Automatic recognition system and method for circuit board gripping position

Abstract

This invention provides an automatic identification system and method for the gripping position of a circuit board. The invention simplifies the layout of a standard circuit board to obtain a training layout that removes identification noise. This training layout, with the gripping position set, is then input into a training model to enable the model to identify circuit boards of the same type. Furthermore, when the training model identifies a target image as a standard circuit board, the invention performs matching processing on the target image to determine the gripping position. This invention can accurately identify the gripping position of a target image and enables unmanned automatic circuit board gripping.

Description

Technical Field

[0001] This invention relates to an automatic identification system and method, and more particularly to an automatic identification system and method for identifying the position of a circuit board gripping device. Background Technology

[0002] To ensure that the circuit board is gripped from the correct gripping position, a method for automatically identifying the gripping position has been proposed.

[0003] Specifically, the existing method involves manually taking a large number of images of the circuit board, using these images to train a recognition model, and then using this recognition model to identify the circuit board.

[0004] In the above methods, since the image content collected by different people and the learning algorithm used are usually different, the accuracy of the recognition model will vary.

[0005] Specifically, because circuit boards contain a large number of complex components, their flexible boards are prone to deformation and reflection. Training images taken manually usually contain a lot of noise, which reduces the accuracy of the recognition model.

[0006] Furthermore, the accuracy of the recognition model is closely related to the selection of recognition regions and features in the training images. Unsuitable learning algorithms and inappropriate recognition features will reduce the accuracy of the recognition model.

[0007] Therefore, the existing methods for automatically identifying the position of the circuit board have the problems caused by the aforementioned human factors, and there is an urgent need for a more effective solution. Summary of the Invention

[0008] The main objective of this invention is to provide an automatic identification system and method for the gripping position of a circuit board, which can completely eliminate human intervention that causes instability in the accuracy of the training model and provide stable identification features.

[0009] In one embodiment, an automatic identification method for a circuit board gripping position includes: a) obtaining multiple layout diagrams of multiple standard circuit boards of different types, and performing a simplification process on the layout diagram of each standard circuit board to obtain a training layout diagram with eliminated identification noise; b) performing a training process on a training model based on each training layout diagram to enable the training model to identify the multiple standard circuit boards of different types, wherein each training layout diagram is configured with a gripping position; c) inputting a target image of a target circuit board into the training model to perform a recognition process on the target image; and d) when the target image is identified as conforming to any of the standard circuit boards, performing a matching process on the target image to determine a gripping position of the target image based on the gripping position of the training layout diagram of the conforming standard circuit board; the recognition process includes: e1) performing a target edge analysis on the target image to obtain multiple target element edge features of the target image; e2) calculating a similarity score of the target image to each standard circuit board based on the target element edge features and multiple identification features of each standard circuit board; and e3) determining a recognition result based on the standard circuit board with the highest score.

[0010] In one embodiment, an automatic identification system for a circuit board gripping position includes: an image acquisition device, a gripping device, and a control device. The image acquisition device is used to capture a target circuit board to obtain a target image. The gripping device is used to grip the circuit board. The control device is connected to the image acquisition device and the gripping device, and includes a simplification module, a training module, and a recognition module. The simplification module is configured to perform a simplification process on a layout diagram of a standard circuit board to remove recognition noise and obtain a training layout diagram. The training module is configured to input the training layout diagram into a training model to enable the training model to recognize circuit boards of the same type, wherein a gripping position is defined on the training layout diagram. The recognition module inputs the target image into the training model to perform recognition. When the target image is identified as the standard circuit board, a matching process is performed on the target image to determine a gripping position of the target image based on the gripping position of the training layout diagram of the matching standard circuit board. The control device is configured to control the gripping device to grip the target circuit board based on the gripping position of the target image.

[0011] This invention can accurately identify the grasping position of the target image and realize unmanned automatic grasping of circuit boards. Attached Figure Description

[0012] Figure 1 This is an architecture diagram of an automatic identification system according to an embodiment of the present invention;

[0013] Figure 2 This is an architecture diagram of an automatic identification system according to an embodiment of the present invention;

[0014] Figure 3 This is an architectural diagram of a control device according to an embodiment of the present invention;

[0015] Figure 4 This is a schematic diagram of an automatically identified data stream according to an embodiment of the present invention;

[0016] Figure 5 This is a flowchart of an automatic identification method according to an embodiment of the present invention;

[0017] Figure 6 This is a simplified flowchart of an embodiment of the present invention;

[0018] Figure 7 This is a flowchart of the training and recognition processes according to an embodiment of the present invention;

[0019] Figure 8 This is a schematic diagram of a circuit board clamping device according to an embodiment of the present invention;

[0020] Figure 9 This is a schematic outline diagram of a layout according to an embodiment of the present invention;

[0021] Figure 10 This is a schematic diagram illustrating the processing result of a fragmented component filtering process according to an embodiment of the present invention;

[0022] Figure 11 This is a schematic diagram of the processing result of a repeating element filtering process according to an embodiment of the present invention;

[0023] Figure 12 This is a schematic diagram of the processing result of a tortuous element filtering process according to an embodiment of the present invention;

[0024] Figure 13 This is a schematic diagram of a target image according to an embodiment of the present invention;

[0025] Figure 14 This is a schematic diagram of another target image according to an embodiment of the present invention.

