Sub-pixel defect detection method and device based on imaging characteristics and medium

Abstract

The invention relates to the technical field of semiconductor detection, discloses a sub-pixel defect detection method, and particularly relates to a sub-pixel defect detection method and device based on imaging characteristics and a medium. According to the method, the features of the image are extracted, the extracted features are subjected to abnormal screening according to the distribution characteristics of the sub-pixel defects to determine the candidate defects, and the target sub-pixel defect in the candidate defects is determined based on the statistical features. The whole method does not depend on manual screening, automatic sub-pixel defect detection is achieved through the characteristic distribution characteristics of the sub-pixels, the overall detection efficiency is improved, and the detection cost is reduced.

Description

Technical Field

[0001] This application relates to the field of semiconductor detection technology, and is a sub-pixel defect detection method, specifically a sub-pixel defect detection method, device and medium based on imaging features. Background Art

[0002] The defect detection ability of CMOS chips has been continuously enhanced with the improvement of shooting requirements. Currently, the manufacturing processes of some mainstream manufacturers can basically reach an accuracy of 0.5um - 1um, which greatly improves the requirements for product defect detection equipment. To detect such extremely small defects, there are generally two methods. The first is to detect under a high-magnification imaging system. The field of view is very small under high magnification, but the resolution is very high, and the pixel size can reach 0.5um / pixel, which is equivalent to detecting defects of 1 - 2 pixels on the image. However, the field of view of this method is too small, resulting in the need to take many photos of a product to capture the whole, which greatly reduces the productivity of the equipment. The second method is to solve it through algorithms. Currently, the mainstream method in the market is to tighten the specifications of the algorithm, then detect a large number of defects, and then manually screen and confirm through high-magnification back-capture images. Since this method will detect a lot of defects, the workload of manual re-inspection will increase greatly, resulting in an increase in labor costs. Summary of the Invention

[0003] To solve the above problems, this application provides a sub-pixel defect detection method, device and medium based on imaging features, which can accurately detect sub-pixel defects according to the imaging features of extremely small pixels.

[0004] To achieve the above object, the technical solutions adopted in the embodiments of this application are as follows:

[0005] In the first aspect, a sub-pixel detection method based on imaging features is provided. The method includes: segmenting the detection area of the acquired panoramic image to obtain a regional image that only contains the functional area; extracting features from the regional image within a target pixel range, and updating based on the extracted features to obtain a feature image of the regional image; determining the abnormal feature distribution on the feature image based on a standard template image, and determining whether there are defects on the regional image according to the abnormal feature distribution; determining the category of the defect according to the statistical features of the defect; the categories of the defect include noise defects and target sub-pixel defects.

[0006] In some specific implementation manners, the segmenting the detection area of the acquired panoramic image includes: determining a mask image corresponding to the size and / or corresponding area based on the detection task, and multiplying the mask image with the panoramic image to obtain an image of interest corresponding to the size and / or corresponding area, and the image of interest is the regional image.

[0007] In some specific implementation manners, the feature extraction of the regional image with the target pixel range includes: summing the pixel values within the target pixel range in the regional image based on a 3×3 convolution kernel to obtain a new feature image.

[0008] In some specific implementation manners, the target pixel range is a 2×2 or 3×3 pixel range within any pixel point area in the regional image.

[0009] In some specific implementation manners, determining the abnormal feature distribution on the feature image based on the standard template image and determining whether there are defects on the regional image according to the abnormal feature distribution includes: performing 3×3 convolution processing on the standard template image, calculating the difference between the convolved template image and the feature image to obtain a difference image, and determining whether there are defects based on the distribution of the differences on the difference image; when the difference exceeds the preset difference value, it is determined as a candidate defect.

[0010] In some specific implementation manners, determining the category of the defect according to the statistical features of the defect, wherein the variance of each candidate defect within the recognition area is determined, and the category of the defect is determined according to the distribution of the variance.

[0011] In some specific implementation manners, determining the category of the defect according to the distribution of the variance includes: when the variance does not exceed the preset variance threshold, the candidate defect is a noise defect; when the variance exceeds the preset variance threshold, the candidate defect is a sub-pixel defect. In some specific implementation manners, obtaining the features in each unit wafer image includes: segmenting the unit wafer image to obtain a plurality of foreground images; calculating the energy value corresponding to each foreground image, and the energy value is used to characterize the surface feature.

