Parameter adjustment method and apparatus for defect detection, and device and storage medium

Abstract

The embodiments of the present application relate to the technical field of machine vision. Provided are a parameter adjustment method and apparatus for defect detection, and a device and a storage medium. The method comprises: in response to a parameter adjustment operation on an initial defect type in a defect display interface, displaying a defect parameter adjustment interface corresponding to the initial defect type, wherein the defect parameter adjustment interface comprises: a plurality of feature values corresponding to a plurality of defect features of an abnormal region corresponding to the initial defect type, a defect selection control, and a data learning control; in response to a trigger operation on the data learning control, acquiring a plurality of initial defect ground truth ranges corresponding to a target defect type in the defect selection control; on the basis of the plurality of feature values of the abnormal region corresponding to the initial defect type, adjusting the plurality of initial defect ground truth ranges corresponding to the target defect type, in order to obtain a plurality of target defect ground truth ranges corresponding to the target defect type; and on the basis of the plurality of target defect ground truth ranges, updating a defect determination condition corresponding to the target defect type.

Description

A defect detection parameter tuning method, apparatus, equipment, and storage medium

[0001] This application claims priority to Chinese Patent Application No. CN202411889806.0, filed on December 19, 2024, entitled "A Defect Detection Parameter Adjustment Method, Apparatus, Device and Storage Medium", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of machine vision technology, and in particular to a defect detection parameter tuning method, apparatus, device, and storage medium. Background Technology

[0003] During the production process, various defects may occur in products or components due to various reasons (such as material problems, process fluctuations, equipment failures, etc.). If these defects are not detected and repaired in a timely manner, they will adversely affect the performance, reliability, and safety of the product. Therefore, it is necessary to collect images of products or components and then perform defect detection on these images to promptly identify and eliminate problems in the products or components, thereby ensuring that the quality of the products or components meets the predetermined standards and requirements.

[0004] When conducting defect detection, it is necessary to set a defect judgment condition for each defect type. The presence of a defect in a product or component is determined by judging whether the image of the product or component meets any defect judgment condition. However, the setting of defect judgment conditions depends on the experience of the staff, and is therefore easily affected by personal subjective consciousness, personal experience and understanding, resulting in low accuracy of defect judgment conditions. Therefore, during the defect detection process, it is necessary to continuously adjust the parameters in the defect judgment conditions to make the judgment more accurate. Summary of the Invention

[0005] This application provides a defect detection parameter adjustment method, apparatus, device, and storage medium, which provides a way to adjust the parameters in the defect judgment conditions during the defect detection process to improve the accuracy of defect detection.

[0006] The first aspect of this application provides a defect detection parameter tuning method, comprising: in response to a parameter tuning operation on an initial defect type in a defect display interface, displaying a defect parameter tuning interface corresponding to the initial defect type; the defect display interface displays an initial defect type obtained by detecting an image to be detected; the defect parameter tuning interface includes: multiple feature values ​​corresponding to multiple defect features of an abnormal region corresponding to the initial defect type, a defect selection control, and a data learning control; in response to a trigger operation on the data learning control, obtaining multiple initial defect truth value ranges corresponding to a target defect type in the defect selection control, and adjusting the multiple initial defect truth value ranges corresponding to the target defect type based on multiple feature values ​​of an abnormal region corresponding to the initial defect type to obtain multiple target defect truth value ranges corresponding to the target defect type; wherein, the target defect truth value range corresponding to each defect feature matches the feature value corresponding to the defect feature; and updating the defect judgment conditions corresponding to the target defect type based on the multiple target defect truth value ranges.

[0007] The defect detection parameter tuning method provided in this application allows, after obtaining the defect detection results, to display the initial defect type obtained from the detection of the image under test through a defect display interface. This facilitates the staff's review of whether the detected initial defect type matches the actual defect type. If they do not match, the staff can trigger the parameter tuning control for the initial defect type in the defect display interface (triggering a parameter tuning operation for the initial defect type in the defect display interface). This displays the corresponding defect parameter tuning interface, allowing the staff to select the target defect type corresponding to the initial defect type based on the defect selection control in the defect tuning interface. The data learning control in the defect tuning interface automatically adjusts the initial defect truth value range corresponding to multiple defect features of the target defect type, obtaining the target defect truth value range that matches the feature value of each defect feature. In this way, when re-detecting defects, the feature value of the abnormal region can match the defect judgment condition corresponding to the target defect type, meaning the defect type of the abnormal region is detected as the target defect type, thus achieving a more accurate defect detection.

[0008] In conjunction with the first implementation of the first aspect, the true value ranges of multiple initial defects corresponding to the target defect type are adjusted based on multiple feature values ​​of the abnormal region corresponding to the initial defect type to obtain multiple true value ranges of target defects corresponding to the target defect type. This includes: inputting multiple feature values ​​of the abnormal region corresponding to the initial defect type and multiple true value ranges of initial defects corresponding to the target defect type into the target parameter prediction model corresponding to the target defect type to obtain multiple true value ranges of target defects corresponding to the target defect type; the target parameter prediction model is used to adjust the true value ranges of multiple initial defects based on multiple feature values ​​to obtain multiple true value ranges of target defects.

[0009] In conjunction with the second implementation method of the first aspect, the target parameter prediction model is trained in the following way: multiple sample data corresponding to the target defect type are obtained, wherein each sample data includes the historical feature value, the initial historical defect true value range and the target historical defect true value range corresponding to each defect feature among multiple defect features; based on the multiple sample data corresponding to the target defect type, the initial parameter prediction model is trained to obtain the target parameter prediction model corresponding to the target defect type.

[0010] In conjunction with the third implementation method of the first aspect, after updating the defect judgment conditions corresponding to the target defect type, the method further includes: obtaining multiple historical defect judgment conditions corresponding to the target defect type; determining the confidence level of the updated defect judgment conditions corresponding to the target defect type based on the multiple historical defect judgment conditions; and outputting a prompt message when the confidence level is less than or equal to a preset confidence threshold. The prompt message is used to prompt staff to determine whether the updated defect judgment conditions are reasonable.

[0011] In conjunction with the fourth implementation method of the first aspect, the method further includes: determining the abnormal region in the image to be detected, and determining the feature values ​​of multiple defect features corresponding to the abnormal region; matching the defect judgment conditions corresponding to the defect type with the feature values ​​of multiple defect features corresponding to the abnormal region according to the defect type order recorded in the defect priority, so as to determine the defect type obtained by detecting the image to be detected; and displaying the defect type obtained by detecting the image to be detected through the defect display interface.

[0012] In conjunction with the fifth implementation method of the first aspect, the defect parameter tuning interface also includes: the feature status corresponding to each defect feature; the feature status is used to indicate whether the feature value corresponding to the defect feature matches the initial defect truth value range; and / or, the display control and adjustment control corresponding to each initial defect truth value range.

[0013] In conjunction with the sixth implementation method of the first aspect, the defect display interface also includes: an image to be detected and a display box for the abnormal region corresponding to the defect type obtained by detecting the image to be detected; and / or, a filtered defect corresponding to the image to be detected, wherein the filtered defect is used to represent the defect of the abnormal region in the image to be detected where the feature values ​​of multiple defect features of the abnormal region do not match the multiple defect judgment conditions corresponding to multiple defect types.

[0014] The second aspect of this application provides a defect detection parameter tuning method, including:

[0015] The interface for adjusting parameters corresponding to the initial defect type is displayed. The initial defect type is the defect type obtained from the preliminary detection of the image to be detected. The interface for adjusting parameters includes a defect selection control and a data learning control. The defect selection control is used to select the actual target defect type corresponding to the image to be detected.

[0016] In response to a trigger operation on the data learning control, the initial defect truth value range corresponding to the target defect type is adjusted based on the feature values ​​detected for the initial defect type to obtain the target defect truth value range; wherein the target defect truth value range matches the feature values.

[0017] The defect detection parameter tuning method provided in the second aspect can directly display the defect tuning interface corresponding to the initial defect type. Operators can select the target defect type corresponding to the initial defect type based on the defect selection controls in the interface. The data learning controls in the interface automatically adjust the initial defect truth range corresponding to the target defect type to obtain the target defect truth range. In this way, when defect detection is performed again, the target defect truth range (i.e., the defect judgment condition for the target defect type) matches the feature values ​​detected for the initial defect type (i.e., the feature values ​​detected for the abnormal areas of the initial defect type), thus achieving a more accurate defect detection.

[0018] A third aspect of this application provides a defect detection parameter tuning device, comprising: a display module configured to display a defect parameter tuning interface corresponding to the initial defect type in response to a parameter tuning operation on an initial defect type in a defect display interface; the defect display interface displays the initial defect type obtained by detecting an image to be detected; the defect parameter tuning interface includes: multiple feature values ​​corresponding to multiple defect features of an abnormal region corresponding to the initial defect type, a defect selection control, and a data learning control; an adjustment module configured to obtain multiple initial defect truth value ranges corresponding to the target defect type in response to a trigger operation on the data learning control, and adjust the multiple initial defect truth value ranges corresponding to the target defect type based on multiple feature values ​​of an abnormal region corresponding to the initial defect type to obtain multiple target defect truth value ranges corresponding to the target defect type; wherein the target defect truth value range corresponding to each defect feature matches the feature value corresponding to the defect feature; and an update module configured to update the defect judgment conditions corresponding to the target defect type based on the multiple target defect truth value ranges.