[0026] Explanation of icon numbers

[0027] 10: Control equipment

[0028] 11: Image acquisition device

[0029] 12: Grabbing equipment

[0030] 13: Storage devices

[0031] 130: Layout Diagram

[0032] 131: Training the model

[0033] 132: Computer program

[0034] 20: Target circuit board

[0035] 30: Processing equipment

[0036] 31: Human-computer interface

[0037] 32: Communication Interface

[0038] 40: Computing Platform

[0039] 500: Shooting Control Module

[0040] 501: Simplified Module

[0041] 502: Preprocessing module

[0042] 503: Contour Extraction Module

[0043] 504: Miscellaneous Component Filtering Module

[0044] 505: Repetitive Component Filtering Module

[0045] 506: Twisted Element Filtering Module

[0046] 507: Training Module

[0047] 508: Recognition Module

[0048] 509: Grab Control Module

[0049] 510: Processing Control Module

[0050] 60: Suction and drive equipment

[0051] 61: Circuit board transport equipment

[0052] 62, 62': Position

[0053] 70-73: Layout Diagram

[0054] 74, 75: Target Images

[0055] 740-748, 830-838: Image Features

[0056] 80-82: Pattern

[0057] S10-S11: Training Steps

[0058] S20-S24: Grabbing Steps

[0059] S30-S35: Simplified Procedures

[0060] S40-S41: Training Processing Steps

[0061] S50-S52: Identification and Processing Steps Detailed Implementation

[0062] A preferred embodiment of the present invention will be described in detail below with reference to the accompanying drawings.

[0063] For automated and unmanned production lines to be realized, the accuracy of automatic circuit board clamping is extremely important.

[0064] To accurately identify the gripping position of a circuit board (i.e., the position of the circuit board and its components during gripping), this invention proposes an automatic recognition system and method for the gripping position of a circuit board. A layout drawing can be used to train a model, enabling the model to correctly identify such circuit boards. Furthermore, this invention uses the trained model to identify the type of the target circuit board (i.e., the physical circuit board to be processed) and determines the gripping position based on the recognition results.

[0065] Please see Figure 1 This is an architecture diagram of an automatic recognition system according to an embodiment of the present invention. The automatic recognition system of this embodiment mainly includes an image acquisition device 11, a grasping device 12, a storage device 13, and a control device 10 connected to the above devices.

[0066] Image acquisition device 11, such as a color camera, infrared camera, 3D camera or other optical camera, is used to photograph the target circuit board 20 to obtain target images (such as color images, grayscale images, black and white images, 3D images, etc.).

[0067] Gripping device 12, such as an automated machine capable of gripping circuit boards, is used to grip the target circuit board 20 and move it to a designated location (such as the processing location described later). Gripping device 12 can grip the circuit board by means including clamping (such as mechanical clamping) and suction (such as magnetic or attractive force).

[0068] Storage device 13 is used to store data. In one embodiment, storage device 13 can store layout diagrams 130 of various standard circuit boards, the layout diagrams 130 being 2D images that can be obtained by converting design drawings (typically vector or 3D diagrams) of various standard circuit boards.

[0069] In one embodiment, storage device 13 may store training model 131, such as a machine learning model or a data model recording image recognition features. Training model 131 is used to learn the ability to recognize various standard circuit boards by receiving data for training.

[0070] Control device 10, such as a processor, a machine control computer or control box equipped with a processor, is used to control the operation of various devices in the automatic identification system, such as executing the training mode and grasping mode described later.

[0071] Please refer to the following: Figure 2 This is an architecture diagram of an automatic identification system according to an embodiment of the present invention. Compared to Figure 1 In this embodiment, the automatic identification system further includes a processing device 30 connected to the control device 10, such as circuit board manufacturing equipment (e.g., soldering / assembly equipment) or inspection equipment (e.g., component / contact inspection equipment). The processing device 30 is used to perform processing (e.g., soldering, assembly, or inspection) on the target circuit board 20 placed at the processing location.

[0072] In one embodiment, the automatic identification system further includes a human-machine interface 31 connected to the control device 10, such as any combination of input / output devices such as a touch screen, buttons, display screen, indicator lights, and buzzers, used to provide information and interact with the user.

