A method, system, device and medium for quantitative evaluation of insulator aging based on hyperspectrum

Abstract

The application discloses a kind of based on hyperspectral insulator aging quantitative evaluation method, system, equipment and medium, belong to insulator evaluation technical field, including: to the change sequence of each waveband spectral reflectance of hyperspectral image along axial direction is established, monotonic spectral attenuation component is extracted from change sequence, obtain aging monotonic spectral component graph;To the band ratio operation of multi-waveband spectral intensity of aging monotonic spectral component graph, obtain the stable band ratio set of pollution;According to the stable band ratio set of pollution, spectral ratio calculation is carried out to each pixel point, and the calculation result is aggregated according to umbrella skirt partition, and aging index distribution graph is obtained;The aging index of each umbrella skirt partition in aging index distribution graph is output, and the quantitative evaluation result of insulator aging is output.The application selects effective sensitive waveband by consistency test, so that aging component in spectral information is dominant, and the interference of random distribution of pollution on aging feature extraction is inhibited.

Description

Technical Field

[0001] This invention relates to the field of insulator evaluation technology, specifically to a method, system, equipment, and medium for quantitative evaluation of insulator aging based on hyperspectral imaging. Background Technology

[0002] Because insulators are exposed to complex service environments such as ultraviolet radiation, temperature and humidity changes, and air pollution for a long time, their silicone rubber shed material will undergo irreversible photo-oxidative degradation, resulting in aging and deterioration phenomena such as decreased hydrophobicity, microcrack propagation, and surface roughening.

[0003] Existing hyperspectral-based insulator aging assessment methods cannot effectively separate the coupling interference between surface contamination deposition and material aging on spectral reflectance.

[0004] Insulator surfaces often exhibit both spectral degradation caused by aging and spectral absorption caused by contamination. These two factors overlap in their spectral responses, making it difficult for existing methods to accurately extract pure aging characteristic information from the mixed spectral signals.

[0005] When a thick layer of contaminant accumulates in a certain area of ​​an insulator's skirt due to geographical location or wind direction, its broadband spectral reflectance will decrease as a whole due to the contaminant absorption effect. However, existing methods suppress this spectrum and misjudge it as a characteristic response of aging, thus outputting an artificially high aging assessment result for that area. However, another skirt that has actually undergone deep photo-oxidative degradation but has a relatively clean surface is underestimated because it lacks the reinforcing effect of contaminant superposition. Ultimately, this leads to a serious distortion of the aging distribution map of the entire insulator, making it impossible to provide a reliable basis for operation and maintenance decisions. Summary of the Invention

[0006] In view of the above-mentioned problems, the present invention provides a method, system, device and medium for quantitative assessment of insulator aging based on hyperspectral imaging.

[0007] Therefore, the technical problem solved by this invention is that existing hyperspectral-based insulator aging assessment methods cannot effectively separate the coupling interference between surface contamination deposition and material aging on spectral reflectance.

[0008] To address the aforementioned technical problems, this invention provides the following technical solution: a method for quantitatively assessing insulator aging based on hyperspectral imaging, comprising the following steps: establishing a variation sequence of spectral reflectance along the axial direction for each band of a hyperspectral image; extracting a monotonically decaying component from the variation sequence to obtain a first aging component map; performing a first-level calculation on the multi-band spectral intensity of the first aging component map; filtering the results of the first-level calculation to obtain a first stable band set; performing a second-level calculation on each pixel based on the first stable band set; aggregating the results of the second-level calculation to obtain an aging index distribution map; and quantitatively assigning a value to the degree of aging using the aging index in the aging index distribution map, and outputting a quantitative assessment result of insulator aging.

[0009] As a preferred embodiment of the quantitative assessment method for insulator aging based on hyperspectral imaging as described in this invention, the step of establishing a variation sequence along the axial direction includes: performing a directional consistency test on the spectral reflectance of each band arranged along the axial direction; statistically analyzing the proportion of the spectral reflectance that meets the preset direction, and determining the bands with a proportion not lower than a preset threshold as effective sensitive bands; and constructing the variation sequence using the spectral reflectance of the effective sensitive bands arranged along the axial direction.

[0010] As a preferred embodiment of the hyperspectral-based quantitative assessment method for insulator aging described in this invention, the step of screening the results of the first-level calculation includes: grouping two pixels of the insulator that are adjacent in the same axial height and circumferential direction into a pixel pair; calculating the difference in the spectral ratio between the two pixels under each candidate band combination using the pixel pair, and determining the band combination with the smallest difference as the first stable band set.

[0011] As a preferred embodiment of the hyperspectral-based quantitative assessment method for insulator aging described in this invention, the step of obtaining the aging index distribution map includes: using the average spectral ratio of the corresponding region pixels of the reference insulator under the first stable band set as the first reference value; performing a difference operation between the spectral ratio calculation results of each pixel and the first reference value to obtain a first correction value for each pixel; and aggregating the first correction values ​​according to the skirt partition to obtain the aging index distribution map.

[0012] As a preferred embodiment of the hyperspectral-based quantitative assessment method for insulator aging described in this invention, the step of calculating by aggregating the values ​​according to the skirt partitions includes: statistically calculating the spectral reflectance of pixels at each axial position in the hyperspectral image to obtain a first spectral curve distributed along the axial direction; calculating the gradient of the first spectral curve and determining the axial position where the gradient value meets a preset condition as the skirt boundary; assigning the first correction value of each pixel to the corresponding skirt partition according to the skirt boundary, and aggregating the first correction values ​​within the same skirt partition to obtain the aging index distribution map.

[0013] As a preferred embodiment of the hyperspectral-based insulator aging quantitative assessment method of the present invention, the step of quantitatively assigning a value to the aging degree includes: subtracting the aging indices of two adjacent skirt sections in the aging index distribution map to obtain a first difference value, wherein the first difference value characterizes the degree of spatial non-uniformity of the aging progress of the insulator; and combining the first difference value with the aging degree assignment results of each skirt section to constitute the insulator aging quantitative assessment result.

