TWI751760B - Image capturing method for electronic component and polarity determining method for capacitor - Google Patents

Abstract

Image capturing method for electronic component comprises obtaining an entity baseboard image, transforming a predetermined component coordinate related to a baseboard design layout into an entity component coordinate of the entity baseboard image by using transformation matrix wherein said predetermined component coordinate is a coordinate of the electronic component in the baseboard design layout, framing an image coverage comprising said entity component coordinate in the entity baseboard image, and storing an entity component image corresponding to the image coverage in a memory component ,wherein the entity component image is a part of the image coverage framed in the entity baseboard image.

Description

Electronic component imaging method and capacitor polarity determination method using this method

The present invention relates to an electronic component imaging method and a capacitor polarity determination method using the method, in particular to an electronic component imaging method for automatically acquiring an electronic component image according to a design drawing file.

When manufacturing circuit boards such as computer motherboards, development boards, etc., various electronic components are combined with the substrate, and whether the electronic components are placed manually or automatically, there are positive and negative polarities of the placed electronic components, and the wrong direction of the pins. possibility. The conventional methods for checking whether there are errors are, for example, manual inspection, connecting some electronic components or some areas of the circuit board with a voltage or current detector, and measuring whether the measured value conforms to the factory setting value, and the like. However, the above-mentioned inspection method is too time-consuming, which tends to reduce the efficiency of the entire circuit board manufacturing process. At present, image recognition technology can also be used to assist the inspection. After obtaining an image of an electronic component including a circuit board, it can be determined whether the electronic component has misplaced pins according to the processed image.

However, due to the large number of electronic components on the circuit board, if it is necessary to manually frame the images of the electronic components, it will obviously cause great labor and time costs to the overall manufacturing process. In addition, due to various environmental factors that cannot be precisely controlled, such as the direction of illumination, the lens angle of the imaging device, and the shading of adjacent components, the image containing electronic components is too distorted or cannot be judged after preprocessing. Identify the position of the electronic component to be tested on the circuit board, etc., so that the method of image recognition occasionally occurs underkill or misjudgment (overkill). If it is detected, misjudgment means that the pin position of the electronic component is correct but it is judged to be reversed.

In view of the above, the present invention provides an electronic component imaging method to meet the above requirements and a capacitor polarity determination method using the method. The imaging method automatically locates the electronic component, assists the establishment of a model for determining the polarity, and utilizes this model. Determine the polarity of capacitors, and try to avoid errors or misjudgments of capacitor polarity caused by environmental factors that cannot be precisely controlled.

The electronic component imaging method according to the first embodiment of the present invention is suitable for acquiring an image of an electronic component on a circuit board, including: acquiring a physical circuit board image associated with the circuit board; A preset component coordinate of a circuit board design file of the circuit board is converted into a physical component coordinate of the physical circuit board image, wherein the preset component coordinate is the coordinate of the electronic component in the circuit board design file ; frame an imaging range including the coordinates of the physical component in the physical circuit board image; and store a physical component image corresponding to the imaging range in a memory element, wherein the physical component image is the imaging The extent is the portion framed by the image of the physical board.

The capacitor polarity determination method according to the second embodiment of the present invention includes: executing a modeling program to obtain an online judgment model capable of judging the polarity state of a capacitor; Outputting a polarity state of a capacitor to be tested, wherein the modeling program includes: obtaining a plurality of physical capacitor images from a circuit board or a plurality of circuit boards using the aforementioned electronic component imaging method as a training image group, wherein the physical capacitor images are Each of the capacitor images is an image of the physical device in the electronic device imaging method; and a machine learning model is trained with the training image group to obtain the online judgment model; wherein the judgment procedure includes: using the aforementioned electronic components The imaging method acquires a pending capacitor image from another circuit board, wherein the pending capacitor image is also the physical element image in the electronic component imaging method; and the online judgment model is used to determine and output the pending capacitor image The polarity state of the capacitor under test in , wherein the capacitor under test is the electronic component in the electronic component imaging method.

