WO2017088553A1

WO2017088553A1 - Method and system for rapidly identifying and marking electronic component polarity direction

Info

Publication number: WO2017088553A1
Application number: PCT/CN2016/098228
Authority: WO
Inventors: 杨铭
Original assignee: 广州视源电子科技股份有限公司
Priority date: 2015-11-23
Filing date: 2016-09-06
Publication date: 2017-06-01
Also published as: CN105469400A; CN105469400B

Abstract

A method and system for rapidly identifying and marking an electronic component polarity direction. The method comprises: acquiring an image containing a target electronic component (S101); performing forward calculation on image data of the target electronic component by using a trained convolutional neural network, acquiring a classification characteristic of a polarity direction category of the target electronic component, and acquiring, according to the classification characteristic, a probability distribution of polarity direction categories of the target electronic component belonging to the all kinds of electronic components (S102); and selecting a polarity direction category having a maximum probability as the polarity direction category of the target electronic component (S103). In the method, a convolutional neural network is used, and accordingly polarity directions of electronic components can be identified automatically and precisely; the method is not aimed at a specified electronic component structure but is suitable for all kinds of electronic components having polarity, and accordingly cross-category polarity direction identification of electronic components is implemented, and the method has broad applicability.

Description

Method and system for quickly identifying and marking the polarity direction of electronic components

Technical field

The invention relates to the field of automatic optical detection, and in particular to a method and a system for quickly identifying and marking the polarity direction of an electronic component.

Background technique

At present, for the PCB board, it is necessary to judge the polarity direction of the electronic components thereon. At present, there are two main methods for judging the polarity direction of electronic components: one is based on the polarity direction judgment of component structural features. This method is based on specific component structure features, and is usually only designed for certain types of components, and the applicability is limited. In addition, some structural features are unstable under different environments (such as different illumination, different shooting angles, noise interference, etc.), resulting in lower accuracy of judgment; second, based on the polarity direction of positive and negative template matching, as long as The components are all set to a template, but this may result in a lower accuracy due to the mixing of multiple categories of templates, and the components have to introduce new templates as long as the appearance is slightly different, the scalability is limited, and the calculation time is also due to the template library. Increase and increase.

Summary of the invention

Based on this, it is necessary to provide a method and system for quickly identifying and marking the polarity direction of electronic components in order to solve the problem of low accuracy and limited applicability for the polarity direction of electronic components.

A method for quickly identifying the polarity direction of an electronic component includes the following steps:

Obtain an image containing the target electronic component;

Using the trained convolutional neural network, the image data of the target electronic component is forwardly calculated, and the classification features of the polarity direction category of the target electronic component are obtained, and the target electronic components are obtained according to the classification features. Probability distribution of sexual direction categories;

The polarity direction category with the highest probability is selected as the polarity direction category of the target electronic component.

According to the above fast identification method, the image is obtained by including the target electronic component, and then subjected to the forward calculation by using the trained convolutional neural network to obtain the classification feature of the polarity direction category of the target electronic component, and then the target electronic component is obtained. Probability distribution of various polar direction categories belonging to various electronic components, The polarity direction category with the highest probability is selected as the polarity direction category of the target electronic component. In this scheme, a convolutional neural network is used. The convolutional neural network can automatically and accurately identify the polarity direction of electronic components, and it is not specific to the structure of electronic components. It is applicable to various electronic components with polarity to achieve cross- The polarity direction identification of the electronic components of the category is widely applicable.

In one embodiment, the step of calculating the image data of the target electronic component by using the trained convolutional neural network to perform the forward calculation of the image data of the target electronic component includes the following steps:

The image data is convoluted by the convolution layer, then nonlinearly transformed by the activation function layer, and then pooled by the pooling layer, and then the classification feature of the polarity direction category of the target electronic component is obtained through the fully connected layer. The convolutional neural network includes a convolution module and a fully connected layer, and the convolution module includes a convolution layer, an activation function layer, and a pooling layer that are sequentially connected.

