A printed circuit board defect detection system and method based on image recognition

Abstract

The present invention discloses a printed circuit board defect detection method and system based on image recognition, which belongs to the field of PCB detection. The method collects image data of a standard board through an image acquisition unit, and compares the image data of the board to be tested with the standard board as a benchmark, and then judges whether the board to be tested is qualified and the specific defect cause of the unqualified board to be tested through the difference in image data, replacing the traditional manual detection method. When used, there is no need to import a standard base-level image, which greatly improves the detection efficiency of PCB. The printed circuit board defect detection method and system based on image recognition of the present invention have a simple overall structure and are easy to operate. They can not only realize fast and accurate detection of PCBs, but also realize fast detection and input of standard base-level PCB images. At the same time, the overall size of the device is small and easy to carry, and has good practical value and application prospects.

Description

Technical Field

[0001] The invention belongs to the field of PCB detection, and in particular relates to a printed circuit board defect detection method and system based on image recognition. Background Art

[0002] PCB, or printed circuit board, is one of the important components of the electronics industry. Almost every electronic device, from electronic watches to calculators, uses printed circuit boards. In the production process of PCB, its quality inspection is a key step in PCB production and the final checkpoint for controlling the qualified rate of PCB finished products.

[0003] The qualified rate of PCB products has always been one of the important factors restricting the profit growth of the electronic foundry industry. However, with the combination of AOI technology and PCB inspection process, the traditional problems of difficulty and poor efficiency of manual inspection have been effectively solved. It is currently one of the most widely used and most promising PCB defect detection methods, solving the problem of surface welding defect detection under component-intensive conditions.

[0004] However, with the diversification of PCB types, the types of defects in mounted components have become more complex, and the types of mounting defects that the AOI inspection system can detect are limited and need to be continuously improved. At the same time, as the density of circuits continues to increase, missed detection and false detection are also inevitable problems of this technology. Nowadays, some defect measuring instruments on the market are bulky and not conducive to handling and measurement. At the same time, when using it, it is necessary to re-import the standard base-level image and then compare it with the board to be tested, which is very troublesome. At the same time, only a single type of PCB can be inspected during inspection, and the types of defects detected are also relatively single, and the inspection efficiency is low, which cannot meet the needs of PCB production and inspection in the existing technology. Summary of the invention

[0005] In response to one or more of the above defects or improvement needs in the prior art, the present invention provides a printed circuit board defect detection method and system based on image recognition, which utilizes an image recognition detection system to simultaneously collect image information of the board to be tested and the standard board, and compares the two, thereby quickly and accurately collecting defects on the board to be tested.

[0006] To achieve the above object, the present invention provides a printed circuit board defect detection method based on image recognition, comprising the following steps:

[0007] S101: using an image acquisition unit to respectively acquire two-dimensional image information of a standard board and a board to be tested;

[0008] S200: transmitting the two-dimensional image information of the two to a data processing unit for processing, and obtaining image data of the board to be tested and image data of the standard board;

[0009] S300: Using the image data of the standard board as a reference, the two-dimensional image data of the board to be tested is compared with the standard board, and the defects of the board to be tested are determined by using the data difference after the comparison.

[0010] As a further improvement of the present invention, in step S200, the process of processing the two-dimensional image information further includes the following steps:

[0011] According to the complexity of the two-dimensional image information, the two-dimensional image information is segmented, and each segmented image is analyzed;

[0012] Binarize the image of each component on the board and obtain the binary image of the component;

[0013] Perform grayscale processing on the two-dimensional image information and draw a grayscale histogram based on the grayscale values.

[0014] As a further improvement of the present invention, in step S300, the comparison process of the two-dimensional image data is as follows:

[0015] Obtaining the position parameters and image parameters of the board according to the binary image of the board;

[0016] Compare the image parameters and position parameters of the standard board and the board to be tested, obtain a correlation coefficient r between the two, and set a minimum threshold T1 of the correlation coefficient;

[0017] If the correlation coefficient r between the two is greater than the minimum threshold T1, it is directly judged as a qualified sample; otherwise, it is judged as a defective sample.