[0073] In one embodiment, the automatic identification system further includes a communication interface 32 for connecting to the control device 10, such as a network interface that can connect to a local area network or the Internet, such as a network card, a Wi-Fi module, or a cellular network module.

[0074] In one embodiment, the communication interface 32 can be used to connect to the computing platform 40 (such as a cloud server or a remote host) and communicate with the computing platform 40.

[0075] In one embodiment, the aforementioned layout diagram 130 and training model 131 can be stored in the computing platform 40. The computing platform 40 can be used to execute the training mode described later to train the training model 131 using the layout diagram 130 (which can be provided by the automatic recognition system, obtained from the circuit board database, or obtained by converting the design drawing itself), and transmit the trained training model 131 to the automatic recognition system to execute the grasping mode.

[0076] In one embodiment, the automatic identification system further includes a power device (not shown) connected to the control device 10, the power device being used to connect to a fixed power source (such as mains power) and convert the received power into the power required for the operation of the automatic identification system.

[0077] Please refer to the following: Figure 3 This is an architectural diagram of a control device according to an embodiment of the present invention. In this embodiment, the control device 10 may include modules 500-510. These modules 500-510 are respectively configured to perform different functions.

[0078] The shooting control module 500 is configured to control the image acquisition device 11 to shoot the target circuit board 20 to obtain the target image.

[0079] The simplification module 501 is configured to perform simplification processing on the image to filter out recognition noise in the image and retain highly recognizable recognition features, such as distinctive patterns, pattern combinations and / or pattern distributions.

[0080] In one embodiment, the simplification module 501 may include modules 502-506.

[0081] The preprocessing module 502 is configured to perform preprocessing on the image, such as cropping, color space conversion, contrast enhancement, etc., to make the image conform to the specified format or improve its recognizability.

[0082] The contour extraction module 503 is configured to identify the contours of multiple components of a circuit board in an image.

[0083] The fragmented component filtering module 504 is configured to filter out small components from the image. These small components (such as tiny copper foils or lines) appear as small specks on the image, cannot be used as identification features, and will become noise for image recognition.

[0084] The repeating element filtering module 505 is configured to identify a large number of recurring elements in an image and filter out these recurring elements. These repeating elements are repetitive or similar designs that are close in location and have the same or similar shape, which may affect the accuracy of subsequent positioning and are not suitable as recognition features.

[0085] The zigzag element filtering module 506 is configured to identify zigzag elements (such as elements with multiple corners) in an image and filter them out. Zigzag elements are unsuitable as identification features because their unstable angles can affect the accuracy of subsequent positioning.

[0086] Training module 507 is configured to execute training mode. Training module 507 can input training images (such as training layout diagrams) to training model 131 to perform training processing on training model 131 so that it has the ability to recognize the same type of circuit board.

[0087] In one embodiment, the training image may be set with a grasping position, so that the trained model 131 can automatically identify the grasping position after training.

[0088] The grasping control module 509 is configured to control the grasping device 12 to grasp the target circuit board 20 based on the grasping position of the target image obtained by recognition.

[0089] In one embodiment, after the gripping is completed, the gripping control module 509 can further control the gripping device 12 to transport the target circuit board 20 to the processing position.

[0090] The processing control module 510 is configured to control the processing equipment 30 to perform processing on the target circuit board 20 at the processing position.

[0091] The aforementioned modules 500-510 are interconnected (which can be electrical or informational connections) and can be hardware modules (such as electronic circuit modules, integrated circuit modules, SoCs, etc.), software modules (such as firmware, operating systems, or applications), or a combination of hardware and software modules, without limitation.

[0092] It is worth mentioning that when the aforementioned module is a software module (such as firmware, operating system or application), the storage device 13 or the computing platform 40 may include a non-transitory computer-readable recording medium (not shown in the figure). The aforementioned non-transitory computer-readable recording medium stores a computer program 132, which records computer-executable program code. When the processor of the control device 10 or the computing platform 40 executes the aforementioned program code, the function of the corresponding module can be implemented.

[0093] Please refer to the following: Figure 4 This is a schematic diagram of an automatically identified data stream according to an embodiment of the present invention. Figure 4 The data flow of the training mode (top figure) and the capture mode (bottom figure) according to an embodiment of the present invention is shown.

[0094] like Figure 4 As shown in the figure above, the present invention first performs a simplification process on the layout diagram 130 of a standard circuit board to generate training data (i.e., a training layout diagram), and then inputs the training data into the training model 131 to perform training to obtain the recognition features of the training data. By performing training on different training data, the recognition features of different standard circuit boards can be obtained, that is, the training model 131 can recognize different types of circuit boards.

[0095] like Figure 4 As shown in the figure below, the present invention then inputs the target image of the target circuit board 20 into the trained model 131 to identify the circuit board type of the target circuit board 20, and obtains the corresponding recognition features (including the grasping position) based on the identified circuit board type. Finally, the obtained recognition features are used to perform matching processing on the target image to confirm the grasping position of the target image by locating these recognition features.

