WO2017092427A1 - 一种电子元件定位方法及装置 - Google Patents
一种电子元件定位方法及装置 Download PDFInfo
- Publication number
- WO2017092427A1 WO2017092427A1 PCT/CN2016/096896 CN2016096896W WO2017092427A1 WO 2017092427 A1 WO2017092427 A1 WO 2017092427A1 CN 2016096896 W CN2016096896 W CN 2016096896W WO 2017092427 A1 WO2017092427 A1 WO 2017092427A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- image
- electronic component
- sub
- predetermined electronic
- target sub
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000001514 detection method Methods 0.000 claims abstract description 71
- 238000012549 training Methods 0.000 claims abstract description 53
- 238000010586 diagram Methods 0.000 description 11
- 230000006872 improvement Effects 0.000 description 7
- 239000003990 capacitor Substances 0.000 description 4
- 230000007547 defect Effects 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 238000013461 design Methods 0.000 description 3
- 238000003780 insertion Methods 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 230000037431 insertion Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000005868 electrolysis reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 239000003550 marker Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012634 optical imaging Methods 0.000 description 1
Images
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
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/28—Testing of electronic circuits, e.g. by signal tracer
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
Definitions
- the invention relates to the field of automatic optical detection, and in particular to a method and a device for positioning electronic components.
- Automatic optical inspection refers to the use of optical imaging to obtain the surface state of the finished product, and image processing to detect the presence of foreign matter or surface defects on the surface of the finished product.
- automatic optical inspection is widely used for quality inspection of circuit boards.
- the relevant detecting device automatically scans the circuit board to acquire an image, extracts a partial image of each electronic component, and uses image processing technology to determine whether the electronic components on the circuit board have defects such as mis-insertion, missing insertion or reverse insertion. Finally, the electronic components with suspected defects are displayed or marked for easy viewing and maintenance.
- an object of the present invention is to provide an electronic component positioning method and apparatus that can quickly and accurately locate the position of an electronic component on an image of a circuit board.
- the invention provides a method for positioning an electronic component, comprising the following steps:
- the detection window based on the preset size intercepts the sub-image on the image of the circuit board to be detected
- the sub-image is detected according to a detection model of the predetermined predetermined electronic component; wherein The detection model of the predetermined electronic component is obtained by training a positive sample picture and a negative sample picture, the positive sample picture being a picture containing the predetermined electronic component, and the negative sample picture is not including the predetermined electronic a picture of the component;
- the sub-image is an image of the predetermined electronic component
- the sub-image is marked as a target sub-image, and position information of the target sub-image on the to-be-detected circuit board image is recorded.
- the method before the detecting the sub-image according to the detection model of the established predetermined electronic component, the method further includes:
- the first weak classifier is learned by using the adaboost cascade classifier algorithm for N training samples, and the training sample determined by the first weak classifier as a positive sample picture together with other new training samples constitutes a new N.
- Training samples, the second weak classifier is learned by the adaboost cascade classifier algorithm for the N training samples; the training samples include a positive sample picture and a negative sample picture, and N is an integer greater than 1;
- the at least two weak classifiers obtained by successive iterations are cascaded to form a detection model of the predetermined electronic component.
- the size of the positive sample picture is a normalized size
- the size of any of the negative sample pictures is not smaller than the normalized size
- the predetermined electronic component is an electronic component having a uniform aspect ratio.
- the sub-image when the sub-image is determined to be an image of the predetermined electronic component, the sub-image is marked as a target sub-image, and the target sub-image is recorded on the to-be-detected circuit board After the location information on the image, it also includes:
- the sub-image when the sub-image is determined to be an image of the predetermined electronic component, the sub-image is marked as a target sub-image, and the target sub-image is recorded on the to-be-detected circuit board After the location information on the image, it also includes:
- the invention also provides an electronic component positioning device, comprising:
- An image intercepting unit configured to intercept a sub-image on the image of the circuit board to be detected based on the detection window of the preset size
- a detecting unit configured to detect the sub-image according to a detection model of a predetermined electronic component; wherein the detection model of the predetermined electronic component is obtained by training a positive sample image and a negative sample image, where the positive sample image is included a picture of the predetermined electronic component, the negative sample picture being a picture not including the predetermined electronic component;
- a position recording unit configured to mark the sub-image as a target sub-image when the sub-image is determined to be an image of the predetermined electronic component, and record the target sub-image on the to-be-detected circuit board image location information.
- the electronic component positioning device further includes:
- An acquisition unit configured to collect a positive sample image and a negative sample image of a predetermined electronic component to be detected
- a training unit for learning the first weak classifier for the N training samples by using the adaboost cascade classifier algorithm, and training the training sample determined to be the positive sample picture by the first weak classifier together with other new training samples Forming a new N training samples, learning a second weak classifier for the N training samples by an adaboost cascade classifier algorithm;
- the training samples include a positive sample picture and a negative sample picture, and N is an integer greater than 1.
- a cascading unit configured to cascade at least two weak classifiers obtained through successive iterations into a detection model of the predetermined electronic component.
- the electronic component positioning device further includes:
- a first label information adding unit configured to add label information to the target sub image
- a first marking unit configured to mark the marking information on the to-be-detected circuit board image according to the location information of the target sub-image.
- the electronic component positioning device further includes:
- a second indicator information adding unit configured to add different label information to the target sub-image detected by the different size detection windows
- a second marking unit configured to mark the marking information on the to-be-detected circuit board image according to the location information of the target sub-image.
- An electronic component positioning method and apparatus detects a captured sub-image based on a generated detection model of a predetermined electronic component to determine whether the sub-image includes the predetermined electronic component and the sub-image. Position information to quickly and accurately locate desired predetermined electronic components on the image of the board to be inspected.
- FIG. 1 is a schematic flow chart of a method for positioning an electronic component according to an embodiment of the present invention.
- FIG. 2 is a schematic diagram of a circuit board to be tested according to an embodiment of the present invention.
- 3(a) to 3(d) are schematic diagrams of positive sample pictures provided by an embodiment of the present invention.
- 4(a) to 4(d) are diagrams normalized to the positive sample picture shown in Figs. 3(a) to 3(d).
- 5(a) to 5(g) are schematic diagrams of negative sample pictures provided by an embodiment of the present invention.
- FIG. 6 is a schematic diagram of detection of a detection model of a predetermined electronic component according to an embodiment of the present invention.
- FIG. 7 is a schematic flow chart of a method for positioning an electronic component according to another embodiment of the present invention.
