Insulator contamination level assessment method based on multispectral analysis

Abstract

The application provides an insulator contamination grade evaluation method based on multispectral analysis, and relates to the technical field of data analysis. The insulator contamination grade evaluation process comprises precise alignment of a camera carried by a UAV body, multispectral camera image acquisition, image processing and image evaluation. The UAV body main control comprises a flight trajectory control system, a flight attitude control system and an automatic navigation positioning calculation system. The UAV carries a multispectral camera. The multispectral camera is installed at the center position of the UAV. Temperature detection probes and humidity detection probes are installed on both sides of the UAV. The UAV can obtain the height image of the insulator in advance. Then, the UAV calculates the vertical level difference between the UAV and the insulator by using the automatic navigation positioning calculation system. After obtaining the precise insulator position information, the UAV flies to the periphery of the insulator to replace manual information acquisition.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to an insulator pollution level assessment method based on multi-spectral analysis. Background Art

[0002] According to Chinese Patent No. "CN 105956377 A," a method for regional pollution assessment based on sampling points discloses a method for obtaining the concentration of each pollutant at each sampling point and the degree of each pollutant at each estimated point. The method then calculates the degree of pollution at each estimated point based on the degree of pollution at the sampling point, the degree of pollution at each sampling point, and the concentration of each pollutant at the estimated point. This method effectively considers the impact of background pollutant concentrations on the degree of pollution at the sampling and estimated points, reducing the need for sampling and achieving accurate regional pollution assessment. This method is also applied to the assessment of the degree of pollution on transmission line insulators. The degree of pollution on the insulators of the sampling pole is used to estimate the degree of pollution on the tower insulator, reducing sampling effort and achieving accurate assessment of the degree of pollution on transmission line insulators.

[0003] The above patent documents and prior art have the following technical problems when used:

[0004] Problem 1: Currently, power systems have numerous transmission lines, and insulators are located high up on towers. The aforementioned solutions are largely manual, so manual sampling of insulator contamination levels is labor-intensive, requiring significant manpower and resources, and also presents risks for the inspectors.

[0005] The second problem is that when existing technologies use spectral cameras to photograph the surface of insulators, the recognition end does not process the images in a detailed and comprehensive manner, cannot identify the main features, and does not perform analysis and evaluation based on the fusion of other multi-source data, resulting in inaccurate evaluation results. Summary of the Invention

[0006] In view of the shortcomings of the existing technology, the present invention provides an insulator pollution level assessment method based on multi-spectral analysis, which solves the problems of heavy workload in manual sampling of insulator pollution and inaccurate pollution level assessment results.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for evaluating the contamination level of insulators based on multispectral analysis. The insulator contamination level evaluation process includes precise positioning of a camera mounted on a drone, multispectral camera image acquisition, image processing, and image evaluation. The specific steps are as follows:

[0008] The main control system of the drone includes a flight trajectory control system, a flight attitude control system, and an automatic navigation and positioning calculation system. The drone is equipped with a multispectral camera, which is installed in the center of the drone. At the same time, temperature detection probes and humidity detection probes are installed on both sides of the drone. The drone will obtain the height image of the insulator in advance, and then the drone calculates the vertical level difference between itself and the insulator through the automatic navigation and positioning calculation system, specifically calculating the image height h of the high-altitude high-voltage line A. a and the image height h of insulator B b As well as the image distance between the high-altitude high-voltage line A and the insulator B, the distance u between the high-altitude high-voltage line A and the insulator B and the drone a and u b As well as the angle θ between the high-altitude high-voltage line A and the insulator B relative to the imaging center, the safe distance between the drone and the transmission line plane is:

[0009]

[0010] SP2: When the drone's built-in computing system calculates the height of the insulator, it transmits the information to the drone's flight drive module, which then controls the drone's body to automatically ascend.

[0011] SP3: When the UAV is in flight, it will automatically search, detect, and monitor various devices around the insulator. During this process, the station staff will calibrate and control the flight speed, flight position, and flight trajectory of the UAV. During the control process, the roll angle and pitch angle control variables will be calculated in detail based on the attitude stability of the UAV. The flight trajectory will be controlled and adjusted based on the calculation results to maximize the stability of the UAV inspection flight and make the multispectral camera more stable when working;

[0012] SP4: The front-end imaging element of the multispectral camera is a filter array obtained by splicing multiple independent filter substrates in sequence. The filter array is installed in front of the imaging device. When the multi-band light obtained by the light splitting by the filter array is dispersed to different imaging positions, multispectral images of corresponding bands in different strip areas are obtained. The imaging device is a planar array detector.

