Transformer equipment image reflection removal method and device, electronic equipment and storage medium

Abstract

The application provides a power transformation equipment image reflection removal method, which comprises the following steps: performing image recognition on power transformation equipment according to pre-acquired power transformation equipment images, and obtaining material category information and reflection type information of the power transformation equipment; controlling a patrol robot to collect multi-modal image data of the power transformation equipment, and aligning the multi-modal image data to obtain aligned image data; performing reflection area detection on the equipment images according to geometric structure information of the power transformation equipment and the aligned image data, and obtaining a reflection area detection result; and removing the reflection from the power transformation equipment images according to the reflection type information and the reflection area detection result, and obtaining the power transformation equipment images after the reflection is removed. Thus, the accurate removal of the reflection from the power transformation equipment images is realized.

Description

Technical Field

[0001] This application relates to the field of image recognition technology, and in particular to a method, apparatus, electronic device and storage medium for removing reflections from images of power equipment. Background Technology

[0002] As a crucial component of the power system, the safe operation of substations is vital to the stability of the power grid. With the advancement of smart grid construction, substation inspection robots have been widely applied in equipment condition monitoring and defect identification. However, the substation environment is complex and variable, characterized by: dense high-voltage equipment; numerous metal surfaces on equipment such as transformers, circuit breakers, and disconnectors, which are prone to strong glare; complex lighting conditions, combining indoor and outdoor environments, varying shadows, and the coexistence of areas with strong direct sunlight and weak light; diverse equipment materials, including metal casings, insulating porcelain insulators, and silicone rubber composite insulators; and complex structural geometry, with varied curved surfaces and high angle sensitivity.

[0003] During actual inspections, the inspection robot's camera frequently encounters reflective interference from materials such as metal and glass, resulting in a significant deterioration in image quality and directly impacting the accuracy of equipment condition monitoring and defect identification. Traditional single-algorithm processing methods are poorly adapted to the unique reflective patterns of power equipment, making it difficult to meet the application requirements of complex substation environments. Simultaneously, real-time performance and high-quality imaging coexist, necessitating a balance between processing speed and image quality. Furthermore, the changing perspective caused by the inspection robot's movement further increases the processing difficulty, with dirt, water stains, and reflective surfaces causing confusion and leading to a high false detection rate. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide a method, apparatus, electronic device and storage medium for removing reflections from images of power equipment.

[0005] To achieve the above objectives, this application provides a method for removing reflections from images of power equipment, comprising: performing image recognition on the power equipment based on pre-acquired images of the power equipment to obtain material category information and reflection type information of the power equipment; controlling an inspection robot to collect multimodal image data of the power equipment and aligning the multimodal image data to obtain aligned image data; performing reflection area detection on the equipment image based on the geometric structure information of the power equipment and the aligned image data to obtain reflection area detection results; and removing reflections from the power equipment image based on the reflection type information and the reflection area detection results to obtain a reflection-removed image of the power equipment.

[0006] As can be seen from the above, the image reflection removal method for power equipment provided in this application performs image recognition on the power equipment to determine the material type and reflection type of the power equipment. This allows for targeted reflection removal based on different materials and reflection types of the power equipment. Furthermore, by controlling an inspection robot to collect multimodal image data and aligning the multimodal image data, a foundation is provided for subsequent reflection recognition, improving the accuracy of reflection recognition. Combining the geometric structure of the power equipment with the aligned image data, reflection areas are detected in the equipment image, thereby improving the accuracy of reflection area detection and avoiding interference. Finally, targeted reflection removal is performed based on different reflection types in different reflection areas, thus improving the reflection removal effect. Attached Figure Description

[0007] To more clearly illustrate the technical solutions in this application or related technologies, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0008] Figure 1 A schematic flowchart illustrating a method for removing reflections from images of power equipment provided in an embodiment of this application;

[0009] Figure 2 This is a schematic diagram of the structure of a device for removing image reflections from power equipment, provided in an embodiment of this application. Figure 3 This is a schematic diagram of a more specific electronic device hardware structure provided in an embodiment of this application. Detailed Implementation

[0010] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with specific embodiments and the accompanying drawings.

[0011] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this application should have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms "first," "second," and similar terms used in the embodiments of this application do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are only used to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0012] As a crucial component of the power system, the safe operation of substations is vital to the stability of the power grid. With the advancement of smart grid construction, substation inspection robots have been widely applied in equipment condition monitoring and defect identification. However, the substation environment is complex and variable, characterized by: dense high-voltage equipment; numerous metal surfaces on equipment such as transformers, circuit breakers, and disconnectors, which are prone to strong glare; complex lighting conditions, combining indoor and outdoor environments, varying shadows, and the coexistence of areas with strong direct sunlight and weak light; diverse equipment materials, including metal casings, insulating porcelain insulators, and silicone rubber composite insulators; and complex structural geometry, with varied curved surfaces and high angle sensitivity.

[0013] During actual inspections, the inspection robot's camera frequently encounters reflective interference from materials such as metal and glass, resulting in a significant deterioration in image quality and directly impacting the accuracy of equipment condition monitoring and defect identification. Traditional single-algorithm processing methods are poorly adapted to the unique reflective patterns of power equipment, making it difficult to meet the application requirements of complex substation environments. Simultaneously, real-time performance and high-quality imaging coexist, necessitating a balance between processing speed and image quality. Furthermore, the changing perspective caused by the inspection robot's movement further increases the processing difficulty, with dirt, water stains, and reflective surfaces causing confusion and leading to a high false detection rate.

[0014] Currently, there is a lack of reflection removal systems tailored to the material characteristics of substation equipment and inspection scenarios. Existing technologies cannot effectively address issues such as the unified identification and classification of various reflection types in power equipment, robust handling under complex lighting conditions in substations, meeting the dual requirements of real-time inspection and accurate analysis, and removing reflection interference in equipment defect identification. Therefore, researching an intelligent reflection removal method for substation inspection robots is of great significance.

[0015] The image reflection elimination method for substation equipment provided in this application is applied to a substation inspection robot. The robot is equipped with a variety of sensors: a multi-angle pan-tilt camera that can rotate 360 ​​degrees to capture images of the equipment from different perspectives; an infrared thermal imager to collect temperature distribution information of the equipment; a polarization imaging device to acquire images under different polarization states; and a positioning system, a light sensor, a temperature sensor, a meteorological data acquisition device, etc., to acquire environmental and location information.

[0016] The robot's edge computing unit processes sensor data in real time, executes reflection removal algorithms, and uploads the processing results to the substation edge cloud and the cloud via 5G / 4G networks. The substation edge cloud is responsible for local data analysis and task scheduling, while the cloud is responsible for model training and knowledge management.

[0017] like Figure 1 As shown, this application provides a method for removing reflections from images of power equipment, including: Step S1: Perform image recognition on the power equipment based on the pre-acquired images of the power equipment to obtain the material category information and reflectivity information of the power equipment.

[0018] In this embodiment, by performing image recognition on the power equipment, the material type and reflection type of the power equipment can be determined, thereby enabling targeted reflection elimination based on the different materials and reflection types of the power equipment.

[0019] Step S2: Control the inspection robot to collect multimodal image data from the substation equipment, and align the multimodal image data to obtain aligned image data.

[0020] In this embodiment, by controlling the inspection robot to collect multimodal image data and aligning the multimodal image data, a foundation is provided for subsequent reflective recognition, thereby improving the accuracy of reflective recognition.

[0021] Step S3: Based on the geometric structure information of the power equipment and the aligned image data, perform reflective area detection on the equipment image to obtain the reflective area detection result.

[0022] In this embodiment, the accuracy of reflective area detection is improved and interference is avoided by combining the geometric structure of the power equipment with the aligned image data to detect reflective areas in the equipment image.

[0023] Step S4: Based on the reflection type information and the reflection area detection results, the reflection of the substation equipment image is removed to obtain the substation equipment image after reflection removal.

[0024] In this embodiment, the effect of reflection elimination is improved by targeting the different reflection types of different reflective areas.

[0025] In some embodiments, step S1 specifically includes: Step S11: Extract features from the substation equipment image according to the preset image recognition model to obtain material category information.

[0026] In this embodiment, after obtaining images of substation equipment, a material recognition engine can be built based on deep learning to identify the materials of typical substation equipment such as transformer metal casings, insulating porcelain insulators, silicone rubber insulators, and surge arresters.

[0027] Step S12: Compare the preset material optical properties with the material category information to obtain the reflection type information.

[0028] In this embodiment, after identifying the transformer structure and different materials, the type of reflection can be identified by recognizing the reflection type corresponding to different materials. Specifically, reflection can be divided into specular reflection (metal surface), diffuse reflection (rough surface), translucent reflection (glass, insulator), and mixed reflection.

[0029] As an optional implementation, material recognition engines and reflection type recognition can further improve the implementation of the large-scale adventure. Specifically, the large model can include: Multimodal coding: This method uses a visual encoder (such as Vision Transformer) and a text encoder (such as BERT) to process image and text information respectively. The visual encoder encodes the image into visual feature vectors. The text encoder encodes the inspection task description into text feature vectors. .

[0030] Cross-modal alignment: Employing contrastive learning or cross-modal attention mechanisms, this achieves the alignment and fusion of visual and textual features. The cross-modal attention calculation method is as follows: ; in, , , , For dimensions.

[0031] Semantic association: Associating reflective areas with the semantic information of the equipment to understand the impact of reflection on equipment status interpretation. For example, identifying whether a reflective area is located in a critical part of the equipment (such as insulator strings, contacts, etc.).

[0032] Reasoning and Decision-Making: Based on the reasoning capabilities of the large model, formulate strategies and priorities for reflective processing. The reasoning process considers factors such as equipment importance, defect risk, and the urgency of inspection tasks.

[0033] Result Interpretation: Generate natural language explanations to illustrate the decision-making process and results of reflection processing. The explanations include: the type of reflection detected, the algorithm used, the processing effect, and potential risks.

[0034] Furthermore, according to the system architecture described in claim 1, the present invention also provides a substation inspection robot camera reflection removal system, the system comprising: Sensor layer: Includes multi-angle gimbal cameras, infrared thermal imagers, polarization imaging equipment, positioning systems, light sensors, temperature sensors, meteorological data acquisition equipment, etc., responsible for multimodal data acquisition.

[0035] Data processing layer: includes feature extraction module, data fusion module, quality assessment module, knowledge management module, uncertainty quantification module, time series analysis module, geometric modeling module, material analysis module, vector storage, knowledge graph, knowledge base management, vector retrieval, etc., responsible for data preprocessing and feature extraction.

[0036] The core algorithm layer includes deep learning algorithms, traditional algorithms, reinforcement learning algorithms, graph neural network algorithms, meta-learning algorithms, Bayesian learning algorithms, image processing algorithms, optical modeling algorithms, multimodal large models, retrieval enhancement generation (RAG), etc., forming an algorithm library.

[0037] Algorithm coordination layer: includes algorithm selection module, parameter optimization module, parallel scheduling module, result fusion module, multimodal fusion module, knowledge reasoning module, adaptive adjustment module, quality assessment module, etc., responsible for the coordinated scheduling and optimization of multiple algorithms.

[0038] Business logic layer: includes modules such as reflection removal, equipment identification, defect detection, meter identification, status analysis, risk assessment, inspection planning, and path optimization, which implement specific business functions.

[0039] Application layer: including reflective processing applications, equipment inspection applications, defect identification applications, meter reading applications, temperature monitoring applications, safety assessment applications, alarm decision-making applications, operation and maintenance reporting applications, etc., providing end services to users.

[0040] The system adopts an edge-cloud collaborative architecture: the robot edge layer is responsible for tasks with high real-time requirements such as real-time reflection processing, algorithm scheduling and execution, immediate defect warning, and inspection path planning; the substation edge cloud is responsible for medium-scale data processing such as inspection task scheduling, equipment status management, local data analysis, and safety warning processing; and the cloud is responsible for large-scale computing and storage tasks such as multi-substation collaborative learning, model training and optimization, big data analysis, equipment knowledge graph management, inspection strategy optimization, large model training and inference, and knowledge management and retrieval.

[0041] In some embodiments, step S11 specifically includes: Step S111: Collect ambient light information around the power equipment.

[0042] Step S112: Determine the direct light component, the diffuse light component, and the shadow area based on the ambient lighting information.

[0043] In this embodiment, by analyzing the ambient light, the specific locations of the direct light component, the scattered light component, and the shadow area are determined, providing an accurate basis for subsequent identification.

[0044] Step S113: Based on the ambient lighting information and geometric structure information, perform physical modeling of the lighting of the substation to obtain an image recognition model.

[0045] Step S114: Extract features from the power equipment based on the image recognition model.

[0046] In this embodiment, physical modeling is performed by combining the geometric structure and optical properties of the power equipment to construct a mapping relationship between equipment, material, and reflective type.

[0047] In some embodiments, step S2 specifically includes: Step S21: The inspection robot performs visible light image acquisition, infrared image acquisition and polarized light image acquisition on the substation equipment to obtain visible light image information, infrared image information and polarized light image information.

[0048] In this embodiment, the inspection robot is equipped with a multi-angle gimbal camera, an infrared thermal imager, and a polarization imaging device to collect image data of the same device at different angles and under different spectra.

[0049] As an optional implementation, to improve the accuracy of aligned image recognition, the acquired images can also be preprocessed, including image denoising, geometric correction, brightness normalization, and other operations, thereby improving the clarity of the aligned images.

[0050] Step S22: Align the visible light image information, infrared image information and polarized light image information to obtain aligned image data.

[0051] In this embodiment, a multimodal image registration method is constructed to spatially align images from different angles and with different spectra, providing a foundation for subsequent multimodal fusion.

[0052] In some embodiments, the reflection type information includes specular reflection, diffuse reflection, translucent reflection, and mixed reflection.

[0053] Step S4 specifically includes: Step S41: In response to the current reflection type being specular reflection, remove the reflection based on the polarized light image information and geometric structure information.

[0054] In this embodiment, for specular reflection, a polarization difference method based on physical modeling can be used. This method utilizes the reflection differences under different polarization states to eliminate specular reflection through multi-angle polarization image fusion. Combined with the device's geometric information, the non-reflective surface texture is reconstructed.

[0055] Step S42: In response to the current reflection type being diffuse reflection, the reflection is removed based on the continuous image time sequence information corresponding to the aligned image data and the preset diffuse reflection removal model.

[0056] In this embodiment, for diffuse reflection, a deep learning-based method can be used to train a reflection removal neural network and generate a reflection-free image using a generative adversarial network (GAN). Combined with temporal information, the true surface information is reconstructed from multiple frames of images.

[0057] Step S43: In response to the current reflection type being semi-transparent reflection, the surface reflection and transmission components are separated based on visible light image information and infrared image information, and the reflection is removed based on polarized light image information.

[0058] In this embodiment, spectral analysis can be used to separate the surface reflective and transmissive components by utilizing the spectral differences between visible light and infrared thermal imaging. Combining this with polarization information improves the removal effect of translucent reflections.

[0059] Step S44: In response to the current reflection type being mixed reflection, remove the reflection based on visible light image information, infrared image information, polarized light image information, and continuous image timing information corresponding to the aligned image data.

[0060] In this embodiment, for mixed reflections, a multimodal fusion method can be used to comprehensively utilize multi-dimensional information such as polarization, spectrum, and time sequence, and to realize the transmission and joint inference of reflection information between devices through graph neural networks.

[0061] As an optional implementation method, specular reflection removal employs a polarization difference-based approach, specifically including: Multi-angle polarization image acquisition: Images are acquired at different polarization angles (0°, 45°, 90°, 135°). Assume the acquired image is... .

[0062] Polarization parameter calculation: Calculate the Stokes parameters, which represent the polarization state of light. ; ; ; Calculation of degree of polarization (DoP): ; Polarization angle calculation: Calculate the angle of polarization (AoP): ; Reflective region separation: Specular reflections typically have a high degree of polarization. A polarization degree threshold is used to separate the specular reflection region. ; in, For reflective mask, This is the polarization degree threshold.

[0063] Reflection Removal: Reflection removal based on polarization difference while preserving diffuse reflection components. ; in, For adjustment coefficients, This indicates the dot product operation.

[0064] In some embodiments, step S3 specifically includes: Step S31: Determine the first reflective area based on the set geometric structure information and the preset illumination direction information.

[0065] Step S32: Based on the aligned image data and the sequential information of the continuous images corresponding to the aligned image data, the first reflective area is verified to obtain the second reflective area.

[0066] Step S33: The credibility of the second reflective area is evaluated according to the preset uncertainty quantification mechanism to obtain the reflective area detection result.

[0067] In this embodiment, reflective areas are detected in the image based on material type and reflectivity type information. Combining device geometry and illumination direction information, a multi-scale feature extraction method is used to identify reflective areas of different sizes and intensities. Utilizing temporal information and combining continuous frame images from the inspection trajectory, the authenticity of reflective areas is determined by motion consistency. An uncertainty quantification mechanism is introduced to assess the reliability of the detection results.

[0068] As can be seen from the above embodiments of this application, the image reflection removal method for power equipment provided by this application performs image recognition on the power equipment to determine the material type and reflection type of the power equipment. This allows for targeted reflection removal based on different materials and reflection types of the power equipment. Furthermore, by controlling an inspection robot to collect multimodal image data and aligning the multimodal image data, a foundation is provided for subsequent reflection recognition, improving the accuracy of reflection recognition. Combining the geometric structure of the power equipment with the aligned image data, reflection areas are detected in the equipment image, thereby improving the accuracy of reflection area detection and avoiding interference. Finally, targeted reflection removal is performed based on different reflection types in different reflection areas, thus improving the reflection removal effect.

[0069] It should be noted that the method in this embodiment can be executed by a single device, such as a computer or server. The method can also be applied in a distributed scenario, where multiple devices cooperate to complete the task. In such a distributed scenario, one of these devices may execute only one or more steps of the method in this embodiment, and the multiple devices will interact with each other to complete the method described.

[0070] It should be noted that the above description describes some embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in a different order than that shown in the above embodiments and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0071] Based on the same inventive concept, corresponding to any of the above embodiments, this application also provides a device for removing reflections from images of power equipment.

[0072] refer to Figure 2 A device for removing reflections from images of power equipment, comprising: The image recognition module 10 is used to perform image recognition on the substation equipment based on the pre-acquired images of the substation equipment, and obtain the material category information and reflectivity information of the substation equipment.

[0073] The image acquisition module 20 is used to control the inspection robot to acquire multimodal image data of the substation and to align the multimodal image data to obtain aligned image data.

[0074] The image alignment module 30 is used to perform reflective area detection on the equipment image based on the geometric structure information of the power equipment and the aligned image data, and obtain the reflective area detection result.

[0075] The reflection removal module 40 is used to remove reflections from the image of the power equipment based on the reflection type information and the reflection area detection results, so as to obtain the image of the power equipment after reflection removal.

[0076] In some embodiments, the image recognition module 10 includes: The material recognition module 101 is used to extract features from the image of the substation equipment according to the preset image recognition model to obtain material category information.

[0077] The comparison module 102 is used to compare the preset material optical properties with the material category information to obtain the reflection type information.

[0078] For ease of description, the above devices are described in terms of function, divided into various modules. Of course, in implementing this application, the functions of each module can be implemented in one or more software and / or hardware.

[0079] The apparatus described above is used to implement the corresponding method for removing reflections from images of power equipment in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0080] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the image reflection removal method for power equipment described in any of the above embodiments.

[0081] Figure 3 This embodiment illustrates a more specific hardware structure of an electronic device. The device may include a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, memory 1020, input / output interface 1030, and communication interface 1040 are interconnected internally via the bus 1050.

[0082] The processor 1010 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0083] The memory 1020 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 1020 and is called and executed by the processor 1010.

[0084] The input / output interface 1030 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components within the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touchscreens, microphones, various sensors, etc., while output devices may include displays, speakers, vibrators, indicator lights, etc.

[0085] The communication interface 1040 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0086] Bus 1050 includes a pathway for transmitting information between various components of the device, such as processor 1010, memory 1020, input / output interface 1030, and communication interface 1040.

[0087] It should be noted that although the above-described device only shows the processor 1010, memory 1020, input / output interface 1030, communication interface 1040, and bus 1050, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0088] The electronic devices described above are used to implement the corresponding image reflection removal method for substation equipment in any of the foregoing embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0089] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this application also provides a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the image reflection removal method for power equipment as described in any of the above embodiments.

[0090] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0091] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to execute the image reflection removal method for power equipment as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0092] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of this application (including the claims) is limited to these examples; within the framework of this application, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of the embodiments of this application as described above, which are not provided in the details for the sake of brevity.

[0093] Additionally, to simplify the description and discussion, and to avoid obscuring the embodiments of this application, the well-known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided drawings. Furthermore, the apparatus may be shown in block diagram form to avoid obscuring the embodiments of this application, and this also takes into account the fact that the details of the implementation of these block diagram apparatuses are highly dependent on the platform on which the embodiments of this application will be implemented (i.e., these details should be fully understood by those skilled in the art). While specific details (e.g., circuits) have been set forth to describe exemplary embodiments of this application, it will be apparent to those skilled in the art that the embodiments of this application can be implemented without these specific details or with variations thereof. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0094] Although this application has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0095] The embodiments of this application are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the embodiments of this application should be included within the protection scope of this application.

Claims

1. A method for removing reflections from images of power equipment, comprising: Based on the pre-acquired images of the power equipment, image recognition is performed on the power equipment to obtain the material category information and reflectivity information of the power equipment; The inspection robot is controlled to collect multimodal image data of the substation equipment, and the multimodal image data is aligned to obtain aligned image data. Based on the geometric structure information of the substation and the aligned image data, reflective areas are detected in the equipment image to obtain reflective area detection results. Based on the reflection type information and the reflection area detection results, the reflection of the power equipment image is removed to obtain a reflection-removed image of the power equipment.

2. The method for removing reflections from images of power equipment according to claim 1, wherein, The step of performing image recognition on the power equipment based on pre-acquired images of the power equipment specifically includes: The material category information is obtained by extracting features from the image of the substation based on a preset image recognition model. The preset optical properties of the material are compared with the material category information to obtain the reflectivity type information.

3. The method for removing reflections from images of power equipment according to claim 2, wherein, The step of extracting features from the substation equipment image according to a preset image recognition model specifically includes: Collect ambient light information around the power equipment; Based on the ambient lighting information, determine the direct light component, the diffuse light component, and the shadow area; Based on the ambient lighting information and the geometric structure information, a physical model of the lighting of the substation is performed to obtain the image recognition model; Based on the image recognition model, feature extraction is performed on the substation.

4. The method for removing reflections from images of power equipment according to claim 1, wherein, The controlled inspection robot collects multimodal image data from the substation equipment and aligns the multimodal image data to obtain aligned image data, specifically including: The inspection robot performs visible light image acquisition, infrared image acquisition, and polarized light image acquisition on the substation to obtain visible light image information, infrared image information, and polarized light image information. The visible light image information, the infrared image information, and the polarized light image information are aligned to obtain the aligned image data.

5. The method for removing reflections from images of power equipment according to claim 4, wherein, The reflection type information includes specular reflection, diffuse reflection, translucent reflection, and mixed reflection; The step of removing reflections from the substation equipment image based on the reflection type information and the reflection area detection result to obtain a reflection-removed substation equipment image specifically includes: In response to the current reflection type being specular reflection, the reflection is removed based on the polarized light image information and the geometric structure information; In response to the current reflection type being diffuse reflection, the reflection is removed based on the continuous image time sequence information corresponding to the aligned image data and a preset diffuse reflection removal model; In response to the current reflection type being semi-transparent reflection, the surface reflection and transmission components are separated based on the visible light image information and the infrared image information, and the reflection is removed based on the polarized light image information; In response to the current reflection type being mixed reflection, the reflection is removed based on the visible light image information, the infrared image information, the polarized light image information, and the continuous image timing information corresponding to the aligned image data.

6. The method for removing reflections from images of power equipment according to claim 1, wherein, The step of detecting reflective areas in the equipment image based on the geometric structure information of the substation and the aligned image data to obtain reflective area detection results specifically includes: The first reflective area is determined based on the aforementioned geometric structure information and the preset illumination direction information; Based on the aligned image data and the sequential image information corresponding to the aligned image data, the first reflective area is verified to obtain the second reflective area. The credibility of the second reflective area is evaluated according to a preset uncertainty quantification mechanism to obtain the detection result of the reflective area.

7. A device for removing image reflections from power equipment, comprising: The image recognition module is used to perform image recognition on the substation equipment based on the pre-acquired images of the substation equipment, and obtain the material category information and reflectivity information of the substation equipment. The image acquisition module is used to control the inspection robot to acquire multimodal image data of the substation and to align the multimodal image data to obtain aligned image data. The image alignment module is used to perform reflective area detection on the equipment image based on the geometric structure information of the substation and the aligned image data, and obtain the reflective area detection result; The reflection removal module is used to remove reflections from the image of the power equipment based on the reflection type information and the reflection area detection result, so as to obtain a reflection-removed image of the power equipment.

8. The apparatus according to claim 7, wherein, The image recognition module includes: The material recognition module is used to extract features from the image of the substation equipment according to a preset image recognition model to obtain the material category information. The comparison module is used to compare the preset optical properties of the material with the material category information to obtain the reflection type information.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method as claimed in any one of claims 1 to 6.

10. A non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the method of any one of claims 1 to 6.

Images