Battery envelope detection method, device and system

Abstract

The invention discloses a battery envelope detection method, device and system, and belongs to the technical field of battery detection. The method comprises the following steps: acquiring an envelope image of a to-be-detected battery envelope; splitting the coating image to obtain a first split image and a second split image corresponding to the coating image; determining a first composite image based on the first split image and the second split image; based on the first split image, the second split image and the first synthetic image, determining a defect detection result of the to-be-detected battery envelope; wherein the first split image comprises a part of pixels in the enveloping image, the second split image comprises a part of pixels in the enveloping image, and the first split image and the second split image comprise different pixels. The method can improve the defect detection efficiency of the battery coating.

Description

Technical Field

[0001] The present application belongs to the field of battery detection technology, and in particular, relates to a battery packaging detection method, device and system. Background Art

[0002] The battery cell is the most basic and important independent unit in the power battery. In the production process, a layer of blue protective film is usually wrapped around the battery cell as a safety barrier for the battery. In order to increase the life of the battery and ensure the safety of the battery, it is necessary to detect whether the blue film of the battery has defects during the battery production process.

[0003] At present, pressure data applied on the surface of the battery blue film is usually collected by a pressure sensor, and the battery blue film is inspected for defects based on the pressure data. However, the efficiency of this method for inspecting defects in the battery blue film is low. Summary of the invention

[0004] The present application aims to solve at least one of the technical problems existing in the prior art. To this end, the present application proposes a battery package film detection method, device and system, which can improve the defect detection efficiency of the battery package film.

[0005] In a first aspect, the present application provides a method for detecting a battery envelope, the method comprising:

[0006] Acquire a membrane image of a battery membrane to be inspected;

[0007] Splitting the envelope image to obtain a first split image and a second split image corresponding to the envelope image;

[0008] Determine a first composite image based on the first split image and the second split image;

[0009] Determine a defect detection result of the battery package to be inspected based on the first split image, the second split image and the first composite image;

[0010] The first split image includes some pixels in the envelope image, the second split image includes some pixels in the envelope image, and the first split image and the second split image include different pixels.

[0011] According to the battery package inspection method of the present application, by acquiring the package image of the battery package to be inspected, the first split image, the second split image and the first composite image corresponding to the package image are obtained, and the defect detection result is determined according to the image recognition results of the first split image, the second split image and the first composite image. There is no need for human intervention to apply pressure to the battery package, etc., and the process of processing complex data is reduced, which can improve the defect detection efficiency of the battery package.

[0012] According to an embodiment of the present application, determining the defect detection result of the battery package to be inspected based on the first split image, the second split image and the first composite image includes:

[0013] Inputting the first split image, the second split image and the first composite image into a first detection model, detecting whether the battery package to be detected has a foreign matter defect or an air bubble defect, and obtaining first detection information of the battery package to be detected output by the first detection model, wherein the first detection information is used to characterize whether the battery package to be detected has a defect and the type of the defect;

[0014] Determine a defect detection result of the battery package to be detected based on the first detection information;

[0015] The first detection model is obtained by training with bubble foreign body sample images.

[0016] According to an embodiment of the present application, determining the defect detection result of the battery package to be detected based on the first detection information includes:

[0017] In the case where it is determined based on the first detection information that the battery package to be detected has the foreign matter defect, the first split image, the second split image, and the first composite image are input into a second detection model to determine whether the foreign matter defect is a hard foreign matter defect or a soft foreign matter defect, and second detection information of the battery package to be detected output by the second detection model is obtained, where the second detection information is used to characterize whether the battery package to be detected has the hard foreign matter defect or the soft foreign matter defect;

[0018] Determine a defect detection result of the battery package to be detected based on the second detection information and the foreign matter threshold information;

[0019] The second detection model is obtained by training with foreign body sample images.

[0020] According to an embodiment of the present application, the second detection model includes a first detection sub-model and a second detection sub-model, and the first split image, the second split image and the first composite image are input into the second detection model, and the foreign body defect is judged to be a hard foreign body defect or a soft foreign body defect, and the second detection information of the battery package to be detected output by the second detection model is obtained, including:

[0021] Inputting the first split image and the second split image into the first detection sub-model, detecting the foreign body defect, and obtaining a first defect area of ​​the battery package to be detected output by the first detection sub-model;

[0022] Inputting the first synthetic image into the second detection sub-model, detecting the foreign body defect, and obtaining a second defect region of the battery package to be detected output by the second detection sub-model;

[0023] The second detection information is determined based on the first defect area and the second defect area.

[0024] According to an embodiment of the present application, determining the second detection information based on the first defect area and the second defect area includes:

[0025] In a case where the first defect region overlaps with the second defect region, determining that the foreign matter defect corresponding to the first defect region is the hard foreign matter defect;

[0026] Alternatively, when the first defect region does not overlap with the second defect region, the foreign matter defect corresponding to the first defect region is determined to be the soft foreign matter defect.

[0027] According to an embodiment of the present application, the first detection model includes a third detection sub-model and a fourth detection sub-model, and the first split image, the second split image, and the first composite image are input into the first detection model to detect whether the battery package to be detected has a foreign matter defect or an air bubble defect, and obtain the first detection information of the battery package to be detected output by the first detection model, including:

[0028] Inputting the first split image and the second split image into the fourth detection sub-model, detecting whether the battery package to be detected has the foreign matter defect or the air bubble defect, and obtaining fourth detection information of the battery package to be detected output by the fourth detection sub-model, wherein the fourth detection information is used to characterize whether the battery package to be detected has defects and the type of defects;

[0029] Inputting the first synthetic image into the third detection sub-model, detecting whether the battery package to be inspected has the foreign matter defect or the air bubble defect, and obtaining fifth detection information of the battery package to be inspected output by the third detection sub-model, wherein the fifth detection information is used to characterize whether the battery package to be inspected has defects and the type of defects;

[0030] The first detection information includes the fourth detection information and the fifth detection information;

[0031] The step of determining the defect detection result of the battery package to be detected based on the first detection information includes:

[0032] Based on the fourth detection information and the fifth detection information, a defect detection result of the battery package to be detected is determined.

[0033] According to an embodiment of the present application, determining the defect detection result of the battery package film to be detected based on the fourth detection information and the fifth detection information includes:

[0034] In a case where it is determined that the battery package to be inspected has the air bubble defect based on the fourth detection information and the fifth detection information, and it is determined that the size of the air bubble defect is smaller than the size corresponding to the first bubble threshold information based on the fourth detection information, the fifth detection information and the first bubble threshold information, the first synthetic image is input into the third detection sub-model to detect the air bubble defect, and the third detection information of the battery package to be inspected output by the third detection sub-model is obtained, where the third detection information is used to characterize the size of the air bubble defect;

[0035] Based on the third detection information and the second bubble threshold information, a defect detection result of the battery package to be detected is determined.

[0036] According to one embodiment of the present application, the step of acquiring a membrane image of a battery membrane to be inspected includes:

[0037] The membrane image of the battery membrane to be inspected is acquired by time-sharing stroboscopic technology.

[0038] According to one embodiment of the present application, the first split image includes even-numbered rows of pixels in the envelope image, and the second split image includes odd-numbered rows of pixels in the envelope image.

[0039] In a second aspect, the present application provides a battery package detection device, the device comprising:

[0040] An acquisition module, used for acquiring a membrane image of a battery membrane to be inspected;

[0041] A first processing module, used for splitting the capsule image to obtain a first split image and a second split image corresponding to the capsule image;

[0042] A second processing module, configured to determine a first composite image based on the first split image and the second split image;

[0043] A third processing module, configured to determine a defect detection result of the battery package to be inspected based on the first split image, the second split image and the first composite image;

[0044] The first split image includes some pixels in the envelope image, the second split image includes some pixels in the envelope image, and the first split image and the second split image include different pixels.

