An automated visual inspection device

Abstract

The present invention relates to the field of transformer detection technology, and specifically to an automated visual inspection device, which includes a processing module, a conveying mechanism, a visual inspection mechanism located around the conveying mechanism, a transfer mechanism, and a temporary storage mechanism; the conveying mechanism and the temporary storage mechanism are both used to convey products to be inspected, the visual inspection mechanism is used to take pictures of the products on the conveying mechanism to obtain image data of the products, the processing module is used to analyze and process the image data of the visual inspection mechanism, and the transfer mechanism is used to transfer the products between the conveying mechanism and the temporary storage mechanism. The purpose of the present invention is to provide an automated visual inspection device, which can effectively reduce misjudgments, realize large-scale product recognition, and improve product inspection efficiency by performing six-sided inspection on the product and performing secondary inspection on the product.

Description

Technical Field

[0001] The present invention relates to the technical field of transformer detection, and in particular to an automatic visual detection device. Background Art

[0002] As the application of fast charger technology matures, the automated production requirements for thin, compact and standardized electronic transformer products have emerged. Under efficient automated production conditions, the production capacity of consumer electronics and home appliance electronic transformer products has increased rapidly, and the workload of external inspection has increased significantly.

[0003] After the production of this type of transformer, the pins need to be inspected for quality before they can be shipped. Traditional visual inspection uses industrial cameras and other visual inspection equipment to take pictures of the product features and analyze the image data in the photos to judge the quality of the product. Conventional manual video inspection takes a long time, the number of product moves increases, visual fatigue is inevitable, and there is no effective way to prevent the appearance of poor products. In addition, the current practice is to take pictures of each product in multiple directions and then analyze the image data. It is impossible to identify a small area by zooming in or comparing twice. Due to the limited processing speed of the processing module and the limited mechanical handling action, this detection method is inefficient and has gradually failed to meet the increasing production needs. Summary of the invention

[0004] In order to overcome the shortcomings and deficiencies in the prior art, the purpose of the present invention is to provide an automated visual inspection equipment that performs six-sided inspection on the product while also performing secondary inspection on the product, thereby effectively reducing misjudgments, realizing large-scale product identification, and improving product inspection efficiency.

[0005] The present invention is achieved through the following technical solutions:

[0006] An automated visual inspection device comprises a processing module, a conveying mechanism, a visual inspection mechanism located around the conveying mechanism, a transfer mechanism and a temporary storage mechanism;

[0007] The conveying mechanism and the temporary storage mechanism are both used to convey the products to be inspected, the visual inspection mechanism is used to take pictures of the products on the conveying mechanism to obtain image data of the products, the processing module is used to analyze and process the image data of the visual inspection mechanism, and the transfer mechanism is used to transfer the products between the conveying mechanism and the temporary storage mechanism.

[0008] In combination with the first aspect, further, a portion of the conveying mechanism close to the visual inspection mechanism is a spiral structure, and the conveying mechanism behind the spiral structure is hollowed out so that the visual inspection mechanism can photograph the bottom of the product to be inspected when the product to be inspected passes by.

[0009] In combination with the first aspect, further, the temporary storage mechanism includes a temporary storage area and a temporary storage track whose conveying direction is opposite to that of the conveying mechanism.

[0010] On the other hand, the present invention also discloses an automated visual inspection method, which comprises the following steps:

[0011] S100. Place the product to be inspected into the conveying mechanism, and the visual inspection mechanism obtains image data of the product to be inspected in six directions: top, bottom, front and back, and both sides;

[0012] S200. The processing module analyzes and processes the image data obtained in step S100, marks the suspected defect area, and then takes out the product by the transfer mechanism and puts it into the temporary storage mechanism after marking it;

[0013] S300. When a certain number of products are placed in the temporary storage mechanism, the products are recorded as a unit, and the transfer mechanism takes out the products of the unit and puts them back into the conveying mechanism;

[0014] S400. The processing module controls the visual inspection mechanism to obtain image data of the marked area corresponding to the product according to the product of the unit according to the marking order or marking value of the product in the unit;

[0015] S500. The processing module analyzes and processes the image data acquired in step S400, determines whether the image data is a good product or a bad product, and removes the bad products.