- FIG. 8 is a schematic diagram of adding flag information to an image of a board to be inspected.
- FIG. 9 is a schematic flow chart of a method for positioning an electronic component according to another embodiment of the present invention.
- FIG. 10 is a schematic structural diagram of an electronic component positioning device according to an embodiment of the present invention.
- FIG. 11 is a schematic structural diagram of an electronic component positioning device according to another embodiment of the present invention.
- FIG. 12 is a schematic structural diagram of an electronic component positioning device according to another embodiment of the present invention.
- FIG. 13 is a schematic structural diagram of an electronic component positioning device according to another embodiment of the present invention.
- Embodiments of the present invention provide an electronic component positioning method and apparatus for positioning predetermined electronic components on a circuit board to be inspected by means of automatic positioning, which are respectively described in detail below.
- FIG. 1 is a schematic flow chart of a method for positioning an electronic component according to an embodiment of the present invention.
- the electronic component positioning method can be performed by an electronic component positioning device and includes at least steps S101 to S103, wherein:
- the sub-image on the image of the circuit board to be detected is intercepted based on the detection window of the preset size.
- various electronic components such as capacitors, resistors, coils, etc.
- the electronic component positioning device In order to locate a predetermined electronic component from the circuit board to be inspected, the electronic component positioning device first acquires an image of the circuit board to be detected (ie, a circuit board image to be detected), and then intercepts the image based on a preset size detection window.
- the step size of the secondary movement can be set according to actual needs, and the invention is not specifically limited.
- the circuit to be detected is relatively large, if the complete image of the circuit board to be detected cannot be obtained at one time, the image may be collected in batches and then spliced.
- the size of the detection window may be set according to the size of the predetermined electronic component, which is not specifically limited in the present invention.
- the sub-image is detected according to a detection model of the established predetermined electronic component; wherein the detection model of the predetermined electronic component is trained by aligning the positive sample image and the negative sample image Obtained, the positive sample picture is a picture including the predetermined electronic component, and the negative sample picture is a picture not including the predetermined electronic component.
- the sub-image may be detected by using a detection model of the established predetermined electronic component to determine whether the sub-image is The predetermined electronic component is included.
- the detection model of the predetermined electronic component can be obtained by training the positive sample image and the negative sample image.
- the sub-image is an image of the predetermined electronic component
- mark the sub-image as a target sub-image and record position information of the target sub-image on the to-be-detected circuit board image.
- the electronic component positioning device inputs the sub-image to the detection model of the predetermined electron for detection, and if the sub-picture passes the detection of the predetermined electronic detection model and finally outputs, Determining that the sub-image is the predetermined electronic component, at this time, the electronic component positioning device marks the sub-image as a target sub-image, and records a position of the target sub-image on the image of the to-be-detected circuit board information.
- the electronic component positioning device when the position information of the target sub image is recorded, can be obtained according to the number of movements in the left and right direction and the number of movements in the up and down direction. For example, assuming that the detection window moves by one pixel at a time, if the target sub-image is a distance in which the detection window is moved by 5 pixels in the up and down direction, and is obtained by moving a distance of 20 pixels in the left and right direction. , the position information of this target sub-image can be recorded as (20, 5).
- the electronic component positioning method generateds a detection model of a predetermined electronic component by training the collected positive sample image and the negative sample image, and then uses the detection model of the predetermined electronic component to intercept the captured model.
- the sub-image is detected to determine whether the sub-image includes a location of the predetermined electronic component and a target sub-image including the predetermined electronic component, thereby positioning the predetermined electronic component on the to-be-detected circuit board.
- the method before step S102, the method further includes:
- S1021 Collect a positive sample image and a negative sample picture of a predetermined electronic component to be detected.
- the electronic component positioning device first collects pictures (positive sample pictures) of different types of electrolytic capacitors, and After normalizing these positive sample pictures, a normalized positive sample picture as shown in Figs. 4(a) to 4(d) is obtained (i.e., the sizes of all positive sample pictures are identical).
- the electronic component positioning device collects the negative sample picture again.
- the negative sample picture does not need to be normalized, but its size must be Not smaller than the size of the normalized positive sample picture.
- the negative sample picture can select those pictures of electronic components or patterns similar in shape to the predetermined electronic component, so that the accuracy of detection can be improved.
- the adaboost cascade classifier algorithm uses the adaboost cascade classifier algorithm to learn the first weak classifier for the N training samples, and the training sample determined by the first weak classifier as the positive sample picture together with other new training samples to form a new one.
- the N training samples are learned by the adaboost cascade classifier algorithm for the N training samples to obtain a second weak classifier; the training samples include a positive sample picture and a negative sample picture, and N is an integer greater than 1.
- the adaboost cascade classifier algorithm is an iterative algorithm, and the core idea is to train different classifiers (weak classifiers) for the same training set, and then combine these weak classifiers to form a Strong classifier.
- the adaboost cascade classifier algorithm itself is implemented by changing the data distribution. It determines the weight of each sample based on whether the classification of each sample in each training set is correct and the accuracy of the last overall classification. .
- the modified new data set is sent to the lower classifier for training, and finally the classifier obtained by each training is cascaded as the final strong classifier, that is, the detection model of the predetermined electronic component.
- the electronic component positioning device sets the sub-image Inputting to a detection model of the predetermined electrons consisting of at least two weak classifier cascades
- the sub-images are sequentially detected by each weak classifier if the sub-picture passes through all weak classifiers Detecting and finally outputting, determining that the sub-image is the predetermined electronic component, at this time, the electronic component positioning device marks the sub-image as a target sub-image, and records the target sub-image in the to-be-detected Location information on the board image.
- the adaboost cascade classifier algorithm is applicable to electronic components with uniform aspect ratio, such as the bottom surface of the electrolytic capacitor is a circle, and thus the outer truncated rectangle is a square, regardless of electrolysis. How big is the capacitance, this aspect ratio is constant. It should be understood that, for other electronic components, if the ratios of the length and the width (outer truncated rectangle) are the same, the technical solutions of the embodiments of the present invention are also applicable.
- the preferred embodiment is based on the adaboost cascade classifier algorithm to train the collected positive sample image and the negative sample image to generate a detection model of the predetermined electronic component, and then use the detection model of the predetermined electronic component to detect the intercepted sub-image, Determining whether the sub-image includes a location of the predetermined electronic component and a target sub-image including the predetermined electronic component, thereby positioning the predetermined electronic component on the to-be-detected circuit board, and the embodiment of the present invention has The positioning is simple, rapid, and the positioning accuracy is high.