[0045] According to the battery package detection device of the present application, by acquiring the package image of the battery package to be detected, the first split image, the second split image and the first composite image corresponding to the package image are obtained, and the defect detection result is determined according to the image recognition results of the first split image, the second split image and the first composite image. There is no need for human intervention to apply pressure to the battery package, etc., and the process of processing complex data is reduced, which can improve the defect detection efficiency of the battery package.

[0046] In a third aspect, the present application provides a battery envelope detection system, the system comprising:

[0047] An image acquisition device, wherein the image acquisition device is used to acquire an image of a battery package to be inspected;

[0048] As in the battery pack detection device described in the second aspect above, the detection device is connected to the image acquisition device.

[0049] According to the battery package inspection system of the present application, by acquiring the package image of the battery package to be inspected, a first split image, a second split image and a first composite image corresponding to the package image are obtained, and the defect detection result is determined according to the image recognition results of the first split image, the second split image and the first composite image. There is no need for human intervention to apply pressure to the battery package, etc., and the process of processing complex data is reduced, which can improve the defect detection efficiency of the battery package.

[0050] According to one embodiment of the present application, the battery package detection system further includes:

[0051] A light source module is used to emit a light beam to the battery package to be inspected by using a time-sharing stroboscopic technique.

[0052] In a fourth aspect, the present application provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the battery pack detection method as described in the first aspect above is implemented.

[0053] In a fifth aspect, the present application provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the battery package detection method as described in the first aspect above is implemented.

[0054] In a sixth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the battery package detection method as described in the first aspect above.

[0055] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0057] Figure 1 This is one of the flow charts of the battery envelope detection method provided in the embodiment of the present application;

[0058] Figure 2 is a schematic diagram of a first split image provided in an embodiment of the present application;

[0059] Figure 3 is a schematic diagram of a first composite image provided in an embodiment of the present application;

[0060] Figure 4 This is the second flow chart of the battery envelope detection method provided in the embodiment of the present application;

[0061] Figure 5 This is the third flow chart of the battery envelope detection method provided in the embodiment of the present application;

[0062] Figure 6 It is a schematic diagram of the process of training the first detection model and the second detection model provided in an embodiment of the present application;

[0063] Figure 7 It is a structural schematic diagram of a battery envelope detection system provided in an embodiment of the present application;

[0064] Figure 8 It is a structural schematic diagram of a battery package detection device provided in an embodiment of the present application;

[0065] Fig. 9 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0066] Reference numerals:

[0067] Battery cell 710 , image acquisition device 720 , light source 730 . DETAILED DESCRIPTION

[0068] The following will be combined with the drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments in the present application belong to the scope of protection of this application.

[0069] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first", "second", etc. are generally of one type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0070] In combination with the accompanying drawings, the battery package detection method, battery package detection device, battery package detection system, electronic device and readable storage medium provided in the embodiments of the present application are described in detail through specific embodiments and their application scenarios.

[0071] The battery package detection method may be applied to a terminal, and may be specifically executed by hardware or software in the terminal.

[0072] The terminal includes but is not limited to a desktop computer.It should be understood that the terminal may include one or more other physical user interface devices such as a physical keyboard, a mouse and a joystick.

[0073] The battery package detection method provided in the embodiment of the present application can be executed by an electronic device or a functional module or functional entity in the electronic device that can implement the battery package detection method. The electronic device mentioned in the embodiment of the present application includes but is not limited to a computer. The battery package detection method provided in the embodiment of the present application is described below using an electronic device as an example of the execution subject.

[0074] The battery package film in this embodiment may be a protective film wrapped around the outside of the battery cell 710 in the power battery, and the battery package film may be a blue film.

[0075] The battery package inspection method provided in the embodiment of the present application is used to perform defect inspection on the battery package, to detect whether the battery package has defects, and if the battery package has defects, to detect the defect type of the battery package.

[0076] like Figure 1As shown, the battery package detection method includes: steps 110 to 140.

[0077] Step 110 : Acquire a coating image of the battery coating to be inspected.

[0078] The battery package to be inspected is a battery package to be inspected for defects, and the package image is an image including the battery package to be inspected.

[0079] In this embodiment, the battery package to be inspected may be photographed by the image acquisition device 720 to obtain a package image of the battery package to be inspected.

[0080] Step 120: split the capsule image to obtain a first split image and a second split image corresponding to the capsule image.

[0081] The first split image includes some pixels in the envelope image, the second split image includes some pixels in the envelope image, and the first split image and the second split image include different pixels.

[0082] In this embodiment, the capsule image can be split into pixels as the smallest unit, a first split image can be obtained based on a part of the pixels in the capsule image, and a second split image can be obtained based on another part of the pixels in the capsule image. The number of pixels included in the first split image and the number of pixels included in the second split image can be the same.

[0083] In actual implementation, the envelope image may be divided into a first target row and a second target row, a first split image is generated according to pixels of the first target row, and a second split image is generated according to pixels of the second target row.

[0084] In actual implementation, the capsule image may also be divided into a first target column and a second target column. A first split image is generated according to pixels of the first target column, and a second split image is generated according to pixels of the second target column.

[0085] It can be understood that the first split image can represent some features of the capsule image, and the second split image can represent some features of the capsule image.

[0086] In this embodiment, the pixels of the capsule image may be divided into two categories according to the brightness characteristics of the capsule image, and the first split image is generated according to one category of pixels, and the second split image is generated according to the other category of pixels.

[0087] Step 130: Determine a first composite image based on the first split image and the second split image.

[0088] The first composite image is an image obtained by synthesizing the first split image and the second split image.

[0089] In this embodiment, the first split image and the second split image may be synthesized with pixels as the minimum unit, and pixels at corresponding positions of the first split image and the second split image may be synthesized to obtain a first synthesized image.

[0090] In actual implementation, mathematical operations may be performed between pixels at corresponding positions of the first split image and the second split image to achieve synthesis of the first split image and the second split image.

[0091] Step 140: Determine a defect detection result of the battery package to be inspected based on the first split image, the second split image and the first composite image.

[0092] The defect detection result is used to characterize whether there are defects in the battery package to be detected. In the case that there are defects in the battery package to be detected, the defect detection result can also characterize the defect type of the battery package to be detected.

[0093] The defect detection results may include the defect type of the battery package to be detected, the number of each type of defects, and the size of the defects.

[0094] In this step, image recognition may be performed on the first split image, the second split image, and the first composite image respectively, and the image recognition results of the first split image, the second split image, and the first composite image are integrated to determine the defect detection result.

[0095] In the related art, pressure data applied on the surface of the battery blue film is collected by a pressure sensor, and defect detection of the battery blue film is performed based on the pressure data. This method of defect detection of the battery blue film is relatively inefficient.

[0096] In an embodiment of the present application, by acquiring a membrane image of a battery membrane to be inspected, a first split image, a second split image, and a first composite image corresponding to the membrane image are obtained, and image recognition is performed on the first split image, the second split image, and the first composite image, respectively. The image recognition results of the first split image, the second split image, and the first composite image are combined to determine a defect detection result to determine whether the battery membrane to be inspected has defects. In the case that the battery membrane to be inspected has defects, the defect detection result can also characterize the defect type of the battery membrane to be inspected.