[0096] This invention achieves extremely high recognition and positioning accuracy because it uses a training layout diagram (an electronic file obtained from a design drawing) with no or very little noise to train recognition features, rather than images obtained by taking pictures (which contain a lot of noise).

[0097] Please refer to the following: Figure 5 This is a flowchart of an automatic identification method according to an embodiment of the present invention. The automatic identification methods of the various embodiments of the present invention can be applied to... Figures 1-4 , Figure 8 Any of the systems shown.

[0098] The automatic identification method in this embodiment includes a training mode (steps S10-S11).

[0099] Step S10: The control device 10 first obtains the layout diagram 130 of the standard circuit board, and performs simplification processing on the layout diagram 130 through the simplification module 501 to filter out identification noise, thereby obtaining a training layout diagram with the identification noise removed.

[0100] The aforementioned layout diagram 130 is a 2D image showing the distribution of electronic components and the circuit layout of a standard circuit board.

[0101] Step S11: The control device 10 inputs the generated training layout diagram to the training model 131 through the training module 507 to perform training processing on the training model 131, thereby enabling the training model 131 to recognize the same type of circuit board.

[0102] In one embodiment, the training layout diagram is configured with suitable gripping positions for this type of circuit board, such as positions with fewer 3D components, positions with fewer or no components, positions with higher flexibility on the circuit board, and positions that are easy to maintain balance during gripping (such as the center of gravity or center of the circuit board). The trained model 131 can identify suitable gripping positions for this type of circuit board from the input image when an image of the same type is input. Furthermore, when gripping the circuit board at this position, because the gripping position has fewer components, higher flexibility, and / or is easy to balance, the probability of damage to the circuit board during gripping can be significantly reduced, such as avoiding damage to components, damage to the circuit board, or the circuit board falling during gripping.

[0103] In one embodiment, the present invention can obtain multiple layout diagrams 130 of multiple standard circuit boards of different types, and perform the above-described training mode on each of the multiple layout diagrams 130, so that the training model 131 can simultaneously have the ability to identify multiple standard circuit boards of different types.

[0104] In one embodiment, the present invention performs a training mode on multiple layout diagrams 130 of multiple standard circuit boards of multiple types, which can generate multiple classification rules for standard circuit boards. These classification rules can be used to classify the image features of the input image to determine the type of circuit board to which the input image belongs.

[0105] The automatic identification method in this embodiment also includes a capture mode executed after training (steps S20-S24).

[0106] Step S20: The control device 10 can first control the image acquisition device 11 to take a picture of the target circuit board 20 that has arrived at the preparation position through the image acquisition control module 500 to obtain the target image, and then input the target image into the training model 131 through the recognition module 508 to perform recognition processing.

[0107] Step S21: The control device 10 determines whether the input target image conforms to any pre-established standard circuit board based on the result of the recognition process.

[0108] In one embodiment, during the recognition process, the training model 131 can compare the target image with the image features of each pre-established standard circuit board, and determine that the target circuit board 20 belongs to the standard circuit board when it is determined that the target image contains all or most of the image features of any standard circuit board (including shape matching and position matching).

[0109] In one embodiment, during the identification process, the training model 131 can extract image features of the target image and compare the image features with pre-established classification rules to determine the standard circuit board to which the target circuit board 20 belongs through classification.

[0110] If the control device 10 determines that the target image conforms to any standard circuit board, then step S22 is executed; otherwise, the control device 10 executes step S24.

[0111] Step S22: The control device 10 obtains the recognition-related data (such as recognition features or training layout diagram, including the grasping position) of the standard circuit board, and performs matching processing on the target image and the recognition-related data through the recognition module 508 to determine the grasping position in the target image. The grasping position in the target image is the grasping position of the training layout diagram set when the training mode is executed.

[0112] In one embodiment, the identification data includes the arrangement relationship between identification features and grasping positions. After the control device 10 identifies these identification features from the target image, it can calculate the grasping position in the target image based on the above arrangement relationship.

[0113] In one embodiment, the relevant identification data includes a training layout diagram with a set grasping position. After the control device 10 identifies multiple identification features from the target image, it can adjust the target image (such as scaling and / or rotating) based on the direction, rotation angle, distance, etc. of these features in the training layout diagram, so that the target image and the training layout diagram present the same direction and / or size, and the grasping position of the training layout diagram can be clearly known to correspond to the position in the target image.

[0114] Step S23: Based on the determined grasping position of the target image, the control device 10 converts the image position into a robot coordinate position through the grasping control module 509, generates a corresponding grasping command, and executes the grasping command to control the grasping device 12 to grasp the target circuit board 20 from the grasping position of the target circuit board 20.

