An automated intelligent defect analysis method and system for UAV image acquisition of power transmission and distribution lines

Abstract

The present application provides an automated intelligent defect analysis method and system based on UAV image acquisition of power transmission and distribution lines. Among them, infrared thermal imaging time series data and multispectral imaging data of the conductor surface are collected by using a UAV equipped with an infrared thermal imager and a multispectral camera; the temperature gradient direction of the conductor body area is calculated, and the abnormal hot spot area with an angle less than a preset angle with the conductor axis is screened, and the spectral absorption attenuation curve of the oxide layer in the multispectral data is used to locate the oxidation area boundary; the hot spot migration direction is spatially matched with the oxidation boundary to generate a visual map; dynamic association rules are established based on the angle change of the hot spot path in the map and the spectral mutation points in the oxidation area; the defect level is determined based on the critical angle threshold and the mutation point number threshold, and a defect analysis report is generated. The technical solution provided by the present application improves the automation level and accuracy of defect detection.

Description

Technical Field

[0001] The present application relates to the field of defect analysis technology, and in particular to an automated intelligent defect analysis method and system based on drone image acquisition of power transmission and distribution lines. Background Art

[0002] Long-term exposure to complex environments (such as high loads, oxidative corrosion, and poor contact) can lead to localized overheating defects at the connection points of power transmission and distribution lines, which can severely cause disconnections or fires. Traditional manual inspections cannot meet real-time requirements, necessitating an automated approach for early warning and dynamic tracking of overheating defects at connection points. This approach must meet the following core requirements: accurately capture subtle thermal anomalies at conductor connection points; support all-weather monitoring under complex lighting and weather conditions; and enable quantitative analysis and risk stratification of defect evolution trends.

[0003] The current mainstream solution is a dual-modal drone inspection system based on visible light cameras and thermal infrared imaging. This system uses a visible light camera mounted on a drone to capture high-definition images of wire connection points, combined with a thermal infrared imager to collect surface temperature data. It then uses image matching technology to locate connection points and a threshold segmentation algorithm to identify areas of abnormal temperature. The system generates preliminary warning signals by comparing historical temperature data, and then combines manual verification to confirm defect levels.

[0004] The defects of existing solutions include: visible light images are severely affected by light, rain and fog, and the location of connection points cannot be effectively identified at night or in bad weather, resulting in inaccurate positioning of thermal infrared data; they rely on fixed temperature threshold segmentation, making it difficult to distinguish between normal temperature rise and abnormal overheating of connection points (such as the false alarm rate in high temperature environments in summer exceeds 40%), and are unable to capture early subtle thermal changes; they only rely on temperature comparisons at discrete time points, lack modeling and analysis of dynamic characteristics such as the diffusion path and rate of thermal anomalies, and are difficult to provide trend prediction basis for operation and maintenance decisions. Summary of the Invention

[0005] The present application provides an automated intelligent defect analysis method and system based on drone image acquisition of power transmission and distribution lines, which is used to solve the problem of insufficient defect detection accuracy in the prior art.

[0006] In a first aspect, the present application provides an automated intelligent defect analysis method based on drone image acquisition of power transmission and distribution lines, comprising:

[0007] Use drones to collect infrared thermal imaging time series data and multispectral imaging data of the conductor surface of power transmission and distribution lines within continuous time windows;

[0008] Calculating the temperature gradient direction of the main area of the wire 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 locating the boundary contour of the wire surface oxidation area based on the spectral absorption attenuation curve corresponding to the wire surface oxide layer in the multispectral imaging data;

[0009] Performing spatial topological matching on 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;

[0010] Establishing a dynamic association rule between the wire overheating defect and the surface oxide layer state based on the change in the angle 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;

[0011] Based on the dynamic association rule, the local overheating defect level caused by oxide layer degradation on the surface of the wire is determined according to the critical angle threshold of the angle change and the number threshold of the mutation points, and a corresponding defect analysis report is generated according to the local overheating defect level.

[0012] Optionally, establishing a dynamic association rule between the wire overheating defect and the surface oxide layer state based on the change in the angle 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, includes:

[0013] In the visual association map, continuous detection intervals are divided along the extension direction of the hot spot diffusion path with the wire axis as the reference axis, and the change in the angle between the hot spot diffusion path and the wire axis within the continuous detection interval is quantified to generate an angle change sequence;

[0014] Divide the coverage area of the boundary contour into detection units of equal area, count the number of jumps of the spectral absorption attenuation rate from a stable state to a sudden change in the detection unit, and generate a jump number distribution graph for each detection unit;

[0015] Marking a section in the angle change sequence where the angle increment between adjacent detection intervals exceeds a preset increment as an abnormal diffusion section, and marking a detection unit in the jump number distribution diagram where the jump number exceeds a preset number as an active degraded unit;

[0016] According to the coverage overlap between the spatial position of the abnormal diffusion section and the spatial density distribution of the active degradation unit, a linkage judgment rule between the angle change of the hot spot diffusion path and the number of jumps of the spectral absorption attenuation rate is set as a dynamic association rule.

[0017] Optionally, calculating the temperature gradient direction of the conductor main body area in each frame of the infrared thermal imaging time series data, extracting the abnormal hot spot area consistent with the conductor axis from the temperature gradient direction, and locating the boundary contour of the conductor surface oxidation area based on the spectral absorption attenuation curve corresponding to the conductor surface oxide layer in the multispectral imaging data, includes:

[0018] In the conductor main body area of each frame of the infrared thermal imaging time series data, temperature detection zones are divided into equal intervals along the axial extension direction of the conductor, and temperature sampling points are arranged in the temperature detection zones along a direction perpendicular to the axial direction of the conductor;

[0019] Within the temperature detection zone, calculating a temperature difference between a temperature sampling point and an adjacent sampling point, generating a temperature gradient direction along an arrangement direction of the temperature sampling points based on the temperature difference, generating a directional distribution histogram of the temperature gradient direction, and screening a temperature gradient direction in the directional distribution histogram whose angle with the conductor axis is less than a preset angle;

[0020] The screened temperature gradient directions are spatially clustered in the main conductor area, and areas showing continuous distribution and having temperature gradient direction consistency exceeding a minimum ratio in the spatial clustering results are merged and marked as abnormal hot spot areas;

[0021] In the multispectral imaging data, spectral sampling bands are divided along the surface contour line of the conductor, and slope change points of the spectral absorption attenuation curve corresponding to the conductor surface oxide layer are extracted within the spectral sampling bands. The slope change points are connected to sampling band boundaries where the length of the curve segment between adjacent slope change points is less than a preset distance, thereby forming a closed boundary contour of the conductor surface oxidation area.

[0022] Optionally, performing spatial topological matching on 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 includes:

[0023] Generating a hot spot diffusion path based on the temporal position change path of the abnormal hot spot area, dividing the hot spot diffusion path into equal-length path segments along the axial extension direction of the wire, and extracting a migration direction line of the abnormal hot spot area within the equal-length path segments, wherein the migration direction line is a central extension trajectory of the hot spot diffusion path within the equal-length path segments;

[0024] Dividing the detection grids into equal widths along the boundary contour, calculating the distribution density of oxidation areas within the detection grids, and generating an oxidation area density distribution map;

[0025] Spatially superimposing the migration direction line and the oxidation area density distribution map, analyzing the extension trend of the migration direction line and the dynamic change trend of the oxidation area density distribution in the corresponding grid based on the spatial superposition result, and marking the path segment where the extension trend direction is consistent with the direction of the density increase area as the heat and oxidation action segment;

[0026] The extension angle change rate of the migration direction line is calculated, and the density change direction of the oxidation area density distribution map is extracted at the same time. The spatial consistency analysis of the extension angle change rate and the density change direction is performed within the action section, and the analysis results are converted into a single visualization layer to generate a visualization correlation map.

[0027] Optionally, the linkage judgment rule between the angle change of the hot spot diffusion path and the number of jumps of the spectral absorption attenuation rate is set as a dynamic association rule based on the coverage overlap between the spatial position of the abnormal diffusion section and the spatial density distribution of the active degradation unit, including:

[0028] Dividing the conductor surface area corresponding to the abnormal diffusion section into equal-area analysis units, calculating the coverage overlap between the number of active degradation units in the equal-area analysis unit and the total number of grids, and marking the equal-area analysis unit as a thermal oxidation-related unit when the coverage overlap exceeds a preset overlap ratio;

[0029] In the abnormal diffusion section corresponding to the thermal oxidation correlation unit, an angle increment sequence of adjacent detection intervals is extracted, and the correlation fluctuation amplitude between the coverage overlap and the angle increment sequence is calculated. When the correlation fluctuation amplitude exceeds the fluctuation upper limit, the detection section is marked as a significant diffusion section;

[0030] In the significant diffusion section, calculating the product of the coverage overlap and the sum of the angle increments, and normalizing the product to obtain a thermal oxidation correlation index;

[0031] Based on the thermal oxidation correlation index, a linkage judgment rule is established between the angle change of the hot spot diffusion path and the number of jumps of the spectral absorption attenuation rate. When the thermal oxidation correlation index exceeds a preset standard value, it is determined that there is a wire overheating defect caused by oxide layer degradation.