[0097] There is no need for human intervention to apply pressure to the battery package. After obtaining the package image of the battery package to be inspected, the three corresponding images in the package image can be obtained through simple image splitting and image synthesis. The three images corresponding to the package image are processed by image recognition to obtain the defect detection results, reducing the process of processing complex data and improving the defect detection efficiency.

[0098] According to the battery package detection method provided in the embodiment of the present application, by acquiring the package image of the battery package to be detected, the first split image, the second split image and the first composite image corresponding to the package image are obtained, and the defect detection result is determined according to the image recognition results of the first split image, the second split image and the first composite image. There is no need for human intervention to apply pressure to the battery package, etc., and the process of processing complex data is reduced, which can improve the defect detection efficiency of the battery package.

[0099] In some embodiments, based on the first split image, the second split image and the first composite image, determining the defect detection result of the battery package to be detected includes:

[0100] Inputting the first split image, the second split image and the first composite image into the first detection model, detecting whether the battery package to be detected has foreign matter defects or air bubble defects, and obtaining first detection information of the battery package to be detected output by the first detection model;

[0101] Based on the first detection information, a defect detection result of the battery package to be detected is determined.

[0102] Among them, the first detection model is obtained by training the bubble foreign matter sample images, and the bubble foreign matter sample images are images that mark air bubble defects and foreign matter defects respectively. The first detection model can be used to detect whether the battery package to be inspected has no defects, air bubble defects or foreign matter defects.

[0103] The air bubble defect is a defect corresponding to bubbles formed due to the loose contact between the battery film and the battery cell 710 and the presence of air. The foreign matter defect is a defect corresponding to the presence of foreign matter between the film and the battery cell 710.

[0104] In this embodiment, multiple battery package sample images can be obtained, and each battery package sample image can be split to obtain a first split sample image and a second split sample image corresponding to the battery package sample image, and based on the first split sample image and the second split sample image, a first composite sample image can be obtained.

[0105] The air bubble defects and foreign matter defects in the first split sample image can be annotated to obtain the corresponding bubble foreign matter sample image; the air bubble defects and foreign matter defects in the second split sample image can be annotated to obtain the corresponding bubble foreign matter sample image; the air bubble defects and foreign matter defects in the first synthetic sample image can be annotated to obtain the corresponding bubble foreign matter sample image.

[0106] In this embodiment, the first detection model can be trained using multiple bubble foreign matter sample images, so that the first detection model can recognize the input first split image, second split image and first composite image to obtain first detection information.

[0107] Among them, the first detection information is used to characterize whether the battery package to be inspected has defects and the type of defects. The first detection information can characterize whether the battery package to be inspected has no defects, air bubble defects or foreign matter defects. The first detection information can also include the location, outline and size of the defects.

[0108] In this embodiment, after obtaining the first detection information, it can be determined whether there are defects in the battery package to be detected. If it is determined that there are defects in the battery package to be detected, further judgment can be made to determine the defect detection result of the battery package to be detected.

[0109] In the case where there is a foreign matter defect in the battery package to be inspected, further determination may include determining what kind of foreign matter defect it is and determining whether the size of the foreign matter defect is large enough to make the battery package to be inspected unqualified.

[0110] In the case where an air bubble defect exists in the battery package to be inspected, further determination may include determining whether the size of the air bubble defect is large enough to make the battery package to be inspected unqualified.

[0111] In some embodiments, based on the first detection information, determining a defect detection result of the battery package to be detected includes:

[0112] In the case where it is determined based on the first detection information that the battery package to be inspected has a foreign matter defect, the first split image, the second split image, and the first composite image are input into the second detection model to determine whether the foreign matter defect is a hard foreign matter defect or a soft foreign matter defect, and obtain second detection information of the battery package to be inspected output by the second detection model;

[0113] Based on the second detection information and the foreign matter threshold information, a defect detection result of the battery package to be detected is determined.

[0114] Among them, the second detection model is obtained by training the foreign body sample image, and the foreign body sample image is an image with foreign body defects marked. The second detection model can be used to detect whether the foreign body defect in the battery package to be inspected is a soft foreign body defect or a hard foreign body defect.

[0115] The soft foreign matter defect is a defect corresponding to the presence of soft foreign matter between the coating and the battery cell 710. The soft foreign matter may be a foreign matter that is easily deformed. The hard foreign matter defect is a defect corresponding to the presence of hard foreign matter between the coating and the battery cell 710. The hard foreign matter may be a foreign matter that is not easily deformed.

[0116] In this embodiment, multiple battery package sample images can be obtained, and each battery package sample image can be split to obtain a first split sample image and a second split sample image corresponding to the battery package sample image, and based on the first split sample image and the second split sample image, a first composite sample image can be obtained.

[0117] The foreign body defects in the first split sample image can be marked to obtain the corresponding foreign body sample image; the foreign body defects in the second split sample image can be marked to obtain the corresponding foreign body sample image; the foreign body defects in the first synthetic sample image can be marked to obtain the corresponding foreign body sample image.

[0118] In this embodiment, the second detection model can be trained using multiple foreign body sample images, so that the second detection model can recognize the input first split image, second split image and first composite image to obtain second detection information.

[0119] The second detection information is used to characterize whether the battery package to be detected has a hard foreign matter defect or a soft foreign matter defect. The second detection information may also include the location, contour and size of the foreign matter defect.

[0120] In this embodiment, after obtaining the second detection information, it can be determined whether a soft foreign matter defect or a hard foreign matter defect exists in the battery package to be detected.

[0121] In this embodiment, the foreign matter threshold information is pre-set, and the foreign matter threshold information may include size information. The foreign matter threshold information is used to determine whether the size of the soft foreign matter defect or the size of the hard foreign matter defect is large enough to make the battery pack to be detected unqualified, that is, when the size of the soft foreign matter defect is larger than the corresponding foreign matter threshold information, the size of the soft foreign matter defect is large enough to make the battery pack to be detected unqualified; when the size of the hard foreign matter defect is larger than the corresponding foreign matter threshold information, the size of the hard foreign matter defect is large enough to make the battery pack to be detected unqualified.

[0122] In this embodiment, the foreign matter threshold information may also include quantity information, and the foreign matter threshold information may be used to determine whether the number of soft foreign matter defects or the number of hard foreign matter defects is large enough to make the battery package to be inspected unqualified.

[0123] It is understandable that the foreign matter threshold information may include both size information and quantity information. When determining whether the battery package to be inspected is qualified, the size and quantity of the soft foreign matter defect or the hard foreign matter defect are comprehensively considered.

[0124] Among them, the foreign matter threshold information corresponding to the soft foreign matter defect and the foreign matter threshold information corresponding to the hard foreign matter defect can be the same or different.

[0125] In this embodiment, when it is determined based on the second detection information that there is a soft foreign matter defect in the battery package to be inspected, the information of the soft foreign matter defect is compared with the foreign matter threshold information corresponding to the soft foreign matter defect to determine the defect detection result, that is, to determine whether the soft foreign matter defect is large enough or the number is large enough to make the battery package to be inspected unqualified.

[0126] In this embodiment, when it is determined based on the second detection information that there is a hard foreign matter defect in the battery package to be inspected, the information of the hard foreign matter defect is compared with the foreign matter threshold information corresponding to the hard foreign matter defect to determine the defect detection result, that is, to judge whether the hard foreign matter defect is large enough or the number is large enough to make the battery package to be inspected unqualified.