[0018] As a further improvement of the present invention, after determining that the sample to be tested is a sample with defects, the defect type is further determined, and the determination process includes the following steps:

[0019] S301: Set the maximum threshold value T2 of the mutual correlation coefficient. If r is less than T2, go to step S302; otherwise, go to step S303:

[0020] S302: According to the binary image of the component area, respectively obtain the proportion R of the white foreground in the board to be tested and the standard board to the whole image; and set the minimum threshold of the proportion. If the difference between the board to be tested and the standard board is greater than the minimum threshold, it is determined that the defect of the board to be tested is a lack of soldering. Otherwise, it goes to step S203;

[0021] S303: Obtain the total areas SP and SD of the connected areas on the board to be tested and the standard board, as well as the numbers of connections NP and SP;

[0022] S304: If NP is less than ND and SP is greater than SD, it is determined that the defect of the board to be tested is a bridge; otherwise, it proceeds to step S305;

[0023] S305: Obtain the deviations ax and ay of the centroid of the component; if ax or ay is greater than 0.25, it can be judged as skewed, otherwise it goes to S206;

[0024] S306: Obtaining GP and GP of the board to be tested and the standard board according to the grayscale histogram, as well as grayscale proportions RH and TH of the two;

[0025] S307: If GP is less than GD and RH is greater than TH, it is determined that the defect of the board to be tested is insufficient tin; otherwise, it proceeds to step S308;

[0026] S308: Setting a minimum threshold of the mutual correlation coefficient, determining whether the mutual correlation coefficients of the remaining samples to be tested are less than the minimum threshold, and if so, determining that the samples are other defects.

[0027] On this basis, the present invention also proposes a printed circuit board defect detection system based on image recognition, which includes a box;

[0028] The box body is provided with at least two workstations, one of which is used to place a standard board, and at least one of which is used to place a board to be tested;

[0029] At least one image acquisition unit and a data processing unit are arranged in the box; the acquisition direction of at least one of the image acquisition units is arranged toward at least two of the workstations on the box, and is used to respectively acquire two-dimensional image information of the standard board and the board to be tested; and

[0030] The data processing unit is electrically connected to at least one image acquisition unit, and is used to transmit the two-dimensional image information of the standard board and the board to be tested to the data processing unit, and to segment, binarize and grayscale the two-dimensional image information of the standard board and the board to be tested in the data processing unit, and then compare the processed two-dimensional image data to obtain the data difference between the board to be tested and the standard board, so as to determine the defects of the board to be tested.

[0031] As a further improvement of the present invention, there are multiple image acquisition units and multiple workstations; and

[0032] The acquisition directions of the multiple image acquisition units correspond to the multiple workstation settings respectively, and are used to correspond to the two-dimensional image information of the boards on the workstations respectively.

[0033] As a further improvement of the present invention, the image acquisition unit has one;

[0034] An adjustment mechanism is provided corresponding to the image acquisition unit, which is used to adjust the acquisition direction of the image acquisition unit, thereby acquiring two-dimensional image information of boards and cards on different workstations.

[0035] As a further improvement of the present invention, a lighting assembly is provided corresponding to the image acquisition unit, and its lighting direction is arranged toward the workstation.

[0036] As a further improvement of the present invention, a storage module is further provided in the box, which is electrically connected to the data processing module and is used to receive and store the two-dimensional image data of the standard board and the board to be tested.

[0037] As a further improvement of the present invention, a data transceiver unit is further provided in the box body for receiving the two-dimensional image data processed by the data processing unit.

[0038] The above-mentioned improved technical features can be combined with each other as long as they do not conflict with each other.

[0039] In general, compared with the prior art, the above technical solution conceived by the present invention has the following beneficial effects:

[0040] (1) The printed circuit board defect detection method and system based on image recognition of the present invention collects image data of a standard board through an image acquisition unit, and compares the image data of the board to be tested with the standard board as a reference, and then judges whether the board to be tested is qualified and the specific defect cause of the unqualified board to be tested based on the difference in image data, replacing the traditional manual detection method. When in use, there is no need to import a standard base-level image, which greatly improves the detection efficiency of PCBs.