[0032] Optionally, within the temperature detection zone, calculating a temperature difference between a temperature sampling point and an adjacent sampling point, generating a temperature gradient direction along an arrangement direction of the temperature sampling points according to the temperature difference, generating a statistically analyzed directional distribution histogram of the temperature gradient direction, and screening a temperature gradient direction in the directional distribution histogram whose angle with the conductor axis is less than a preset angle, includes:

[0033] In the temperature detection zone, the temperature difference between the temperature sampling point and the adjacent sampling points on the left and right is calculated, and the temperature gradient direction from low temperature to high temperature is generated according to the positive and negative signs and absolute values of the temperature difference;

[0034] Taking the axial direction of the conductor as a reference direction, statistically analyzing the direction angle distribution of the temperature gradient direction, and dividing the direction angle distribution according to preset angle intervals to generate a direction distribution histogram;

[0035] Screening out, from the direction distribution histogram, a temperature gradient direction whose angle with the axis of the conductor is less than a preset angle, wherein the preset angle is set according to a maximum allowable thermal diffusion offset during normal operation of the conductor;

[0036] The spatial distribution and directional angle deviation of the screened temperature gradient direction are quantitatively analyzed, and the quantitative analysis includes statistics of intervals between adjacent vectors and cumulative distribution calculation of the directional angle deviation.

[0037] Optionally, determining the level of local overheating defects on the surface of the wire caused by oxide layer degradation by using the dynamic association rule according to a critical angle threshold of the angle variation and a threshold of the number of mutation points, and generating a corresponding defect analysis report according to the level of local overheating defects includes:

[0038] By using the dynamic association rule, a range of a critical angle threshold value of the angle variation is set, wherein the range is determined according to a physical upper limit of an abnormal rate of axial heat diffusion of the conductor, and an interval of a threshold value of the number of mutation points is set, wherein the interval is determined according to a spectral response characteristic of the oxide layer degradation activity;

[0039] Establishing a combination relationship table of the ranges and the intervals, wherein the combination relationship table defines matching intervals of critical angle thresholds and quantity thresholds corresponding to defects of different grades;

[0040] According to the range and the interval, a section in the conductor surface detection area in which the angle variation and the number of mutation points fall within the corresponding threshold intervals is selected as a section to be determined, and the section to be determined is mapped according to the combination relationship table to obtain an initial defect level;

[0041] The initial defect level is continuously corrected. When the defect levels of three consecutive segments are the same and the increasing trend of the angle change is consistent, the three segments are merged and the defect level is increased by one to output the local overheating defect level. A defect analysis report including the defect location distribution, level intensity and diffusion trend is generated based on the local overheating defect level.

[0042] In a second aspect, the present application provides an automated intelligent defect analysis system based on drone image acquisition of power transmission and distribution lines, comprising:

[0043] An acquisition module is used to collect infrared thermal imaging time series data and multispectral imaging data of the conductor surface of the power transmission and distribution line within a continuous time window through a drone;

[0044] a calculation module for calculating the temperature gradient direction of the conductor main body area in each frame of the infrared thermal imaging time series data, extracting the abnormal hot spot area consistent with the conductor axis from the temperature gradient direction, and locating the boundary contour of the conductor surface oxidation area based on the spectral absorption attenuation curve corresponding to the conductor surface oxide layer in the multispectral imaging data;

[0045] A matching module is used to perform spatial topological matching between the migration direction of the abnormal hot spot area and the boundary contour to generate a visual correlation map that integrates the hot spot diffusion path and the oxidation area distribution;

[0046] An establishment module is used to establish a dynamic association rule between the wire overheating defect and the surface oxide layer state based on the change in the angle 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;

[0047] A generation module is used to determine, based on the dynamic association rule, a local overheating defect level caused by oxide layer degradation on the surface of the wire according to a critical angle threshold of the angle change and a threshold of the number of mutation points, and generate a corresponding defect analysis report according to the local overheating defect level.

[0048] In a third aspect, an embodiment of the present application provides a computing device comprising 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 an automated intelligent defect analysis method based on drone image acquisition of power transmission and distribution lines as described in the first aspect above.

[0049] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program. When the computer program is executed by a computer, it implements an automated intelligent defect analysis method based on drone image acquisition of power transmission and distribution lines as described in the first aspect.

[0050] In an embodiment of the present application, infrared thermal imaging time series data and multispectral imaging data of the conductor surface of a power transmission and distribution line within a continuous time window are collected by an unmanned aerial vehicle; the temperature gradient direction of the conductor main body area in each frame of the infrared thermal imaging time series data is calculated, and the abnormal hot spot area consistent with the conductor axis is extracted from the temperature gradient direction. At the same time, based on the spectral absorption attenuation curve corresponding to the conductor surface oxide layer in the multispectral imaging data, the boundary contour of the conductor surface oxide area is located; the migration direction of the abnormal hot spot area is spatially topologically matched with the boundary contour to generate a visual association map that integrates the hot spot diffusion path and the oxidation area distribution; based on the change in the angle between the hot spot diffusion path and the conductor axis in the visual association map, and the mutation point of the spectral absorption attenuation rate within the boundary contour, a dynamic association rule between the conductor overheating defect and the surface oxide layer state is established; based on the dynamic association rule, the local overheating defect level caused by the degradation of the oxide layer on the conductor surface is determined according to the critical angle threshold of the angle change and the number threshold of the mutation points, and a corresponding defect analysis report is generated according to the local overheating defect level.

[0051] The technical solution of this application has the following beneficial effects:

[0052] By synchronously acquiring infrared thermal imaging time-series data and multispectral imaging data, the multi-dimensional information fusion of the dynamic changes in the conductor surface temperature field and the spectral characteristics of the oxide layer is realized, thereby improving the comprehensiveness of defect detection; based on the temperature gradient direction, abnormal hot spot areas consistent with the conductor axis are screened to eliminate non-axial heat diffusion interference; the spectral absorption attenuation curve is combined to accurately locate the oxidation boundary and enhance the specificity of material degradation identification; through the spatial topological matching of the hot spot migration direction and the oxidation area, the spatiotemporal coupling relationship between the hot spot diffusion path and the oxidation distribution is intuitively presented, providing a visual carrier for defect cause analysis; based on the quantitative relationship between the hot spot path angle change and the spectral mutation point, a causal judgment model for oxide layer degradation and overheating defects is constructed to improve the physical interpretability of defect classification; through the linkage rules of critical thresholds and quantity thresholds, automatic defect level classification is realized, and a structured report is generated based on the defect location, level and diffusion trend to support accurate operation and maintenance decision-making.

[0053] Furthermore, in the visual correlation map, continuous detection intervals are divided along the hot spot diffusion path and the sequence of angle changes with the conductor axis is quantified. Within the oxidation boundary contour, equal-area detection units are divided and the distribution of jump times of the spectral absorption attenuation rate is statistically analyzed. Abnormal diffusion sections with excessive angle increments and active degradation units with excessive jump times are screened, and dynamic correlation rules between angle changes and jump times are established based on the spatial overlap between the two. By collaboratively analyzing the dynamic characteristics of hot spot diffusion in infrared time series data and the spatial distribution of multi-spectral oxidation degradation activity, the problem of traditional methods being unable to correlate defect thermal manifestations with the root causes of material degradation is resolved, and the overheating defect of the conductor is accurately linked to the state of the oxide layer, improving defect detection accuracy and reducing false alarm rates.

[0054] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0056] Figure 1 A flowchart of an automated intelligent defect analysis method based on UAV image acquisition of power transmission and distribution lines provided by the present application is shown;

[0057] Figure 2 The present invention provides a schematic diagram of an automated intelligent defect analysis system for power transmission and distribution line image acquisition based on drones;

[0058] Figure 3 A schematic structural diagram of a computing device provided by the present application is shown. DETAILED DESCRIPTION

[0059] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0060] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. 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 of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to being different types.

[0061] Researchers have found that existing methods for detecting overheating defects in power transmission and distribution line conductors often rely on a single data source (such as static infrared temperature measurement or visible light imaging). This makes it difficult to effectively distinguish between local overheating caused by oxidation and degradation of the conductor surface and transient environmental interference. Furthermore, a dynamic correlation model between the abnormal thermal diffusion path and the state of the oxide layer cannot be established, resulting in a high rate of false positives and delayed early warnings. Based on this, the present invention provides an automated intelligent defect analysis method based on drone image acquisition of power transmission and distribution lines. Specifically, the method uses a drone to synchronously collect infrared thermal imaging time-series data and multispectral imaging data from the conductor surface. It then screens axial abnormal hot spot areas based on the temperature gradient direction and locates the oxidation boundary contour. The hot spot migration direction and oxidation distribution are then integrated to generate a visual correlation map. Dynamic correlation rules are then established between the change in the hot spot diffusion angle and the spectral mutation point, enabling defect level determination and trend analysis. This method can accurately identify local overheating defects caused by oxide layer degradation, reduce false alarm rates, and support visual tracking and risk grading of defect evolution paths, providing a reliable decision-making basis for intelligent operation and maintenance of power transmission lines.

[0062] The technical solution of the present application can be applied to scenarios of early warning and dynamic tracking of overheating defects at wire connection points.

[0063] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0064] Figure 1 The present invention provides a flowchart of an automated intelligent defect analysis method based on drone image acquisition of power transmission and distribution lines, as shown in FIG. Figure 1 As shown, the method includes:

[0065] 101. Collect infrared thermal imaging time series data and multispectral imaging data of the conductor surface of the power transmission and distribution line within a continuous time window using drones;

[0066] In this step, infrared thermal imaging time-series data is obtained by the drone continuously scanning along the conductor's extension direction, showing the temperature field on the conductor surface changing with the flight path. Multispectral imaging data is obtained by the drone hovering at a fixed point, showing the difference in spectral reflectance intensity between the oxide layer on the conductor surface and the normal metal surface in selected wavelength bands.

