Thin film capacitor electrode defect detection method, device, equipment and medium

Abstract

The invention discloses a thin-film capacitor electrode defect detection method, device and equipment and a medium. The method comprises the following steps: acquiring a thin-film capacitor electrode image, and cutting the thin-film capacitor electrode image; performing manual defect labeling on the cut image, and generating an XML labeling file in a VOC format; performing first processing on the film capacitor electrode image with the defect manually marked and the XML marking file in the VOC format to obtain a first processing image and the XML marking file in the VOC format; correspondingly combining the thin-film capacitor electrode image of which the defect is manually labeled, the XML labeling file in the VOC format, the first processing image and the XML labeling file in the VOC format, performing data set division to generate a training set and a verification set of a detection model, and completing training of the detection model through the training set and the verification set; and through the training model, completing defect detection of a newly collected film capacitor electrode image. According to the invention, the complex image preprocessing and feature extraction steps are reduced, and the feature capture of the small target is more accurate.

Description

Technical Field

[0001] The present disclosure generally relates to the field of equipment detection technology, and specifically relates to a thin film capacitor electrode defect detection method, device, equipment and medium. Background Art

[0002] The electrode sheets of thin film capacitors are used to produce capacitors. Many rectangular electrode sheets are arrayed on the base film. In the process of making the electrode sheets, defects such as hypertrophy, defects, discontinuity, burrs, serrations, adhesions and white spots are generated. These defects will seriously affect the quality and performance of the capacitor, and then affect the use effect of downstream products. Therefore, it is of great significance to quickly and accurately detect electrode sheet defects to improve production efficiency and product quality. Some companies use manual inspection for electrode sheet defect detection, and manual inspection is carried out through manually operated inspection equipment. Manual inspection is relatively subjective, time-consuming and labor-intensive, with high cost, low efficiency, easy to miss small defects, and susceptible to fatigue, which affects the results of electrode sheet defect detection.

[0003] Therefore, the existing technology needs to be improved. Summary of the invention

[0004] In view of the above-mentioned defects or deficiencies in the prior art, it is desirable to provide a thin film capacitor electrode defect detection method, device, equipment and medium that can meet the needs of the art.

[0005] Based on one aspect of an embodiment of the present invention, the embodiment of the present application provides a thin film capacitor electrode defect detection method, comprising:

[0006] Acquire the thin film capacitor electrode image and crop the image;

[0007] Manually mark defects on the cropped thin-film capacitor electrode image and generate an XML annotation file in VOC format;

[0008] Performing a first processing on the thin-film capacitor electrode image with manually marked defects and the XML marking file in the VOC format to obtain a first processed image and the XML marking file in the VOC format;

[0009] The thin-film capacitor electrode image with manually marked defects and the XML annotation file in VOC format are correspondingly merged with the first processed image and the XML annotation file in VOC format, the XML annotation file in VOC format of the merged image is converted into a TXT annotation file in YOLO format, the data set is divided to generate a training set and a verification set of the detection model, and the training set and the verification set are used to complete the training of the detection model;

[0010] By training the model, defect detection of newly collected thin-film capacitor electrode images is completed.

[0011] In another embodiment, the first processing of the thin-film capacitor electrode image with manually marked defects and the XML annotation file in VOC format includes:

[0012] The thin-film capacitor electrode image with manually marked defects is flipped, the brightness is adjusted, and the blur is processed to generate the processed image and the XML annotation file in VOC format.

[0013] In another embodiment, generating a training set and a validation set of a detection model comprises:

[0014] The thin-film capacitor electrode image with manually marked defects and the XML annotation file in VOC format are correspondingly merged with the first processed image and the XML annotation file in VOC format to generate a new data set;

[0015] The data in the new dataset is divided into a training set and a validation set for the detection model in proportion.

[0016] In another embodiment, completing the training of the detection model through the training set and the validation set includes:

[0017] Select the architecture of the detection model, calculate the loss function, build the training environment of the detection model, modify the training parameters of the detection model, select the optimizer, and start training the detection model.

[0018] In another embodiment, the architecture of the selected detection model is a YOLOv8-P2 detection model.