The above description of the present disclosure and the following description of the embodiments are used to demonstrate and explain the spirit and principle of the present invention, and provide further explanation of the scope of the patent application of the present invention.

The detailed features and advantages of the present invention are described in detail below in the embodiments, and the content is sufficient to enable any person skilled in the relevant art to understand the technical content of the present invention and implement it accordingly, and according to the content disclosed in this specification, the scope of the patent application and the drawings , any person skilled in the related art can easily understand the related objects and advantages of the present invention. The following examples further illustrate the viewpoints of the present invention in detail, but do not limit the scope of the present invention in any viewpoint.

The first embodiment of the present invention provides an electronic component imaging method, which can obtain an electronic component image from a circuit board image according to the coordinates of the electronic component, and the method can be implemented in any computer system that needs to perform electronic component imaging. The capacitor polarity determination method provided by the second embodiment of the present invention can utilize the electronic component imaging method to assist in constructing a model for recognizing capacitor polarity and perform real-time determination on production line quality control. Therefore, the method is preferably implemented in a quality control computer system. However, the present invention is not limited to this. The two embodiments will be described in detail below with reference to the accompanying drawings.

Please refer to FIG. 1 , which is a flowchart of a method for capturing an image of an electronic component according to the first embodiment of the present invention. The electronic component imaging method according to the first embodiment of the present invention may include the following steps: step A1, obtaining a physical circuit board image associated with the circuit board; step A2, using a transformation matrix to convert a circuit board associated with the circuit board A preset component coordinate of the design file is converted into a physical component coordinate of the physical circuit board image, wherein the preset component coordinate is the coordinate of the electronic component in the circuit board design file; Step A3, in the physical An imaging range including the coordinates of the physical element is framed in the circuit board image; and in step A4, a physical element image corresponding to the imaging range is stored in a memory element, wherein the physical element image is the imaging range The portion framed in the image of the physical circuit board. The following are detailed explanations and examples for the above steps.

In practice, step A1 can, for example, use an imaging device to take a picture of the circuit board, record it, etc. to obtain a physical circuit board image, or first obtain an initial circuit board image in the above manner, and then perform the initial circuit board image. Image preprocessing, such as binarization, erosion, and/or dilation, is used to obtain physical circuit board images. The circuit board needs to contain at least one electronic component, that is, the image of the physical circuit board obtained by the imaging device at least contains the electronic component. For the convenience of subsequent descriptions, the present application takes capacitors as an example of the above electronic components for the following description. However, the present invention does not limit the electronic components to be capacitors, and this electronic component imaging method can also be applied to other types of electronic components. Please refer to FIG. 2 , which illustrates a capacitor 10 suitable for the electronic device imaging method of the present application. The upper surface of the capacitor 10 has a solid pattern 11 , and the solid pattern 11 is used to represent the polarity direction of the capacitor 10 . Usually, the solid pattern 11 is in the shape of an arc, and the orientation of the arc of the arc relative to its chord indicates the negative electrode of the capacitor 10 Orientation relative to the positive pole. That is, if the arc of the arcuate solid pattern 11 is located on the right side of its chord, the negative electrode of the capacitor 10 is located on the right side of the positive electrode of the capacitor 10 . Please refer to FIG. 3 , which is a schematic top view of a circuit board with capacitors in general. The circuit board MB has capacitors arranged in different positions and with different sizes, and the physical patterns on the upper surface of each capacitor can be in different orientations. .

Step A2 is to convert the preset component coordinates of the circuit board design file into the physical component coordinates of the physical circuit board image by using the conversion matrix. Specifically, the circuit board design drawing file is the aforementioned circuit board design drawing file, especially an electronic design drawing file. The circuit board design file represents a plurality of electronic components of the circuit board, such as various capacitors, resistors, microchips, processing chips, etc., and the circuit board design file preferably includes a plurality of preset component coordinates, wherein Each preset component coordinate may correspond to an electronic component. The circuit board design file and the physical circuit board image can be regarded as being in different coordinate systems. The conversion matrix between the physical circuit board image and the circuit board design file is obtained, and the aforementioned predetermined component coordinates are converted into the physical component coordinates on the physical circuit board image by the conversion matrix.