[0016] In combination with the second aspect, further, in step S100, the number of products in a unit is single or multiple.

[0017] In combination with the second aspect, further, in step S100, after the visual inspection mechanism obtains image data in five directions of the top, front and back, and both sides of the product to be inspected, the visual inspection mechanism obtains image data of the bottom of the product to be inspected after the product to be inspected passes through the spiral structure of the conveying mechanism.

[0018] In combination with the second aspect, further, in step S200, the marked products are placed in a temporary storage area, and when a certain number of products are placed in the temporary storage area, the products are taken out and placed in a temporary storage track, and after the products reach the end of the temporary storage track, they are placed in the head end of the conveying mechanism.

[0019] In combination with the second aspect, further, in step S200, the steps of analyzing and processing the image data are as follows:

[0020] S201. Perform edge detection on the image data and calculate the gradient amplitude of the image. The gradient calculation formula is:

[0021]

[0022] The calculation formula of gradient amplitude is:

[0023]

[0024] Among them, G x and G y are the gradients of the image in the x and y directions respectively;

[0025] S202. Binarize the gradient amplitude image and set a threshold T;

[0026] S203. Use morphological operations to remove noise and fill defect areas;

[0027] In step S300, the defective area is marked using the connected component marking method, and an information matrix M of the marking information is established.

[0028] In combination with the second aspect, further, in step S400, the processing module obtains the marking sequence or marking value of the product by reading the information matrix M, where M i =(x i ,y i ,w i ,h i ) represents the marking information of the i-th product, (x i ,y i ) is the coordinate of the upper left corner of the defect area, (w i ,h i ) are the width and height of the defect area.

[0029] In combination with the second aspect, further, in step S500, a convolutional neural network is used to classify the product to determine whether the product is a good product or a defective product, and the classification step includes:

[0030] S501. Input image L ′ i,j Perform convolution operation. The calculation formula of convolution operation is: C 1 =ReLU(Conv(L ′ i,j ,W 1 )+b 1 ), where C 1 is the convolution result, Conv represents the convolution operation, W 1 is the convolution kernel, b 1 is the bias, ReLU is the activation function;

[0031] S502. Perform a pooling operation on the output of the convolutional layer. The calculation formula for the pooling operation is: 1 =MaxPool(C 1 ), where P 1It is the pooling result, and MaxPool represents the pooling operation;

[0032] S503. Expand the output of the pooling layer and pass it through the fully connected layer. The calculation formula for expanding the pooling layer is: F = ReLU (W 2 ·flatten(P 1 )+b 2 ), where F is the expansion result, W 2 and b 2 are the weights and biases of the fully connected layer, and flatten represents the expansion operation;

[0033] S504. The classification result is obtained through the Softmax layer. The calculation formula of the Softmax layer is: i =Softmax(W 3 ·F+b 3 ), where W 3 and b 3 are the weights and biases of the output layer, and Softmax represents the normalized exponential function.

[0034] In combination with the second aspect, further, in step S400, the processing module controls the industrial camera to zoom to obtain the marked product and obtain image data of the marked area corresponding to the product.

[0035] In combination with the second aspect, further, the automated visual inspection equipment also includes step S600: the processing module uses the AI ​​big model to analyze and learn the inspection results according to the accuracy of judging good products or defective products, and corrects the inspection algorithms in steps S200 and S500, and then re-executes steps S100-S500.

[0036] Beneficial effects of the present invention:

[0037] An automated visual inspection device and inspection method of the present invention is provided with a processing module, a conveying mechanism, a visual inspection mechanism, a transfer mechanism and a temporary storage mechanism. When in use, the processing module and the visual inspection mechanism photograph and analyze the product, mark the suspected defective area and put the suspected defective product into the temporary storage mechanism to complete the screening; when the number of products in the temporary storage mechanism reaches a certain number, the transfer mechanism takes these products out and puts them into the conveying mechanism, and the processing module controls the visual inspection mechanism to perform secondary photographic analysis on these areas according to the suspected defective area recorded corresponding to the product, so as to ensure that there are no misjudged products and to eliminate confirmed defective products. Compared with the prior art, the present invention can avoid misjudgment of products while reducing the workload of the processing module in analyzing defects by performing preliminary and reselection on products, thereby effectively improving the inspection efficiency of products. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0039] Figure 1 It is a schematic diagram of the three-dimensional structure of the detection device of the present invention.