[0019] In another embodiment, the formula for calculating the loss function is:

[0020] ;

[0021] Where c is the diagonal length of the smallest enclosing rectangle containing the predicted box and the real box in the detection model, α is an adjustment coefficient, IoU is the intersection over union ratio, which is used to measure the overlap between the predicted box and the real box, v represents the consistency of the aspect ratio between the predicted box and the real box, gt represents the real box, b is the center point of the predicted box, is the center point of the real frame, is the Euclidean distance between the center point of the predicted box and the true box:

[0022] ;

[0023] The constraint on the aspect ratio of the bounding box makes the shape of the predicted box closer to the real box by controlling the consistency of the aspect ratio. This penalty term is based on the similarity of the aspect ratio:

[0024] ;

[0025] , , w, h are the width, height of the real box and the width, height of the predicted box respectively.

[0026] In another aspect of the embodiments of the present invention, a thin film capacitor electrode defect detection device is disclosed, comprising:

[0027] A processing module is used to obtain a thin film capacitor electrode image and crop the image; manually mark defects on the cropped thin film capacitor electrode image and generate an XML annotation file in VOC format; correspondingly merge the thin film capacitor electrode image with manually marked defects and the XML annotation file in VOC format with the first processed image and the XML annotation file in VOC format, convert the XML annotation file in VOC format of the merged image into a TXT annotation file in YOLO format, and divide the data set to generate a training set and a verification set for the detection model;

[0028] The training module is used to complete the training of the detection model through the training set and the validation set;

[0029] The detection module is used to complete defect detection of thin film capacitor electrode images using the trained model.

[0030] According to another aspect of an embodiment of the present invention, an electronic device is disclosed, characterized in that the device includes one or more processors and a memory, and the memory is used to store one or more programs;

[0031] When the one or more programs are executed by the processor, the processor implements the thin film capacitor electrode defect detection method provided by each embodiment of the present invention.

[0032] According to another aspect of the embodiments of the present invention, a computer-readable storage medium storing a computer program is disclosed. When the computer program is executed, the method for detecting thin film capacitor electrode defects provided by the embodiments of the present invention is implemented.

[0033] In an embodiment of the present application, by acquiring a thin film capacitor electrode image and cropping the image; manually marking defects on the cropped thin film capacitor electrode image and generating an XML annotation file in VOC format; performing a first processing on the thin film capacitor electrode image with manually marked defects and the XML annotation file in VOC format, obtaining a first processed image and an XML annotation file in VOC format; correspondingly merging the thin film capacitor electrode image with manually marked defects and the XML annotation file in VOC format with the first processed image and the XML annotation file in VOC format, converting the XML annotation file in VOC format of the merged image into a TXT annotation file in YOLO format, performing data set division to generate a training set and a verification set of the detection model, and completing the training of the detection model through the training set and the verification set; completing the defect detection of the newly acquired thin film capacitor electrode image through the training model. The present application reduces the complex image preprocessing and feature extraction steps, and more accurately captures the features of small targets, making the entire detection process simpler and faster. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0035] Figure 1 is a flow chart of a thin film capacitor electrode defect detection method provided by an embodiment of the present application;

[0036] Figure 2 A schematic diagram of the structure of a thin film capacitor electrode defect detection device provided in one embodiment of the present application;

[0037] Figure 3 It is a diagram of the internal structure of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0038] The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.

[0039] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0040] Please refer to Figure 1 , which shows an exemplary process of a thin film capacitor electrode defect detection method that can be applied in an embodiment of the present application.

[0041] like Figure 1 As shown, the method includes:

[0042] Step 110, obtaining a thin film capacitor electrode image and cropping the image;

[0043] Specifically, the image resolution of the original collected thin-film capacitor electrode image is 33000*8192. Since the resolution of the thin-film capacitor electrode image is too large and the electrode product defects are too small, the thin-film capacitor electrode image needs to be cropped into multiple small images to facilitate subsequent model training.

[0044] Step 120, manually marking defects on the cropped thin-film capacitor electrode image, and generating an XML marking file in VOC format;

[0045] Specifically, since there are a large number of thin-film capacitor electrode images, the images are manually annotated in small batches in advance with the aid of annotation software to obtain XML annotation files in VOC format.

[0046] Step 130, performing a first processing on the thin-film capacitor electrode image with manually marked defects and the XML marking file in VOC format to obtain a first processed image and an XML marking file in VOC format;

[0047] Specifically, data enhancement is performed on the annotated first processed image, and operations such as flipping and adjusting the brightness of the first processed image are performed on the first processed image. At the same time, the data-enhanced first processed image will obtain a corresponding XML annotation file in the VOC format.