Step A3 is to frame the imaging range including the coordinates of the physical element in the image of the physical circuit board. The coordinates of the physical component can be set at the center of the electronic component on the circuit board, and in this case, the coordinates of the physical component can be used as the center, and a predetermined length can be expanded outward to frame a circular imaging range. , wherein the imaging range includes the electronic component. However, in addition to being a circle, the imaging range can also be a regular polygon such as a regular quadrilateral. To be more specific, taking the coordinates of the physical element as the center, the imaging range can be obtained by expanding to a first predetermined length and a second predetermined length along the two axes of the physical circuit board image, respectively. It should be noted that the above manner of obtaining the physical capacitor image is only an illustration of an embodiment, and the present invention is not limited thereto.

Then, step A4 stores the framed portion of the image range in the physical circuit board image in the memory element, wherein the framed portion can be defined as the physical element image, and the memory element can be the image of the electronic element. A memory element of an electronic device, such as random access memory or read-only memory.

However, if the related procedures of image processing have not been performed in step A1, image preprocessing, such as binarization, erosion and/or dilation, may also be performed before storing the framed portion in the memory device. (dilation) and other morphological image processing, and then use the content that has undergone the above image preprocessing as the aforementioned entity element image. In this way, unnecessary noise can be removed or the image can be enhanced, so as to improve the image quality and facilitate subsequent judgment.

With the above-mentioned method for capturing images of electronic components according to the first embodiment of the present invention, especially using a conversion matrix to convert the preset coordinates of electronic components into coordinates of physical components and then capturing images, it is possible to effectively generate images that are not caused by various environmental factors. In the case of expected changes, for example, when there are changes in the direction of illumination, the angle of the lens of the imaging device, and the shading of adjacent components, the image of the electronic component can still be accurately acquired, effectively maintaining the accuracy of the imaging.

Please refer to FIG. 4A , which is a flowchart of a capacitor polarity determination method according to a second embodiment of the present invention. The capacitor polarity determination method of the second embodiment of the present invention may include the following steps: step S1, executing a modeling program to obtain an on-line determination model capable of determining the polarity state of a capacitor; and step S2, executing a determination program to Judging and outputting the polarity state of a capacitor under test through the online judgment model. Wherein, the polarity state of the capacitor to be tested can be "the negative electrode of the capacitor is located on the left side of the positive electrode of the capacitor", "the negative electrode of the capacitor is located on the right side of the positive electrode of the capacitor", or other relative orientations.

Please refer to 4B, step S1 may include sub-step S11 and sub-step S12. Sub-step S11 is to obtain a plurality of physical capacitor images from a circuit board or a plurality of circuit boards by using the electronic component imaging method of the first embodiment, and use these physical capacitor images as a training image group, and these physical capacitor images Each of them is a solid component image in the electronic component imaging method of the first embodiment. In detail, since step S1 is a modeling procedure of a machine learning model for judging the polarity state of capacitors, it is preferable to obtain a plurality of physical capacitor images according to steps A1 to A4 of the first embodiment of the present invention, so as to train the machine learning model. Basic information of the model. In practice, as shown in FIG. 3 , since there may be more than one capacitor on a circuit board, when a circuit board including a plurality of capacitors is imaged by the imaging method according to the first embodiment of the present invention, a plurality of physical capacitors can be obtained. image. Alternatively, a plurality of circuit boards can be imaged to obtain more physical capacitor images, so as to improve the recognition accuracy of the trained machine learning model.

Sub-step S12 is to train a machine learning model with the training image group to obtain an online judgment model. The present invention does not limit the machine learning model that uses the training image group for training to obtain the online judgment model in step S12, but preferably uses a convolutional neural network algorithm to build the model. The online judgment model obtained in step S12 may be a static model that no longer performs subsequent training based on new data, but is preferably a dynamic model that can be updated and trained as needed. Through the above training, the online judgment model can be used to judge the polarity state of the capacitors according to the obtained images.