In one of the embodiments, the trained convolutional neural network is obtained by the following steps:

Establishing image sample sets of various polarity directions of various electronic components;

Using the convolution module and the fully connected layer in the convolutional neural network to perform forward calculation on each image sample data of the image sample set, and obtain classification features of various polarity direction categories of various electronic components, according to each classification The feature training convolutional neural network enables the convolutional neural network to identify various polar directions of various electronic components.

In one of the embodiments, the step of establishing a sample set of various polarity directions of various types of electronic components includes the following steps:

Obtaining the PCB board image and the PCB template map, and using the PCB template map as a reference to position the PCB board image;

Intercepting the image of various electronic components on the PCB board image after registration, and matching the corresponding electronic components in the image of various electronic components with various electronic components in the PCB template image, and obtaining corresponding electrons in the image of various electronic components. The precise position of the components, the corresponding electronic components are adjusted according to the precise position of the corresponding electronic components, so that the corresponding electronic components are located at the center of the image of various electronic components, and image sample sets of various polar directions of various electronic components are obtained.

In one embodiment, there are five convolutional modules, wherein the number of convolution kernels of the five convolutional layers is 24, 64, 96, 96, and 64, respectively, and the convolution kernel sizes of the five convolutional layers are respectively 7×7, 5×5, 3 ×3, 3×3, 3×3, 5 convolutional layers have a step size of 1; the total connection layer is 2, and the number of hidden nodes of the 2 fully connected layers is 512 and 4, respectively.

In one of the embodiments, after the step of acquiring the image containing the target electronic component, the following steps are included:

Obtaining a template image of the target electronic component, matching the target electronic component in the image with a template image of the target electronic component, obtaining an accurate position of the target electronic component in the image, and adjusting the target electronic component in the image according to the precise position, so that The target electronic component in the image is located at the center of the image, and the adjusted image is used for forward calculation by the trained convolutional neural network.

A method for marking the polarity direction of an electronic component includes the following steps:

According to the polarity direction of the target electronic component determined by the fast identification method of the polarity direction of the electronic component, the polarity direction information of the target electronic component is marked in the panel file, and the panel file is used to store various attribute information of the electronic component.

The above-mentioned method for labeling the polarity direction of the electronic component can be applied to the AOI board production, and the polarity direction of the components on the PCB board is automatically and accurately marked, thereby improving the automation level of the panel production and improving the efficiency and accuracy of the panel production.

A rapid identification system for the polarity direction of electronic components, comprising the following units:

Acquiring a unit to obtain an image containing the target electronic component;

The calculating unit is configured to perform forward calculation on the image data including the target electronic component by using the trained convolutional neural network, obtain a classification feature of the polarity direction category of the target electronic component, and acquire the target electronic component according to the classification feature to belong to various types of electronic Probability distribution of various polar direction categories of components;

The selection unit is configured to select the polarity direction category with the highest probability as the polarity direction category of the target electronic component.

According to the above fast recognition system, the image is obtained by including the target electronic component, and then subjected to the forward calculation by using the trained convolutional neural network to obtain the classification feature of the polarity direction category of the target electronic component, and then the target electronic component is obtained. The probability distribution of various polarity direction categories belonging to various electronic components is selected as the polarity direction category of the target electronic component. In this scheme, a convolutional neural network is used, which can automatically and accurately identify the polarity direction of electronic components through a convolutional neural network, and is not specific to a specific electronic component structure, and is applicable to various electronic components with polarity. It realizes the polarity direction identification of electronic components across categories and has wide applicability.

In one embodiment, the computing unit performs convolution operation on the image data through the convolution layer, then performs nonlinear transformation through the activation function layer, performs pooling operation through the pooling layer, and then obtains the target electronic component through the fully connected layer. The classification feature of the polarity direction category, wherein the convolutional neural network comprises a convolution module and a fully connected layer, and the convolution module comprises a convolution layer, an activation function layer and a pooling layer connected in sequence.

An annotation system for the polarity direction of an electronic component, comprising a labeling unit and a quick identification system for the polarity direction of the electronic component, wherein the labeling unit is used to mark the polarity direction information of the target electronic component in the panel file, and the panel file is used for Save various attribute information of electronic components.

