Automatic intelligent defect analysis method and system based on power transmission and distribution line unmanned aerial vehicle image acquisition

Through the drone, infrared thermal imaging and multispectral imaging data are collected, combined with the temperature gradient direction and oxidation boundary profile, visual correlation maps are generated, and dynamic correlation rules are established, which solves the problem of insufficient detection accuracy of wire connection points in transmission and distribution lines, and accurately identify and risk classification of wire overheating defects.

CN120219390AActive Publication Date: 2025-06-27YUNNAN HUADIAN FUXIN ENERGY POWER GENERATION CO LTD +1

Patent Information

Application Number
CN202510696345.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-05-28
Publication Date
2025-06-27
Estimated Expiration
2045-05-28

Smart Images

  • Figure CN120219390A_ABST
    Figure CN120219390A_ABST
Patent Text Reader

Abstract

The invention provides an automatic intelligent defect analysis method and system based on power transmission and distribution line unmanned aerial vehicle image acquisition. The method comprises the following steps: acquiring infrared thermal imaging time sequence data and multispectral imaging data of the surface of a wire through an unmanned aerial vehicle carrying an infrared thermal imager and a multispectral camera; calculating a temperature gradient direction of a conductor main body area, screening an abnormal hot spot area whose axial included angle with the conductor is smaller than a preset angle, and positioning an oxidation area boundary by using a spectral absorption attenuation curve of an oxidation layer in the multispectral data; performing space matching on the hot spot migration direction and the oxidation boundary to generate a visual map; establishing a dynamic association rule according to an included angle variable quantity of a hot spot path in the map and a spectrum abrupt change point in the oxidation region; according to the critical angle threshold value and the sudden change point number threshold value, the defect grade is judged, and a defect analysis report is generated. According to the technical scheme provided by the invention, the automation degree and accuracy of defect detection are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the technical field of defect analysis, and particularly to an automated intelligent defect analysis method and system based on UAV image acquisition for power transmission and distribution lines. Background Art

[0002] Due to long-term exposure to complex environments (such as high load, oxidation corrosion, poor contact, etc.), the wire connection points of power transmission and distribution lines are prone to local overheating defects, which may lead to wire breakage or fire in severe cases. Traditional manual inspections cannot meet the real-time requirements. There is an urgent need for an automated method to achieve early warning and dynamic tracking of overheating defects at connection points, which needs to meet the following core requirements: accurately capture the minute thermal anomaly characteristics of wire connection points; support all-weather monitoring under complex lighting and weather conditions; and achieve quantitative analysis of defect evolution trends and risk grading.

[0003] Currently, the mainstream solution is a UAV dual-modal inspection system based on visible light cameras and thermal infrared imaging. This system uses a UAV equipped with a visible light camera to obtain high-definition images of wire connection points, combines a thermal infrared imager to collect surface temperature data, uses image matching technology to locate the connection point positions, and identifies temperature anomaly regions through threshold segmentation algorithms. The system generates preliminary warning signals by comparing historical temperature data and combines manual review to confirm the defect level.

[0004] Defects of the existing solutions include: visible light images are severely interfered by light, rain, and fog, and it is impossible to effectively identify the connection point positions at night or in bad weather, resulting in inaccurate positioning of thermal infrared data; relying on fixed temperature threshold segmentation, it is difficult to distinguish normal temperature rise and abnormal overheating at connection points (for example, the false alarm rate exceeds 40% in a high-temperature environment in summer), and it is impossible to capture early minute thermal changes; only relying on temperature comparison at discrete time points, lacking modeling analysis of dynamic characteristics such as the diffusion path and rate of thermal anomalies, it is difficult to provide a basis for trend prediction for operation and maintenance decisions. Summary of the Invention

[0005] This application provides an automated intelligent defect analysis method and system based on UAV image acquisition for power transmission and distribution lines to solve the problem of insufficient accuracy in defect detection in the prior art.

[0006] In a first aspect, this application provides an automated intelligent defect analysis method based on UAV image acquisition for power transmission and distribution lines, including: Collecting infrared thermal imaging time-series data and multi-spectral imaging data of the wire surface of the power transmission and distribution line within a continuous time window through a UAV; Calculate the temperature gradient direction of the wire body region in each frame of the infrared thermal imaging time series data, extract the abnormal hot spot region consistent with the wire axis from the temperature gradient directions, and at the same time, based on the spectral absorption attenuation curve corresponding to the wire surface oxide layer in the multispectral imaging data, locate the boundary contour of the wire surface oxidation region; Perform spatial topological matching between the migration direction of the abnormal hot spot region and the boundary contour to generate a visual association map integrating the hot spot diffusion path and the oxidation region distribution; According to the change amount of the angle between the hot spot diffusion path and the wire axis in the visual association map, and the mutation points of the spectral absorption attenuation rate within the boundary contour, establish a dynamic association rule between the wire overheating defect and the state of the surface oxide layer; Based on the dynamic association rule, determine the local overheating defect level caused by the deterioration of the oxide layer on the wire surface according to the critical angle threshold of the change amount of the angle and the number threshold of the mutation points, and generate a corresponding defect analysis report according to the local overheating defect level.

[0007] Optionally, the establishing a dynamic association rule between the wire overheating defect and the state of the surface oxide layer according to the change amount of the angle between the hot spot diffusion path and the wire axis in the visual association map, and the mutation points of the spectral absorption attenuation rate within the boundary contour includes: In the visual association map, along the extension direction of the hot spot diffusion path, divide continuous detection intervals with the wire axis as the reference axis, quantify the change amount of the angle between the hot spot diffusion path and the wire axis within the continuous detection intervals, and generate an angle change sequence; Within the coverage area of the boundary contour, divide equal-area detection units, count the number of jumps of the spectral absorption attenuation rate from a stable state to a mutation within the detection units, and generate a jump number distribution map of each detection unit; Mark the section where the angle increment between adjacent detection intervals in the angle change sequence exceeds the preset increase amplitude as an abnormal diffusion section, and at the same time, mark the detection units with the number of jumps exceeding the preset number in the jump number distribution map as active deterioration units; According to the coverage overlap degree between the spatial position of the abnormal diffusion section and the spatial density distribution of the active deterioration units, set the linkage determination rule of the angle change amount of the hot spot diffusion path and the number of jumps of the spectral absorption attenuation rate as the dynamic association rule.

[0008] Optionally, calculating the temperature gradient direction of the wire body region in each frame of the infrared thermal imaging time-series data, extracting the abnormal hot spot region consistent with the wire axis direction from the temperature gradient directions, and at the same time, based on the spectral absorption attenuation curve corresponding to the wire surface oxide layer in the multispectral imaging data, positioning the boundary contour of the wire surface oxidation region, including: In the wire body region of each frame in the infrared thermal imaging time-series data, equally spaced temperature detection bands are divided along the wire axis extension direction, and temperature sampling points are arranged in the temperature detection bands along the direction perpendicular to the wire axis; In the temperature detection band, calculate the temperature difference between the temperature sampling point and its adjacent sampling point, generate the temperature gradient direction along the arrangement direction of the temperature sampling points according to the temperature difference, statistically analyze the direction distribution histogram of the temperature gradient direction, and screen out the temperature gradient directions with an angle less than the preset angle with the wire axis in the direction distribution histogram; Perform spatial clustering on the filtered temperature gradient directions in the wire body region, merge the regions with continuous distribution and temperature gradient direction consistency exceeding the lowest ratio in the spatial clustering results, and mark them as abnormal hot spot regions; In the multispectral imaging data, divide spectral sampling bands along the wire surface contour line, extract the slope change points of the spectral absorption attenuation curve corresponding to the wire surface oxide layer in the spectral sampling bands, and connect the sampling band boundaries with a curve segment length less than the preset distance between adjacent slope change points to form a closed boundary contour of the wire surface oxidation region.

[0009] Optionally, performing spatial topological matching between the migration direction of the abnormal hot spot region and the boundary contour to generate a visual association map integrating the hot spot diffusion path and the oxidation region distribution, including: Generate a hot spot diffusion path according to the time-series position change path of the abnormal hot spot region, divide the hot spot diffusion path into equal-length path segments along the wire axis extension direction, and extract the migration direction line of the abnormal hot spot region in the equal-length path segments, where the migration direction line is the central extension trajectory of the hot spot diffusion path in the equal-length path segments; Divide equal-width detection grids along the boundary contour, calculate the oxidation region distribution density in the equal-width detection grids, and generate an oxidation region density distribution map; Perform spatial superposition of the migration direction line and the oxidation region density distribution map, analyze the extension trend of the migration direction line and the dynamic change trend of the oxidation region density distribution in the corresponding grid according to the spatial superposition result, and mark the path segments with the extension trend direction consistent with the density increase region direction as the interaction sections of heat and oxidation; Calculate the change rate of the extension angle of the migration direction line, and at the same time extract the density change direction of the density distribution map of the oxidation region. Perform spatial consistency analysis on the change rate of the extension angle and the density change direction within the action section, and convert the analysis result into a single visualization layer to generate a visualization correlation map.

[0010] Optionally, setting the linkage determination rule of the angle change amount of the hot spot diffusion path and the jump times of the spectral absorption attenuation rate as a dynamic correlation rule according to the coverage coincidence degree between the spatial position of the abnormal diffusion section and the spatial density distribution of the active deterioration unit, includes: Divide the wire surface area corresponding to the abnormal diffusion section into equal-area analysis units, calculate the coverage coincidence degree between the number of active deterioration units and the total number of grids in the equal-area analysis unit. When the coverage coincidence degree exceeds the preset coincidence ratio, mark the equal-area analysis unit as a thermal oxidation correlation unit; In the abnormal diffusion section corresponding to the thermal oxidation correlation unit, extract the angle increment sequence of adjacent detection intervals, calculate the correlation fluctuation amplitude between the coverage coincidence degree and the angle increment sequence. When the correlation fluctuation amplitude exceeds the fluctuation upper limit, mark the detection section as a significant diffusion section; In the significant diffusion section, calculate the product of the coverage coincidence degree and the total angle increment, and obtain the thermal oxidation correlation index after normalizing the product; Establish the linkage determination rule of the angle change amount of the hot spot diffusion path and the jump times of the spectral absorption attenuation rate according to the thermal oxidation correlation index. When the thermal oxidation correlation index exceeds the preset standard value, it is determined that there is a wire overheating defect caused by the deterioration of the oxide layer.