Please refer to FIG. 4B again, step S2 includes sub-step S21 and sub-step S22. Sub-step S21 is to obtain a pending capacitor image from another circuit board using the electronic component imaging method of the first embodiment, wherein the pending capacitor image is also a physical component in the electronic component imaging method of the first embodiment image, and the so-called other circuit board may be the circuit board under test on the production line. In other words, in this embodiment, the electronic component imaging method of the aforementioned first embodiment is not only suitable for obtaining the physical capacitor image required by the modeling program for training the online judgment model, but also for obtaining the judgment program executed on the production line. Desired pending capacitor image. Sub-step S22 is to input the pending capacitor image to the online judgment model, so as to judge and output the polarity state of the capacitor under test in the pending capacitor image. In this way, in the process of producing circuit boards, there is no need to manually select each capacitor (or other electronic components to be tested) on the circuit board to be tested repeatedly, which can effectively implement the capacitor installation orientation for all products in the production line. quality control.

In practical situations, in addition to the implementation of determining the polarity of the capacitor in steps S1 and S2, the second embodiment of the present invention can further refer to the data of the preset state that the capacitor to be tested should have, according to the determination in step S2. The discrepancy between the obtained polarity state and the preset state can timely adjust the online judgment model. The following will continue to describe this embodiment.

Referring to FIG. 5 , the capacitor polarity determination method according to the second embodiment of the present invention may further include steps S3 to S6 , and the aforementioned determination procedure is defined as the first determination procedure, and the aforementioned pending capacitor image is defined as the first pending Capacitive image. In step S3, the polarity state of the capacitor under test is compared with a preset state in a circuit board design file to determine whether the polarity state conforms to the preset state. When the polarity state is consistent with the preset state, it can be considered that the polarity state obtained in step S2 is correct, that is, the capacitor to be tested is installed in a correct manner. However, when the polarity state does not match the preset state, a warning signal for warning is generated. At this time, the judgment procedure of step S2 may be wrong, or the installation method of the capacitor to be tested is wrong. At this time, the present embodiment continues to execute step S4, but in production practice, the circuit board to be tested including the capacitor to be tested can also be directly checked manually for confirmation.

In step S4, when the polarity state does not match the preset state, a second determination procedure is further executed to obtain a second pending capacitor image associated with the aforementioned capacitor under test from the circuit board under test, and determine and output the second pending capacitor image Determines the polarity state of the capacitor under test in the capacitor image. To be more specific, step S4 is to perform the entire judgment procedure again on the capacitor to be tested when the polarity state is not as expected to obtain the second capacitor image to be tested and the polarity state obtained from the image. Further, in step S5, when the polarity states output by the first and second judgment programs are different, record the preset component coordinates corresponding to the capacitor under test in the circuit board design file, and accumulate a total number of misjudgments, and The first pending capacitor image is stored in a temporary storage element. In other words, when the polarity states obtained by the two judgment procedures are the same (that is, the two polarity states are different from the preset state), it can be considered that the installation method of the capacitor to be tested is very likely to be wrong, and the defective products can be corrected later. Correct this installation error; and if the polarity states obtained by the two judgment procedures are different (that is, the same polarity status as the default status is obtained after the judgment procedure is performed again), the first pending capacitor image is regarded as prone to misjudgment. , so add one to the total number of false positives, and store this first pending capacitor image. For the total number of misjudgments, preferably, each preset component coordinate of the circuit board design file has a total number of misjudgments, and the total number of misjudgments means that the capacitor under test corresponding to its preset component coordinates has been misjudged. frequency. In this setting, when the total number of misjudgments of the coordinates of a preset component is high, the factors that cause this situation are more likely to be insufficient light source at this position, a skewed shooting angle, and the physical pattern of the capacitor is blocked by surrounding components. due to differences in individual locations. However, it is also possible that a single circuit board design file has only a single total number of false positives, and the total number of false positives may represent factors other than the influence of individual locations, and may also represent a poor overall imaging environment. In addition, the total number of errors can be presented not only in the form of judging the number of error occurrences, but also in the form of dividing the number of occurrences by the total number of judging times of error occurrence rate, which is not limited in the present invention.