[0019] A fourth aspect of this application provides a defect detection parameter tuning device, comprising:

[0020] The first display module is set to display the defect parameter tuning interface corresponding to the initial defect type; wherein, the initial defect type is the defect type obtained by preliminary detection of the image to be detected; the defect parameter tuning interface includes a defect selection control and a data learning control; the defect selection control is used to select the actual target defect type corresponding to the image to be detected;

[0021] The target adjustment module is configured to respond to a trigger operation of the data learning control, and adjust the initial defect truth value range corresponding to the target defect type based on the feature values ​​detected for the initial defect type to obtain the target defect truth value range; wherein the target defect truth value range matches the feature values.

[0022] A fifth aspect of this application provides an electronic device, including: one or more processors; one or more memories; wherein the one or more memories are used to store computer program code, the computer program code including computer instructions, and when the one or more processors execute the computer instructions, the electronic device performs the defect detection parameter tuning method provided in the first aspect and its possible implementations, or performs the defect detection parameter tuning method provided in the second aspect.

[0023] The sixth aspect of this application provides a computer-readable storage medium storing computer-executable instructions. When the computer-executable instructions are executed on the computer, the computer executes the defect detection parameter tuning method provided in the first aspect and its possible implementations, or executes the defect detection parameter tuning method provided in the second aspect.

[0024] A seventh aspect of this application provides a computer program product, including a computer program / instruction that, when executed by a processor, implements the defect detection parameter tuning method provided by the first aspect and its possible implementations, or executes the defect detection parameter tuning method provided by the second aspect.

[0025] The eighth aspect of this application provides a computer program that, when run on a computer, causes the computer to execute the defect detection parameter tuning method provided in the first aspect and its possible implementations, or to execute the defect detection parameter tuning method provided in the second aspect.

[0026] The beneficial effects of the solutions provided in aspects three through eight can be found in the analysis of the beneficial effects in aspect one or two, and will not be elaborated here. Attached Figure Description

[0027] The accompanying drawings are used to provide a further understanding of the technical solutions of this application and constitute a part of the specification. They are used together with the embodiments of this application to explain the technical solutions of this application and do not constitute a limitation on the technical solutions of this application.

[0028] Figure 1 is a schematic diagram of a defect detection parameter tuning system provided in an embodiment of this application;

[0029] Figure 2 is a schematic diagram of a defect display interface provided in an embodiment of this application;

[0030] Figure 3 is a flowchart of a defect detection parameter tuning method provided in an embodiment of this application;

[0031] Figure 4 is a schematic diagram of a defect parameter tuning interface provided in an embodiment of this application;

[0032] Figure 5 is a second flowchart of a defect detection parameter tuning method provided in an embodiment of this application;

[0033] Figure 6 is a flowchart of a defect detection parameter tuning method provided in an embodiment of this application.

[0034] Figure 7 is a schematic diagram of a defect detection parameter tuning device provided in an embodiment of this application;

[0035] Figure 8 is a schematic diagram of the structure of an electronic device provided in an embodiment of this application;

[0036] Figure 9 is a flowchart of a defect detection parameter tuning method provided in an embodiment of this application;

[0037] Figure 10 is a schematic diagram of another defect detection parameter adjustment device provided in an embodiment of this application. Detailed Implementation

[0038] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0039] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0040] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "connected" and "linked" should be interpreted broadly, for example, as a fixed connection, a detachable connection, or an integral connection. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances. Furthermore, when describing pipelines, the terms "connected" and "linked" as used in this application have the meaning of establishing electrical connection. The specific meaning needs to be understood in conjunction with the context.

[0041] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0042] During the production process, various defects may occur in products or components due to various reasons (such as material problems, process fluctuations, equipment failures, etc.). If these defects are not detected and repaired in a timely manner, they will adversely affect the performance, reliability, and safety of the product. Therefore, it is necessary to collect images of products or components and then perform defect detection on these images to promptly identify and eliminate problems in the products or components, thereby ensuring that the quality of the products or components meets the predetermined standards and requirements.

[0043] When performing defect detection, the first step is to use hardware equipment such as industrial cameras, lenses, and light sources to acquire images of the product (or component) to be inspected. The images are then preprocessed to improve image quality, remove noise, and enhance defect features, resulting in a preprocessed image. Next, defect-related feature information is extracted from the preprocessed image. For example, features can be extracted as follows: edge detection algorithms such as the Canny operator and Sobel operator are used to extract image edge information; shape parameters of the defects, such as area, perimeter, and roundness, are extracted; and features such as the mean gray level, gray level variance, and gray level entropy of the defect area are extracted. Finally, based on the extracted feature information and pre-set defect judgment conditions, defects are identified and classified to determine whether a defect exists in the product to be inspected.

[0044] However, the setting of defect judgment criteria relies on the experience of the staff, and is therefore easily influenced by personal subjectivity, experience, and understanding, leading to low accuracy and a high likelihood of false detections. For example, an actual defect type A might be mistakenly detected as defect type B. Therefore, during defect detection, the parameters in the defect judgment criteria need to be continuously adjusted to ensure more accurate defect type detection.

[0045] Based on this, this application provides a defect detection parameter tuning method. After detecting the image to be detected based on the defect judgment conditions corresponding to each of multiple defect types, the initial defect type obtained from the detection of the image to be detected is displayed through a defect display interface. If the operator finds that the detected initial defect type is inconsistent with the actual defect type, they can trigger the parameter tuning operation of the initial defect type in the defect display interface to adjust the parameters.

[0046] In response to the triggering operation of the parameter tuning control for the initial defect type in the defect display interface, the defect parameter tuning interface corresponding to the initial defect type is displayed. The defect display interface shows the initial defect type obtained from the detection of the image to be inspected. The defect parameter tuning interface includes: multiple feature values ​​corresponding to multiple defect features of the abnormal region corresponding to the initial defect type, a defect selection control, and a data learning control. Then, based on the defect selection control in the defect parameter tuning interface, the operator can select the target defect type actually corresponding to the initial defect type, and select the data learning control for automatic parameter adjustment.

[0047] In response to a trigger operation on the data learning control, multiple initial defect truth value ranges corresponding to the target defect type in the defect selection control are obtained. Based on multiple feature values ​​of the abnormal region corresponding to the initial defect type, the multiple initial defect truth value ranges corresponding to the target defect type are adjusted to obtain multiple target defect truth value ranges corresponding to the target defect type. Among them, the target defect truth value range corresponding to each defect feature matches the feature value corresponding to the defect feature. Based on the multiple target defect truth value ranges, the defect judgment conditions corresponding to the target defect type are updated so that defect detection can be performed on the image to be detected again.

[0048] As can be seen, the defect detection parameter tuning method provided in this application, after obtaining the defect detection result, can display the initial defect type obtained from detecting the image to be detected through a defect display interface, so that staff can check whether the detected initial defect type is consistent with the actual defect type. Furthermore, if they are inconsistent, staff can trigger the parameter tuning control for the initial defect type in the defect display interface (triggering a parameter tuning operation for the initial defect type in the defect display interface) to display the defect parameter tuning interface corresponding to the initial defect type. This allows staff to select the target defect type actually corresponding to the initial defect type based on the defect selection control in the defect parameter tuning interface, and automatically adjust the initial defect truth value range corresponding to multiple defect features of the target defect type through the data learning control in the defect parameter tuning interface, obtaining the target defect truth value range corresponding to the feature value of each defect feature that matches the target defect type.

[0049] In this way, when performing defect detection again, the feature values ​​of the abnormal area can match the defect judgment conditions corresponding to the target defect type. This means the defect type in the abnormal area is detected as the target defect type, thus achieving more accurate defect detection. Instead of requiring staff to manually adjust the true range of the target defect based on their experience, this avoids errors and delays caused by human factors, thereby improving the accuracy and reliability of defect detection. Simultaneously, it reduces the time spent on manual operations and judgments during defect detection, improving detection efficiency.

[0050] It should be understood that, by way of example, the defect detection parameter tuning method provided in this application embodiment can be applied to the defect detection parameter tuning system shown in FIG1. ​​Referring to FIG1, the defect detection parameter tuning system 10 may include: an acquisition device 100 and a detection device 200. The acquisition device 100 and the detection device 200 can be connected via a wired network or a wireless network. In some embodiments, the acquisition device 100 may be a device with image acquisition capabilities, such as an industrial area scan camera or a line scan camera, used to acquire images of the product to be inspected and send them to the detection device 200 for defect detection.

[0051] In some embodiments, the detection device 200 may be equipped with a defect detection algorithm, which can determine whether there is a defect in the image of the product to be inspected (hereinafter referred to as the image to be inspected) by using the defect judgment conditions corresponding to each defect type, and display the defect types obtained by detecting the image to be inspected through a defect display interface; the defect types obtained by detecting the image to be inspected are the initial defect types obtained by detecting the image to be inspected based on the defect judgment conditions corresponding to each of the multiple defect types. For an image to be inspected, one or more defect types can be detected, that is, one or more initial defect types can be detected.

[0052] When staff determine that the actual defect type in the image to be inspected does not match the initial defect type detected in the defect display interface, they can adjust the defect judgment conditions of the actual defect type to make the defect judgment conditions of the actual defect type more accurate, so as to more accurately determine the actual defect type in the image to be inspected during defect detection.