[0011] Optionally, within the temperature detection band, calculate the temperature difference between the temperature sampling point and its adjacent sampling point, generate the temperature gradient direction according to the temperature difference along the arrangement direction of the temperature sampling points, and statistically analyze the direction distribution histogram of the temperature gradient direction, and screen out the temperature gradient direction with an angle less than the preset angle with the wire axis direction, including: Within the temperature detection band, calculate the temperature difference between the temperature sampling point and the left and right adjacent sampling points, and generate the temperature gradient direction pointing from low temperature to high temperature according to the positive and negative signs and absolute value sizes of the temperature differences; Taking the wire axis direction as the reference direction, statistically analyze the direction angle distribution of the temperature gradient direction, and divide the direction angle distribution into a direction distribution histogram according to the preset angle interval; Screen out the temperature gradient direction with an angle less than the preset angle with the wire axis direction from the direction distribution histogram, and the preset angle is set according to the maximum allowable thermal diffusion offset amount during the normal operation of the wire; Quantitatively analyze the spatial distribution and direction angle deviation of the selected temperature gradient direction. The quantitative analysis includes the spacing statistics of adjacent vectors and the cumulative distribution calculation of the direction angle deviation.

[0012] Optionally, through the dynamic association rule, according to the critical angle threshold of the included angle change amount and the number threshold of the mutation points, determine the local overheating defect level caused by the deterioration of the oxide layer on the wire surface, and generate a corresponding defect analysis report according to the local overheating defect level, including: Through the dynamic association rule, set the range of the critical angle threshold of the included angle change amount, which is determined according to the physical upper limit value of the abnormal rate of axial heat diffusion of the wire. At the same time, set the interval of the number threshold of the mutation points, which is determined according to the spectral response characteristics of the activity of the oxide layer deterioration; Establish a combination relationship table between the range and the interval. The combination relationship table defines the matching intervals of the critical angle threshold and the number threshold corresponding to different levels of defects; According to the range and the interval, screen the sections in the detected area of the wire surface where the included angle change amount and the number of mutation points respectively fall into the corresponding threshold intervals as the sections to be determined. The sections to be determined are mapped to the initial defect level according to the combination relationship table; Perform continuity correction on the initial defect level. When the defect levels of three consecutive sections are the same and the increasing trend of the included angle change amount is consistent, merge the three sections and upgrade one defect level to output the local overheating defect level, and generate a defect analysis report including the defect position distribution, level intensity and diffusion trend according to the local overheating defect level.

[0013] In a second aspect, the present application provides an automated intelligent defect analysis system for drone image acquisition of power transmission and distribution lines, including: An acquisition module for acquiring infrared thermal imaging time-series data and multispectral imaging data of the wire surface of the power transmission and distribution line through a drone within a continuous time window; A calculation module for calculating the temperature gradient direction of the main wire area in each frame of the infrared thermal imaging time-series data, extracting the abnormal hot spot area consistent with the wire axis from the temperature gradient direction, and at the same time, based on the spectral absorption attenuation curve corresponding to the oxide layer on the wire surface in the multispectral imaging data, positioning the boundary contour of the oxidized area on the wire surface; A matching module for performing spatial topological matching between the migration direction of the abnormal hot spot area and the boundary contour to generate a visual association map integrating the hot spot diffusion path and the oxidized area distribution; A building module is configured to establish a dynamic association rule between the overheating defect of the wire and the state of the surface oxide layer according to the change amount of the angle between the hot spot diffusion path in the visualization association map and the axial direction of the wire, and the mutation point of the spectral absorption attenuation rate within the boundary contour; A generation module is configured to determine the local overheating defect level caused by the deterioration of the oxide layer on the wire surface based on the dynamic association rule, according to the critical angle threshold of the change amount of the angle and the number threshold of the mutation points, and generate a corresponding defect analysis report according to the local overheating defect level.

[0014] In a third aspect, an embodiment of the present application provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a method for automatically analyzing intelligent defects based on UAV image acquisition of power transmission and distribution lines as described in the first aspect above.

[0015] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program, and when the computer program is executed by a computer, it implements a method for automatically analyzing intelligent defects based on UAV image acquisition of power transmission and distribution lines as described in the first aspect.

[0016] In the embodiment of the present application, infrared thermal imaging time-series data and multi-spectral imaging data of the wire surface of the power transmission and distribution line are collected by a UAV within a continuous time window; the temperature gradient direction of the wire main body region in each frame of the infrared thermal imaging time-series data is calculated, and an abnormal hot spot region consistent with the axial direction of the wire is extracted from the temperature gradient direction. At the same time, based on the spectral absorption attenuation curve corresponding to the surface oxide layer of the wire in the multi-spectral imaging data, the boundary contour of the oxidized region on the wire surface is located; the migration direction of the abnormal hot spot region is topologically matched with the boundary contour to generate a visualization association map integrating the hot spot diffusion path and the oxidized region distribution; according to the change amount of the angle between the hot spot diffusion path in the visualization association map and the axial direction of the wire, and the mutation point of the spectral absorption attenuation rate within the boundary contour, a dynamic association rule between the overheating defect of the wire and the state of the surface oxide layer is established; based on the dynamic association rule, according to the critical angle threshold of the change amount of the angle and the number threshold of the mutation points, the local overheating defect level caused by the deterioration of the oxide layer on the wire surface is determined, and a corresponding defect analysis report is generated according to the local overheating defect level.

[0017] The technical solution of the present application has the following beneficial effects: By synchronously acquiring infrared thermal imaging timing data and multispectral imaging data, multi-dimensional information fusion of the dynamic changes in the surface temperature field of the wire and the spectral characteristics of the oxide layer is achieved, enhancing the comprehensiveness of defect detection; screening abnormal hot spot areas consistent with the wire axis based on the temperature gradient direction to exclude non-axial heat diffusion interference; accurately positioning the oxidation boundary in combination with the spectral absorption attenuation curve to enhance the specificity of material degradation identification; intuitively presenting the spatio-temporal coupling relationship between the hot spot diffusion path and the oxidation distribution through the spatial topology matching of the hot spot migration direction and the oxidation area, providing a visual carrier for defect cause analysis; constructing a causal determination model for oxide layer degradation and overheating defects based on the quantitative relationship between the change amount of the hot spot path angle and the spectral mutation point to enhance the physical interpretability of defect classification; realizing automatic grading of defect levels through the linkage rule of the critical threshold and the quantity threshold, and generating a structured report in combination with the defect location, level, and diffusion trend to support precise operation and maintenance decisions.

[0018] Furthermore, in the visual association map, continuous detection intervals are divided along the hot spot diffusion path and the sequence of the angle change with the wire axis is quantified; equal-area detection units are divided within the oxidation boundary contour and the distribution of the jump times of the spectral absorption attenuation rate is statistically analyzed; abnormal diffusion sections with excessive angle increment and active degradation units with excessive jump times are screened, and a dynamic association rule between the angle change amount and the jump times is established according to the spatial coverage coincidence degree between the two. Through the collaborative analysis of the dynamic characteristics of hot spot diffusion in the infrared timing data and the spatial distribution of multispectral oxidation degradation activity, the problem that traditional methods cannot associate the thermal performance of defects with the root cause of material degradation is solved, realizing the precise binding of wire overheating defects and the state of the oxide layer, enhancing the defect detection accuracy and reducing the false alarm rate.

[0019] These aspects or other aspects of the present application will be more clearly understood in the following description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0021] Figure 1 Shows a flowchart of an automated intelligent defect analysis method based on UAV image acquisition for power transmission and distribution lines provided by the present application; Figure 2 Shows a schematic structural diagram of an automated intelligent defect analysis system based on UAV image acquisition for power transmission and distribution lines provided by the present application; Figure 3The structural schematic diagram of a computing device provided by the present application is shown. Specific embodiments

[0022] In order to enable those skilled in the art to better understand the solution of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application.

[0023] In some processes described in the specification, claims and above-mentioned drawings of the present application, a plurality of operations appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order in which they appear in this text or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions such as "first" and "second" in this text are used to distinguish different messages, devices, modules, etc., do not represent the sequence, and do not limit that "first" and "second" are different types.

[0024] Researchers have found that existing methods for detecting overheating defects in transmission and distribution line conductors mostly rely on a single data source (such as static infrared thermography or visible light imaging), making it difficult to effectively distinguish local overheating caused by surface oxidation deterioration of the conductor from instantaneous environmental interference, and unable to establish a dynamic correlation model between the thermal anomaly diffusion path and the state of the oxide layer, resulting in a high defect misjudgment rate and early warning lag. Based on this, the present invention provides an automated intelligent defect analysis method based on unmanned aerial vehicle (UAV) image acquisition of transmission and distribution lines. Specifically, by synchronously collecting infrared thermal imaging time-series data and multi-spectral imaging data of the conductor surface through a UAV, screening the axial abnormal hot spot area based on the temperature gradient direction and locating the oxidation boundary contour, generating a visual correlation map by fusing the hot spot migration direction and oxidation distribution, and establishing a dynamic correlation rule between the change amount of the hot spot diffusion angle and the spectral mutation point, to achieve defect level determination and trend analysis. This method can accurately identify local overheating defects caused by oxide layer deterioration, reduce the false alarm rate and support the visual tracking and risk grading of the defect evolution path, providing a reliable decision-making basis for the intelligent operation and maintenance of transmission lines.

[0025] The technical solution of the present application is applicable to the early warning and dynamic tracking scenarios of overheating defects at conductor connection points.

[0026] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative efforts shall fall within the protection scope of the present application.

[0027] Figure 1 The flowchart of an automated intelligent defect analysis method for UAV image acquisition of transmission and distribution lines provided by an embodiment of the present application is as follows. As Figure 1 shown, the method includes: 101. Collect infrared thermal imaging time-series data and multispectral imaging data of the surface of the conductor of the transmission and distribution line within a continuous time window through a UAV; In this step, the infrared thermal imaging time-series data is continuous data of the temperature field of the conductor surface changing with the flight path obtained by the UAV continuously scanning along the extension direction of the conductor. The multispectral imaging data is the data of the spectral reflectance intensity difference between the oxide layer on the surface of the conductor and the normal metal surface under the selected wavelength band obtained by the UAV hovering and taking pictures at a fixed point.