[0014] As a preferred embodiment of the hyperspectral-based insulator aging quantitative assessment method of the present invention, the step of outputting the insulator aging quantitative assessment result includes: extracting a unique aging index corresponding to each skirt partition from the aging index distribution map; arranging all the extracted aging indices sequentially according to the axial arrangement order of each skirt partition on the insulator from the high-voltage end to the low-voltage end to generate a one-dimensional axial aging index sequence; initializing the maximum value variable to the first element of the axial aging index sequence and initializing the maximum value index variable to 0; starting from the second element of the axial aging index sequence, comparing the current element with the maximum value variable sequentially; in response to the value of the current element being greater than the value of the maximum value variable, updating the maximum value variable to the value of the current element and simultaneously updating the maximum value index variable to the sequence index corresponding to the current element; after traversing all elements of the axial aging index sequence, querying the axial pixel coordinate range of the corresponding skirt partition in the hyperspectral image of the insulator according to the maximum value index variable; determining the center point coordinate of the axial pixel coordinate range as the aging concentration area location; and outputting the aging concentration area location together with the insulator aging quantitative assessment result.

[0015] This invention provides a quantitative assessment system for insulator aging based on hyperspectral imaging.

[0016] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a quantitative assessment system for insulator aging based on hyperspectral imaging, comprising: The image extraction module is used to establish a change sequence of spectral reflectance of each band of the hyperspectral image along the axis, and extract the monotonic attenuation component from the change sequence to obtain the first aging component map; The ratio calculation module performs a first-level calculation on the multi-band spectral intensity of the first aging component map, filters the results of the first-level calculation to obtain a first stable band set, performs a second-level calculation on each pixel based on the first stable band set, and aggregates the results of the second-level calculation to obtain an aging index distribution map. The results output module assigns a quantitative value to the degree of aging based on the aging index in the aging index distribution chart and outputs the quantitative assessment result of insulator aging.

[0017] The present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the aforementioned method for quantitative assessment of insulator aging based on hyperspectral imaging.

[0018] The present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the aforementioned method for quantitative assessment of insulator aging based on hyperspectral imaging.

[0019] The beneficial effects of this invention are as follows: By using the screening mechanism of the first stable band set, the difference in the spectral ratio between adjacent pixel pairs at the same axial height is used to quantify the sensitivity of each candidate band combination to contamination. The band combination least sensitive to contamination distribution is selected to construct the aging index, so that the final aging assessment result can remain stable when the contamination thickness changes laterally. This solves the problem of misjudgment caused by the coupling of contamination and aging spectrum in existing methods.

[0020] By performing gradient calculation on the first spectral curve to identify the skirt boundary, and aggregating the first correction value according to the skirt partition, an aging index distribution map is obtained. By combining the traversal comparison of the aging index distribution map to locate the location of the aging concentration area, a complete quantitative characterization from pixel-level spectral information to the overall aging spatial distribution of the insulator can be achieved, providing maintenance personnel with skirt-by-skirt aging assessment results and aging concentration locations. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1This is an overall timeline diagram of a hyperspectral-based quantitative assessment method for insulator aging, provided as an embodiment of the present invention.

[0023] Figure 2 The flowchart of the aging monotonic spectral component diagram of a method for quantitative assessment of insulator aging based on hyperspectral imaging, provided as an embodiment of the present invention, is shown.

[0024] Figure 3 This invention provides an application scenario for a quantitative insulator aging assessment system based on a hyperspectral method, which is an embodiment of the present invention. Detailed Implementation

[0025] To make the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0026] Example 1, referring to Figure 1 and Figure 3 This is one embodiment of the present invention, which provides a method for quantitative assessment of insulator aging based on hyperspectral imaging, comprising the following steps: S1. Establish a variation sequence of spectral reflectance of each band of the hyperspectral image along the axis, and extract the monotonic attenuation component from the variation sequence to obtain the first aging component map.

[0027] S2. Perform a first-level calculation on the multi-band spectral intensity of the first aging component map, and filter the results of the first-level calculation to obtain the first stable band set.

[0028] S3. Perform secondary calculations on each pixel based on the first stable band set, and aggregate the results of the secondary calculations to obtain the aging index distribution map.

[0029] S4. Quantitatively assign values ​​to the degree of aging using the aging index in the aging index distribution chart, and output the quantitative assessment results of insulator aging.

[0030] like Figure 3 As shown, in actual service scenarios, the surface of an insulator often exhibits both spectral degradation caused by aging and spectral absorption caused by contamination. These two factors overlap in the spectral response, making it difficult for existing methods to accurately extract pure aging feature information from the mixed spectral signal.

[0031] When a thick layer of contaminant accumulates in a certain area of ​​an insulator's skirt due to geographical location or wind direction, its broadband spectral reflectance will decrease as a whole due to the contaminant absorption effect. The commonly used method is to suppress this spectrum and misjudge it as a characteristic response of severe aging, thus outputting an artificially high aging assessment result for that area. At the same time, another skirt that has actually undergone deep photo-oxidative degradation but has a relatively clean surface is underestimated because it lacks the reinforcing effect of contaminant superposition, ultimately leading to a serious distortion of the aging distribution map of the entire insulator.

[0032] According to steps S1 to S4, through the screening mechanism of the first stable band set, the difference in the spectral ratio between adjacent pixel pairs at the same axial height is used to quantify the sensitivity of each candidate band combination to contamination. The band combination least sensitive to contamination distribution is selected to construct the aging index, so that the final aging assessment result can remain stable when the contamination thickness changes laterally. This solves the problem of misjudgment caused by the coupling of contamination and aging spectrum in existing methods.

[0033] By performing gradient calculation on the first spectral curve to identify the skirt boundary, and aggregating the first correction value according to the skirt partition, an aging index distribution map is obtained. By combining the traversal comparison of the aging index distribution map to locate the location of the aging concentration area, a complete quantitative characterization from pixel-level spectral information to the overall aging spatial distribution of the insulator can be achieved, providing maintenance personnel with skirt-by-skirt aging assessment results and aging concentration locations.

[0034] Example 2, refer to Figures 1-3 As an embodiment of the present invention, based on the previous embodiment, a method for quantitative assessment of insulator aging based on hyperspectral imaging is provided, comprising the following steps: like Figure 2 As shown, S1, establish a variation sequence of spectral reflectance of each band of the hyperspectral image along the axis, and extract the monotonic attenuation component from the variation sequence to obtain the first aging component map.