[0053] Specifically, when the initial defect type of the detected abnormal area does not match the actual defect type in the defect display interface, the staff can select the parameter adjustment control of the initial defect type in the defect display interface, so that the detection equipment 200 displays the defect parameter adjustment interface corresponding to the initial defect type. The staff can determine the target defect type (actual defect type) corresponding to the abnormal area in the defect selection control in the defect parameter adjustment interface, and then select the data learning control.

[0054] The detection device 200 can acquire multiple initial defect truth value ranges corresponding to the target defect type in the defect selection control, and adjust these ranges based on multiple feature values ​​of the abnormal region corresponding to the initial defect type, thus obtaining multiple target defect truth value ranges corresponding to the target defect type. Specifically, the target defect truth value range corresponding to each defect feature matches the feature value corresponding to that defect feature. Furthermore, based on these multiple target defect truth value ranges, the defect judgment conditions corresponding to the target defect type are updated to re-detect the image.

[0055] In this way, when performing defect detection again, the feature value of the abnormal area can match the defect judgment condition corresponding to the target defect type, that is, the defect type of the abnormal area is detected as the target defect type, thereby achieving the purpose of more accurate defect detection.

[0056] It should be understood that the detection device 200 in the embodiments of this application can be an electronic device such as a personal computer (PC), laptop computer, mobile device, tablet computer, or laptop computer, and the embodiments of this application do not limit the specific form of the electronic device. Alternatively, the detection device 200 can also be a single server or a server cluster composed of multiple servers. In some implementations, the server cluster can be a distributed cluster server. The embodiments of this application do not impose any restrictions in this regard.

[0057] It should be understood that the defect detection parameter tuning method provided in this application embodiment can be applied to the detection device 200 in the defect detection parameter tuning system shown in FIG1. ​​In other embodiments, the defect detection parameter tuning method provided in this application embodiment can also be applied to other devices that communicate with the detection device and have deployed functional software that implements the defect detection parameter tuning method, which is also reasonable.

[0058] In some embodiments, the defect detection parameter tuning method provided in this application can apply the detection results (such as defect type) and intermediate data (such as feature values ​​of defect features) obtained during the detection process to achieve parameter tuning; wherein, when performing defect detection, it is done by the following method: based on the defect judgment conditions corresponding to each defect type among multiple defect types, the defect type corresponding to the abnormal area in the image to be detected is determined, and the detected defect type in the image to be detected is displayed through the defect display interface.

[0059] The defect judgment criteria include the truth value range of each defect feature among multiple defect features. That is, for each defect type, there can be a truth value range of each defect feature among multiple defect features. Furthermore, the multiple defect features involved in the defect judgment criteria for different defect types can be the same or different.

[0060] It is understood that the image to be detected can be an image of the product to be detected acquired by the acquisition device 100, or an image of the product to be detected pre-stored in the detection device 200 or other memory. The source of the image to be detected is not limited in this application embodiment.

[0061] It should be understood that the types of defects corresponding to different products in the image to be inspected will also differ. When the image to be inspected is an image of a lithium electrode sheet, i.e., when performing defect inspection on a lithium electrode sheet, multiple defect types can include: seams, missing foil, black spots, white spots, dark marks, scratches, etc. Each defect type corresponds to a defect judgment condition. During defect inspection, the feature values ​​of multiple defect features in the abnormal area of ​​the image to be inspected can be matched with multiple defect judgment conditions to obtain the defect type obtained from the inspection of the image.

[0062] For example, as a feasible implementation, multiple defect features of the abnormal region may include: area, rectangularity, shorter side, average gray level, and polarity; wherein, area can be considered as the area of ​​the abnormal region, rectangularity is used to measure the similarity between the abnormal region and a rectangle, the larger the rectangularity, the more similar the shape of the abnormal region is to a rectangle, and polarity represents the difference in charge properties between the positive and negative poles of the abnormal region; this application embodiment does not specifically limit this. During defect detection, the feature value of each defect feature among the multiple defect features of the abnormal region is determined, and then the multiple feature values ​​are matched with the defect judgment conditions corresponding to each defect type among the multiple defect types to determine the initial defect type of the abnormal region.

[0063] As a feasible approach, defect detection of the image to be detected can be performed through the following steps: identify abnormal regions in the image to be detected and determine the feature values ​​of multiple defect features corresponding to the abnormal regions; match the defect judgment conditions corresponding to the defect types with the feature values ​​of multiple defect features corresponding to the abnormal regions according to the defect type order recorded in the defect priority, so as to determine the defect type obtained by detecting the image to be detected.

[0064] In other words, during defect detection, abnormal regions in the image to be detected are first identified. Then, the feature values ​​of the defect features corresponding to each abnormal region are determined. Following the defect type order recorded in the defect priority table, multiple defect judgment conditions corresponding to multiple defect types are matched one-to-one with the feature values ​​of multiple defect features corresponding to the abnormal regions. If any defect judgment condition matches, that is, if the feature values ​​of multiple defect features corresponding to the abnormal region match the condition content of one of the defect judgment conditions, it indicates that a defect of the defect type corresponding to that defect judgment condition exists in the image to be detected. It can be understood that "any defect judgment condition matching" can mean that the feature values ​​of multiple defect features corresponding to the abnormal region are all within the true value range of the corresponding defect feature in that defect judgment condition, or it can mean that the feature value of at least one defect feature corresponding to the abnormal region is within the true value range of the corresponding defect feature in that defect judgment condition. This is also reasonable, and the "at least one defect feature" can be set according to the actual situation.

[0065] It is understandable that some defects are more important among the various defect detection types. For example, in the defect detection of lithium electrode sheets, defects such as bonding and foil leakage are more important than defects such as scratches and bright lines. Therefore, when performing defect detection, the order of defect types recorded in the defect priority list can be used to determine the defect detection of the image to be inspected.

[0066] The "leaky foil" problem refers to gaps between the metal foils on the battery's separator and the positive and negative electrodes, which can lead to electrolyte leakage or short circuits. Electrolyte leakage directly affects the battery's internal structure and chemical balance, while short circuits can cause serious safety issues such as overheating, fire, or even explosion. Therefore, the "leaky foil" problem poses a serious threat to the safety and stability of the battery. Poor or broken connections affect the integrity and continuity of the electrode sheets, thus impacting battery performance and lifespan. These defects directly relate to battery safety and performance; therefore, they must be strictly controlled during the production process. Once these defects are discovered, it may be necessary to immediately stop the production line for repairs, which will severely impact production efficiency and costs.

[0067] While defects such as scratches and bright lines can affect battery performance (e.g., reducing the contact area between the electrodes and the electrolyte, thus affecting discharge performance), these defects usually do not immediately lead to battery failure or safety incidents. Therefore, during the testing process, these defects may be considered minor issues, but they still require attention and measures to improve them.

[0068] In summary, in lithium electrode defect detection, defects such as bonding defects and foil leaks are more important than defects such as scratches and bright streaks. This is because they have a more direct and significant impact on battery performance, safety, production efficiency, and cost. Therefore, these defects should be given higher priority and stricter control standards during the inspection process. In other words, inspecting according to the defect type order listed in the defect priority list can identify more important defect types first, facilitating the handling of critical defects and ensuring the performance, safety, production efficiency, and cost of the inspected products.

[0069] The defect display interface includes the defect types obtained from the inspection of the images to be inspected, allowing staff to quickly identify the defects in the products corresponding to the images. Furthermore, as a feasible implementation, the defect display interface also includes parameter adjustment controls for each defect type, allowing staff to select and adjust the parameters for any defect type if they determine that it is not a genuine defect.

[0070] For example, please refer to Figure 2, which is a schematic diagram of a defect display interface provided in an embodiment of this application. As shown in Figure 2, when performing defect detection on the image A to be inspected, abnormal region 1 and abnormal region 2 are detected. Therefore, the defect type corresponding to each abnormal region is displayed through the defect display interface. For example, the defect type corresponding to abnormal region 1 represented by number 1 is white spot, and the defect type corresponding to abnormal region 2 represented by number 2 is classification filtering. To facilitate staff review, the defect display interface can also display the feature values ​​of the defect features of each abnormal region. For example, the defect features corresponding to abnormal region 1 are: area 225; long side 15.4; short side 15; grayscale... The defect features corresponding to abnormal region 2 are: area 0.19; long side 0.4; short side 0.5; grayscale...

[0071] In some embodiments, to facilitate staff in adjusting defect types, as shown in Figure 2, each defect type corresponds to a parameter adjustment control, namely the circular control under the parameter adjustment field shown in Figure 2, so that staff can perform parameter adjustment operations.

[0072] In some embodiments, to facilitate staff to check whether the defect type detected in the image to be inspected is a real defect, as shown in Figure 2, the defect display interface may also include the image to be inspected and a display box for the abnormal area corresponding to the defect type detected in the image to be inspected.

[0073] In this way, staff can better determine whether a defect type is a real defect and whether parameter adjustments are needed based on the defect type present in the image to be inspected and the corresponding display box in the image.