[0013] Preferably, when the multispectral camera on the drone is close to the insulator, the light source is turned on and the light intensity and angle are adjusted according to the light reflection condition of the surface of the insulator to be evaluated. The parallel light reaches the surface of the insulator to be evaluated and generates reflected light, and a multi-band image is used to collect a multi-band image of the reflected light on the insulator surface.

[0014] Preferably, the multispectral camera needs to be preheated for half an hour before being placed close to the insulator for image acquisition to maintain the stability of the light source. Parameters are set after the preheating is completed. Due to the current of the device itself and the light source image in the environment, black and white correction is required to obtain hyperspectral images.

[0015] Preferably, after acquiring the insulator surface image, the multispectral camera transmits it to the image processing system, performs principal component analysis (PCA) and minimum noise factorization (MNF) dimensionality reduction analysis on the hyperspectral full-band image, finds the most obvious contamination component on the insulator surface, and then performs characteristic wavelength screening based on the competitive adaptive reweighting (CARS) algorithm and the continuous projection (SPA) algorithm, and combines the support vector machine (SVM) to perform contamination adhesion degree classification.

[0016] Preferably, the image processing system performs denoising on the image data, and removes noise in the image signal through wavelet denoising and convolution smoothing technology.

[0017] Preferably, after the image information noise is processed, the image processing system will perform refraction correction on the image, and use MSC for spectral analysis to effectively suppress the scattering phenomenon, thereby more accurately capturing the characteristics of each component in the "ideal spectrum" and improving the accuracy of the analysis.

[0018] Preferably, a multispectral detection and analysis system is built in the image processing system. The multispectral detection and analysis system is a multispectral analysis system established by the image evaluation system based on the acquired hyperspectral image. The multispectral detection and analysis system includes the design of the imaging system, the establishment of the classification model and the software development, and uses the image to identify the location of contamination attachment and evaluate the contamination level through SVM.

[0019] Preferably, auxiliary detectors are installed on both sides of the multi-spectral camera on the front of the drone, and the auxiliary detectors are temperature sensors, humidity sensors, pressure sensors and flammable gas concentration sensors.

[0020] Preferably, the multispectral camera and the auxiliary detector jointly obtain the surface information of the insulator, and the detection value obtained by the auxiliary detector is obtained by x1, x2, ..., x n Indicates that these test values ​​are all unbiased judgment results of the true value x, and at the same time, each test value is independent of each other. Indicates its corresponding variance, with X m (i) and w1, w2, …, w n Represent the detection value of the mth sensor at time i and the weights of different sensors, thereby determining the fused judgment value of the surrounding environment of the transmission line The sum weights must conform to the following formula.

[0021]

[0022] Preferably, the threshold value for detecting the external safety state of the insulator by the auxiliary detector is specifically set as follows:

[0023] Highly safe state: temperature not higher than 15°C, humidity not higher than 5%, pressure not higher than 30MPa, flammable gas concentration not higher than 80mg / m 3 ;

[0024] Moderate safety state: temperature not higher than 30℃, humidity not higher than 20%, pressure not higher than 60MPa, flammable gas concentration not higher than 120mg / m 3 ;

[0025] Critical safety state: temperature not higher than 45℃, humidity not higher than 35%, pressure not higher than 90MPa, flammable gas concentration not higher than 160mg / m 3 Moderately dangerous state: temperature not higher than 60℃, humidity not higher than 50%, pressure not higher than 120MPa, flammable gas concentration not higher than 190mg / m 3 ;

[0026] Highly dangerous state: temperature not higher than 75℃, humidity not higher than 65%, pressure not higher than 150MPa, flammable gas concentration not higher than 220mg / m 3 .

[0027] The present invention provides an insulator contamination level assessment method based on multispectral analysis. It has the following beneficial effects:

[0028] 1. The insulator contamination level assessment process of the present invention includes precise positioning of the drone body's mounted camera, multispectral camera image acquisition, image processing and image evaluation. The drone uses the multispectral camera mounted on the drone to capture images of the insulator's surroundings. The drone includes a flight trajectory control system, a flight attitude control system, and an automatic navigation and positioning calculation system. After accurately acquiring the insulator's position information, it will fly to the insulator's surroundings, replacing manual information collection.

[0029] 2. The present invention performs denoising and refraction correction preprocessing on the collected images to make the collected images clearer. A multispectral detection and analysis system is built into the image processing system. The multispectral detection and analysis system is a multispectral analysis system established by the image evaluation system based on the acquired hyperspectral images. The multispectral detection and analysis system includes the design of the imaging system, the establishment of the classification model and the development of software, and uses the support vector machine (SVM) to evaluate the contamination level by identifying the contamination attachment site in the image.