[0115] In one embodiment, the control device 10 may further control the gripping device 12 to move the gripped target circuit board 20 to the processing position of the processing device 30.

[0116] Furthermore, the control device 10 can perform processing on the target circuit board 20 at the processing position through the processing control module 510 and the processing equipment 30.

[0117] If, in step S21, the control device 10 determines that the target image does not conform to all standard circuit boards, then step S24 is executed: the control device 10 issues a warning through the human-machine interface 31 to remind the user of the recognition failure.

[0118] In one embodiment, when the target image does not conform to all standard circuit boards, the control device 10 can automatically switch to machine learning to identify a suitable gripping position of the target circuit board 20, such as a position with low component density, the center or center of gravity, and then perform gripping at the identified gripping position.

[0119] In one embodiment, if the target image does not conform to all standard circuit boards, the control device 10 may accept manual operation to perform the grasping.

[0120] This invention can accurately identify the grasping position of the target image and realize unmanned automatic grasping of circuit boards.

[0121] Please see Figure 8 This is a schematic diagram of a circuit board gripping device according to an embodiment of the present invention. In this embodiment, the gripping device 12 includes a suction and driving device 60 and a circuit board transport device 61 connected to the control device 10.

[0122] First, the circuit board transport device 61 can move the target circuit board 20 to the gripping range (i.e., the ready position 62) of the suction and drive device 60. Here, the target circuit board 20 is transported by a trackless automated guided vehicle and the lifting platform is used to send the target circuit board 20 to a designated height for taking a picture to trigger the image acquisition device 11 to take a picture of the target. Alternatively, a conveyor belt or other circuit board transport device can be used.

[0123] Next, the suction and drive device 60 can move the suction nozzle to the preparation position 62 by a drive mechanism (such as a robotic arm) and suck up the target circuit board 20 from the gripping position and leave the preparation position 62, as shown in position 62'.

[0124] Next, the suction and drive device 60 can be moved to the processing position by the drive mechanism and the target circuit board 20 can be released to the processing position.

[0125] Therefore, the present invention enables unmanned automatic grasping and placement of circuit boards.

[0126] Please refer to the following: Figure 5 and Figure 6 , Figure 6 This is a simplified flowchart of a process according to an embodiment of the present invention. Compared to Figure 5 In this embodiment, the simplified processing of the automatic identification method includes the following steps S30-S35.

[0127] Step S30: Control device 10 obtains the layout diagram 130 of the standard circuit board.

[0128] In one embodiment, the control device 10 may first obtain a design drawing of a standard circuit board and perform 2D conversion processing on the design drawing to obtain a 2D layout drawing.

[0129] The aforementioned design drawings can be in Gerber format or computer-aided design (CAD) format. The obtained layout drawing 130 can be a 2D image in color, black and white, or grayscale.

[0130] Next, the control device 10 can execute steps S31-S35 to move from the layout Figure 13 Noise is filtered out to obtain a training layout map with little or no noise.

[0131] It is worth mentioning that steps S31-S35 described later can be executed in their entirety or only partially, and there is no specific order of execution among them. Whether they are executed or not, and the order of execution, can be changed arbitrarily according to the user's needs.

[0132] Step S31: The control device 10 performs preprocessing on the obtained layout diagram 130 through the preprocessing module 502 to make the layout diagram 130 conform to the specified format or improve its recognizability.

[0133] Steps S32-S35 perform noise detection by analyzing the component profile.

[0134] Step S32: The control device 10 performs contour extraction processing on the layout drawing 130 through the contour extraction module 503 to identify the outermost contour of the target circuit board 20 in the layout drawing 130 to define the range of the target circuit board 20, and can further identify the contours of multiple components of the target circuit board 20.

[0135] In one embodiment, the aforementioned contour extraction process may include the Canny Algorithm, the Sobel Operator, or other edge detection algorithms, without limitation.

[0136] Step S33: The control device 10 performs a component filtering process on the layout diagram 130 through the component filtering module 504 to filter out the pattern of components with an area smaller than the component threshold from the layout diagram 130.

[0137] In one embodiment, the fragmentation threshold is determined based on the copper foil area of ​​the target circuit board 20 and a first ratio, such as multiplying the copper foil area by the first ratio (which may be greater than 1) as the fragmentation threshold.

[0138] Step S34: The control device 10 performs repeating element filtering processing on the layout diagram 130 through the repeating element processing module 505 to identify multiple repeating elements within a specified search range that exceed the repeating threshold, and filters out the pattern of multiple repeating elements from the layout diagram 130.

[0139] In one embodiment, the duplicate element filtering process includes expanding a specified search radius outward from each element as the search range.

[0140] In one embodiment, the aforementioned search radius is determined based on the side length of each element and a second ratio, such as multiplying the side length (e.g., the length of the longest side) by the second ratio (which may be greater than 1) to obtain the search radius.