[0127] In some embodiments, the second detection model includes a first detection sub-model and a second detection sub-model, and the first split image, the second split image, and the first composite image are input into the second detection model, and the foreign body defect is judged to be a hard foreign body defect or a soft foreign body defect, and the second detection information of the battery package to be detected output by the second detection model is obtained, including:

[0128] Inputting the first split image and the second split image into the first detection sub-model to detect foreign matter defects, and obtaining a first defect area of ​​the battery package to be detected output by the first detection sub-model;

[0129] Inputting the first synthetic image into the second detection sub-model to detect foreign matter defects, and obtaining a second defect area of ​​the battery package to be detected output by the second detection sub-model;

[0130] Based on the first defect area and the second defect area, second detection information is determined.

[0131] The first detection sub-model is a model that can identify foreign body defects in the second detection model and can identify the area where the foreign body defects are located.

[0132] In this embodiment, a plurality of battery package film sample images may be acquired, and each battery package film sample image may be split to obtain a first split sample image and a second split sample image corresponding to the battery package film sample image.

[0133] The foreign matter defects in the first split sample image can be marked to obtain a corresponding foreign matter sample image; the foreign matter defects in the second split sample image can be marked to obtain a corresponding foreign matter sample image.

[0134] In this embodiment, the first detection sub-model can be trained by using a plurality of foreign matter sample images corresponding to the first split sample images and a plurality of foreign matter sample images corresponding to the second split sample images, so that the first detection sub-model can recognize the input first split image and the second split image, detect foreign matter defects, and obtain the first defect area of ​​the battery package to be detected.

[0135] Among them, the first defect area is the area where the foreign matter defect in the battery package to be detected is located.

[0136] In this embodiment, the first detection sub-model can perform reasoning through the semantic segmentation algorithm of the deep learning algorithm to capture information such as the location and contour of the foreign body defect and obtain the first defect area.

[0137] In this embodiment, the second detection sub-model is a model in the second detection model that can identify hard foreign body defects.

[0138] In this embodiment, multiple battery package sample images can be obtained, and each battery package sample image can be split to obtain a first split sample image and a second split sample image corresponding to the battery package sample image, and based on the first split sample image and the second split sample image, a first composite sample image can be obtained.

[0139] The foreign matter defects in the first synthetic sample image may be marked to obtain a corresponding foreign matter sample image.

[0140] In this embodiment, the second detection sub-model can be trained using multiple foreign body sample images corresponding to the first synthetic sample images, so that the second detection sub-model can recognize the input first synthetic image, detect foreign body defects, and obtain the second defect area of ​​the battery package to be detected.

[0141] Among them, the second defect area is the area where the hard foreign matter defect in the battery package to be detected is located.

[0142] In this embodiment, it is possible to determine whether the foreign matter defect is a hard foreign matter defect or a soft foreign matter defect based on the positional relationship between the first defect region and the second defect region in the battery package to be detected, and obtain second detection information.

[0143] In some embodiments, determining second detection information based on the first defect area and the second defect area includes:

[0144] In the case where the first defect region overlaps with the second defect region, determining that the foreign matter defect corresponding to the first defect region is a hard foreign matter defect;

[0145] Alternatively, when the first defect region does not overlap with the second defect region, it is determined that the foreign matter defect corresponding to the first defect region is a soft foreign matter defect.

[0146] In this embodiment, the characteristics of the soft foreign matter defect in the first synthetic image are not obvious, and the characteristics of the soft foreign matter defect in the first synthetic image are relatively obvious. When the foreign matter defect is a soft foreign matter defect, the second detection sub-model cannot identify the existence of the foreign matter defect in the first synthetic image. When the foreign matter defect is a hard foreign matter defect, the second detection sub-model can identify the existence of the foreign matter defect in the first synthetic image.

[0147] In this embodiment, the defect corresponding to the first defect area of ​​the battery package to be inspected may be a soft foreign matter defect or a hard foreign matter defect, and the defect corresponding to the second defect area of ​​the battery package to be inspected is a hard foreign matter defect. When the defect corresponding to the first defect area is a hard foreign matter defect, the first defect area and the second defect area may overlap. When the defect corresponding to the first defect area is a soft foreign matter defect, the second defect area cannot be detected in the first defect area.

[0148] In this embodiment, the first detection sub-model can simultaneously identify hard foreign body defects and soft foreign body defects. The foreign body defects corresponding to the first defect area can be hard foreign body defects or soft foreign body defects. The second detection sub-model can identify hard foreign body defects. The foreign body defects corresponding to the second defect area can be hard foreign body defects.

[0149] The type of foreign matter defect corresponding to the first defect area can be distinguished by the positional relationship between the second defect area and the first defect area. When the first defect area overlaps with the second defect area, that is, the second defect area is detected in the first defect area at the same time, the foreign matter defect corresponding to the first defect area is a hard foreign matter defect.

[0150] When the first defect region does not overlap with the second defect region, that is, the second defect region is not detected in the first defect region, the foreign matter defect corresponding to the first defect region is a soft foreign matter defect.

[0151] In some embodiments, the first detection model includes a third detection sub-model and a fourth detection sub-model, and the first split image, the second split image, and the first composite image are input into the first detection model to detect whether the battery package to be detected has a foreign matter defect or an air bubble defect, and obtain the first detection information of the battery package to be detected output by the first detection model, including:

[0152] Input the first split image and the second split image into the fourth detection sub-model to detect whether the battery package to be detected has a foreign matter defect or an air bubble defect, and obtain fourth detection information of the battery package to be detected output by the fourth detection sub-model;

[0153] The first synthetic image is input into the third detection sub-model to detect whether the battery package to be inspected has foreign matter defects or air bubble defects, and the fifth detection information of the battery package to be inspected output by the third detection sub-model is obtained.

[0154] In this embodiment, the fourth detection information is used to characterize whether the battery pack to be inspected has defects and the type of defects, the fifth detection information is used to characterize whether the battery pack to be inspected has defects and the type of defects, and the first detection information includes the fourth detection information and the fifth detection information.

[0155] Among them, the fourth detection sub-model is a model in the first detection model that can detect whether the battery package to be inspected has no defects, air bubble defects or foreign body defects. The fourth detection sub-model can also identify the location, contour and size of the defects.

[0156] In this embodiment, the foreign matter defects and air bubble defects in the first split sample image can be marked separately to obtain the corresponding bubble foreign matter sample image, and the foreign matter defects and air bubble defects in the second split sample image can be marked separately to obtain the corresponding bubble foreign matter sample image.

[0157] In this embodiment, the fourth detection sub-model can be trained using multiple bubble foreign matter sample images corresponding to the first split sample images and multiple bubble foreign matter sample images corresponding to the second split sample images, so that the fourth detection sub-model can identify defects in the first split image and the second split image and information corresponding to the defects, i.e., the fourth detection information.

[0158] In this embodiment, the third detection sub-model is a model in the first detection model that can detect whether the battery package to be inspected has no defects, air bubble defects or foreign matter defects. The third detection sub-model can also identify the location, contour and size of the defect.

[0159] In this embodiment, multiple battery package sample images can be obtained, and each battery package sample image can be split to obtain a first split sample image and a second split sample image corresponding to the battery package sample image, and based on the first split sample image and the second split sample image, a first composite sample image can be obtained.

[0160] The foreign matter defects and air bubble defects in the first synthetic sample image can be marked respectively to obtain the corresponding air bubble foreign matter sample image.

[0161] In this embodiment, the third detection sub-model can be trained using multiple bubble foreign matter sample images corresponding to the first synthetic sample images, so that the third detection sub-model can identify defects in the first synthetic image and information corresponding to the defects, namely, the fifth detection information.

[0162] In this embodiment, based on the first detection information, determining the defect detection result of the battery package to be detected includes:

[0163] Based on the fourth detection information and the fifth detection information, a defect detection result of the battery package to be detected is determined.