- the method further includes:
- S105 Mark the indication information on the to-be-detected circuit board image according to location information of the target sub-image.
- the electronic component positioning device in order to display a predetermined electronic component on the image of the to-be-detected circuit board, the electronic component positioning device further adds marking information to the target sub-image (increasing at the edge of the target sub-image) a bezel), and marking the indication information on the image of the to-be-detected circuit board according to the position of the target sub-image, thereby displaying all the indication information on a complete image of the to-be-detected circuit board (FIG. 8) Shown) for easy viewing and comparison.
- the method further includes:
- the electronic component positioning device adds different marking information to the target sub-image detected by the different size detection windows, for example, adding different color borders to different size target sub-images. Performing a distinction, and then marking the indication information on the to-be-detected circuit board image according to the location information of the target sub-image.
- different types of predetermined electronic components are marked with different indication information, and the operator can quickly distinguish different types of predetermined electronic components according to the indication information, which facilitates further viewing, detection and comparison.
- FIG. 10 is a schematic structural diagram of an electronic component positioning apparatus according to an embodiment of the present invention.
- the electronic component positioning device 100 includes:
- the image intercepting unit 10 is configured to intercept a sub-image on the image of the board to be detected based on the detection window of the preset size.
- the size of the detection window may be set according to the size of the predetermined electronic component, which is not specifically limited in the present invention.
- the detecting unit 20 is configured to detect the sub-image according to a detection model of a predetermined electronic component, wherein the detection model of the predetermined electronic component is obtained by training a positive sample image and a negative sample image, where the positive sample image is A picture containing the predetermined electronic component, the negative sample picture being a picture not including the predetermined electronic component.
- a position recording unit 30 configured to mark the sub-image as a target sub-image when the sub-image is determined to be an image of the predetermined electronic component, and record the target sub-image on the to-be-detected circuit board image Location information.
- the detecting unit 20 may detect the sub-image captured by the image capturing unit 10 according to the generated detection model of the predetermined electronic component to determine whether the sub-image includes Determining, by the position recording unit 30, position information of a target sub-image including the predetermined electronic component by the position recording unit 30, thereby positioning a predetermined electronic component on the to-be-detected circuit board, and implementing the present invention
- the example has the advantages of simple positioning, rapid positioning and high positioning accuracy.
- the electronic component positioning apparatus 100 further includes:
- the collecting unit 40 is configured to collect a positive sample image and a negative sample picture of a predetermined electronic component to be detected.
- the training unit 50 is configured to learn, by using the adaboost cascade classifier algorithm, the first weak classifier for the N training samples, and the training samples determined by the first weak classifier as the positive sample picture and other new training samples. Forming a new N training samples together, learning the N weak classifiers by using the adaboost cascade classifier algorithm; the training samples include a positive sample picture and a negative sample picture, where N is greater than 1. Integer.
- the cascading unit 60 is configured to cascade at least two weak classifiers obtained through successive iterations to form a detection model of the predetermined electronic component.
- the detection model of the predetermined electronic component is obtained by training the positive sample picture and the negative sample picture by an adaboost cascade classifier algorithm.
- the electronic component positioning apparatus 100 further includes:
- the first indication information adding unit 70 is configured to add the indication information to the target sub-image.
- the first marking unit 80 is configured to mark the indication information on the to-be-detected circuit board image according to the location information of the target sub-image.
- the first indication information adding unit 70 further adds marking information to the target sub-image (adding a border at the edge of the target sub-image), and the first marking unit 80 according to the position of the target sub-image
- the indication information is marked on the image of the circuit board to be detected, so that all the indication information (shown in FIG. 8) is displayed on a complete image of the circuit board to be inspected, which facilitates subsequent viewing and comparison.
- the electronic component positioning apparatus 100 further includes:
- the second indication information adding unit 90 is configured to add different indication information to the target sub-images detected by the detection windows of different sizes.
- the second marking unit 91 is configured to mark the indication information on the to-be-detected circuit board image according to the location information of the target sub-image.
- the second indication information adding unit 90 adds different indication information to the target sub-images detected based on the detection windows of different sizes, for example, adding different colors to target sub-images of different sizes.
- the frame is further distinguished, and then the second indicator unit 91 further marks the indication information on the to-be-detected circuit board image according to the location information of the target sub-image.
- different types of predetermined electronic components are marked with different indication information, and the operator can quickly distinguish different types of predetermined electronic components according to the indication information, which facilitates further viewing, detection and comparison.
- the storage medium may be a magnetic disk, an optical disk, or a read-only memory (Read-Only Memory, ROM) or random access memory (RAM).
