Infrared image and visible light image distinguishing method and device in live detection

Abstract

The invention provides a method and a device for distinguishing an infrared image and a visible light image in live detection, relates to the technical field of live detection, and solves the problems of low manual judgment efficiency and possible errors caused by non-uniform distinguishing standards of infrared images and visible light images shot by different manufacturers in the field. The type of the to-be-detected image can be accurately and efficiently determined, so that a background system can accurately judge the operation condition of equipment based on the accurate image type. The method comprises the following steps: acquiring a to-be-detected image of power equipment in a running state; according to the filename of the to-be-detected image, determining whether a database contains an association relationship matched with the key information of the filename, the association relationship being used for associating the key information with the type of the image, the type being an infrared image or a visible light image; if the database contains the association relationship matched with the key information, determining the type of the to-be-detected image according to the association relationship; otherwise, determining the type of the to-be-detected image according to the pixel features of the to-be-detected image.

Description

Technical Field

[0001] The present application relates to the technical field of charge detection, and in particular to a method and device for distinguishing between infrared images and visible light images in charge detection. Background Art

[0002] Live inspection technology is a non-destructive testing method performed on operating electrical equipment. It aims to promptly detect potential equipment failures and ensure the safe and stable operation of the power system. Routine live inspection and testing of primary and secondary equipment in normal operation is a crucial component of power grid operations and maintenance. This work involves a wide variety of main equipment types, a wide range of sensor types used in operation, a large amount of data required for manual recording and export, and a high level of complexity in inspection scheduling and management.

[0003] Inspection equipment is required at sites within power systems, including power generation, transmission, and distribution systems, to collect on-site images. These images can be used to predict power system faults, for example by inputting them into fault models. To improve the accuracy and efficiency of fault prediction, different types of on-site images reflecting different site characteristics can be registered and fused. For example, image fusion can be used to combine the thermal detail information of an infrared image with the texture information of a visible light image to generate a composite image containing both temperature and structural information. Finally, image recognition is performed on this composite image to identify abnormalities such as partial discharge and mechanical failures in power equipment, enabling timely detection and resolution of potential safety hazards. Therefore, the backend system must clearly identify the image type, whether infrared or visible light, in order to accurately carry out the subsequent processing steps.

[0004] Currently, various inspection equipment manufacturers exhibit significant discrepancies in how they distinguish between infrared and visible light images, with some manufacturers even failing to differentiate between the two. Due to the lack of a unified standard for differentiation, backend systems receiving image data from inspection equipment from multiple manufacturers must rely on manual differentiation of image types. However, manual differentiation is not only inefficient but also labor-intensive and time-consuming when dealing with large volumes of inspection images. Therefore, a solution for efficiently distinguishing between infrared and visible light images is urgently needed. Summary of the Invention

[0005] The present application provides a method and device for distinguishing between infrared images and visible light images in live detection, which solves the problem in this field that inspection equipment from different manufacturers have inconsistent standards for distinguishing between infrared images and visible light images. It can accurately and efficiently determine the type of image to be detected, effectively reducing the possible misclassification that may occur during manual judgment, and helps the background system to more accurately judge the equipment operation status based on the accurate image type, timely discover potential problems, thereby effectively ensuring the safe and stable operation of the equipment.

[0006] In a first aspect, a method for distinguishing infrared images from visible light images in charge detection is provided, comprising: Acquire images of power equipment to be inspected in an operating state; According to the file name of the image to be detected, determine whether the database contains an association relationship that matches the key information of the file name, and the association relationship is used to associate the key information with the type of the image, which is an infrared image or a visible light image; If the database contains an association relationship that matches the key information, the type of the image to be detected is determined based on the association relationship; Otherwise, the type of the image to be detected is determined according to the pixel features of the image to be detected.

[0007] In a feasible design, determining the type of the image to be detected based on the pixel features of the image to be detected includes: Obtain the target grayscale histogram of the image to be detected; The type of the image to be detected is determined based on the skewness and / or kurtosis of the target grayscale histogram.

[0008] In a feasible design, the type of the image to be detected is determined based on the skewness and / or kurtosis of the target grayscale histogram, including: If the skewness of the target grayscale histogram is within a predetermined first range, or the kurtosis of the target grayscale histogram is within a predetermined second range, determining that the type of the image to be detected is an infrared image; Otherwise, it is determined that the type of the image to be detected is a visible light image.

[0009] In a feasible design, if the skewness of the target grayscale histogram is outside the first range and the kurtosis of the target grayscale histogram is outside the second range, the method further includes: The type of the image to be detected is determined according to the green channel value of the pixel of the image to be detected.

[0010] In one possible design, the method further includes: By traversing pixels of the visible light images in the database, a first average value is determined, where the first average value is an average value of green channel values of each visible light image in the database.

