Spectral image analysis method and system for wettability of dirt on surface of silicone rubber

Abstract

The invention discloses a silicone rubber surface dirt wettability spectral image analysis method and system, and belongs to the technical field of high-voltage power transmission and transformation equipment state monitoring, and the method comprises the following steps: preparing a dry dirt sample and a saturated wet dirt sample of a vulcanized silicone rubber sample; acquiring reflection spectrum data of the dry dirty sample of the vulcanized silicone rubber sample and the saturated wet dirty sample of the vulcanized silicone rubber sample through a near-infrared reflection spectrum imager; normalizing the pixel gray value of the reflection spectrum imager by using the calibration white board and the standard blackboard, and determining a reflectivity calculation formula; respectively calculating curves of the reflectivity average values of the dry dirty sample area and the saturated wet dirty sample area along with the wavelength; and constructing a pollution wettability evaluation model which comprises pollution wettability evaluation model parameters and a wettability evaluation coefficient, calculating the pollution wettability of the surface of the vulcanized silicone rubber sample sheet through the pollution wettability evaluation model, and visually presenting the pollution wettability of the surface of the vulcanized silicone rubber sample sheet.

Description

Technical Field

[0001] The present invention belongs to the technical field of high-voltage power transmission and transformation equipment status monitoring, and more specifically, relates to a silicone rubber surface contamination wetness spectrum image analysis method and system. Background Art

[0002] In order to achieve large-scale energy transmission and resource allocation optimization, the country has proposed a strategic layout of west-to-east power transmission with high-voltage transmission lines as the skeleton, aiming to build a cross-regional power energy transmission network. Among them, insulators play a vital role. They not only provide the necessary mechanical support, but also ensure electrical isolation. However, due to the long-term exposure of insulators to the outdoors, their surfaces are prone to accumulate pollutants brought by atmospheric pollution. According to the operation data of the power system, up to 40% of external insulation failures are caused by pollution flashover. If these pollution problems are not handled in time, they may cause large-scale power outages and cause serious economic losses. In the development of insulator pollution flashover, surface pollution and the wetness of the pollution layer are the key factors leading to pollution flashover. The wet pollution layer will lead to an increase in the leakage current on the surface of the insulator, thereby promoting the generation and extension of the arc. For the vulcanized silicone rubber composite insulators commonly used in the line, when its hydrophobicity decreases, it will not only cause a decrease in pollution resistance, but also affect the hydrophobicity of the pollution layer. Therefore, accurate monitoring of the wetness of the pollution layer on the surface of the composite insulator is crucial to understanding the pollution flashover mechanism, assessing the operation risk, and optimizing the maintenance strategy of the insulator.

[0003] However, for a long time, there has been a lack of effective detection methods for the wetness of the pollution layer. Currently, the methods used to characterize the wetness of the pollution layer mainly include the pollution layer conductivity method, leakage current method, phase angle difference method and differential weight method.

[0004] The conductivity of the pollution layer increases significantly after being dampened, which can characterize the pollution wetness of the pollution layer at a fixed pollution level. However, there is currently no public literature or patent on how to separate the effects of pollution level and pollution wetness on conductivity from the measurement results. The leakage current method is also an effective method for characterizing the pollution wetness of the pollution layer. The ratio of the fundamental component of the leakage current to the maximum value of the total current characterizes the ratio of the residual pollution layer resistance to the total resistance of the pollution layer, but this method is difficult to separate the contribution of the pollution level and pollution wetness to the leakage current. The phase difference method characterizes the pollution wetness by measuring the phase difference between the leakage current and the externally applied voltage. The problem with this method is that it is difficult to measure the voltage phase and the small current phase in a complex electromagnetic environment. The differential weight method characterizes the pollution wetness by measuring the mass difference of the polluted insulator before and after being damp, but this method is difficult to actually apply to the engineering site and cannot provide the distribution characteristics of the pollution wetness, which is of little significance for actually judging whether the pollution flashover voltage has decreased. In recent years, various spectral imaging methods have been used to evaluate the surface condition of insulators, such as infrared thermal imaging, ultraviolet discharge imaging, and hyperspectral imaging. Although the former two can detect heating or discharge phenomena under severe contamination, the contamination that can cause heating or discharge is relatively rare in practice. Although the latter hyperspectral imaging can obtain the contamination distribution or composition of the insulator surface in the laboratory, there are few analysis methods specifically for the contamination wetness. At present, although there are papers and literature specifically studying the hyperspectral visualization analysis of the contamination wetness of glass insulator surfaces, there are essential differences between glass substrates and vulcanized silicone rubber substrates in the mechanism of contamination accumulation and spectral analysis methods, which makes it difficult to guide the analysis of the contamination wetness of composite insulators.