Landscapes
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Analytical Chemistry (AREA)
- Health & Medical Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
一种电子元件定位方法及装置，可快速准确的在待检测电路板图像上定位出预定电子元件。所述方法包括如下步骤：基于预设尺寸的检测窗截取待检测电路板图像上的子图像（S101）；根据已建立好的预定电子元件的检测模型对所述子图像进行检测；其中，所述预定电子元件的检测模型通过对正样本图片及负样本图片进行训练获得，所述正样本图片为包含有所述预定电子元件的图片，所述负样本图片为不包含有所述预定电子元件的图片（S102）；当确定所述子图像为所述预定电子元件的图像时，将所述子图像标记为目标子图像，并记录所述目标子图像在所述待检测电路板图像上的位置信息（S103）。
Description
本发明涉及自动光学检测领域，尤其涉及一种电子元件的定位方法及装置。
自动光学检测是指利用光学成像的方式取得成品的表面状态，并通过影像处理来检测成品的表面是否存在异物或表面瑕疵。目前，自动光学检测被广泛应用于电路板的质量检测。检测时，相关的检测装置通过摄像头自动扫描电路板获取图像，提取每个电子元件的局部图像，并通过图像处理技术，判断电路板上的电子元件是否存在错插、漏插或反插等缺陷，最后将疑似缺陷的电子元件显示或标记出来，方便查看与检修。
在检测电子元件缺陷之前，需先制作电路板的标准版式，特别地，需要标记电路板上每个电子元件的位置。现有的方案主要有两种：一是根据电路板的设计文件导出电子元件的位置信息；二是采用人工操作的方法在电路板上设置每个电子元件的位置。但是第一个方案需要使用到电路板的设计文件，保密性不好，并且有时候无法获取到设计文件；而第二个方案在电子元件数目较多时，不仅耗时，而且容易出现漏设电子元件的现象，无法满足使用需求。
发明内容
针对上述问题，本发明的目的在于提供一种电子元件定位方法及装置，其可快速、准确的在电路板的图像上定位出电子元件的位置。
本发明提供了一种电子元件定位方法，包括如下步骤：
基于预设尺寸的检测窗截取待检测电路板图像上的子图像；
根据已建立好的预定电子元件的检测模型对所述子图像进行检测；其中，
所述预定电子元件的检测模型通过对正样本图片及负样本图片进行训练获得，所述正样本图片为包含有所述预定电子元件的图片，所述负样本图片为不包含有所述预定电子元件的图片；
当确定所述子图像为所述预定电子元件的图像时，将所述子图像标记为目标子图像，并记录所述目标子图像在所述待检测电路板图像上的位置信息。
作为上述方案的改进，在所述根据已建立好的预定电子元件的检测模型对所述子图像进行检测之前，还包括：
采集待检测的预定电子元件的正样本图像和负样本图片；
利用adaboost级联分类器算法对N个训练样本学习得到第一个弱分类器，将被第一个弱分类器判定为正样本图片的训练样本和其他的新的训练样本一起构成一个新的N个训练样本，通过adaboost级联分类器算法对这N个训练样本学习得到第二个弱分类器；所述训练样本包括正样本图片和负样本图片，N为大于1的整数；
将通过连续迭代获得的至少两个弱分类器级联组成预定电子元件的检测模型。
作为上述方案的改进，所述正样本图片的尺寸为归一化后的尺寸，且任一所述负样本图片的尺寸不小于所述归一化后的尺寸。
作为上述方案的改进，所述预定电子元件为宽高比一致的电子元件。
作为上述方案的改进，在所述当确定所述子图像为所述预定电子元件的图像时，将所述子图像标记为目标子图像，并记录所述目标子图像在所述待检测电路板图像上的位置信息之后，还包括：
对所述目标子图像添加标示信息；
根据所述目标子图像的位置信息将所述标示信息标示在所述待检测电路板图像上。
作为上述方案的改进，在所述当确定所述子图像为所述预定电子元件的图像时，将所述子图像标记为目标子图像，并记录所述目标子图像在所述待检测电路板图像上的位置信息之后，还包括：
对基于不同尺寸的检测窗检测到的目标子图像添加不同的标示信息；
根据所述目标子图像的位置信息将所述标示信息标示在所述待检测电路板图像上。
本发明还提供一种电子元件定位装置，包括：
图像截取单元，用于基于预设尺寸的检测窗截取待检测电路板图像上的子图像；
检测单元，用于根据预定电子元件的检测模型对所述子图像进行检测；其中，所述预定电子元件的检测模型通过对正样本图片及负样本图片进行训练获得，所述正样本图片为包含有所述预定电子元件的图片，所述负样本图片为不包含有所述预定电子元件的图片；
位置记录单元，用于当确定所述子图像为所述预定电子元件的图像时，将所述子图像标记为目标子图像，并记录所述目标子图像在所述待检测电路板图像上的位置信息。
作为上述方案的改进，所述电子元件定位装置还包括：
采集单元，用于采集待检测的预定电子元件的正样本图像和负样本图片；
训练单元，用于利用adaboost级联分类器算法对N个训练样本学习得到第一个弱分类器，将被第一个弱分类器判定为正样本图片的训练样本和其他的新的训练样本一起构成一个新的N个训练样本，通过adaboost级联分类器算法对这N个训练样本学习得到第二个弱分类器；所述训练样本包括正样本图片和负样本图片，N为大于1的整数；
级联单元，用于将通过连续迭代获得的至少两个弱分类器级联组成预定电子元件的检测模型。
作为上述方案的改进，所述电子元件定位装置还包括：
第一标示信息添加单元，用于对所述目标子图像添加标示信息；
第一标示单元，用于根据所述目标子图像的位置信息将所述标示信息标示在所述待检测电路板图像上。
作为上述方案的改进，所述电子元件定位装置还包括：
第二标示信息添加单元，用于对基于不同尺寸的检测窗检测到的目标子图像添加不同的标示信息；
第二标示单元，用于根据所述目标子图像的位置信息将所述标示信息标示在所述待检测电路板图像上。
本发明实施例提供的电子元件定位方法及装置，基于生成的预定电子元件的检测模型对截取的子图像进行检测，以确定所述子图像是否包含有所述预定电子元件及所述子图像的位置信息，从而在所述待检测电路板图像上快速准确的定位出所需的预定电子元件。
为了更清楚地说明本发明的技术方案，下面将对实施方式中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施方式，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1是本发明实施例提供的电子元件定位方法的流程示意图。
图2是本发明实施例提供的待检测电路板的示意图。
图3(a)至图3(d)是本发明实施例提供的正样本图片的示意图。
图4(a)至图4(d)是对图3(a)至图3(d)所示的正样本图片进行归一化后的示意图。
图5(a)至图5(g)是本发明实施例提供的负样本图片的示意图。
图6是本发明实施例提供的预定电子元件的检测模型的检测示意图。
图7是本发明另一实施例提供的电子元件定位方法的流程示意图。
图8是在待检测电路板图像上添加标示信息的示意图。
图9是本发明另一实施例提供的电子元件定位方法的流程示意图。
图10是本发明实施例提供的电子元件定位装置的结构示意图。
图11是本发明另一实施例提供的电子元件定位装置的结构示意图。
图12是本发明另一实施例提供的电子元件定位装置的结构示意图。
图13是本发明另一实施例提供的电子元件定位装置的结构示意图。
下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
本发明实施例提供一种电子元件定位方法及装置，用于通过自动定位的方式定位出待检测电路板上的预定电子元件，以下分别进行详细描述。