[0041] (2) The printed circuit board defect detection method and system based on image recognition of the present invention utilizes the optimization of the graphic algorithm and cooperates with the image processing technology to enable the device to realize the rapid and accurate detection of the board to be tested, and each unit is contained in the box, so that the volume of the device is smaller and convenient for carrying and measuring operations. At the same time, the device can also be carried on the production line for use, which has good practicality.

[0042] (3) The printed circuit board defect detection method and system based on image recognition of the present invention has a simple overall structure and is easy to operate. It can not only realize fast and accurate detection of PCBs, but also realize fast detection and input of standard base-level PCB images. At the same time, the overall size of the device is small and easy to carry, and it has good practical value and application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 Schematic diagram of the overall structure of a printed circuit board defect detection system and method thereof in an embodiment of the present invention;

[0044] Figure 2 1 is an overall isometric diagram of a printed circuit board defect detection system and method thereof according to an embodiment of the present invention;

[0045] Figure 3 It is a logic block diagram of a printed circuit board defect detection method in an embodiment of the present invention;

[0046] Figure 4 It is a logic block diagram of defect judgment in the printed circuit board defect detection method in an embodiment of the present invention;

[0047] In all the drawings, the same reference numerals represent the same technical features, specifically: 100, housing; 101, work station; 103, operation panel; 104, power switch; 105, heat dissipation structure; 106, image acquisition unit; 107, lighting unit; 108, data processing unit. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0049] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0050] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0051] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements, unless otherwise clearly defined. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0052] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may mean that the first and second features are in direct contact, or the first and second features are in indirect contact through an intermediate medium. Moreover, a first feature being "above", "above" or "above" a second feature may mean that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is higher in level than the second feature. A first feature being "below", "below" or "below" a second feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is lower in level than the second feature.

[0053] Example:

[0054] See also Figure 3 and Figure 4 The printed circuit board defect detection method of the present invention uses image acquisition to collect two-dimensional image data of the standard board and the board to be tested in real time, and compares the two-dimensional image data of the two to obtain the difference between the image data of the board to be tested and the standard board, and judges the defects of the board to be tested by the data difference, so as to achieve fast, accurate and diverse PCB detection.

[0055] Specifically, the method comprises the following steps:

[0056] S100: Put the standard board and the board to be tested on the two workstations 101, and use the image acquisition unit 106 to respectively acquire two-dimensional image information of the standard board and the board to be tested.

[0057] More specifically, the process of placing the board to be tested and the standard board also includes the following steps:

[0058] S101: Place the standard board on station 101.

[0059] S102: Align the image acquisition unit 106 with the standard board, and turn on the illumination component and the light source acquisition unit.

[0060] S103: The image acquisition unit 106 acquires the two-dimensional image information of the standard board.

[0061] S104: Place the board to be tested and repeat the above operation.

[0062] S200: Transmitting the two-dimensional image information of both to the data processing unit 108 for processing, and obtaining the image data of the board to be tested and the image data of the standard board.

[0063] In actual use, for example, when testing boards of the same series, the image information of the standard board is first collected, processed, and then stored in the storage unit. Then, the image data of the board to be tested is collected, processed, and compared.

[0064] S300: Using the image data of the standard board as a reference, the two-dimensional image data of the board to be tested is compared with the standard board, and the defects of the board to be tested are determined according to the data difference after the comparison.

[0065] S301: Set the maximum threshold value T2 of the mutual correlation coefficient. If r is less than T2, go to step 302; otherwise, go to step S303:

[0066] S302: According to the binary image of the component area, respectively obtain the proportion R of the white foreground in the board to be tested and the standard board to the whole image; and set a minimum threshold of the proportion. If the difference between the board to be tested and the standard board is greater than the minimum threshold, it is determined that the defect of the board to be tested is a lack of soldering. Otherwise, it goes to step S303;

[0067] S303: Obtain the total areas SP and SD of the connected areas on the board to be tested and the standard board, as well as the numbers of connections NP and SP;

[0068] S304: If NP is less than ND and SP is greater than SD, it is determined that the defect of the board to be tested is a bridge; otherwise, it proceeds to step S305;