[0040] Figure 2 It is a top view of the detection device of the present invention.

[0041] Figure 3 It is a schematic diagram of the structure of the spiral structure.

[0042] Figure 4 Flow chart of the automated visual inspection method of the present invention.

[0043] Reference numerals

[0044] Conveying mechanism--1, spiral structure--11, temporary storage mechanism--2, temporary storage area--21, temporary storage track--22, transfer mechanism--3, visual inspection mechanism--4. DETAILED DESCRIPTION

[0045] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present invention, so the present invention is not limited by the specific embodiments disclosed below.

[0046] 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.

[0047] 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.

[0048] As the application of fast charger technology matures, the automated production requirements for thin, compact and standardized electronic transformer products have emerged. Under efficient automated production conditions, the production capacity of consumer electronics and home appliance electronic transformer products has increased rapidly, and the workload of external inspection has increased significantly.

[0049] After the production of this type of transformer, the pins need to be inspected for quality before they can be shipped. Traditional visual inspection uses industrial cameras and other visual inspection equipment to take pictures of the product features and analyze the image data in the photos to judge the quality of the product. Conventional manual video inspection takes a long time, the number of product moves increases, visual fatigue is inevitable, and there is no effective way to prevent the appearance of poor products. In addition, the current practice is to take pictures of each product in multiple directions and then analyze the image data. It is impossible to identify a small area by zooming in or comparing twice. Due to the limited processing speed of the processing module and the limited mechanical handling action, this detection method is inefficient and has gradually failed to meet the increasing production needs.

[0050] Example 1

[0051] In order to solve the above problems, this embodiment discloses an automated visual inspection device, the structure of which is as follows: Figure 1 and Figure 3 As shown, the detection equipment includes a processing module, a conveying mechanism, a visual inspection mechanism 4 located around the conveying mechanism, a transfer mechanism and a temporary storage mechanism; the conveying mechanism and the temporary storage mechanism are both used to convey the products to be inspected, the visual inspection mechanism 4 is used to take pictures of the products on the conveying mechanism to obtain image data of the products, the processing module is used to analyze and process the image data of the visual inspection mechanism 4, and the transfer mechanism is used to transfer the products between the conveying mechanism and the temporary storage mechanism.

[0052] In this embodiment, the visual inspection mechanism 4 is preferably six industrial cameras, and the conveying mechanism is preferably a conveying track. Furthermore, a portion of the conveying mechanism close to the visual inspection mechanism 4 is a spiral structure, which can flip certain products that need to be flipped for inspection; in addition, the conveying mechanism behind the spiral structure is hollowed out, so that after the product to be inspected is flipped by the spiral structure, the visual inspection mechanism 4 can take a picture of the bottom of the product to be inspected, which can greatly improve the efficiency of product transportation and inspection. Figure 3 As shown, through the conveying mechanism of the above structure, the six industrial cameras of this embodiment are respectively arranged at corresponding positions of the conveying mechanism to shoot six sides of the product, thereby improving the efficiency of the processing module in acquiring product image data.

[0053] In addition, corresponding sensors and limit strips are provided at corresponding positions of the conveying mechanism to improve the stability of product conveying.

[0054] It should be noted that the structures and working principles of the processing module, the visual detection mechanism 4 and the conveying mechanism in this embodiment belong to the prior art and will not be described in detail here.