[0028] In an embodiment of the present application, first, a uniform flight path is planned along the axial direction of the conductor through the UAV to ensure that the infrared thermal imager continuously collects the temperature field data of the conductor surface at a frame rate of 30 Hz, and the multispectral camera is synchronously triggered to take spectral reflectance images in the preset wavelength band to obtain multispectral imaging data; second, radiation correction is performed on the infrared data, and the original radiation value is converted into an absolute temperature value based on the environmental temperature and material emissivity parameters (such as the emissivity of aluminum conductor is 0.95), and at the same time, reflectance calibration is performed on the multispectral imaging data, and the illumination interference is eliminated by using the whiteboard reference value; finally, the GPS / INS system carried by the UAV records the spatial coordinates and timestamps of each frame of data, and the linear interpolation algorithm is used to align the spatio-temporal positions of the infrared and multispectral data to ensure that the spatial error between the two is less than 5 cm.

[0029] Assume a scenario of detecting overheating defects at the connection point of the conductor of a 500 kV transmission line. The UAV flies 1.5 km along the conductor, the flight speed is 1 m / s, the infrared thermal imager collects 15 minutes of time-series data (temperature resolution 0.1 °C), and the multispectral camera obtains high-resolution images in 5 wavelength bands. The calibrated data shows that there are temperature anomaly regions (maximum temperature difference ±4 °C) and spectral reflectance anomaly regions (reflectance in the 650 nm wavelength band is 30% lower than the normal value) in the middle section of the conductor.

[0030] 102. Calculate the temperature gradient direction of the main body region of the conductor in each frame of the infrared thermal imaging time-series data, extract the abnormal hot spot region consistent with the axial direction of the conductor from the temperature gradient directions, and at the same time, based on the spectral absorption attenuation curve corresponding to the oxide layer on the surface of the conductor in the multispectral imaging data, locate the boundary contour of the oxide region on the surface of the conductor; In this step, the temperature gradient direction is the vector direction calculated by the temperature difference between adjacent pixels, which characterizes the heat diffusion trend. The spectral absorption attenuation curve is the curve of the reflection intensity of the oxide layer region changing with the wavelength, and the slope mutation point of it reflects the activity of oxidation deterioration.

[0031] In the embodiments of the present application, first, the Sobel operator is applied to each frame of infrared thermal imaging time-series data to calculate the temperature gradient direction. The calculation formula is θ = arctan(ΔTy / ΔTx), where ΔTx and ΔTy are the temperature differences of adjacent pixels in the X / Y direction. Secondly, vectors with an included angle less than 10° between the temperature gradient direction and the wire axis are screened, adjacent regions are merged through morphological closing operation (3×3 rectangular kernel), isolated noise points with an area less than 10 cm² are removed, and marked as abnormal hot spot regions. At the same time, for the hyperspectral imaging data, the Spectral Angle Mapping (SAM) algorithm is used. Taking the normal metal spectrum as the reference vector, the spectral similarity of each pixel is calculated (a threshold <0.9 is determined as the oxide layer), and the contour of the oxidation region is extracted by combining Canny edge detection. Finally, the boundary contour of the wire surface oxidation region is generated through polygon fitting.

[0032] For example, continuing with the above example, abnormal hot spots are found at position 3 in the middle section of the wire during the calculation of the temperature gradient direction, with a maximum gradient direction deviation of 7.5°. After morphological processing, they are merged into two continuous regions (with areas of 0.3 m² and 0.5 m² respectively). The hyperspectral SAM analysis locates the oxidation region with an area of 1.1 m², and the fitting accuracy of the boundary contour is ±2 cm, partially overlapping with the temperature anomaly region.

[0033] 103. Perform spatial topological matching between the migration direction of the abnormal hot spot region and the boundary contour to generate a visual correlation map that integrates the hot spot diffusion path and the oxidation region distribution. In this step, the visual correlation map is a spatial superposition map that integrates the hot spot diffusion path (time-series migration direction) and the oxidation region distribution, and is used to intuitively display the correlation between defects and the oxide layer.

[0034] In the embodiments of the present application, first, the central point coordinates of the migration direction of the abnormal hot spot region are subjected to least squares linear fitting to generate the hot spot diffusion path, where the path slope reflects the diffusion rate (such as 0.25 m / s). Secondly, the hot spot diffusion path and the boundary contour are imported into the same coordinate system, and a spatial correlation network is established through Delaunay triangulation, and the overlap ratio between the path and the boundary contour is calculated (such as a coincidence degree >70% is marked as a strong correlation). Thirdly, a visualization engine (such as Matplotlib) is used to render the map. The hot spot diffusion path is represented by a red gradient arrow to indicate the diffusion direction and rate, and the oxidation region is filled with semi-transparent blue, and a visual correlation map in PNG format is generated by overlaying.

[0035] For example, continuing with the above example, the hot spot diffusion path shows that the two abnormal regions diffuse along the wire axis at rates of 0.2 m / s and 0.35 m / s respectively. Among them, the path with a rate of 0.35 m / s coincides with the oxidation boundary by 82%, and is marked as a "high-risk correlation area" through the generated visual correlation map, while the coincidence degree of the other path is 45% and is marked as "medium risk".

[0036] 104. Establish a dynamic association rule between the overheating defect of the wire and the state of the surface oxide layer based on the change amount of the angle between the hot spot diffusion path and the axial direction of the wire in the visualized association map, and the mutation point of the spectral absorption attenuation rate within the boundary contour; In this step, the change amount of the angle is the real-time angle difference between the tangent direction of the hot spot diffusion path and the axial direction of the wire, reflecting the deviation degree of the abnormal heat diffusion. The spectral mutation point is the mutation position where the absolute value of the slope in the spectral absorption attenuation curve of the oxidation region is > 0.5, characterizing the activity of oxidation deterioration.

[0037] In the embodiment of the present application, first, divide the detection interval at intervals of 1 m along the hot spot diffusion path, and calculate the average value of the change amount of the angle between the path tangent and the axial direction in each interval (for example, interval 1: 5°, interval 2: 8°, increment 3°); second, divide a 10 cm × 10 cm grid within the oxidation boundary, apply a sliding window slope detection to the spectral curve in each grid (window width 10 nm), and count the number of mutation points where the absolute value of the slope exceeds 0.5; third, establish a dynamic rule based on a logistic regression model, with the input features being the average value of the change amount of the angle and the number of mutation points, and the output being the defect level (low / medium / high risk), and the model weights are optimized and determined through historical training data (such as 100 sets of samples).

[0038] For example, continuing the above example, the average value of the change amount of the angle in the high-risk association area is 6.2° / m, and the number of mutation points is 5 times per grid. The logistic regression model establishes a dynamic association rule, and the output defect level is "high risk" (confidence level 91%). The average value of the change amount of the angle in the medium-risk area is 3.8° / m, and the number of mutation points is 2 times per grid, and it is determined as "medium risk" (confidence level 76%).

[0039] 105. Based on the dynamic association rule, determine the local overheating defect level of the wire surface caused by the deterioration of the oxide layer according to the critical angle threshold of the change amount of the angle and the number threshold of the mutation points, and generate a corresponding defect analysis report according to the local overheating defect level.

[0040] In this step, the critical angle threshold is the upper limit of the change amount of the angle set according to the wire material, and an alarm is triggered if it is exceeded. The number threshold is the lower limit of the number of mutation points in the oxidation region grid, used to determine the activity of deterioration.

[0041] In the embodiments of the present application, first, a critical angle threshold (e.g., 5° / m for aluminum wires) and a mutation point quantity threshold (≥3 times / grid) are set according to the wire material. If the change in the included angle exceeds the threshold and the number of mutation points meets the standard, it is marked as "confirmed defect"; second, a structured defect analysis report is generated, integrating the defect location (GPS coordinates), a dynamic GIF of the thermal spot diffusion (an animation of sequential thermal spots superimposed), a high-definition screenshot of the oxidation area, and maintenance suggestions (such as "re-inspection within 48 hours"); finally, the report is automatically pushed to the responsible unit via email or the operation and maintenance platform.

[0042] For example, continuing with the above example, in the high-risk associated area, "confirmed defect" is triggered. The defect analysis report indicates that the defect is located 1.5 km from the wire (GPS: E118°45'36", N32°03'12"), and it is recommended to immediately cut off the power for maintenance and replace the oxidized section of the wire; in the medium-risk area, it is marked as "suspected defect", and it is recommended to conduct a re-inspection within 2 weeks and apply an anti-corrosion coating.

[0043] Steps 101 - 105 achieve precise correlation determination of wire overheating defects and oxidation deterioration through collaborative acquisition and fusion analysis of multi-source data by drones. Infrared sequential data dynamically tracks the thermal spot diffusion path, multi-spectral data locates the oxidation boundary, and spatial topology matching generates a visual map to support defect causality analysis; based on dynamic rule modeling of the change in the included angle and spectral mutation, it breaks through the limitations of traditional single-threshold criteria; automated report generation and pushing significantly improve the operation and maintenance response efficiency, and are applicable to intelligent inspection and preventive maintenance of high-voltage transmission lines.

[0044] To solve the problem of dynamic correlation determination of local overheating defects caused by oxidation deterioration at the wire connection points of power transmission and distribution lines, and further improve the accuracy of defect early warning and the ability to predict trends, in some embodiments, based on the change in the included angle between the thermal spot diffusion path in the visual correlation map and the wire axis, and the mutation points of the spectral absorption attenuation rate within the boundary contour, a dynamic correlation rule between the wire overheating defect and the surface oxide layer state is established, including: 201. In the visual correlation map, along the extension direction of the thermal spot diffusion path, continuous detection intervals are divided with the wire axis as the reference axis, and the change in the included angle between the thermal spot diffusion path and the wire axis within the continuous detection intervals is quantified to generate an included angle change sequence; In step 201, the continuous detection intervals are continuous analysis units equally spaced along the wire axis along the thermal spot diffusion path, used to quantify local diffusion characteristics. The change in the included angle is the real-time included angle difference between the tangent direction of the thermal spot diffusion path and the wire axis, reflecting the deviation degree of the heat diffusion direction.