[0048] Specifically, in one embodiment of the present application, the first processing of the thin-film capacitor electrode image with manually marked defects and the XML annotation file in VOC format includes:

[0049] The thin-film capacitor electrode image with manually marked defects is flipped, the brightness is adjusted, and the blur is processed to generate the processed image and the XML annotation file in VOC format.

[0050] Step 140, the thin-film capacitor electrode image with manually marked defects and the XML annotation file in VOC format are correspondingly merged with the first processed image and the XML annotation file in VOC format, the XML annotation file in VOC format of the merged image is converted into a TXT annotation file in YOLO format, the data set is divided to generate a training set and a verification set of the detection model, and the detection model is trained through the training set and the verification set;

[0051] Specifically, in one embodiment of the present application, the generating of the training set and the validation set of the detection model includes:

[0052] The thin-film capacitor electrode image with manually marked defects and the XML annotation file in VOC format are correspondingly merged with the first processed image and the XML annotation file in VOC format to generate a new data set;

[0053] The data in the new dataset is divided into a training set and a validation set for the detection model in proportion.

[0054] Specifically, in one embodiment of the present application, the training of the detection model by using the training set and the validation set includes:

[0055] Select the architecture of the detection model, calculate the loss function, build the training environment of the detection model, modify the training parameters of the detection model, select the optimizer, and start training the detection model.

[0056] In an embodiment of the present application, the architecture of the selected detection model is the YOLOv8-P2 detection model.

[0057] The formula for calculating the loss function is:

[0058] ;

[0059] Where c is the diagonal length of the smallest enclosing rectangle containing the predicted box and the real box in the detection model, α is an adjustment coefficient, IoU is the intersection over union ratio, which is used to measure the overlap between the predicted box and the real box, v represents the consistency of the aspect ratio between the predicted box and the real box, gt represents the real box, b is the center point of the predicted box, is the center point of the real frame, is the Euclidean distance between the center point of the predicted box and the true box:

[0060] ;

[0061] The constraint on the aspect ratio of the bounding box makes the shape of the predicted box closer to the real box by controlling the consistency of the aspect ratio. This penalty term is based on the similarity of the aspect ratio:

[0062] ;

[0063] , , w, h are the width, height of the real box and the width, height of the predicted box respectively.

[0064] Specifically, YOLOv8 consists of a backbone network Backbone, a feature fusion layer Neck, and a detection head Head. The backbone network Backbone is responsible for extracting the basic features of the input image. It extracts multi-level feature representations from the original pixels of the image through a series of convolutional layers, pooling layers, and activation functions. The main purpose of the feature fusion layer Neck is to fuse features from different levels of the backbone network Backbone in order to better identify and locate targets of different scales. The detection head Head is responsible for performing specific target detection tasks on the features fused by the feature fusion layer Neck, that is, regressing the location and classification of the target.

[0065] The purpose of calculating the loss function is to optimize the bounding box regression, and the loss function is calculated by distance, overlap area and aspect ratio parameters.

[0066] The original structure of YOLO includes three detection heads: P3, P4, and P5, which correspond to different detection scales from top to bottom, and can detect small-scale, medium-scale, and large-scale targets. In order to enhance the detection capability of small targets, YOLOv8 is constantly improving and adding a detection head to the P2 layer to improve the recognition effect of small targets.

[0067] Use the divided data set to train YOLOv8-P2, build the model training environment, modify the YOLOv8-P2 training parameters, including the number of training rounds and batches, select the optimizer, and start model training. Crop the test basement membrane image, use the trained model for inference, generate a result map with defects marked, compare the result map generated by the model with the original image, and check whether the model correctly identifies the defects in the electrode sheet image. Record the time required for the model to infer each image to determine whether the model's inference speed meets the requirements.

[0068] Step 150, completing defect detection on the newly collected thin-film capacitor electrode image by training the model.

[0069] The thin film capacitor electrode defect detection method of the present application obtains the thin film capacitor electrode image and crops the image; manually marks defects on the cropped thin film capacitor electrode image and generates an XML annotation file in VOC format; performs a first processing on the thin film capacitor electrode image with manually marked defects and the XML annotation file in VOC format to obtain the first processed image and the XML annotation file in VOC format; the thin film capacitor electrode image with manually marked defects and the XML annotation file in VOC format and the first processed image and the XML annotation file in VOC format are correspondingly merged, and the XML annotation file in VOC format of the merged image is converted into a TXT annotation file in YOLO format, and the data set is divided to generate the training set and verification set of the detection model, and the training of the detection model is completed through the training set and the verification set; through the training model, the defect detection of the newly collected thin film capacitor electrode image is completed. The present application reduces the complex image preprocessing and feature extraction steps, and the feature capture of small targets is more accurate, making the entire detection process simpler and faster.