[0069] S305: Obtain the deviations ax and ay of the centroid of the component; if ax or ay is greater than 0.25, it can be judged as skewed, otherwise it goes to S306;

[0070] S306: Obtaining GP and GP of the board to be tested and the standard board according to the grayscale histogram, as well as grayscale proportions RH and TH of the two;

[0071] S307: If GP is less than GD and RH is greater than TH, it is determined that the defect of the board to be tested is insufficient tin; otherwise, it proceeds to step S308;

[0072] S308: Setting a maximum threshold of the mutual correlation coefficient, determining whether the mutual correlation coefficients of the remaining samples to be tested are less than the maximum threshold, and if so, determining that the samples are other defects.

[0073] In addition, the present invention also provides a printed circuit board defect detection system based on image recognition, please refer to Figure 1 and Figure 2 The image acquisition unit 106 is used to respectively acquire the two-dimensional image information of the standard board and the board to be tested, and the processor is used to process the two-dimensional image information of the board to be tested and the standard board, thereby obtaining the detection indicators of the two boards, and comparing the detection indicators of the board to be tested with the standard board based on the detection of the standard board, so as to determine the defects of the board to be tested, and realize fast and accurate PCB detection operation.

[0074] Specifically, the defect detection system in the preferred embodiment of the present invention includes a box 100, on which at least two workstations 101 are arranged, one of which is used to place a standard board, and at least one is used to place a board to be tested; a data processing unit 108 and at least one image acquisition unit 106 are arranged in the box 100, and the acquisition direction of the image acquisition unit 106 is set toward the workstation 101. The image acquisition unit 106 is electrically connected to the data processing unit 108, and is used to respectively collect two-dimensional image information of the standard board and the board to be tested on the two workstations 101, and transmit it to the data processing unit 108 for processing, and after processing, the data of the standard board is used as a reference to compare it with the data of the board to be tested, so as to obtain the difference data between the two, and finally judge whether the board to be tested has defects according to the difference data.

[0075] It can be understood that the above-mentioned defects may be information such as the number of defects, the number of bridges, the number of misalignments, and the defect ratio.

[0076] In more detail, when processing the two-dimensional image information, the data processing unit 108 first segments the image according to the complexity of the board, and then analyzes each image separately, then performs image binarization processing on the components on the board, and finally calculates the relevant position parameters and image parameters of the components.

[0077] In the actual processing process, the standard board is first placed in the workstation 101 for testing. After obtaining various data of the standard parts, the board to be tested is placed in other workstations 101, its image is captured, and its image data is obtained. The data of the board to be tested is then compared with the data of the standard board to obtain the defect indicators of the board to be tested.

[0078] Further preferably, a storage unit is provided in the housing 100 , which is electrically connected to the data processing unit 108 and is used to receive and store various data of the standard board and use the data as a standard to compare with other boards of the same type to be tested.

[0079] More specifically, after obtaining the relevant position parameters and image parameters of the image data, the mutual correlation coefficient r between the component acquisition images of the board to be tested and the standard board is obtained, |r|∈[0,1]. The larger the mutual correlation coefficient, the higher the matching degree between the image data of the standard board and the image data of the board to be tested. A minimum threshold T1 is set. When |r|>T1, the board to be tested is a qualified sample, otherwise it is an unqualified sample.

[0080] In a specific embodiment, the minimum threshold value T1 of the image correlation coefficient is set to 0.7. When the detected sample image correlation coefficient r is greater than 0.7, it is a qualified sample, and when it is less than 0.7, it is an unqualified sample.

[0081] Furthermore, if the sample is an unqualified sample, the specific defects of the sample are determined:

[0082] First, determine a maximum threshold T2, T2 is 0.4. Determine whether the board image correlation coefficient r is less than T2, and obtain the proportion of white foreground in the binary image of the component area of ​​the standard board and the board to be tested, which are recorded as and respectively, and determine whether the difference between the two is greater than the minimum threshold of the proportion; if the above two conditions are met at the same time, the board to be tested is judged to be defective.