[0045] In the embodiments of the present application, first, based on the spatial coordinates of the hotspot diffusion path in the visualization association graph, with the wire axis as the reference axis, continuous detection intervals are divided at a preset interval (such as 1 meter); second, cubic spline interpolation fitting is performed on the path within each interval, and the real-time angle between the tangent of the fitting curve and the axis is calculated (such as the starting angle of the interval is 5°, the ending angle is 9°, and the change amount is 4°); finally, an angle change sequence is generated in the order of the intervals (such as [5°, 9°, 12°, ...]) for subsequent abnormal diffusion determination.

[0046] 202. Within the coverage area of the boundary contour, equal-area detection units are divided, and the number of jumps of the spectral absorption attenuation rate from the stable state to the mutation within the detection units is counted to generate a jump number distribution map for each detection unit; In step 202, the equal-area detection units are fixed-area analysis grids divided within the coverage area of the oxidation boundary contour for local spectral mutation statistics. The jump number distribution map is a heat map of the number distribution of the spectral absorption attenuation rate from stability to mutation within each detection unit.

[0047] In the embodiments of the present application, first, the oxidation boundary contour is divided into equal-area grids of 10 cm × 10 cm; second, a sliding window mean filter (window width 5 nm) is performed on the spectral absorption attenuation curve within each grid, and the rate fluctuation amplitude within the window is calculated. If the rate difference between adjacent windows exceeds 15%, it is marked as one jump; finally, the number of jumps of all grids is counted to generate a two-dimensional number distribution heat map (such as red represents the high-jump area and blue represents the low-jump area).

[0048] 203. Sections in the angle change sequence where the angle increment between adjacent detection intervals exceeds a preset increase amplitude are marked as abnormal diffusion sections, and at the same time, detection units in the jump number distribution map with the number of jumps exceeding a preset number are marked as active deterioration units; In step 203, the abnormal diffusion section is a section in the continuous detection intervals where the adjacent angle increment exceeds the preset increase amplitude, representing the area of abnormal acceleration of heat diffusion. The active deterioration unit is a detection unit in the jump number distribution map with the number of jumps exceeding the preset threshold, reflecting the active core area of oxidation deterioration.

[0049] In the embodiments of the present application, first, differential calculation is performed on the angle change sequence to extract the angle increment between adjacent intervals (such as the increment from interval 1 to 2 is 3°, and the increment from interval 2 to 3 is 6°), and continuous sections with an increment ≥ 5° / m are screened (such as interval 2 - 3) and marked as abnormal diffusion sections; at the same time, grids with the number of jumps ≥ 4 are extracted from the jump number distribution map, and isolated noise points are removed through morphological erosion operations, and continuous distributed active deterioration units are retained.

[0050] 204. Set the linkage determination rule between the angle change amount of the hot spot diffusion path and the jump count of the spectral absorption attenuation rate as a dynamic correlation rule according to the coverage coincidence degree between the spatial position of the abnormal diffusion section and the spatial density distribution of the active deterioration units.

[0051] In step 204, the coverage coincidence degree is the overlapping ratio of the abnormal diffusion section and the active deterioration units in the spatial distribution, and is used to quantify the heat-oxidation correlation intensity.

[0052] In the embodiment of the present application, first calculate the area of the wire surface area covered by the abnormal diffusion section, and count the proportion of the number of active deterioration units in this area to quantify the coverage coincidence degree (such as the coverage area is 1 square meter and the proportion of active units is 70%); secondly, set the linkage determination rule based on the proportion (such as when the proportion ≥ 60%, for every 1° / m increase in the angle change amount, the corresponding jump count ≥ 2 times) to form a dynamic correlation rule library.

[0053] The following is a specific example: Suppose that the connection point of a 500 kV transmission line wire is locally overheated due to the deterioration of the oxide layer. The operation and maintenance party needs to track the heat diffusion trend in real time and determine the driving effect of the oxidation deterioration on the defect. Through the inspection by an unmanned aerial vehicle, it is found that the central temperature of the connection point has abnormally increased by 9 °C, and the infrared thermal imaging time-series data shows that the heat diffuses along the wire axis to both sides. In the visual correlation map, first divide the continuous detection interval at intervals of 1 meter along the hot spot diffusion path (step 201), and it is found that the angle change amount in the interval of 1 - 3 meters from the connection point rapidly increases from 7° to 19°, generating an angle change sequence [7°, 13°, 19°], indicating that the heat diffusion direction significantly deviates from the wire axis; then, divide the detection units of 10 cm × 10 cm within the oxidation boundary coverage area (step 202). The analysis of the spectral absorption attenuation curve of the core oxidation area (a circular area with a diameter of 0.6 m) of the connection point shows that the jump count exceeds 5 times in 60% of the units, forming a high-density active deterioration unit; subsequently, screen out the abnormal diffusion section (the 1 - 3 meter interval) with an angle increment exceeding 5° / m and the active deterioration units with a jump count ≥ 5 times (step 203). The spatial superposition analysis shows that the coverage coincidence degree between the two reaches 88%, and the heat diffusion rate is positively correlated with the jump count; finally, based on the coverage coincidence degree, set the dynamic correlation rule (step 204) to determine that the overheating defect of this connection point is dominated by oxidation deterioration, and the defect level is "critical".

[0054] In order to accurately identify the local overheating defects caused by oxidation deterioration of the wire connection points and solve the problems of misjudgment of the heat diffusion direction and fuzzy positioning of the oxidation boundary in the traditional methods, in some embodiments, calculate the temperature gradient direction of the wire main body region in each frame of the infrared thermal imaging time-series data, extract the abnormal hot spot region consistent with the wire axis from the temperature gradient directions, and at the same time, based on the spectral absorption attenuation curve corresponding to the wire surface oxide layer in the multispectral imaging data, locate the boundary contour of the wire surface oxidation region, including: 301. In the wire main body region of each frame of the infrared thermal imaging time-series data, divide equally spaced temperature detection bands along the extending direction of the wire axis, and arrange temperature sampling points in the temperature detection bands along the direction perpendicular to the wire axis; In step 301, the temperature sampling points are sampling positions evenly distributed in the temperature detection bands along the direction perpendicular to the wire axis, and the interval matches the wire diameter.

[0055] In the embodiments of the present application, first, according to the spatial coordinates of the wire axis, divide the wire main body region in the infrared thermal imaging data into equally spaced temperature detection bands at a preset interval (0.5 m); second, evenly set temperature sampling points in each temperature detection band along the direction perpendicular to the wire axis (i.e., the circumferential direction), and the sampling point interval is adaptively adjusted according to the wire surface curvature (for example, the interval is 1 cm in the straight section and 0.5 cm in the curved section), forming a transverse temperature sampling sequence.

[0056] 302. In the temperature detection bands, calculate the temperature difference between the temperature sampling points and their adjacent sampling points, generate a temperature gradient direction along the arrangement direction of the temperature sampling points according to the temperature difference, statistically analyze the direction distribution histogram of the temperature gradient direction, and screen the temperature gradient directions in the direction distribution histogram whose included angle with the wire axis is less than a preset angle; In step 302, the direction distribution histogram is a histogram statistically analyzing the direction angle distribution of all temperature gradient vectors and divided into preset intervals.

[0057] In the embodiments of the present application, first, for the temperature sampling points in each temperature detection band, calculate the temperature difference between it and its left and right adjacent points (such as the temperature difference of the left point is +2 °C, and the temperature difference of the right point is -3 °C), and generate a temperature gradient direction pointing from the low temperature to the high temperature; second, statistically analyze the direction angles of all temperature gradient directions, take the wire axis as the 0° reference, and generate a direction distribution histogram at 5° intervals; finally, screen the gradient vectors in the histogram whose included angle with the axis is less than 10° (such as the intervals of 0° - 10° and 170° - 180°), and exclude non-axial thermal interference.

[0058] 303. In the wire body region, perform spatial clustering on the filtered temperature gradient directions, merge the regions that show continuous distribution and whose temperature gradient direction consistency exceeds the lowest ratio in the spatial clustering results, and mark them as abnormal hot spot regions. In step 303, spatial clustering combines the filtered temperature gradient directions into continuous regions according to spatial proximity.

[0059] In the embodiment of the present application, first, use a density clustering algorithm (such as DBSCAN) to perform spatial clustering on the filtered temperature gradient directions and merge the gradient directions that are adjacent and have a direction deviation less than 5°; second, calculate the direction consistency ratio within each clustering region (for example, the vector direction deviation within the region <5° for 80%), and if the ratio exceeds 75%, mark it as an abnormal hot spot region.

[0060] 304. In the multispectral imaging data, divide spectral sampling bands along the wire surface contour line, extract the slope change points of the spectral absorption attenuation curve corresponding to the wire surface oxide layer within the spectral sampling bands, and connect the sampling band boundaries where the curve segment length between adjacent slope change points is less than a preset distance to form the boundary contour of the closed wire surface oxidation region.

[0061] In step 304, the spectral sampling band is a strip-shaped region divided along the wire surface contour line for local spectral analysis. The slope change point is the mutation position where the absolute value of the slope in the spectral absorption attenuation curve exceeds the threshold.

[0062] In the embodiment of the present application, first, divide spectral sampling bands with a width of 1 cm along the wire surface contour line; second, perform first-order derivative calculation on the spectral absorption attenuation curve within each sampling band and extract the mutation points where the absolute value of the slope exceeds 0.5; finally, if the curve segment length between adjacent mutation points is less than a preset distance (such as 5 mm), connect the two points to form the boundary contour of the oxidation region.