[0055] Please refer to Figure 4 , shows an automated visual inspection method provided by some embodiments of the present application, which can be used to inspect the above transformer. Figure 4 As shown, the method comprises the following steps:

[0056] S100. Place the product to be inspected into the conveying mechanism, and the visual inspection mechanism 4 obtains image data in six directions: top, bottom, front and back, and both sides of the product to be inspected;

[0057] S200. The processing module analyzes and processes the image data obtained in step S100, marks the suspected defect area, and then takes out the product through the transfer mechanism and puts it into the temporary storage mechanism after marking the product;

[0058] S300. When a certain number of products are placed in the temporary storage mechanism, they are recorded as a unit, and the transfer mechanism takes out the product of the unit and puts it back into the conveying mechanism. The number of products in a unit can be single or multiple;

[0059] S400. The processing module controls the visual inspection mechanism 4 to obtain image data of the marked area corresponding to the product according to the marking order or marking value of the product in the unit;

[0060] S500. The processing module analyzes and processes the image data acquired in step S400, determines whether the image data is a good product or a bad product, and removes the bad products.

[0061] That is to say, the processing module in this embodiment has a certain memory function. For the products placed in the temporary storage mechanism, the processing module will match the position of the defective area of ​​the product one by one to ensure that there will be no errors in the subsequent re-inspection stage; in addition, the processing module controls the focal length of the industrial camera at the corresponding position according to the direction of the corresponding defective area of ​​the product to magnify the defective area of ​​the product to improve the accuracy of the judgment of the defective area, and after the above-mentioned re-inspection process is completed, the industrial camera is adjusted back to the initial focal length.

[0062] Furthermore, in step S100, after the visual inspection mechanism 4 obtains image data in five directions, namely, top, front, back, and sides of the product to be inspected, the visual inspection mechanism 4 obtains image data of the bottom of the product to be inspected when the product to be inspected passes through the spiral structure of the conveying mechanism.

[0063] In this embodiment, the positions of the six industrial cameras of the visual inspection mechanism 4 can be adjusted according to different product types.

[0064] Furthermore, the temporary storage mechanism includes a temporary storage area and a temporary storage track with a conveying direction opposite to the conveying mechanism. In step S200, the marked products are placed in the temporary storage area, and when a certain number of products are placed in the temporary storage area, the products are taken out and placed in the temporary storage track, and when the products reach the end of the temporary storage track, they are placed in the head end of the conveying mechanism. Figure 1 and Figure 2 It can be seen that the products that have undergone the initial inspection on six sides will reach the end of the conveying track. At this time, the transfer mechanism will take out the suspected defective products in the initial inspection and put them into the temporary storage area. It should be noted that for different products, the number of products in a unit in the temporary storage area can be determined, which can be a single or multiple products. After the number of products in the temporary storage area reaches the set value of the unit, the transfer mechanism can take the products out of the temporary storage area and put them into the temporary storage track, and re-convey the products of the unit to the end of the temporary storage track and put the products of the unit back to the head end of the conveying mechanism by the transfer mechanism for subsequent re-inspection.

[0065] It should be noted that the transfer mechanism in this embodiment may preferably be a multi-axis manipulator or other structure that can achieve the same function. The number and position of the transfer mechanism may be set according to actual conditions and will not be described in detail here.

[0066] Furthermore, in step S200, the steps of analyzing and processing the image data are as follows:

[0067] S201. Perform edge detection on the image data and calculate the gradient amplitude of the image. The gradient calculation formula is:

[0068]

[0069] The calculation formula of gradient amplitude is:

[0070]

[0071] Among them, G x and G y are the gradients of the image in the x and y directions respectively;

[0072] S202. Binarize the gradient amplitude image and set a threshold T;

[0073] S203. Use morphological operations to remove noise and fill defect areas;

[0074] In step S300, the defective area is marked using the connected component marking method, and an information matrix M of the marking information is established.

[0075] Further, in step S400, the processing module obtains the marking sequence or marking value of the product by reading the information matrix M, where M i =(x i ,y i ,w i ,h i ) represents the marking information of the i-th product, (x i ,y i ) is the coordinate of the upper left corner of the defect area, (w i ,h i ) are the width and height of the defect area.