[0070] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0071] Figure 2 FIG. 1 is a schematic diagram of a thin film capacitor electrode defect detection device provided by an embodiment of the present application. Figure 2 As shown, the thin film capacitor electrode defect detection device comprises:

[0072] Processing module, training module, detection module;

[0073] A processing module is used to obtain a thin film capacitor electrode image and crop the image; manually mark defects on the cropped thin film capacitor electrode image and generate an XML annotation file in VOC format; correspondingly merge the thin film capacitor electrode image with manually marked defects and the XML annotation file in VOC format with the first processed image and the XML annotation file in VOC format, convert the XML annotation file in VOC format of the merged image into a TXT annotation file in YOLO format, and divide the data set to generate a training set and a verification set for the detection model;

[0074] The training module is used to complete the training of the detection model through the training set and the validation set;

[0075] The detection module is used to complete defect detection of thin film capacitor electrode images using the trained model.

[0076] In another embodiment of the present application, the processing module is used to:

[0077] The thin-film capacitor electrode image with manually marked defects is flipped, the brightness is adjusted, and the blur is processed to generate the processed image and the XML annotation file in VOC format.

[0078] In another embodiment of the present application, the processing module is further used for:

[0079] Convert the XML annotation file in VOC format of the first processed image into a TXT file in YOLO format.

[0080] In another embodiment of the present application, the processing module is further used for:

[0081] The thin-film capacitor electrode image with manually marked defects and the XML annotation file in VOC format are correspondingly merged with the first processed image and the XML annotation file in VOC format to generate a new data set;

[0082] The data in the new dataset is divided into a training set and a validation set for the detection model in proportion.

[0083] In another embodiment of the present application, the processing module is further used for:

[0084] Select the architecture of the detection model, calculate the loss function, build the training environment of the detection model, modify the training parameters of the detection model, select the optimizer, and start training the detection model.

[0085] In an embodiment of the present application, the thin film capacitor electrode defect detection device obtains a thin film capacitor electrode image through a processing module and crops the image; manually marks defects on the cropped thin film capacitor electrode image and generates an XML annotation file in VOC format; performs a first processing on the thin film capacitor electrode image with manually marked defects and the XML annotation file in VOC format to obtain a first processed image and an XML annotation file in VOC format; the thin film capacitor electrode image with manually marked defects and the XML annotation file in VOC format and the first processed image and the XML annotation file in VOC format are correspondingly merged, and the XML annotation file in VOC format of the merged image is converted into a TXT annotation file in YOLO format, and the data set is divided to generate a training set and a verification set of the detection model, and the training of the detection model is completed through the training set and the verification set; the detection module completes the defect detection of the newly collected thin film capacitor electrode image through the training model. The present application reduces the complex image preprocessing and feature extraction steps, and the feature capture of small targets is more accurate, making the entire detection process simpler and faster.

[0086] For the specific definition of the thin film capacitor electrode defect detection device, please refer to the definition of the thin film capacitor electrode defect detection method mentioned above, which will not be repeated here. Each module in the above-mentioned thin film capacitor electrode defect detection device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0087] In particular, according to the embodiments of the present disclosure, Figure 3 As shown, the present invention discloses an electronic device, which includes one or more processors and a memory, wherein the memory is used to store one or more programs; when the one or more programs are executed by the processor, the processor implements the thin film capacitor electrode defect detection method described in an embodiment of the present invention.

[0088] In particular, according to an embodiment of the present disclosure, the method for software upgrade described in any of the above embodiments can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program tangibly contained on a machine-readable medium, and the computer program includes a program code for executing the method for software upgrade. In such an embodiment, the computer program can be downloaded and installed from a network through a communication part, and / or installed from a removable medium.

[0089] The one or more programs are stored in a read-only memory ROM or a random access memory RAM to perform various appropriate actions and processes. The random access memory RAM includes software programs for the server to complete corresponding services, as well as various programs and data required for vehicle driving operations. The server and its controlled hardware devices, read-only memory ROM, and random access memory RAM are connected to each other through a bus, and various input / output interfaces are also connected to the bus.