The above-mentioned electronic component polarity direction labeling system can be applied to AOI board production, and the polarity direction of the components on the PCB board is automatically and accurately marked, thereby improving the automation level of the panel production and improving the efficiency and accuracy of the panel production.

DRAWINGS

1 is a schematic flow chart of a method for quickly identifying a polarity direction of an electronic component in one embodiment;

2 is a schematic structural view of a convolution module in one embodiment;

3 is a schematic structural view of a convolutional neural network in one embodiment;

4 is a schematic structural view of a rapid identification system for the polarity direction of an electronic component in one embodiment;

5 is a schematic structural view of a rapid identification system for the polarity direction of an electronic component in one embodiment;

6 is a partial structural schematic view of a rapid identification system for the polarity direction of an electronic component in one embodiment;

7 is a schematic structural view of a rapid identification system for the polarity direction of an electronic component in one embodiment;

8 is a schematic structural view of a rapid identification system for the polarity direction of an electronic component in one embodiment;

Figure 9 is a block diagram showing the structure of the polarity of the electronic component in one embodiment.

detailed description

The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein The invention is only intended to be illustrative, and does not limit the scope of the invention.

Referring to FIG. 1, an embodiment of a method for quickly identifying a polarity direction of an electronic component of the present invention is shown. As shown in FIG. 1, the method for quickly identifying the polarity direction of an electronic component in this embodiment includes the following steps:

Step S101: Acquire an image including the target electronic component;

In this step, the image containing the electronic component may be a PCB card image that needs to identify the polarity of the electronic component, or other images that need to identify the polarity of the electronic component.

Step S102: Perform forward calculation on the image data including the target electronic component by using the trained convolutional neural network to obtain a classification feature of the polarity direction category of the target electronic component, and obtain the target electronic component belonging to various electronic components according to the classification feature. Probability distribution of various polar direction categories;

In this step, the image data including the target electronic component is forwardly calculated by using the convolution module and the fully connected layer in the convolutional neural network after training, and the image data including the target electronic component is passed through the convolution module. And the fully connected layer can obtain the classification characteristics of the polarity direction category of the target electronic component; the various polarity direction categories refer to various polarity directions of various electronic components, including various electronic components, and the probability distribution obtained can be obtained. Applicable to all kinds of electronic components; the trained convolutional neural network can operate on the image, identify the polarity direction of the target electronic components, and obtain the probability distribution.

Step S103: Select the polarity direction category with the highest probability as the polarity direction category of the target electronic component.

The above steps S101, S102 and S103 are processes for performing online testing using the trained convolutional neural network.

The method for quickly identifying the polarity direction of an electronic component according to the present embodiment is to acquire an image including a target electronic component, and then perform forward calculation using the trained convolutional neural network to obtain a polarity direction category of the target electronic component. The classification feature is obtained, and then the probability distribution of the target electronic components belonging to various polarity direction categories of various electronic components is obtained, and the polarity direction category with the highest probability is selected as the polarity direction category of the target electronic component. In this scheme, a convolutional neural network is used. The convolutional neural network can automatically and accurately identify the polarity direction of electronic components, and it is not specific to the structure of electronic components. It is applicable to various electronic components with polarity to achieve cross- Polarity identification of electronic components of the category, applicability wide.

In this embodiment, better classification features can be obtained by convolution layers, activation function layers and pooling layers sequentially connected in the convolution module, and fully connected layers connected to the convolution module.

Preferably, the activation function layer may be a ReLU function layer.

As shown in FIG. 2, the convolution module includes a convolution layer, a ReLU function layer, and a pooling layer that are sequentially connected, and convolves the image data of the target electronic component through the convolution layer, and then performs nonlinear transformation through the ReLU function layer. Then, the pooling operation is performed through the pooling layer, and then the classification feature of the polarity direction category of the target electronic component is obtained through the fully connected layer.

The ReLU function in the convolutional module of the convolutional neural network is an activation function that can be used for nonlinear transformation; the pooling layer is used for pooling operations, and can aggregate statistics on features of different positions of image samples, and This will cause the final result to be over-fitting.

The convolutional neural network has a very strong expression ability, which can effectively solve the problem of low precision in multi-classification tasks. Even in the cross-class electronic component polarity direction recognition task, a very high accuracy can be achieved.