[0067] In the embodiment of the present application, first, a uniform flight path is planned along the axis of the conductor by the drone to ensure that the infrared thermal imager continuously collects the surface temperature field data of the conductor at a frame rate of 30 Hz, and the multispectral camera is synchronously triggered to capture the spectral reflectance image in the preset band to obtain multispectral imaging data; secondly, the infrared data is subjected to radiation correction, and the original radiation value is converted into an absolute temperature value based on the ambient temperature and the material emissivity parameter (such as the emissivity of the aluminum conductor is 0.95). At the same time, the reflectivity of the multispectral imaging data is calibrated, and the whiteboard reference value is used to eliminate light interference; finally, the spatial coordinates and timestamp of each frame of data are recorded by the GPS / INS system carried by the drone, and the spatiotemporal positions of the infrared and multispectral data are aligned using a linear interpolation algorithm to ensure that the spatial error between the two is less than 5 cm.

[0068] In a scenario involving overheating defects at a conductor connection point on a 500 kV transmission line, a drone flew 1.5 km along the conductor at a speed of 1 m / s. The infrared thermal imager collected 15 minutes of time-series data (temperature resolution 0.1°C), while the multispectral camera acquired high-resolution images in five bands. Calibrated data revealed an abnormal temperature region (maximum temperature difference ±4°C) and spectral reflectance abnormality (reflectance in the 650 nm band was 30% below normal) in the middle of the conductor.

[0069] 102. Calculate the temperature gradient direction of the conductor main body area in each frame of the infrared thermal imaging time series data, extract the abnormal hot spot area consistent with the conductor axis from the temperature gradient direction, and locate the boundary contour of the conductor surface oxidation area based on the spectral absorption attenuation curve corresponding to the conductor surface oxide layer in the multispectral imaging data;

[0070] In this step, the temperature gradient direction is the vector direction calculated from the temperature difference between adjacent pixels, representing the heat diffusion trend. The spectral absorption attenuation curve is a curve showing the reflection intensity of the oxide layer area as a function of wavelength. The slope mutation point reflects the activity of oxidation degradation.

[0071] In the embodiment 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, and 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 whose temperature gradient direction has an angle less than 10° with the conductor axis are screened, and adjacent areas are merged through morphological closing operation (3×3 rectangular kernel), and isolated noise points with an area less than 10 cm² are eliminated and marked as abnormal hot spot areas; at the same time, the spectral angle classification (SAM) algorithm is used for multispectral imaging data, and the normal metal spectrum is used as the reference vector to calculate the spectral similarity of each pixel (threshold <0.9 is determined to be an oxide layer), and the contour of the oxide area is extracted by combining Canny edge detection, and finally the boundary contour of the oxide area on the conductor surface is generated by polygon fitting.

[0072] For example, continuing with the above example, temperature gradient direction calculation found three abnormal hot spots in the middle section of the conductor, with a maximum gradient direction deviation of 7.5°. After morphological processing, they were merged into two continuous areas (with areas of 0.3㎡ and 0.5㎡ respectively); multi-spectral SAM analysis located the oxidation area of 1.1㎡, with a boundary contour fitting accuracy of ±2cm, which partially overlapped with the temperature anomaly area.

[0073] 103. Perform spatial topological matching on 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;

[0074] In this step, the visual correlation map is a spatial overlay map that integrates the hot spot diffusion path (time migration direction) and the oxidation area distribution, which is used to intuitively show the correlation between defects and the oxide layer.

[0075] In an embodiment of the present application, first, the coordinates of the center point of the migration direction of the abnormal hot spot area are fitted with least squares linear fitting to generate a hot spot diffusion path, wherein the path slope reflects the diffusion rate (e.g., 0.25 m / s); secondly, the hot spot diffusion path and the boundary contour are imported into the same coordinate system, a spatial association network is established through Delaunay triangulation, and the overlap ratio of the path and the boundary contour is calculated (e.g., an overlap > 70% is marked as a strong association); thirdly, a visualization engine (e.g., 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, the oxidized area is filled with translucent blue, and the visual association map in PNG format is superimposed.

[0076] For example, continuing with the above example, the hot spot diffusion path shows that two abnormal areas diffused along the conductor axis at rates of 0.2m / s and 0.35m / s respectively. The 0.35m / s path had an 82% overlap with the oxidation boundary and was marked as a "high-risk association area" through the generated visual association map. The other path had a 45% overlap and was marked as "medium risk."

[0077] 104. Establish a dynamic association rule between the wire overheating defect and the surface oxide layer state based on the change in the angle 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;

[0078] In this step, the angle change is the real-time angle difference between the tangent direction of the hot spot diffusion path and the conductor axis, reflecting the degree of deviation from the abnormal thermal diffusion. The spectral mutation point is the mutation position in the spectral absorption attenuation curve of the oxidation region where the absolute value of the slope exceeds 0.5, indicating the activity of oxidative degradation.

[0079] In an embodiment of the present application, first, the detection intervals are divided at intervals of 1m along the hot spot diffusion path, and the average value of the angle change between the path tangent and the axial direction in each interval is calculated (such as interval 1: 5°, interval 2: 8°, increment 3°); secondly, a 10cm×10cm grid is divided within the oxidation boundary, and a sliding window slope detection (window width 10nm) is applied to the spectral curve in each grid, and the number of mutation points with an absolute value of the slope exceeding 0.5 is counted; thirdly, a dynamic rule is established based on a logistic regression model, the input features are the average value of the angle change and the number of mutation points, and the defect level (low / medium / high risk) is output, and the model weight is optimized and determined through historical training data (such as 100 groups of samples).

[0080] For example, continuing with the previous example, the average angle change in the high-risk association area is 6.2° / m, and the number of mutation points is 5 times / grid. The logistic regression model establishes dynamic association rules and outputs a defect level of "high risk" (confidence level 91%). The average angle change in the medium-risk area is 3.8° / m, and the number of mutation points is 2 times / grid, which is judged as "medium risk" (confidence level 76%).

[0081] 105. Based on the dynamic association rule, according to the critical angle threshold of the angle change and the threshold of the number of mutation points, determine the local overheating defect level caused by oxide layer degradation on the surface of the wire, and generate a corresponding defect analysis report according to the local overheating defect level.

[0082] In this step, the critical angle threshold is the upper limit of the angle change set based on the wire material; exceeding this limit triggers an alert. The quantity threshold is the lower limit of the number of mutation points within the oxidation region grid, which is used to determine the degradation activity.

[0083] In the embodiment of the present application, first, a critical angle threshold (such as 5° / m for aluminum wire) and a mutation point number threshold (≥3 times / grid) are set according to the wire material. If the angle change exceeds the threshold and the number of mutation points meets the standard, it is marked as "confirmed defect"; secondly, a structured defect analysis report is generated, integrating the defect location (GPS coordinates), hot spot diffusion dynamic GIF (time-series hot spot overlay animation), high-definition screenshots of the oxidation area and maintenance suggestions (such as "re-inspect within 48 hours"); finally, the report is automatically pushed to the responsible unit via email or the operation and maintenance platform.

[0084] For example, continuing with the previous example, the high-risk associated area triggers a "confirmed defect" and the defect analysis report indicates that the defect is located 1.5 km from the conductor (GPS: E118°45'36", N32°03'12"). It is recommended to immediately shut down the power supply for maintenance and replace the oxidized section of the conductor. The medium-risk area is marked as a "suspected defect" and it is recommended to re-inspect and apply anti-corrosion coating within 2 weeks.

[0085] Steps 101-105 utilize drone-based collaborative multi-source data collection and fusion analysis to accurately correlate conductor overheating defects with oxidation degradation. Infrared time-series data dynamically tracks the diffusion path of hot spots, multispectral data locates oxidation boundaries, and spatial topology matching generates visual maps to support causal analysis of defects. Dynamic rule modeling based on angle variation and spectral mutations overcomes the limitations of traditional single threshold criteria. Automated report generation and push notification significantly improves operational response efficiency, making it suitable for intelligent inspection and preventive maintenance of high-voltage transmission lines.

[0086] In order to solve the problem of dynamic correlation determination of local overheating defects caused by oxidation degradation at the connection points of transmission and distribution line conductors, and further improve the accuracy of defect warnings and trend prediction capabilities, in some embodiments, the dynamic correlation rules between conductor overheating defects and the state of the surface oxide layer are established based on the change in the angle between the hot spot diffusion path and the conductor axis in the visual correlation map, and the mutation point of the spectral absorption attenuation rate within the boundary contour, including:

[0087] 201. In the visual association map, divide continuous detection intervals along the extension direction of the hot spot diffusion path with the wire axis as a reference axis, quantify the change in the angle between the hot spot diffusion path and the wire axis within the continuous detection intervals, and generate an angle change sequence;

[0088] In step 201, the continuous detection intervals are continuous analysis units divided into evenly spaced intervals along the wire axis along the hot spot diffusion path, used to quantify local diffusion characteristics. The angle change is the real-time angle difference between the tangent direction of the hot spot diffusion path and the wire axis, reflecting the degree of deviation in the heat diffusion direction.