[0074] As a feasible implementation method, as shown in Figure 2, the defect display interface can also include a basic settings area, which is used to select the acquisition method of the image to be inspected (mode selection shown in Figure 2). The operator can choose to use images captured by a camera as the image to be inspected, or they can use local data (images stored locally) as the image to be inspected. Furthermore, the operator can select the camera and the output to facilitate defect inspection. Simultaneously, the operator can choose to execute the inspection once or continuously to inspect a single image or multiple images simultaneously.

[0075] Understandably, during defect detection, the setting of defect judgment criteria relies on the experience of the staff, making it susceptible to influence from personal subjective opinions, experience, and understanding, leading to misjudgments. Therefore, it is necessary to adjust the parameters of the defect judgment criteria in case of misjudgments.

[0076] Based on this, please refer to Figure 3. The defect detection parameter tuning method provided in this embodiment includes the following steps:

[0077] S201. In response to the parameter tuning operation for the initial defect type in the defect display interface, display the defect parameter tuning interface corresponding to the initial defect type.

[0078] The defect parameter tuning interface includes: multiple feature values ​​corresponding to multiple defect features in the abnormal area corresponding to the initial defect type, a defect selection control, and a data learning control. For example, the parameter tuning operation for the initial defect type in the defect display interface can be: the worker triggers the parameter tuning control corresponding to the initial defect type on the defect display interface to perform the parameter tuning operation; this embodiment does not specifically limit this. Furthermore, each defect type shown in the defect display interface can be used as a separate initial defect type.

[0079] Understandably, since the defect judgment conditions used in defect detection are usually set by staff based on their own work experience, they are easily affected by personal subjective consciousness, experience, and understanding, resulting in low accuracy of defect judgment conditions. That is, the defect type of the abnormal area detected may not match the actual defect type in the product to be inspected. Staff need to make a second judgment to check whether the detected defect type matches the actual defect type. If they do not match, the parameters in the defect judgment conditions need to be adjusted to make the detection results of the initial defect type more accurate.

[0080] In some embodiments, as shown in Figure 2, the defect display interface includes parameter tuning controls corresponding to each defect type. If an operator determines that the initial defect type in the defect display interface does not match its corresponding actual defect type, they can select the parameter tuning control corresponding to the initial defect type to perform parameter tuning.

[0081] Upon receiving a parameter tuning operation for the initial defect type displayed on the defect display interface, the detection equipment displays the corresponding defect parameter tuning interface. As shown in Figure 4, the defect parameter tuning interface includes: multiple feature values ​​corresponding to multiple defect features in the abnormal area corresponding to the initial defect type, a defect selection control, and a data learning control.

[0082] It should be understood that a defect type includes multiple defect features, such as area, long side, short side, rectangularity, roundness, grayscale, etc. Each defect feature corresponds to a detected feature value. The defect judgment conditions corresponding to the defect type include the initial defect truth value range corresponding to each defect feature among the multiple defect features of the defect type.

[0083] For example, as shown in Figure 4, when the initial defect type is white spot, the corresponding defect features include: area, rectangularity, shorter side, average gray level, and polarity. Each defect feature corresponds to a feature value and a judgment specification (the true value range of the defect). For example, as shown in Figure 4, the feature value corresponding to the defect feature "area" is 225.2; the feature value corresponding to the defect feature "rectangularity" is 1; the feature value corresponding to the defect feature "short side" is 15; the feature value corresponding to the defect feature "average gray level" is 123.9; and the feature value corresponding to the defect feature "polarity" is 0.

[0084] As a feasible implementation method, as shown in Figure 4, the defect display interface can also include a defect truth value range display control corresponding to each defect feature. It should be understood that if the operator does not select a target defect type in the defect selection control, the defect truth value range display control can display the defect truth value range corresponding to the initial defect type. If the operator selects a target defect type in the defect selection control, the defect truth value range display control can display the defect truth value range corresponding to the target defect type.

[0085] As a feasible implementation method, as shown in Figure 4, the defect display interface can also include the feature status (OK / NG as shown in Figure 4) corresponding to each defect feature. It should be understood that the feature status indicates whether the feature value corresponding to the defect feature matches the initial defect truth value range. As shown in Figure 4, when the feature value corresponding to the defect feature matches the initial defect truth value range, its corresponding feature status is OK; when the feature value corresponding to the defect feature does not match the initial defect truth value range, its corresponding feature status is NG. This embodiment of the application does not specifically limit this.

[0086] As a feasible implementation method, as shown in Figure 4, the defect display interface can also include an adjustment control for the initial defect truth value range corresponding to each defect feature, so that staff can manually adjust the initial defect truth value range corresponding to the defect feature based on the adjustment control.

[0087] S202. In response to a trigger operation on the data learning control, obtain the range of multiple initial defect truth values ​​corresponding to the target defect type in the defect selection control.

[0088] It should be noted that when the staff does not select a target defect type based on the defect selection control, the data learning control can be grayed out to prevent the staff from accidentally touching the data learning control. After the staff selects a target defect type based on the defect selection control, the data learning control can be restored to normal so that the staff can trigger the data learning control. This application embodiment does not make specific limitations in this regard.

[0089] Understandably, the target defect type is the actual defect type of the abnormal area determined by the staff. During defect detection, due to inaccurate defect judgment conditions (multiple initial defect truth value ranges) for the target defect type, the abnormal area is detected as the initial defect type. Therefore, it is necessary to obtain and adjust the multiple initial defect truth value ranges corresponding to the target defect type in the defect selection control. Furthermore, the target defect type is the actual defect type corresponding to the initial defect type. Both the target defect type and the initial defect type are one of multiple preset defect types, and different names are used to distinguish between the actual defect type and the defect type detected by the defect detection algorithm (i.e., the defect type displayed in the defect display interface). Moreover, the multiple initial defect truth value ranges corresponding to the target defect type are the defect truth value ranges corresponding to each defect feature in the current defect judgment conditions of the target defect type.

[0090] S203. Based on the multiple feature values ​​of the abnormal region corresponding to the initial defect type, adjust the multiple initial defect truth value ranges corresponding to the target defect type to obtain the multiple target defect truth value ranges corresponding to the target defect type.

[0091] In this context, the true value range of the target defect corresponding to each defect feature is matched with the feature value corresponding to the defect feature.

[0092] Understandably, when the true defect type of the abnormal region is determined to be the target defect type, the feature values ​​of multiple defect features in the abnormal region should match the true value range of the target defect type. Therefore, it is necessary to adjust the multiple initial true value ranges of the target defect type based on the multiple feature values ​​of the abnormal region corresponding to the initial defect type, to obtain the target defect true value range that matches the feature value of each defect feature for the target defect type. For example, the feature value of the defect feature "area" corresponding to the initial defect type 1 is 225.2, that is, the feature value of the defect feature "area" of the abnormal region corresponding to the initial defect type 1 is 225.2. Then, the initial true value range of the defect feature "area" of the target defect type can be adjusted using the feature value of the defect feature "area" of the initial defect type 1, to obtain the target defect true value range of the defect feature "area" of the target defect type, which includes 225.2.

[0093] To ensure that the feature values ​​of multiple defect features in the abnormal region match the true value range of the target defect type, there are multiple ways to adjust the true value range of multiple initial defects corresponding to the target defect type based on the multiple feature values ​​of the abnormal region corresponding to the initial defect type. The following will introduce these methods with examples.

[0094] S204. Based on the true value range of multiple target defects, update the defect judgment conditions corresponding to the target defect type.

[0095] After obtaining the true value range of multiple target defects corresponding to multiple defect features in the target defect type, the defect judgment conditions corresponding to the target defect type can be updated to make the defect judgment conditions corresponding to the target defect type more accurate. Then, when performing defect detection on the image to be detected based on the updated defect judgment conditions, the abnormal region can be accurately identified as the target defect type.

[0096] As can be seen, the defect detection parameter tuning method provided in this application, after obtaining the defect detection result, can display the defect type detected in the image to be detected through a defect display interface, so that staff can check whether the detected initial defect type is consistent with the actual defect type. Furthermore, if they are inconsistent, staff can trigger the parameter tuning control for the initial defect type in the defect display interface (triggering a parameter tuning operation for the initial defect type in the defect display interface) to display the defect parameter tuning interface corresponding to the initial defect type. This allows staff to select the target defect type actually corresponding to the initial defect type based on the defect selection control in the defect parameter tuning interface, and automatically adjust the initial defect truth value range corresponding to multiple defect features of the target defect type through the data learning control in the defect parameter tuning interface, obtaining the target defect truth value range corresponding to the feature value of each defect feature that matches the target defect type.

[0097] In this way, when performing defect detection again, the feature values ​​of the abnormal area can match the defect judgment conditions corresponding to the target defect type, that is, the defect type of the abnormal area is detected as the target defect type, thus achieving the goal of more accurate defect detection. The defect detection parameter tuning method provided in this application embodiment does not require manual adjustment by staff based on their own experience, which can avoid errors and delays caused by human factors, thereby improving the accuracy and reliability of defect detection. At the same time, it can reduce the manual operation and judgment time of staff in defect detection, thereby improving the detection efficiency in the defect detection process. Furthermore, the data learning control in the defect parameter tuning interface automatically adjusts the initial defect truth value range corresponding to multiple defect features of the target defect type, providing users with a one-click parameter tuning function, which has high parameter tuning convenience and speed.