In step S6, when the total number of misjudgments exceeds a predetermined value, the training image group is updated with the first pending capacitor image obtained by the first judgment procedure, and the online judgment model is trained with the updated training image group. And the total number of misjudgments mentioned above can be reset to zero. Corresponding to the presentation manner of the total number of misjudgments, the preset value may be an integer value corresponding to the number of error occurrences, or may be a percentage value or a decimal value corresponding to the error occurrence rate. The first pending capacitor image for updating the training image group may be the pending capacitor image obtained when the total number of misjudgments just exceeds the preset value, but preferably a plurality of first pending capacitor images stored in the temporary storage component are used. Capacitor images update the training image ensemble, thereby enhancing the importance of pending capacitor images in the training image ensemble that caused false positives.

In another embodiment, when each preset component coordinate of the circuit board design file has a separate total number of false positives, a dedicated training image can also be set for the coordinates of the preset component whose total false positives exceed the preset value group (that is, the training image group updated with the first pending capacitor image corresponding to this preset element coordinate only affects the judgment of the polarity state of the under-test capacitor corresponding to this preset element coordinate), and after this update Another online judgment model generated by retraining the training image group is a dedicated model for judging the capacitor under test of the preset component coordinates. By independently creating an exclusive model and an exclusive training image group that are different from the original online judgment model and the original image training group, it is not only possible to make exclusive judgments for the capacitors on the coordinates of the preset components that are prone to misjudgment, but also to make exclusive judgments in relation to other presets. When the total number of misjudgments of another capacitor on the component coordinates also exceeds the preset value, another dedicated model can also be generated based on another dedicated training image group, and the original online judgment model continues to make misjudgments that do not exceed the preset value. The total number of capacitors is judged. Under the permission of computing performance, these models can be applied at the same time to make adaptive judgments for capacitors in different conditions, so as to better overcome various environments such as capacitors in the same circuit board design file due to different positions, illumination, and shading of adjacent components. misjudgment caused by factors.

With the above-mentioned capacitor polarity determination method according to the second embodiment of the present invention, it is not only unnecessary to manually select each capacitor on the circuit board to be tested repeatedly in the process of producing the circuit board, but also the performance of the online judgment model can be selectively judged. When it is poor, update the training image group to improve the judgment accuracy of the online judgment model, and achieve the effect of providing customized judgment effects for different types of circuit boards under test.

Although the present invention is disclosed in the foregoing embodiments, it is not intended to limit the present invention. Changes and modifications made without departing from the spirit and scope of the present invention belong to the scope of patent protection of the present invention. For the protection scope defined by the present invention, please refer to the attached patent application scope.

10: Capacitors 11: Solid Pattern MB: circuit board

FIG. 1 is a flowchart of a method for capturing an image of an electronic component according to the first embodiment of the present invention. FIG. 2 is a perspective view of a capacitor suitable for the imaging method of electronic components according to the first embodiment of the present invention. FIG. 3 is a schematic top view of a circuit board of the electronic component imaging method according to the first embodiment of the present invention. 4A is a flowchart of an embodiment of a capacitor polarity determination method according to the second embodiment of the present invention. FIG. 4B is a detailed flowchart of an embodiment of the capacitor polarity determination method according to the second embodiment of the present invention. FIG. 5 is a partial flowchart of another embodiment of the capacitor polarity determination method according to the second embodiment of the present invention.