[0089] In an embodiment of the present application, first, based on the spatial coordinates of the hot spot diffusion path in the visual association map, the wire axis is used as the reference axis and continuous detection intervals are divided according to preset intervals (such as 1 meter); secondly, cubic spline interpolation fitting is performed on the path in each interval, and the real-time angle between the tangent of the fitting curve and the axis is calculated (such as the angle at the starting point of the interval is 5°, the angle at the end point is 9°, and the change is 4°); finally, an angle change sequence (such as [5°, 9°, 12°, ...]) is generated in the order of the intervals for subsequent abnormal diffusion judgment.

[0090] 202. Divide the coverage area of the boundary contour into detection units of equal area, count the number of transitions of the spectral absorption attenuation rate from a stable state to a sudden change in the detection units, and generate a distribution graph of the number of transitions for each detection unit;

[0091] In step 202, the equal-area detection cells are fixed-area analysis grids divided within the oxidation boundary contour coverage area for local spectral mutation statistics. The jump frequency distribution diagram is a heat map of the number of times the spectral absorption decay rate changes from stable to sudden changes within each detection cell.

[0092] In the embodiment of the present application, the oxidation boundary contour is first divided into equal-area grids of 10 cm × 10 cm; secondly, the spectral absorption attenuation curve in each grid is subjected to sliding window mean filtering (window width 5 nm), and the rate fluctuation amplitude in the window is calculated. If the rate difference between adjacent windows exceeds 15%, it is marked as a jump; finally, the number of jumps in all grids is counted to generate a two-dimensional frequency distribution heat map (such as red represents a high jump area and blue represents a low jump area).

[0093] 203. Mark a section in the angle change sequence where the angle increment between adjacent detection intervals exceeds a preset increment as an abnormal diffusion section, and mark a detection unit in the jump frequency distribution diagram where the jump frequency exceeds a preset number as an active degraded unit;

[0094] In step 203, the abnormal diffusion section is the section in the continuous detection interval where the increment of adjacent angles exceeds a preset increment, indicating an area of abnormally accelerated thermal diffusion. The active degradation unit is the detection unit in the jump number distribution diagram where the jump number exceeds a preset threshold, reflecting the active core area of oxidative degradation.

[0095] In an embodiment of the present application, a differential calculation is first performed on the angle change sequence to extract the angle increments of adjacent intervals (such as the increment of interval 1→2 is 3°, and the increment of interval 2→3 is 6°), and continuous segments with increments ≥ 5° / m (such as interval 2-3) are screened and marked as abnormal diffusion segments; at the same time, grids with jump times ≥ 4 times are extracted from the jump number distribution diagram, and isolated noise points are eliminated through morphological corrosion operations to retain continuously distributed active degraded units.

[0096] 204. According to the coverage overlap between the spatial position of the abnormal diffusion section and the spatial density distribution of the active degradation unit, a linkage judgment rule between the angle change of the hot spot diffusion path and the number of jumps of the spectral absorption attenuation rate is set as a dynamic association rule.

[0097] In step 204 , the coverage coincidence is the overlap ratio between the abnormal diffusion section and the active degradation unit in spatial distribution, which is used to quantify the thermal-oxidation correlation strength.

[0098] In the embodiment of the present application, the surface area of the conductor covered by the abnormal diffusion section is first calculated, and the proportion of the number of active degraded units in the area is counted to quantify the coverage overlap (for example, if the coverage area is 1 m2, the proportion of active units is 70%); secondly, a linkage judgment rule is set based on the proportion of the number (for example, when the proportion is ≥60%, the number of jumps must be ≥2 for every increase of 1° / m in the angle change), forming a dynamic association rule base.

[0099] Here's a specific example:

[0100] Assume that a connection point of a 500kV transmission line conductor causes local overheating due to oxide layer degradation. The operation and maintenance personnel need to track the heat diffusion trend in real time and determine the driving effect of oxidation degradation on the defect. Through drone inspection, it was found that the center temperature of the connection point abnormally increased by 9°C. Infrared thermal imaging time series data showed that the heat diffused to both sides along the axis of the conductor. In the visual correlation map, the continuous detection interval was first divided into 1-meter intervals along the diffusion path of the hot spot (step 201). It was found that the angle change in the interval of 1-3 meters from the connection point increased rapidly from 7° to 19°, generating an angle change sequence [7°, 13°, 19°], indicating that the direction of heat diffusion significantly deviates from the conductor axis; then, 10cm×10cm detection units are divided within the oxidation boundary coverage area (step 202), and the spectral absorption attenuation curve analysis of the core oxidation area of the connection point (annular area with a diameter of 0.6m) shows that the number of jumps in 60% of the units exceeds 5 times, forming a high-density active degradation unit; then, abnormal diffusion sections (1-3 meters range) with an angle increment exceeding 5° / m and active degradation units with a jump number of ≥5 times are screened out (step 203). Spatial superposition analysis shows that the coverage overlap between the two reaches 88%, and the heat diffusion rate is positively correlated with the number of jumps; finally, dynamic association rules are set based on the coverage overlap (step 204), and it is determined that the overheating defect of the connection point is dominated by oxidation degradation, and the defect level is "critical".

[0101] In order to accurately identify local overheating defects caused by oxidation degradation at wire connection points and solve the problems of misjudgment of heat diffusion direction and fuzzy oxidation boundary positioning in traditional methods, in some embodiments, the temperature gradient direction of the wire main body area in each frame of the infrared thermal imaging time series data is calculated, and the abnormal hot spot area consistent with the wire axis is extracted from the temperature gradient direction. At the same time, based on the spectral absorption attenuation curve corresponding to the wire surface oxide layer in the multispectral imaging data, the boundary contour of the wire surface oxidation area is located, including:

[0102] 301. In a conductor main body region of each frame in the infrared thermal imaging time series data, divide temperature detection zones at equal intervals along the axial extension direction of the conductor, and arrange temperature sampling points in the temperature detection zones in a direction perpendicular to the axial direction of the conductor;

[0103] In step 301 , the temperature sampling points are sampling positions evenly distributed in a direction perpendicular to the axial direction of the conductor within the temperature detection zone, with the intervals matching the diameter of the conductor.

[0104] In the embodiment of the present application, first, the main area of the conductor in the infrared thermal imaging data is divided into equally spaced temperature detection zones at a preset interval (0.5 meters) based on the spatial coordinates of the conductor axis. Secondly, temperature sampling points are evenly set in each temperature detection zone along a direction perpendicular to the conductor axis (i.e., the circumferential direction). The spacing between the sampling points is adaptively adjusted according to the curvature of the conductor surface (e.g., 1 cm interval for straight sections and 0.5 cm interval for curved sections), forming a transverse temperature sampling sequence.

[0105] 302. Within the temperature detection zone, calculate the temperature difference between a temperature sampling point and its adjacent sampling points, generate a temperature gradient direction along the arrangement direction of the temperature sampling points based on the temperature difference, calculate a direction distribution histogram of the temperature gradient directions, and select a temperature gradient direction in the direction distribution histogram whose angle with the conductor axis is less than a preset angle;

[0106] In step 302 , the directional distribution histogram is a histogram obtained by statistically analyzing the directional angle distribution of all temperature gradient vectors and dividing the data into preset intervals.

[0107] In the embodiment of the present application, first, for each temperature sampling point in the temperature detection zone, the temperature difference between it and the adjacent points on the left and right is calculated (such as the temperature difference of the left point is +2°C, and the temperature difference of the right point is -3°C), and the temperature gradient direction from low temperature to high temperature is generated; secondly, the direction angles of all temperature gradient directions are counted, and the direction distribution histogram is generated in 5° intervals with the wire axis as the 0° reference; finally, the gradient vectors in the histogram whose angle with the axis is less than 10° (such as the 0°-10° and 170°-180° intervals) are screened to eliminate non-axial thermal interference.

[0108] 303. Perform spatial clustering on the screened temperature gradient directions in the conductor main body region, merge the regions showing continuous distribution and having temperature gradient direction consistency exceeding a minimum ratio in the spatial clustering results, and mark them as abnormal hot spot regions;

[0109] In step 303 , spatial clustering is to merge the filtered temperature gradient directions into continuous regions according to spatial proximity.

[0110] In an embodiment of the present application, a density clustering algorithm (such as DBSCAN) is first used to spatially cluster the screened temperature gradient directions, and adjacent gradient directions with a directional deviation of less than 5° are merged; secondly, the directional consistency ratio within each cluster area is calculated (such as 80% of the vector direction deviations in the area are <5°). If the ratio exceeds 75%, it is marked as an abnormal hot spot area.

[0111] 304. In the multispectral imaging data, a spectral sampling band is divided along the surface contour line of the conductor, and the slope change points of the spectral absorption attenuation curve corresponding to the conductor surface oxide layer are extracted within the spectral sampling band. The slope change points are connected to the sampling band boundary where the length of the curve segment between adjacent slope change points is less than a preset distance, so as to form a closed boundary contour of the conductor surface oxidation area.

[0112] In step 304, the spectrum sampling band is a strip area divided along the surface contour of the conductor for local spectrum analysis. The slope change point is the sudden change position where the absolute value of the slope in the spectrum absorption attenuation curve exceeds the threshold.

[0113] In the embodiment of the present application, first, a spectral sampling band with a width of 1 cm is divided along the surface contour line of the conductor; second, the first-order derivative of the spectral absorption attenuation curve within each sampling band is calculated, and the mutation points with an absolute value of the slope exceeding 0.5 are extracted; finally, if the length of the curve segment between adjacent mutation points is less than a preset distance (such as 5 mm), the two points are connected to form the boundary contour of the oxidation area.