[0076] In combination with the second aspect, further, in step S500, a convolutional neural network is used to classify the product to determine whether the product is a good product or a defective product, and the classification step includes:

[0077] S501. Input image L ′ i,j Perform convolution operation. The calculation formula of convolution operation is: C 1 =ReLU(Conv(L ′ i,j ,W 1 )+b 1 ), where C 1 is the convolution result, Conv represents the convolution operation, W 1 is the convolution kernel, b 1 is the bias, ReLU is the activation function;

[0078] S502. Perform a pooling operation on the output of the convolutional layer. The calculation formula for the pooling operation is: 1 =MaxPool(C 1 ), where P 1 It is the pooling result, and MaxPool represents the pooling operation;

[0079] S503. Expand the output of the pooling layer and pass it through the fully connected layer. The calculation formula for expanding the pooling layer is: F = ReLU (W 2 ·flatten(P 1 )+b 2 ), where F is the expansion result, W 2 and b 2 are the weights and biases of the fully connected layer, and flatten represents the expansion operation;

[0080] S504. The classification result is obtained through the Softmax layer. The calculation formula of the Softmax layer is: i =Softmax(W 3 ·F+b 3 ), where W 3 and b 3 are the weights and biases of the output layer, and Softmax represents the normalized exponential function.

[0081] In addition, the automated visual inspection method of this embodiment also includes step S600: the processing module uses the AI ​​big model to analyze and learn the inspection results according to the accuracy of judging good products or defective products, and corrects the inspection algorithms in steps S200 and S500, and then re-executes steps S100-S500.

[0082] By using the AI ​​big model to provide negative feedback on the judgment results of the image data in this embodiment, the accuracy of identifying good or defective products can be gradually improved.

[0083] Example 2

[0084] This embodiment provides a computer device, including a processor and a storage medium;

[0085] The storage medium is used to store instructions;

[0086] The processor is used to operate according to the instructions to execute the steps of any method described in Example 1.

[0087] Example 3

[0088] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the method described in Embodiment 1 are implemented.

[0089] In summary, the present invention provides an automated visual inspection device, method, device and storage medium. By setting a processing module, a conveying mechanism, a visual inspection mechanism 4, a transfer mechanism and a temporary storage mechanism, when in use, the processing module and the visual inspection mechanism 4 take photos and analyze the products, mark the suspected defective areas and put the suspected bad products into the temporary storage mechanism to complete the screening; when the number of products in the temporary storage mechanism reaches a certain number, the transfer mechanism takes these products out and puts them into the conveying mechanism, and the processing module controls the visual inspection mechanism 4 to take photos and analyze these areas again according to the suspected defective areas recorded corresponding to the products, so as to ensure that there are no misjudged products and to eliminate the confirmed bad products. Compared with the prior art, the present invention can avoid misjudgment of products while reducing the workload of the processing module in analyzing defects by performing preliminary and re-selection on products, thereby effectively improving the inspection efficiency of products.

[0090] The above-mentioned automated visual inspection equipment and inspection methods can effectively achieve:

[0091] 1. High efficiency, 3-5 times the efficiency of existing technology;

[0092] 2. Effectively reduce the misjudgment rate, reducing misjudgment by 60-80% compared with existing technologies;

[0093] 3. AI deep learning can continuously improve efficiency and reduce the misjudgment rate.

[0094] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0095] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0096] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0097] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0098] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention, rather than to limit the protection scope of the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the essence and scope of the technical solution of the present invention.

[0099] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention, rather than to limit the scope of protection of the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the essence and scope of the technical solution of the present invention.