A1~A4: Steps

Claims (12)

An electronic component imaging method, suitable for acquiring an image of an electronic component on a circuit board, includes: obtaining a physical circuit board image associated with the circuit board; Converting a predetermined component coordinate of a circuit board design file associated with the circuit board to a physical component coordinate of the physical circuit board image by a conversion matrix, wherein the predetermined component coordinate is the electronic component on the circuit board. The coordinates in the design file; framing an imaging range including the coordinates of the physical element in the image of the physical circuit board; and A physical device image corresponding to the imaging range is stored in a memory element, wherein the physical device image is the part of the imaging range framed by the physical circuit board image. The electronic component imaging method of claim 1, wherein framing the imaging range including the physical component coordinates in the physical circuit board image includes: taking the physical component coordinates as the center, along two axes of the physical circuit board image It is expanded to a first predetermined length and a second predetermined length respectively, so as to obtain the quadrilateral image capturing range. The electronic component imaging method of claim 1, wherein framing the imaging range including the physical component coordinates in the physical circuit board image includes: taking the physical component coordinates as the center, expanding outward by a predetermined length, to The acquisition range in the form of a circle is obtained. The electronic component imaging method of claim 1, wherein acquiring the physical circuit board image associated with the circuit board comprises: photographing the circuit board with an imaging device to obtain an initial circuit board image; and At least one of binarization, erosion, and dilation is performed on the initial circuit board image to form the physical circuit board image. A capacitor polarity determination method, comprising: performing a modeling procedure to obtain an on-line judgment model capable of judging the polarity state of a capacitor; and Execute a judgment program to judge and output the polarity state of a capacitor to be tested through the online judgment model, Where the modeling program includes: Obtain a plurality of physical capacitor images from a circuit board or a plurality of circuit boards as a training image group using the electronic component imaging method of the aforementioned claim 1, wherein each of the physical capacitor images is an image of the electronic component. the physical element image in the method; and training a machine learning model with the training image group to obtain the online judgment model; The judgment procedure includes: Obtaining a pending capacitor image from another circuit board using the electronic component imaging method of the aforementioned claim 1, wherein the pending capacitor image is also the physical component image in the electronic component imaging method; and The on-line judgment model is used to determine and output the polarity state of the capacitor to be tested in the capacitor image to be tested, wherein the capacitor to be tested is the electronic component in the electronic component imaging method. The capacitor polarity determination method of claim 5, wherein the capacitor polarity determination method further comprises: comparing the polarity state of the capacitor under test with a preset state in a circuit board design file, and when the polarity state does not conform to the preset state A reminder signal is generated when the state is set. The capacitor polarity determination method of claim 5, wherein the determination process is defined as a first determination process, and the pending image is defined as a first pending image, the capacitor polarity determination method further comprises: comparing the capacitor under test polarity state and a preset state in a circuit board design file, and when the polarity state does not match the preset state, a second judgment procedure is executed, and the second judgment procedure is obtained by the other circuit board. In a second pending capacitor image of the capacitor under test, the polarity state of the capacitor under test in the second pending capacitor image is determined and output. The capacitor polarity determination method of claim 7, wherein when the polarity states output by the first and second determination programs are different, the preset component coordinates in the circuit board design file corresponding to the capacitor to be tested are recorded, And add up the total number of false positives. The capacitor polarity determination method of claim 8, wherein when the total number of misjudgments exceeds a predetermined value, the training image group is updated with the first pending capacitor image obtained by the first determination procedure, and the training image group is updated with the updated training The image group trains the online judgment model. The capacitor polarity determination method of claim 8, wherein when the polarity states output by the first and second determination programs are different, the first pending capacitor image is further stored in a temporary storage element. The capacitor polarity determination method of claim 10, wherein when the total number of false positives exceeds a predetermined value, the training image group is updated with a plurality of first pending capacitor images stored in the temporary storage component, and the training image group is updated with the updated training The image group trains the online judgment model. The capacitor polarity determination method of claim 10, wherein the total number of false positives corresponds to the predetermined component coordinates, and when the total number of false positives exceeds a predetermined value, an exclusive model associated with the predetermined component coordinates is established, and the temporary The training image group is updated with at least one first pending capacitor image associated with the predetermined element coordinates in the storage component, and the dedicated model is trained with the updated training image group.

Images