[0114] Here's a specific example:

[0115] Suppose a 330kV transmission line conductor connection point is experiencing localized overheating due to oxidation corrosion. The defective area needs to be precisely located and the correlation between the heat diffusion path and oxidation degradation needs to be analyzed. In step 301, four temperature monitoring zones (one every 0.5 meters) are divided around the connection point and the 2-meter conductor area on either side. Twenty temperature sampling points (1 cm apart) are set within each zone. Analysis in step 302 shows that the temperature gradient in the central zone of the connection point is concentrated between 5° and 8° (angle with the axial direction), while the gradient directions in the adjacent zones are dispersed (15° to 30°). Gradient directions with axial angles less than 10° are selected. In step 303, spatial clustering analysis reveals a continuous region (0.12 m2) at the center of the connection point, with a directional consistency rate of 85%. This region is marked as an abnormal hot spot. In step 304, spectral sampling zones are divided along the conductor surface contour. Eight slope mutation points are extracted within the central spectral sampling zone of the connection point. The spacing between adjacent mutation points is less than 3 mm. After connection, the boundary outline of a closed oxidation zone (a 0.25 m diameter ring) is formed. Ultimately, the abnormal hot spot overlapped with the oxidation boundary by 90%, and was determined to be a critical defect driven by oxidation. A report was generated recommending immediate power outage, maintenance, and coating of anti-corrosion materials.

[0116] Steps 301-304 effectively distinguish axial heat diffusion from non-axial environmental interference at conductor connection points through refined division of 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 material deterioration areas. The spatial overlap analysis of abnormal hot spots and oxidation boundaries enables physically explainable determination of the cause of defects, providing full-process support for transmission lines from early warning to precise operation and maintenance, significantly reducing the risk of false detection and missed detection.

[0117] In order to accurately characterize the dynamic coupling relationship between thermal diffusion and oxidative degradation in overheating defects at wire connection points and to address the issues of strong subjectivity and insufficient visualization in traditional methods for determining thermal-oxidative correlation, in some embodiments, the migration direction of the abnormal hot spot area is spatially topologically matched with the boundary contour to generate a visual correlation map that integrates the hot spot diffusion path and the oxidation area distribution, including:

[0118] 401. Generate a hot spot diffusion path based on 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 a migration direction line of the abnormal hot spot area within the equal-length path segments. The migration direction line is a central extension trajectory of the hot spot diffusion path within the equal-length path segments.

[0119] In step 401, the temporal position change path is the trajectory of the center point of the abnormal hot spot region within a continuous time window, reflecting the spatiotemporal evolution of heat diffusion. The migration direction line is an extended trajectory line fitted to the center point of the hot spot diffusion path within a path segment of equal length, representing the local diffusion direction and rate.

[0120] In an embodiment of the present application, first, cubic spline interpolation is performed on the center point coordinates of the time-series position change path of the abnormal hot spot area to generate a smooth diffusion path; secondly, the diffusion path is divided into equal-length segments along the axis of the wire (such as every 0.5 meters), and least squares linear fitting is applied to each segment of the path, and the center extension trajectory line is extracted as the migration direction line, and its slope reflects the diffusion rate of the segment (such as 0.3m / s).

[0121] 402. Divide the detection grids into equal widths along the boundary contour, calculate the distribution density of oxidation areas within the detection grids, and generate an oxidation area density distribution map;

[0122] In step 402, a uniform-width detection grid is formed by dividing the oxidized area into strips of fixed-width grid cells along the oxidized boundary. This grid is used to quantify the spatial distribution density of the oxidized area. The oxidized area density distribution map is a heat map of the area ratio of the oxidized area within each grid cell, with red representing high-density areas and blue representing low-density areas.

[0123] In the embodiment of the present application, first, an equal-width detection zone is generated by expanding 1 cm along the oxidation boundary contour line; secondly, the detection zone is divided into 1 cm × 1 cm grid units, and the proportion of oxidized area pixels in each unit is calculated (for example, the oxidized pixels in unit A account for 70%); finally, a continuous oxidation area density distribution map is generated by performing bilinear interpolation on the oxidation area pixel proportions.

[0124] 403. Spatially superimpose 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 based on the spatial superposition result. Mark the path segment where the extension trend direction is consistent with the direction of the density increase region as the heat and oxidation action segment.

[0125] In step 403, the extension trend direction is the angular variation trend of the migration direction line within the path segment. The density increase region direction is the spatial variation trend of the oxidation region density along the conductor axis.

[0126] In an embodiment of the present application, the migration direction line and the density distribution map are first imported into the same coordinate system; secondly, the angular change trend of the migration direction line (such as the linear regression slope) is calculated for each path segment in the coordinate system, and the density gradient direction of the corresponding grid area is extracted (such as the density increases from east to west); finally, the path segments with the same direction of the two (such as both showing positive increase) are marked as the action sections of heat and oxidation.

[0127] 404. Calculate the extension angle change rate of the migration direction line, and simultaneously 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 results into a single visualization layer to generate a visualization correlation map.

[0128] In step 404, the extension angle change rate is the angle change per unit length of the migration direction line. The density change direction is the density gradient direction inside the oxidation region.

[0129] In an embodiment of the present application, the angular change rate of the migration direction line within the action section is first calculated (for example, the angle in the 0-0.5 meter interval increases from 10° to 15°, with a change rate of 10° / m); secondly, the density gradient direction of the corresponding grid area is extracted (for example, the density increases along the 30° direction); finally, the directional consistency scores of the two are calculated through spatial convolution (for example, cosine similarity > 0.8), and the results are rendered as arrow layers (red indicates high consistency, and yellow indicates low consistency), which are superimposed to generate a visual association map.

[0130] Here's a specific example:

[0131] The connection point of a 500kV transmission line conductor caused local overheating due to the peeling of the oxide layer. It was necessary to analyze the spatiotemporal correlation between the heat diffusion path and oxidation degradation and generate a visual decision 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 the conductors on both sides. After dividing it into 6 0.5-meter path segments, the fitting found that the migration direction line angle of the left path segment (0-1.5 meters) increased from 5° to 20°, and the diffusion rate increased from 0.2m / s to 0.5m / s; step 402 divides the equal-width detection grid along the oxidation boundary, and generates a density distribution map showing that the oxidation density on the west side of the connection point is 80%, and on the east side it is only 30%; step 403 analyzes the migration direction line trend (increasing angle on the left) and the density The direction of the density gradient (decreasing from west to east) is shown, and the 0-1.5 meters on the left side is marked as the heat and oxidation zone. Step 404 calculates the angle change rate of this zone as 12° / m and the density gradient direction as 150°, with a spatial consistency score of 0.85 (>0.8 threshold). The high consistency area is marked with a red arrow in the visualization map. After superimposing the hot spot path and the oxidation density map, it is shown that the high heat density area on the west side completely overlaps with the fast diffusion path, which is determined to be an oxidation-dominated defect. The generated report recommends replacing the west side conductor first and applying an anti-corrosion layer.

[0132] Steps 401-404 achieve a dynamic visualization of the correlation between overheating defects and oxidation degradation at conductor connection points through refined segmented analysis of hot spot diffusion paths and spatial coupling determination of oxidation density gradient directions. The collaborative analysis of migration direction lines and density gradients breaks through traditional static threshold limitations and accurately identifies high-risk areas for oxidation-driven heat diffusion. The visual map intuitively presents defect evolution trends and operation and maintenance priorities, providing full-chain decision support for transmission lines from defect tracing, risk classification to precise disposal, significantly improving operation and maintenance efficiency and reliability in complex defect scenarios.

[0133] In order to solve the problem of insufficient accuracy in the dynamic correlation determination between oxidation degradation and thermal diffusion in overheating defects at wire connection points and further improve the reliability of defect root cause tracing and early warning, in some embodiments, the dynamic correlation rule is set as follows: the linkage determination rule of the angle change of the hot spot diffusion path and the number of jumps of the spectral absorption attenuation rate is set based on the overlap between the spatial position of the abnormal diffusion section and the spatial density distribution of the active degradation unit, including:

[0134] 501. Divide the conductor surface area corresponding to the abnormal diffusion section into equal-area analysis units, calculate the coverage overlap between the number of active degraded units in the equal-area analysis unit and the total number of grids, and mark the equal-area analysis unit as a thermal oxidation-related unit when the coverage overlap exceeds a preset overlap ratio;

[0135] In step 501 , equal-area analysis cells are fixed-area analysis grids divided within the conductor surface region corresponding to the abnormal diffusion section, for quantifying the spatial correlation between heat and oxidation.

[0136] In the embodiment of the present application, the surface area of the conductor covered by the abnormal diffusion section is first divided into 1m×0.1m rectangular analysis units; secondly, the number of active degraded units in each unit is counted (for example, unit A has 8 active units), and its ratio to the total number of unit grids (for example, 20) is calculated (8 / 20=40%); finally, if the ratio exceeds a preset ratio (for example, 30%), the unit is marked as a thermal oxidation-related unit.

[0137] 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 overlap and the angle increment sequence, and mark the detection section as a significant diffusion section when the correlation fluctuation amplitude exceeds the fluctuation upper limit;

[0138] In step 502, the correlation fluctuation amplitude is the covariance value of the overlap and angle increment series, reflecting the consistency of their dynamic changes. The significant diffusion section is a continuous detection interval where the correlation fluctuation amplitude exceeds the preset upper limit, indicating an active area of thermal-oxidation synergy.