The use of external data sets to train convolutional neural networks can enhance the identification of convolutional neural networks, but High hardware computing power requirements, usually require high-end GPUs to meet the actual computing speed requirements, thus increasing hardware costs. The technical solution of the invention does not need to use the external data set to train the convolutional neural network, and can accurately and quickly determine the automatic polarity direction of the electronic component under the low-cost hardware configuration, and is suitable for various electronic components (such as capacitors and sockets). , resistance, etc.). The main feature of the scheme is that it does not depend on special hardware mechanisms, and the cost is low, which greatly reduces the requirements of the algorithm for hardware computing power and storage space, and solves the problem of high cost of high-precision technical solutions.

At present, existing methods for identifying the polarity direction of electronic components often utilize low-level image features such as color information of electronic components. This conventional method is too simple for many application scenarios, has low robustness, and is limited in scope of application. The effect is poor; in this scheme, the convolutional neural network is used to learn multi-layer feature representation (including low-level, middle-level, high-level image features, not just low-level image features), and such multi-layer features are combined to identify electronic components. The polarity direction greatly enhances the recognition accuracy and extends the scope of application.

Through this step, the convolutional neural network can be offlinely trained according to the image sample sets of various polarity directions of various electronic components, so that the convolutional neural network can recognize the polarity direction information of the electronic components in the image, thereby being online The trained convolutional neural network is used in the test to identify the polarity direction of the target electronic component.

The image sample sets obtained by the above methods are directly taken from the PCB card image, and are images of actual electronic components, which are more instructive.

Preferably, the camera can be set up on the PCB board production line, and different types of PCBs can be collected in batches. Card image, and board tracking technology to avoid repeating a certain PCB board, so each type of PCB board contains multiple image samples, each image sample corresponds to a certain type of PCB board; The corresponding PCB template map needs to be saved when collecting the PCB board image;

The position of the PCB card image in multiple image samples may be offset. It is necessary to use the corresponding template PCB template map as a reference, perform image registration for each image sample, and use the electronic component position information of the model PCB card (from The panel file or manual labeling automatically intercepts the electronic component image and automatically labels the electronic component image according to the electronic component category information (from the panel file or manual labeling);

The precise position of the electronic components in the image sample is further obtained by matching the electronic components in the PCB template image, and the electronic component images are aligned and adjusted to ensure that the electronic components are located at the center of the image, and image samples of various polar directions of the electronic components are obtained. set.

In one embodiment, as shown in FIG. 3, there are five convolutional modules, wherein the number of convolution kernels of the five convolutional layers is 24, 64, 96, 96, and 64, respectively, and 5 convolutional layers. The size of the convolution kernel is 7×7, 5×5, 3×3, 3×3, 3×3, and the steps of the five convolution layers are all 1; the total connection layer is 2, of which 2 are all The number of hidden nodes in the connection layer is 512 and 4, respectively.

With the convolutional neural network in this embodiment, the rapid identification of the polarity direction of the electronic component can be better achieved.

Obtaining a template image of the target electronic component, matching the target electronic component in the image with a template image of the target electronic component, obtaining an accurate position of the target electronic component in the image, and adjusting the target electronic component in the image according to the precise position, so that The target electronic component is located at the center of the image, and the adjusted image is used for forward calculation by the trained convolutional neural network.

The target electronic components in the image are adjusted by the template image of the target electronic component, so that the trained convolutional neural network can more easily identify the target electronic component.

In one of the embodiments, after the step of selecting the polarity direction category having the highest probability as the polarity direction category of the electronic component, the following steps are further included:

If the polarity direction category with the highest probability of selection is different from the preset polarity direction category of the electronic component, an error alarm message is given.

The above steps are a process of detecting the polarity direction, and it is possible to judge whether or not the electronic component is correctly mounted.

In one embodiment, the image of the target electronic component acquired by the online test is a PCB card image, and the corresponding PCB template is matched and matched to obtain the precise position of the target electronic component, and the target electronic component in the image is aligned and adjusted. To ensure that the target electronic component is at the center of the image.