[0083] Furthermore, in the samples that do not meet the above two conditions, the number NP and the total area SP of the connected areas on the sample to be tested are judged with the number ND and the total area SD of the connected areas on the standard board (the connected area refers to the number of connected welding points on the pad); if the number of connected areas NP of the board to be tested is less than the number of connected areas ND of the standard board, and the total area SP of the connected areas of the board to be tested is greater than the total area SD of the connected areas of the standard board, it can be judged as a bridge;

[0084] When judging the remaining samples, judge the deviation ax and ay of the centroid of the sample; if ax or ay is greater than 0.25, it can be judged as skewed; finally, after the image is grayscale processed, a grayscale histogram is drawn according to the distribution of grayscale values. After many experiments, combined with the actual situation, the grayscale mean credible range of the standard welding point is 85-135. If the grayscale mean GP of the board to be tested is smaller than the grayscale mean GD of the standard board, the proportion of the grayscale value of the board to be tested within the credible range and the proportion RH of the grayscale mean of the board to be tested within ±3α is greater than the proportion threshold TH of the standard board is 0.65, it can be judged as insufficient tin.

[0085] In addition, if the image correlation coefficient is less than 0.1, it is other defects, and the remaining samples are qualified samples.

[0086] It can be understood that how to obtain relevant position parameters, image parameters, centroid acceleration, grayscale histogram and other parameters can be directly calculated by those skilled in the art according to the prior art, and will not be elaborated here.

[0087] Furthermore, in the actual setting process, the number of workstations 101 is preferably two. Of course, multiple workstations 101 can also be added according to actual use requirements to place the boards to be tested. At the same time, it is preferred that an image acquisition unit 106 is provided corresponding to each workstation 101, so as to simultaneously acquire image data of multiple boards to be tested, and detect defects of the boards to be tested according to the above processing method.

[0088] In the actual setting process, multiple workstations 101 can be set up to place multiple boards to be tested. Only one workstation 101 needs to be reserved to place a standard board, and multiple image acquisition units 106 are set corresponding to the multiple workstations 101, so that the device can detect multiple boards to be tested at the same time, thereby greatly improving the detection efficiency.

[0089] Furthermore, a lighting unit 107 is provided corresponding to each image acquisition unit 106 , and its lighting direction is arranged toward the workstation 101 , so that the image acquisition unit 106 can better acquire the two-dimensional image data of the board to be tested or the standard board.

[0090] Furthermore, a display screen and an operation panel 103 electrically connected thereto are also provided on the box 100. The operation panel 103 is used to control the entire device to be turned on or off. The display screen is electrically connected to a data processing unit 108, and is used to receive and display the two-dimensional image data of the board to be tested. At the same time, a heat dissipation structure 105 is also provided on the box 100, and is used to quickly dissipate heat for various components provided inside the box 100. Corresponding to the power supply in the box 100, a power switch 104 is provided on the surface of the box 100, and is used to control the entire device to be turned on or off.

[0091] In addition, the defect detection system in the preferred embodiment of the present invention further includes a data transceiver unit, which is electrically connected to the data processing unit 108 and is used to transmit the processed board data to the terminal device.

[0092] The printed circuit board defect detection method and system based on image recognition of the present invention has a simple overall structure and is easy to operate. It can not only realize fast and accurate detection of PCBs, but also realize fast detection and entry of standard base-level PCB images. At the same time, the overall size of the device is small and easy to carry, and it has good practical value and application prospects.