[0139] In an embodiment of the present application, first, a sequence of angle increments of all detection intervals in the abnormal diffusion segment corresponding to the thermal oxidation association unit is extracted (such as [3°, 5°, 7°]); secondly, the covariance of the sequence and the coverage overlap sequence (such as [30%, 40%, 50%]) is calculated (such as covariance > 0.8); finally, if the covariance exceeds the upper limit of fluctuation (such as 0.7), the segment is marked as a significant diffusion segment.

[0140] 503. In the significant diffusion section, calculate the product of the coverage overlap and the sum of the angle increments, and obtain a thermal oxidation correlation index after normalizing the product;

[0141] In step 503 , the thermal-oxidation coupling index is a normalized product of the overlap and the sum of the angle increments, and is used to quantify the thermal-oxidation coupling strength.

[0142] In the embodiment of the present application, the product of the mean coverage overlap (e.g., 45%) and the sum of the angle increments (e.g., 15°) in the significant diffusion section is first calculated (45%×15=6.75); secondly, the product value is normalized to the range of 0-1 (6.75 / 10=0.675) through maximum and minimum value normalization (e.g., the global maximum product value is 10, the minimum is 0) to generate a thermal oxidation correlation index.

[0143] 504. Establish a linkage judgment rule between the angle change of the hot spot diffusion path and the number of jumps of the spectral absorption attenuation rate based on the thermal oxidation correlation index. When the thermal oxidation correlation index exceeds a preset standard value, it is determined that there is a wire overheating defect caused by oxide layer degradation.

[0144] In step 504, the linkage judgment rule is a defect judgment condition set based on the thermal oxidation correlation index, and an oxide layer degradation defect alarm is triggered when the index exceeds the limit.

[0145] In the embodiment of the present application, a linkage judgment rule is set: a standard value (such as 0.6) is set according to historical data. If the thermal oxidation correlation index exceeds this value, it is determined that there is an overheating defect caused by the oxide layer; secondly, a defect level label is generated (such as 0.6-0.8 is "medium risk", >0.8 is "high risk"), and is linked to the operation and maintenance disposal suggestion library (such as high risk requires power outage and maintenance within 48 hours).

[0146] Here's a specific example:

[0147] Assume that oxidation corrosion causes abnormal local temperature rise at the connection point of a 220kV transmission line conductor. It is necessary to accurately determine the defect level and formulate an emergency repair strategy. In step 501, the 1.2m abnormal diffusion zone west of the connection point was divided into 12 analysis units (1m x 0.1m). Calculations revealed that the coverage overlap of units 3-6 reached 35%-50% (preset threshold 30%), and these units were marked as thermal oxidation-related units. In step 502, the angle increment sequence [4°, 6°, 9°] and the coverage overlap sequence [35%, 42%, 50%] for this section were extracted, and the covariance was calculated to be 0.85 (>0.7 threshold), marking this section as a significant diffusion section. In step 503, the product of the mean coverage overlap of 45% and the total angle increment of 19° was calculated to be 8.55, resulting in a normalized thermal oxidation correlation index of 0.855. In step 504, based on the index of 0.855 (>0.8), this defect was identified as a "high-risk defect," triggering a Level 3 alarm and generating a report. This report required immediate power outage and replacement of the connection point and the 1.5m conductors on both sides. A high-definition map of the oxidation area and a thermal diffusion prediction model were simultaneously delivered to a mobile terminal.

[0148] Steps 501-504 achieve precise quantitative correlation between overheating defects at conductor connection points and oxide layer degradation through dynamic collaborative analysis of coverage overlap and angle increment, breaking through the limitations of traditional single-parameter threshold judgment; the standardized design of the thermal oxidation correlation index supports the comparability of defect levels across scenarios and reduces the risk of misjudgment caused by environmental interference; the deep binding of linkage judgment rules and operation and maintenance disposal recommendations provides full-process closed-loop support for transmission lines from defect detection, root cause analysis to emergency disposal, significantly improving the timeliness and reliability of operation and maintenance under complex working conditions.

[0149] In order to solve the problems of high misjudgment rate of heat diffusion direction and difficulty in distinguishing between normal temperature rise and abnormalities caused by oxidation degradation in traditional wire overheating defect detection, and to further improve the accuracy of early warning, in some embodiments, within the temperature detection zone, the temperature difference between a temperature sampling point and its adjacent sampling points is calculated, and a temperature gradient direction is generated along the arrangement direction of the temperature sampling points based on the temperature difference. A directional distribution histogram of the temperature gradient direction is statistically analyzed, and the temperature gradient direction in the directional distribution histogram whose angle with the wire axis is less than a preset angle is screened, including:

[0150] 601. Within the temperature detection zone, calculate the temperature difference between the temperature sampling point and the left and right adjacent sampling points, and generate a temperature gradient direction from low temperature to high temperature based on the positive and negative signs and absolute values of the temperature differences;

[0151] In step 601, the temperature gradient direction is a vector direction from low temperature to high temperature, which is generated by the sign and magnitude of the temperature difference and represents the heat diffusion direction.

[0152] In an embodiment of the present application, first, sampling points are arranged along the circumference of the conductor within the temperature detection zone (e.g., at intervals of 1 cm), and the temperature difference between each sampling point and its left and right adjacent points is calculated (e.g., the left difference is +2°C, the right difference is -3°C); secondly, the temperature gradient direction is generated based on the difference sign (if the left difference > the right difference, the direction is to the left, otherwise it is to the right) and the absolute value. 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).

[0153] 602. Taking the conductor axis as a reference direction, statistically analyzing the direction angle distribution of the temperature gradient direction, and dividing the direction angle distribution into preset angle intervals to generate a direction distribution histogram;

[0154] In step 602, the direction angle distribution is the statistical distribution of all temperature gradient direction angles, with the conductor axis as 0°. The direction distribution histogram is a distribution diagram of the number of gradient directions divided into preset angle intervals.

[0155] In the embodiment of the present application, first, the axial direction of the wire is taken as the 0° reference, and the direction angles of all temperature gradient vectors are converted into the range of 0°-180° (such as 0°-90° for left deflection and 90°-180° for right deflection); secondly, the number of vectors in each interval is counted at 5° intervals (such as there are 15 vectors in the 0°-5° interval) to generate a directional distribution histogram; finally, Gaussian smoothing filtering is used to eliminate noise interference and enhance distribution continuity.

[0156] 603. Filter out, from the direction distribution histogram, temperature gradient directions whose angles with the conductor axis are smaller than a preset angle, where the preset angle is set according to a maximum allowable thermal diffusion offset during normal operation of the conductor;

[0157] In step 603 , the preset angle is a maximum allowable directional deviation angle set according to the thermal diffusion characteristics of the wire material.

[0158] In an embodiment of the present application, the temperature gradient directions in the ranges of 0°-10° and 170°-180° (corresponding to an axial offset of <10°) are first extracted from the direction distribution histogram; secondly, adjacent temperature gradient directions are merged through connected domain analysis, and isolated noise points (such as areas with an area of <5 cm²) are eliminated; finally, a filtered set of axial gradient directions is generated.

[0159] 604. Perform quantitative analysis on the spatial distribution and directional angle deviation of the screened temperature gradient direction, wherein the quantitative analysis includes statistics on the distances between adjacent vectors and calculation of the cumulative distribution of the directional angle deviation.

[0160] In step 604 , the direction angle deviation is the difference between the filtered vector direction and the regional average direction, and is used to quantify the direction consistency.

[0161] In the embodiment of the present application, the Euclidean distance between adjacent vectors is first calculated (for example, the distance between vectors A and B is 2 cm), and the spacing distribution is statistically analyzed (for example, 80% of the spacing is <3 cm); secondly, the deviation of each vector direction from the average direction of the region is calculated (for example, the average direction is 5°, and the deviation of a certain vector is 2°), and a cumulative distribution curve is generated (for example, 90% of the vector deviation is <5°); finally, the KS test is used to evaluate whether the distribution conforms to the expected thermal diffusion pattern.

[0162] Here's a specific example:

[0163] Suppose a 330 kV transmission line conductor connection point experiences localized temperature anomalies due to oxidation corrosion. It is necessary to determine whether this is an oxidation-driven overheating defect. In step 601, the temperature difference between sampling points within the temperature monitoring zone west of the connection point is calculated. The left-side difference in the central region is consistently greater than the right-side difference (difference ranges from +3°C to +5°C), generating a cluster of leftward-facing gradient vectors. In step 602, the directional angle distribution is calculated. The histogram shows that 80% of the vectors (axial deviation <8°) are concentrated in the 0°-8° range. In step 603, the vectors in this range are filtered and merged into a continuous region (0.25 m2 in area), excluding scattered vectors on either side. In step 604, analysis shows that the average spacing between adjacent vectors is 2.5 cm (95% <3 cm), and 90% of the vectors in the cumulative distribution of directional deviations have deviations <4°, consistent with an axial heat diffusion model. Multispectral data confirms that the oxidation boundary overlap in this region is 85%, indicating an oxidation-driven defect. A report is generated recommending that the oxidized conductor be replaced within 72 hours.