In this embodiment, the corresponding PCB template is used for the matching, and the PCB template may be a PCB template corresponding to the image sample set image, or may be a PCB template different from the PCB template corresponding to the image sample set image, if In the former, the PCB template corresponding to the image sample set image has been saved during offline training, and can be directly used. In the latter case, in addition to obtaining the PCB board image, the online PCB template needs to obtain the corresponding PCB template from the outside to achieve the subsequent matching match. The implementation of the solution can identify the polarity direction of all electronic components in the acquired PCB card image.

In one of the embodiments, the steps of establishing a sample set of various polarity directions of various electronic components and the step of acquiring an image containing the target electronic component include the following steps:

The size of the adjusted image is normalized.

In the step of establishing a sample set of various polarity directions of various electronic components, normalizing the size of the adjusted image may facilitate training of the convolutional neural network to make the convolutional neural network image sample The learning of multi-layer features is more accurate.

After the step of acquiring the image including the target electronic component, the normalized processing of the size of the adjusted image facilitates the forward calculation processing of the image of the target electronic component by the convolutional neural network, and accelerates the recognition process of the electronic component.

The method for quickly identifying the polarity direction of the electronic component of the present invention uses a convolutional neural network to automatically and accurately identify the polarity direction of the electronic component, and is not specific to a specific electronic component structure, and is applicable to various polarities. As for the electronic component, as long as the image sample set for training the convolutional neural network covers a plurality of types of electronic components, the polarity direction identification of the electronic components across the categories can be realized, and the applicability is wide. Running on the current mainstream GPU can also judge the regular PCB board (including about 20-30 components with polarity), and judge the average calculation time of a component on the GPU even within 1 millisecond. Meet a variety of real-time judgment scenarios; whether it is offline training or online testing, minimize manual intervention, and does not rely on special hardware mechanisms, the cost is lower; no need Powerful computing resources can even run on inexpensive embedded platforms in some scenarios.

The invention also provides a method for marking the polarity direction of an electronic component, comprising the following steps:

According to the above-described rapid identification method for the polarity direction of the electronic component, the present invention also provides a rapid identification system for the polarity direction of the electronic component. Hereinafter, an embodiment of the rapid identification system for the polarity direction of the electronic component of the present invention will be described in detail.

Referring to Fig. 4, there is shown an embodiment of a rapid identification system for the polarity direction of an electronic component of the present invention. The fast identification system of the polarity direction of the electronic component in this embodiment includes the acquisition unit 210, the calculation unit 220 and the selection unit 230 in FIG. 4;

An acquiring unit 210, configured to acquire an image that includes the target electronic component;

The calculating unit 220 is configured to perform forward calculation on the image data including the target electronic component by using the trained convolutional neural network, obtain a classification feature of the polarity direction category of the target electronic component, and obtain the target electronic component according to the classification feature. Probability distribution of various polar direction categories of electronic components;

The selecting unit 230 is configured to select a polarity direction category with the highest probability as the polarity direction category of the target electronic component.

According to the above fast recognition system, the image is obtained by including the target electronic component, and then subjected to the forward calculation by using the trained convolutional neural network to obtain the classification feature of the polarity direction category of the target electronic component, and then the target electronic component is obtained. The probability distribution of various polarity direction categories belonging to various electronic components is selected as the polarity direction category of the target electronic component. This scheme uses a convolutional neural network to automatically and accurately identify the poles of electronic components through convolutional neural networks. Sexual orientation, and is not specific to the structure of electronic components. It is applicable to various electronic components with polarity, and realizes the polarity direction identification of electronic components across categories, and has wide applicability.

In one embodiment, the calculation unit 220 performs convolution operation on the image data through the convolution layer, then performs nonlinear transformation through the activation function layer, performs pooling operation through the pooling layer, and then obtains target electrons through the fully connected layer. A classification feature of a polarity direction category of the component, wherein the convolutional neural network includes a convolution module and a fully connected layer, and the convolution module includes a convolution layer, an activation function layer, and a pooling layer that are sequentially connected.