[0035] In this embodiment, the step of establishing a variation sequence along the axial direction includes S1.1 to S1.3: S1.1 Perform a directional consistency check on the spectral reflectance of each band arranged along the axial direction.

[0036] For the λ-th band in the hyperspectral data cube, the spectral reflectance of each pixel is first arranged along the axial direction to obtain the axial spectral reflectance sequence of that band.

[0037] The steps of consistency verification include: Compare each adjacent element in the sequence one by one to determine whether the first element is greater than the second element, that is, whether the reflectance value of the later position is less than the reflectance value of the previous position. Adjacent element pairs that meet this condition are recorded as a monotonically decreasing response event with the same direction.

[0038] By traversing all adjacent element pairs in the sequence, the number of response events that satisfy the monotonically decreasing direction is counted, thereby quantitatively assessing the degree to which the band exhibits an overall monotonically decreasing trend in the axial direction.

[0039] In this embodiment, for example, a hyperspectral system collected 200 bands, and there are 100 pixel positions in the axial direction of the insulator.

[0040] Taking the 85th band as an example, the axial reflectivity sequence contains 99 pairs of adjacent elements consisting of 100 adjacent elements. After comparing them one by one, it was found that 82 of the pairs of elements met the monotonically decreasing condition, while the remaining 17 pairs of elements showed a local increase in reflectivity due to local pollution or noise interference.

[0041] The same test was performed on band 120, and it was found that only 41 of its 99 adjacent element pairs met the monotonically decreasing condition. This means that the axial reflectivity of this band does not have a monotonically decreasing law and does not carry effective aging information.

[0042] S1.2 Statistical analysis of the proportion of spectral reflectance that meets the preset direction, and determination of the bands with a proportion not lower than the preset threshold as effective sensitive bands.

[0043] For each band, calculate the ratio of the number of adjacent element pairs that satisfy the monotonically decreasing direction to the total number of all adjacent element pairs, which is the reflectance value ratio.

[0044] The reflectance value ratio is compared with a preset threshold. In this embodiment, the preset threshold is 0.75, which means that at least 75% of the adjacent pixel pairs in the axial reflectance sequence of a certain band must show a decreasing trend in order to determine that the band carries sufficient intensity of aging monotonic attenuation information.

[0045] Specifically, for example, the monotonically decreasing proportion of band 85 is 82 / 99≈0.828, which is higher than the preset threshold of 0.75. Therefore, band 85 is identified as an effective sensitive band.

[0046] The monotonically decreasing proportion of band 120 is 41 / 99≈0.414, which is lower than the preset threshold, so band 120 is removed.

[0047] S1.3 Construct the variation sequence using the spectral reflectance of the effective sensitive bands arranged along the axial direction.

[0048] For example, there are 143 effective sensitive bands. From the original 200-band hyperspectral data cube, the data layers corresponding to these 143 bands are extracted according to the band index and combined to form a first aging component map with the same spatial resolution as the original image and the spectral dimension compressed from 200 to 143. In this embodiment, the first aging component map is an aging monotonic spectral component map.

[0049] In the aging monotonic spectral component diagram, taking band 85 as an example, the corresponding data layer shows an overall decreasing distribution along the axial direction, from approximately 0.62 reflectance at pixel 1 (high voltage end) to approximately 0.38 reflectance at pixel 100 (low voltage end). This reflects the spatial gradient of the insulator aging degree along the axial direction in this band. The 120th band, which was originally excluded, does not appear in the aging monotonic spectral component diagram because its axial reflectance fluctuates irregularly and does not carry effective aging information.

[0050] S2. Perform a first-level calculation on the multi-band spectral intensity of the first aging component map, and filter the results of the first-level calculation to obtain the first stable band set.

[0051] The steps for filtering the results of the first-level operation include S2.1 to S2.2: S2.1. Group two pixels that are circumferentially adjacent at the same axial height of the insulator into a pixel pair.

[0052] The insulator has an axisymmetric rotational structure, and the pixel positions are arranged in the circumferential direction, which is along the circumference of the insulator. Circumferentially adjacent pixels at the same axial height mean that they are on the same cross-sectional ring of the insulator, have the same axial coordinates, and differ only in circumferential coordinates by one pixel unit.

[0053] When constructing an insulator, in the aging monotonic spectral component map, for each axial pixel position i, all circumferential pixels at that axial height are extracted, and two adjacent pixels with circumferential coordinates j and j+1 are combined into a pixel pair. By traversing all axial and circumferential positions, a complete set of pixel pairs is obtained.

[0054] For example, the spatial size of the aging monotonic spectral component map is 100 (axial) × 60 (circumferential) pixels, with a total of 143 effective sensitive bands.

[0055] At axial position i=35 (corresponding to the middle area of ​​a certain skirt of the insulator), there are 60 pixels in the circumferential direction. The circumferential pixels (35,1) and (35,2), (35,2) and (35,3), ..., (35,59) and (35,60) are combined into a pixel pair in turn, and a total of 59 pixel pairs are generated at this axial position.

[0056] S2.2 Calculate the difference in spectral ratio between two pixels under each candidate band combination using the pixel pair, and determine the band combination with the smallest difference as the first stable band set. In this embodiment, the first stable band set is the pollution stable band ratio set.

[0057] In this step, candidate band combinations are generated by pairwise combinations of the effective sensitive bands determined in step S1.2. Each candidate band combination consists of a numerator band and a denominator band, and the corresponding spectral ratio is defined as the reflectance ratio of a certain pixel point under that combination.

[0058] For each pixel pair in the pixel pair set, calculate the difference in the spectral ratio of the two pixels under the candidate band combination. Since the aging state of adjacent pixels in the same axial height tends to be consistent, the aging contribution in the difference cancels each other out, and the residual difference is mainly caused by the difference in dirt thickness between the two pixels.

[0059] The average difference between all pixel pairs is used to obtain the pollution sensitivity quantification index of the candidate band combination.