[0164] Steps 601-604 accurately identify abnormal heat diffusion caused by oxidation degradation at conductor connection points through axial gradient directional screening and spatial continuity analysis, effectively distinguishing short-term temperature rises caused by environmental interference; the combination of directional distribution histograms and cumulative deviation distributions enhances the interpretability of heat diffusion patterns and reduces the misjudgment rate; the quantitative analysis results provide a basis for the strength of the correlation between heat and oxidation for operation and maintenance decisions, supporting the intelligent upgrade of transmission lines from defect detection to root cause tracing.

[0165] In order to solve the problems of strong subjectivity in threshold setting and insufficient tracking of defect evolution trends in determining the overheating defect level of wire connection points, and to further improve the scientific nature of defect classification and operation and maintenance decision-making, in some embodiments, the dynamic association rule is used to determine the local overheating defect level caused by oxide layer degradation on the wire surface based on the critical angle threshold of the angle change and the threshold of the number of mutation points, and a corresponding defect analysis report is generated based on the local overheating defect level, including:

[0166] 701. Using the dynamic association rule, set a range of a critical angle threshold for the angle variation, where the range is determined based on a physical upper limit of the abnormal axial thermal diffusion rate of the conductor. Also, set an interval for a threshold of the number of mutation points, where the interval is determined based on a spectral response characteristic of oxide layer degradation activity.

[0167] In step 701, the critical angle threshold range is the range of the angle variation allowed according to the physical limit of thermal diffusion of the wire material. The number threshold range is based on the number of mutations divided by the activity of the oxide layer spectrum response.

[0168] In the embodiment of the present application, the critical angle threshold range is first determined by using experimental data on thermal conductivity of the conductor material (for example, the maximum diffusion rate of an aluminum conductor under normal heat dissipation corresponds to an upper threshold limit of 8° / m). Secondly, the relationship between the oxide layer degradation rate and the number of spectral mutation points is analyzed based on multi-spectral historical data to set the mutation point threshold range (for example, ≥4 times in every 10cm×10cm grid is determined to be high activity).

[0169] 702. Establish a combination relationship table of the ranges and the intervals, wherein the combination relationship table defines matching intervals of critical angle thresholds and quantity thresholds corresponding to defects of different grades;

[0170] In step 702, the combination relationship table is a matching rule table that defines the critical angle threshold and the mutation point number threshold under different defect levels.

[0171] In the embodiment of the present application, a two-dimensional matrix is first established, with the horizontal axis representing the critical angle threshold range and the vertical axis representing the number of mutation points. Secondly, the defect level corresponding to each matrix unit is labeled based on historical defect samples (e.g., an angle of 4°-6° / m and 4-6 mutation points are considered secondary defects). Finally, the matrix weights are optimized through expert experience and machine learning to generate a combination relationship table.

[0172] 703. Based on the range and the interval, select a section in the conductor surface inspection area where the angle variation and the number of mutation points fall within the corresponding threshold intervals as a section to be determined, and obtain an initial defect level for the section to be determined according to the combination relationship table.

[0173] In step 703 , the section to be determined is a continuous section in the conductor surface detection area that satisfies both the critical angle and the mutation point threshold.

[0174] In an embodiment of the present application, the detection sections are first divided along the surface of the conductor (e.g., every 1 meter), and the mean angle change and the number of mutation points of each section are extracted; secondly, the combination relationship table is traversed to match sections that meet the threshold interval (e.g., if the mean angle of a section is 5° / m and there are 5 mutation points, it matches a secondary defect); finally, the initial defect level label is output.

[0175] 704. Perform continuity correction on the initial defect level. When the defect levels of three consecutive segments are the same and the increasing trend of the angle change is consistent, merge the three segments and increase the defect level by one to output a local overheating defect level. Generate a defect analysis report including defect location distribution, level intensity, and diffusion trend based on the local overheating defect level.

[0176] In step 704, continuity modification: enhancing the judgment rules for the consistency and diffusion trend of defect levels in adjacent sections.

[0177] In the embodiment of the present application, first, it is detected whether three consecutive segments have the same defect level (for example, all are level two); secondly, it is analyzed whether the change in their angles shows a monotonically increasing trend (for example, 4° / m→5° / m→6° / m); finally, the segments are merged and upgraded by one level (for example, level two is upgraded to level three), and a defect analysis report including the location, level and diffusion trend is generated.

[0178] Here's a specific example:

[0179] Assume that a conductor connection point on a 500kV transmission line experiences local overheating due to oxide layer spalling. It is necessary to dynamically determine the defect level and develop a prevention and control strategy. In step 701, the critical angle threshold range for the aluminum conductor is set to 3°-8° / m, and the number of mutation points is set to low activity (1-3 times), medium activity (4-6 times), and high activity (≥7 times). In step 702, a combination relationship table is established, defining angles of 4°-6° / m and 4-6 mutation points as "medium risk," and angles ≥6° / m and ≥7 mutation points as "high risk." In step 703, detection revealed that the average angle of the 1-3 meter section west of the connection point was 5.2° / m and had 5 mutation points, marking it as "medium risk." The angle of the 4-6 meter section east of the connection point was 7.1° / m and had 8 mutation points, marking it as "high risk." In step 704, analysis showed that three consecutive sections on the west side all had "medium risk" defect levels, with increasing angle changes (4.8 → 5.2 → 5.5° / m). These sections were combined and upgraded to "high risk." The generated defect analysis report marked the west side as the core diffusion area, recommending priority replacement and coating of the anti-corrosion coating, and designated the east side as an emergency disposal area.

[0180] Steps 701-704 achieve objective quantitative determination of the defect level at the conductor connection point through threshold setting driven by physical properties and spectral response data. The combination of the combined relationship table and continuity correction rules enhances the dynamic tracking capability of defect evolution trends and avoids misjudgment of discrete sections. The multi-dimensional information integration of defect reports provides full-chain support for transmission lines, from risk warning and grading to prevention and control decisions, significantly improving the accuracy and timeliness of operation and maintenance in complex defect scenarios.

[0181] Figure 2The present application provides a schematic diagram of a system for analyzing power transmission and distribution line defects based on automated intelligent defect analysis using drone image acquisition. Figure 2 As shown, the system includes:

[0182] The acquisition module 21 is used to synchronously acquire infrared thermal imaging time-series data and multispectral imaging data of the conductor surface within a continuous time window using an infrared thermal imager and a multispectral camera carried by the drone. The infrared thermal imaging time-series data covers the entire axial length of the conductor through the drone's flight trajectory, capturing the dynamic migration trajectory of the surface temperature field. The multispectral imaging data obtains the spectral absorption difference between the oxide layer on the conductor surface and normal metal in a selected band through the drone's hovering posture.

[0183] a positioning module 22 for calculating the temperature gradient direction of the conductor main body region in each frame of the infrared thermal imaging time series data, extracting the abnormal hot spot region where the temperature gradient direction is consistent with the conductor axis, and locating the boundary contour of the oxidized region on the conductor surface based on the spectral absorption attenuation curve corresponding to the oxide layer in the multispectral imaging data;

[0184] A 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 correlation map integrating the hot spot diffusion path and the oxidation area distribution;

[0185] An establishing module 24 is configured to establish a dynamic association rule between the conductor overheating defect and the surface oxide layer state based on the change in the angle between the hot spot diffusion path and the conductor axis in the visual association map and the mutation point of the spectral absorption attenuation rate within the boundary contour;

[0186] The discrimination module 25 is configured to determine the level of the local overheating defect on the surface of the conductor caused by the degradation of the oxide layer according to the dynamic association rule, the critical angle threshold of the angle variation and the threshold of the number of mutation points.

[0187] Figure 2 The automated intelligent defect analysis system based on UAV image acquisition of power transmission and distribution lines can perform Figure 1 The implementation principles and technical effects of the automated intelligent defect analysis method for power transmission and distribution lines based on drone image acquisition, as described in the illustrated embodiment, will not be elaborated upon. The specific manner in which each module and unit performs operations in the automated intelligent defect analysis system for power transmission and distribution lines based on drone image acquisition, as described in the aforementioned embodiment, has been described in detail in the relevant embodiments of the method and will not be elaborated upon here.

[0188] In one possible design, Figure 2The embodiment shown is an automated intelligent defect analysis system based on UAV image acquisition of power transmission and distribution lines, which can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0189] 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 .

[0190] The processing component 32 is used for the above Figure 1 The embodiment provides an automated intelligent defect analysis method based on drone image acquisition of power transmission and distribution lines.

[0191] 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 method. Of course, the processing component may also be implemented as 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 to perform the above method.

[0192] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile memory 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.

[0193] Of course, a computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0194] The input / output interface provides an interface between the processing component and the peripheral interface module, which can be an output device, an input device, etc.

[0195] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0196] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0197] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The embodiment shown is an automated intelligent defect analysis method based on drone image acquisition of power transmission and distribution lines.

[0198] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0199] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0200] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion 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, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0201] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An automated intelligent defect analysis method based on UAV image acquisition of power transmission and distribution lines, characterized in that: include: Use drones to collect infrared thermal imaging time series data and multispectral imaging data of the conductor surface of power transmission and distribution lines within continuous time windows; Calculating the temperature gradient direction of the main area of the wire 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 locating the boundary contour of the wire surface oxidation area based on the spectral absorption attenuation curve corresponding to the wire surface oxide layer in the multispectral imaging data; Performing spatial topological matching on 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; Establishing a dynamic association rule between the wire overheating defect and the surface oxide layer state based on the change in the angle 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; Based on the dynamic association rule, the local overheating defect level caused by oxide layer degradation on the surface of the wire is determined according to the critical angle threshold of the angle change and the number threshold of the mutation points, and a corresponding defect analysis report is generated according to the local overheating defect level.