请参阅图1，图1是本发明实施例提供的电子元件定位方法的流程示意图。所述电子元件定位方法可由电子元件定位装置来执行，并至少包括步骤S101至S103，其中：
S101，基于预设尺寸的检测窗截取待检测电路板图像上的子图像。
请一并参阅图2，在本发明实施例中，所述待检测电路板上设置有各种电子元件，如电容、电阻、线圈等。为了从所述待检测电路板上定位出预定电子元件，所述电子元件定位装置先获取所述待检测电路板的图像(即待检测电路板图像)，然后基于预设尺寸的检测窗截取待检测电路板图像上的子图像，其中，所述电子元件定位装置在截取完当前区域的子图像后，会按照预定的方向移动所述检测窗，且每移动一次后会截取一次子图像，每次移动的步长可根据实际的需要进行设置，本发明不做具体限定。
在本发明实施例中，对于待检测电路板比较大的情况，若无法一次性获取所述待检测电路板的完整图像，可采用分批采集图像，再对图像进行拼接来获得。
在本发明实施例中，所述检测窗的尺寸可根据所述预定电子元件的大小进行设定，本发明不做具体限定。
S102，根据已建立好的预定电子元件的检测模型对所述子图像进行检测；其中，所述预定电子元件的检测模型通过对正样本图片及负样本图片进行训练
获得，所述正样本图片为包含有所述预定电子元件的图片，所述负样本图片为不包含有所述预定电子元件的图片。
在本发明实施例中，所述电子元件定位装置在截取到所述子图像后，可利用已建立好的预定电子元件的检测模型对所述子图像进行检测，以确定所述子图像中是否包含有所述预定电子元件。
在本发明实施例中，所述预定电子元件的检测模型可通过对正样本图片及负样本图片进行训练获得。
S103，当确定所述子图像为所述预定电子元件的图像时，将所述子图像标记为目标子图像，并记录所述目标子图像在所述待检测电路板图像上的位置信息。
在本发明实施例中，所述电子元件定位装置将所述子图像输入到所述预定电子的检测模型进行检测，若所述子图片通过所述预定电子的检测模型的检测并最终输出，则确定所述子图像为所述预定电子元件，此时，所述电子元件定位装置将所述子图像标记为目标子图像，并记录所述目标子图像在所述待检测电路板图像上的位置信息。
在本发明的一个实施例中，在记录所述目标子图像的位置信息时，所述电子元件定位装置可根据在左右方向上的移动次数和上下方向上的移动次数来获得。例如，假设所述检测窗每次移动一个像素的距离，若所述目标子图像为所述检测窗在上下方向上移动5个像素的距离，在左右方向上移动20个像素的距离后获得的，则可将这个目标子图像的位置信息记录为(20，5)。
综上所述，本发明实施例提供的电子元件定位方法，通过对采集的正样本图片和负样本图像进行训练生成预定电子元件的检测模型，再利用所述预定电子元件的检测模型对截取的子图像进行检测，以确定所述子图像是否包含有所述预定电子元件及包含有所述预定电子元件的目标子图像的位置，从而在所述待检测电路板上定位出所述预定电子元件，本发明实施例具有定位简单、迅速，定位准确率高等优点。
为了进一步说明本发明实施例的技术方案，下面将对本发明的一些优选实施例做进一步说明。
在一个优选实施例中，在步骤S102之前，还包括：
S1021，采集待检测的预定电子元件的正样本图像和负样本图片。
例如，如图3(a)至图3(d)所示，假设所述预定电子元件为电解电容，则所述电子元件定位装置首先收集不同型号的电解电容的图片(正样本图片)，并对这些正样本图片进行归一化后，得到如图4(a)至图4(d)所示的归一化后的正样本图片(即所有正样本图片的尺寸是一致的)。接着，如图5(a)至图5(g)所示，所述电子元件定位装置再收集所述负样本图片，这里，所述负样本图片不要求必须进行归一化，但其尺寸必须不小于归一化后的正样本图片的尺寸。且较佳地，所述负样本图片可选择那些与所述预定电子元件形状类似的电子元件或图案的图片，如此可提高检测的准确率。
S1022，利用adaboost级联分类器算法对N个训练样本学习得到第一个弱分类器，将被第一个弱分类器判定为正样本图片的训练样本和其他的新的训练样本一起构成一个新的N个训练样本，通过adaboost级联分类器算法对这N个训练样本学习得到第二个弱分类器；所述训练样本包括正样本图片和负样本图片，N为大于1的整数。
S1023，将通过连续迭代获得的至少两个弱分类器级联组成预定电子元件的检测模型。
在本发明实施例中，adaboost级联分类器算法是一种迭代算法，其核心思想是针对同一个训练集训练不同的分类器(弱分类器)，然后把这些弱分类器集合起来，构成一个强分类器。adaboost级联分类器算法本身是通过改变数据分布来实现的，它根据每次训练集之中每个样本的分类是否正确，以及上次的总体分类的准确率，来确定每个样本的权值。将修改过权值的新数据集送给下层分类器进行训练，最后将每次训练得到的分类器级联起来，作为最终的强分类器，即所述预定电子元件的检测模型。
如图6所示，在本发明实施例中，当所述电子元件定位装置将所述子图像
输入到由至少两个弱分类器级联组成的所述预定电子的检测模型进行检测，所述子图像会依次经过每个弱分类器的检测，若所述子图片通过所有的弱分类器的检测并最终输出，则确定所述子图像为所述预定电子元件，此时，所述电子元件定位装置将所述子图像标记为目标子图像，并记录所述目标子图像在所述待检测电路板图像上的位置信息。
需要说明的是，在本发明实施例中，adaboost级联分类器算法适用于宽高比一致的电子元件，如所述电解电容的底面是一个圆形，因而其外截矩形是正方形，不论电解电容有多大，这个宽高比都是不变的。应当理解的是，针对其它电子元件，若其的长和宽(外截矩形)的比例一致，也适用于本发明实施例的技术方案。
本优选实施例基于adaboost级联分类器算法对采集的正样本图片和负样本图像进行训练生成预定电子元件的检测模型，再利用所述预定电子元件的检测模型对截取的子图像进行检测，以确定所述子图像是否包含有所述预定电子元件及包含有所述预定电子元件的目标子图像的位置，从而在所述待检测电路板上定位出所述预定电子元件，本发明实施例具有定位简单、迅速，定位准确率高等优点。
请一并参阅图7及图8，在一个优选实施例中，在步骤S103之后，还包括：
S104，对所述目标子图像添加标示信息。
S105，根据所述目标子图像的位置信息将所述标示信息标示在所述待检测电路板图像上。
在本优选实施例中，为了可以在所述待检测电路板图像上显示出预定电子元件，所述电子元件定位装置还对所述目标子图像添加标示信息(在所述目标子图像的边缘增加边框)，并根据所述目标子图像的位置将所述标示信息标示在所述待检测电路板图像上，从而在一张待检测电路板的完整图像上显示出所有的标示信息(如图8所示)，方便后续的查看和比对。
需要说明的是，对于具有不同型号的电子元件，由于其尺寸大小也不同，
因而可能需要采用不同尺寸的检测窗反复检测才能将所述待检测电路板上的所有预定电子元件检测出来，此时只需改变所述检测窗的尺寸即可，这些方案也在本发明的保护范围之内。
请一并参阅图9，在一个优选实施例中，在步骤S103之后，还包括：
S106，对基于不同尺寸的检测窗检测到的目标子图像添加不同的标示信息。
S107，根据所述目标子图像的位置信息将所述标示信息标示在所述待检测电路板图像上。
如上所述，对于具有不同型号的电子元件，由于其尺寸大小也不同，因而可能需要采用不同尺寸的检测窗来检测，在一些实施例中，有时也需要对这些不同型号的电子元件做进一步的区分。
具体地，在本优选实施例中，所述电子元件定位装置对基于不同尺寸的检测窗检测到的目标子图像添加不同的标示信息，例如，对不同尺寸的目标子图像添加不同颜色的边框来进行区分，然后，再根据所述目标子图像的位置信息将所述标示信息标示在所述待检测电路板图像上。如此，不同型号的预定电子元件被以不同的标示信息标示出来，操作人员可以根据这些标示信息快速分辨出不同型号的预定电子元件，方便了进一步的查看、检测和比对。