In one embodiment, as shown in FIG. 5, the rapid identification system of the polarity direction of the electronic component further includes an establishing unit 240 and a training unit 250;

The establishing unit 240 is configured to establish an image sample set of various polarity directions of various electronic components;

The training unit 250 is configured to perform forward calculation on each image sample data of the image sample set by using the convolution module and the fully connected layer in the convolutional neural network, and obtain classifications of various polarity direction categories of various electronic components. The feature trains the convolutional neural network according to each classification feature, so that the convolutional neural network can recognize various polar directions of various electronic components.

In one embodiment, as shown in FIG. 6, the establishing unit 240 includes a registration unit 241 and an intercept processing unit 242;

The registration unit 241 is configured to acquire a PCB card image and a PCB template image, and perform position registration on the PCB card image by using the PCB template image as a reference;

The intercepting processing unit 242 is configured to intercept the electronic component image on the PCB card image after the position registration, and match various electronic components in the electronic component images by various electronic components in the PCB template image to obtain various electronic components. The precise position of the corresponding electronic components in the image, the corresponding electronic components are adjusted according to the precise position of the corresponding electronic components, so that the corresponding electronic components are located at the center of the image of various electronic components, and image samples of various polar directions of various electronic components are obtained. set.

In one embodiment, there are five convolutional modules, wherein the number of convolution kernels of the five convolutional layers is 24, 64, 96, 96, and 64, respectively, and the convolution kernel sizes of the five convolutional layers are respectively 7×7, 5×5, 3×3, 3×3, 3×3, the steps of the five convolutional layers are all 1; the total connection layer is 2, wherein the number of hidden nodes of the 2 fully connected layers They are 512 and 4 respectively.

In one embodiment, as shown in FIG. 7, the rapid identification system of the polarity direction of the electronic component is further Include a pre-processing unit 260;

The pre-processing unit 260 is configured to match the target electronic component in the image with the template image of the target electronic component after acquiring the image of the target electronic component, to obtain the precise position of the target electronic component in the image, according to the precise position The target electronic components in the image are adjusted such that the target electronic components in the image are located at the center of the image, and the adjusted images are used for forward calculation by the trained convolutional neural network.

In one embodiment, as shown in FIG. 8, the fast identification system of the polarity direction of the electronic component further includes an alarm unit 270 for selecting the polarity direction category with the highest probability and the preset polarity direction of the electronic component. Different, the error alarm information is given.

The rapid identification system of the polarity direction of the electronic component of the present invention corresponds to the rapid identification method of the polarity direction of the electronic component of the present invention, and the technical features described in the embodiment of the rapid identification method of the polarity direction of the electronic component are beneficial. The effects are all applicable to embodiments of the fast identification system for the polarity direction of the electronic components.

The present invention also provides an annotation system for the polarity direction of an electronic component, as shown in FIG. 9, comprising an identification unit 300 and the above-described identification system for the polarity direction of the electronic component, and the labeling unit 300 is configured to be based on the polarity direction of the electronic component. The polar identification direction of the target electronic component determined by the system is quickly identified, the polarity direction information of the electronic component is marked in the panel file, and the panel file is used to store various attribute information of the electronic component.

The technical features of the above-described embodiments may be arbitrarily combined. For the sake of brevity of description, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction between the combinations of these technical features, All should be considered as the scope of this manual.

The above described embodiments merely express several embodiments of the present invention, and the description thereof is more specific and detailed. However, it cannot be construed as limiting the scope of the invention patent. It should be noted that a number of variations and modifications may be made by those skilled in the art without departing from the spirit and scope of the invention. Therefore, the scope of the invention should be determined by the appended claims.

Claims

A method for quickly identifying a polarity direction of an electronic component, comprising the steps of:

Obtain an image containing the target electronic component;

Performing forward calculation on the image data including the target electronic component by using the trained convolutional neural network to obtain a classification feature of the polarity direction category of the target electronic component, and acquiring the target electronic component according to the classification feature Probability distribution of various polar direction categories of electronic components;

The polarity direction category with the highest probability is selected as the polarity direction category of the target electronic component.
The method for quickly identifying a polarity direction of an electronic component according to claim 1, wherein said using said trained convolutional neural network performs forward calculation on image data including target electronic components to obtain said target electronic component The steps of the classification feature of the polarity direction category include the following steps:

Convoluting the image data by a convolution layer, then performing nonlinear transformation through an activation function layer, performing a pooling operation through the pooling layer, and then obtaining a polarity direction category of the target electronic component through the fully connected layer The classification feature, wherein the convolutional neural network comprises a convolution module and a fully connected layer, the convolution module comprising the convolution layer, the activation function layer and the pooling layer connected in sequence.
The method for quickly identifying a polarity direction of an electronic component according to claim 1, wherein the trained convolutional neural network is obtained by the following steps:

Establishing image sample sets of various polarity directions of various electronic components;

Performing forward calculation on each image sample data of the image sample set by using a convolution module and a fully connected layer in the convolutional neural network to obtain classification features of various polarity direction categories of various electronic components, according to Each of the classification features trains the convolutional neural network to cause the convolutional neural network to identify various polar directions of the various types of electronic components.
The method for quickly identifying a polarity direction of an electronic component according to claim 3, wherein the step of establishing a sample set of various polarity directions of the various types of electronic components comprises the following steps:

Obtaining a PCB card image and a PCB template image, and performing position registration on the PCB card image by using the PCB template image as a reference;

Obtaining various types of electronic component images on the PCB card image after the position registration, and matching the corresponding electronic components in the image of the various types of electronic components with various electronic components in the PCB template image to obtain The precise position of the corresponding electronic component in the image of the various types of electronic components, the corresponding electronic component is adjusted according to the precise position of the corresponding electronic component, so that the corresponding electronic component is located at the center of the image of each electronic component, and obtains A collection of image samples of various polarity directions of various types of electronic components.
The method for quickly identifying the polarity direction of an electronic component according to claim 2, wherein the number of convolution modules is five, wherein the number of convolution kernels of the five convolution layers is 24, 64, respectively. 96, 96 and 64, the convolution kernel sizes of the five convolutional layers are 7×7, 5×5, 3×3, 3×3, 3×3, and the steps of the five convolution layers are respectively All of the two are connected to each other, and the number of hidden nodes of the two fully connected layers is 512 and 4, respectively.
The method for quickly identifying a polarity direction of an electronic component according to any one of claims 1 to 5, characterized in that after the step of acquiring an image including the target electronic component, the following steps are included:

Obtaining a template image of the target electronic component, matching a target electronic component in the image with a template image of the target electronic component, obtaining an accurate position of the target electronic component in the image, according to the precise position The target electronic component in the image is adjusted such that the target electronic component in the image is located at the center of the image, and the adjusted image is used for forward calculation by the trained convolutional neural network.
A method for marking a polarity direction of an electronic component, comprising the steps of:

The polarity direction of the target electronic component determined by the rapid identification method of the polarity direction of the electronic component according to any one of claims 1 to 6, wherein the polarity direction information of the target electronic component is marked in the panel file, A panel file is used to store various attribute information of electronic components.
A rapid identification system for the polarity direction of electronic components, characterized in that it comprises the following units:

Acquiring a unit to obtain an image containing the target electronic component;

a calculating unit, configured to perform forward calculation on the image data including the target electronic component by using the trained convolutional neural network, obtain a classification feature of the polarity direction category of the target electronic component, and acquire the target according to the classification feature Electronic components belong to the probability distribution of various polarity direction categories of various electronic components;

The selecting unit is configured to select a polarity direction category with the highest probability as the polarity direction category of the target electronic component.
A rapid identification system for polarity direction of an electronic component according to claim 8, wherein

The calculating unit performs convolution operation on the image data through a convolution layer, then performs nonlinear transformation through an activation function layer, performs a pooling operation through the pooling layer, and then obtains the target electronic component through the fully connected layer. a classification feature of a polarity direction category, wherein the convolutional neural network comprises a convolution module and a fully connected layer, the convolution module comprising the convolution layer sequentially connected, the activation function layer, and the pooling Floor.
An annotation system for polarity direction of an electronic component, comprising: a labeling unit and a rapid identification system for polarity direction of an electronic component according to claim 8 or 9, wherein the labeling unit is used in a panel file Polarity direction information of the target electronic component is marked, and the panel file is used to store various attribute information of the electronic component.