[0164] In this embodiment, the fourth detection information and the fifth detection information in the first detection information are combined to determine the defect detection result of the battery package to be detected.

[0165] In some embodiments, based on the fourth detection information and the fifth detection information, determining the defect detection result of the battery package to be detected includes:

[0166] When it is determined that the battery package to be inspected has an air bubble defect based on the fourth detection information and the fifth detection information, and it is determined that the size of the air bubble defect is smaller than the size corresponding to the first bubble threshold information based on the fourth detection information, the fifth detection information and the first bubble threshold information, the first synthetic image is input into the third detection sub-model to detect the air bubble defect, and the third detection information of the battery package to be inspected output by the third detection sub-model is obtained;

[0167] Based on the third detection information and the second bubble threshold information, a defect detection result of the battery package to be detected is determined.

[0168] The third detection information is used to characterize the size of the air bubble defect.

[0169] In this embodiment, the first bubble threshold information is pre-set, and the first bubble threshold information may include size information. The first bubble threshold information is used to determine whether the size of the air bubble defect is large enough to make the battery package to be inspected unqualified. When the size of the air bubble defect is larger than the first bubble threshold information, the size of the air bubble defect is large enough to make the battery package to be inspected unqualified.

[0170] In this embodiment, the first bubble threshold information may also include quantity information, and the first bubble threshold information may be used to determine whether the number of air bubble defects is large enough to make the battery package to be inspected unqualified.

[0171] It is understandable that the first bubble threshold information may include both size information and quantity information. When judging whether the battery package to be inspected is qualified, the size and quantity of the air bubble defect are comprehensively considered.

[0172] In this embodiment, when it is determined that the battery pack to be inspected has an air bubble defect based on the fourth detection information and the fifth detection information, the size of the air bubble defect can be compared with the size corresponding to the first bubble threshold information to determine the defect detection result, that is, to determine whether the air bubble defect is large enough to make the battery pack to be inspected unqualified.

[0173] In this embodiment, the size of the air bubble defect is compared with the first bubble threshold information. The size of the air bubble defect is smaller than the first bubble threshold information, indicating that the size of the air bubble defect is small. The first synthetic image is input into the third detection sub-model of the first detection model for further identification and judgment. The third detection sub-model detects the air bubble defect, obtains third detection information, and determines the size of the air bubble defect.

[0174] In this embodiment, the second bubble threshold information is pre-set, and the second bubble threshold information is used to determine whether the size of the air bubble defect is large enough to make the battery package to be inspected unqualified. When the size of the air bubble defect is larger than the second bubble threshold information, the size of the air bubble defect is large enough to make the battery package to be inspected unqualified.

[0175] In this embodiment, the size of the air bubble defect identified by the third detection sub-model is compared with the second bubble threshold information to determine the defect detection result, that is, to determine whether the air bubble defect is large enough to make the battery package to be detected unqualified.

[0176] It should be noted that, when it is determined based on the fourth detection information, the fifth detection information and the first bubble threshold information that the size of the air bubble defect is greater than the first bubble threshold information, it can be directly determined that the battery package to be inspected is unqualified and the corresponding defect detection result can be output.

[0177] In some embodiments, obtaining a membrane image of a battery membrane to be inspected includes:

[0178] The membrane image of the battery membrane to be inspected is obtained by time-sharing stroboscopic technology.

[0179] In this embodiment, a plurality of light sources 730 may be provided, and the plurality of light sources 730 may be controlled to emit light beams alternately at a first frequency, so that the surface features of the battery package to be inspected under different angles and lighting conditions may be obtained.

[0180] In this embodiment, multiple membrane images can be taken for a battery membrane to be inspected, each membrane image can include surface features of the battery membrane to be inspected under corresponding angles and lighting conditions, and defect detection is performed based on each membrane image. By combining the defect detection results corresponding to each membrane image, possible defects in the battery membrane to be inspected can be detected more comprehensively.

[0181] It should be noted that when acquiring sample images of the battery package, the multiple light sources 730 can also be controlled to alternately emit light beams to the battery package at a first frequency, so as to obtain sample images of the battery package under different angles and lighting conditions. After the sample images of the battery package are processed and labeled, the first detection model and the second detection model are trained to improve the accuracy of the first detection model and the second detection model.

[0182] In some embodiments, the first split image includes even-numbered rows of pixels in the envelope image, and the second split image includes odd-numbered rows of pixels in the envelope image.

[0183] In this embodiment, the capsule image is split into even-numbered rows of pixels and odd-numbered rows of pixels. According to the even-numbered rows of pixels, a first split image is obtained, i.e., the first split image includes the even-numbered rows of pixels of the capsule image. According to the odd-numbered rows of pixels, a second split image is obtained, i.e., the second split image includes the odd-numbered rows of pixels of the capsule image.

[0184] For example, the resolution of the envelope image O is X×Y, where Y represents the length of the envelope image, and X represents the width of the envelope image. The envelope image O includes Y rows and X columns of pixels, and Y may be an even number. When Y is an odd number, the last row of the envelope image O is deleted before processing.

[0185] The first split image A is a combination of all even-numbered rows of pixels in the envelope image O, and the first split image B is a combination of all odd-numbered rows of pixels in the envelope image O.

[0186] A(0...x-1,0...y / 2)=O(0...x-1,z)(z=0,2,4...y-2)

[0187] B(0...x-1,0...y / 2)=O(0...x-1,h)(h=1,3,5...y-1)

[0188] Among them, x represents the pixel in the xth column, and y represents the pixel in the yth row.

[0189] In this embodiment, the capsule image is split according to even-numbered rows of pixels and odd-numbered rows of pixels to obtain a first split image and a second split image. The first split image and the second split image can correspond to the imaging effects under different angles of illumination of the light source 730, and the brightness characteristics in the first split image and the second split image are different.

[0190] In this embodiment, pixels corresponding to the first split image and the second split image may be added or subtracted to determine the first composite image.

[0191] In this embodiment, pixels at corresponding positions of the first split image and the second split image are added or subtracted in a matrix form to obtain a first composite image.

[0192] For example, the first composite image C is the pixels of the first split image A minus the pixels of the second split image B at the corresponding position.

[0193] C(x,y)=A(x,y)-B(x,y)

[0194] In this embodiment, the brightness features in the first split image and the second split image are different, for example, Figure 7 As shown, two light sources 730 emit light to the battery package to be inspected on the left and right sides above the battery package to be inspected, and the emitted light is at a 45° angle to the plane where the battery package to be inspected is located. The image acquisition device 720 is set just above the battery package to be inspected to capture the package image, and the line of the package image can be perpendicular to the direction of the line connecting the two light sources 730.

[0195] The two light sources 730 are symmetrical on the left and right. The light and shadow characteristic effects of the first split image are opposite to those of the second split image. The shadow surface area of ​​the first split image corresponds to the bright surface area of ​​the second split image, and the bright surface area of ​​the first split image corresponds to the shadow surface area of ​​the second split image.

[0196] When the pixels of the first split image are subtracted from the pixels of the corresponding positions in the second split image, the bright surface with a larger grayscale value in the first split image is subtracted from the shadow surface with a smaller grayscale value in the second split image, and the shadow surface in the first split image is subtracted from the bright surface in the second split image. The grayscale value of the bright surface of the first composite image obtained is larger than that of the first split image, and the grayscale value of the dark surface of the first composite image is smaller than that of the first split image. The shadow feature of the first composite image is more prominent, thereby achieving the superposition effect of light and shadow features of two light sources 730 at different angles.

[0197] like Figure 2 As shown in, it is the first split image corresponding to the envelope image, such as Figure 3 As shown, this is a first composite image obtained based on the first split image, and the shadow features of the first composite image are more prominent.