[0012] In a second aspect, a detection device is provided, including: a region segmentation unit for performing detection region segmentation on the acquired panoramic image to obtain a region image only including a functional region; a feature extraction unit for performing feature extraction on the region image with a target pixel range and updating based on the extracted features to obtain a feature image of the region image; a defect screening unit for determining the abnormal feature distribution on the feature image based on the standard template image and determining whether there are defects on the region image according to the abnormal feature distribution; and a defect classification unit for determining the category of the defect according to the statistical features of the defect.

[0013] In a third aspect, a terminal device is provided, including: a processor and a memory connected to the processor, where the memory stores instructions executed by the processor, and the instructions cause the processor to perform operations to carry out the sub-pixel defect detection method as described in any one of the above.

[0014] In a fourth aspect, a readable medium is provided, where the readable medium stores computer-readable instructions, and the computer-readable instructions include instructions for executing the sub-pixel defect detection method as described in any one of the above.

[0015] In the technical solution provided by the embodiments of the present application, by extracting features from an image, and based on the distribution characteristics of sub-pixel defects, the extracted features are screened for anomalies to determine candidate defects, and the target sub-pixel defects among the candidate defects are determined based on statistical features. The overall method does not rely on manual screening, and realizes automated sub-pixel defect detection through the feature distribution characteristics of sub-pixels, improving the overall detection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0017] The methods, systems, and / or programs in the drawings will be further described according to exemplary embodiments. These exemplary embodiments will be described in detail with reference to the drawings. These exemplary embodiments are non-limiting exemplary embodiments, where the example numbers represent similar mechanisms in each view of the drawings.

[0018] Figure 1 is a schematic diagram of the sub-pixel defect distribution characteristics provided by the embodiments of the present application.

[0019] Figure 2 is a schematic diagram of the sub-pixel defect detection process provided by the embodiments of the present application.

[0020] Figure 3 is a schematic diagram of the structure of the detection device provided by the embodiments of the present application.

[0021] Figure 4 is a schematic diagram of the structure of the server provided by the embodiments of the present application.

[0022] Figure 5 is a schematic diagram of the structure of the computer-readable storage medium provided by the embodiments of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0023] In order to better understand the above technical scheme, the technical scheme of the present application is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical scheme of the present application, rather than limitations on the technical scheme of the present application. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other.

[0024] In the following detailed description, numerous specific details are set forth by way of example in order to provide a comprehensive understanding of the relevant guidance. However, it will be apparent to those skilled in the art that the present application may be practiced without these details. In other cases, well-known methods, procedures, systems, compositions and / or circuits have been described at a relatively high level, without detail, in order to avoid unnecessarily obscuring aspects of the present application.

[0025] Flowcharts are used in the present application to illustrate the execution process performed by the system according to the embodiment of the present application. It should be clearly understood that the execution process of the flowchart may not be performed in order. On the contrary, these execution processes may be performed in reverse order or simultaneously. In addition, at least one other execution process may be added to the flowchart. One or more execution processes may be deleted from the flowchart.

[0026] Before further describing the embodiments of the present invention in detail, the nouns and terms involved in the embodiments of the present invention are described. The nouns and terms involved in the embodiments of the present invention are subject to the following interpretations.

[0027] (1) In response, it is used to indicate the conditions or states on which the executed operations depend. When the dependent conditions or states are met, one or more operations executed may be in real time or with a set delay. Unless otherwise specified, there is no restriction on the order in which the multiple operations executed are executed.

[0028] (2) Based on is used to indicate the conditions or states on which the executed operations depend. When the dependent conditions or states are met, one or more operations executed may be in real time or have a set delay. Unless otherwise specified, there is no restriction on the order in which the multiple operations executed are executed.

[0029] The present application provides a sub-pixel defect detection method based on extremely small pixel imaging features. The present applicant has found that the characteristic distribution of sub-pixel defects can basically present a diffuse distribution after imaging. That is, a sub-pixel defect will cause the values ​​of the surrounding 2*2 or 3*3 pixels to change, but the degree of change of each pixel is small, which is equivalent to the energy being shared. Figure 1, which is a schematic diagram of the distribution of sub-pixel defects. Referring to this figure, it can be seen that for a single sub-pixel defect, the gray-scale difference of the sub-pixel defect is very small, and it will be impossible to screen out the defect when screening at the single-pixel level. However, when screening according to the pixels within a certain range, the defect can be detected through an algorithm. For example Figure 1 As shown, for a very small defect that affects four pixels, the ability is dispersed to the four pixels.