[0098] In some embodiments, since different defect types correspond to different defect judgment conditions, the basis for parameter adjustment also differs. Therefore, in practical applications, a parameter prediction model can be trained for each defect type. When staff need to adjust parameters after selecting the data learning control, the target parameter prediction model corresponding to that defect type is used to determine the adjusted target defect truth value range.

[0099] Specifically, as a feasible implementation method, S203 can be implemented as follows: inputting multiple feature values ​​of the abnormal region corresponding to the initial defect type and multiple initial defect truth value ranges corresponding to the target defect type into the target parameter prediction model corresponding to the target defect type to obtain multiple target defect truth value ranges corresponding to the target defect type.

[0100] Among them, the target parameter prediction model is used to adjust the true value range of multiple initial defects based on multiple feature values ​​to obtain multiple true value ranges of target defects.

[0101] As a feasible approach, the target parameter prediction model can be a linear regression model. Linear regression is a statistical analysis method used to establish a linear relationship between independent variables (inputs) and dependent variables (outputs). In linear regression, it is assumed that the dependent variable (the true range of the target defect) is a linear combination of the independent variables (eigenvalues ​​and the true range of the initial defect), and the optimal linear relationship is found through data fitting to predict new dependent variable values.

[0102] The feature values ​​corresponding to multiple defect features represent multiple factors influencing the true range of the target defect. In the target parameter prediction model, each defect feature can have a corresponding weight coefficient, which represents the degree of influence of that feature on the variable. Then, based on the data input to the model and the corresponding weight coefficients, the true range of multiple target defects is calculated.

[0103] Specifically, as a feasible implementation method, the feature value corresponding to each defect feature among multiple defect features, the upper limit and lower limit of the initial defect true value range can be used as input to the target parameter prediction model, so that the target parameter prediction model outputs the upper limit and lower limit of the target defect true value range.

[0104] In some embodiments, referring to Figure 5, the target parameter prediction model provided in this application embodiment can be trained in the following manner:

[0105] S501. Obtain multiple sample data corresponding to the target defect type.

[0106] Each sample data includes the historical feature value, the initial historical defect true value range, and the target historical defect true value range for each defect feature among multiple defect features.

[0107] When training the target parameter prediction model corresponding to the initial defect type, it is necessary to obtain multiple sample data corresponding to the initial defect type.

[0108] It should be understood that, as a feasible approach, when staff manually adjust the parameters of the target defect type, the historical feature values ​​corresponding to each detected defect feature, as well as the unadjusted defect judgment conditions (multiple initial historical defect true value ranges) and the manually adjusted defect judgment conditions (multiple target historical defect true value ranges) can be saved as sample data corresponding to the target defect type for training the initial parameter prediction model.

[0109] S502. Train the initial parameter prediction model based on multiple sample data corresponding to the target defect type to obtain the target parameter prediction model corresponding to the target defect type.

[0110] After obtaining multiple sample data corresponding to the target defect type, the historical feature values ​​and the initial historical defect true value range corresponding to multiple defect features in the sample data can be input into the initial parameter prediction model. Then, based on the error between the predicted value and the expected value (the true value range of multiple target historical defects) output by the initial parameter prediction model, the loss function is calculated. The initial parameter prediction model is trained by minimizing the loss function until the preset iteration stopping condition is met, and the target parameter prediction model corresponding to the initial defect type is obtained.

[0111] As a feasible approach, the loss function can be calculated by measuring the squared difference between the predicted value and the expected value output by the initial parameter prediction model. Of course, other methods can also be used to calculate the loss function, such as using the difference between the predicted value and the expected value output by the initial parameter prediction model as the loss function; however, this application does not specifically limit this approach.

[0112] It should be understood that the preset iteration stopping conditions can be reaching a preset number of iterations, reaching a preset training time, or the loss function value being lower than a preset threshold, etc., and the embodiments of this application do not limit these conditions.

[0113] As can be seen, the defect detection parameter tuning method provided in this embodiment trains the target parameter prediction model corresponding to the target defect type by acquiring multiple sample data corresponding to the target defect type. The obtained target parameter prediction model can adjust the parameters more accurately, thereby obtaining more accurate defect judgment conditions for the target defect type.

[0114] In some embodiments, since the decision on whether to adjust parameters is based on the staff’s own experience, there may be errors in judgment. In this case, the updated defect judgment conditions for the target defect type obtained after adjusting the initial defect truth value range may not be reasonable.

[0115] Therefore, as a feasible implementation method, to avoid the phenomenon of incorrect parameter tuning, please refer to Figure 6. The defect detection parameter tuning method provided in this application embodiment, after updating the defect judgment conditions corresponding to the target defect type, further includes:

[0116] S601. Obtain multiple historical defect judgment conditions corresponding to the target defect type.

[0117] It is understood that the historical defect judgment conditions corresponding to the target defect type are the range of multiple initial defect truth values ​​corresponding to the target defect type each time the defect judgment conditions of the target defect type are updated; this application embodiment does not specifically limit this.

[0118] S602. Determine the confidence level of the updated defect judgment condition corresponding to the target defect type based on multiple historical defect judgment conditions.

[0119] It should be understood that confidence level can be used to represent the degree of similarity between the updated defect judgment condition and multiple historical defect judgment conditions, thereby characterizing the credibility of the updated defect judgment condition corresponding to the target defect type.

[0120] As a feasible approach, the confidence level of each individual defect feature in the updated defect judgment conditions can be calculated first: for each defect feature, the confidence level between the true value range of the defect in the updated defect judgment conditions and the true value range of the defect in each historical defect judgment condition is calculated. The confidence level can then be calculated using methods such as interval overlap or Jaccard similarity; this application does not specifically limit this method. For example, interval overlap can be defined as the ratio of the length of the intersection of two intervals to the length of their union.

[0121] Furthermore, since different features may have varying importance in defect judgment conditions, a weight can be assigned to each defect feature. The weight can be determined based on the actual situation, such as through expert scoring or historical data analysis; this embodiment does not impose specific limitations on this. Then, the similarity of each defect feature is multiplied by its corresponding weight, and the results are summed to obtain the confidence level of the updated defect judgment condition compared to each historical defect judgment condition. Finally, by averaging or other methods, the confidence level of the updated defect judgment condition is obtained based on the confidence levels of the updated defect judgment condition and multiple historical defect judgment conditions.

[0122] S603. When the confidence level is less than or equal to the preset confidence threshold, output a prompt message.

[0123] The prompt message is used to prompt staff to determine whether the updated defect judgment conditions are reasonable; of course, the prompt message can also be highlighted, for example, by rendering the border of the prompt message to achieve the effect of highlighting; this application embodiment does not specifically limit this.

[0124] It should be understood that the preset reliability threshold is pre-set and can be set according to needs in actual application. The preset reliability threshold can be an empirical value, and the embodiments of this application do not limit it.

[0125] When the calculated confidence level is less than or equal to the preset confidence threshold, it usually means that the result may not be reliable enough or requires further analysis. In this case, a prompt message needs to be output to remind staff to intervene manually and avoid unreasonable defect judgment conditions. On the other hand, when the calculated confidence level is greater than the preset confidence threshold, it indicates that the result is relatively reliable, and this application embodiment does not specifically limit this.

[0126] It should be understood that the prompt message should clearly and explicitly convey that the confidence level is less than or equal to a preset confidence threshold, and may include suggested actions (such as collecting more data, reconsidering the analysis method, etc.). The specific content of the prompt message is not limited in the embodiments of this application.

[0127] It is understandable that the defect detection parameter tuning method provided in the above embodiments is mainly applied to false detection, that is, in the defect detection process, abnormal areas that do not belong to the initial defect type are mistakenly identified as the initial defect type. Therefore, the defect judgment conditions of the target defect type can be adjusted so that the initial defect type can be more accurately identified in the defect detection process. However, in the defect detection process, not only false detections may occur, but also missed detections may occur, that is, multiple feature values ​​of multiple defect features of the abnormal area do not match multiple defect judgment conditions. In this case, the defect type corresponding to the abnormal area can be determined to be a classification filtering defect.

[0128] In other words, as a feasible approach, if the feature values ​​of multiple defect characteristics in an abnormal region do not match the defect judgment conditions of other defect types among multiple defect types, the abnormal region is identified as a classification filtering defect. This abnormal region is then filtered out, and during the defect detection process, it is considered not to belong to other defect types, resulting in a missed detection.

[0129] If a worker believes that any type of defect exists in the abnormal area, they can select the parameter adjustment control corresponding to the defect classification filter on the defect display interface. This will display the defect parameter adjustment interface for the classified filter defect, allowing the worker to select the target defect type corresponding to the abnormal area and adjust the parameters. Specific adjustment methods can be found in the above embodiment, and will not be repeated here.

[0130] This application also provides another defect detection parameter tuning method, which can be applied to an electronic device. For example, the electronic device can be a detection device in the aforementioned defect detection parameter tuning system, or other devices communicating with the detection device. For a description of the detection device and other devices, please refer to the relevant content in the above embodiments, which will not be repeated here. As shown in Figure 9, the other defect detection parameter tuning method provided in this application embodiment may include the following steps:

[0131] S901 displays the defect parameter tuning interface corresponding to the initial defect type;

[0132] The initial defect type is the defect type obtained from the preliminary detection of the image to be detected; the defect parameter tuning interface includes a defect selection control and a data learning control; the defect selection control is used to select the actual target defect type corresponding to the image to be detected.