[0060] The pollution sensitivity index was calculated for all candidate band combinations. The band combination with the smallest pollution sensitivity index was determined as the pollution stable band ratio set, that is, the spectral ratio under this combination fluctuates the least when the pollution distribution on the insulator surface changes laterally. In this embodiment, there are a total of 143 effective sensitive bands, a total of 10,153 candidate band combinations, and a total of 5,900 pixel pairs in the pixel pair set.

[0061] With candidate combinations (λ) 85 ,λ 102 For example, for pixel pairs (p) 35,22 ,p 35,23 Calculate the spectral ratio difference: Assume p 35,22 The thick layer of dirt caused R 85 =0.31、R 102 =0.27, R is the reflectance value, and the ratio is 0.31 / 0.27≈1.148; p 35,23 The dirt is relatively thin, R 85 =0.38, R 102 =0.33, the ratio is 0.38 / 0.33≈1.152; The difference between the two is |1.148-1.152|=0.004, indicating that even if there is a difference in the dirt thickness between the two pixels, the ratio under this band combination is still very close.

[0062] The difference between all 5900 pixel pairs was calculated and the average was taken, resulting in a value of 0.006.

[0063] Then another candidate combination (λ) 85 ,λ 67 For example, the difference was 0.031 after the same calculation, indicating that the combination is more sensitive to horizontal changes in filth and the ratio fluctuates more.

[0064] After traversing all 10153 candidate combinations, assuming that the minimum value of 0.006 is obtained among all candidate combinations, the band combination (λ) is... 85 ,λ 102 It was determined to be the set of ratio values ​​for the stable pollutant bands.

[0065] In another alternative implementation, the first-level calculation can also be a band difference calculation.

[0066] Specifically, the spectral intensities of the two bands under each candidate band combination in the first aging component image are subtracted to obtain the band difference result. Subsequently, for the pixel pairs formed in the screening step, the difference of the band difference between two circumferentially adjacent pixels at the same axial height under each candidate band combination is calculated. The mean of the differences of all pixel pairs is used as the quantification index of the dirt sensitivity of the candidate band combination, and the band combination with the smallest index is determined as the first stable band set.

[0067] For example, candidate combinations (λ) are selected from the effective sensitive bands. 85 , λ 102 ), for pixel pairs (p 35,22 p 35,23 Calculate the band difference: p 35,22 R 85 =0.31、R 102 =0.27, the band difference is 0.31-0.27=0.04; p 35,23 R 85 =0.38, R 102 =0.33, the band difference is 0.38-0.33=0.05; The difference between the two is |0.04-0.05|=0.01.

[0068] The above differences are calculated for all pixel pairs, and the average value is taken to obtain the pollution sensitivity quantification index of the candidate combination. After traversing all candidate combinations, the band combination with the smallest index is determined as the first stable band set.

[0069] Compared to band ratio calculation, band difference calculation can also offset aging contributions and preserve dirt difference information through pixel pair mechanism, thereby selecting the band combination least sensitive to lateral changes in dirt. S3. Perform secondary calculations on each pixel based on the first stable band set, and aggregate the results of the secondary calculations to obtain the aging index distribution map.

[0070] The steps for obtaining the aging index distribution map include S3.1 to S3.3: S3.1. The average spectral ratio of the corresponding pixel points of the reference insulator in the first stable band set is used as the first reference value. Specifically, the first reference value is the spectral reference value without aging.

[0071] In this step, the establishment of the aging-free spectral reference value is achieved using a brand-new reference insulator sample that is identical in model and material to the insulator under test, or one that has been verified in the laboratory to be in an aging-free state. The reference insulator is subjected to hyperspectral scanning under the same imaging conditions to acquire its hyperspectral data cube. Based on the pollution stability band ratio set determined in step S2, the spectral ratio is calculated for each pixel in the region corresponding to the insulator under test in the hyperspectral image of the reference insulator at the axial position.

[0072] The average spectral ratio of all pixels within the corresponding region of the reference insulator is taken to obtain the aging-free spectral reference value for that region.

[0073] Specifically, the set of ratios for polluted stable bands is (λ 85 ,λ 102 A brand-new reference insulator of the same model as the insulator to be tested is selected, and a hyperspectral scan is performed under the same illumination conditions to obtain its hyperspectral data cube.

[0074] Extract all pixels corresponding to the axial position range of the reference insulator and the complete coverage area of ​​the insulator under test. Assuming that the area contains a total of 5800 pixels, calculate the reflectance ratio of band 85 to band 102 for each pixel. The obtained ratios are concentrated between 1.148 and 1.156. Take the average value to obtain the aging-free spectral reference value of 1.152.

[0075] This benchmark indicates that when the insulator is in a non-aging state, the expected level of the spectral ratio of each pixel is approximately 1.152 under the pollution stable band ratio set.

[0076] S3.2. Perform a difference operation between the calculated spectral ratio of each pixel and the first reference value to obtain the first correction value of each pixel. That is, subtract the aging-free spectral reference value from the calculated spectral ratio to obtain the corrected spectral ratio of each pixel.

[0077] In this step, the spectral ratios of all spatial pixels in the aging monotonic spectral component map of the insulator under test are calculated one by one according to the pollution stable band ratio set, resulting in the original spectral ratio map of each pixel of the insulator under test. Subsequently, the original spectral ratio of each pixel is subtracted from the aging-free spectral reference value calculated in step S3.1 to obtain the corrected spectral ratio of that pixel.

[0078] The deviation of the spectral ratio of the insulator material at the current pixel from its initial state without aging. Since aging causes directional degradation of reflectivity in specific bands, the deeper the aging, the greater the deviation of the spectral ratio from the unaged baseline. Therefore, the absolute value of the corrected spectral ratio directly corresponds to the relative aging degree of the material at that pixel.

[0079] For regions that have not yet shown significant aging, the corrected spectral ratio approaches zero; For regions with a high degree of aging, the corrected spectral ratio will show positive or negative deviations.