[0030] Therefore, the applicant provides a sub-pixel defect detection method. For the detection of sub-pixel defects, the pixel values within a certain range can be summed, and then the potential sub-pixel defects can be determined according to the change in the sum of the pixel values. This method is based on the characteristic distribution of sub-pixel defects and realizes the automatic detection of sub-pixel defects through pixel calculation, improving the problem that sub-pixel defect screening depends on manual work. Moreover, for this method, only the characteristic distribution of the pixels in the image needs to be obtained to screen the defects. Compared with the existing sub-pixel defect detection methods, on the basis of the same detection accuracy, its processing cost and processing efficiency are higher than those of the existing sub-pixel defect automatic detection methods.

[0031] Specifically, referring to Figure 2 , the method in this embodiment includes the following steps:

[0032] Step S21. Segment the detection area of the obtained panoramic image to obtain a regional image that only contains the functional area.

[0033] In this embodiment, for the detection object being a wafer, the panoramic image is a complete wafer image obtained by scanning the entire surface of the wafer through an optical system. According to the structural distribution of the wafer, it includes dies, functional areas composed of multiple dies, and other non-functional areas such as edge areas. The detection area in this embodiment is the above-mentioned functional area composed of dies. Therefore, the initially obtained complete image is not the target image to be detected, and the detection area in the complete initial image needs to be segmented to obtain a target image that only contains the functional area. This process in image processing can be called the region of interest, that is, the ROI region extraction process.

[0034] Among them, for this process, first determine the size and / or the range of the corresponding area to be detected according to the detection task, and sample a pre-made mask image according to this range. Multiply this mask image with the panoramic image to obtain the corresponding image of interest. In this embodiment, the image values within the region of interest remain unchanged, and the image values outside the region are 0.

[0035] Step S22. Extract features from the regional image within the target pixel range, and update based on the extracted features to obtain a feature image of the regional image.

[0036] In this embodiment, the target pixel range refers to the 2×2 or 3×3 pixel range around any pixel point in the regional image. In essence, this pixel range is the energy distribution range of sub-pixel defects, because after imaging, sub-pixel defects appear as diffuse dispersions. For one sub-pixel defect, it will change the pixel values of the surrounding 2×2 or 3×3 pixels, but the degree of change of each pixel is very small.

[0037] In this embodiment, by summing the pixel values within this range, the feature distribution within this range is obtained, and then it is determined whether there are sub-pixel defects according to the distribution. Therefore, in this embodiment, for this process, it is necessary to sum the features, that is, the pixel values, within the 2×2 or 3×3 pixel range around any pixel point to form a new feature image, and in the subsequent process, defects are identified according to the distribution of the features of the feature image.

[0038] In this embodiment, this process is completed by using a 3×3 convolution kernel, which can cover the influence range of the defect and improve the accuracy of feature extraction.

[0039] Step S23. Determine the abnormal feature distribution on the feature image based on the standard template image, and determine whether there are defects on the regional image according to the abnormal feature distribution.

[0040] For the feature image within the region of interest obtained by summing the pixel values within the pixel point range in step S22, it is used to statistically analyze the feature distribution. And for whether there are abnormal features and defects in the above-mentioned region of interest image, in this embodiment, it is determined by the method of template matching. For the features in this embodiment, they are pixel values, and whether the features are abnormal is judged according to whether the pixel values are abnormal. Therefore, the process of template matching in this embodiment is to subtract the feature map from the standard template image to obtain the difference of pixel values, that is, to form a difference map. Then, it is determined whether there is an abnormality according to the distribution of the pixel differences.

[0041] Specifically, when the pixel difference exceeds the preset difference, it is considered a candidate defect. In this embodiment, the image difference method is the most direct defect detection method, and its principle is to subtract the pixel gray value at the corresponding coordinate (x, y) in the image to be detected from the standard image, and output the absolute value of the gray difference to the difference result image.

[0042] In this embodiment, the standard template image is a perfect image obtained by image acquisition of a standard wafer or a perfect image constructed by image processing technology for images corresponding to multiple wafers, and there are no defects in the standard template image. In order to calculate the difference between the standard template image and the feature map, the standard template image is also subjected to a 3*3 convolution, and the difference map is obtained by calculating the convolution-processed template image and the feature map.

[0043] Among them, before calculating the difference between the standard template image and the feature map in this embodiment, the standard template image and the feature map need to be aligned.