[0011] In a feasible design, determining the type of the image to be detected according to the green channel value of the pixel of the image to be detected includes: Determine a second average value by traversing pixels of the image to be detected, where the second average value is an average value of each green channel value of the image to be detected; comparing the second mean with the first mean; If the second average value is less than the first average value, determining that the type of the image to be detected is an infrared image; Otherwise, the type of the image to be detected is determined to be a visible light image, or the type of the image to be detected is determined according to the file size of the image to be detected.

[0012] In a feasible design, the type of the image to be detected is determined based on the file size of the image to be detected, including: Acquire another image to be detected corresponding to the image to be detected, wherein the other image to be detected and the image to be detected form a pair of visible light images and infrared images for image registration processing; Comparing the file size of the image to be detected with the file size of another image to be detected; If the file size of the image to be detected is larger than the file size of another image to be detected, determining that the type of the image to be detected is an infrared image; Otherwise, it is determined that the type of the image to be detected is a visible light image.

[0013] In a feasible design, the method further includes: when the image to be detected is one of a pair of an infrared image and a visible light image, if the type of the image to be detected is an infrared image, determining that the type of the other image to be detected is a visible light image; If the type of the image to be detected is a visible light image, it is determined that the type of the other image to be detected is an infrared image.

[0014] In a feasible design, the key information is a character or code in the main file name that indicates the type of the image, or the key information is an extension. Based on the file name of the image to be detected, it is determined whether a database contains an association relationship that matches the key information in the file name. If the database contains an association relationship that matches the key information, the type of the image to be detected is determined based on the association relationship, including: Match the main file name of the image to be tested with the characters or codes used by each manufacturer in the database according to the type of image; If the main file name of the image to be detected contains characters or codes in the database, determining, based on the characters or codes contained in the main file name, an association relationship in the database that matches the characters or codes contained in the main file name, and determining the type of the image to be detected based on the association relationship; If the main file name of the image to be detected does not contain characters or codes in the database, and the file name of the image to be detected contains an extension, determining whether the database contains an association relationship that matches the extension of the image to be detected; If the database contains an association relationship that matches the extension of the image to be detected, the type of the image to be detected is determined according to the association relationship.

[0015] In a second aspect, a device for distinguishing infrared images from visible light images during charge detection is provided, comprising: An image acquisition module is used to acquire images of the power equipment to be tested in an operating state; An image type recognition module is used to determine, based on the file name of the image to be detected, whether the database contains an association relationship that matches the key information of the file name. The association relationship is used to associate the key information with the type of the image, which can be an infrared image or a visible light image. The image type recognition module is further configured to, if the database contains an association relationship that matches the key information, determine the type of the image to be detected based on the association relationship; Otherwise, the image type recognition module is further configured to determine the type of the image to be detected according to pixel features of the image to be detected.

[0016] The database of the embodiment of the present application includes a large number of infrared images and visible light images taken by inspection equipment from existing manufacturers. These images come from a wide range of sources and cover a variety of practical application scenarios and equipment models. Therefore, the database of the present application is highly representative and comprehensive. On this basis, based on the file name of the image to be detected, it is basically possible to find the corresponding association in the database, thereby accurately and efficiently determining whether the image to be detected is an infrared image or a visible light image. In rare cases, if the image type cannot be determined based on the image file name and the database, the present application also provides a solution for determining the type of the image to be detected based on the pixel features of the image to be detected. Therefore, no matter how complex the method used by the manufacturer of the inspection equipment to distinguish image types is, or even whether it does not distinguish at all, the solution of the present application can automatically adapt and accurately identify the image type. It can be seen that compared with traditional manual judgment methods, this solution has excellent adaptive capabilities and can achieve accurate distinction of image types without manual intervention. This feature greatly improves data processing efficiency, significantly saves manpower and time costs, enables detection work to be carried out more efficiently and smoothly, and provides strong support for the large-scale development of detection work.

[0017] In an actual live detection scenario, when the solution of the present application is applied to a handheld live detection mobile terminal, after acquiring the infrared image and the visible light image, the solution of the present application is used to quickly and accurately distinguish and process the images, and the type of the image is displayed in the file name using a specific field according to the preset rules. Then, the distinguished image is accurately uploaded to the background system through the live detection function. The background system can accurately determine the type of image based on the preset rules, and efficiently process and analyze the image, so as to achieve accurate flow and effective utilization of image data in the entire live detection workflow. The live detection mobile terminal using this solution can help the members of the power grid operation and maintenance team to realize the real-time processing and uploading of maintenance and experimental data in a digital and wireless manner during routine inspections, which not only reduces the workload of the staff, but also greatly improves the timeliness and accuracy of the data.

[0018] In summary, this application solution effectively reduces the possible misclassification that may occur during manual judgment, efficiently and accurately identifies image types, and provides a more reliable data basis for equipment status assessment; it helps the background system to more accurately judge the equipment operating status based on accurate image types and discover potential problems in a timely manner, thereby effectively ensuring the safe and stable operation of the equipment, improving the quality and reliability of detection work, and reducing the risk of equipment failure. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solution of the present application, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0020] Figure 1 This is a schematic flow chart of a method for distinguishing infrared images from visible light images in charge detection provided by an exemplary embodiment of the present application; Figure 2 This is a schematic flow chart of another method for distinguishing infrared images from visible light images in charge detection provided by an exemplary embodiment of the present application; Figure 3 This is a schematic diagram of a device for distinguishing infrared images from visible light images in charge detection provided by an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0022] Infrared images, acquired through infrared thermal imaging technology, reflect the surface temperature distribution of an object, typically using specific colors or grayscale to represent different temperature zones. Visible light images, captured using visible light, reveal the true color and appearance of an object. Infrared and visible light images have different characteristics, such as resolution, viewing angle, and field of view. Clarifying the image type helps the backend system employ appropriate registration algorithms. Only after the image type is determined can the backend system effectively fuse the thermal details of the infrared image with the texture information of the visible light image.

[0023] It can be seen that distinguishing between infrared and visible light images is essential for the smooth processing of inspection equipment and backend systems. From image upload by inspection equipment to a series of image processing operations in the backend system, each step relies on accurate image type determination. If errors occur in this step, subsequent operations such as image registration, fusion, and recognition will not be performed correctly, and the entire live inspection system will be unable to effectively discover and address safety hazards in power equipment.

[0024] However, currently, after image data captured by inspection equipment is uploaded to the backend system, due to the lack of a unified standard for distinguishing images, the system must rely on manual identification of image types. This approach is not only extremely inefficient but also consumes a significant amount of manpower and time when faced with a massive volume of inspection images. Furthermore, human judgment is highly subjective and can easily lead to misclassification. For example, misidentifying an infrared image as a visible light image can lead to an inaccurate assessment of the equipment's heating condition, affecting the accuracy of equipment maintenance decisions and jeopardizing its safe operation. The inability of the backend to accurately identify the image type will cause the entire data processing process to stall, seriously disrupting the consistency and accuracy of live inspection work.

[0025] Based on the above problems, Figure 1 As shown, the present application provides a method for distinguishing infrared images from visible light images in charge detection, comprising: S110, obtaining an image of the power equipment to be inspected in an operating state.

[0026] It should be noted that during the live detection process, the images of the power equipment to be detected in the operating state can be obtained by taking pictures of the inspection equipment. Usually, the infrared images and visible light images containing the images of the power equipment taken by the inspection equipment appear in pairs. After the images to be detected taken by multiple inspection devices are aggregated into the live detection mobile terminal, the images are processed by type differentiation and then uploaded to the background system. It is understandable that the images to be detected taken by multiple inspection devices can also be uploaded directly to the background system, which will first differentiate the images by type and then determine the operating status of the power equipment through calculation.

[0027] Inspection equipment can be drones moving in the air and inspection robots moving on the ground, such as patrol robots or drones equipped with multimodal (visible light image and infrared image) sensors.

[0028] It should be understood that this application solution can be applied not only to mobile terminals for live-line inspection but also to back-end systems for live-line inspection. When applied to mobile terminals for live-line inspection, differentiation and processing can be performed directly on-site. When the mobile terminals are equipped with a back-end system, fault diagnosis and repair can be performed in real time on-site based on differentiated infrared and visible light images, thereby improving inspection efficiency.

[0029] The live-line inspection mobile terminal is a dedicated device for live-line inspections. Inspectors can use it to perform routine inspections, operations, and maintenance of power equipment. In practical applications, multiple inspection devices can capture a pair of infrared and visible light images of the same power equipment in operation and upload them to the live-line inspection mobile terminal. The mobile terminal then uploads the captured infrared and visible light images to a backend system. The backend system performs image processing on the infrared and visible light images, including image registration, image fusion, and image recognition. Image registration is a key prerequisite for subsequent image processing, as it matches and overlays multiple images of the same scene with different resolutions, perspectives, and fields of view. The backend system aligns the infrared and visible light images through image registration. Then, through image fusion, the thermal details of the infrared image are combined with the texture information of the visible light image to generate a composite image containing temperature and structural information. Finally, image recognition is performed on this composite image to identify abnormalities such as partial discharge and mechanical failures in the power equipment, enabling timely detection and resolution of potential safety hazards.

[0030] Since the images aggregated into the live detection mobile terminal often come from infrared cameras or visible light cameras of inspection equipment from different manufacturers and lack a unified standard for distinguishing image types, the live detection mobile terminal applies the solution of this application to distinguish the image types of the images to be detected. This can improve the work efficiency of the power grid operation and maintenance team members, as well as the accuracy of image type distinction, thereby improving the ability of the background system to identify abnormal conditions such as partial discharge and mechanical failure in power equipment.

[0031] S120 , determining, based on the file name of the image to be detected, whether a database contains an association relationship that matches key information of the file name.

[0032] The association relationship is used to associate key information with the type of image, which is infrared image or visible light image.

[0033] This application constructs the database in the following ways: First, a large number of infrared images and visible light images taken by inspection equipment from existing commonly used manufacturers are widely collected and stored in a database. Then, these images are systematically organized and analyzed, and key information related to distinguishing between infrared images and visible light images is extracted and saved. For example, the characters or codes used by each manufacturer to name the image according to the image type, or the specific format suffix (i.e., extension) used to distinguish the image type, etc. Finally, various associations are created to associate key information with the image type and saved in the database.

[0034] The images in the database come from a wide range of sources, covering a variety of practical application scenarios and device models, ensuring that the database is highly representative and comprehensive. The database will serve as a solid foundation for subsequent image type recognition and provide a rich reference for algorithms.

[0035] S130: If the database contains an association relationship that matches the key information, determine the type of the image to be detected according to the association relationship.

[0036] In one feasible design, the image's file name can be used to quickly determine the image's type. The file name includes the main file name and may also include an extension. The key information is the characters or code in the main file name that indicates the image's type, or the key information is the extension. Accordingly, an association is used to associate the characters in the main file name with the image's type, or an association is used to associate the code in the main file name with the image's type, or an association is used to associate the extension with the image's type. In this case, this is achieved by determining, based on the file name of the image to be detected, whether a database contains an association that matches the file name. If the database contains an association that matches the file name, the type of the image to be detected is determined based on the association: Match the main file name of the image to be tested with the characters or codes used by each manufacturer in the database according to the type of image; If the main file name of the image to be detected contains characters or codes in the database, determining, based on the characters or codes contained in the main file name, an association relationship in the database that matches the characters or codes contained in the main file name, and determining the type of the image to be detected based on the association relationship; If the main file name of the image to be detected does not contain characters or codes in the database, and the file name of the image to be detected contains an extension, determining whether the database contains an association relationship that matches the extension of the image to be detected; If the database contains an association relationship that matches the extension of the image to be detected, the type of the image to be detected is determined according to the association relationship.

[0037] The above example first matches the main file name of the image to be detected with characters or codes in a database. If the main file name of the image to be detected contains characters or codes in the database, it means that there is a matching association in the database for the characters or codes in the main file name. The type of the image to be detected can be efficiently determined directly based on this association. For example, if the file name contains the word "infrared" or a specific infrared-related code, it is determined to be an infrared image; if it contains the word "visible light" or a corresponding code, it is determined to be a visible light image. If the main file name of the image to be detected does not contain characters or codes in the database, it means that the main file name does not contain the characters or codes indicating the image type, or the database does not store the characters or codes in the main file name and the corresponding association. The image type can be further determined by identifying a matching association based on the extension. For example, if the image extension is ".IR.jpg," ".ir.jpg," or ".irp.jpg," the image type is infrared; if the image extension is ".DC.jpg" or ".vi.jpg," the image type is visible light. This example first determines the image type by the main file name and then by the extension, achieving multi-dimensional determination of the image type and increasing the accuracy of determining the image type.

[0038] Furthermore, in practice, information related to the image type is more clearly reflected in the main file name, while a wide variety of extensions exist to reflect the image type. For example, the main file name of an infrared image may contain the word "infrared," while the extensions may include ".IR.jpg," ".ir.jpg," ".irp.jpg," and so on. Therefore, this method of first determining the image type by the main file name and then by the extension can minimize unnecessary subsequent determination steps and improve overall work efficiency.

[0039] It should be understood that if the file name of the image to be detected does not contain an extension, or if the file name of the image to be detected contains an extension but the database does not contain an association that matches the extension of the image to be detected, a subsequent step of determining the type of the image based on the pixel features of the image is required.

[0040] It should be noted that, since some characters or suffixes in the file name may be different but the manufacturer is the same, the association relationship in the above example may have multiple characters or codes associated with the same manufacturer.

[0041] In practical applications, different manufacturers may associate the same character with different types of images. Because the image type associated with a character is related to its semantics, in this case, different manufacturers may define the same image type for the same character. Therefore, if the database contains multiple associations matching the same character, the type of the image to be detected can be determined based on any one of the associations corresponding to that character.

[0042] It's also possible for different manufacturers to associate the same code. Because the image types corresponding to manufacturer-defined codes are random, different manufacturers may define different image types for the same code. Therefore, if there are multiple associations matching the same code in the database, determine whether the image types corresponding to the codes in these multiple associations are the same. If they are, the image types in these multiple associations are determined as the type of the image to be detected. If they are different, proceed to the subsequent step of determining the image type based on the image's pixel features.

[0043] S140: Otherwise, determine the type of the image to be detected according to the pixel features of the image to be detected.

[0044] The database also includes a large number of images collected, including images of power equipment in operation, of either infrared or visible light types. In this case, the present application analyzes the grayscale histograms of a large number of infrared and visible light images in the database and finds that the grayscale histograms of different types of images have differences in grayscale distribution. Based on these differences, the image type can be further determined. Specifically, the process includes the following steps: According to the grayscale histograms of each infrared image and the grayscale histograms of each visible light image in the database, it is determined whether the skewness of the image grayscale histogram belongs to the first range and the kurtosis belongs to the second range.

[0045] After analyzing the grayscale histograms of a large number of two types of images in a database, the inventors discovered that infrared images typically exhibit a higher concentration of grayscale values in high-temperature areas, and their grayscale histograms may have obvious peaks or distribution biases at higher grayscale levels. In contrast, the grayscale distribution of visible light images is relatively uniform, with their grayscale histograms being relatively flat or having a relatively uniform distribution over a wider grayscale range. Therefore, illustratively, the first range is [0.5, +∞), and the second range is (3, +∞). Accordingly, images with a skewness within the first range or a kurtosis within the second range can be determined to be infrared images; images with a skewness outside the first range and a kurtosis outside the second range are determined to be visible light images.

[0046] Therefore, for the image to be detected whose type cannot be determined based on the file name, in a feasible design, the following method is used to further analyze the pixel features of the image to be detected to determine the type of the image to be detected based on the pixel features of the image to be detected: Obtain the target grayscale histogram of the image to be detected; The type of the image to be detected is determined based on the skewness and / or kurtosis of the target grayscale histogram.

[0047] Taking into account the differences in grayscale distribution between infrared images and visible light images, when the type of the image cannot be directly determined based on the file name, the above example can further accurately determine the type of the image to be detected based on the skewness and / or kurtosis of the target grayscale histogram.

[0048] In a feasible design, the type of the image to be detected is determined based on the skewness and / or kurtosis of the target grayscale histogram in the following manner: If the skewness of the target grayscale histogram is within a predetermined first range, or the kurtosis of the target grayscale histogram is within a predetermined second range, determining that the type of the image to be detected is an infrared image; Otherwise, it is determined that the type of the image to be detected is a visible light image.

[0049] In the above example, if the skewness of the target grayscale histogram is not within the first range and the kurtosis is not within the second range, it can be directly determined that the type of the image to be detected is a visible light image.

[0050] Through research, the present applicant has discovered that, because infrared light filters green the most, the green channel value of infrared images is relatively low among the three colors of red, green, and blue. Therefore, if computing resources permit or based on actual needs, in order to further accurately determine the type of the image to be detected, if the skewness of the target grayscale histogram is outside the first range and the kurtosis of the target grayscale histogram is outside the second range, the type of the image to be detected can also be accurately determined based on the green channel value of the pixels of the image to be detected.

[0051] This application analyzes the channel value characteristics of pixels in a large number of infrared images and visible light images in the database to determine the average value of the green channel value of the visible light image, so as to subsequently determine the type of the image to be detected based on the average value. The specific steps include: By traversing pixels of the visible light images in the database, a first average value is determined, where the first average value is an average value of green channel values of each visible light image in the database.

[0052] For example, when determining the first average value, the average value of the green channel value of each pixel of each visible light image in the database may be determined first, and then the average values are averaged to obtain the first average value.

[0053] Therefore, the type of the image to be detected can be determined based on the green channel value of the pixel of the image to be detected in the following way: Determine a second average value by traversing pixels of the image to be detected, where the second average value is an average value of each green channel value of the image to be detected; comparing the second mean with the first mean; If the second average value is less than the first average value, determining that the type of the image to be detected is an infrared image; Otherwise, the type of the image to be detected is determined to be a visible light image; or, the type of the image to be detected is determined according to the file size of the image to be detected.

[0054] In the above example, if the second average value is greater than the first average value, it can be directly determined that the type of image to be tested is a visible light image. This application analyzes a large amount of collected image data and finds that during the live detection process, images are captured as a pair of infrared images and visible light images, and the file size of infrared images generally shows a larger size than that of visible light images.

[0055] Therefore, if computing resources allow or according to actual needs, the type of the image to be detected can be accurately determined based on the file size of the image to be detected. In a feasible design, the type of the image to be detected can be determined based on the file size of the image to be detected in the following way: Acquire another image to be detected corresponding to the image to be detected, wherein the other image to be detected and the image to be detected form a pair of visible light images and infrared images for image registration processing; Comparing the file size of the image to be detected with the file size of another image to be detected; If the file size of the image to be detected is larger than the file size of another image to be detected, determining that the type of the image to be detected is an infrared image; Otherwise, it is determined that the type of the image to be detected is a visible light image.

[0056] The file size is used to indicate the number of bytes contained in the image data.

[0057] Based on the rule that the images captured during the live detection process are a pair of infrared images and visible light images, the type of the other image to be detected in a pair of visible light images and infrared images can be quickly and efficiently determined by the following methods: When the image to be detected is one of a pair of an infrared image and a visible light image, if the type of the image to be detected is an infrared image, determining the type of the other image to be detected is a visible light image; If the type of the image to be detected is a visible light image, it is determined that the type of the other image to be detected is an infrared image.

[0058] For example, to adapt to new image type classification methods from emerging manufacturers or changes in existing manufacturers' classification methods, new image type classification methods can be updated through the cloud or manually imported, and new associations can be established based on the new classification methods. This continuously expands and improves the database, ensuring that the solution can always accurately identify image types from different manufacturers and further improving its ability to automatically adapt to the complex classification methods of different manufacturers.

[0059] Furthermore, whenever a certain number of newly added images is added to the database, the first average value of the green channel value of the visible light image and the skewness of the grayscale histogram of the infrared image belonging to the first range and the kurtosis belonging to the second range are re-determined to further improve the accuracy of determining the image type.

[0060] Based on the above embodiments, Figure 2 As shown, this application also provides a method for distinguishing infrared images from visible light images, as shown in the following steps: Step a, determining whether the main file name of the image to be detected contains the word "infrared", if yes, proceed to step b, otherwise, proceed to step c; Step b, determining that the type of the image to be detected is an infrared image and outputting the result; Step c, determining whether the extension of the image to be detected matches the extension of the infrared image stored in the database, if so, executing step b, otherwise executing step d; Step d, determining the grayscale histogram of the image to be detected. If the skewness of the grayscale histogram is greater than or equal to 0.5, or the kurtosis is greater than 3, proceed to step b; otherwise, proceed to step e; Step e, traversing the pixels of the image to be detected, calculating a second average value of the green channel value of the image to be detected, and determining whether the second average value is lower than the first average value of the green channel value of the visible light image in the database. If so, executing step b; otherwise, executing step f; Step f, obtaining another image to be detected corresponding to the image to be detected, and determining whether the file size of the image to be detected is larger than the file size of the other image to be detected. If so, proceed to step b; otherwise, proceed to step g; Step g: determine whether the type of the image to be detected is visible light and output the result.

[0061] After completing various analyses, the above example combines the results of primary file name matching, extension matching, and pixel feature analysis (grayscale histogram, color information, and image size) to make a final comprehensive judgment. By weighing the reliability and consistency of each judgment basis, the accurate type of each image is determined. Finally, the judgment results are output, providing accurate image classification information for subsequent processing steps.

[0062] The database of the embodiment of the present application includes a large number of infrared images and visible light images taken by inspection equipment from existing manufacturers. These images come from a wide range of sources and cover a variety of practical application scenarios and equipment models. Therefore, the database of the present application is highly representative and comprehensive. On this basis, based on the file name of the image to be detected, it is basically possible to find the corresponding association in the database, thereby accurately and efficiently determining whether the image to be detected is an infrared image or a visible light image. In rare cases, if the image type cannot be determined based on the image file name and the database, the present application also provides a solution for determining the type of the image to be detected based on the pixel features of the image to be detected. Therefore, no matter how complex the method used by the manufacturer of the inspection equipment to distinguish image types is, or even whether it does not distinguish at all, the solution of the present application can automatically adapt and accurately identify the image type. It can be seen that compared with traditional manual judgment methods, this solution has excellent adaptive capabilities and can achieve accurate distinction of image types without manual intervention. This feature greatly improves data processing efficiency, significantly saves manpower and time costs, enables detection work to be carried out more efficiently and smoothly, and provides strong support for the large-scale development of detection work.

[0063] The solution of this application has excellent integrability and can be easily integrated into various types of patrol equipment, live detection mobile terminals, and background systems. It is compatible with different operating systems and hardware platforms to ensure that the solution can run stably on various common patrol equipment, terminals, or systems, and seamlessly cooperate with other live detection functional modules without affecting the overall performance of the patrol equipment, terminals, or systems and the normal progress of detection work. This enables on-site inspection personnel to conveniently use the solution to distinguish the types of images to be inspected in actual work, optimizes the entire live detection workflow, enhances the convenience and reliability of work, improves the efficiency and quality of on-site inspection work, and provides great convenience for inspection personnel.

[0064] In an actual live detection scenario, when the solution of the present application is applied to a handheld live detection mobile terminal, after acquiring the infrared image and the visible light image, the solution of the present application is used to quickly and accurately distinguish and process the images, and the type of the image is displayed in the file name using a specific field according to the preset rules. Then, the distinguished image is accurately uploaded to the background system through the live detection function. The background system can accurately determine the type of image based on the preset rules, and efficiently process and analyze the image, so as to achieve accurate flow and effective utilization of image data in the entire live detection workflow. The live detection mobile terminal using this solution can help the members of the power grid operation and maintenance team to realize the real-time processing and uploading of maintenance and experimental data in a digital and wireless manner during routine inspections, which not only reduces the workload of the staff, but also greatly improves the timeliness and accuracy of the data.

[0065] In summary, this application solution effectively reduces the potential for misclassification during manual judgment, efficiently and accurately identifies image types, and provides a more reliable data basis for equipment status assessment. This helps the backend system more accurately judge the equipment's operating status based on the accurate image type and promptly identify potential problems, thereby effectively ensuring the safe and stable operation of the equipment, improving the quality and reliability of detection work, and reducing the risk of equipment failure.

[0066] like Figure 3 As shown, the present application also provides a device for distinguishing infrared images from visible light images during charge detection, comprising: An image acquisition module is used to acquire images of the power equipment to be tested in an operating state; An image type recognition module is used to determine, based on the file name of the image to be detected, whether the database contains an association relationship that matches the key information of the file name. The association relationship is used to associate the key information with the type of the image, which can be an infrared image or a visible light image. The image type recognition module is further configured to, if the database contains an association relationship that matches the key information, determine the type of the image to be detected based on the association relationship; Otherwise, the image type recognition module is further configured to determine the type of the image to be detected according to pixel features of the image to be detected.

[0067] In a feasible design, the image type recognition module includes an irregular recognition submodule, which is implemented in the following manner to determine the type of the image to be detected based on the pixel features of the image to be detected: Obtain the target grayscale histogram of the image to be detected; The type of the image to be detected is determined based on the skewness and / or kurtosis of the target grayscale histogram.

[0068] In a feasible design, the irregularity recognition submodule is implemented in the following manner to determine the type of the image to be detected based on the skewness and / or kurtosis of the target grayscale histogram: If the skewness of the target grayscale histogram is within a predetermined first range, or the kurtosis of the target grayscale histogram is within a predetermined second range, determining that the type of the image to be detected is an infrared image; Otherwise, it is determined that the type of the image to be detected is a visible light image.

[0069] In a feasible design, if the skewness of the target grayscale histogram is outside the first range and the kurtosis of the target grayscale histogram is outside the second range, the irregular recognition submodule is further used to determine the type of the image to be detected based on the green channel value of the pixel of the image to be detected.

[0070] In a feasible design, the image data statistical analysis module is further configured to determine a first average value by traversing pixels of visible light images in the database, where the first average value is an average value of green channel values of each visible light image in the database.

[0071] In a feasible design, the irregularity recognition submodule is implemented in the following manner to determine the type of the image to be detected based on the green channel value of the pixel of the image to be detected: Determine a second average value by traversing pixels of the image to be detected, where the second average value is an average value of each green channel value of the image to be detected; comparing the second mean with the first mean; If the second average value is less than the first average value, determining that the type of the image to be detected is an infrared image; Otherwise, the type of the image to be detected is determined to be a visible light image, or the type of the image to be detected is determined according to the file size of the image to be detected.

[0072] In a feasible design, the irregular recognition submodule is implemented in the following manner to determine the type of the image to be detected based on the file size of the image to be detected: Acquire another image to be detected corresponding to the image to be detected, wherein the other image to be detected and the image to be detected form a pair of visible light images and infrared images for image registration processing; Comparing the file size of the image to be detected with the file size of another image to be detected; If the file size of the image to be detected is larger than the file size of another image to be detected, determining that the type of the image to be detected is an infrared image; Otherwise, it is determined that the type of the image to be detected is a visible light image.

[0073] In a feasible design, when the image to be detected is one of a pair of infrared image and visible light image, the irregular recognition submodule is further configured to: If the type of the image to be detected is an infrared image, determining that the type of the other image to be detected is a visible light image; If the type of the image to be detected is a visible light image, it is determined that the type of the other image to be detected is an infrared image.

[0074] In a feasible design, the image type identification module further includes a file name identification submodule and an extension identification submodule. When the key information is a character or code in the main file name that indicates the type of the image, or when the key information is an extension: The file name recognition submodule is used to match the main file name of the image to be detected with the characters or codes used by each manufacturer according to the type of image in the database; if the main file name of the image to be detected contains characters or codes in the database, based on the characters or codes contained in the main file name, determine the association relationship in the database that matches the characters or codes contained in the main file name, and determine the type of the image to be detected based on the association relationship; The extension recognition submodule is used to determine whether the database contains an association relationship that matches the extension of the image to be detected if the main file name of the image to be detected does not contain characters or codes in the database and the file name of the image to be detected contains an extension; if the database contains an association relationship that matches the extension of the image to be detected, determine the type of the image to be detected based on the association relationship.

[0075] For other implementations and effects of the above-mentioned device, please refer to the description in the embodiment of the method for distinguishing infrared images from visible light images, which will not be repeated here.

[0076] The basic principles of the present application have been described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, and effects mentioned in this application are merely illustrative and not restrictive, and it should not be assumed that these advantages, strengths, and effects are required of each embodiment of this application. In addition, the specific details disclosed above are merely illustrative and facilitating understanding, and are not restrictive. The above details do not limit this application to necessarily being implemented using the above specific details.

[0077] It should be understood that although the steps in the flowcharts of the accompanying drawings are shown in sequence as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some of the steps in the flowcharts of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0078] The block diagrams of the devices, devices, equipment, and systems involved in this application are intended only as illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As will be appreciated by those skilled in the art, these devices, devices, equipment, and systems can be connected, arranged, or configured in any manner. Words such as "include," "comprise," "have," and the like are open-ended words, meaning "including but not limited to," and can be used interchangeably therewith. The words "or" and "and" used herein refer to the words "and / or" and can be used interchangeably therewith, unless the context clearly indicates otherwise. The word "such as" used herein refers to the phrase "such as but not limited to," and can be used interchangeably therewith.

[0079] It should also be noted that in the apparatus, device, and method of the present application, each component or each step can be decomposed and / or recombined, and such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0080] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0081] The above description has been provided for the purpose of illustration and description. Furthermore, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although a number of example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A method for distinguishing infrared images from visible light images in charge detection, characterized in that: include: Acquire images of power equipment to be inspected in an operating state; Determine, based on the file name of the image to be detected, whether a database contains an association relationship that matches key information of the file name, wherein the association relationship is used to associate the key information with the type of the image, the type being an infrared image or a visible light image; If the database contains an association relationship that matches the key information, determining the type of the image to be detected according to the association relationship; Otherwise, the type of the image to be detected is determined according to the pixel features of the image to be detected.

2. The method according to claim 1, characterized in that The determining the type of the image to be detected according to pixel features of the image to be detected includes: Obtaining a target grayscale histogram of the image to be detected; The type of the image to be detected is determined according to the skewness and / or kurtosis of the target grayscale histogram.

3. The method according to claim 2, characterized in that Determining the type of the image to be detected according to the skewness and / or kurtosis of the target grayscale histogram includes: If the skewness of the target grayscale histogram is within a predetermined first range, or the kurtosis of the target grayscale histogram is within a predetermined second range, determining that the type of the image to be detected is an infrared image; Otherwise, it is determined that the type of the image to be detected is a visible light image.

4. The method according to claim 2, characterized in that Determining the type of the image to be detected according to the skewness and / or kurtosis of the target grayscale histogram includes: If the skewness of the target grayscale histogram is outside a predetermined first range and the kurtosis of the target grayscale histogram is outside a predetermined second range, the type of the image to be detected is determined according to the green channel value of the pixel of the image to be detected.

5. The method according to claim 4, characterized in that The method further comprises: By traversing pixels of the visible light images in the database, a first average value is determined, where the first average value is an average value of green channel values of each visible light image in the database.

6. The method according to claim 5, characterized in that The determining the type of the image to be detected according to the green channel value of the pixel of the image to be detected includes: Determine a second average value by traversing pixels of the image to be detected, where the second average value is an average value of each green channel value of the image to be detected; comparing the second average value with the first average value; If the second average value is less than the first average value, determining that the type of the image to be detected is an infrared image; Otherwise, it is determined that the type of the image to be detected is a visible light image, or the type of the image to be detected is determined according to the file size of the image to be detected.

7. The method according to claim 6, characterized in that The determining the type of the image to be detected according to the file size of the image to be detected includes: Acquire another image to be detected corresponding to the image to be detected, wherein the another image to be detected and the image to be detected form a pair of visible light images and infrared images for image registration processing; Comparing the file size of the image to be detected with the file size of another image to be detected; If the file size of the image to be detected is larger than the file size of the other image to be detected, determining that the type of the image to be detected is an infrared image; Otherwise, it is determined that the type of the image to be detected is a visible light image.

8. The method according to claim 1, characterized in that The method further includes: when the image to be detected is one of a pair of an infrared image and a visible light image, if the type of the image to be detected is an infrared image, determining that the type of the other image to be detected is a visible light image; If the type of the image to be detected is a visible light image, it is determined that the type of another image to be detected is an infrared image.

9. The method according to any one of claims 1 to 8, characterized in that The key information is a character or code in the main file name that indicates the type of the image, or the key information is an extension. According to the file name of the image to be detected, determining whether a database contains an association relationship that matches the key information of the file name; if the database contains an association relationship that matches the key information, determining the type of the image to be detected according to the association relationship includes: Matching the main file name of the image to be detected with the characters or codes used by each manufacturer according to the type of image in the database; If the main file name of the image to be detected contains characters or codes in the database, determining, based on the characters or codes contained in the main file name, an association relationship in the database that matches the characters or codes contained in the main file name, and determining the type of the image to be detected based on the association relationship; If the main file name of the image to be detected does not contain the characters or codes in the database, and the file name of the image to be detected contains an extension, determining whether the database contains an association relationship that matches the extension of the image to be detected; If the database contains an association relationship that matches the extension of the image to be detected, the type of the image to be detected is determined according to the association relationship.

10. A device for distinguishing infrared images from visible light images in charge detection, characterized in that: include: An image acquisition module is used to acquire images of the power equipment to be tested in an operating state; An image type recognition module is used to determine, based on the file name of the image to be detected, whether a database contains an association relationship that matches the key information of the file name, wherein the association relationship is used to associate the key information with the type of the image, and the type is an infrared image or a visible light image; The image type recognition module is further configured to, if the database contains an association relationship matching the key information, determine the type of the image to be detected according to the association relationship; Otherwise, the image type recognition module is further configured to determine the type of the image to be detected according to pixel features of the image to be detected.

Images