请参阅图10，图10是本发明实施例提供的电子元件定位装置的结构示意图。所述电子元件定位装置100包括：
图像截取单元10，用于基于预设尺寸的检测窗截取待检测电路板图像上的子图像。
在本发明实施例中，所述检测窗的尺寸可根据所述预定电子元件的大小进行设定，本发明不做具体限定。
检测单元20，用于根据预定电子元件的检测模型对所述子图像进行检测；其中，所述预定电子元件的检测模型通过对正样本图片及负样本图片进行训练获得，所述正样本图片为包含有所述预定电子元件的图片，所述负样本图片为不包含有所述预定电子元件的图片。
位置记录单元30，用于当确定所述子图像为所述预定电子元件的图像时，将所述子图像标记为目标子图像，并记录所述目标子图像在所述待检测电路板图像上的位置信息。
本发明实施例提供的电子元件定位装置100，所述检测单元20可根据生成的预定电子元件的检测模型对所述图像截取单元10截取的子图像进行检测，以确定所述子图像是否包含有所述预定电子元件，并通过所述位置记录单元30对包含有所述预定电子元件的目标子图像进行位置信息的记录，从而在所述待检测电路板上定位出预定电子元件，本发明实施例具有定位简单、迅速，定位准确率高等优点。
请一并参阅图11，在一个优选实施例中，所述电子元件定位装置100还包括：
采集单元40，用于采集待检测的预定电子元件的正样本图像和负样本图片。
训练单元50，用于利用adaboost级联分类器算法对N个训练样本学习得到第一个弱分类器，将被第一个弱分类器判定为正样本图片的训练样本和其他的新的训练样本一起构成一个新的N个训练样本，通过adaboost级联分类器算法对这N个训练样本学习得到第二个弱分类器；所述训练样本包括正样本图片和负样本图片，N为大于1的整数。
级联单元60，用于将通过连续迭代获得的至少两个弱分类器级联形成预定电子元件的检测模型。
本优选实施例中，所述预定电子元件的检测模型是由adaboost级联分类器算法通过对所述正样本图片和负样本图片训练获得的。
请一并参阅图12，在一个优选实施例中，所述电子元件定位装置100还包括：
第一标示信息添加单元70，用于对所述目标子图像添加标示信息。
第一标示单元80，用于根据所述目标子图像的位置信息将所述标示信息标示在所述待检测电路板图像上。
在本优选实施例中，为了可以在所述待检测电路板图像上显示出预定电子
元件，所述第一标示信息添加单元70还对所述目标子图像添加标示信息(在所述目标子图像的边缘增加边框)，所述第一标示单元80根据所述目标子图像的位置将所述标示信息标示在所述待检测电路板图像上，从而在一张待检测电路板的完整图像上显示出所有的标示信息(如图8所示)，方便后续的查看和比对。
请一并参阅图13，在一个优选实施例中，所述电子元件定位装置100还包括：
第二标示信息添加单元90，用于对基于不同尺寸的检测窗检测到的目标子图像添加不同的标示信息。
第二标示单元91，用于根据所述目标子图像的位置信息将所述标示信息标示在所述待检测电路板图像上。
在本优选实施例中，对于具有不同型号的电子元件，由于其尺寸大小也不同，因而可能需要采用不同尺寸的检测窗来检测，有时也需要对这些不同型号的电子元件进行进一步的区分。
具体地，在本优选实施例中，所述第二标示信息添加单元90对基于不同尺寸的检测窗检测到的目标子图像添加不同的标示信息，例如，对不同尺寸的目标子图像添加不同颜色的边框来进行区分，然后，所述第二标示单元91再根据所述目标子图像的位置信息将所述标示信息标示在所述待检测电路板图像上。如此，不同型号的预定电子元件被以不同的标示信息标示出来，操作人员可以根据这些标示信息快速分辨出不同型号的预定电子元件，方便了进一步的查看、检测和比对。
以上所揭露的仅为本发明一种较佳实施例而已，当然不能以此来限定本发明之权利范围，本领域普通技术人员可以理解实现上述实施例的全部或部分流程，并依本发明权利要求所作的等同变化，仍属于发明所涵盖的范围。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过计算机程序来指令相关的硬件来完成，所述的程序可存储于一计算机可读取存储介质中，该程序在执行时，可包括如上述各方法的实施例的流程。其中，所述的存储介质可为磁碟、光盘、只读存储记忆体(Read-Only Memory，
ROM)或随机存储记忆体(Random Access Memory，RAM)等。
Claims (10)
- 一种电子元件定位方法，其特征在于，包括如下步骤：基于预设尺寸的检测窗截取待检测电路板图像上的子图像；根据已建立好的预定电子元件的检测模型对所述子图像进行检测；其中，所述预定电子元件的检测模型通过对正样本图片及负样本图片进行训练获得，所述正样本图片为包含有所述预定电子元件的图片，所述负样本图片为不包含有所述预定电子元件的图片；当确定所述子图像为所述预定电子元件的图像时，将所述子图像标记为目标子图像，并记录所述目标子图像在所述待检测电路板图像上的位置信息。
- 根据权利要求1所述的电子元件定位方法，其特征在于，在所述根据已建立好的预定电子元件的检测模型对所述子图像进行检测之前，还包括：采集待检测的预定电子元件的正样本图像和负样本图片；利用adaboost级联分类器算法对N个训练样本学习得到第一个弱分类器，将被第一个弱分类器判定为正样本图片的训练样本和其他的新的训练样本一起构成一个新的N个训练样本，通过adaboost级联分类器算法对这N个训练样本学习得到第二个弱分类器；所述训练样本包括正样本图片和负样本图片，N为大于1的整数；将通过连续迭代获得的至少两个弱分类器级联形成预定电子元件的检测模型。
- 根据权利要求1所述的电子元件定位方法，其特征在于，所述正样本图片的尺寸为归一化后的尺寸，且任一所述负样本图片的尺寸不小于所述归一化后的尺寸。
- 根据权利要求1所述的电子元件定位方法，其特征在于，所述预定电子元件为宽高比一致的电子元件。
- 根据权利要求1至4任意一项所述的电子元件定位方法，其特征在于，在所述当确定所述子图像上存在所述预定电子元件时，将所述子图像标记为目标子图像，并记录所述目标子图像在所述待检测电路板图像上的位置信息之后，还包括：对所述目标子图像添加标示信息；根据所述目标子图像的位置信息将所述标示信息标示在所述待检测电路板图像上。
- 根据权利要求1至4任意一项所述的电子元件定位方法，其特征在于，在所述当确定所述子图像为所述预定电子元件的图像时，将所述子图像标记为目标子图像，并记录所述目标子图像在所述待检测电路板图像上的位置信息之后，还包括：对基于不同尺寸的检测窗检测到的目标子图像添加不同的标示信息；根据所述目标子图像的位置信息将所述标示信息标示在所述待检测电路板图像上。
- 一种电子元件定位装置，其特征在于，包括：图像截取单元，用于基于预设尺寸的检测窗截取待检测电路板图像上的子图像；检测单元，用于根据预定电子元件的检测模型对所述子图像进行检测；其中，所述预定电子元件的检测模型通过对正样本图片及负样本图片进行训练获得，所述正样本图片为包含有所述预定电子元件的图片，所述负样本图片为不包含有所述预定电子元件的图片；位置记录单元，用于当确定所述子图像为所述预定电子元件的图像时，将所述子图像标记为目标子图像，并记录所述目标子图像在所述待检测电路板图像上的位置信息。