[0005] In summary, the existing pollution wetness monitoring still faces problems such as low detection efficiency, difficulty in separating multiple factors, and poor feasibility of on-site application.

[0006] Prior art document CN118424945A discloses a method for visualizing the contaminated moisture content on the surface of an insulator based on hyperspectral imaging technology, the method comprising the following steps: S1, preparing a contaminated test sample; S2, building a hyperspectral imaging platform and a high-voltage test platform, and connecting the test sample to the high-voltage test platform; S3, measuring the moisture content of the test sample and obtaining a hyperspectral image of the test sample through the hyperspectral imaging platform; S4, processing the hyperspectral image obtained in step S3, wherein the hyperspectral three-dimensional data is composed of one-dimensional spectral information (λ) and two-dimensional spatial information (x, y). For the test sample, its effective features are contained in a few characteristic bands. Based on the low-dimensional data in the characteristic bands, a quantitative solution model for the moisture content is proposed by partial least squares regression. Finally, quantitative visualization inversion of the moisture content of the test sample is achieved by pseudo-color processing. Its shortcomings are: 1. The influence of moisture on the reflectance spectrum is mainly reflected in the near-infrared band. The spectral band that can effectively reflect the wetness in the visible light range is relatively sparse and narrow. In addition, how to use the information of a few characteristic bands in hyperspectral imaging to reflect the wetness, and which spectral parameters and parameter combinations can be used for modeling are not described in detail; 2. Vulcanized silicone rubber materials generally use RTV hydrophobic coatings, and their hydrophobic state will affect the judgment of the water content in the contamination layer. It is necessary to construct model parameters specifically for the entire dynamic range of contamination wetness on the surface of vulcanized silicone rubber (i.e., from completely dry to saturated wetness); 3. Hyperspectral imaging is a general analysis method that is mostly used in laboratories. For actual line insulators, the shooting distance, angle and stability are difficult to guarantee, especially for drone mounting applications. Therefore, the model training is completely based on hyperspectral curves, which deviates from reality. Summary of the invention

[0007] In order to solve the deficiencies in the prior art, the present invention provides a method and system for analyzing the spectral image of contamination and wetness on the surface of silicone rubber.

[0008] The present invention adopts the following technical solution.

[0009] The first aspect of the present invention provides a method for analyzing the spectral image of contamination wetness on the surface of silicone rubber, comprising the following steps:

[0010] Prepare dry dirty samples and saturated wet dirty samples of vulcanized silicone rubber specimens;

[0011] The reflectance spectrum data of the dry dirty sample of the vulcanized silicone rubber sample and the saturated wet dirty sample of the vulcanized silicone rubber sample are obtained by a near-infrared reflectance spectrum imager in the 900nm-1700nm band;

[0012] The pixel grayscale value of the reflectance spectrum imager is normalized using a calibration white board and a standard black board to determine the reflectance calculation formula;

[0013] Calculate the curves of the average reflectance of the dry dirty sample area and the saturated wet dirty sample area versus wavelength respectively;

[0014] Constructing a pollution wetness assessment model, which includes pollution wetness assessment model parameters and wetness estimation coefficients, setting and calculating pollution wetness assessment model parameters based on the curves of the average reflectance of the dry pollution sample area and the saturated wet pollution sample area versus wavelength, including the filling rate, difference rate, change rate and change difference rate, and calculating the wetness estimation coefficients corresponding to the pollution wetness assessment model parameters;

[0015] Based on the pollution wetness assessment model parameters and wetness estimation coefficients, the pollution wetness on the surface of the vulcanized silicone rubber sample is calculated using the pollution wetness assessment model and visualized.

[0016] Optionally, the preparation of the dry dirty sample and the saturated wet dirty sample of the vulcanized silicone rubber sample comprises:

[0017] Uniformly prepare a foul coating on the surface of the vulcanized silicone rubber sample;

[0018] Place the contaminated samples in an artificial fog chamber, and set the relative humidity of the atmosphere in the artificial fog chamber to no higher than 5% and no lower than 95% respectively. The relative humidity of the atmosphere no higher than 5% is the dry humidity condition, and the relative humidity of the atmosphere no lower than 95% is the saturated humidity condition. The sample standing time under dry and saturated humidity conditions is no less than 60 minutes.

[0019] Optionally, the reflection spectrum data includes the gray value P0(λ) of each pixel in the sample area, where λ is the wavelength;

[0020] The calculation formula of reflectivity R(λ) is:

[0021]

[0022] Where P B (λ) is the standard black image data, P W (λ) is the standard white image data.