[0080] For example, the baseline value for the aging-free spectrum is 1.152. The reflectance ratio of band 85 to band 102 is calculated for each of the 6000 pixels (100 × 60 = 6000 pixels) in the monotonic spectral component diagram of the insulator under test. Taking a pixel p at axial position i = 10, i.e., near the high-voltage end, which is severely aged... 10,30 For example, its R 85 =0.29, R 102 =0.23, the original spectral ratio is 0.29 / 0.23≈1.261, and the corrected spectral ratio is 1.261-1.152=0.109, indicating that the deviation of the spectral ratio caused by material aging at this pixel is 0.109.

[0081] After all 6000 pixels have been calculated, a complete corrected spectral ratio spatial distribution map of the insulator under test is obtained.

[0082] S3.3. Aggregate the first correction value according to the umbrella skirt partition to obtain the aging index distribution map.

[0083] The steps for calculating the aggregation results based on the umbrella skirt partitions include A1~A3: A1. Statistically calculate the spectral reflectance of pixels at each axial position in the hyperspectral image to obtain a first spectral curve distributed along the axial direction. In this embodiment, the statistical calculation adopts the mean calculation, and the obtained first spectral curve is the average spectral reflectance curve.

[0084] In this step, for each axial position i in the original hyperspectral image, all reflectance values ​​of all circumferential pixels at that axial position are extracted across all acquisition bands, and the overall average is taken to obtain the average spectral reflectance corresponding to that axial position.

[0085] By traversing all N axial positions, a one-dimensional average spectral reflectance curve of length N is obtained.

[0086] In this embodiment, the original hyperspectral image has a spatial size of 100×60 pixels and a total of 200 bands. At the axial position i=20, all 60×200=12000 reflectance values ​​of 60 circumferential pixels in the 200 bands are extracted, and the average value is 0.412.

[0087] Performing the same calculation at axial position i=21 yields 0.438, and at i=22 yields 0.451. By repeating this process for all 100 axial positions, the complete average spectral reflectance curve is obtained.

[0088] A2. Perform gradient calculation on the first spectral curve, and determine the axial position where the gradient value meets the preset condition as the umbrella skirt boundary; In this embodiment, the preset condition is that the gradient value appears as a local maximum.

[0089] The first-order gradient of the average spectral reflectance curve along the axial direction is calculated to obtain a gradient sequence, where the gradient value at each position is defined as the absolute value of the reflectance difference between adjacent axial positions.

[0090] At the edge of the umbrella skirt, due to the geometric transformation of the insulator from the umbrella skirt disc to the narrow neck or from the narrow neck to the next level umbrella skirt disc, the light reception conditions and surface orientation change abruptly. This results in the average reflectivity difference between adjacent axial positions at this location being greater than that in the gradually changing region inside the umbrella skirt, which manifests as a local maximum in the gradient sequence.

[0091] The criteria for determining local maxima are: the gradient sequence G(i) is greater than its two adjacent gradient values ​​G(i-1) and G(i+1), and the gradient sequence G(i) exceeds the threshold formed by the sum of the global mean and standard deviation of the gradient sequence, so as to exclude the case where small gradient fluctuations caused by noise are misjudged as umbrella skirt boundaries.

[0092] All axial positions that meet the conditions constitute the set of umbrella skirt boundary positions.

[0093] Specifically, the average spectral reflectance curve has a length of 100. Gradient calculation is performed on the curve to obtain a gradient sequence of length 99.

[0094] Assume the global mean of the gradient sequence is 0.018, the standard deviation is 0.009, and the threshold is 0.018 + 0.009 = 0.027. Checking the gradient sequence, we find that at the axial position i = 15, G(15) = 0.043, which is greater than G(14) = 0.011 and G(16) = 0.008, and exceeds the threshold of 0.027, so it is determined to be the umbrella skirt boundary.

[0095] A3. Based on the umbrella skirt boundary, the first correction value of each pixel is assigned to the corresponding umbrella skirt partition, and the first correction values ​​within the same umbrella skirt partition are aggregated to obtain the aging index distribution map.

[0096] Using the set of umbrella skirt boundary positions determined in step A2, all pixels in the image of the insulator under test are assigned to the corresponding umbrella skirt partitions according to their axial coordinates. For the k-th umbrella skirt partition, the set of pixels it contains is [formula missing]. The aging index of the umbrella skirt partition is obtained by taking the average of the corrected spectral ratios of all pixels within the same umbrella skirt partition.

[0097] Aggregation calculations are performed sequentially on all K+1 umbrella skirt partitions to obtain an ordered set of aging indices for each partition. An aging index distribution map is constructed using the umbrella skirt partition as the spatial unit and the aging index as the attribute value.

[0098] In this embodiment, the umbrella skirt boundary set B={15,31,47,63,79} is divided into 6 umbrella skirt partitions. Taking partition 1 (i=1~15, near the high voltage end) as an example, this partition contains 15×60=900 pixels. The corrected spectral ratio of these 900 pixels is extracted, and the average value is 0.103. Taking partition 6 (i=80~100, near the low voltage end) as an example, it contains a total of 21×60=1260 pixels, with an average value of 0.007.

[0099] Aggregation was performed on all 6 umbrella skirt partitions to obtain the aging index sequence {0.103,0.081,0.059,0.038,0.021,0.007}, which constitutes the aging index distribution map.

[0100] S4. Quantitatively assign values ​​to the degree of aging using the aging index in the aging index distribution chart, and output the quantitative assessment results of insulator aging.

[0101] Specifically, the steps for quantitatively assigning values ​​to the aging degree of each umbrella skirt section include S4.1~S4.2: S4.1. Subtract the aging indices of two adjacent skirt sections in the aging index distribution map to obtain a first difference value, wherein the first difference value characterizes the degree of spatial unevenness of the aging progress of the insulator.

[0102] Using the aging index distribution map output in step S3, extract the ordered set of aging indices for each skirt section. For two axially adjacent skirt sections k and k+1, calculate the difference in their aging indices to obtain the first difference between adjacent skirt sections. This first difference is the aging index difference.

[0103] Since the degree of aging generally decreases from the high-pressure end to the low-pressure end along the axial direction, the difference in aging index should be positive under normal aging gradient distribution. When the difference in aging index is large, it indicates that there is a large leap in the degree of aging between two adjacent umbrella skirt sections, and there is an abrupt interface of aging degree near this location; When the aging index difference value approaches zero, it indicates that the aging degree of two adjacent skirt sections is similar, and the aging distribution in this axial region is relatively uniform. The difference calculation is performed sequentially on all K adjacent skirt sections to obtain the aging index difference sequence.