- 根据权利要求7所述的电子元件定位装置，其特征在于，所述电子元件定位装置还包括：采集单元，用于采集待检测的预定电子元件的正样本图像和负样本图片；训练单元，用于利用adaboost级联分类器算法对N个训练样本学习得到第一个弱分类器，将被第一个弱分类器判定为正样本图片的训练样本和其他的新的训练样本一起构成一个新的N个训练样本，通过adaboost级联分类器算法对这N个训练样本学习得到第二个弱分类器；所述训练样本包括正样本图片和负样本图片，N为大于1的整数；级联单元，用于将通过连续迭代获得的至少两个弱分类器级联形成预定电子元件的检测模型。
- 根据权利要求7或8所述的电子元件定位装置，其特征在于，所述电子元件定位装置还包括：第一标示信息添加单元，用于对所述目标子图像添加标示信息；第一标示单元，用于根据所述目标子图像的位置信息将所述标示信息标示在所述待检测电路板图像上。
- 根据权利要求7或8所述的电子元件定位装置，其特征在于，所述电子元件定位装置还包括：第二标示信息添加单元，用于对基于不同尺寸的检测窗检测到的目标子图像添加不同的标示信息；第二标示单元，用于根据所述目标子图像的位置信息将所述标示信息标示在所述待检测电路板图像上。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510890921.4A CN105588840B (zh) | 2015-12-04 | 2015-12-04 | 一种电子元件定位方法及装置 |
| CN201510890921.4 | 2015-12-04 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2017092427A1 true WO2017092427A1 (zh) | 2017-06-08 |
Family
ID=55928583
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2016/096896 WO2017092427A1 (zh) | 2015-12-04 | 2016-08-26 | 一种电子元件定位方法及装置 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN105588840B (zh) |
| WO (1) | WO2017092427A1 (zh) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111861690A (zh) * | 2020-07-23 | 2020-10-30 | 金蝶软件（中国）有限公司 | 账务数据核对方法及账务数据核对装置 |
| CN113591965A (zh) * | 2021-07-26 | 2021-11-02 | 格力电器(南京)有限公司 | 一种aoi检测图像处理方法、装置、存储介质及计算机设备 |
| CN117058241A (zh) * | 2023-10-10 | 2023-11-14 | 轩创(广州)网络科技有限公司 | 一种基于人工智能的电子元件定位方法及系统 |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105588840B (zh) * | 2015-12-04 | 2019-01-25 | 广州视源电子科技股份有限公司 | 一种电子元件定位方法及装置 |
| CN106803249A (zh) * | 2016-12-19 | 2017-06-06 | 广州视源电子科技股份有限公司 | 电子器件在板卡图中的突显方法及系统 |
| CN108428247B (zh) * | 2018-02-27 | 2020-06-12 | 广州视源电子科技股份有限公司 | 焊锡点方向的检测方法和系统 |
| CN108480811A (zh) * | 2018-03-30 | 2018-09-04 | 四川斐讯信息技术有限公司 | 一种焊接设备和焊接方法 |
| CN109977783B (zh) * | 2019-02-28 | 2021-06-04 | 浙江新再灵科技股份有限公司 | 基于直梯场景婴儿车单独乘梯检测的方法 |
| CN111721779B (zh) * | 2020-05-27 | 2023-02-28 | 联宝(合肥)电子科技有限公司 | 一种产品重工方法、装置及存储介质 |
| CN111562488A (zh) * | 2020-06-18 | 2020-08-21 | 汪向丽 | 一种线路板维修检测平台 |
| CN111871848A (zh) * | 2020-06-23 | 2020-11-03 | 安徽信息工程学院 | 一种基于串口通信控制的机器视觉零件分拣方法和系统 |
| TWI730918B (zh) * | 2020-10-29 | 2021-06-11 | 技嘉科技股份有限公司 | 電容器極性狀態判定方法 |
| CN113394141B (zh) * | 2021-08-18 | 2021-10-29 | 湖南省计量检测研究院 | 一种芯片结构缺陷的质量评估系统及方法 |
| CN117351485B (zh) * | 2023-12-04 | 2024-03-08 | 深圳赛陆医疗科技有限公司 | 基因测序训练数据集的增强方法及装置、设备及存储介质 |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060029265A1 (en) * | 2004-08-04 | 2006-02-09 | Samsung Electronics Co., Ltd. | Face detection method based on skin color and pattern match |
| KR100755800B1 (ko) * | 2006-06-21 | 2007-09-07 | 한국과학기술원 | 얼굴색과 에이다부스트를 이용한 얼굴 검출기 및 검출방법 |
| CN101297305A (zh) * | 2004-06-21 | 2008-10-29 | 卡姆特有限公司 | 用于映像缺陷印刷电路的系统 |
| CN102914549A (zh) * | 2012-09-10 | 2013-02-06 | 中国航天科技集团公司第五研究院第五一三研究所 | 针对星载表露型pcb焊点质量的光学图像匹配检测方法 |
| CN103559501A (zh) * | 2013-10-25 | 2014-02-05 | 公安部第三研究所 | 基于图像分析的车辆遮阳板检测方法及装置 |
| CN105588840A (zh) * | 2015-12-04 | 2016-05-18 | 广州视源电子科技股份有限公司 | 一种电子元件定位方法及装置 |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5436028A (en) * | 1992-07-27 | 1995-07-25 | Motorola, Inc. | Method and apparatus for selectively applying solder paste to multiple types of printed circuit boards |
| CN103971118A (zh) * | 2013-01-25 | 2014-08-06 | 北京明日时尚信息技术有限公司 | 静态图片中葡萄酒瓶的检测方法 |
| CN103208008B (zh) * | 2013-03-21 | 2015-11-18 | 北京工业大学 | 基于机器视觉的交通视频监控目标检测快速适应方法 |
| CN103743760A (zh) * | 2013-10-28 | 2014-04-23 | 广州杰赛科技股份有限公司 | 一种pcb板的线路检测方法 |