[0023] Optionally, the curve of the average reflectance of the dry dirty sample area and the wavelength λ is recorded as R0(λ); the curve of the average reflectance of the saturated wet dirty sample area and the wavelength λ is recorded as R1(λ);

[0024] The calculation of the filling rate k1 includes:

[0025] k1=R p (λ1) / R max

[0026] In the formula, R maxis the maximum absolute value of the difference between R1(λ) and R0(λ), λ1 is the wavelength corresponding to the maximum absolute value of the difference between R1(λ) and R0(λ);

[0027] Calculating the difference rate k2 includes:

[0028] k2=R p (λ1)-R p (λ2)

[0029] Where λ2 is the wavelength corresponding to the minimum absolute value of the difference between R1(λ) and R0(λ);

[0030] Calculating the rate of change k3 includes:

[0031]

[0032] Where λ3 is the central wavelength corresponding to the maximum value of the absolute value of the difference between R1(λ) and R0(λ) at the dλ step change rate;

[0033] Calculating the change difference rate k4 includes:

[0034]

[0035] Where λ4 is the central wavelength corresponding to the minimum value of the absolute value of the difference between R1(λ) and R0(λ) at the dλ step change rate;

[0036] R p (λ) is the reflectivity of the current test result at each pixel.

[0037] Optionally, the R max The calculation includes:

[0038] R max =max(|R1(λ)-R0(λ)|)

[0039] The r max The calculation includes:

[0040]

[0041] The value of dλ is 5 times the spectral resolution of the near-infrared reflectance spectrometer.

[0042] Optionally, the calculation of the wetness estimation coefficient corresponding to the pollution wetness assessment model parameter includes:

[0043] Set the initial wetness coefficient α 0 (t):

[0044]

[0045] Among them, Yi is the ith element of the measured pollution wetness vector, X i is the i-th row of the parameter matrix of the pollution wetness assessment model, α is the wetness estimation coefficient vector, and n is the number of samples;

[0046] Calculate the model deviation value δ of the mth iteration at time t m (t), the calculation formula is:

[0047] δ m (t) = Y - Xα m (t)

[0048] Among them, Y is the measured pollution wetness vector, X is the pollution wetness assessment model parameter matrix, α m (t) is the estimated wetness coefficient of the mth iteration at time t;

[0049] The deviation σ is defined as an indicator describing the degree of deviation of the model output:

[0050] σ=1.4826EA

[0051] Among them, EA is the average deviation absolute value, which is the absolute value of the median deviation value of all models;

[0052] Defining Standardized Deviation is the model deviation value δ of the mth iteration at time t m The ratio of (t) to the deviation σ:

[0053]

[0054] Calculate the indicator weight ρ of the mth iteration at time t m (t), using the indicator weight ρ m (t) Perform weighted regression and update the estimated coefficient α of the corrected humidity m (t):

[0055] α m+1 (t) = (X T ρ * X) -1 X T ρ * Y

[0056] Among them, ρ * It is a diagonal matrix, and the elements on its diagonal are the index weights ρ m (t);

[0057] Repeat the above steps until the change of the modified wetness estimation coefficient is less than the preset iterative convergence deviation threshold s:

[0058] |α m+1 (t)-αm (t)|

[0059] At this time, α m (t) is the final wetness estimation coefficient vector.

[0060] Optionally, the calculation of the index weight ρ of the mth iteration at time t is m (t) include:

[0061]

[0062] Among them, b is the weight scaling value, which ranges from 1.05 to 1.4.

[0063] Optionally, the calculating the dirt wettability of the surface of the vulcanized silicone rubber sample by using the dirt wettability evaluation model and visually presenting it specifically includes:

[0064] Extracting one-dimensional reflection spectrum data corresponding to each pixel from the reflection spectrum data corrected for black and white;

[0065] According to the calculation results of the pollution wetness evaluation model parameters k1~k4 and the corresponding wetness estimation coefficients α1~α4, the pollution wetness is calculated. The expression of the pollution wetness evaluation model is:

[0066] Dirt wetness = α1·k1+α2·k2+α3·k3+α4·k4

[0067] The wetness value of each pixel point (x, y) in the test area is calculated, and the characterization color scale is defined for different result data. The detection results obtained by all coordinates are pseudo-colored according to the color scale in turn to obtain distribution images of different wetness.

[0068] Optionally, the method further includes, when using a near-infrared reflectance spectrum imager to acquire reflectance spectrum data of the sample, using a uniform halogen light source to irradiate the surface of the sample.