[0133] It is understood that in this embodiment, after the initial defect type (initial defect type) is obtained through preliminary detection of the image to be inspected, the defect parameter tuning interface corresponding to the initial defect type can be directly displayed. Compared with the above embodiment, in this embodiment, it is not necessary to first display the defect display interface, but to directly display the defect parameter tuning interface corresponding to the initial defect type, thereby improving the convenience and speed of parameter tuning. Furthermore, the initial defect type is specifically any defect type obtained by the detection device through preliminary detection of the image to be inspected based on the defect judgment conditions corresponding to multiple defect types. A defect type can include multiple defect features, such as one or more features from areas, long sides, short sides, rectangularity, roundness, grayscale, etc., and each defect feature corresponds to a detected feature value. A defect type can correspond to defect judgment conditions, which include the defect truth value range corresponding to each defect feature among the multiple defect features. For the specific implementation method of the detection device performing preliminary detection of the image to be inspected to obtain the defect type, as well as examples and detailed content of each defect type, the defect judgment conditions for each defect type, and the defect features included in each defect type, please refer to the relevant content of the above embodiments, which will not be repeated here.

[0134] Furthermore, the defect parameter tuning interface may include a defect selection control and a data learning control; wherein, the defect selection control is used to select the target defect type corresponding to the image to be inspected. Optionally, in one implementation, the data learning control can be grayed out (in an inoperable state) when the operator has not selected a target defect type based on the defect selection control, to prevent the operator from accidentally touching the data learning control; after the operator has selected the target defect type based on the defect selection control, the data learning control can be restored to normal (i.e., in an operable state), so that the operator can trigger the data learning control.

[0135] Alternatively, the defect parameter tuning interface shown may be the defect parameter tuning interface shown in Figure 4. For details, please refer to the above embodiments, which will not be elaborated further here.

[0136] S902, in response to the trigger operation of the data learning control, adjusts the initial defect truth value range corresponding to the target defect type based on the feature values ​​detected for the initial defect type to obtain the target defect truth value range;

[0137] Among them, the true value range of the target defect matches the feature value.

[0138] Among them, the feature value detected for the initial defect type refers to the feature value corresponding to the defect feature of the abnormal area corresponding to the initial detection type. The number of defect features in the abnormal area can be one or more.

[0139] It is understandable that the target defect type is the true defect type (also known as the actual defect type) of the abnormal area determined by the worker through the defect selection control. That is, for an abnormal area, the target defect type is misjudged as the initial defect type. Because the initial true value range of the target defect type is inaccurate during defect detection, the feature values ​​of the defect features in the abnormal area do not match (are contained within) the initial true value range of the target defect type, leading to inaccurate detection. Therefore, after manually selecting the target defect type, the initial true value range corresponding to the target defect type can be adjusted using the feature values ​​detected by the initial defect type. The number of initial true value ranges for the target defect type can be one or more, with each initial true value range corresponding to a defect feature. Furthermore, when the true defect type of any abnormal area is determined to be the target defect type, the feature values ​​of the defect features in that abnormal area (the feature values ​​detected for the initial defect type) should match the target defect true value range. Therefore, the initial true value range corresponding to the target defect type can be adjusted based on the feature values ​​detected for the initial defect type to obtain the target defect true value range that matches the feature values.

[0140] In addition, the obtained target defect truth value range can be used as the adjusted defect judgment condition for the target defect type. Then, the target defect truth value range can be directly determined as the defect judgment condition for the target defect type, making the defect judgment condition corresponding to the target defect type more accurate. Thus, when performing defect detection on the image to be detected based on the defect judgment condition, the target defect type can be accurately identified.

[0141] The defect detection parameter tuning method provided in this application embodiment can directly display the defect parameter tuning interface corresponding to the initial defect type. Based on the defect selection control in the defect parameter tuning interface, the staff can select the target defect type that actually corresponds to the initial defect type, and automatically adjust the initial defect truth range corresponding to the target defect type through the data learning control in the defect parameter tuning interface to obtain the target defect truth range.

[0142] In this way, when re-performing defect detection, the true value range of the target defect (i.e., the defect judgment condition of the target defect type) matches the feature values ​​detected for the initial defect type (i.e., the feature values ​​detected for the abnormal areas of the initial defect type), thus achieving a more accurate defect detection. Furthermore, this defect detection parameter tuning method eliminates the need for manual adjustments based on staff experience, avoiding errors and delays caused by human factors, thereby improving the accuracy and reliability of defect detection. Simultaneously, it reduces the time spent on manual operations and judgments during defect detection, improving detection efficiency. Moreover, the automated adjustment of the initial defect true value range corresponding to multiple defect features of the target defect type through the data learning controls in the defect parameter tuning interface provides users with a one-click parameter tuning function, offering high convenience and speed. Additionally, this embodiment can directly display the defect parameter tuning interface corresponding to the initial defect type, further enhancing the convenience and speed of parameter tuning.

[0143] Optionally, in another embodiment, before displaying the defect parameter tuning interface corresponding to the initial defect type, the method further includes:

[0144] The defect display interface shows the initial defect type.

[0145] If a parameter tuning operation is detected for the initial defect type in the defect display interface, the step of displaying the defect parameter tuning interface corresponding to the initial defect type is triggered.

[0146] It is understood that in this embodiment, before displaying the defect parameter tuning interface corresponding to the initial defect type, a defect display interface can be displayed first, which can display the initial defect type; and if the initial defect type displayed on the defect display interface does not match the target defect type actually corresponding to the image to be detected, the operator can perform parameter tuning operations on the initial defect type in the defect display interface. Accordingly, when a parameter tuning operation on the initial defect type in the defect display interface is detected, the step of displaying the defect parameter tuning interface corresponding to the initial defect type can be triggered, thereby realizing the interface jump from the defect display interface to the defect parameter tuning interface.

[0147] As can be seen, the defect detection parameter tuning method provided in this embodiment can display the initial defect type through a defect display interface, allowing staff to check whether the detected initial defect type matches the actual defect type. If they do not match, staff can trigger a parameter tuning operation for the initial defect type in the defect display interface, thereby redirecting to the corresponding defect parameter tuning interface and providing a foundation for defect detection parameter tuning.

[0148] Optionally, in another embodiment, based on the feature values ​​detected for the initial defect type, the initial defect truth range corresponding to the target defect type is adjusted to obtain the target defect truth range, including:

[0149] The feature values ​​detected for the initial defect type and the initial defect truth range corresponding to the target defect type are input into the pre-trained target parameter prediction model corresponding to the target defect type to obtain the target defect truth range. The target parameter prediction model is used to adjust the initial defect truth range corresponding to the target defect type based on the feature values ​​detected for the initial defect type to obtain the target defect truth range.

[0150] The target parameter prediction model can be a linear regression model, which is a statistical analysis method used to establish a linear relationship between independent variables (inputs) and dependent variables (outputs). In linear regression, it is assumed that the dependent variable (the true range of the target defect) is a linear combination of the independent variables (eigenvalues ​​and the true range of the initial defect), and the optimal linear relationship is found through data fitting to predict the new value of the dependent variable.

[0151] The feature values ​​of each defect feature detected for the initial defect type represent factors influencing the true range of the target defect. In the target parameter prediction model, each defect feature has a corresponding weight coefficient, which represents the degree of influence of that defect feature on the variable. The true range of the target defect is then calculated based on the input model data and the corresponding weight coefficients. It should be noted that feature values ​​belonging to the same defect feature correspond to the initial defect true range. Therefore, when multiple defect feature values ​​are detected for anomaly regions for the initial defect type, the feature values ​​of these multiple defect features and the multiple initial defect true ranges corresponding to the target defect type can be input into the target parameter prediction model corresponding to the target defect type to obtain multiple target defect true ranges. The target parameter prediction model is then used to adjust these multiple initial defect true ranges based on these feature values ​​to obtain multiple target defect true ranges.

[0152] As an optional implementation, the feature values, the upper and lower limits of the initial defect truth value range can be used as inputs to the target parameter prediction model, so that the target parameter prediction model outputs the upper and lower limits of the target defect truth value range. This application does not specifically limit this implementation. For details regarding the target parameter prediction model, such as training methods and specific inference applications, please refer to the relevant content in the above embodiments, which will not be elaborated upon here.

[0153] As can be seen, in this embodiment, the feature values ​​detected for the initial defect type and the initial defect truth value range corresponding to the target defect type can be input into the target parameter prediction model, so that the target parameter prediction model can quickly output the target defect truth value range. Compared with determining the target defect truth value range based on the experience of the staff, the target defect truth value range output by the target parameter prediction model is more accurate and more efficient.

[0154] Optionally, in another embodiment, the defect display interface further includes: an image to be detected and a display box for the abnormal area corresponding to the initial defect type; and / or,

[0155] The filtered defects corresponding to the image to be detected are used to represent abnormal regions in the image to be detected where the feature values ​​of multiple defect features do not match the defect judgment conditions corresponding to each of the multiple defect types.