[0044] Specifically, in this embodiment, to align them, first, the relationship between the region of interest image and the standard target image without convolution processing is obtained through a template matching algorithm. This relationship is the affine transformation matrix between the two, which is used to indicate the relative position relationship between the region of interest image and the standard template image. Since the position relationship between the feature map and the region of interest image remains the same, the position of the feature map is updated based on this affine transformation matrix, so that the feature map can be aligned with the convolution-processed standard template image.

[0045] Step S24. Determine the category of the defect according to the statistical characteristics of the defect.

[0046] In this embodiment, the categories of defects include noise defects and target sub-pixel defects. The abnormal distribution detected in step S23 includes not only the sub-pixel defects to be determined but also the defects corresponding to noise. Therefore, the noise defects need to be screened in the final defect determination stage.

[0047] Since the distribution of noise is generally independent, in this embodiment, according to the characteristics of noise, the variance within the 3*3 region of each candidate defect in step S23 is calculated in this embodiment. When the variance size exceeds the preset variance threshold, it indicates that there is an obvious pixel value jump within this range, indicating that this defect is a noise defect. When the variance size does not exceed the preset variance threshold, it indicates that the gray value is uniform within this range, indicating that this defect is a sub-pixel defect.

[0048] To better implement the above method, the embodiment of the present application also provides a detection device, which can be specifically integrated in an electronic device, and the electronic device can be a terminal, a server, etc. Among them, the terminal can be a mobile phone, a tablet computer, a smart Bluetooth device, a laptop computer, a personal computer, etc.; the server can be a single server or a server cluster composed of multiple servers.

[0049] For example, in this embodiment, taking the detection device being specifically integrated into an electronic device as an example, the method of the embodiment of the present application will be described in detail.

[0050] For example, as Figure 3 shown, the detection device 30 may include a region segmentation unit 31, a feature extraction unit 32, a defect screening unit 33, and a defect classification unit 34, where:

[0051] The region segmentation unit 31 is configured to perform detection region segmentation on the acquired panoramic image to obtain a region image that only contains the functional region.

[0052] In some specific embodiments, for the detection region segmentation, a mask image with a corresponding size or / and a corresponding region is determined based on the detection task, and the mask image is multiplied by the panoramic image to obtain an image of interest with a corresponding size or / and a corresponding region, and the image of interest is the region image.

[0053] The feature extraction unit 32 is configured to extract features from the region image within a target pixel range, and update based on the extracted features to obtain a feature image of the region image.

[0054] In some specific embodiments, for the feature extraction of the region image, the pixel values within the target pixel range in the region image are summed based on a 3*3 convolutional kernel to obtain a new feature image.

[0055] In some specific embodiments, the target pixel range is a 2*2 or 3*3 pixel range within any one pixel point region of the region image.

[0056] The defect screening unit 33 is configured to determine the abnormal feature distribution on the feature image based on a standard template image, and determine whether there are defects on the region image according to the abnormal feature distribution.

[0057] In some specific embodiments, defect screening is performed by performing 3*3 convolution processing on the standard template image, calculating the difference between the convolved template image and the feature image to obtain a difference image, and determining whether there are defects based on the distribution of the differences on the difference image; when the difference exceeds the preset difference value, it is determined as a candidate defect.

[0058] The defect classification unit 34 is configured to determine the category of the defect according to the statistical features of the defect.

[0059] In some specific embodiments, the screening of the defect category is performed by determining the variance of each candidate defect within the recognition region, and determining the category of the defect according to the distribution of the variance.

[0060] In some specific embodiments, when the variance does not exceed a preset variance threshold, the candidate defect is a noise defect, and when the variance exceeds the preset variance threshold, the candidate defect is a sub-pixel defect.

[0061] Regarding the sub-pixel defect detection method provided in the embodiments of the present application, by extracting features from an image, and based on the distribution characteristics of sub-pixel defects, screening the extracted features for anomalies to determine candidate defects, and determining the target sub-pixel defects among the candidate defects based on statistical features. The overall method does not rely on manual screening, and realizes automated sub-pixel defect detection through the distribution characteristics of sub-pixel features, improving the efficiency of the overall detection.

[0062] The embodiments of the present application further provide an electronic device, which can be a device such as a terminal, a server, etc. Among them, the terminal can be a mobile phone, a tablet computer, a smart Bluetooth device, a laptop computer, a personal computer, and so on; the server can be a single server or a server cluster composed of multiple servers, and so on.