[0069] The second aspect of the present invention provides a silicone rubber surface contamination wetness spectrum image analysis system, based on the silicone rubber surface contamination wetness spectrum image analysis method described in the first aspect of the present invention, the system comprises:

[0070] Near-infrared reflectance spectrum imager, pixel gray value normalization module, reflectance calculation module, model parameter calculation module, wetness estimation coefficient calculation module, pollution wetness calculation module and visualization module;

[0071] Near-infrared reflectance spectrum imager, pixel gray value normalization module, reflectance calculation module, model parameter calculation module, wetness estimation coefficient calculation module, pollution wetness calculation module and visualization module; ​

[0072] The near-infrared reflectance spectrum imager is used to obtain the reflectance spectrum data of dry contaminated samples of vulcanized silicone rubber samples and saturated wet contaminated samples of vulcanized silicone rubber samples in the 900nm~1700nm band; the pixel grayscale value normalization module normalizes the pixel grayscale value of the reflectance spectrum imager; the reflectance calculation module is used to calculate the reflectance; the model parameter calculation module is used to calculate the contamination wetness evaluation model parameters; the wetness estimation coefficient calculation module is used to calculate the wetness estimation coefficient corresponding to the contamination wetness evaluation model parameters; the contamination wetness calculation module is used to calculate the contamination wetness on the surface of the vulcanized silicone rubber sample; the visualization module is used to visualize the contamination wetness on the surface of the vulcanized silicone rubber sample.

[0073] Compared with the prior art, the beneficial effects of the present invention include at least:

[0074] High detection efficiency: The present invention utilizes reflective spectrum imaging technology to quickly obtain spectrum data of the contamination layer, thus significantly improving detection efficiency.

[0075] Strong field applicability: This method does not require complex equipment and environmental conditions and is suitable for rapid deployment and use in various field environments.

[0076] Rich information: The present invention can provide accurate quantitative analysis of the wetness of the pollution layer and visualize the distribution characteristics of the pollution wetness, providing important data support for insulator design optimization and pollution flashover risk assessment.

[0077] High precision and high stability: The reflectance spectrum imaging technology and pollution wetness assessment model algorithm adopted in the present invention can effectively reduce the influence of measurement errors and outliers, and improve the accuracy and stability of the detection results. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 It is a schematic diagram of the visualized distribution of dirt on the surface of vulcanized silicone rubber at different dirt wetness levels provided in accordance with an embodiment of the present invention;

[0079] Figure 2 It is a flow chart of a method provided according to an embodiment of the present invention. DETAILED DESCRIPTION

[0080] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The described embodiments are only embodiments of a part of the present invention, not all embodiments. Based on the spirit of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the protection scope of the present invention.

[0081] The present invention provides a method for analyzing the spectral image of the contamination wetness of a silicone rubber surface in Example 1, such as Figure 2 As shown, the following steps are included:

[0082] Step 1, preparing a dry dirty sample and a saturated wet dirty sample of a vulcanized silicone rubber sample;

[0083] In a further preferred but non-limiting embodiment, the step 1 specifically comprises:

[0084] Uniformly prepare a foul coating on the surface of the vulcanized silicone rubber sample;

[0085] The contaminated samples are placed in an artificial fog chamber, and the relative humidity of the atmosphere in the artificial fog chamber is set to no higher than 5% (dry) and no lower than 95% (saturated). The sample is allowed to stand for no less than 60 minutes under dry and saturated humidity conditions.

[0086] Step 2, obtaining the reflectance spectrum data of the dry dirty sample of the vulcanized silicone rubber sample and the saturated wet dirty sample of the vulcanized silicone rubber sample in the 900nm to 1700nm band by a near-infrared reflectance spectrum imager;

[0087] In a further preferred but non-limiting embodiment, in step 2, the reflection spectrum data includes the grayscale value P0(λ) of each pixel in the sample area, wherein λ is the wavelength in the range of 900nm to 1700nm.

[0088] In a further preferred but non-limiting embodiment, step 2 further includes, when using a near-infrared reflectance spectroscopic imager to acquire reflectance spectrum data of the sample, using a uniform halogen light source to irradiate the surface of the sample.

[0089] Step 3, using a calibration white board with a reflectivity greater than 95% and a standard black board with a reflectivity less than 5% to normalize the pixel grayscale value of the reflectance spectrum imager, and determine the reflectivity calculation formula;

[0090] In a further preferred but non-limiting embodiment, in step 3, the calculation formula of the reflectivity R(λ) is:

[0091]

[0092] Where P0(λ) is the gray value of each pixel in the sample area, P B (λ) is the standard black image data, P W (λ) is the standard white image data.

[0093] Step 4, based on the reflectance spectrum data obtained in step 2 and the reflectance calculation formula determined in step 3, respectively calculate the curves of the reflectance average value of the dry dirty sample area and the saturated wet dirty sample area versus wavelength;

[0094] In a further preferred but non-limiting embodiment, in step 4, the curve of the average reflectance of the dry dirty sample area versus the wavelength λ is recorded as R0(λ); the curve of the average reflectance of the saturated wet dirty sample area versus the wavelength λ is calculated and recorded as R1(λ).