[0156] It is understandable that the defect display interface may also display: the image to be inspected and the display box of the abnormal area corresponding to the initial defect type. In the defect display interface, the operator can better determine whether the initially detected defect type (i.e., the initial defect type) is a real defect and whether parameter adjustments are needed, based on the defect type present in the image to be inspected and its corresponding display box. Furthermore, the defect display interface may also include filtered defects corresponding to the image to be inspected. Filtered defects are a special type of defect, used to represent abnormal areas in the image to be inspected where the feature values ​​of multiple defect features do not match the defect judgment conditions corresponding to each of the multiple defect types; this embodiment does not specifically limit this aspect.

[0157] As can be seen, in this embodiment, the defect display interface may also include the image to be detected and a display box for the abnormal area corresponding to the initial defect type; and / or, the filtered defect corresponding to the image to be detected, so that the staff can more clearly see the image to be detected and the abnormal area corresponding to the initial defect type.

[0158] Optionally, in another embodiment, the defect parameter tuning interface also includes feature values ​​detected for the initial defect type. For example, the feature values ​​detected for the initial defect type may specifically include: feature values ​​of multiple defect features detected for the abnormal region corresponding to the initial defect type.

[0159] And / or,

[0160] The defect parameter tuning interface also includes: feature status, and / or, display controls and adjustment controls corresponding to the initial defect truth value range; feature status is used to indicate whether the feature value detected for the initial defect type matches the initial defect truth value range.

[0161] It is understood that, for any defect feature, the feature status is used to indicate whether the detected feature value for the initial defect type matches the initial defect truth value range. If the feature value matches the initial defect truth value range, the corresponding feature status is OK; if the feature value does not match the initial defect truth value range, the corresponding feature status is NG. This application embodiment does not specifically limit this. Furthermore, the defect display interface may also include adjustment controls for the initial defect truth value range corresponding to each defect feature, so that staff can manually adjust the initial defect truth value range corresponding to the defect feature based on the adjustment controls.

[0162] As can be seen, the defect parameter tuning interface can also include display controls and adjustment controls for the feature values ​​detected for the initial defect type, and / or feature status, and / or the initial defect true value range, so that staff can see the mismatch between the feature values ​​and the initial defect true value range, and facilitate staff to manually adjust the initial defect true value range corresponding to the defect feature by adjusting the control.

[0163] Optionally, in another embodiment, after adjusting the initial defect truth range corresponding to the target defect type based on the feature values ​​detected for the initial defect type, and obtaining the target defect truth range, that is, after updating the defect judgment conditions corresponding to the target defect type, the method further includes:

[0164] Obtain multiple historical defect judgment conditions corresponding to the target defect type;

[0165] The confidence level of the updated defect judgment condition corresponding to the target defect type is determined based on multiple historical defect judgment conditions.

[0166] When the confidence level is less than or equal to the preset confidence threshold, a prompt message is output. The prompt message is used to prompt staff to determine whether the updated defect judgment conditions are reasonable.

[0167] The relevant descriptions of this embodiment can be found in the corresponding content of the above embodiments, and will not be repeated here.

[0168] This application embodiment also provides a defect detection parameter tuning device, as shown in FIG7. The defect detection parameter tuning device 70 includes: a display module 71, an adjustment module 72 and an update module 73.

[0169] The display module 71 is configured to respond to the parameter tuning operation of the initial defect type in the defect display interface and display the defect parameter tuning interface corresponding to the initial defect type. The defect display interface displays the initial defect type obtained by detecting the image to be detected. The defect parameter tuning interface includes: multiple feature values ​​corresponding to multiple defect features of the abnormal area corresponding to the initial defect type, a defect selection control, and a data learning control.

[0170] The adjustment module 72 is configured to respond to the trigger operation of the data learning control, obtain multiple initial defect truth value ranges corresponding to the target defect type in the defect selection control, and adjust the multiple initial defect truth value ranges corresponding to the target defect type based on multiple feature values ​​of the abnormal region corresponding to the initial defect type to obtain multiple target defect truth value ranges corresponding to the target defect type; wherein, the target defect truth value range corresponding to each defect feature matches the feature value corresponding to the defect feature.

[0171] Update module 73 to update the defect judgment conditions corresponding to the target defect type based on the true value range of multiple target defects.

[0172] In some embodiments, the adjustment module 72 is specifically configured to input multiple feature values ​​of the abnormal region corresponding to the initial defect type and multiple initial defect truth value ranges corresponding to the target defect type into the target parameter prediction model corresponding to the target defect type to obtain multiple target defect truth value ranges corresponding to the target defect type; the target parameter prediction model is used to adjust multiple initial defect truth value ranges based on multiple feature values ​​to obtain multiple target defect truth value ranges.

[0173] In some embodiments, the defect detection parameter tuning device 70 further includes: an acquisition module configured to acquire multiple historical defect judgment conditions corresponding to the target defect type; a determination module further configured to determine the confidence level of the updated defect judgment conditions corresponding to the target defect type based on the multiple historical defect judgment conditions; and an output module configured to output a prompt message when the confidence level is less than or equal to a preset confidence threshold, the prompt message being used to prompt staff to determine whether the updated defect judgment conditions are reasonable.

[0174] In some embodiments, the defect detection parameter tuning device 70 further includes: a determination module, configured to determine an abnormal region in the image to be detected and determine the feature values ​​of multiple defect features corresponding to the abnormal region; a matching module, configured to match the defect judgment conditions corresponding to the defect type with the feature values ​​of multiple defect features corresponding to the abnormal region according to the defect type order recorded in the defect priority, so as to determine the defect type obtained by detecting the image to be detected; and a display module 71, further configured to display the defect type obtained by detecting the image to be detected through a defect display interface.

[0175] In some embodiments, the defect parameter tuning interface further includes: a feature state corresponding to each defect feature; the feature state is used to indicate whether the feature value corresponding to the defect feature matches the initial defect truth value range; and / or, a display control and an adjustment control corresponding to each initial defect truth value range.

[0176] In some embodiments, the defect display interface further includes: an image to be detected and a display box for the abnormal region corresponding to the defect type obtained by detecting the image to be detected; and / or, a filtered defect corresponding to the image to be detected, wherein the filtered defect is used to represent a defect in an abnormal region in the image to be detected where the feature values ​​of multiple defect features of the abnormal region do not match the multiple defect judgment conditions corresponding to the multiple defect types.

[0177] Based on the other defect detection parameter tuning method provided above, this application embodiment also provides a defect detection parameter tuning device, as shown in FIG10, the defect detection parameter tuning device 100 includes:

[0178] The first display module 1001 is configured to display the defect parameter tuning interface corresponding to the initial defect type; wherein, the initial defect type is the defect type obtained by preliminary detection of the image to be detected; the defect parameter tuning interface includes a defect selection control and a data learning control; the defect selection control is used to select the actual target defect type corresponding to the image to be detected;

[0179] The target adjustment module 1002 is configured to respond to a trigger operation of the data learning control, and adjust the initial defect truth value range corresponding to the target defect type based on the feature values ​​detected for the initial defect type to obtain the target defect truth value range; wherein the target defect truth value range matches the feature values.

[0180] Optionally, the device further includes:

[0181] The second display module is configured to display a defect display interface before the first display module displays the defect parameter tuning interface corresponding to the initial defect type; the defect display interface displays the initial defect type.

[0182] The trigger module is configured to trigger the first display module to display the parameter tuning interface corresponding to the initial defect type when a parameter tuning operation for the initial defect type in the defect display interface is detected.

[0183] Optionally, the target adjustment module 1002 is specifically configured as follows:

[0184] The feature values ​​detected for the initial defect type and the initial defect truth range corresponding to the target defect type are input into the pre-trained target parameter prediction model corresponding to the target defect type to obtain the target defect truth range. The target parameter prediction model is used to adjust the initial defect truth range corresponding to the target defect type based on the feature values ​​detected for the initial defect type to obtain the target defect truth range.

[0185] Optionally, the defect display interface may also include: the image to be inspected and a display box for the abnormal area corresponding to the initial defect type; and / or,

[0186] The filtered defects corresponding to the image to be detected are used to represent abnormal regions in the image to be detected where the feature values ​​of multiple defect features do not match the defect judgment conditions corresponding to each of the multiple defect types.

[0187] Optionally, the defect parameter tuning interface also includes feature values ​​detected for the initial defect type;

[0188] And / or,

[0189] The defect parameter tuning interface also includes: feature status, and / or, display controls and adjustment controls corresponding to the initial defect truth value range; feature status is used to indicate whether the feature value detected for the initial defect type matches the initial defect truth value range.

[0190] This application also provides an electronic device, as shown in FIG8. The electronic device 201 includes: one or more memories 121, one or more processors 122, a communication bus 123, and a communication interface 124. The processors 122 and the memories 121 are connected via the communication bus 123. The one or more memories 121 are used to store computer program code, which includes computer instructions. When the one or more processors 122 execute the computer instructions, the electronic device 201 performs the defect detection parameter tuning method provided in the above embodiment.

[0191] Optionally, the memory 121 may be a non-transitory computer-readable storage medium, such as a read-only memory (ROM), random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc. The embodiments of this application do not impose any limitations on this.

[0192] The processor 122 may be a central processing unit (CPU), a general-purpose processor, a network processor (NP), a digital signal processor (DSP), a microprocessor, a microcontroller, a programmable logic device (PLD), or any combination thereof, and the embodiments of this application do not impose any limitations on this.