[0198] A specific embodiment of a battery packaging detection method is given below.

[0199] like Figure 4 As shown, in step one, during the assembly process of the battery cell 710 of the power battery, an image of the battery cell 710 after being coated is captured, the battery cell 710 may be a square battery cell 710, a coating image of the battery coating to be detected is obtained, the coating image is preprocessed, the coating image is split, and a first split image and a second split image corresponding to the coating image are obtained, and a first composite image is determined based on the first split image and the second split image.

[0200] Step 2: Input the first split image, the second split image and the first composite image into the first detection model. When the third detection sub-model and the fourth detection sub-model of the first detection model are used to infer and judge whether there is a defect in the battery package to be inspected, the first detection model is used to judge whether the defect is a foreign body defect.

[0201] Step 3: When the defect is a foreign body defect, the first split image and the second split image are input into the first detection sub-model of the second detection model, and the first defect area of ​​the battery package to be detected is identified by the first detection sub-model.

[0202] Step 4: input the first composite image into the second detection sub-model of the second detection model, and identify the second defect area of ​​the battery package to be inspected through the second detection sub-model.

[0203] Step 5: When the first defect area overlaps with the second defect area, determine that the foreign matter defect corresponding to the first defect area is a hard foreign matter defect, compare the size and quantity of the hard foreign matter defect with the corresponding foreign matter threshold information, and determine whether the battery package to be inspected is qualified.

[0204] Step 6: When the first defect area does not overlap with the second defect area, determine that the foreign body defect corresponding to the first defect area is a soft foreign body defect, compare the size and quantity of the soft foreign body defect with the corresponding foreign body threshold information, and determine whether the battery package to be tested is qualified.

[0205] Step seven, after step two, if it is determined that the defect is not a foreign body defect, the defect is determined to be an air bubble defect, and the size of the air bubble defect is compared with the first bubble threshold information. When the size of the air bubble defect is larger than the size corresponding to the first bubble threshold information, it is determined that the battery package to be inspected is unqualified, including an air bubble defect.

[0206] Step 8. When the size of the air bubble defect is smaller than the size corresponding to the first bubble threshold information, the first synthetic image is input into the third detection sub-model of the first detection model, the size of the air bubble defect is identified again by the third detection sub-model, and compared with the size corresponding to the second bubble threshold information, and the number of air bubble defects is compared with the number corresponding to the second bubble threshold information to determine whether the battery package to be inspected is qualified.

[0207] In this embodiment, air bubble defects, hard foreign matter defects and soft foreign matter defects can be detected according to corresponding steps and models, and the air bubble defects, hard foreign matter defects and soft foreign matter defects are classified and counted according to size and quantity to obtain the final defect detection result.

[0208] Another specific embodiment of the battery packaging detection method is given below.

[0209] In this embodiment, by Figure 7 The battery pack inspection system shown is used to inspect the battery pack.

[0210] like Figure 5 As shown, in step 1, the coated battery cell 710, i.e., the battery coating to be inspected, is grabbed and placed at the inspection starting position, the image acquisition device 720 is triggered to take a photo and the image acquisition device 720 or the battery cell 710 is moved, and a plurality of light sources 730 at different angles are controlled to alternately emit light beams to the battery coating to be inspected at a first frequency to obtain a coating image of the battery coating to be inspected.

[0211] Step 2: Split the envelope image according to different light sources 730, that is, split it according to odd-numbered rows of pixels and even-numbered rows of pixels, obtain a first split image and a second split image under different angles of illumination, and synthesize the first split image and the second split image to obtain a first synthesized image. The first synthesized image can highlight the light and shadow characteristics of the concave and convex areas on the surface of the battery cell 710.

[0212] Step 3: Infer different defects in the first split image, the second split image and the first composite image to obtain defect detection results.

[0213] The following is a specific example of training the first detection model and the second detection model.

[0214] like Figure 6 As shown, step one, collect a certain number of automatically or artificially manufactured defective battery cells 710, including battery cells 710 with defects such as blue film air bubble defects in different forms, different types of metal foreign matter defects, sponge debris foreign matter defects, and plastic debris foreign matter defects, and collect a certain number of automatically or artificially manufactured non-defective battery cells 710.

[0215] Step 2: Capture images of each battery cell 710 collected in step 1 to obtain multiple battery package sample images, split each battery package sample image to obtain a first split sample image and a second split sample image corresponding to the battery package sample image, and obtain a first composite sample image based on the first split sample image and the second split sample image.

[0216] Step three, mark the foreign matter defects and air bubble defects in the first split sample image respectively to obtain the corresponding bubble foreign matter sample images, mark the air bubble defects and foreign matter defects in the second split sample image to obtain the corresponding bubble foreign matter sample images, and train the fourth detection sub-model in the first detection model through multiple bubble foreign matter sample images corresponding to the first split sample images and multiple bubble foreign matter sample images corresponding to the second split sample images. The fourth detection sub-model can be used to detect and distinguish air bubble defects and foreign matter defects.

[0217] Step 4: Label the foreign matter defects and air bubble defects in the first synthetic sample image respectively to obtain corresponding bubble foreign matter sample images, and train the third detection sub-model in the first detection model through multiple bubble foreign matter sample images corresponding to the first synthetic sample images. The third detection sub-model can be used to identify and distinguish air bubble defects and foreign matter defects.

[0218] Step 5. Label the foreign body defects in the first split sample image to obtain the corresponding foreign body sample image, label the foreign body defects in the second split sample image to obtain the corresponding foreign body sample image, and train the first detection sub-model in the second detection model through multiple foreign body sample images corresponding to the first split sample images and multiple foreign body sample images corresponding to the second split sample images. The first detection sub-model can be used to identify soft foreign body defects and hard foreign body defects.

[0219] Step six: mark the foreign body defects in the first synthetic sample image to obtain the corresponding foreign body sample image, and train the second detection sub-model in the second detection model through multiple foreign body sample images corresponding to the first synthetic sample image. The second detection sub-model can be used to identify hard foreign body defects.

[0220] In this embodiment, defects may be labeled using a deep learning algorithm.

[0221] In the related art, the method for detecting defects in the blue film of the battery cell includes using a pressure sensor to collect pressure data changes on the surface of the blue film during the process of applying pressure to the blue film of the battery cell to determine whether there is foreign matter and the type of foreign matter. However, the detection efficiency of this method is low.

[0222] The blue film image on the surface of the battery cell is collected through a two-dimensional camera, and the presence and size of foreign matter are determined by changes in the grayscale value of the image. However, this detection method is difficult and has a high misjudgment rate.

[0223] The blue film surface height data is collected by a three-dimensional camera, and the size of the foreign matter is determined by the height change of the blue film surface caused by the foreign matter. However, the cost of three-dimensional cameras is high and they cannot accurately classify defects.

[0224] In the embodiment of the present application, through time-sharing stroboscopic technology and image synthesis technology, the defect type and defect size of the battery package are accurately identified through the two-dimensional grayscale value information of the defect and the light and shadow characteristics at the three-dimensional height, combined with multiple detection models. The two-dimensional grayscale value information and the light and shadow characteristics at the three-dimensional height of the defect can be obtained without a three-dimensional camera. The two-dimensional grayscale value information and the three-dimensional light and shadow characteristics are combined and input into multiple trained detection models for defect detection, which reduces the process of processing complex data, improves the defect detection efficiency of the battery package, and has high detection accuracy, and can detect the type of defect.

[0225] The battery package film detection method provided in the embodiment of the present application can be executed by a battery package film detection device. In the embodiment of the present application, the battery package film detection method performed by the battery package film detection device is taken as an example to illustrate the battery package film detection device provided in the embodiment of the present application.