Claims

1. An automated visual inspection device, characterized in that: It includes a processing module, a conveying mechanism, a visual inspection mechanism located around the conveying mechanism, a transfer mechanism, and a temporary storage mechanism; The conveying mechanism and the temporary storage mechanism are both used to convey the products to be inspected, the visual inspection mechanism is used to take pictures of the products on the conveying mechanism to obtain image data of the products, the processing module is used to analyze and process the image data of the visual inspection mechanism, and the transfer mechanism is used to transfer the products between the conveying mechanism and the temporary storage mechanism; The automated visual inspection equipment is also used to perform the following steps: S100. The product to be inspected is placed into the conveying mechanism, and the visual inspection mechanism obtains image data of the product to be inspected in six directions: top, bottom, front and back, and both sides; S200. The processing module analyzes and processes the image data acquired in step S100, marks suspected defective areas, and removes the product from the transfer mechanism and places it into a temporary storage mechanism after marking it; S300. After a certain number of products are placed in the temporary storage mechanism, they are recorded as a unit, and the transfer mechanism takes out the product of the unit and puts it back into the conveying mechanism; S400. The processing module controls the visual inspection mechanism to obtain image data of the marked area corresponding to the product according to the marking order or marking value of the product in the unit; S500. The processing module analyzes and processes the image data acquired in step S400, determines whether the image data is a good product or a bad product, and removes the bad products.

2. The automated visual inspection device according to claim 1, characterized in that: In step S300, the quantity of a product in a unit is single or multiple.

3. The automated visual inspection device according to claim 1, characterized in that: A portion of the conveying mechanism close to the visual inspection mechanism is a spiral structure, and the conveying mechanism behind the spiral structure is hollowed out so that the visual inspection mechanism can take a picture of the bottom of the product to be inspected when the product to be inspected passes by; In step S100, after the visual inspection mechanism obtains image data of the top, front and back, and both sides of the product to be inspected, the visual inspection mechanism obtains image data of the bottom of the product to be inspected after the product to be inspected passes through the spiral structure of the conveying mechanism.

4. The automated visual inspection device according to claim 1, characterized in that: The temporary storage mechanism includes a temporary storage area and a temporary storage track having a conveying direction opposite to that of the conveying mechanism; In step S200, the marked products are placed in the temporary storage area, and when a certain number of products are placed in the temporary storage area, the products are taken out and placed in the temporary storage track, and when the products reach the end of the temporary storage track, they are placed in the head end of the conveying mechanism.

5. The automated visual inspection device according to claim 1, characterized in that: In step S200, the steps of analyzing and processing the image data are as follows: S201. Perform edge detection on the image data and calculate the gradient amplitude of the image. The gradient calculation formula is: The calculation formula of gradient amplitude is: in, and The images are and Directional gradient; S202. Binarize the gradient magnitude image and set the threshold ; S203. Using morphological operations to remove noise and fill defect areas; In step S300, the defect area is marked using the connected component marking method, and an information matrix of the marking information is established. .

6. The automated visual inspection device according to claim 5, characterized in that: In step S400, the processing module reads the information matrix To obtain the product's tag order or tag value, where: Indicates Product marking information, is the coordinate of the upper left corner of the defect area, are the width and height of the defect area.

7. The automated visual inspection device according to claim 1, characterized in that: In step S500, a convolutional neural network is used to classify the product to determine whether the product is a good product or a defective product. The classification step includes: S501. Input image Perform a convolution operation. The calculation formula for the convolution operation is: ,in, is the convolution result, represents the convolution operation, is the convolution kernel, is the bias, is the activation function; S502. Perform a pooling operation on the output of the convolutional layer. The calculation formula of the pooling operation is: ,in, is the pooling result, Represents a pooling operation; S503. Expand the output of the pooling layer and pass it through the fully connected layer. The calculation formula for expanding the pooling layer is: ,in, is the expansion result, and are the weights and biases of the fully connected layer, Indicates the expansion operation; S504. Pass The classification result is obtained by layer. The calculation formula of the layer is: ,in, and are the weights and biases of the output layer, represents the normalized exponential function.

8. The automated visual inspection device according to claim 1, characterized in that: The visual inspection mechanism is an industrial camera; In step S400 , the processing module controls the industrial camera to zoom in order to obtain image data of the marked area corresponding to the product.

9. The automated visual inspection device according to claim 1, characterized in that: The automated visual inspection equipment also includes step S600: the processing module uses the AI ​​big model to analyze and learn the inspection results according to the accuracy of judging good or defective products, and corrects the inspection algorithms in steps S200 and S500, and then re-executes steps S100-S500.

Images