for surface defects of hot-rolled strip steel
US20220163461A1 (en) * 2020-11-24 2022-05-26 Inventec (Pudong) Technology Corporation Defective Soldering Point Intensive Extent Analysis System For Solder Paste Inspection And Method Thereof
WO2024087640A1 (en) * 2022-10-26 2024-05-02 上海第二工业大学 Printed circuit board welding spot defect detection method based on digital image processing
KR102590040B1 (en) * 2022-12-09 2023-10-18 크레플 주식회사 Automated system and method for real-time detection of process defects using a machine learning model
CN117405703A (en) * 2023-10-24 2024-01-16 湖北东禾电子科技有限公司 PCBA circuit board welding spot detection system based on image recognition
CN117635620A (en) * 2024-01-26 2024-03-01 深圳市嘉熠精密自动化科技有限公司 Circuit board defect detection method and system based on image processing
CN118637738A (en) * 2024-07-29 2024-09-13 龙南鼎泰电子科技有限公司 An intelligent monitoring method and monitoring system for the reaction progress of circuit board etching waste liquid for automobiles

Also Published As

Publication number Publication date
CN118961755B (en) 2024-12-20

Similar Documents

Publication Publication Date Title
CN110021005B (en) Method and device for screening defects of circuit board and computer readable recording medium
CN117152123B (en) Detection positioning optimization method, system and storage medium for solder paste printing
US7127099B2 (en) Image searching defect detector
CN111242899B (en) Image-based flaw detection method and computer-readable storage medium
JP2006098151A (en) Pattern inspection device and pattern inspection method
Thielen et al. A machine learning based approach to detect false calls in SMT manufacturing
TWI648766B (en) Automated calibration sample selection for die-to-database mask inspection
TWI669519B (en) Board defect filtering method and device thereof and computer-readabel recording medium
CN113486622A (en) Chip failure analysis method and device, electronic equipment and storage medium
CN114078109A (en) Image processing method, electronic device, and storage medium
CN113469944A (en) Product quality inspection method and device and electronic equipment
CN118798233A (en) Scanning anomaly recognition method and system based on image analysis
CN113758524A (en) Circuit board processing monitoring system
CN119574582B (en) SMT surface defect visual inspection device based on AI image recognition
CN118037640B (en) Product detection method, system and computer medium based on intelligent design drawing recognition
CN118961755B (en) An automatic defect recognition method for LED circuit boards based on machine learning
JP2013205320A (en) Inspection condition determination method, inspection method, and inspection device
CN118761997A (en) A Printed Circuit Board Quality Assessment Method Based on Data Analysis
CN112702905B (en) Backtracking method and system for printed circuit board yield and production equipment error rate
CN120031870B (en) AI image detection method and system for electronic circuit board
KR20230138869A (en) Inspection system, training data generation apparatus, training data generation method, and program recorded on recording medium
CN115510888A (en) Two-dimensional code identification method and device, electronic equipment and storage medium
CN118302790A (en) Teaching data creation support device and teaching data creation support method
CN118761989B (en) Industrial product quality detection method and system for artificial intelligence
JP4155497B2 (en) Defect classification method, program, and defect classification apparatus

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant