High-voltage equipment imaging method and system based on ultraviolet light multispectral fusion and medium

Abstract

The application discloses a kind of high-voltage equipment imaging method, system and medium based on ultraviolet light multispectral fusion, the method of the application includes obtaining the ultraviolet light radiation image of target high-voltage equipment, infrared radiation image and visible light image;From infrared radiation image, extract infrared abnormal temperature rise area, from ultraviolet light radiation image, extract ultraviolet light spot area, calculate the spatial overlap of infrared abnormal temperature rise area and ultraviolet light spot area;According to spatial overlap, judge whether the fusion image of ultraviolet light radiation image and visible light image is qualified, if spatial overlap is less than preset overlap, then determine that qualified fusion image cannot be obtained, otherwise determine that qualified fusion image can be obtained, and the fusion image of ultraviolet light radiation image and visible light image obtains the fusion image of target high-voltage equipment.The application aims at improving the accuracy of multispectral fusion imaging, improving the inspection diagnosis capability of high-voltage equipment, improving the efficiency of inspection and the accuracy of inspection result.

Description

Technical Field

[0001] This invention relates to the fields of power line inspection technology and image processing technology, specifically to a high-voltage equipment imaging method, system, and medium based on ultraviolet multispectral fusion. Background Technology

[0002] Traditional power equipment inspections primarily rely on manual visual inspection or handheld infrared and ultraviolet (UV) detection instruments, which suffers from inherent drawbacks such as low efficiency, high labor intensity, significant limitations imposed by terrain and weather, and high safety risks associated with close-range operations. In recent years, inspection drones equipped with single visible light or infrared thermal imaging cameras have been initially applied, achieving a mechanized upgrade of the inspection mode. However, existing technologies still have significant limitations: single visible light imaging can only identify physical defects such as equipment deformation, damage, and obvious foreign objects, but cannot detect early, potential insulation discharge faults; while single UV imaging, although sensitive to capturing solar-blind UV signals generated by corona discharge, lacks intuitive information about the equipment's spatial location and environmental background, making it difficult to quickly and accurately locate fault points, and diagnostic results highly dependent on the experience and interpretation of professionals. Chinese patent application CN110672980A discloses an online monitoring method and system for power line inspection based on ultraviolet, infrared, and visible imaging. This method organically combines visible light, infrared, and ultraviolet detection techniques to detect and predict thermal defects, discharge defects, and foreign object intrusion in ultra-high voltage / extra-high voltage substations and transmission lines. However, this prior art does not consider the differential impact of optical lens distortion on multispectral registration strategies, resulting in low image fusion efficiency and consequently affecting inspection and diagnostic capabilities. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a high-voltage equipment imaging method, system and medium based on ultraviolet multispectral fusion, which addresses the above-mentioned problems of the prior art. The present invention aims to improve the accuracy of multispectral fusion imaging, enhance the inspection and diagnosis capabilities of high-voltage equipment, and improve inspection efficiency and the accuracy of inspection results.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: An imaging method for high-voltage equipment based on ultraviolet multispectral fusion includes the following steps: S101, acquire ultraviolet radiation image, infrared radiation image and visible light image of the target high-voltage equipment; S102, extract the infrared abnormal temperature rise region from the infrared radiation image, extract the ultraviolet spot region from the ultraviolet radiation image, and calculate the spatial overlap between the infrared abnormal temperature rise region and the ultraviolet spot region. S103, determine whether the fused image of the ultraviolet radiation image and the visible light image is qualified based on the spatial overlap. If the spatial overlap is less than the preset overlap, it is determined that a qualified fused image cannot be obtained. Otherwise, it is determined that a qualified fused image can be obtained, and the fused image of the ultraviolet radiation image and the visible light image is used to obtain the fused image of the target high-voltage equipment.

[0005] Optionally, in step S102, extracting the infrared abnormal temperature rise region from the infrared radiation image means extracting the infrared abnormal temperature rise region whose temperature exceeds a preset threshold from the infrared radiation image, and extracting the ultraviolet spot region from the ultraviolet radiation image means extracting the ultraviolet spot region whose ultraviolet radiation value is greater than a preset threshold from the ultraviolet radiation image.

[0006] Optionally, the spatial overlap of the infrared abnormal temperature rise region and the ultraviolet spot region calculated in step S102 refers to the overlap ratio of the coverage areas of the infrared abnormal temperature rise region and the ultraviolet spot region. The overlap ratio is the ratio of the intersection area of ​​the ultraviolet spot region and the infrared abnormal temperature rise region to the union area.

[0007] Optionally, step S103, which involves fusing the ultraviolet radiation image and the visible light image to obtain the fused image of the target high-voltage equipment, includes: S201, Calculate the overall distortion intensity index of the visible light image. The functional expression for calculating the overall distortion intensity index of the visible light image is: ; ; ; in, This is the overall distortion intensity index for visible light images. This is the radial distortion offset. This is the tangential distortion offset. , , The radial distortion coefficient is... , The tangential distortion coefficient is... The normalized radial distance of the pre-selected points. The normalized xy coordinates of the pre-selected points; S202, Select the corresponding registration method from the preset candidate registration methods based on whether the comprehensive distortion intensity index of the visible light image is less than the preset distortion intensity index; S203, Register the ultraviolet radiation image and the visible light image to the same size according to the selected registration method; S204, the registered ultraviolet radiation image and the visible light image are weighted and fused to obtain the fused image of the target high-voltage equipment.

[0008] Optionally, in step S202, when selecting a corresponding registration method from preset candidate registration methods based on whether the comprehensive distortion intensity index of the visible light image is less than a preset distortion intensity index, the selected registration method is full-image affine transformation registration based on feature points when the comprehensive distortion intensity index of the visible light image is less than the preset distortion intensity index. S301, extract feature points of SIFT or ORB features from ultraviolet radiation images and visible light images; S302, establish the correspondence between matching feature point pairs in the ultraviolet radiation image and the visible light image through a preset feature descriptor matching algorithm, and obtain a set of matching feature point pairs; S303, a global affine transformation matrix is ​​obtained by fitting the set of matching feature point pairs using a preset robust estimation algorithm. The global affine transformation matrix contains some or all of the linear transformation parameters in translation, rotation, scaling and shearing. S304 uses a global affine transformation matrix to perform a uniform coordinate transformation on all pixel coordinates in the ultraviolet radiation image, mapping them to the pixel coordinate system of the visible light image, thereby obtaining the registered ultraviolet radiation image and the visible light image.

[0009] Optionally, in step S202, when selecting a registration method from preset candidate registration methods based on whether the comprehensive distortion intensity index of the visible light image is less than a preset distortion intensity index, the selected registration method is nonlinear geometric correction and local optimization registration based on pre-calibrated parameters when the comprehensive distortion intensity index of the visible light image is greater than or equal to the preset distortion intensity index. S401, the ultraviolet radiation image and the visible light image are subjected to nonlinear geometric transformation using a pre-calibrated intrinsic parameter matrix and distortion coefficient vector to eliminate the inherent radial and tangential distortion of the lens and generate a preliminarily aligned distortion-free image pair. S402, for the ultraviolet radiation image and visible light image in the initially aligned distortion-free image pair, divide them into several equal-sized sub-regions, and in each sub-region, iteratively solve a local affine or homography transformation parameter with the optimization objective of maximizing the mutual information or structural similarity between the ultraviolet image block and the visible light image block in that region. S403 integrates local affine or homography transformation parameters to obtain global affine or homography transformation parameters; S404 uses global affine or homography transformation parameters to resample all pixel coordinates in the ultraviolet radiation image, obtaining the registered ultraviolet radiation image and visible light image, achieving pixel-level spatial alignment between the two.

[0010] Optionally, after determining in step S103 that a qualified fused image cannot be obtained, the method further includes: S501, Calculate the ultraviolet light discharge characteristic value of the ultraviolet spot region: ; in, These are characteristic values ​​of ultraviolet light discharge. and These are the weighting coefficients. The maximum pixel intensity in the ultraviolet light spot area. The preset pixel intensity threshold, The area fluctuation rate of the ultraviolet light spot region. This is a preset volatility threshold; S502, If the ultraviolet discharge characteristic value is greater than or equal to the preset discharge characteristic value, then proceed to step S503; otherwise, proceed to step S504. S503, calculate the difference between the ultraviolet discharge characteristic value and the preset discharge characteristic value as the ultraviolet discharge difference value. If the ultraviolet discharge difference value is less than the preset ultraviolet discharge difference value, multiply the initial integration time of the ultraviolet imaging sensor by the first integration time adjustment coefficient or use it as the new integration time of the ultraviolet imaging sensor. If the ultraviolet discharge difference value is greater than or equal to the preset ultraviolet discharge difference value, multiply the initial integration time of the ultraviolet imaging sensor by the second integration time adjustment coefficient or use it as the new integration time of the ultraviolet imaging sensor. The first integration time adjustment coefficient and the second integration time adjustment coefficient are both constants greater than 1, and the first integration time adjustment coefficient is less than the second integration time adjustment coefficient. The integration time of the ultraviolet imaging sensor refers to the length of time that the photosensitive element of the ultraviolet imaging sensor accumulates charge on the incident ultraviolet light during one image acquisition process. Jump to step S505. S504, calculate the circularity of the ultraviolet spot region. If the circularity of the ultraviolet spot region is less than a preset circularity, multiply the weight of the initial ultraviolet radiation image by a first weight adjustment coefficient to obtain the weight of the new ultraviolet radiation image. If the circularity of the ultraviolet spot region is greater than or equal to the preset circularity, multiply the weight of the initial ultraviolet radiation image by a second weight adjustment coefficient to obtain the weight of the new ultraviolet radiation image. Both the first and second weight adjustment coefficients are constants less than 1, and the first weight adjustment coefficient is less than the second weight adjustment coefficient. Skip to step S505. S505: The UAV reacquires an ultraviolet radiation image at the current location and calculates the ultraviolet discharge characteristic value of the ultraviolet spot area. If the ultraviolet discharge characteristic value is still greater than or equal to the preset threshold, the UAV is controlled to adjust the observation posture and jump to step S101 to re-inspect the target high-voltage equipment.

[0011] The present invention also provides a high-voltage equipment imaging system based on ultraviolet multispectral fusion, comprising a microprocessor and a memory interconnected thereto, wherein the microprocessor is programmed or configured to execute the high-voltage equipment imaging method based on ultraviolet multispectral fusion.

[0012] The present invention also provides a computer-readable storage medium storing a computer program or instructions that are programmed or configured to execute the high-voltage equipment imaging method based on ultraviolet multispectral fusion by a processor.

[0013] The present invention also provides a computer program product, including a computer program or instructions, which are programmed or configured to execute the high-voltage equipment imaging method based on ultraviolet multispectral fusion via a processor.

[0014] Compared with existing technologies, the present invention mainly achieves the following beneficial effects: The method of the present invention includes acquiring ultraviolet radiation images, infrared radiation images, and visible light images of the target high-voltage equipment; extracting the infrared abnormal temperature rise region from the infrared radiation image and the ultraviolet spot region from the ultraviolet radiation image; calculating the spatial overlap between the infrared abnormal temperature rise region and the ultraviolet spot region; judging whether the fused image of the ultraviolet radiation image and the visible light image is qualified based on the spatial overlap; if the spatial overlap is less than the preset overlap, it is determined that a qualified fused image cannot be obtained; otherwise, it is determined that a qualified fused image can be obtained, and the fused image of the ultraviolet radiation image and the visible light image is obtained as the fused image of the target high-voltage equipment. The present invention can improve the accuracy of multispectral fusion imaging, enhance the inspection and diagnosis capabilities of high-voltage equipment, and improve inspection efficiency and the accuracy of inspection results. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the basic process of the method in an embodiment of the present invention.

[0016] Figure 2 This is a schematic diagram of the system structure in an embodiment of the present invention.

[0017] Figure 3 This is a schematic diagram of the process for generating a fused image in an embodiment of the present invention.

[0018] Figure 4 This is a schematic diagram of the processing flow after a qualified fused image cannot be obtained in an embodiment of the present invention. Detailed Implementation

[0019] To enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions of the present invention will be further described in detail below with reference to the accompanying drawings in the embodiments of the present invention.

[0020] like Figure 1 As shown, the high-voltage equipment imaging method based on ultraviolet multispectral fusion in this embodiment includes the following steps: S101, acquire ultraviolet radiation image, infrared radiation image and visible light image of the target high-voltage equipment; S102, extract the infrared abnormal temperature rise region from the infrared radiation image, extract the ultraviolet spot region from the ultraviolet radiation image, and calculate the spatial overlap between the infrared abnormal temperature rise region and the ultraviolet spot region. S103, if the spatial overlap is less than the preset overlap, it is determined that a qualified fused image cannot be obtained; otherwise, it is determined that a qualified fused image can be obtained, and the fused image of the target high-voltage equipment is obtained by fusing the ultraviolet radiation image and the visible light image.

[0021] like Figure 2 As shown, the hardware system structure in this embodiment includes: a multispectral collaborative detection module, comprising sensors for acquiring ultraviolet radiation images, infrared radiation images, and visible light images, for acquiring ultraviolet radiation images, infrared radiation images, and visible light images of the target high-voltage equipment; an image processing module, connected to the multispectral collaborative detection module, for extracting the infrared abnormal temperature rise region from the infrared radiation image, extracting the ultraviolet spot region from the ultraviolet radiation image, calculating the spatial overlap between the infrared abnormal temperature rise region and the ultraviolet spot region, and determining that a qualified fused image cannot be obtained if the spatial overlap is less than a preset overlap, otherwise determining that a qualified fused image can be obtained, and fusing the ultraviolet radiation image and the visible light image to obtain the fused image of the target high-voltage equipment; and a control module, connected to the multispectral collaborative detection module and the image processing module, for controlling the multispectral collaborative detection module to adjust inspection parameters or strategies.

[0022] The image processing and control modules can be implemented using independent or integrated chips as needed. For example, a low-power system-on-a-chip (SoC) can be used as the core processing unit to implement the functions of the image processing and control modules. This SoC chip directly receives and processes the raw MIPI signals output by the visible light imaging unit and the ultraviolet imaging sensor. After receiving the raw image stream, the SoC chip first calls the pre-stored distortion parameters of each imaging unit and performs real-time geometric distortion correction of the multi-channel images in parallel during the raw data domain or preprocessing stage. Subsequently, using the graphics processing unit (GPU) and neural network processing unit (NPU) integrated within the SoC chip, it performs rapid image registration, feature extraction, fusion, and overlay processing in parallel based on the corrected image data. This technical solution reduces the number of dedicated image processing chips in the system, thereby effectively reducing the overall power consumption of the system. At the same time, since distortion correction is completed at the front end of the raw data stream, and multiple encoding / decoding and format conversion processes of the video signal are avoided, the data transmission and processing links in the image processing chain are significantly reduced, thereby fundamentally reducing the end-to-end latency from image acquisition to display of the fusion results.

[0023] In step S102 of this embodiment, extracting the infrared abnormal temperature rise region from the infrared radiation image means extracting the infrared abnormal temperature rise region whose temperature exceeds a preset threshold from the infrared radiation image, and extracting the ultraviolet spot region from the ultraviolet radiation image means extracting the ultraviolet spot region whose ultraviolet radiation value is greater than a preset threshold from the ultraviolet radiation image.

[0024] In step S102 of this embodiment, the spatial overlap of the infrared abnormal temperature rise region and the ultraviolet spot region refers to the overlap ratio of the coverage area of ​​the infrared abnormal temperature rise region and the ultraviolet spot region. The coverage area overlap ratio is the ratio of the intersection area of ​​the ultraviolet spot region and the infrared abnormal temperature rise region to the union area.

[0025] In step S103 of this embodiment, if the spatial overlap is less than the preset overlap, it is determined that a qualified fused image cannot be obtained; otherwise, it is determined that a qualified fused image can be obtained, and the fused image of the target high-voltage equipment is obtained by fusing the ultraviolet radiation image and the visible light image. The preset overlap ranges from 0.3 to 0.7, for example, it can be 0.5. It can be understood that a discharge defect in equipment is a continuous energy process that simultaneously releases ultraviolet photons and infrared thermal radiation, and the two have a high degree of spatial consistency. Therefore, by determining the qualification of fusion based on the spatial overlap of abnormal signals, typical environmental interference with only a single spectral feature can be effectively identified and filtered out, thereby improving the accuracy of defect identification and reducing the false alarm rate of the system. In addition, a qualified determination result is a prerequisite for triggering subsequent in-depth analysis, while an unqualified determination immediately drives the system to execute an adaptive adjustment strategy, so that limited onboard computing resources and inspection time can be dynamically and preferentially allocated to high-confidence targets, thereby optimizing the overall system performance.

[0026] To overcome the problem that existing technologies suffer from low image fusion efficiency and thus affect inspection and diagnostic capabilities due to the failure to consider the differential impact of optical lens distortion on multispectral registration strategies, such as... Figure 3 As shown, in step S103 of this embodiment, fusing the ultraviolet radiation image and the visible light image to obtain the fused image of the target high-voltage equipment includes: S201, Calculate the overall distortion intensity index of the visible light image. The functional expression for calculating the overall distortion intensity index of the visible light image is: ; ; ; in, This is the overall distortion intensity index for visible light images. This is the radial distortion offset. This is the tangential distortion offset. , , The radial distortion coefficient is... , The tangential distortion coefficient is... The normalized radial distance of the pre-selected points. The xy coordinates of the pre-selected point are the normalized image coordinates. In this embodiment, the pre-selected point is the endpoint of the diagonal of the image sensor's image plane, and its corresponding physical pixel coordinates can be directly determined based on the sensor's image plane size, and then converted into normalized image coordinates through lens intrinsic parameters. Then, the normalized radial distance of the pre-selected point is calculated. .

[0027] S202, select the corresponding registration method from the preset candidate registration methods based on whether the comprehensive distortion intensity index of the visible light image is less than the preset distortion intensity index; the preset distortion intensity index ranges from 0.01 to 0.05, and can be 0.02. S203, Register the ultraviolet radiation image and the visible light image to the same size according to the selected registration method; S204, the registered ultraviolet radiation image and the visible light image are weighted and fused to obtain the fused image of the target high-voltage equipment.

[0028] In step S202, when selecting a registration method from preset candidate registration methods based on whether the comprehensive distortion intensity index of the visible light image is less than a preset distortion intensity index, the selected registration method is feature point-based full-image affine transformation registration when the comprehensive distortion intensity index of the visible light image is less than the preset distortion intensity index. S301, extract feature points of SIFT or ORB features from ultraviolet radiation images and visible light images; SIFT or ORB features are existing known features, and the extraction algorithm is an existing known algorithm, so its implementation details will not be described here. S302, establish the correspondence between matching feature point pairs in the ultraviolet radiation image and the visible light image through a preset feature descriptor matching algorithm to obtain a set of matching feature point pairs; wherein, the feature descriptor matching algorithm is the basic algorithm for image matching, and known algorithms such as the FLANN algorithm can be used as needed; S303, a global affine transformation matrix is ​​obtained by fitting the set of matching feature point pairs using a preset robust estimation algorithm. The global affine transformation matrix contains some or all of the linear transformation parameters in translation, rotation, scaling and shearing. The robust estimation algorithm is the basic algorithm for fitting the global affine transformation matrix, and known algorithms such as RANSAC can be used as needed. S304 uses a global affine transformation matrix to perform a unified coordinate transformation on all pixel coordinates in the ultraviolet radiation image, mapping them to the pixel coordinate system of the visible light image, thereby obtaining the registered ultraviolet radiation image and the visible light image, thus achieving fast and global alignment of the two images.

[0029] Understandably, when the overall distortion intensity index is less than the preset distortion intensity index, it indicates that the geometric distortion of the ultraviolet and visible light imaging units is slight, and its impact on the spatial consistency of the multispectral image is within an acceptable linear error range. Under this condition, the selection of a feature-point-based full-image affine transformation registration method is based on the consideration of optimal efficiency. This method, by extracting and matching global features between images and solving a unified linear transformation model, can minimize computational complexity and processing latency while meeting the requirements of fusion positioning accuracy, thereby optimizing the overall energy efficiency and real-time performance of the system in routine inspection tasks.

[0030] In step S202, when selecting a registration method from preset candidate registration methods based on whether the comprehensive distortion intensity index of the visible light image is less than a preset distortion intensity index, the selected registration method is nonlinear geometric correction and local optimization registration based on pre-calibrated parameters when the comprehensive distortion intensity index of the visible light image is greater than or equal to the preset distortion intensity index. S401, the ultraviolet radiation image and the visible light image are subjected to nonlinear geometric transformation using a pre-calibrated intrinsic parameter matrix and distortion coefficient vector to eliminate the inherent radial and tangential distortion of the lens, generating a preliminarily aligned distortion-free image pair; the pre-calibrated intrinsic parameter matrix and distortion coefficient vector of the ultraviolet and visible light imaging channels are called, and nonlinear geometric transformation is performed pixel by pixel on the two original images according to the camera imaging geometry model to eliminate the inherent radial and tangential distortion of the lens, generating a preliminarily aligned distortion-free image pair. The above method is an existing registration method, so its implementation details will not be described in detail here. S402: The ultraviolet radiation image and the visible light image in the initially aligned distortion-free image pair are divided into several equal-sized sub-regions. Within each sub-region, a local affine or homography transformation parameter is iteratively solved with the optimization objective of maximizing the mutual information or structural similarity (SSIM) between the ultraviolet image patch and the visible light image patch in that region. Here, mutual information or structural similarity (SSIM) are well-known index parameters, so their implementation details will not be described in detail here. S403 integrates local affine or homography transformation parameters to obtain global affine or homography transformation parameters; S404 uses global affine or homography transformation parameters to resample all pixel coordinates in the ultraviolet radiation image, obtaining the registered ultraviolet radiation image and visible light image, achieving pixel-level spatial alignment between the two.

[0031] When the overall distortion intensity index is greater than or equal to the preset distortion intensity index, it indicates that the geometric distortion of the imaging unit has reached a non-negligible nonlinear level. If a linear registration model is continued, significant registration errors will occur in the image edge regions, thereby compromising the reliability of multispectral information. Using a nonlinear geometric correction and local optimization registration method based on pre-calibrated parameters can perform rigorous distortion correction according to the lens's optical physical model, eliminating systematic geometric deviations at the source. Then, through iterative optimization of local regions, residual minor deformations can be compensated.

[0032] like Figure 4 As shown, in step S103 of this embodiment, after determining that a qualified fused image cannot be obtained, the following steps are also included: S501, Calculate the ultraviolet light discharge characteristic value of the ultraviolet spot region: ; in, These are characteristic values ​​of ultraviolet light discharge. and These are the weighting coefficients. The maximum pixel intensity in the ultraviolet light spot area. The preset pixel intensity threshold, The area fluctuation rate of the ultraviolet light spot region. The ultraviolet discharge characteristic value is determined by the maximum pixel intensity and the spot area fluctuation rate, with a preset volatility threshold. In this embodiment, the weighting coefficient... The value is 0.7, which is the preset pixel intensity threshold. The value is set to 90% of the sensor's maximum recordable value, with a weighting coefficient. The value is 0.3, which is the preset volatility threshold. The value is 0.2; where the maximum pixel intensity is the maximum gray value among all pixels identified as light spots in a single frame of ultraviolet image; the calculation process of the light spot area fluctuation rate of the ultraviolet light spot region includes: capturing N frames (e.g., N=10) of registered and ultraviolet radiation images in a continuous time series and extracting the ultraviolet light spot region, calculating the pixel area of ​​the ultraviolet light spot region in each frame respectively, thereby obtaining a time series of the pixel area of ​​the ultraviolet light spot region for N frames, calculating the standard deviation and mean of the time series of the pixel area of ​​the ultraviolet light spot region respectively, and calculating the ratio of the standard deviation to the mean as the obtained light spot area fluctuation rate; S502, if the ultraviolet discharge characteristic value is greater than or equal to the preset discharge characteristic value, then proceed to step S503; otherwise, proceed to step S504. The preset threshold value is 0.8, but this value is not limited to this. Those skilled in the art can adjust this value according to actual needs. S503, Improve the integration time of the ultraviolet imaging sensor: Calculate the difference between the ultraviolet discharge characteristic value and the preset discharge characteristic value as the ultraviolet discharge difference value. If the ultraviolet discharge difference value is less than the preset ultraviolet discharge difference value, multiply the initial integration time of the ultraviolet imaging sensor by the first integration time adjustment coefficient or use it as the new integration time of the ultraviolet imaging sensor; if the ultraviolet discharge difference value is greater than or equal to the preset ultraviolet discharge difference value, multiply the initial integration time of the ultraviolet imaging sensor by the second integration time adjustment coefficient or use it as the new integration time of the ultraviolet imaging sensor; the first integration time adjustment coefficient and the second integration time adjustment coefficient are both constants greater than 1, and the first integration time adjustment coefficient is less than the second integration time adjustment coefficient; the integration time of the ultraviolet imaging sensor refers to the length of time that the photosensitive element accumulates charge on the incident ultraviolet light during one image acquisition process; jump to step S505. S504, reduce the weight of the ultraviolet radiation image when the registered image is fused with the visible light image pixel by pixel: calculate the circularity of the ultraviolet spot region; if the circularity of the ultraviolet spot region is less than a preset circularity, multiply the weight of the initial ultraviolet radiation image by a first weight adjustment coefficient to obtain the weight of the new ultraviolet radiation image; if the circularity of the ultraviolet spot region is greater than or equal to the preset circularity, multiply the weight of the initial ultraviolet radiation image by a second weight adjustment coefficient to obtain the weight of the new ultraviolet radiation image; both the first and second weight adjustment coefficients are constants less than 1, and the first weight adjustment coefficient is less than the second weight adjustment coefficient; jump to step S505; S505: The UAV reacquires an ultraviolet radiation image at the current location and calculates the ultraviolet discharge characteristic value of the ultraviolet spot area. If the ultraviolet discharge characteristic value is still greater than or equal to the preset threshold, the UAV is controlled to adjust the observation posture and jump to step S101 to re-inspect the target high-voltage equipment.

[0033] Understandably, the maximum pixel intensity directly reflects the instantaneous energy density of the discharge point, serving as the basis for determining whether the sensor's operating range is saturated or whether the signal is valid. The spot area fluctuation rate quantifies the signal's temporal stability, effectively distinguishing between fixed discharge sources and random motion interference. Determining the ultraviolet discharge characteristic value based on the maximum pixel intensity and spot area fluctuation rate enables a unified quantitative assessment of signal quality and reliability. When the ultraviolet discharge characteristic value is greater than or equal to the preset discharge characteristic value, it indicates that the signal faces the risk of saturation or distortion. The system prioritizes adjusting the integration time of the ultraviolet imaging sensor, calibrating the sensor's operating state from the data acquisition source and ensuring the validity and linearity of the original information. When the ultraviolet discharge characteristic value is less than the preset discharge characteristic value, it indicates that the current ultraviolet discharge signal is within the sensor's linear response range and is relatively stable, without exhibiting typical interference patterns such as saturation risk or severe jitter caused by excessively strong signals. Therefore, the system reduces the weight of this channel in the fusion decision based on the morphological characteristics of the ultraviolet spot, achieving real-time assessment of information reliability and optimal reallocation of decision resources.

[0034] Specifically, the adjustment range of the integration time of the ultraviolet imaging sensor is positively correlated with the ultraviolet discharge difference, where the ultraviolet discharge difference is the difference between the ultraviolet discharge characteristic value and a preset discharge characteristic value. If the ultraviolet discharge difference is less than the preset ultraviolet discharge difference, the integration time of the ultraviolet imaging sensor is increased to the corresponding value using a first integration time adjustment coefficient; if the ultraviolet discharge difference is greater than or equal to the preset ultraviolet discharge difference, the integration time of the ultraviolet imaging sensor is increased to the corresponding value using a second integration time adjustment coefficient. In this embodiment, the preset ultraviolet discharge difference is 0.1, the initial ultraviolet imaging sensor integration time is 1ms, the first integration time adjustment coefficient is 1.2, the second integration time adjustment coefficient is 1.5, and the adjusted ultraviolet imaging sensor integration time is the product of the initial ultraviolet imaging sensor integration time and the integration time adjustment coefficient. Specifically, the adjustment range of the decision weight of the ultraviolet channel in multispectral feature fusion is negatively correlated with the circularity of the ultraviolet spot region. If the circularity of the ultraviolet spot region is less than a preset circularity, the decision weight of the ultraviolet channel is reduced to the corresponding value using a first weight adjustment coefficient; if the circularity of the ultraviolet spot region is greater than or equal to the preset circularity, the decision weight of the ultraviolet channel is reduced to the corresponding value using a second weight adjustment coefficient. In this embodiment, the circularity of the ultraviolet spot region is calculated by extracting the binarized region of the target ultraviolet spot from a single frame of registered ultraviolet image using an image segmentation algorithm. The preset circularity value is 0.7, the initial decision weight of the ultraviolet channel is 0.5, the first weight adjustment coefficient is 0.6, and the second weight adjustment coefficient is 0.8. It should be noted that circularity is a shape descriptor used in image processing and computer vision to measure how close a region's shape is to an ideal circle. It has wide applications in target detection, cell analysis, particle recognition, and other fields. Various existing known methods can be used to calculate the circularity of the ultraviolet spot region. In this embodiment, the control module can determine a strategy for adjusting the observation pose of the UAV for re-inspection based on the comparison result that the ultraviolet discharge characteristic value after adjusting the integration time of the ultraviolet imaging sensor or adjusting the weight of multispectral feature fusion is greater than or equal to a preset discharge characteristic value. In this embodiment, the way to control the UAV to adjust the observation pose is, for example, to control the UAV to move at a constant speed to the side (such as to the left or right) for a preset distance and then hover, the preset distance being, for example, 3 to 5 meters, or to control the UAV gimbal or fuselage to yaw and rotate around the current target point by a preset angle and then hover, the preset angle being, for example, 15 to 30 degrees.

[0035] In this embodiment of the invention, the control module employs a two-axis FOC compact gimbal to achieve active, high-precision mechanical stabilization and angle adjustment of the imaging module's pitch and roll axes. Its yaw axis control is synchronized with the yaw rotation of the UAV itself. During the re-inspection process, this gimbal effectively compensates for additional roll and pitch jitter generated during the UAV's translation or rotation, ensuring that the multispectral image sequences acquired in the new pose remain clear and stable, thereby guaranteeing the accuracy and effectiveness of the re-inspection comparison analysis.

[0036] Those skilled in the art will understand that the technical solutions provided by this invention can take the form of a method, system, or computer program product. For example, this invention can provide a high-voltage equipment imaging system based on ultraviolet multispectral fusion, including a microprocessor and a memory interconnected, wherein the microprocessor is programmed or configured to execute the high-voltage equipment imaging method based on ultraviolet multispectral fusion. This invention can provide a computer-readable storage medium storing a computer program or instructions programmed or configured to execute the high-voltage equipment imaging method based on ultraviolet multispectral fusion via a processor. This invention can provide a computer program product including a computer program or instructions programmed or configured to execute the high-voltage equipment imaging method based on ultraviolet multispectral fusion via a processor. Furthermore, this invention can also take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, this invention can take the form of a computer program product embodied on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each block of a flowchart and / or block diagram, and combinations of blocks in a flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0037] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A high-voltage equipment imaging method based on ultraviolet multispectral fusion, characterized in that, Includes the following steps: S101, acquire ultraviolet radiation image, infrared radiation image and visible light image of the target high-voltage equipment; S102, extract the infrared abnormal temperature rise region from the infrared radiation image, extract the ultraviolet spot region from the ultraviolet radiation image, and calculate the spatial overlap between the infrared abnormal temperature rise region and the ultraviolet spot region. S103, determine whether the fused image of the ultraviolet radiation image and the visible light image is qualified based on the spatial overlap. If the spatial overlap is less than the preset overlap, it is determined that a qualified fused image cannot be obtained. Otherwise, it is determined that a qualified fused image can be obtained, and the fused image of the ultraviolet radiation image and the visible light image is used to obtain the fused image of the target high-voltage equipment.

2. The high-voltage equipment imaging method based on ultraviolet multispectral fusion according to claim 1, characterized in that, In step S102, extracting the infrared abnormal temperature rise region from the infrared radiation image means extracting the infrared abnormal temperature rise region whose temperature exceeds a preset threshold from the infrared radiation image. Extracting the ultraviolet spot region from the ultraviolet radiation image means extracting the ultraviolet spot region whose ultraviolet radiation value is greater than a preset threshold from the ultraviolet radiation image.

3. The high-voltage equipment imaging method based on ultraviolet multispectral fusion according to claim 1, characterized in that, In step S102, the spatial overlap of the infrared abnormal temperature rise region and the ultraviolet spot region is calculated as the coverage area overlap ratio of the infrared abnormal temperature rise region and the ultraviolet spot region. The coverage area overlap ratio is the ratio of the intersection area of ​​the ultraviolet spot region and the infrared abnormal temperature rise region to the union area.

4. The high-voltage equipment imaging method based on ultraviolet multispectral fusion according to claim 1, characterized in that, Step S103, which involves fusing the ultraviolet radiation image and the visible light image to obtain the fused image of the target high-voltage equipment, includes: S201, calculate the comprehensive distortion intensity index of the visible light image according to the following formula: ； ； ； in, This is the overall distortion intensity index for visible light images. This is the radial distortion offset. This is the tangential distortion offset. , , The radial distortion coefficient is... , The tangential distortion coefficient is... The normalized radial distance of the pre-selected points. The normalized xy coordinates of the pre-selected points; S202, Select the corresponding registration method from the preset candidate registration methods based on whether the comprehensive distortion intensity index of the visible light image is less than the preset distortion intensity index; S203, Register the ultraviolet radiation image and the visible light image to the same size according to the selected registration method; S204, the registered ultraviolet radiation image and the visible light image are weighted and fused to obtain the fused image of the high-voltage equipment.

5. The high-voltage equipment imaging method based on ultraviolet multispectral fusion according to claim 4, characterized in that, In step S202, when selecting a registration method from preset candidate registration methods based on whether the comprehensive distortion intensity index of the visible light image is less than a preset distortion intensity index, the selected registration method is feature point-based full-image affine transformation registration when the comprehensive distortion intensity index of the visible light image is less than the preset distortion intensity index. S301, extract feature points of SIFT or ORB features from ultraviolet radiation images and visible light images; S302, establish the correspondence between matching feature point pairs in the ultraviolet radiation image and the visible light image through a preset feature descriptor matching algorithm, and obtain a set of matching feature point pairs; S303, a global affine transformation matrix is ​​obtained by fitting the set of matching feature point pairs using a preset robust estimation algorithm. The global affine transformation matrix contains some or all of the linear transformation parameters in translation, rotation, scaling and shearing. S304 uses a global affine transformation matrix to perform a uniform coordinate transformation on all pixel coordinates in the ultraviolet radiation image, mapping them to the pixel coordinate system of the visible light image, thereby obtaining the registered ultraviolet radiation image and the visible light image.

6. The high-voltage equipment imaging method based on ultraviolet multispectral fusion according to claim 4, characterized in that, In step S202, when selecting a registration method from preset candidate registration methods based on whether the comprehensive distortion intensity index of the visible light image is less than a preset distortion intensity index, the selected registration method is nonlinear geometric correction and local optimization registration based on pre-calibrated parameters when the comprehensive distortion intensity index of the visible light image is greater than or equal to the preset distortion intensity index. S401, the ultraviolet radiation image and the visible light image are subjected to nonlinear geometric transformation using a pre-calibrated intrinsic parameter matrix and distortion coefficient vector to eliminate the inherent radial and tangential distortion of the lens and generate a preliminarily aligned distortion-free image pair. S402, for the ultraviolet radiation image and visible light image in the initially aligned distortion-free image pair, divide them into several equal-sized sub-regions, and in each sub-region, iteratively solve a local affine or homography transformation parameter with the optimization objective of maximizing the mutual information or structural similarity between the ultraviolet image block and the visible light image block in that region. S403 integrates local affine or homography transformation parameters to obtain global affine or homography transformation parameters; S404 uses global affine or homography transformation parameters to resample all pixel coordinates in the ultraviolet radiation image, obtaining the registered ultraviolet radiation image and visible light image, achieving pixel-level spatial alignment between the two.

7. The high-voltage equipment imaging method based on ultraviolet multispectral fusion according to claim 1, characterized in that, After determining in step S103 that a qualified fused image cannot be obtained, the following steps are also included: S501, Calculate the ultraviolet light discharge characteristic value of the ultraviolet spot region: ； in, These are characteristic values ​​of ultraviolet light discharge. and These are the weighting coefficients. The maximum pixel intensity in the ultraviolet light spot area. The preset pixel intensity threshold, The area fluctuation rate of the ultraviolet light spot region. This is a preset volatility threshold; S502, If the ultraviolet discharge characteristic value is greater than or equal to the preset discharge characteristic value, then proceed to step S503; otherwise, proceed to step S504. S503, calculate the difference between the ultraviolet discharge characteristic value and the preset discharge characteristic value as the ultraviolet discharge difference value. If the ultraviolet discharge difference value is less than the preset ultraviolet discharge difference value, multiply the initial integration time of the ultraviolet imaging sensor by the first integration time adjustment coefficient or use it as the new integration time of the ultraviolet imaging sensor. If the ultraviolet discharge difference value is greater than or equal to the preset ultraviolet discharge difference value, multiply the initial integration time of the ultraviolet imaging sensor by the second integration time adjustment coefficient or use it as the new integration time of the ultraviolet imaging sensor. The first integration time adjustment coefficient and the second integration time adjustment coefficient are both constants greater than 1, and the first integration time adjustment coefficient is less than the second integration time adjustment coefficient. The integration time of the ultraviolet imaging sensor refers to the length of time that the photosensitive element of the ultraviolet imaging sensor accumulates charge on the incident ultraviolet light during one image acquisition process. Jump to step S505. S504, calculate the circularity of the ultraviolet spot region. If the circularity of the ultraviolet spot region is less than a preset circularity, multiply the weight of the initial ultraviolet radiation image by a first weight adjustment coefficient to obtain the weight of the new ultraviolet radiation image. If the circularity of the ultraviolet spot region is greater than or equal to the preset circularity, multiply the weight of the initial ultraviolet radiation image by a second weight adjustment coefficient to obtain the weight of the new ultraviolet radiation image. Both the first and second weight adjustment coefficients are constants less than 1, and the first weight adjustment coefficient is less than the second weight adjustment coefficient. Skip to step S505. S505: The UAV reacquires an ultraviolet radiation image at the current location and calculates the ultraviolet discharge characteristic value of the ultraviolet spot area. If the ultraviolet discharge characteristic value is still greater than or equal to the preset threshold, the UAV is controlled to adjust the observation posture and jump to step S101 to re-inspect the target high-voltage equipment.

8. A high-voltage equipment imaging system based on ultraviolet multispectral fusion, comprising a microprocessor and a memory interconnected, characterized in that, The microprocessor is programmed or configured to execute the high-voltage equipment imaging method based on ultraviolet multispectral fusion as described in any one of claims 1 to 7.

9. A computer-readable storage medium storing a computer program or instructions, characterized in that, The computer program or instructions are programmed or configured to execute, via a processor, the high-voltage equipment imaging method based on ultraviolet multispectral fusion as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program or instructions, characterized in that, The computer program or instructions are programmed or configured to execute, via a processor, the high-voltage equipment imaging method based on ultraviolet multispectral fusion as described in any one of claims 1 to 7.

Images