[0090] The following components are connected to the input / output interface: an input part including a keyboard, a mouse, etc.; an output part including a cathode ray tube CRT, a liquid crystal display LCD, etc., and a speaker, etc.; and a communication part including a network interface card such as a LAN card, a modem, etc. The communication part performs communication processing via a network such as the Internet. A drive is also connected to the input / output interface as needed. Removable media, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., are installed on the drive as needed so that a computer program read therefrom is installed into the memory as needed.

[0091] In particular, according to an embodiment of the present disclosure, the thin film capacitor electrode defect detection method described in any of the above embodiments can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program tangibly contained on a machine-readable medium, and the computer program includes a program code for executing a method for software upgrading. In such an embodiment, the computer program can be downloaded and installed from a network through a communication part, and / or installed from a removable medium.

[0092] The units or modules involved in the embodiments described in the present application may be implemented by software or hardware. The units or modules described may also be arranged in a processor. The names of these units or modules do not constitute limitations on the units or modules themselves in certain circumstances.

[0093] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the inventive concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the present application.

Claims

1. A thin film capacitor electrode defect detection method, characterized in that: include: Acquire the thin film capacitor electrode image and crop the image; Manually mark defects on the cropped thin-film capacitor electrode image and generate an XML annotation file in VOC format; Performing a first processing on the thin-film capacitor electrode image with manually marked defects and the XML marking file in the VOC format to obtain a first processed image and the XML marking file in the VOC format; The thin-film capacitor electrode image with manually marked defects and the XML annotation file in VOC format are correspondingly merged with the first processed image and the XML annotation file in VOC format, the XML annotation file in VOC format of the merged image is converted into a TXT annotation file in YOLO format, the data set is divided to generate a training set and a verification set of the detection model, and the training set and the verification set are used to complete the training of the detection model; By training the model, defect detection of newly collected thin-film capacitor electrode images is completed.

2. The method according to claim 1, characterized in that The first processing of the thin-film capacitor electrode image with manually marked defects and the XML annotation file in VOC format includes: The thin-film capacitor electrode image with manually marked defects is flipped, the brightness is adjusted, and the blur is processed to generate the processed image and the XML annotation file in VOC format.

3. The method according to claim 1, characterized in that The training set and validation set of the generated detection model include: The thin-film capacitor electrode image with manually marked defects and the XML annotation file in VOC format are correspondingly merged with the first processed image and the XML annotation file in VOC format to generate a new data set; The data in the new dataset is divided into a training set and a validation set for the detection model in proportion.

4. The method according to claim 1, characterized in that The training of the detection model is completed by using the training set and the validation set, including: Select the architecture of the detection model, calculate the loss function, build the training environment of the detection model, modify the training parameters of the detection model, select the optimizer, and start training the detection model.

5. The method according to claim 4, characterized in that The architecture of the selected detection model is the YOLOv8-P2 detection model.

6. The method according to claim 4, characterized in that The formula for calculating the loss function is: ； Where c is the diagonal length of the smallest enclosing rectangle containing the predicted box and the real box in the detection model, α is an adjustment coefficient, IoU is the intersection over union ratio, which is used to measure the overlap between the predicted box and the real box, v represents the consistency of the aspect ratio between the predicted box and the real box, gt represents the real box, b is the center point of the predicted box, is the center point of the real frame, is the Euclidean distance between the center point of the predicted box and the true box: ； The constraint on the aspect ratio of the bounding box makes the shape of the predicted box closer to the real box by controlling the consistency of the aspect ratio. This penalty term is based on the similarity of the aspect ratio: ； , , w, h are the width, height of the real box and the width, height of the predicted box respectively.

7. A thin film capacitor electrode defect detection device, characterized in that: include: A processing module, used for acquiring the thin film capacitor electrode image and cropping the image; Manually mark defects on the cropped thin-film capacitor electrode image and generate an XML annotation file in VOC format; The thin-film capacitor electrode image with manually marked defects and the XML annotation file in VOC format are correspondingly merged with the first processed image and the XML annotation file in VOC format, and the XML annotation file in VOC format of the merged image is converted into a TXT annotation file in YOLO format, and the data set is divided to generate a training set and a verification set of the detection model; The training module is used to complete the training of the detection model through the training set and the validation set; The detection module is used to complete defect detection of thin film capacitor electrode images using the trained model.

8. An electronic device, characterized in that: The device includes one or more processors and a memory, wherein the memory is used to store one or more programs; When the one or more programs are executed by the processor, the processor is caused to implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed, the method according to any one of claims 1 to 6 is implemented.

Images