2. The method according to claim 1, characterized in that The method of establishing a dynamic association rule between the wire overheating defect and the surface oxide layer state based on the change in the angle 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 includes: In the visual association map, continuous detection intervals are divided along the extension direction of the hot spot diffusion path with the wire axis as the reference axis, and the change in the angle between the hot spot diffusion path and the wire axis within the continuous detection interval is quantified to generate an angle change sequence; Divide the coverage area of the boundary contour into detection units of equal area, count the number of jumps of the spectral absorption attenuation rate from a stable state to a sudden change in the detection unit, and generate a jump number distribution graph for each detection unit; Marking a section in the angle change sequence where the angle increment between adjacent detection intervals exceeds a preset increment as an abnormal diffusion section, and marking a detection unit in the jump number distribution diagram where the jump number exceeds a preset number as an active degraded unit; According to the coverage overlap between the spatial position of the abnormal diffusion section and the spatial density distribution of the active degradation unit, a linkage judgment rule between the angle change of the hot spot diffusion path and the number of jumps of the spectral absorption attenuation rate is set as a dynamic association rule.

3. The method according to claim 1, characterized in that The step of calculating the temperature gradient direction of the main conductor area in each frame of the infrared thermal imaging time series data, extracting an abnormal hot spot area consistent with the conductor axis from the temperature gradient direction, and locating the boundary contour of the conductor surface oxidation area based on a spectral absorption attenuation curve corresponding to the conductor surface oxide layer in the multispectral imaging data, includes: In the conductor main body area of each frame of the infrared thermal imaging time series data, temperature detection zones are divided into equal intervals along the axial extension direction of the conductor, and temperature sampling points are arranged in the temperature detection zones along a direction perpendicular to the axial direction of the conductor; Within the temperature detection zone, calculating a temperature difference between a temperature sampling point and an adjacent sampling point, generating a temperature gradient direction along an arrangement direction of the temperature sampling points based on the temperature difference, generating a directional distribution histogram of the temperature gradient direction, and screening a temperature gradient direction in the directional distribution histogram whose angle with the conductor axis is less than a preset angle; The screened temperature gradient directions are spatially clustered in the main conductor area, and areas showing continuous distribution and having temperature gradient direction consistency exceeding a minimum ratio in the spatial clustering results are merged and marked as abnormal hot spot areas; In the multispectral imaging data, spectral sampling bands are divided along the surface contour line of the conductor, and slope change points of the spectral absorption attenuation curve corresponding to the conductor surface oxide layer are extracted within the spectral sampling bands. The slope change points are connected to sampling band boundaries where the length of the curve segment between adjacent slope change points is less than a preset distance, thereby forming a closed boundary contour of the conductor 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 correlation map integrating the hot spot diffusion path and the oxidation area distribution includes: Generating a hot spot diffusion path based on the temporal position change path of the abnormal hot spot area, dividing the hot spot diffusion path into equal-length path segments along the axial extension direction of the wire, and extracting a migration direction line of the abnormal hot spot area within the equal-length path segments, wherein the migration direction line is a central extension trajectory of the hot spot diffusion path within the equal-length path segments; Dividing the detection grids into equal widths along the boundary contour, calculating the distribution density of oxidation areas within the detection grids, and generating an oxidation area density distribution map; The migration direction line is spatially superimposed with the oxidation area density distribution map, and the extension trend of the migration direction line and the dynamic change trend of the oxidation area density distribution in the corresponding grid are analyzed based on the spatial superposition result. The path segment where the extension trend direction is consistent with the direction of the density increase area is marked as the action section of heat and oxidation; The extension angle change rate of the migration direction line is calculated, and the density change direction of the oxidation area density distribution map is extracted at the same time. The spatial consistency analysis of the extension angle change rate and the density change direction is performed within the action section, and the analysis results are converted into a single visualization layer to generate a visualization correlation map.

5. The method according to claim 2, characterized in that The method of setting a linkage judgment rule between the angle change of the hot spot diffusion path and the number of jumps of the spectral absorption attenuation rate as a dynamic association rule based on the overlap between the spatial position of the abnormal diffusion section and the spatial density distribution of the active degradation unit includes: Dividing the conductor surface area corresponding to the abnormal diffusion section into equal-area analysis units, calculating the coverage overlap between the number of active degradation units in the equal-area analysis unit and the total number of grids, and marking the equal-area analysis unit as a thermal oxidation-related unit when the coverage overlap exceeds a preset overlap ratio; In the abnormal diffusion section corresponding to the thermal oxidation correlation unit, an angle increment sequence of adjacent detection intervals is extracted, and the correlation fluctuation amplitude between the coverage overlap and the angle increment sequence is calculated. When the correlation fluctuation amplitude exceeds the fluctuation upper limit, the detection section is marked as a significant diffusion section; In the significant diffusion section, calculating the product of the coverage overlap and the sum of the angle increments, and normalizing the product to obtain a thermal oxidation correlation index; Based on the thermal oxidation correlation index, a linkage judgment rule is established between the angle change of the hot spot diffusion path and the number of jumps of the spectral absorption attenuation rate. When the thermal oxidation correlation index exceeds a preset standard value, it is determined that there is a wire overheating defect caused by oxide layer degradation.

6. The method according to claim 3, characterized in that The method comprises: calculating a temperature difference between a temperature sampling point and an adjacent sampling point within the temperature detection zone, generating a temperature gradient direction along an arrangement direction of the temperature sampling points according to the temperature difference, generating a directional distribution histogram of the temperature gradient direction, and screening a temperature gradient direction in the directional distribution histogram whose angle with the conductor axis is less than a preset angle. In the temperature detection zone, the temperature difference between the temperature sampling point and the adjacent sampling points on the left and right is calculated, and the temperature gradient direction from low temperature to high temperature is generated according to the positive and negative signs and absolute values of the temperature difference; Taking the axial direction of the conductor as a reference direction, statistically analyzing the direction angle distribution of the temperature gradient direction, and dividing the direction angle distribution according to preset angle intervals to generate a direction distribution histogram; Screening out, from the direction distribution histogram, a temperature gradient direction whose angle with the axis of the conductor is less than a preset angle, wherein the preset angle is set according to a maximum allowable thermal diffusion offset during normal operation of the conductor; The spatial distribution and directional angle deviation of the screened temperature gradient direction are quantitatively analyzed, and the quantitative analysis includes statistics of intervals between adjacent vectors and cumulative distribution calculation of the directional angle deviation.

7. The method according to claim 1, characterized in that The method of determining the level of local overheating defects caused by oxide layer degradation on the surface of the conductor based on the dynamic association rule and the critical angle threshold of the angle change and the number threshold of the mutation points, and generating a corresponding defect analysis report based on the level of local overheating defects includes: By using the dynamic association rule, a range of a critical angle threshold value of the angle variation is set, wherein the range is determined according to a physical upper limit of an abnormal rate of axial heat diffusion of the conductor, and an interval of a threshold value of the number of mutation points is set, wherein the interval is determined according to a spectral response characteristic of the oxide layer degradation activity; Establishing a combination relationship table of the ranges and the intervals, wherein the combination relationship table defines matching intervals of critical angle thresholds and quantity thresholds corresponding to defects of different grades; According to the range and the interval, a section in the conductor surface detection area in which the angle variation and the number of mutation points fall within the corresponding threshold intervals is selected as a section to be determined, and the section to be determined is mapped according to the combination relationship table to obtain an initial defect level; The initial defect level is continuously corrected. When the defect levels of three consecutive segments are the same and the increasing trend of the angle change is consistent, the three segments are merged and the defect level is increased by one to output the local overheating defect level. A defect analysis report including the defect location distribution, level intensity and diffusion trend is generated based on the local overheating defect level.

8. An automated intelligent defect analysis system based on UAV image acquisition of power transmission and distribution lines, characterized by: include: An acquisition module is used to synchronously collect infrared thermal imaging time-series data and multispectral imaging data of the conductor surface within a continuous time window using an infrared thermal imager and a multispectral camera carried by a drone. The infrared thermal imaging time-series data covers the entire axial length of the conductor through the drone's flight trajectory, capturing the dynamic migration trajectory of the surface temperature field. The multispectral imaging data obtains the spectral absorption difference between the oxide layer on the conductor surface and normal metal in a selected band through the drone's hovering posture. a positioning module for calculating the temperature gradient direction of the conductor main body 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 conductor axis, and locating the boundary contour of the oxidized area on the conductor surface based on the spectral absorption attenuation curve corresponding to the oxide layer in the multispectral imaging data; A matching module is used to perform spatial topological matching between the migration direction of the abnormal hot spot area and the boundary contour to generate a visual correlation map that integrates the hot spot diffusion path and the oxidation area distribution; An establishment module is used to establish a dynamic association rule between the wire overheating defect and the surface oxide layer state based on the change in the angle 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; The discrimination module is used to determine the level of local overheating defects on the surface of the wire caused by degradation of the oxide layer according to the dynamic association rule, the critical angle threshold of the angle change amount and the threshold of the number of mutation points.

9. A computing device, characterized in that It includes 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 an automated intelligent defect analysis method based on drone 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 A computer program is stored, and when the computer program is executed by a computer, an automated intelligent defect analysis method based on UAV image acquisition of power transmission and distribution lines as described in any one of claims 1 to 7 is implemented.

Images