[0063] The following is a specific example: Suppose that the wire connection point of a 330 kV transmission line is locally overheated due to oxidation corrosion. It is necessary to accurately locate the defective area and analyze the correlation between the heat diffusion path and oxidation deterioration. In step 301, 4 temperature detection zones (one every 0.5 m) are divided in the connection point and the 2-meter wire areas on both sides. 20 temperature sampling points (spaced 1 cm apart) are set in each detection zone. The analysis in step 302 shows that the temperature gradient directions in the central detection zone of the connection point are concentrated in the range of 5°-8° (the angle with the axial direction), while the gradient directions in the detection zones on both sides are dispersed (15°-30°). The gradient directions with an axial angle <10° are screened out; in step 303, through spatial clustering, a continuous area (with an area of 0.12 ㎡) is found at the center of the connection point, and the direction consistency ratio reaches 85%, which is marked as an abnormal hot spot area; in step 304, spectral sampling zones are divided along the surface contour line of the wire, and 8 slope mutation points are extracted in the spectral sampling zone at the center of the connection point. The distance between adjacent mutation points is less than 3 mm. After connection, the boundary contour of the closed oxidation area (a circular area with a diameter of 0.25 m) is formed. Finally, the coincidence degree between the abnormal hot spot and the oxidation boundary reaches 90%, which is determined as a critical defect driven by oxidation, and a report is generated recommending immediate power-off maintenance and coating with anti-corrosion materials.

[0064] Steps 301-304 effectively distinguish the axial heat diffusion of the wire connection point from the non-axial environmental interference through the refined division of the axial temperature detection zones and spectral sampling zones, combined with gradient direction screening and spatial clustering; the oxidation boundary formed by connecting the spectral slope mutation points improves the positioning accuracy of the material deterioration area; the spatial coincidence degree analysis between the abnormal hot spot and the oxidation boundary realizes the physically interpretable determination of the defect cause, provides full-process support for the transmission line from early warning to precise operation and maintenance, and significantly reduces the risk of false detection and missed detection.

[0065] In order to accurately depict the dynamic coupling relationship between heat diffusion and oxidation deterioration in the overheating defect of the wire connection point, and to solve the problems of strong subjectivity in the determination of the heat-oxidation correlation and insufficient visualization expression in traditional methods, in some embodiments, the migration direction of the abnormal hot spot area is topologically matched with the boundary contour to generate a visual correlation map integrating the hot spot diffusion path and the oxidation area distribution, including: 401. Generate a hot spot diffusion path according to the temporal position change path of the abnormal hot spot area. Divide the hot spot diffusion path into equal-length path segments along the axial extension direction of the wire, and extract the migration direction lines of the abnormal hot spot area in the equal-length path segments. The migration direction line is the central extension trajectory of the hot spot diffusion path in the equal-length path segment; In step 401, the temporal position change path is the moving trajectory of the center point of the abnormal hot spot area within a continuous time window, reflecting the spatio-temporal evolution characteristics of heat diffusion. The migration direction line is the extended trajectory line that fits the center point of the hot spot diffusion path within the equal-length path segment, characterizing the local diffusion direction and rate.

[0066] In the embodiments of the present application, first, cubic spline interpolation is performed on the central point coordinates of the temporal position change path of the abnormal hot spot area to generate a smooth diffusion path; second, the diffusion path is divided into equal-length segments (such as every 0.5 meters) along the axial direction of the wire, and the least squares linear fitting is applied to each segment of the path to extract the central extension trajectory line as the migration direction line, and its slope reflects the diffusion rate of this segment (such as 0.3 m / s).

[0067] 402. Divide an equal-width detection grid along the boundary contour, calculate the distribution density of the oxidation area within the equal-width detection grid, and generate an oxidation area density distribution map; In step 402, the equal-width detection grid is a strip-shaped grid unit divided along the oxidation boundary contour line at a fixed width, and is used to quantify the spatial distribution density of the oxidation area. The oxidation area density distribution map is a heat map of the proportion of the oxidation area in each grid unit, where red represents the high-density area and blue represents the low-density area.

[0068] In the embodiments of the present application, first, an equal-width detection band is generated by expanding 1 cm outside the oxidation boundary contour line; second, the detection band is divided into grid units of 1 cm × 1 cm, and the proportion of oxidation area pixels in each unit is calculated (such as 70% of the oxidation pixels in unit A); finally, a continuous oxidation area density distribution map is generated by performing bilinear interpolation on the proportion of oxidation area pixels.

[0069] 403. Perform spatial superposition of the migration direction line and the oxidation area density distribution map, analyze the extension trend of the migration direction line and the dynamic change trend of the oxidation area density distribution in the corresponding grid according to the spatial superposition result, and mark the path segment whose extension trend direction is consistent with the density increase area direction as the action section of heat and oxidation; In step 403, the extension trend direction is the angle change trend of the migration direction line within the path segment. The density increase area direction is the spatial change trend of the oxidation area density along the axial direction of the wire.

[0070] In the embodiments of the present application, first, the migration direction line and the density distribution map are imported into the same coordinate system; second, the angle change trend (such as the linear regression slope) of the migration direction line is calculated for each path segment in this coordinate system, and the density gradient direction of the corresponding grid area is extracted (such as the density increasing from east to west); finally, the path segment with the same direction (such as both showing a positive increase) of the two is marked as the action section of heat and oxidation.

[0071] 404. Calculate the extension angle change rate of the migration direction line, and at the same time extract the density change direction of the oxidation area density distribution map, perform spatial consistency analysis on the extension angle change rate and the density change direction within the action section, and convert the analysis result into a single visualization layer to generate a visualization correlation map.

[0072] In step 404, the extension angle change rate is the amount of angle change per unit length of the migration direction line. The density change direction is the density gradient direction inside the oxidation region.

[0073] In the embodiment of the present application, first, calculate the angle change rate of the migration direction line in the action section (for example, the angle increases from 10° to 15° in the 0 - 0.5 m interval, and the change rate is 10° / m); second, extract the density gradient direction of the corresponding grid area (for example, the density increases along the 30° direction); finally, calculate the direction consistency score of the two through spatial convolution (for example, the cosine similarity > 0.8), render the result as an arrow layer (red indicates high consistency, yellow indicates low consistency), and overlay to generate a visual association map.

[0074] The following is a specific example: Due to the peeling of the oxide layer at the conductor connection point of a 500 kV transmission line, local overheating occurs. It is necessary to analyze the spatio - temporal correlation between the heat diffusion path and the oxidation deterioration and generate a visual decision - making support map. In step 401, the trajectory of the center point of the abnormal hot spot shows that the heat diffuses from the connection point to both sides of the conductor. After being divided into 6 path segments of 0.5 m, it is found by fitting that the angle of the migration direction line of the left path segment (0 - 1.5 m) increases from 5° to 20°, and the diffusion rate increases from 0.2 m / s to 0.5 m / s; in step 402, equal - width detection grids are divided along the oxidation boundary, and the generated density distribution map shows that the oxidation density on the west side of the connection point reaches 80%, while on the east side it is only 30%; in step 403, analyze the trend of the migration direction line (the angle increases on the left) and the density gradient direction (decreases from west to east), and mark the left 0 - 1.5 m as the heat and oxidation action section; in step 404, calculate the angle change rate of this section as 12° / m and the density gradient direction as 150°, and the spatial consistency score is 0.85 (> 0.8 threshold). In the visual map, the high - consistency area is marked with a red arrow. After overlaying the hot spot path and the oxidation density map, it shows that the high - heat - density area on the west side completely coincides with the fast - diffusion path. It is determined as an oxidation - dominated defect, and a report is generated recommending replacing the west - side conductor first and applying an anti - corrosion coating.

[0075] Steps 401 - 404 realize the visual expression of the dynamic association between the overheating defect of the conductor connection point and the oxidation deterioration through the refined segmented analysis of the hot spot diffusion path and the spatial coupling determination of the oxidation density gradient direction; the collaborative analysis of the migration direction line and the density gradient breaks through the traditional static threshold limit and accurately identifies the high - risk area of heat diffusion driven by oxidation; the visual map intuitively presents the defect evolution trend and the operation and maintenance priority level, providing full - chain decision - making support for the transmission line from defect traceability, risk grading to precise disposal, and significantly improving the operation and maintenance efficiency and reliability in complex defect scenarios.

[0076] In order to solve the problem of insufficient determination accuracy of the dynamic correlation between oxidation deterioration and heat diffusion in the overheating defect of wire connection points and further improve the reliability of defect root cause tracing and early warning, in some embodiments, the linkage determination rule of the angle change amount of the hot spot diffusion path and the jump times of the spectral absorption attenuation rate is set as the dynamic correlation rule according to the coverage coincidence degree between the spatial position of the abnormal diffusion section and the spatial density distribution of the active deterioration units, including: 501. Divide the wire surface area corresponding to the abnormal diffusion section into equal-area analysis units, calculate the coverage coincidence degree of the number of active deterioration units in the equal-area analysis unit and the total number of grids. When the coverage coincidence degree exceeds the preset coincidence ratio, mark the equal-area analysis unit as a thermal oxidation correlation unit; In step 501, the equal-area analysis unit is a fixed-area analysis grid divided within the wire surface area corresponding to the abnormal diffusion section, and is used to quantify the spatial correlation between heat and oxidation.

[0077] In the embodiment of the present application, first, the wire surface area covered by the abnormal diffusion section is divided into rectangular analysis units of 1m×0.1m; second, the number of active deterioration units in each unit is counted (for example, unit A has 8 active units), and the ratio of it to the total number of grids in the unit (for example, 20) is calculated (8 / 20 = 40%); finally, if the ratio exceeds the preset ratio (for example, 30%), then mark the unit as a thermal oxidation correlation unit.

[0078] 502. In the abnormal diffusion section corresponding to the thermal oxidation correlation unit, extract the angle increment sequence of adjacent detection intervals, calculate the correlation fluctuation amplitude between the coverage coincidence degree and the angle increment sequence. When the correlation fluctuation amplitude exceeds the fluctuation upper limit, mark the detection section as a significant diffusion section; In step 502, the correlation fluctuation amplitude is the covariance value between the coverage coincidence degree and the angle increment sequence, and reflects the consistency degree of their dynamic changes. The significant diffusion section is a continuous detection interval where the correlation fluctuation amplitude exceeds the preset upper limit, and represents the active area of the heat-oxidation synergistic effect.

[0079] In the embodiment of the present application, first, extract the angle increment sequence of all detection intervals in the abnormal diffusion section corresponding to the thermal oxidation correlation unit (for example, [3°, 5°, 7°]); second, calculate the covariance of this sequence and the coverage coincidence degree sequence (for example, [30%, 40%, 50%]) (for example, covariance>0.8); finally, if the covariance exceeds the fluctuation upper limit (for example, 0.7), then mark this section as a significant diffusion section.