[0141] Step S35: The control device 10 performs zigzag element filtering processing on the layout diagram 130 through the zigzag element filtering module 506 to identify elements (zigzag elements) with multiple corner points in the layout diagram 130 and filter out the pattern of the elements from the layout diagram 130.

[0142] In one embodiment, the zigzag element filtering process identifies elements with a number of corners greater than a corner threshold as zigzag elements.

[0143] In one embodiment, the zigzag element filtering process involves calculating the corner-to-contour ratio (i.e., the ratio of corners to all edges) between the number of corners and contours of an element, and identifying elements with a corner-to-contour ratio greater than a third ratio (which may be less than 1) as zigzag elements.

[0144] Please refer to the following: Figures 9 to 12 , Figure 9 This is a schematic outline diagram of a layout according to an embodiment of the present invention. Figure 10 This is a schematic diagram illustrating the processing result of a fragmented component filtering process according to an embodiment of the present invention. Figure 11 This is a schematic diagram illustrating the processing result of a repeating element filtering process according to an embodiment of the present invention. Figure 12 This is a schematic diagram of the processing result of the tortuous element filtering process according to an embodiment of the present invention.

[0145] like Figure 9As shown, the original unprocessed layout diagram 70 contains a large number of fragmented component patterns 80, which will become identification noise.

[0146] like Figure 10 As shown, the layout diagram 71 obtained after filtering out fragmented elements has filtered out the pattern 80 of fragmented elements, but there are a large number of pattern 81 of repeated elements. These pattern 81 of repeated elements will reduce the recognition accuracy.

[0147] like Figure 11 As shown, the layout diagram 72 obtained after the filtering of repetitive elements has filtered out the pattern 81 of repetitive elements, but there are a few patterns 82 of zigzag elements. These patterns 82 of zigzag elements will reduce the positioning accuracy.

[0148] like Figure 12 As shown, the layout diagram 73 obtained after the zigzag component filtering process has filtered out the zigzag component pattern 82 and has no other patterns that could become noise. Therefore, the layout diagram 73 can be directly used as a training layout diagram, and the image features 830-838 (including position and shape) in the layout diagram 73 can be used later to identify whether the target image belongs to the same type of circuit board (such as having the same component arrangement and circuit layout).

[0149] It is worth mentioning that since different types of circuit boards usually have very different component arrangements and circuit layouts, filtering the above components will not cause a decrease in recognition accuracy. On the contrary, because the recognition features are more obvious, it can improve recognition speed and accuracy.

[0150] Please refer to the following: Figure 9 and Figures 12-14 , Figure 13 This is a schematic diagram of a target image according to an embodiment of the present invention. Figure 14 This is a schematic diagram of another target image according to an embodiment of the present invention. For example... Figure 12 Through training, image features 830-838 of a standard circuit board can be obtained.

[0151] At Figure 13 In the example shown, after capturing the target image of the target circuit board, the target image can be processed, such as performing preprocessing or contour extraction processing to extract the contours of multiple components, to obtain the processed target image 74.

[0152] In one embodiment, simplification processing may also be performed on the target image 74 to reduce noise components.

[0153] Next, the control device 10 performs a matching process between the image features 740-748 of the target image 74 and the image features 830-838 of the standard circuit board, and finds a perfect match. Therefore, the control device 10 can determine that the target circuit board is... Figure 9 The standard circuit board shown.

[0154] At Figure 14 In the example, after capturing another target image of another target circuit board, the other target image can be processed, such as performing preprocessing, contour extraction processing and / or simplification processing, to obtain the processed other target image 75.

[0155] Next, the control device 10 can display the image features of another target image 75 (such as...) Figure 14 The control device 10 performs a matching process between the target circuit board (multiple squares) and the image features 830-838 of the standard circuit board, and finds any mismatches. Therefore, the control device 10 can determine that the target circuit board does not belong to the standard circuit board. Figure 9 The standard circuit board shown.

[0156] Please refer to the following: Figure 5 and Figure 7 , Figure 7 This is a flowchart of the training and recognition processes according to an embodiment of the present invention. Compared to... Figure 5 In this embodiment, the training process includes steps S40-S41.

[0157] Step S40: The control device 10 performs training edge analysis on the input training image (i.e., the training layout image corresponding to the standard circuit board) through the training module 507 to obtain the edge features of multiple training elements in the training image.

[0158] Step S41: The control device 10 calculates multiple recognition features of the standard circuit board based on the edge features of multiple training elements through the training module 507.

[0159] The aforementioned image edge computing feature recognition technology has many existing technologies in the field of image recognition processing. The present invention applies it to circuit board feature recognition.

[0160] Compared to Figure 5 In this embodiment, the identification process includes steps S50-S52.