[0193] The communication bus 123 can be an industry standard architecture (ISA) bus, a peripheral component interconnect (PCI) bus, or an extended industry standard architecture (EISA) bus, etc. This communication bus 123 can be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, Figure 8 uses only one thick line to represent it, but this does not indicate that there is only one bus or one type of communication bus.

[0194] Communication interface 124 uses any transceiver-like device for communicating with other devices or communication networks, such as control systems, radio access networks (RAN), wireless local area networks (WLAN), etc.

[0195] This application also provides a computer program product comprising one or more instructions, which are stored in the memory of a computer device and executed by a processor to complete the various processes described in the above embodiments.

[0196] This application also provides a computer-readable storage medium, which includes computer-executable instructions. When the computer-executable instructions are run on a computer, the computer performs the defect detection parameter tuning method provided in the above embodiments.

[0197] In another embodiment provided in this application, a computer program is also provided, which, when run on a computer, causes the computer to execute the defect detection parameter tuning method provided in the above embodiments.

[0198] Through the above description of the embodiments, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0199] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another apparatus, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0200] The units described as separate components may or may not be physically separate. A component shown as a unit can be one or more physical units; that is, it can be located in one place or distributed in multiple different locations. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0201] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0202] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, in essence, or the parts that contribute to the prior art, or all or part of the technical solutions, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a device or processor to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0203] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A defect detection parameter tuning method, comprising: In response to the parameter tuning operation for the initial defect type in the defect display interface, the defect parameter tuning interface corresponding to the initial defect type is displayed; The defect display interface shows the initial defect types obtained from the detection of the image to be inspected; The defect parameter tuning interface includes: multiple feature values ​​corresponding to multiple defect features of the abnormal region corresponding to the initial defect type, a defect selection control, and a data learning control; In response to a trigger operation on the data learning control, multiple initial defect truth value ranges corresponding to the target defect type in the defect selection control are obtained, and the multiple initial defect truth value ranges corresponding to the target defect type are adjusted based on multiple feature values ​​of the abnormal region corresponding to the initial defect type to obtain multiple target defect truth value ranges corresponding to the target defect type; wherein, the target defect truth value range corresponding to each defect feature matches the feature value corresponding to the defect feature; Based on the true value range of the multiple target defects, update the defect judgment conditions corresponding to the target defect type.

2. The method of claim 1, wherein, The step of adjusting the multiple initial defect truth ranges corresponding to the target defect type based on multiple feature values ​​of the abnormal region corresponding to the initial defect type to obtain multiple target defect truth ranges corresponding to the target defect type includes: Multiple feature values ​​of the abnormal region corresponding to the initial defect type and multiple initial defect truth value ranges corresponding to the target defect type are input into the target parameter prediction model corresponding to the target defect type to obtain multiple target defect truth value ranges corresponding to the target defect type; the target parameter prediction model is used to adjust the multiple initial defect truth value ranges based on the multiple feature values ​​to obtain the multiple target defect truth value ranges.

3. The method of claim 2, wherein, The target parameter prediction model is trained in the following manner: Obtain multiple sample data corresponding to the target defect type, wherein each sample data includes historical feature value, initial historical defect true value range and target historical defect true value range corresponding to each defect feature among multiple defect features; The initial parameter prediction model is trained based on multiple sample data corresponding to the target defect type to obtain the target parameter prediction model corresponding to the target defect type.

4. The method of claim 1, wherein, After updating the defect judgment conditions corresponding to the target defect type, the method further includes: Obtain multiple historical defect judgment conditions corresponding to the target defect type; The confidence level of the updated defect judgment condition corresponding to the target defect type is determined based on the multiple historical defect judgment conditions. When the confidence level is less than or equal to a preset confidence threshold, a prompt message is output. The prompt message is used to prompt staff to determine whether the updated defect judgment conditions are reasonable.

5. The method of claim 1, wherein, The method further includes: Identify the abnormal regions in the image to be detected, and determine the feature values ​​of multiple defect features corresponding to the abnormal regions; According to the defect type order recorded in the defect priority, the defect judgment conditions corresponding to the defect type are matched with the feature values ​​of multiple defect features corresponding to the abnormal region to determine the defect type obtained by detecting the image to be detected. The defect display interface shows the types of defects detected in the image to be inspected.

6. The method according to any one of claims 1 to 5, wherein, The defect parameter tuning interface further includes: a feature state corresponding to each defect feature, wherein the feature state is used to indicate whether the feature value corresponding to the defect feature matches the initial defect truth value range; and / or, a display control and an adjustment control corresponding to each initial defect truth value range.

7. The method according to any one of claims 1-5, wherein, The defect display interface further includes: the image to be detected and a display box for the abnormal area corresponding to the defect type obtained by detecting the image to be detected; and / or, The filtered defects corresponding to the image to be detected are used to represent defects in abnormal regions where the feature values ​​of multiple defect features in the abnormal region of the image to be detected do not match the multiple defect judgment conditions corresponding to the multiple defect types.

8. A defect detection parameter tuning method, comprising: The interface displays the parameter tuning interface corresponding to the initial defect type; wherein, the initial defect type is the defect type obtained by preliminary detection of the image to be detected; the parameter tuning interface includes a defect selection control and a data learning control; the defect selection control is used to select the actual target defect type corresponding to the image to be detected; In response to a trigger operation on the data learning control, the initial defect truth value range corresponding to the target defect type is adjusted based on the feature value detected for the initial defect type to obtain the target defect truth value range; wherein the target defect truth value range matches the feature value.

9. The method of claim 8, wherein, Before displaying the defect parameter tuning interface corresponding to the initial defect type, the method further includes: A defect display interface is provided; the defect display interface displays the initial defect type. If a parameter tuning operation is detected for the initial defect type in the defect display interface, the step of displaying the defect parameter tuning interface corresponding to the initial defect type is triggered.

10. The method of claim 8, wherein, The step of adjusting the initial defect truth range corresponding to the target defect type based on the feature values ​​detected for the initial defect type to obtain the target defect truth range includes: The feature values ​​detected for the initial defect type and the initial defect truth value range corresponding to the target defect type are input into the pre-trained target parameter prediction model corresponding to the target defect type to obtain the target defect truth value range; the target parameter prediction model is used to adjust the initial defect truth value range corresponding to the target defect type based on the feature values ​​detected for the initial defect type to obtain the target defect truth value range.

11. The method of claim 9, wherein, The defect display interface further includes: the image to be detected and a display box for the abnormal area corresponding to the initial defect type; and / or, The filtered defects corresponding to the image to be detected are used to represent defects in abnormal regions of the image to be detected where the feature values ​​of multiple defect features do not match the defect judgment conditions corresponding to each of the multiple defect types.

12. The method according to any one of claims 8-11, wherein, The defect parameter tuning interface also includes feature values ​​detected for the initial defect type; And / or, The defect parameter tuning interface further includes: feature status, and / or, display controls and adjustment controls corresponding to the initial defect truth value range; the feature status is used to indicate whether the feature value detected for the initial defect type matches the initial defect truth value range.

13. A defect detection and parameter tuning device, comprising: The display module is configured to display the defect parameter tuning interface corresponding to the initial defect type in response to the parameter tuning operation of the initial defect type in the defect display interface. The defect display interface shows the initial defect types obtained from the detection of the image to be inspected; The defect parameter tuning interface includes: multiple feature values ​​corresponding to multiple defect features of the abnormal region corresponding to the initial defect type, a defect selection control, and a data learning control; The adjustment module is configured to respond to a trigger operation on the data learning control, obtain multiple initial defect truth value ranges corresponding to the target defect type in the defect selection control, and adjust the multiple initial defect truth value ranges corresponding to the target defect type based on multiple feature values ​​of the abnormal region corresponding to the initial defect type, thereby obtaining multiple target defect truth value ranges corresponding to the target defect type; wherein, the target defect truth value range corresponding to each defect feature matches the feature value corresponding to the defect feature; The update module is configured to update the defect judgment conditions corresponding to the target defect type based on the truth value range of the multiple target defects.

14. A defect detection and parameter tuning device, comprising: The first display module is configured to display the defect parameter tuning interface corresponding to the initial defect type; wherein, the initial defect type is the defect type obtained by preliminary detection of the image to be detected; the defect parameter tuning interface includes a defect selection control and a data learning control; the defect selection control is used to select the target defect type actually corresponding to the image to be detected; The target adjustment module is configured to respond to a trigger operation of the data learning control, and adjust the initial defect truth value range corresponding to the target defect type based on the feature values ​​detected for the initial defect type, to obtain the target defect truth value range; wherein the target defect truth value range matches the feature values.

15. An electronic device comprising: One or more processors; one or more memories; The one or more memories are used to store computer program code, which includes computer instructions. When the one or more processors execute the computer instructions, the electronic device performs the defect detection parameter tuning method as described in any one of claims 1 to 12.

16. A computer-readable storage medium storing computer-executable instructions that, when executed on a computer, cause the computer to perform the defect detection parameter tuning method as described in any one of claims 1 to 12.

17. A computer program product comprising a computer program / instructions that, when executed by a processor, implement the defect detection parameter tuning method according to any one of claims 1-12.

18. A computer program, which, when run on a computer, causes the computer to perform the defect detection parameter tuning method according to any one of claims 1-12.

Images