[0080] 503. In the significant diffusion section, calculate the product of the coverage coincidence degree and the sum of the angle increments, and obtain the thermal oxidation correlation index after normalizing the product; In step 503, the thermal oxidation correlation index is the normalized product value of the sum of the coverage overlap degree and the included angle increment, which is used to quantify the coupling strength of heat and oxidation.

[0081] In the embodiments of the present application, first, calculate the product of the average coverage overlap degree (such as 45%) and the sum of the included angle increments (such as 15°) in the significant diffusion section (45%×15 = 6.75); secondly, normalize the product value to the range of 0-1 (6.75 / 10 = 0.675) through maximum-minimum normalization (such as the global maximum product value of 10 and the minimum of 0) to generate the thermal oxidation correlation index.

[0082] 504. Establish a linkage determination rule between the change amount of the included angle of the hot spot diffusion path and the number of jumps in the spectral absorption attenuation rate according to the thermal oxidation correlation index. When the thermal oxidation correlation index exceeds the preset standard value, it is determined that there is a wire overheating defect caused by the deterioration of the oxide layer.

[0083] In step 504, the linkage determination rule is a defect determination condition set based on the thermal oxidation correlation index, and a defect warning of oxide layer deterioration is triggered when the index exceeds the limit.

[0084] In the embodiments of the present application, set the linkage determination rule: set the standard value according to historical data (such as 0.6). If the thermal oxidation correlation index exceeds this value, it is determined that there is an overheating defect caused by the oxide layer; secondly, generate a defect level label (such as 0.6-0.8 is "medium risk", >0.8 is "high risk"), and associate it with the operation and maintenance disposal suggestion library (such as high risk requires power outage and maintenance within 48 hours).

[0085] The following is a specific example: Suppose that the local temperature rise of the wire connection point of a 220 kV transmission line is abnormally high due to oxidation corrosion, and it is necessary to accurately determine the defect level and formulate a repair strategy. In step 501, divide the abnormal diffusion area 1.2 m west of the connection point into 12 analysis units (1 m×0.1 m). It is calculated that the coverage overlap degree of the 3rd to 6th units reaches 35%-50% (the preset threshold is 30%), and it is marked as a thermal oxidation correlation unit; in step 502, extract the included angle increment sequence [4°, 6°, 9°] and the coverage overlap degree sequence [35%, 42%, 50%] in this section, and calculate the covariance to be 0.85 (>0.7 threshold), and mark it as a significant diffusion section; in step 503, calculate the product of the average coverage overlap degree of 45% and the sum of the included angle increments of 19° to be 8.55, and the thermal oxidation correlation index after normalization is 0.855; in step 504, it is determined to be a "high-risk defect" according to the index of 0.855 (>0.8), trigger a level-three warning and generate a report, requiring immediate power outage and replacement of the connection point and 1.5 m wires on both sides, and synchronously push the high-definition map of the oxidation area and the thermal diffusion prediction model to the mobile terminal.

[0086] Steps 501 - 504 achieve precise quantitative correlation between overheating defects at wire connection points and deterioration of oxide layers through dynamic collaborative analysis of coverage overlap and angle increment, breaking through the limitations of traditional single - parameter threshold determination; the standardized design of the thermal - oxidation correlation index supports defect - level comparability across scenarios and reduces the risk of misjudgment caused by environmental interference; the deep binding of the linkage determination rule and the operation and maintenance disposal suggestions provides full - process closed - loop support for transmission lines from defect detection, root - cause analysis to emergency disposal, significantly improving the operation and maintenance timeliness and reliability under complex working conditions.

[0087] To solve the problems of high misjudgment rate of heat diffusion direction and difficulty in distinguishing normal temperature rise from abnormal temperature rise caused by oxidation deterioration in traditional wire overheating defect detection, and to further improve the accuracy of early warning, in some embodiments, within the temperature detection band, calculate the temperature difference between a temperature sampling point and its adjacent sampling points, generate a temperature gradient direction according to the temperature difference along the arrangement direction of the temperature sampling points, and statistically analyze the direction distribution histogram of the temperature gradient direction, and screen out the temperature gradient directions in the direction distribution histogram whose included angle with the wire axis is less than a preset angle, including: 601. Within the temperature detection band, calculate the temperature differences between a temperature sampling point and its left and right adjacent sampling points, and generate a temperature gradient direction pointing from the low - temperature to the high - temperature according to the positive and negative signs and absolute values of the temperature differences; In step 601, the temperature gradient direction is a vector direction pointing from the low - temperature to the high - temperature, generated by the sign and magnitude of the temperature difference, representing the heat diffusion direction.

[0088] In the embodiments of the present application, first, sampling points are arranged along the circumferential direction of the wire within the temperature detection band (e.g., at an interval of 1 cm), and the temperature differences between each sampling point and its left and right adjacent points are calculated (e.g., left difference +2 °C, right difference - 3 °C); secondly, the temperature gradient direction is generated according to the difference sign (if the left difference > the right difference, the direction is to the left, otherwise to the right) and the absolute value magnitude, and the vector length of the direction is proportional to the absolute value of the temperature difference (e.g., 3 °C corresponds to a length of 3 units).

[0089] 602. Taking the wire axis as the reference direction, statistically analyze the direction - angle distribution of the temperature gradient direction, and divide the direction - angle distribution into a direction distribution histogram according to a preset angle interval; In step 602, the direction - angle distribution is the statistical distribution of all temperature gradient direction angles, with the wire axis as the 0 ° reference. The direction distribution histogram is a distribution graph of the number of gradient directions divided according to a preset angle interval.

[0090] In the embodiments of the present application, first, taking the axial direction of the wire as the 0° reference, the direction angles of all temperature gradient vectors are converted to the range of 0° - 180° (for example, deflection to the left is 0° - 90°, and to the right is 90° - 180°); second, the number of vectors in each interval is counted at 5° intervals (for example, there are 15 vectors in the 0° - 5° interval) to generate a direction distribution histogram; finally, noise interference is eliminated through Gaussian smoothing filtering to enhance the distribution continuity.

[0091] 603. Screen out the temperature gradient directions with an angle less than a preset angle with the axial direction of the wire from the direction distribution histogram, where the preset angle is set according to the maximum allowable thermal diffusion offset when the wire is operating normally; In step 603, the preset angle is the maximum allowable direction offset angle set according to the thermal diffusion characteristics of the wire material.

[0092] In the embodiments of the present application, first, the temperature gradient directions in the intervals of 0° - 10° and 170° - 180° (corresponding to an axial offset <10°) are extracted from the direction distribution histogram; second, adjacent temperature gradient directions are merged through connected component analysis, and isolated noise points (such as regions with an area <5 cm²) are removed; finally, a set of filtered axial gradient directions is generated.

[0093] 604. Quantitatively analyze the spatial distribution and direction angle deviation of the filtered temperature gradient directions, where the quantitative analysis includes the spacing statistics of adjacent vectors and the cumulative distribution calculation of the direction angle deviation.

[0094] In step 604, the direction angle deviation is the difference between the direction of the filtered vector and the regional average direction, which is used to quantify the direction consistency.

[0095] In the embodiments of the present application, first, the Euclidean distance between adjacent vectors is calculated (for example, the distance between vector A and B is 2 cm), and the spacing distribution is counted (for example, 80% of the spacings <3 cm); second, the deviation between the direction of each vector and the regional average direction is calculated (for example, the average direction is 5°, and the deviation of a certain vector is 2°) to generate a cumulative distribution curve (for example, 90% of the vectors have a deviation <5°); finally, the K-S test is used to evaluate whether the distribution conforms to the expected thermal diffusion pattern.

[0096] The following is a specific example: Suppose the local temperature of a wire connection point on a 330 kV transmission line is abnormally high due to oxidation corrosion, and it is necessary to determine whether it is an overheating defect driven by oxidation. In step 601, the temperature difference of the sampling points is calculated within the temperature detection zone on the west side of the connection point, and it is found that the left difference in the central area is continuously greater than the right difference (the difference is +3°C to +5°C), generating a gradient vector cluster pointing to the left; in step 602, the direction angle distribution is statistically analyzed, and the histogram shows that 80% of the vectors are concentrated in the 0°-8° interval (axial offset <8°); in step 603, the vectors in this interval are screened and merged into a continuous area (area 0.25 m²), excluding the scattered vectors on both sides; in step 604, the analysis shows that the average distance between adjacent vectors is 2.5 cm (95% <3 cm), and 90% of the vector deviations in the cumulative distribution of direction deviations are <4°, which conforms to the axial heat diffusion mode. Combining with the multispectral data, it is confirmed that the coincidence degree of the oxidation boundary in this area reaches 85%, and it is determined as an oxidation-driven defect, and a report is generated recommending replacing the oxidized section of the wire within 72 hours.

[0097] Steps 601-604 accurately identify the abnormal heat diffusion caused by the oxidation deterioration of the wire connection point through axial gradient direction screening and spatial continuity analysis, and effectively distinguish the transient temperature rise caused by environmental interference; the combination of the direction distribution histogram and the cumulative deviation distribution enhances the interpretability of the heat diffusion mode and reduces the misjudgment rate; the quantitative analysis results provide a basis for the heat and oxidation correlation strength for the operation and maintenance decision-making, supporting the intelligent upgrade of the transmission line from defect detection to root cause tracing.

[0098] In order to solve the problems of strong subjectivity in setting the threshold value and insufficient tracking of the defect evolution trend in the determination of the overheating defect level of the wire connection point, and further improve the scientificity of defect classification and operation and maintenance decision-making, in some embodiments, through the dynamic association rule, according to the critical angle threshold of the included angle change amount and the number threshold of the mutation points, the local overheating defect level caused by the deterioration of the oxide layer on the wire surface is determined, and a corresponding defect analysis report is generated according to the local overheating defect level, including: 701. Through the dynamic association rule, set the range of the critical angle threshold of the included angle change amount, and the range is determined according to the physical upper limit value of the abnormal rate of axial heat diffusion of the wire, and at the same time set the interval of the number threshold of the mutation points, and the interval is determined according to the spectral response characteristics of the activity of the oxide layer deterioration; In step 701, the range of the critical angle threshold is the allowable interval of the included angle change amount set according to the physical limit of heat diffusion of the wire material. The interval of the number threshold is the range of the number of mutations divided based on the activity of the spectral response of the oxide layer.