[0095] Step 5, based on the curves of the average reflectance of the dry dirty sample area and the saturated wet dirty sample area versus wavelength calculated in step 4, define and calculate the parameters k1 to k4 of the pollution wetness assessment model;

[0096] In a further preferred but non-limiting embodiment, step 5 specifically comprises:

[0097] Based on the curve R0(λ) of the average reflectance of the dry dirty sample area and the wavelength λ and the curve R1(λ) of the average reflectance of the saturated wet dirty sample area and the wavelength λ calculated in step 4, it can be determined that:

[0098] At time t, within the detection band, the maximum absolute value of the difference between R1(λ) and R0(λ) is:

[0099] R max =max(|R1(λ)-R0(λ)|)

[0100] The wavelength corresponding to the maximum absolute value of the difference between R1(λ) and R0(λ) is λ1;

[0101] At time t, within the detection band, the wavelength corresponding to the minimum absolute value of the difference between R1(λ) and R0(λ) is λ2;

[0102] At time t, within the detection band, the maximum value of the absolute value of the difference between R1(λ) and R0(λ) at the dλ step change rate is:

[0103]

[0104] The central wavelength corresponding to the maximum value of the absolute value of the difference between R1(λ) and R0(λ) at the dλ step change rate is λ3; where dλ is 5 times the spectral resolution of the reflectance spectrum imager.

[0105] Step 5.4, calculate that at time t, within the detection band, the central wavelength corresponding to the minimum value of the absolute value of the difference between R1(λ) and R0(λ) at the dλ step change rate is λ4;

[0106] Step 5.5, calculate, at time t, the reflectivity of the current test result at each pixel is R p (λ), define model parameters k1~k4:

[0107] Filling rate k1:

[0108] k1=R p (λ1) / R max

[0109] Difference rate k2:

[0110] k2=R p (λ1)-R p (λ2)

[0111] Rate of change k3:

[0112]

[0113] Change difference rate k4:

[0114]

[0115] Step 6, calculating the wetness estimation coefficients α1-α4 corresponding to the pollution wetness assessment model parameters k1-k4;

[0116] In a further preferred but non-limiting embodiment, in step 6, the calculation of the wetness estimation coefficient specifically includes:

[0117] The calculation of the wetness estimation coefficient includes:

[0118] Step 6.1, initial parameter setting:

[0119] Set the initial wetness coefficient α 0 (t):

[0120]

[0121] Among them, Y i is the ith element of the measured pollution wetness vector, X i is the i-th row of the model parameter matrix, α is the wetness estimation coefficient vector, and n is the number of samples;

[0122] Step 6.2, model deviation value:

[0123] Calculate the model deviation value δ of the mth iteration at time t m (t), the calculation formula is:

[0124] δ m (t) = Y - Xα m (t)

[0125] Among them, Y is the measured pollution wetness vector, X is the model parameter matrix, X = [k1k2k3k4] T , α m(t) is the estimated wetness coefficient of the mth iteration at time t;

[0126] Step 6.3, Deviation and Standardized Deviation:

[0127] The deviation σ is defined as an indicator describing the degree of deviation of the model output:

[0128] σ=1.4826EA

[0129] Among them, EA is the average deviation absolute value, which is the absolute value of the median deviation value of all models;

[0130] Defining Standardized Deviation is the model deviation value δ of the mth iteration at time t m The ratio of (t) to the deviation σ:

[0131]

[0132] Step 6.4, calculate the indicator weight:

[0133] Calculate the indicator weight ρ of the mth iteration at time t m (t):

[0134]

[0135] Among them, b is the weight scaling value, which ranges from 1.05 to 1.4;

[0136] Step 6.5, weighted regression:

[0137] Use indicator weight ρ m (t) Perform weighted regression and update the estimated coefficient α of the corrected humidity m (t):

[0138] α m+1 (t) = (X T ρ * X) -1 X T ρ * Y

[0139] Among them, ρ * It is a diagonal matrix, and the elements on its diagonal are the index weights ρ m (t);

[0140] Step 6.6, iterate until convergence:

[0141] Repeat steps 6.2-6.5 until the change in the corrected wettability estimation coefficient is less than the preset iterative convergence deviation threshold s:

[0142] |α m+1 (t)-αm (t)|

[0143] At this time, α m (t) is the final wetness estimation coefficient vector, α m (t) = [α1 α2 α3 α4].

[0144] In a further preferred but non-limiting embodiment, the measured contamination wetness can be obtained in a variety of ways, such as using a precision electronic scale to monitor the mass difference of the insulator compared to a dry state, or estimating the moisture content of the contaminants after equilibrium by monitoring the relative humidity of the environment.