[0104] Specifically, the aging index sequence for the six umbrella skirt sections in the aging index distribution map is {0.103, 0.081, 0.059, 0.038, 0.021, 0.007}. Subtracting axially adjacent sections sequentially yields the aging index difference sequence, which is as follows: D1=0.103-0.081=0.022, D2=0.081-0.059=0.022, D3=0.059-0.038=0.021, D4=0.038-0.021=0.017, D5=0.021-0.007=0.014, and the complete aging index difference sequence is {0.022,0.022,0.021,0.017,0.014}.

[0105] The difference sequence shows that the adjacent differences between partitions 1 and 3 are large and close, indicating that the axial gradient of aging degree between the umbrella skirt partitions near the high-pressure end is steep. The small difference between adjacent values ​​in partitions 4 and 6 indicates that the axial gradient of aging degree between the various umbrella skirt partitions near the low-pressure end tends to be gentle, and the aging progress has been relatively weakened.

[0106] The aging index difference sequence characterizes the spatial non-uniformity of the aging progression of insulators along the axial direction from two dimensions.

[0107] When characterizing the overall dispersion of the difference sequence, the standard deviation of the difference sequence is calculated. The larger the standard deviation, the more uneven the distribution of the aging gradient between adjacent partitions, that is, the aging rate along the axial direction varies significantly at different locations, and there are local rapid degradation regions. A smaller standard deviation indicates that the aging rate along the axial direction is more consistent and the spatial uniformity of the aging distribution is better.

[0108] When characterizing the location of local maxima in the difference sequence, when a certain aging index difference is greater than the aging index difference of its axially adjacent region, it indicates that there is a local abrupt change in the degree of aging between the k-th and k+1-th umbrella skirt sections.

[0109] The standard deviation of the difference sequence and the values ​​of each element in the difference sequence are incorporated into the evaluation results as a quantitative characterization index of the unevenness of the spatial distribution of aging. This complements the absolute aging index of each skirt zone and together constitutes a comprehensive description of the aging state of the insulator.

[0110] S4.2 The first difference and the aging degree assignment results of each skirt zone are combined to form the quantitative assessment result of the insulator aging.

[0111] The aging degree assignment result of each umbrella skirt section is directly mapped from the aging index of that section. Specifically, based on the pre-established aging level classification standard, the aging index is mapped to the corresponding aging degree level. The aging level classification standard is determined by statistical analysis of a large amount of measured data of the same type of insulator under different service years and environmental conditions, forming a correspondence table between the aging index and the aging degree level.

[0112] In this embodiment, the degree of aging is divided, specifically as follows: An aging index of <0.02 corresponds to a normal state and is denoted as Level I. A score of 0.02 ≤ aging index < 0.05 corresponds to mild aging and is designated as Level II. A aging index of 0.05 ≤ aging index < 0.09 corresponds to moderate aging and is designated as Level III. An aging index ≥ 0.09 corresponds to severe aging and is designated as Level IV.

[0113] After assigning aging degree levels to all K+1 umbrella skirt sections in sequence, the section number, corresponding aging index, aging degree level, aging index difference sequence obtained in step S4.1, and standard deviation of the difference sequence obtained in step S4.2 are integrated to obtain the quantitative assessment result of insulator aging.

[0114] The steps for outputting quantitative assessment results of insulator aging include B1~B8: B1. Extract the unique aging index corresponding to each umbrella skirt section from the aging index distribution map.

[0115] It is important to know that each umbrella skirt section corresponds to a unique scalar value in the aging index distribution map. This value is the aging index obtained by taking the average of the spectral ratios of all pixels in the section after correction in step A3, which represents a comprehensive quantitative characterization of the overall aging state of the umbrella skirt section.

[0116] The extraction process is carried out sequentially according to the umbrella skirt partition number to ensure that each partition corresponds to only one and only one aging index value. There is no situation where one partition corresponds to multiple values ​​or multiple partitions share the same value, thus ensuring the determinism and uniqueness of the subsequent sequence generation.

[0117] In this embodiment, there are 6 umbrella skirt partitions in the aging index distribution map. The unique aging index of each partition is extracted one by one: partition 1 corresponds to 0.103, partition 2 corresponds to 0.081, partition 3 corresponds to 0.059, partition 4 corresponds to 0.038, partition 5 corresponds to 0.021, and partition 6 corresponds to 0.007. A total of 6 unique aging index values ​​are extracted.

[0118] B2. According to the axial arrangement order of each skirt section on the insulator from the high voltage end to the low voltage end, arrange all the extracted aging indices in sequence to generate a one-dimensional axial aging index sequence.

[0119] Using the extracted aging indexes of each skirt section as elements, and following the physical arrangement order of each skirt section from the high-voltage end to the low-voltage end in the axial direction of the insulator, all K+1 aging index values ​​are arranged sequentially to form a one-dimensional axial aging index sequence.

[0120] In this embodiment, the sequence is arranged from the high-pressure end to the low-pressure end, that is, from partition 1 to partition 6, which can generate a one-dimensional axial aging index sequence, specifically represented as [0.103,0.081,0.059,0.038,0.021,0.007], with corresponding sequence indices of 0,1,2,3,4,5 respectively; where 0 corresponds to partition 1, that is, the high-pressure end, and 5 corresponds to partition 6, that is, the low-pressure end.

[0121] B3. Initialize the maximum value variable to the first element of the axial aging index sequence, and initialize the maximum value index variable to 0.

[0122] Before performing a linear traversal of the axial aging index sequence to find the maximum value, two state variables required for the traversal process are initialized. The maximum value variable is initialized to the first element of the axial aging index sequence; the maximum value index variable is initialized to 0.

[0123] The initialization operation ensures that when traversing from the second element onwards and comparing each element with the current known maximum value, there is an initial reference base, avoiding comparison logic errors caused by uninitialization. At the same time, it guarantees that even when the sequence contains only one element, it can still correctly return the maximum value and index corresponding to that element.