[0099] In the embodiments of the present application, first, the critical angle threshold range is determined through the experimental data of the heat conduction of the wire material (for example, the upper threshold corresponding to the maximum diffusion rate of the aluminum wire under normal heat dissipation is 8° / m); second, based on the analysis of the multi-spectral historical data, the relationship between the deterioration rate of the oxide layer and the number of spectral mutation points is analyzed, and the mutation point threshold interval is set (for example, ≥4 times in each 10 cm×10 cm grid is determined as a high activity level).

[0100] 702. Establish a combined relationship table between the above range and the above interval, and define the matching intervals of the critical angle threshold and the quantity threshold corresponding to different levels of defects in the combined relationship table; In step 702, the combined relationship table is a matching rule table that defines the critical angle threshold and the mutation point quantity threshold under different defect levels.

[0101] In the embodiments of the present application, first, a two-dimensional matrix is established, with the horizontal axis being the critical angle threshold interval and the vertical axis being the mutation point quantity interval; second, based on the historical defect samples, the defect level corresponding to each matrix unit is labeled (for example, an angle of 4°-6° / m and 4-6 mutation points is a secondary defect); finally, the matrix weights are optimized through expert experience and machine learning to generate a combined relationship table.

[0102] 703. According to the above range and the above interval, screen the sections in the wire surface detection area where the change amount of the included angle and the number of mutation points respectively fall into the corresponding threshold intervals as the sections to be determined, and the initial defect level is mapped according to the combined relationship table for the sections to be determined; In step 703, the section to be determined is a continuous section in the wire surface detection area that simultaneously satisfies the critical angle and the mutation point threshold.

[0103] In the embodiments of the present application, first, the detection sections are divided along the wire surface (for example, every 1 meter), and the average value of the included angle change amount and the number of mutation points in each section are extracted; second, the combined relationship table is traversed to match the sections that meet the threshold intervals (for example, the average angle of a certain section is 5° / m and there are 5 mutation points, which matches the secondary defect); finally, the initial defect level label is output.

[0104] 704. Continuously correct the initial defect level. When the defect levels of three consecutive sections are the same and the increasing trend of the included angle change amount is the same, merge the three sections and upgrade one defect level to output the local overheating defect level, and generate a defect analysis report including the defect position distribution, level intensity, and diffusion trend according to the local overheating defect level.

[0105] In step 704, the continuous correction: the enhanced determination rule for the consistency of the defect levels of adjacent sections and the diffusion trend.

[0106] In the embodiments of the present application, first, it is detected whether three consecutive sections have the same defect level (such as all being level two); second, it is analyzed whether the change amount of their included angle shows a monotonically increasing trend (such as 4° / m → 5° / m → 6° / m); finally, the sections are merged and upgraded by one level (such as from level two to level three), and a defect analysis report including the position, level, and diffusion trend is generated.

[0107] The following is a specific example: Suppose that due to the peeling of the oxide layer, local overheating occurs at the conductor connection point of a 500 kV transmission line, and it is necessary to dynamically determine the defect level and formulate a prevention and control strategy. In step 701, the critical angle threshold range of the aluminum conductor is set to 3° - 8° / m, and the number range of mutation points is low activity (1 - 3 times), medium activity (4 - 6 times), and high activity (≥7 times); in step 702, a combined relationship table is established, and it is defined that the angle of 4° - 6° / m and 4 - 6 mutation points are matched as "medium risk", and the angle ≥6° / m and mutation points ≥7 times are "high risk"; in step 703, it is detected that the average angle of the section 1 - 3 meters west of the connection point is 5.2° / m and the number of mutation points is 5, which is marked as "medium risk", and the angle of the section 4 - 6 meters east of the connection point is 7.1° / m and the number of mutation points is 8, which is marked as "high risk"; in step 704, the analysis shows that the defect levels of three consecutive sections on the west side are all "medium risk" and the included angle change amount increases (4.8 → 5.2 → 5.5° / m), and they are merged and upgraded to "high risk". The generated defect analysis report marks the west side as the diffusion core area, and it is recommended to replace and coat the anti-corrosion layer first, and the east side is listed as the emergency disposal area.

[0108] Steps 701 - 704 achieve an objective quantitative determination of the defect level of the conductor connection point through the threshold setting driven by physical characteristics and spectral response data; the combination of the combined relationship table and the continuity correction rule enhances the dynamic tracking ability of the defect evolution trend and avoids misjudgment of discrete sections; the multi-dimensional information integration of the defect report provides full-chain support for the transmission line from risk warning, level division to prevention and control decision-making, and significantly improves the operation and maintenance accuracy and timeliness in complex defect scenarios.

[0109] Figure 2 The structure diagram of an automated intelligent defect analysis system based on UAV image acquisition for transmission and distribution lines is provided for the embodiments of the present application, as Figure 2 shown, the system includes: An acquisition module 21, configured to synchronously acquire the infrared thermal imaging time-series data and multi-spectral imaging data of the conductor surface within a continuous time window through an infrared thermal imager and a multi-spectral camera carried on the UAV. The infrared thermal imaging time-series data covers the entire length of the conductor axis through the UAV flight trajectory to capture the dynamic migration trajectory of the surface temperature field, and the multi-spectral imaging data obtains the spectral absorption difference between the oxide layer and the normal metal on the conductor surface at a selected band through the UAV hovering attitude; The positioning module 22 is configured to calculate the temperature gradient direction of the wire body region in each frame of the infrared thermal imaging timing data, extract the abnormal hot spot region where the temperature gradient direction is consistent with the wire axis, and at the same time, based on the spectral absorption attenuation curve corresponding to the oxide layer in the multispectral imaging data, locate the boundary contour of the oxidized region on the wire surface; The matching module 23 is configured to perform spatial topological matching between the migration direction of the abnormal hot spot region and the boundary contour, and generate a visual association map that combines the hot spot diffusion path and the oxidized region distribution; The establishing module 24 is configured to establish a dynamic association rule between the wire overheating defect and the state of the surface oxide layer according to the change amount of the angle between the hot spot diffusion path in the visual association map and the wire axis, and the mutation point of the spectral absorption attenuation rate within the boundary contour; The discrimination module 25 is configured to determine the local overheating defect level caused by the deterioration of the oxide layer on the wire surface through the dynamic association rule, according to the critical angle threshold of the change amount of the angle and the number threshold of the mutation points.

[0110] Figure 2 The described automatic intelligent defect analysis system based on UAV image acquisition for power transmission and distribution lines can execute Figure 1 The described automatic intelligent defect analysis method based on UAV image acquisition for power transmission and distribution lines in the illustrated embodiment, its implementation principle and technical effects will not be elaborated. For the automatic intelligent defect analysis system based on UAV image acquisition for power transmission and distribution lines in the above embodiment, the specific ways for each module and unit to perform operations have been described in detail in the embodiment related to the method, and will not be elaborated here.

[0111] In a possible design, Figure 2 The automatic intelligent defect analysis system based on UAV image acquisition for power transmission and distribution lines in the illustrated embodiment can be implemented as a computing device, such as Figure 3 shown, the computing device may include a storage component 31 and a processing component 32; The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32.

[0112] The processing component 32 is used for the Figure 1 automatic intelligent defect analysis method based on UAV image acquisition for power transmission and distribution lines in the above

[0113] Among them, the processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above methods. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components for executing the above methods.

[0114] The storage component 31 is configured to store various types of data to support the operation of the terminal. The storage component may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0115] Of course, the computing device may also necessarily include other components, such as input / output interfaces, display components, communication components, etc.

[0116] The input / output interface provides an interface between the processing component and the peripheral interface module, and the above peripheral interface module may be an output device, an input device, etc.

[0117] The communication component is configured to facilitate communication between the computing device and other devices in a wired or wireless manner, etc.

[0118] Among them, the computing device may be a physical device or an elastic computing host provided by a cloud computing platform, etc. At this time, the computing device may refer to a cloud server, and the above processing component, storage component, etc. may be basic server resources leased or purchased from a cloud computing platform.

[0119] The embodiment of the present application also provides a computer storage medium storing a computer program, and when the computer program is executed by a computer, it can implement the above Figure 1 shown embodiment of an automated intelligent defect analysis method based on UAV image acquisition of power transmission and distribution lines.

[0120] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0121] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative work.

[0122] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0123] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of each embodiment of the present application.

Claims

1. An automated intelligent defect analysis method for UAV image acquisition of power transmission and distribution lines, characterized in that, Including: Collecting infrared thermal imaging time-series data and multispectral imaging data of the surface of the conductor of the power transmission and distribution line by means of a drone within a continuous time window; Calculating the temperature gradient direction of the main body area of the conductor in each frame of the infrared thermal imaging time-series data, extracting the abnormal hot spot area consistent with the axial direction of the conductor from the temperature gradient direction, and at the same time, based on the spectral absorption attenuation curve corresponding to the surface oxide layer in the multispectral imaging data, positioning the boundary contour of the surface oxidation area of the conductor; Performing spatial topological matching between the migration direction of the abnormal hot spot area and the boundary contour to generate a visual correlation map integrating the hot spot diffusion path and the oxidation area distribution; According to the change amount of the angle between the hot spot diffusion path and the axial direction of the conductor in the visual correlation map, and the mutation points of the spectral absorption attenuation rate within the boundary contour, establishing a dynamic correlation rule between the overheating defect of the conductor and the state of the surface oxide layer; Based on the dynamic correlation rule, according to the critical angle threshold of the change amount of the angle and the number threshold of the mutation points, determining the local overheating defect level caused by the deterioration of the oxide layer on the surface of the conductor, and generating a corresponding defect analysis report according to the local overheating defect level.