[0145] Step 7, substitute the pollution wetness assessment model parameters k1~k4 and wetness estimation coefficients α1~α4 into the pollution wetness assessment model, calculate the pollution wetness of the surface of the vulcanized silicone rubber sample, and visualize it.

[0146] In a further preferred but non-limiting embodiment, in step 7, calculating the dirt wettability of the surface of the vulcanized silicone rubber sample and visually presenting it specifically includes:

[0147] Extracting one-dimensional reflection spectrum data corresponding to each pixel from the reflection spectrum data corrected for black and white;

[0148] According to the calculation results of the four model parameters (k1~k4) and the corresponding wetness estimation coefficients α1~α4, the pollution wetness is calculated:

[0149] Dirt wetness = α1·k1+α2·k2+α3·k3+α4·k4

[0150] The wetness value of each pixel point (x, y) in the test area is calculated, and the characterization color scale is defined for different result data. The detection results obtained by all coordinates are pseudo-colored according to the color scale in turn to obtain distribution images of different wetness.

[0151] The object of this embodiment is a vulcanized silicone rubber sample, and a certain degree of dirt has accumulated on the surface of these insulators. Our goal is to detect the dirt wetness of the dirt layer on the surface of these insulators by using reflectance spectral imaging technology.

[0152] (a) Implementation environment: A reflectance spectroscopy imaging platform was built in the laboratory, including a reflectance spectroscopy imager, a set of symmetrical light sources, and an optical darkroom.

[0153] Implementation steps:

[0154] (b) Sample preparation:

[0155] ​A clean vulcanized silicon rubber sheet with a diameter of 4 cm was used as a base material to prepare an artificial contamination sample according to the method of the present invention.

[0156] (c) Reflectance spectrum data collection:

[0157] The prepared contaminated sample is placed on the reflectance spectrum imaging platform, and the near-infrared reflectance spectrum data of the sample is obtained in the 900-1700nm band using a reflectance spectrum imager.

[0158] (d) Pollution wetness reflection model parameter set:

[0159] Taking a certain pixel as an example, according to the method provided by the present invention, the difference index (DI), normalized index (NI) and band ratio index (BRI) are calculated. The wavelengths involved in the calculation of the four model parameters are as follows:

[0160]

[0161] (e) Pollution wetness assessment model:

[0162] For time t,

[0163] Use the least squares method to get the initial coefficient α (0) ;

[0164] Calculate the model deviation value δ and deviation σ, and calculate the standardized deviation σ N .

[0165] Calculate the indicator weight ρ for weighted regression, where b is 1.3224.

[0166] Perform weighted regression and continuously iterate and update the modified moisture estimation coefficient vector α m+1 .

[0167] Iterate until convergence to obtain the final moisture estimation coefficient vector α. Taking a certain pixel as an example, assuming that the moisture estimation coefficient vector α obtained after iteration of the robust regression model is: α1=0.12, α2=0.11, α3=0.38, α4=0.09

[0168] (f) Quantitative inversion and visualization of pollution wetness:

[0169] Calculate the pollution wetness data wetness (x, y) corresponding to each pixel point (x, y) at time t. Take a certain pixel point as an example:

[0170] (x,y)=α1·k1+α2·k2+α3·k3+α4·k4=0.12×0.331+0.11×0.0726+0.38×0.465+0.09×0.0331=0.2274

[0171] According to the above calculation, the estimated wettability of the dirt layer on the surface of the vulcanized silicone rubber is 0.2274. Figure 1 As shown, by pseudo-colorizing the wetness data of all pixels, a visualized image of the wetness of the contamination layer is obtained, which provides important information for further analysis and maintenance of insulators.

[0172] The present invention provides a silicone rubber surface contamination wetness spectrum image analysis system in embodiment 3. Based on the silicone rubber surface contamination wetness spectrum image analysis method described in embodiment 1, the system includes:

[0173] Near-infrared reflectance spectrum imager, pixel gray value normalization module, reflectance calculation module, model parameter calculation module, wetness estimation coefficient calculation module, pollution wetness calculation module and visualization module;

[0174] The near-infrared reflectance spectrum imager is used to obtain the reflectance spectrum data of the dry contaminated samples and the saturated wet contaminated samples of the vulcanized silicone rubber samples in the 900nm~1700nm band; the pixel gray value normalization module normalizes the pixel gray value of the reflectance spectrum imager; the reflectance calculation module is used to calculate the reflectance; the model parameter calculation module is used to calculate the contamination wetness assessment model parameters k1~k4; the wetness estimation coefficient calculation module is used to calculate the wetness estimation coefficients α1~α4 corresponding to the contamination wetness assessment model parameters k1~k4; the contamination wetness calculation module is used to calculate the contamination wetness on the surface of the vulcanized silicone rubber sample; the visualization module is used to visualize the contamination wetness on the surface of the vulcanized silicone rubber sample.