[0226] The embodiment of the present application also provides a battery packaging detection device.

[0227] like Figure 8 As shown, the battery package film detection device includes:

[0228] An acquisition module 810 is used to acquire a coating image of a battery coating to be inspected;

[0229] A first processing module 820 is used to split the capsule image to obtain a first split image and a second split image corresponding to the capsule image;

[0230] A second processing module 830 is used to determine a first composite image based on the first split image and the second split image;

[0231] A third processing module 840 is used to determine a defect detection result of the battery package to be detected based on the first split image, the second split image and the first composite image;

[0232] The first split image includes some pixels in the envelope image, the second split image includes some pixels in the envelope image, and the first split image and the second split image include different pixels.

[0233] According to the battery package detection device provided in the embodiment of the present application, by acquiring the package image of the battery package to be detected, a first split image, a second split image and a first composite image corresponding to the package image are obtained, and the defect detection result is determined according to the image recognition results of the first split image, the second split image and the first composite image. There is no need for human intervention to apply pressure to the battery package, etc., and the process of processing complex data is reduced, which can improve the defect detection efficiency of the battery package.

[0234] In some embodiments, the third processing module 840 is used to input the first split image, the second split image and the first composite image into the first detection model, detect whether the battery package to be detected has a foreign matter defect or an air bubble defect, and obtain first detection information of the battery package to be detected output by the first detection model, where the first detection information is used to characterize whether the battery package to be detected has a defect and the type of defect;

[0235] Determine a defect detection result of the battery package to be detected based on the first detection information;

[0236] The first detection model is trained through bubble foreign body sample images.

[0237] In some embodiments, the third processing module 840 is used to input the first split image, the second split image and the first composite image into the second detection model when it is determined that the battery package to be detected has a foreign body defect based on the first detection information, to determine whether the foreign body defect is a hard foreign body defect or a soft foreign body defect, and obtain second detection information of the battery package to be detected output by the second detection model, where the second detection information is used to characterize whether the battery package to be detected has a hard foreign body defect or a soft foreign body defect;

[0238] Determine a defect detection result of the battery package to be detected based on the second detection information and the foreign matter threshold information;

[0239] The second detection model is trained through foreign body sample images.

[0240] In some embodiments, the second detection model includes a first detection sub-model and a second detection sub-model, and a third processing module 840 is used to input the first split image and the second split image into the first detection sub-model, detect foreign body defects, and obtain a first defect area of ​​the battery package to be detected output by the first detection sub-model;

[0241] Inputting the first synthetic image into the second detection sub-model to detect foreign matter defects, and obtaining a second defect area of ​​the battery package to be detected output by the second detection sub-model;

[0242] Determine the second detection information based on the first defect area and the second defect area. In some embodiments, the third processing module 840 is used to determine that the foreign matter defect corresponding to the first defect area is a hard foreign matter defect when the first defect area overlaps with the second defect area;

[0243] Alternatively, when the first defect region does not overlap with the second defect region, it is determined that the foreign matter defect corresponding to the first defect region is a soft foreign matter defect.

[0244] In some embodiments, the first detection model includes a third detection sub-model and a fourth detection sub-model, and the third processing module 840 is used to input the first split image and the second split image into the fourth detection sub-model, detect whether the battery package to be detected has a foreign matter defect or an air bubble defect, and obtain fourth detection information of the battery package to be detected output by the fourth detection sub-model, and the fourth detection information is used to characterize whether the battery package to be detected has a defect and the type of defect information;

[0245] Input the first synthetic image to the third detection sub-model to detect whether the battery package to be detected has a foreign matter defect or an air bubble defect, and obtain fifth detection information of the battery package to be detected output by the third detection sub-model, where the fifth detection information is used to characterize whether the battery package to be detected has a defect and the type of the defect;

[0246] The first detection information includes fourth detection information and fifth detection information;

[0247] The third processing module 840 is used to determine the defect detection result of the battery package to be detected based on the fourth detection information and the fifth detection information.

[0248] In some embodiments, the third processing module 840 is used to input the first synthetic image into the third detection sub-model to detect the air bubble defect, and obtain third detection information of the battery package to be detected output by the third detection sub-model, when it is determined that the battery package to be detected has an air bubble defect based on the fourth detection information and the fifth detection information, and when it is determined that the size of the air bubble defect is smaller than the size corresponding to the first bubble threshold information based on the fourth detection information, the fifth detection information and the first bubble threshold information, and obtain the third detection information of the battery package to be detected output by the third detection sub-model, and the third detection information is used to characterize the size of the air bubble defect;

[0249] Based on the third detection information and the second bubble threshold information, a defect detection result of the battery package to be detected is determined.

[0250] In some embodiments, the acquisition module 810 is used to acquire the membrane image of the battery membrane to be inspected by using time-sharing stroboscopic technology.

[0251] In some embodiments, the first split image includes even-numbered rows of pixels in the envelope image, and the second split image includes odd-numbered rows of pixels in the envelope image.

[0252] The battery pack detection device in the embodiment of the present application can be an electronic device, or a component in the electronic device, such as an integrated circuit or a chip. The electronic device can be a terminal, or other devices other than a terminal. Exemplarily, the electronic device can be a notebook computer, an ultra-mobile personal computer (UMPC), etc., or a server, a personal computer (PC), etc., which is not specifically limited in the embodiment of the present application.

[0253] The battery pack detection device in the embodiment of the present application may be a device having an operating system. The operating system may be an Android operating system, an IOS operating system, or other possible operating systems, which are not specifically limited in the embodiment of the present application.

[0254] The battery pack detection device provided in the embodiment of the present application can achieve Figure 1 and Figure 4-Figure 6 To avoid repetition, the various processes implemented by the method embodiment are not described here.

[0255] The embodiment of the present application also provides a battery packaging detection system.

[0256] The battery package film detection system includes an image acquisition device 720 and the above-mentioned battery package film detection device, and the detection device is connected to the image acquisition device 720.

[0257] The image acquisition device 720 is used to acquire the packaging image of the battery packaging to be inspected.

[0258] According to the battery package inspection system provided in the embodiment of the present application, by acquiring the package image of the battery package to be inspected, a first split image, a second split image and a first composite image corresponding to the package image are obtained, and the defect detection result is determined according to the image recognition results of the first split image, the second split image and the first composite image. There is no need for human intervention to apply pressure to the battery package, etc., and the process of processing complex data is reduced, which can improve the defect detection efficiency of the battery package.

[0259] In some embodiments, the battery package detection system further includes a light source module.

[0260] In this embodiment, the light source module is used to emit a light beam to the battery package to be inspected by using a time-sharing stroboscopic technique.

[0261] For example, Figure 7 As shown, the light source module includes two light sources 730, and the two light sources 730 emit light to the battery package to be inspected on the left and right sides above the battery package to be inspected, and the emitted light is at a 45° angle to the plane where the battery package to be inspected is located. The image acquisition device 720 is arranged directly above the battery package to be inspected to capture the package image, and the two light sources 730 can alternately emit light beams to the battery package to be inspected at a first frequency.

[0262] In this embodiment, the image acquisition device 720 may be a line scan camera with a lens, the light source 730 may be a line light source 730, and the battery package detection system may further include a stroboscopic light source controller and a processor.