[0124] In this example, the axial aging index sequence is [0.103, 0.081, 0.059, 0.038, 0.021, 0.007], with the first element being 0.103. The maximum value variable is initialized to 0.103, and the maximum value index variable is initialized to 0. Both state variables are now initialized.

[0125] B4. Starting from the second element of the axial aging index sequence, compare the current element with the maximum value variable in sequence.

[0126] Iterate from the second element of the sequence, 0.081, that is, from k=1.

[0127] During the first comparison, the current element is 0.081, the maximum value variable is initialized to 0.103, and since 0.081 < 0.103, the update condition is not met, so the iteration continues.

[0128] During the second comparison, the traversal starts from k=2. The current element is 0.059, and the maximum value variable is initialized to 0.103. Since 0.059 < 0.103, the update condition is not met, so the traversal continues.

[0129] Then, the same comparison operation is performed on 0.038, 0.021, and 0.007 in sequence. None of them meet the update condition, so the traversal continues until the end of the sequence.

[0130] B5. In response to the fact that the value of the current element is greater than the value of the maximum value variable, update the maximum value variable to the value of the current element, and at the same time update the maximum value index variable to the sequence index corresponding to the current element.

[0131] In each comparison operation in step B4, if the value of the current element is greater than the value of the current maximum value variable, the state variable update is triggered.

[0132] Update the maximum value variable to the value of the current element; at the same time, update the maximum value index variable to the index of the current element in the sequence.

[0133] When the value of the current element is less than or equal to the maximum value variable, both state variables remain unchanged, and the traversal continues to the next element.

[0134] For example, as can be seen from the traversal process in step B4, all subsequent elements in the sequence [0.103,0.081,0.059,0.038,0.021,0.007] are less than the initial maximum value of 0.103, and the update condition in step B5 is never triggered during the traversal process.

[0135] After the traversal is complete, the maximum value of the axial aging index sequence is 0.103, corresponding to sequence index 0, that is, partition 1 is the umbrella skirt partition with the most severe aging.

[0136] For example, if the sequence is [0.059, 0.103, 0.081, 0.038, 0.021, 0.007], then when processing the element 0.103 at index k=1, 0.103>0.059, triggering an update, the final maximum value index variable is output as 1, and the corresponding partition 2 is the aging concentration area.

[0137] B6. After traversing all elements of the axial aging index sequence, query the axial pixel coordinate range of the corresponding skirt partition in the hyperspectral image of the insulator based on the maximum value index variable.

[0138] In this step, the maximum value index variable after the traversal in step B5 is used as the query key to query the correspondence table between the skirt partitions and the axial pixel coordinate ranges established in step A2, and obtain the axial pixel coordinate range spanned by the skirt partition with the most severe aging in the hyperspectral image of the insulator. ,in, , , b is the corresponding element in the set of umbrella skirt boundary positions determined in step A2. It is the maximum value index variable.

[0139] This axial pixel coordinate range defines the exact pixel coverage area of ​​the aging concentrated umbrella skirt partition in the hyperspectral image spatial coordinate system.

[0140] In this embodiment, the maximum value index variable is 0 after traversal, corresponding to umbrella skirt partition 1. The table of correspondence between umbrella skirt boundaries and partition coordinates established in step A2 is queried. The axial pixel coordinate range of partition 1 is [1, 15], meaning that this partition covers the axial pixel positions from row 1 to row 15 in the hyperspectral image, a total of 15 axial pixel positions, corresponding to the physical area of ​​the insulator from the starting position of the high-voltage end to the first umbrella skirt boundary.

[0141] B7. Determine the center point coordinates of the axial pixel coordinate range as the location of the aging concentration area.

[0142] The obtained axial pixel coordinate range Calculate the axial coordinates of the center point of this range in the axial direction, expressed by the formula: ; in, This indicates a floor operation, ensuring that the center point coordinates are integer pixel indices. The coordinates of the center point are axial coordinates.

[0143] In the circumferential direction, the circumferential coordinates of the aging concentration area are taken as the center of the circumferential pixel range of the insulator, i.e. Where W is the total number of pixels in the circumferential direction.

[0144] Finally, the location of the aging concentration area is determined using two-dimensional pixel coordinates. This indicates the geometric center of the skirt section with the most severe aging in the hyperspectral image. This coordinate can be directly mapped to the physical axial distance of the insulator, providing precise spatial positioning reference for on-site maintenance personnel.

[0145] For example, the axial pixel coordinate range of the aging concentrated umbrella skirt partition 1 is [1, 15], and the calculated axial coordinate of the center point is expressed as follows: The total number of pixels in the circumferential direction is 60, and the coordinates of the circumferential center are... .

[0146] Therefore, the location of the aging concentration area was determined to be the pixel coordinates (8,30) in the hyperspectral image, which corresponds to the position of the 8th axial pixel from the high-voltage end of the insulator, that is, the geometric center of the first shed of the high-voltage end of the insulator. This position is the area with the most concentrated aging and deterioration in this test.

[0147] B8. Output the location of the aging concentration area together with the quantitative assessment results of the insulator aging.

[0148] The coordinates of the concentrated aging area are combined with the quantitative assessment results of insulator aging to form a complete final output report.

[0149] Specifically, the final complete output report includes: The aging index and grade assignment for each umbrella skirt zone are as follows: Zone 1 (AI1=0.103, Grade IV severe aging) and Zone 2 (AI2=0.081, Grade III moderate aging). The aging index difference sequence between adjacent umbrella skirts is {0.022, 0.022, 0.021, 0.017, 0.014}, with a standard deviation of 0.0034, indicating that aging progresses uniformly and gradually along the axial direction without any local abrupt changes. The location coordinates of the aging concentration area are (8,30), which corresponds to the center of the first shed of the high-voltage end of the insulator. This is the aging and deterioration concentration area that needs to be focused on and prioritized in this inspection.