[0161] Step S50: The control device 10 performs target edge analysis on the target image through the recognition module 508 to obtain the edge features of multiple target elements in the target image.

[0162] In one embodiment, the edge analysis in steps S40 and S50 may include the Canny Algorithm, the Sobel Operator, or other edge detection algorithms. The training edge analysis and the target edge analysis may use the same or different algorithms, without limitation.

[0163] Step S51: The control device 10 calculates the similarity score of the target image to each standard circuit board based on the edge features of the target element and multiple recognition features of each standard circuit board through the recognition module 508.

[0164] Step S52: The control device 10 determines the recognition result based on the highest score standard circuit board through the recognition module 508, and locates multiple components of the target image based on the corresponding similarity score.

[0165] In one embodiment, if the highest similarity score is lower than a preset score threshold, the control device 10 may determine that the identification result is all mismatches; if the highest similarity score is not lower than the preset score threshold, the control device 10 may set the identification result to the standard circuit board corresponding to the highest similarity score.

[0166] In one embodiment, when the identification result is all mismatches, since the standard circuit board with the highest score has the highest similarity to the target circuit board 20, the control device 10 can use the standard circuit board with the highest score to locate multiple components of the target image.

[0167] In one embodiment, during the training process, the control device 10 can perform training edge analysis on the training image (such as a training layout diagram) to obtain the positions of multiple training sampling edge points of the training image and the gradient direction and centroid of each training sampling edge point (i.e., training element edge features, which are image features), and calculate multiple training similarity scores (i.e., the recognition features of such a standard circuit board) based on these training sampling edge points and the preset edge detection range and allowable difference range.

[0168] For example, the control device 10 can expand the aforementioned edge detection range for each training sampled edge point, set the allowable difference range of the gradient direction angle, and compare it with the sampled edge points and gradient direction of a hypothetical real-time image to obtain the distance difference and gradient difference of each training sampled edge point, and use it as the aforementioned training similarity score.

[0169] Furthermore, in the recognition process, the control device 10 can sequentially select the recognition features of each standard circuit board to perform the following processing. The control device 10 can perform target analysis on the target image to obtain the position of multiple target sampling edge points and the gradient direction of each target sampling edge point (i.e., target element edge features, which are image features), calculate the predicted centroid corresponding to the target sampling edge points, and compare multiple training similarity scores with multiple centroid similarity scores corresponding to multiple predicted centroids (similar scores can be obtained by performing summation or averaging operations on multiple centroid similarity scores) to compare the similarity between the image features of the target image and the standard circuit board.

[0170] Furthermore, the control device 10 can use the predicted centroid with a score higher than the score threshold as the centroid of the corresponding training image, thereby achieving comparison, matching and localization of multiple image features between the target image and the training image.

[0171] In one embodiment, the control device 10 may filter out a portion of the target sampling edge points based on a plurality of target sampling edge points and their gradient directions (and may further refer to the allowable range of differences in the angle of the gradient directions), and calculate the predicted centroid of these target sampling edge points.

[0172] The above description is merely a preferred embodiment of the present invention and is not intended to limit the claims of the present invention. Therefore, all equivalent variations made using the content of the present invention are similarly included within the scope of the present invention and are hereby stated.

Claims

1. An automatic identification method for the gripping position of a circuit board, comprising: a) Obtain multiple layout diagrams of multiple standard circuit boards of different types, and perform simplification processing on the layout diagrams of each of the standard circuit boards to obtain a training layout diagram that removes identification noise, wherein the layout diagram is a 2D image; b) Based on each of the training layout diagrams, the training model is trained to enable the training model to recognize different types of the plurality of standard circuit boards, wherein each of the training layout diagrams is configured with a gripping position; c) Input the target image of the target circuit board into the training model to perform recognition processing on the target image; and d) When the target image is identified as conforming to any of the standard circuit boards, a matching process is performed on the target image to determine the grab position of the target image based on the grab position of the training layout diagram of the conforming standard circuit board; The identification process includes: e1) Perform target edge analysis on the target image to obtain edge features of multiple target elements in the target image; e2) Calculate the similarity score of the target image to each of the standard circuit boards based on the edge features of the multiple target elements and the multiple recognition features of each of the standard circuit boards; and e3) The recognition result is determined based on the standard circuit board with the highest score; The simplification process includes step f1) performing contour extraction processing on the layout diagram to identify the contours of multiple components of the target circuit board in the layout diagram; the simplification process also includes at least one of the following steps: f2) Perform a fragmented element filtering process on the layout diagram to filter out elements with an area smaller than the fragmented threshold from the layout diagram; f3) Perform a repeating element filtering process on the layout diagram to identify repeating elements within the search range that exceed a repeating threshold, and filter out the repeating elements from the layout diagram; and f4) Perform zigzag element filtering on the layout diagram to identify elements with multiple corner points and filter out the elements from the layout diagram.