[0063] In some embodiments, the detection device can also be integrated in multiple electronic devices. For example, the chip detection device can be integrated in multiple servers, and the chip detection method of the present application can be implemented by multiple servers.

[0064] In this embodiment, the electronic device in this embodiment will be described in detail taking the server as an example. For example, as Figure 4 shown, it shows a schematic structural diagram of the server involved in the embodiments of the present application. Specifically:

[0065] The server may include a processor 401 with one or more processing cores, a memory 402 with one or more computer-readable storage media, a power supply 403, an input module 404, and a communication module 405 and other components. Those skilled in the art can understand that Figure 4 the server structure shown in

[0066] The processor 401 is the control center of the server, connecting various parts of the entire server through various interfaces and lines. By running or executing software programs and / or modules stored in the memory 402, and by invoking the data stored in the memory 402, it performs various functions of the server and processes data. In some embodiments, the processor 401 may include one or more processing cores; in some embodiments, the processor 401 may integrate an application processor and a modem processor. Among them, the application processor mainly processes the operating system, user interface, application programs, etc., and the modem processor mainly processes wireless communications. It can be understood that the above-mentioned modem processor may not be integrated into the processor 401 either.

[0067] The memory 402 can be used to store software programs and modules. The processor 401 executes various functional applications and data processing by running the software programs and modules stored in the memory 402. The memory 402 mainly includes a program storage area and a data storage area. Among them, the program storage area can store the operating system, application programs required for at least one function (such as the sound playback function, image playback function, etc.); the data storage area can store data created according to the use of the server, etc. In addition, the memory 402 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid-state storage devices. Correspondingly, the memory 402 may also include a memory controller to provide the processor 401 with access to the memory 402.

[0068] The server also includes a power supply 403 that powers each component. In some embodiments, the power supply 403 can be logically connected to the processor 401 through a power management system, so as to realize functions such as management of charging, discharging, and power consumption management through the power management system. The power supply 403 may also include any components such as one or more DC or AC power supplies, a recharge system, a power failure detection circuit, a power converter or inverter, and a power status indicator.

[0069] The server may also include an input module 404, which can be used to receive input digital or character information, and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function controls.

[0070] The server may also include a communication module 405. In some embodiments, the communication module 405 may include a wireless module. The server can perform short-range wireless transmission through the wireless module of the communication module 405, thereby providing users with wireless broadband Internet access. For example, the communication module 405 can be used to help users send and receive emails, browse web pages, and access streaming media, etc.

[0071] Although not shown, the server may further include a display unit and the like, which will not be elaborated herein. Specifically, in this embodiment, the processor 401 in the server will load the executable files corresponding to the processes of one or more application programs into the memory 402 according to the following instructions, and the processor 401 will run the application programs stored in the memory 402, so as to implement the steps in the methods of the embodiments of the present application.

[0072] For the specific implementation of each of the above operations, reference may be made to the previous embodiments, which will not be elaborated herein.

[0073] As can be seen from the above, the method provided in this embodiment extracts features from the image, and determines candidate defects by abnormally screening the extracted features according to the distribution characteristics of sub-pixel defects, and determines the target sub-pixel defects in the candidate defects based on statistical features. The overall method does not rely on manual screening, and realizes automatic sub-pixel defect detection through the feature distribution characteristics of sub-pixels, improving the overall detection efficiency.

[0074] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions, or by controlling related hardware through instructions. The instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0075] For this reason, refer to Figure 5 , the embodiment of the present application provides a computer-readable storage medium 50. The readable storage medium 50 stores computer-readable instructions 501, and the computer-readable instructions 501 can be loaded by a processor to execute the steps in any sub-pixel defect detection method provided by the embodiment of the present application. For example, the instructions can execute the following steps:

[0076] Perform detection area segmentation on the acquired panoramic image to obtain a region image that only contains the functional area;

[0077] Extract features from the region image within the target pixel range, and update based on the extracted features to obtain a feature image of the region image;

[0078] Determine the abnormal feature distribution on the feature image based on the standard template image, and determine whether there are defects on the region image according to the abnormal feature distribution;

[0079] Determine the category of the defect according to the statistical features of the defect.

[0080] Among them, the storage medium may include: read-only memory (ROM, Read Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disc, etc.

[0081] According to one aspect of the present application, a computer program product or a computer program is provided. The computer program product or the computer program includes computer programs / instructions, and the computer programs / instructions are stored in a computer-readable storage medium. A processor of an electronic device reads the computer programs / instructions from the computer-readable storage medium, and the processor executes the computer programs / instructions, so that the electronic device executes the methods provided in various alternative implementations of the chip detection aspect in the above embodiments.