[0263] Among them, the line scan camera is used to collect the coating image of the blue film related area of ​​the battery cell 710 during the preparation process of the battery cell 710, the light source 730 is used to provide a stable lighting environment during the process of collecting the coating image, and the stroboscopic light source controller controls different light sources 730 to periodically provide lighting at different angles during the camera collecting the coating image. The processor is used to split and synthesize the collected coating images, and establish a deep learning model, namely a first detection model and a second detection model, to infer and determine the defects on the coating of the battery cell 710, and the battery coating detection device can be integrated in the processor.

[0264] In some embodiments, Fig. 9As shown, an embodiment of the present application further provides an electronic device 900, including a processor 901, a memory 902, and a computer program stored in the memory 902 and executable on the processor 901. When the program is executed by the processor 901, each process of the above-mentioned battery package detection method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0265] It should be noted that the electronic devices in the embodiments of the present application include the mobile electronic devices and non-mobile electronic devices mentioned above.

[0266] The embodiment of the present application also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, each process of the above-mentioned battery package detection method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0267] The processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk.

[0268] An embodiment of the present application also provides a computer program product, including a computer program, which implements the above-mentioned battery package detection method when executed by a processor.

[0269] The processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk.

[0270] An embodiment of the present application further provides a chip, which includes a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the various processes of the above-mentioned battery package detection method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0271] It should be understood that the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

[0272] It should be noted that, in this article, the terms "comprise", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise one..." do not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

[0273] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM / RAM, a disk, or an optical disk), and includes a number of instructions for a terminal (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0274] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.

[0275] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0276] Although the embodiments of the present application have been shown and described, those skilled in the art will appreciate that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present application, and that the scope of the present application is defined by the claims and their equivalents.

Claims

1. A method for detecting a battery envelope, characterized in that: include: Acquire a membrane image of a battery membrane to be inspected; Splitting the envelope image to obtain a first split image and a second split image corresponding to the envelope image; Determine a first composite image based on the first split image and the second split image; Determine a defect detection result of the battery package to be inspected based on the first split image, the second split image and the first composite image; The first split image includes some pixels in the envelope image, the second split image includes some pixels in the envelope image, and the first split image and the second split image include different pixels.

2. The method for detecting a battery package according to claim 1, characterized in that: The step of determining the defect detection result of the battery package to be detected based on the first split image, the second split image and the first composite image includes: Inputting the first split image, the second split image and the first composite image into a first detection model, detecting whether the battery package to be detected has a foreign matter defect or an air bubble defect, and obtaining first detection information of the battery package to be detected output by the first detection model, wherein the first detection information is used to characterize whether the battery package to be detected has a defect and the type of the defect; Determine a defect detection result of the battery package to be detected based on the first detection information; The first detection model is obtained by training with bubble foreign body sample images.

3. The method for detecting a battery package according to claim 2, characterized in that: The step of determining the defect detection result of the battery package to be detected based on the first detection information includes: In the case where it is determined based on the first detection information that the battery package to be detected has the foreign matter defect, the first split image, the second split image, and the first composite image are input into a second detection model to determine whether the foreign matter defect is a hard foreign matter defect or a soft foreign matter defect, and second detection information of the battery package to be detected output by the second detection model is obtained, where the second detection information is used to characterize whether the battery package to be detected has the hard foreign matter defect or the soft foreign matter defect; Determine a defect detection result of the battery package to be detected based on the second detection information and the foreign matter threshold information; The second detection model is obtained by training with foreign body sample images.

4. The method for detecting a battery package according to claim 3, characterized in that: The second detection model includes a first detection sub-model and a second detection sub-model, and the first split image, the second split image and the first composite image are input into the second detection model, and the foreign body defect is judged to be a hard foreign body defect or a soft foreign body defect, and the second detection information of the battery package to be detected output by the second detection model is obtained, including: Inputting the first split image and the second split image into the first detection sub-model, detecting the foreign body defect, and obtaining a first defect area of ​​the battery package to be detected output by the first detection sub-model; Inputting the first synthetic image into the second detection sub-model, detecting the foreign body defect, and obtaining a second defect region of the battery package to be detected output by the second detection sub-model; The second detection information is determined based on the first defect area and the second defect area.

5. The method for detecting a battery package according to claim 4, characterized in that: The determining the second detection information based on the first defect area and the second defect area includes: In a case where the first defect region overlaps with the second defect region, determining that the foreign matter defect corresponding to the first defect region is the hard foreign matter defect; Alternatively, when the first defect region does not overlap with the second defect region, the foreign matter defect corresponding to the first defect region is determined to be the soft foreign matter defect.

6. The method for detecting a battery package according to claim 2, characterized in that: The first detection model includes a third detection sub-model and a fourth detection sub-model. The first split image, the second split image, and the first composite image are input into the first detection model to detect whether the battery package to be detected has a foreign matter defect or an air bubble defect, and first detection information of the battery package to be detected output by the first detection model is obtained, including: Inputting the first split image and the second split image into the fourth detection sub-model, detecting whether the battery package to be detected has the foreign matter defect or the air bubble defect, and obtaining fourth detection information of the battery package to be detected output by the fourth detection sub-model, wherein the fourth detection information is used to characterize whether the battery package to be detected has defects and the type of defects; Inputting the first synthetic image into the third detection sub-model, detecting whether the battery package to be inspected has the foreign matter defect or the air bubble defect, and obtaining fifth detection information of the battery package to be inspected output by the third detection sub-model, wherein the fifth detection information is used to characterize whether the battery package to be inspected has defects and the type of defects; The first detection information includes the fourth detection information and the fifth detection information; The step of determining the defect detection result of the battery package to be detected based on the first detection information includes: Based on the fourth detection information and the fifth detection information, a defect detection result of the battery package to be detected is determined.

7. The method for detecting a battery package according to claim 6, characterized in that: The step of determining the defect detection result of the battery package film to be detected based on the fourth detection information and the fifth detection information includes: In a case where it is determined that the battery package to be inspected has the air bubble defect based on the fourth detection information and the fifth detection information, and it is determined that the size of the air bubble defect is smaller than the size corresponding to the first bubble threshold information based on the fourth detection information, the fifth detection information and the first bubble threshold information, the first synthetic image is input into the third detection sub-model to detect the air bubble defect, and the third detection information of the battery package to be inspected output by the third detection sub-model is obtained, where the third detection information is used to characterize the size of the air bubble defect; Based on the third detection information and the second bubble threshold information, a defect detection result of the battery package to be detected is determined.

8. The method for detecting a battery package membrane according to any one of claims 1 to 7, characterized in that: The step of obtaining a coating image of a battery coating to be inspected includes: The membrane image of the battery membrane to be inspected is acquired by time-sharing stroboscopic technology.

9. The method for detecting a battery package membrane according to any one of claims 1 to 7, characterized in that: The first split image includes even-numbered rows of pixels in the envelope image, and the second split image includes odd-numbered rows of pixels in the envelope image.

10. A battery film detection device, characterized in that: include: An acquisition module, used for acquiring a membrane image of a battery membrane to be inspected; A first processing module, used for splitting the capsule image to obtain a first split image and a second split image corresponding to the capsule image; A second processing module, configured to determine a first composite image based on the first split image and the second split image; A third processing module, configured to determine a defect detection result of the battery package to be inspected based on the first split image, the second split image and the first composite image; The first split image includes some pixels in the envelope image, the second split image includes some pixels in the envelope image, and the first split image and the second split image include different pixels.

11. A battery film detection system, characterized in that: include: An image acquisition device, wherein the image acquisition device is used to acquire an image of a battery package to be inspected; The battery pack detection device as claimed in claim 10, wherein the detection device is connected to the image acquisition device.

12. The battery package film detection system according to claim 11, characterized in that: Also includes: A light source module is used to emit a light beam to the battery package to be inspected by using a time-sharing stroboscopic technique.

Images