[0093] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A printed circuit board defect detection method based on image recognition, characterized in that: The following steps are involved: S101: using an image acquisition unit to respectively acquire two-dimensional image information of a standard board and a board to be tested; S200: transmitting the two-dimensional image information of the two to a data processing unit for processing, and obtaining the image data of the board to be tested and the image data of the standard board; wherein the processing of the two-dimensional image information includes the following steps: According to the complexity of the two-dimensional image information, the two-dimensional image information is segmented, and each segmented image is analyzed; Binarize the image of each component on the board and obtain the binary image of the component; Perform grayscale processing on the two-dimensional image information and draw a grayscale histogram according to the grayscale value; S300: Using the image data of the standard board as a reference, the two-dimensional image data of the board to be tested is compared with the image data of the standard board, and the defects of the board to be tested are determined by using the data difference after the comparison; wherein the comparison process of the two-dimensional image data is as follows: Obtaining the position parameters and image parameters of the board according to the binary image of the board; Compare the image parameters and position parameters of the standard board and the board to be tested, obtain a correlation coefficient r between the two, and set a minimum threshold T1 of the correlation coefficient; If the correlation coefficient r between the two is greater than the minimum threshold T1, it is directly judged as a qualified sample; otherwise, it is judged as a defective sample; After determining that the sample to be tested is a sample with defects, further determination is made on its defect type, and the determination process includes the following steps: S301: Set the maximum threshold value T2 of the mutual correlation coefficient. If r is less than T2, go to step S302; otherwise, go to step S303: S302: According to the binary image of the component area, respectively obtain the proportion R of the white foreground in the board to be tested and the standard board to the whole image; and set the minimum threshold of the proportion. If the difference between the board to be tested and the standard board is greater than the minimum threshold, it is determined that the defect of the board to be tested is a lack of soldering. Otherwise, it goes to step S203; S303: Obtain the total areas SP and SD of the connected areas on the board to be tested and the standard board, as well as the numbers of connections NP and SP; S304: If NP is less than ND and SP is greater than SD, it is determined that the defect of the board to be tested is a bridge; otherwise, it proceeds to step S305; S305: Obtain the deviations ax and ay of the centroid of the component; if ax or ay is greater than 0.25, it can be judged as skewed, otherwise it goes to S206; S306: Obtaining GP and GP of the board to be tested and the standard board according to the grayscale histogram, as well as grayscale proportions RH and TH of the two; S307: If GP is less than GD and RH is greater than TH, it is determined that the defect of the board to be tested is insufficient tin; otherwise, it proceeds to step S308; S308: Setting a minimum threshold of the mutual correlation coefficient, determining whether the mutual correlation coefficients of the remaining samples to be tested are less than the minimum threshold, and if so, determining that the samples are other defects.

2. A printed circuit board defect detection system based on image recognition, characterized in that: It uses the printed circuit board defect detection method based on image recognition as described in claim 1 to implement the detection operation, and includes a box; The box body is provided with at least two workstations, one of which is used to place a standard board, and at least one of which is used to place a board to be tested; At least one image acquisition unit and a data processing unit are arranged in the box; the acquisition direction of at least one of the image acquisition units is arranged toward at least two of the workstations on the box, and is used to respectively acquire two-dimensional image information of the standard board and the board to be tested; and The data processing unit is electrically connected to at least one image acquisition unit, and is used to transmit the two-dimensional image information of the standard board and the board to be tested to the data processing unit, and to segment, binarize and grayscale the two-dimensional image information of the standard board and the board to be tested in the data processing unit, and then compare the processed two-dimensional image data to obtain the data difference between the board to be tested and the standard board, so as to determine the defects of the board to be tested.

3. The printed circuit board defect detection system based on image recognition according to claim 2 is characterized in that: There are multiple image acquisition units and multiple workstations; and The acquisition directions of the multiple image acquisition units correspond to the multiple workstation settings respectively, and are used to correspond to the two-dimensional image information of the boards on the workstations respectively.

4. The printed circuit board defect detection system based on image recognition according to claim 2 is characterized in that: The image acquisition unit has one; An adjustment mechanism is provided corresponding to the image acquisition unit, which is used to adjust the acquisition direction of the image acquisition unit, thereby acquiring two-dimensional image information of boards and cards on different workstations.

5. The printed circuit board defect detection system based on image recognition according to any one of claims 2 to 4, characterized in that: Corresponding to the image acquisition unit, an illumination component is arranged, and its illumination direction is arranged toward the workstation.

6. The printed circuit board defect detection system based on image recognition according to claim 5 is characterized in that: The box body is also provided with a storage module which is electrically connected to the data processing module and is used for receiving and storing the two-dimensional image data of the standard board and the board to be tested.

7. The printed circuit board defect detection system based on image recognition according to claim 6 is characterized in that: The box is also provided with a data transceiver unit for receiving the two-dimensional image data processed by the data processing unit.

Images