[0082] Since the instructions stored in the storage medium can execute the steps in any of the sub-pixel defect detection methods provided in the embodiments of the present application, therefore, the beneficial effects that can be achieved by any of the sub-pixel defect detection methods provided in the embodiments of the present application can be realized. For details, refer to the previous embodiments and will not be elaborated here.

[0083] In the present application, "at least one" means one or more, and "a plurality" means two or more. "At least one of the following items (pieces)" or its similar expressions refer to any combination of these items, including any combination of single item (piece) or plural items (pieces). For example, at least one of a, b, or c can represent: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, c can be single or multiple.

[0084] It should be understood that in various embodiments of the present application, the magnitudes of the serial numbers of the above processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0085] The above has introduced in detail a sub-pixel defect detection method, device, terminal, storage medium, and program product based on imaging features provided by the embodiments of the present application. Specific examples are used in this article to elaborate on the principle and implementation manner of the present application. The description of the above embodiments is only used to help understand the method and its core idea of the present application; at the same time, for those skilled in the art, according to the idea of the present application, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present application.

Claims

1. A sub-pixel defect detection method based on imaging features, characterized in that The method includes: Performing detection region segmentation on the acquired panoramic image to obtain a region image containing only the functional region; Performing feature extraction on the region image within a target pixel range, and updating based on the extracted features to obtain a feature image of the region image; Determining the abnormal feature distribution on the feature image based on a standard template image, and determining whether there are defects on the region image according to the abnormal feature distribution; Determining the category of the defect according to the statistical features of the defect; the categories of the defect include noise defect and target sub-pixel defect.

2. The sub-pixel defect detection method based on imaging features according to claim 1, wherein The performing detection region segmentation on the acquired panoramic image includes: determining a mask image of a corresponding size or / and a corresponding region based on the detection task, and multiplying the mask image with the panoramic image to obtain an image of interest of a corresponding size or / and a corresponding region, and the image of interest is the region image.

3. The sub-pixel defect detection method based on imaging features according to claim 1, characterized in that The performing feature extraction on the region image within a target pixel range includes: summing the pixel values within the target pixel range in the region image based on a 3×3 convolution kernel to obtain a new feature image.

4. The sub-pixel defect detection method based on imaging features according to claim 3, wherein The target pixel range is a 2×2 or 3×3 pixel range within any pixel point region in the region image.

5. The sub-pixel defect detection method based on imaging features according to claim 4, characterized in that, The determining the abnormal feature distribution on the feature image based on a standard template image, and determining whether there are defects on the region image according to the abnormal feature distribution includes: performing 3×3 convolution processing on the standard template image, calculating the difference between the convolved template image and the feature image to obtain a difference image, and determining whether there are defects based on the distribution of the differences on the difference image; when the difference exceeds a preset difference value, it is determined as a candidate defect.

6. The sub-pixel defect detection method based on imaging features according to claim 5, wherein The determining the category of the defect according to the statistical features of the defect is characterized in that the variance of each candidate defect within the recognition region is determined, and the category of the defect is determined according to the distribution of the variance.

7. The sub-pixel defect detection method based on imaging features according to claim 6, characterized in that, The determining the category of the defect according to the distribution of the variance includes: when the variance does not exceed a preset variance threshold, the candidate defect is a noise defect; when the variance exceeds the preset variance threshold, the candidate defect is a sub-pixel defect.

8. A detection device, characterized in that, It includes: A region segmentation unit, configured to perform detection region segmentation on the acquired panoramic image to obtain a region image containing only the functional region; A feature extraction unit, configured to perform feature extraction on the region image within a target pixel range, and update based on the extracted features to obtain a feature image of the region image; A defect screening unit, configured to determine the abnormal feature distribution on the feature image based on a standard template image, and determine whether there are defects on the region image according to the abnormal feature distribution; A defect classification unit, configured to determine the category of the defect according to the statistical features of the defect.

9. A terminal device, characterized in that, It includes: A processor and a memory connected to the processor, the memory stores instructions executed by the processor, and the instructions cause the processor to perform operations to perform the detection method according to any one of claims 1-7.

10. A readable medium, characterized in that, The readable medium stores computer-readable instructions, the computer-readable instructions including instructions for performing the detection method according to any one of claims 1-7.

Images