[0175] It is worth noting that the technical means adopted in the present invention include: 1. 900nm~1700nm near-infrared hyperspectral imaging data is used as the source of parameter extraction, which can more sensitively reflect the moisture content; 2. For vulcanized silicone rubber, the optimal parameters and parameter combinations are found in the entire range of dry pollution and saturated wet pollution, providing more robust parameter input for model establishment; 3. The bands and band combinations corresponding to the obtained parameters can be used separately, and distortion can also be achieved using multispectral imaging with a limited number of bands, providing a reference for practical applications.

[0176] Based on this technical means, the technical effects actually achieved by the present invention include at least: 1. Improving the efficiency, accuracy and dynamic range of the visualization inversion of the wettability of the contamination layer on the surface of silicone rubber insulators; 2. Reducing the general requirements for the extraction of characteristic spectral parameters for performance such as high spectral resolution and spectral response range; 3. The proposed spectral parameters and combinations are conducive to improving the robustness of the wettability analysis model.

[0177] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for analyzing the spectral image of contamination wetness on the surface of silicone rubber, characterized in that: The steps include: Prepare dry dirty samples and saturated wet dirty samples of vulcanized silicone rubber specimens; The reflectance spectrum data of the dry dirty sample of the vulcanized silicone rubber sample and the saturated wet dirty sample of the vulcanized silicone rubber sample are obtained by a near-infrared reflectance spectrum imager in the 900nm-1700nm band; The pixel grayscale value of the reflectance spectrum imager is normalized using a calibration white board and a standard black board to determine the reflectance calculation formula; Calculate the curves of the average reflectance of the dry dirty sample area and the saturated wet dirty sample area versus wavelength respectively; Constructing a pollution wetness assessment model, which includes pollution wetness assessment model parameters and wetness estimation coefficients, setting and calculating pollution wetness assessment model parameters based on the curves of the average reflectance of the dry pollution sample area and the saturated wet pollution sample area versus wavelength, including the filling rate, difference rate, change rate and change difference rate, and calculating the wetness estimation coefficients corresponding to the pollution wetness assessment model parameters; Based on the pollution wetness assessment model parameters and wetness estimation coefficients, the pollution wetness on the surface of the vulcanized silicone rubber sample is calculated through the pollution wetness assessment model and visualized.

2. The method for analyzing the spectral image of contamination wetness on the surface of silicone rubber according to claim 1, characterized in that: The dry dirty sample and saturated wet dirty sample for preparing the vulcanized silicone rubber sample include: Uniformly prepare a foul coating on the surface of the vulcanized silicone rubber sample; Place the contaminated samples in an artificial fog chamber, and set the relative humidity of the atmosphere in the artificial fog chamber to no higher than 5% and no lower than 95% respectively. The relative humidity of the atmosphere no higher than 5% is the dry humidity condition, and the relative humidity of the atmosphere no lower than 95% is the saturated humidity condition. The sample standing time under dry and saturated humidity conditions is no less than 60 minutes.

3. The method for analyzing the spectral image of contamination wetness on the surface of silicone rubber according to claim 1, characterized in that: The reflection spectrum data includes the gray value P0(λ) of each pixel in the sample area, where λ is the wavelength; The calculation formula of reflectivity R(λ) is: Where P B (λ) is the standard black image data, P W (λ) is the standard white image data.

4. A method for analyzing the spectral image of contamination wetness on the surface of silicone rubber according to claim 1 or 3, characterized in that: The curve of the average reflectivity of the dry dirty sample area and the wavelength λ is recorded as R0(λ); the curve of the average reflectivity of the saturated wet dirty sample area and the wavelength λ is recorded as R1(λ); The calculation of filling rate k1 include: k1=R p (λ1) / R max In the formula, R max is the maximum absolute value of the difference between R1(λ) and R0(λ), λ1 is the wavelength corresponding to the maximum absolute value of the difference between R1(λ) and R0(λ); Calculating the difference rate k2 includes: k2=R p (λ1)-R p (λ2) Where λ2 is the wavelength corresponding to the minimum absolute value of the difference between R1(λ) and R0(λ); Calculating the rate of change k3 includes: Where λ3 is the central wavelength corresponding to the maximum value of the absolute value of the difference between R1(λ) and R0(λ) at the dλ step change rate; Calculating the change difference rate k4 includes: Where λ4 is the central wavelength corresponding to the minimum value of the absolute value of the difference between R1(λ) and R0(λ) at the dλ step change rate; R p (λ) is the reflectivity of the current test result at each pixel.