2. The automatic identification method for the gripping position of a circuit board according to claim 1, wherein the fragmented threshold is determined based on the copper foil area of ​​the target circuit board and a first ratio; in, The repeated element filtering process includes expanding the search radius outward from each of the elements as the search range, and the search radius is determined based on the side length of each of the elements and a second ratio; The zigzag element filtering process refers to the element whose number of corner points is greater than the corner point threshold or whose corner point-to-contour point ratio is greater than a third ratio between the number of corner points and the number of contour points.

3. The automatic identification method for circuit board gripping position according to claim 1, wherein the training process includes: g1) Perform training edge analysis on the training layout graph to obtain edge features of multiple training elements of the training layout graph; and g2) Calculate the multiple recognition features of the standard circuit board based on the edge features of the multiple training elements.

4. The automatic identification method for the circuit board gripping position according to claim 3, wherein the training process further includes: h1) Perform the training edge analysis to obtain the position, gradient direction and centroid of multiple training sampling edge points of the training layout map, as edge features of the multiple training elements; and h2) Calculate multiple training similarity scores based on the multiple training sampling edge points, edge detection range, and allowable difference range, as the multiple recognition features.

5. The automatic identification method for the circuit board gripping position according to claim 1, wherein the identification process further includes: e4) Locate multiple elements of the target image using the highest similarity score of the standard circuit board.

6. The automatic identification method for the circuit board gripping position according to claim 5, wherein the identification process further includes: i1) Perform the target edge analysis to obtain multiple target sampling edge points of the target image and the gradient direction of each target sampling edge point, as edge features of the multiple target elements; and i2) Calculate the predicted centroid of each of the target sampling edge points based on the position and gradient direction of each target sampling edge point, compare multiple training similarity scores with multiple centroid similarity scores corresponding to the multiple predicted centroids, and determine the similarity score of the standard circuit board based on the multiple centroid similarity scores.

7. An automatic identification system for the gripping position of a circuit board, comprising: An image acquisition device used to photograph a target circuit board to obtain a target image; A gripping device used to grip circuit boards; and A control device, connected to the image acquisition device and the grasping device, includes: A simplification module is configured to perform simplification processing on a standard circuit board layout to obtain a training layout by removing identification noise, the layout being a 2D image; A training module is configured to input the training layout diagram into a training model to enable the training model to recognize the same type of circuit board, wherein the training layout diagram has a gripping position set on it. The recognition module inputs the target image into the training model to perform recognition. When the target image is recognized as the standard circuit board, the module performs matching processing on the target image to determine the grasping position of the target image based on the grasping position of the training layout diagram of the standard circuit board. The control device is configured to control the gripping device to grip the target circuit board based on the gripping position of the target image; The control device further includes: A contour extraction module is configured to identify the contours of multiple components of the target circuit board in the layout drawing; and A fragmented component filtering module is configured to filter out components with an area smaller than a fragmented threshold from the layout diagram, wherein the fragmented threshold is determined based on the copper foil area of ​​the target circuit board and a first ratio; The control device further includes: A contour extraction module is configured to identify the contours of multiple components of the target circuit board in the layout drawing; and A repeating element filtering module is configured to identify repeating elements in the layout diagram whose number exceeds a repeating threshold within the search range, and to filter out repeating elements from the layout diagram. The control device further includes: A contour extraction module is configured to identify the contours of multiple components of the target circuit board in the layout drawing; and A tortuous element filtering module is configured to identify elements with multiple corner points in a layout diagram and filter out the elements from the layout diagram.

8. The automatic identification system for the gripping position of a circuit board according to claim 7, wherein the duplicate element filtering module is further configured to expand the search radius outward from each of the elements as the search range, the search radius being determined based on the side length of each of the elements and a second ratio.

9. The automatic identification system for the gripping position of a circuit board according to claim 7, wherein the tortuous element filtering module is further configured to filter elements whose number of corner points is greater than a corner point threshold or whose corner point-to-contour point ratio between the number of the plurality of corner points and the number of the plurality of contour points is greater than a third ratio.

10. The automatic identification system for the gripping position of a circuit board according to claim 7, wherein the training module is further configured to perform training edge analysis on the training layout diagram to obtain edge features of multiple training elements of the training layout diagram, and to calculate multiple identification features based on the edge features of the multiple training elements.

11. The automatic identification system for the gripping position of a circuit board according to claim 7, wherein the identification module is further configured to perform edge analysis on the target image to obtain edge features of multiple target elements of the target image, calculate a similarity score of the target image to each of the standard circuit boards based on the edge features of the multiple target elements and multiple identification features of each of the standard circuit boards, determine the identification result based on the standard circuit board with the highest score, and locate the multiple elements of the target image using the corresponding similarity score.

Images