[0150] Example 3 is an embodiment of the present invention, which provides a hyperspectral-based quantitative assessment system for insulator aging, comprising: The image extraction module is used to establish a change sequence of spectral reflectance of each band of the hyperspectral image along the axis, and extract the monotonic attenuation component from the change sequence to obtain the first aging component map; The ratio calculation module performs a first-level calculation on the multi-band spectral intensity of the first aging component map, filters the results of the first-level calculation to obtain a first stable band set, performs a second-level calculation on each pixel based on the first stable band set, and aggregates the results of the second-level calculation to obtain an aging index distribution map. The results output module assigns a quantitative value to the degree of aging based on the aging index in the aging index distribution chart and outputs the quantitative assessment result of insulator aging.

[0151] This embodiment also provides an electronic device applicable to a hyperspectral-based quantitative assessment method for insulator aging, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the hyperspectral-based quantitative assessment method for insulator aging as proposed in the above embodiment.

[0152] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements a hyperspectral-based quantitative assessment method for insulator aging as proposed in the above embodiments.

[0153] The storage medium proposed in this embodiment belongs to the same inventive concept as the method for quantitative assessment of insulator aging based on hyperspectral imaging proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0154] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0155] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for quantitative assessment of insulator aging based on hyperspectral imaging, characterized in that, Includes the following steps: A sequence of spectral reflectance variation along the axis is established for each band of the hyperspectral image, and a monotonic attenuation component is extracted from the sequence to obtain the first aging component map. A first-level calculation is performed on the multi-band spectral intensity of the first aging component map, and the results of the first-level calculation are filtered to obtain the first stable band set. A secondary calculation is performed on each pixel based on the first stable band set, and the results of the secondary calculation are aggregated to obtain the aging index distribution map. The degree of aging is quantitatively assigned using the aging index in the aging index distribution chart, and the quantitative assessment result of insulator aging is output.

2. The method for quantitative assessment of insulator aging based on hyperspectral imaging as described in claim 1, characterized in that, The steps for establishing a variation sequence along the axial direction include: The directional consistency of the spectral reflectance of each band arranged along the axial direction is checked. The statistical analysis of the proportion of spectral reflectance that meets the preset direction determines the bands with a proportion not lower than the preset threshold as effective sensitive bands. The variation sequence is constructed using the spectral reflectance of the effective sensitive bands arranged along the axial direction.

3. The method for quantitative assessment of insulator aging based on hyperspectral imaging as described in claim 2, characterized in that, The steps for filtering the results of the first-level operation include: Two pixels that are circumferentially adjacent at the same axial height of an insulator are grouped together to form a pixel pair. The difference in spectral ratio between two pixels under each candidate band combination is calculated using the pixel pairs, and the band combination with the smallest difference is determined as the first stable band set.

4. The method for quantitative assessment of insulator aging based on hyperspectral imaging as described in claim 3, characterized in that, The steps to obtain the aging index distribution map include: The average spectral ratio of the corresponding region pixels of the reference insulator under the first stable band set is used as the first reference value; The difference between the spectral ratio of each pixel and the first reference value is calculated to obtain the first correction value of each pixel. The aging index distribution map is obtained by aggregating the first correction value according to the umbrella skirt partition.

5. The method for quantitative assessment of insulator aging based on hyperspectral imaging as described in claim 4, characterized in that, The steps for calculating based on the umbrella skirt partition aggregation include: The spectral reflectance of pixels at each axial position in the hyperspectral image is statistically calculated to obtain a first spectral curve distributed along the axial direction. Gradient calculation is performed on the first spectral curve, and the axial position where the gradient value meets the preset condition is determined as the umbrella skirt boundary. Based on the umbrella skirt boundary, the first correction value of each pixel is assigned to the corresponding umbrella skirt partition, and the first correction values ​​within the same umbrella skirt partition are aggregated to obtain the aging index distribution map.

6. The method for quantitative assessment of insulator aging based on hyperspectral imaging as described in claim 5, characterized in that, The steps for quantitatively assigning values ​​to the degree of aging include: The aging index of two adjacent skirt sections in the aging index distribution map is subtracted to obtain a first difference value, wherein the first difference value characterizes the degree of spatial non-uniformity of the aging progress of the insulator. The first difference and the aging degree assignment results of each skirt section together constitute the quantitative assessment result of insulator aging.

7. The method for quantitative assessment of insulator aging based on hyperspectral imaging as described in claim 6, characterized in that, The steps for quantitatively assessing the aging results of output insulators include: Extract the unique aging index corresponding to each umbrella skirt section from the aging index distribution map; According to the axial arrangement order of each skirt section on the insulator from the high voltage end to the low voltage end, all the extracted aging indices are arranged in sequence to generate a one-dimensional axial aging index sequence. The maximum value variable is initialized to the first element of the axial aging index sequence, and the maximum value index variable is initialized to 0. Starting from the second element of the axial aging index sequence, the current element is compared with the maximum value variable in turn; In response to the fact that the value of the current element is greater than the value of the maximum value variable, the maximum value variable is updated to the value of the current element, and the maximum value index variable is updated to the sequence index corresponding to the current element. After traversing all elements of the axial aging index sequence, the axial pixel coordinate range of the corresponding skirt partition in the hyperspectral image of the insulator is queried according to the maximum value index variable. The center point coordinates of the axial pixel coordinate range are determined as the location of the aging concentration area; The location of the aging concentration area and the quantitative assessment results of the insulator aging are output together.

8. A hyperspectral-based quantitative assessment system for insulator aging, employing the hyperspectral-based quantitative assessment method for insulator aging as described in any one of claims 1 to 7, characterized in that, include: The image extraction module is used to establish a change sequence of spectral reflectance of each band of the hyperspectral image along the axis, and extract the monotonic attenuation component from the change sequence to obtain the first aging component map; The ratio calculation module performs a first-level calculation on the multi-band spectral intensity of the first aging component map, filters the results of the first-level calculation to obtain a first stable band set, performs a second-level calculation on each pixel based on the first stable band set, and aggregates the results of the second-level calculation to obtain an aging index distribution map. The results output module assigns a quantitative value to the degree of aging based on the aging index in the aging index distribution chart and outputs the quantitative assessment result of insulator aging.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method for quantitative assessment of insulator aging based on hyperspectral imaging as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for quantitative assessment of insulator aging based on hyperspectral imaging as described in any one of claims 1 to 7.

Images