| CN105069477B (zh) * | 2015-08-13 | 2018-06-19 | 天津津航技术物理研究所 | AdaBoost级联分类器检测图像目标的方法 |
-
2015
- 2015-12-04 CN CN201510890921.4A patent/CN105588840B/zh active Active
-
2016
- 2016-08-26 WO PCT/CN2016/096896 patent/WO2017092427A1/zh active Application Filing
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101297305A (zh) * | 2004-06-21 | 2008-10-29 | 卡姆特有限公司 | 用于映像缺陷印刷电路的系统 |
| US20060029265A1 (en) * | 2004-08-04 | 2006-02-09 | Samsung Electronics Co., Ltd. | Face detection method based on skin color and pattern match |
| KR100755800B1 (ko) * | 2006-06-21 | 2007-09-07 | 한국과학기술원 | 얼굴색과 에이다부스트를 이용한 얼굴 검출기 및 검출방법 |
| CN102914549A (zh) * | 2012-09-10 | 2013-02-06 | 中国航天科技集团公司第五研究院第五一三研究所 | 针对星载表露型pcb焊点质量的光学图像匹配检测方法 |
| CN103559501A (zh) * | 2013-10-25 | 2014-02-05 | 公安部第三研究所 | 基于图像分析的车辆遮阳板检测方法及装置 |
| CN105588840A (zh) * | 2015-12-04 | 2016-05-18 | 广州视源电子科技股份有限公司 | 一种电子元件定位方法及装置 |
Non-Patent Citations (1)
| Title |
|---|
| JI XIU-XIA: "SMT Defection Recognition Algorithm Based on AdaBoost Classification", ELECTRONIC S QUALITY, no. 3, 31-03-2017, pages 1 - 3, ISSN: 1003-0107 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111861690A (zh) * | 2020-07-23 | 2020-10-30 | 金蝶软件（中国）有限公司 | 账务数据核对方法及账务数据核对装置 |
| CN111861690B (zh) * | 2020-07-23 | 2024-04-26 | 金蝶软件（中国）有限公司 | 账务数据核对方法及账务数据核对装置 |
| CN113591965A (zh) * | 2021-07-26 | 2021-11-02 | 格力电器(南京)有限公司 | 一种aoi检测图像处理方法、装置、存储介质及计算机设备 |
| CN117058241A (zh) * | 2023-10-10 | 2023-11-14 | 轩创(广州)网络科技有限公司 | 一种基于人工智能的电子元件定位方法及系统 |
| CN117058241B (zh) * | 2023-10-10 | 2024-03-29 | 轩创(广州)网络科技有限公司 | 一种基于人工智能的电子元件定位方法及系统 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105588840A (zh) | 2016-05-18 |
| CN105588840B (zh) | 2019-01-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2017092427A1 (zh) | 一种电子元件定位方法及装置 | |
| WO2017215241A1 (zh) | 电路板元件漏件检测方法和系统 | |
| US11774735B2 (en) | System and method for performing automated analysis of air samples | |
| CN114140679B (zh) | 一种缺陷融合方法、装置、识别系统及存储介质 | |
| CN101498592B (zh) | 指针式仪表的读数方法及装置 | |
| Mahalingam et al. | Pcb-metal: A pcb image dataset for advanced computer vision machine learning component analysis | |
| CN112053318B (zh) | 基于深度学习的二维pcb缺陷实时自动检测与分类装置 | |
| WO2017181724A1 (zh) | 电子元件漏件检测方法和系统 | |
| CN109918523B (zh) | 一种基于yolo9000算法的电路板元器件检测方法 | |
| WO2018072334A1 (zh) | 一种电子元件定位方法及装置 | |
| TWI635462B (zh) | 一種評估影像品質之方法 | |
| TW201915792A (zh) | 智慧型的缺陷校正系統與其實施方法 | |
| CN111310826B (zh) | 样本集的标注异常检测方法、装置及电子设备 | |
| US11315229B2 (en) | Method for training defect detector | |
| CN106033535A (zh) | 电子阅卷方法 | |
| CN111951210A (zh) | 数据的处理方法、装置及设备 | |
| CN108445010A (zh) | 自动光学检测方法及装置 | |
| JP5995756B2 (ja) | 欠陥検出装置、欠陥検出方法および欠陥検出プログラム | |
| CN113176275B (zh) | 一种用于显示面板复检的方法、装置及系统 | |
| CN116580026B (zh) | 精密零部件外观缺陷自动光学检测方法、设备及存储介质 | |
| JP2014126445A (ja) | 位置合せ装置、欠陥検査装置、位置合せ方法、及び制御プログラム | |
| CN115205288B (zh) | 工业缺陷检测方法和装置 | |
| CN115546219B (zh) | 检测板式生成方法、板卡缺陷检测方法、装置及产品 | |
| WO2017000827A1 (zh) | 一种自动筛选图片的方法及装置 | |
| Zhu et al. | Image quality evaluation method for surface crack detection based on standard test chart |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 16869738 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 16869738 Country of ref document: EP Kind code of ref document: A1 |