2. The method according to claim 1, wherein The establishing of the dynamic correlation rule between the overheating defect of the conductor and the state of the surface oxide layer according to the change amount of the angle between the hot spot diffusion path and the axial direction of the conductor in the visual correlation map, and the mutation points of the spectral absorption attenuation rate within the boundary contour includes: In the visual correlation map, along the extension direction of the hot spot diffusion path, dividing continuous detection intervals with the axial direction of the conductor as the reference axis, quantifying the change amount of the angle between the hot spot diffusion path and the axial direction of the conductor within the continuous detection intervals, and generating an angle change sequence; Within the coverage area of the boundary contour, dividing detection units with equal areas, and counting the number of jumps of the spectral absorption attenuation rate from a stable state to a mutation within the detection units to generate a jump number distribution map of each detection unit; Marking the section where the angle increment between adjacent detection intervals in the angle change sequence exceeds the preset increase amplitude as an abnormal diffusion section, and at the same time, marking the detection units with the number of jumps exceeding the preset number in the jump number distribution map as active deterioration units; According to the coverage coincidence degree between the spatial position of the abnormal diffusion section and the spatial density distribution of the active deterioration units, setting the linkage determination rule of the change amount of the angle of the hot spot diffusion path and the number of jumps of the spectral absorption attenuation rate as the dynamic correlation rule.

3. The method according to claim 1, characterized in that The calculating of the temperature gradient direction of the main body area of the conductor in each frame of the infrared thermal imaging time-series data, extracting the abnormal hot spot area consistent with the axial direction of the conductor from the temperature gradient direction, and at the same time, based on the spectral absorption attenuation curve corresponding to the surface oxide layer in the multispectral imaging data, positioning the boundary contour of the surface oxidation area of the conductor includes: In the main body area of the conductor in each frame of the infrared thermal imaging time-series data, dividing equally spaced temperature detection bands along the axial extension direction of the conductor, and arranging temperature sampling points in the temperature detection bands along the direction perpendicular to the axial direction of the conductor; Within the temperature detection band, calculate the temperature difference between a temperature sampling point and its adjacent sampling point, generate a temperature gradient direction along the arrangement direction of the temperature sampling points according to the temperature difference, statistically analyze the direction distribution histogram of the temperature gradient direction, and filter out the temperature gradient directions in the direction distribution histogram whose included angle with the axial direction of the wire is less than a preset angle; In the wire main body area, perform spatial clustering on the filtered temperature gradient directions, merge the areas that show continuous distribution and whose temperature gradient direction consistency exceeds the lowest ratio in the spatial clustering results, and mark them as abnormal hot spot areas; In the multispectral imaging data, divide spectral sampling bands along the wire surface contour line, extract the slope change points of the spectral absorption attenuation curve corresponding to the wire surface oxide layer within the spectral sampling bands, and connect the sampling band boundaries where the curve segment length between adjacent slope change points is less than a preset distance to form the boundary contour of the closed wire surface oxidation area.

4. The method according to claim 1, wherein The spatial topological matching of the migration direction of the abnormal hot spot area with the boundary contour to generate a visual association map integrating the hot spot diffusion path and the oxidation area distribution includes: Generate a hot spot diffusion path according to the temporal position change path of the abnormal hot spot area, divide the hot spot diffusion path into equal-length path segments along the axial extension direction of the wire, and extract the migration direction line of the abnormal hot spot area within the equal-length path segments. The migration direction line is the central extension trajectory of the hot spot diffusion path within the equal-length path segments; Divide equal-width detection grids along the boundary contour, calculate the oxidation area distribution density within the equal-width detection grids, and generate an oxidation area density distribution map; Perform spatial superposition of the migration direction line and the oxidation area density distribution map, analyze the extension trend of the migration direction line and the dynamic change trend of the oxidation area density distribution within the corresponding grid according to the spatial superposition result, and mark the path segments whose extension trend direction is consistent with the direction of the density increase area as the action sections of heat and oxidation; Calculate the extension angle change rate of the migration direction line, and at the same time extract the density change direction of the oxidation area density distribution map. Perform spatial consistency analysis of the extension angle change rate and the density change direction within the action section, and convert the analysis result into a single visual layer to generate a visual association map.

5. The method according to claim 2, characterized in that According to the coverage overlap degree between the spatial position of the abnormal diffusion section and the spatial density distribution of the active deterioration units, set the linkage determination rule of the included angle change amount of the hot spot diffusion path and the jump times of the spectral absorption attenuation rate as the dynamic association rule, including: Divide the wire surface area corresponding to the abnormal diffusion section into equal-area analysis units, calculate the coverage overlap degree of the number of active deterioration units and the total number of grids within the equal-area analysis units. When the coverage overlap degree exceeds the preset overlap ratio, mark the equal-area analysis unit as a heat oxidation association unit; In the abnormal diffusion section corresponding to the thermal oxidation correlation unit, extract the included angle increment sequence of adjacent detection intervals, and calculate the correlation fluctuation amplitude between the coverage overlap degree and the included angle increment sequence. When the correlation fluctuation amplitude exceeds the upper limit of fluctuation, mark the detection section as a significant diffusion section; In the significant diffusion section, calculate the product of the coverage overlap degree and the total included angle increment, and obtain the thermal oxidation correlation index after normalizing the product; According to the thermal oxidation correlation index, establish a linkage determination rule between the included angle change amount of the hot spot diffusion path and the number of jumps in the spectral absorption attenuation rate. When the thermal oxidation correlation index exceeds the preset standard value, it is determined that there is a wire overheating defect caused by the deterioration of the oxide layer.

6. The method according to claim 3, wherein In the temperature detection band, calculate the temperature difference between a temperature sampling point and its adjacent sampling point, generate a temperature gradient direction along the arrangement direction of the temperature sampling points according to the temperature difference, and statistically analyze the direction distribution histogram of the temperature gradient direction. Screening the temperature gradient directions in the direction distribution histogram whose included angle with the wire axis is less than a preset angle includes: In the temperature detection band, calculate the temperature differences between a temperature sampling point and its left and right adjacent sampling points, and generate a temperature gradient direction pointing from low temperature to high temperature according to the positive and negative signs and absolute value magnitudes of the temperature differences; Taking the wire axis as the reference direction, statistically analyze the direction angle distribution of the temperature gradient direction, and divide the direction angle distribution into a direction distribution histogram according to a preset angle interval; Screen out the temperature gradient directions in the direction distribution histogram whose included angle with the wire axis is less than a preset angle, and the preset angle is set according to the maximum allowable thermal diffusion offset amount during normal operation of the wire; Perform quantitative analysis on the spatial distribution and direction angle deviation of the screened temperature gradient directions, and the quantitative analysis includes the spacing statistics of adjacent vectors and the cumulative distribution calculation of the direction angle deviation.

7. The method according to claim 1, wherein Based on the dynamic correlation rule, according to the critical angle threshold of the included angle change amount and the number threshold of mutation points, determine the local overheating defect level on the wire surface caused by the deterioration of the oxide layer, and generate a corresponding defect analysis report according to the local overheating defect level, including: Through the dynamic correlation rule, set the range of the critical angle threshold of the included angle change amount, and the range is determined according to the physical upper limit value of the abnormal thermal diffusion rate along the wire axis. At the same time, set the interval of the number threshold of mutation points, and the interval is determined according to the spectral response characteristics of the oxide layer deterioration activity; Establish a combination relationship table between the range and the interval, and define the matching intervals of the critical angle threshold and the number threshold corresponding to different levels of defects in the combination relationship table; According to the range and the interval, screen out the sections in the wire surface detection area where the included angle change amount and the number of mutation points respectively fall into the corresponding threshold intervals as the sections to be determined, and the sections to be determined are mapped to the initial defect level according to the combination relationship table; Perform continuous correction on the initial defect level. When the defect levels of three consecutive sections are the same and the increasing trend of the included angle change amount is consistent, merge the three sections and upgrade the defect level by one to output the local overheating defect level, and generate a defect analysis report including the defect position distribution, level intensity, and diffusion trend based on the local overheating defect level.

8. An automated intelligent defect analysis system for UAV image acquisition of power transmission and distribution lines, characterized in that, Including: A collection module for synchronously collecting infrared thermal imaging time-series data and multi-spectral imaging data on the surface of the wire within a continuous time window through an infrared thermal imager and a multi-spectral camera mounted on a drone. The infrared thermal imaging time-series data covers the entire axial length of the wire through the flight trajectory of the drone to capture the dynamic migration trajectory of the surface temperature field, and the multi-spectral imaging data obtains the spectral absorption difference between the oxide layer and the normal metal on the wire surface at a selected band through the hovering attitude of the drone. A positioning module for calculating the temperature gradient direction of the main wire area in each frame of the infrared thermal imaging time-series data, extracting the abnormal hot spot area where the temperature gradient direction is consistent with the wire axis, and at the same time, based on the spectral absorption attenuation curve corresponding to the oxide layer in the multi-spectral imaging data, positioning the boundary contour of the oxide area on the wire surface. A matching module for performing spatial topological matching between the migration direction of the abnormal hot spot area and the boundary contour to generate a visual association map integrating the hot spot diffusion path and the oxide area distribution. An establishment module for establishing a dynamic association rule between the wire overheating defect and the state of the surface oxide layer according to the included angle change amount between the hot spot diffusion path and the wire axis in the visual association map, and the mutation point of the spectral absorption attenuation rate within the boundary contour. A discrimination module for determining the local overheating defect level caused by the deterioration of the oxide layer on the wire surface through the dynamic association rule according to the critical angle threshold of the included angle change amount and the number threshold of the mutation points.

9. A computing device, characterized in that, Including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a method for automatic intelligent defect analysis based on UAV image acquisition of power transmission and distribution lines as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that, Stores a computer program, and when the computer program is executed by a computer, it implements a method for automatic intelligent defect analysis based on UAV image acquisition of power transmission and distribution lines as described in any one of claims 1 to 7.

Citation Information

Patent Citations

  • Image heating apparatus and image forming apparatus

    CN107664943A

  • Regional heat supply pipe network optimization design method and device

    CN118839457A

  • Defect detection method, device and equipment for high altitude fall prevention system

    CN119780639A

  • Power transmission line galloping early warning method, system, equipment and medium

    CN119962974A

  • Optical-electrical web inspection system

    US3843890A

Cited By

  • Power transmission and distribution line iron accessory temperature abnormity control method and system

    CN120747152A

  • Automatic identification method for subfissure defect of assembly

    CN120876463A

  • Ink printing defect detection method based on image processing

    CN121414680A

  • Power line inspection method

    CN122044190A

  • Method, device and equipment for detecting quality of conductor of power transmission line

    CN122545532A