5. The method for analyzing the spectral image of contamination wetness on the surface of silicone rubber according to claim 4, characterized in that: The R max The calculation includes: R max =max(|R1(λ)-R0(λ)|) The r max The calculation includes: The value of dλ is 5 times the spectral resolution of the near-infrared reflectance spectrometer.

6. A method for analyzing the spectral image of contamination wetness on the surface of silicone rubber according to claim 1 or 4, characterized in that: The wetness estimation coefficient corresponding to the pollution wetness assessment model parameter includes: Set the initial wetness coefficient α 0 (t): Among them, Y i is the ith element of the measured pollution wetness vector, X i is the i-th row of the parameter matrix of the pollution wetness assessment model, α is the wetness estimation coefficient vector, and n is the number of samples; Calculate the model deviation value δ of the mth iteration at time t m (t), the calculation formula is: δ m (t)=Y-Xα m (t) Among them, Y is the measured pollution wetness vector, X is the pollution wetness assessment model parameter matrix, α m (t) is the estimated wetness coefficient of the mth iteration at time t; The deviation σ is defined as an indicator describing the degree of deviation of the model output: σ=1.4826EA Among them, EA is the average deviation absolute value, which is the absolute value of the median deviation value of all models; Defining Standardized Deviation is the model deviation value δ of the mth iteration at time t m The ratio of (t) to the deviation σ: Calculate the indicator weight ρ of the mth iteration at time t m (t), using the indicator weight ρ m (t) Perform weighted regression and update the estimated coefficient α of the corrected humidity m (t): a m+1 (t)=(X T r * X) -1 X T r * Y Among them, ρ * It is a diagonal matrix, and the elements on the diagonal are the index weights ρ m (t); Repeat the above steps until the change of the modified wetness estimation coefficient is less than the preset iterative convergence deviation threshold s: |a m+1 (t)-a m (t)| <s At this time, α m (t) is the final wetness estimation coefficient vector.

7. The method for analyzing the spectral image of contamination wetness on the surface of silicone rubber according to claim 6, characterized in that: The calculation of the index weight ρ of the mth iteration at time t m (t) include: Among them, b is the weight scaling value, which ranges from 1.05 to 1.

4.

8. The method for analyzing the spectral image of contamination wetness on the surface of silicone rubber according to claim 1, characterized in that: The method of calculating the dirt wettability of the surface of the vulcanized silicone rubber sample by using the dirt wettability evaluation model and visually presenting the dirt wettability of the surface of the vulcanized silicone rubber sample specifically includes: Extracting one-dimensional reflection spectrum data corresponding to each pixel from the reflection spectrum data corrected for black and white; According to the calculation results of the pollution wetness evaluation model parameters k1~k4 and the corresponding wetness estimation coefficients α1~α4, the pollution wetness is calculated. The expression of the pollution wetness evaluation model is: Dirt wetness = α1·k1+α2·k2+α3·k3+α4·k4 The wetness value of each pixel point (x, y) in the test area is calculated, and the characterization color scale is defined for different result data. The detection results obtained by all coordinates are pseudo-colored according to the color scale in turn to obtain distribution images of different wetness.

9. The method for analyzing the spectral image of contamination wetness on the surface of silicone rubber according to claim 1, characterized in that: The method further includes, when using a near-infrared reflectance spectrum imager to obtain reflectance spectrum data of the sample, using a uniform halogen light source to irradiate the surface of the sample.

10. A silicone rubber surface contamination wetness spectrum image analysis system, based on a silicone rubber surface contamination wetness spectrum image analysis method according to any one of claims 1 to 9, the system comprising: Near-infrared reflectance spectrum imager, pixel gray value normalization module, reflectance calculation module, model parameter calculation module, wetness estimation coefficient calculation module, pollution wetness calculation module and visualization module; The near-infrared reflectance spectrum imager is used to obtain the reflectance spectrum data of dry contaminated samples of vulcanized silicone rubber samples and saturated wet contaminated samples of vulcanized silicone rubber samples in the 900nm~1700nm band; the pixel grayscale value normalization module normalizes the pixel grayscale value of the reflectance spectrum imager; the reflectance calculation module is used to calculate the reflectance; the model parameter calculation module is used to calculate the contamination wetness evaluation model parameters; the wetness estimation coefficient calculation module is used to calculate the wetness estimation coefficient corresponding to the contamination wetness evaluation model parameters; the contamination wetness calculation module is used to calculate the contamination wetness on the surface of the vulcanized silicone rubber sample; the visualization module is used to visualize the contamination wetness on the surface of the vulcanized silicone rubber sample.

Images