[0030] 3. The present invention evaluates the contamination on the surface of insulators mainly through multi-spectral image analysis and information fusion of auxiliary detectors. The contamination on the surface of insulators is evaluated based on the fused information. When evaluating the contamination on the surface of insulators, it is mainly divided into three levels according to the degree of pollution. At the same time, the system can detect and evaluate the safety status of the insulator through four auxiliary detectors during the evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a schematic diagram of the UAV body of the present invention;

[0032] Figure 2 It is a detection process framework diagram of the present invention;

[0033] Figure 3 This is a structural diagram of the main control system of the drone body of the present invention;

[0034] Figure 4 This is a schematic diagram of the multispectral camera detection process of the present invention;

[0035] Figure 5 This is a diagram showing the internal module structure of the multispectral camera of the present invention;

[0036] Figure 6 This is a flow chart of the insulator spectrum analysis system of the present invention;

[0037] Figure 7 This is a framework diagram of the online evaluation method of the present invention;

[0038] Figure 8 This is a denoising flowchart of the present invention;

[0039] Figure 9 This is a flowchart of the multi-source data fusion of the present invention. DETAILED DESCRIPTION

[0040] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention. Specific embodiment one:

[0042] like Figure 1-9 As shown in the figure, the insulator contamination level assessment method based on multispectral analysis is proposed. The insulator contamination level assessment process includes precise positioning of the drone-mounted camera, multispectral camera image acquisition, image processing, and image evaluation. The specific steps are as follows:

[0043] The main control system of the drone includes a flight trajectory control system, a flight attitude control system, and an automatic navigation and positioning calculation system. The drone is equipped with a multispectral camera, which is installed in the center of the drone. At the same time, temperature detection probes and humidity detection probes are installed on both sides of the drone. The drone will obtain the height image of the insulator in advance, and then the drone calculates the vertical level difference between itself and the insulator through the automatic navigation and positioning calculation system. Specifically, it calculates the image height h of the high-altitude high-voltage line A. a and the image height h of insulator B b As well as the image distance between the high-altitude high-voltage line A and the insulator B, the distance u between the high-altitude high-voltage line A and the insulator B and the drone a and u b As well as the angle θ between the high-altitude high-voltage line A and the insulator B relative to the imaging center, the safe distance between the drone and the transmission line plane is:

[0044]

[0045] SP2: When the drone's built-in computing system calculates the height of the insulator, it will transmit it to the drone's flight drive module, and then the drive module will control the drone body to automatically rise.

[0046] SP3: During flight, the drone automatically searches, inspects, and monitors various devices surrounding the insulators. During this process, station staff calibrate and control the drone's flight speed, position, and trajectory. This control process involves detailed calculations of roll and pitch angle control variables based on the drone's attitude stability. The flight trajectory is then controlled and adjusted based on these calculations, ensuring maximum stability during the drone's inspection flight and ensuring more stable operation of the multispectral camera.

[0047] SP4: The front-end imaging element of a multispectral camera mainly splices multiple independent filter substrates in sequence to obtain a filter array, and installs the filter array in front of the imaging device. When the multi-band light obtained by the filter array is dispersed to different imaging positions, multispectral images of corresponding bands in different strip areas are obtained. The imaging device is a planar array detector.

[0048] In the present invention, a multispectral camera is used to collect dirt images on the surface of the insulator, and then a drone is flown to the insulator. In order to facilitate stable collection of the spectral camera, a series of control methods for the drone are provided above to make the collection more stable. Specific embodiment two:

[0050] like Figure 1-9As shown in the figure, when the multispectral camera on the drone is close to the insulator, the light source is turned on and the light intensity and angle are adjusted according to the light reflection of the insulator surface to be evaluated. The parallel light reaches the surface of the insulator to be evaluated and generates reflected light. The multi-band image is used to collect the multi-band image of the reflected light on the insulator surface.

[0051] Before the multispectral camera is placed close to the insulator for image acquisition, it needs to be preheated for half an hour to maintain the stability of the light source. Parameters can be set after the preheating is complete. Due to the current of the device itself and the light source image in the environment, black and white correction is required to obtain hyperspectral images.

[0052] After acquiring images of the insulator surface, the multispectral camera transmits them to an image processing system. Principal component analysis (PCA) and minimum noise factoring (MNF) dimensionality reduction are performed on the full-band hyperspectral image to identify the most obvious contamination components on the insulator surface. Characteristic wavelengths are then screened using the Competitive Adaptive Reweighting (CARS) algorithm and the Successive Projection (SPA) algorithm, respectively, and the contamination adhesion level is classified using a support vector machine (SVM). Finally, images corresponding to the characteristic wavelengths are combined into a multi-wavelength image and subjected to an MNF transformation. The resulting image is similar to the full-band transformation, demonstrating that images based on characteristic wavelengths can replace full-band images for early detection of latent insulator contamination.

[0053] Image processing systems remove noise from image data. Image data collected by electronic devices is often contaminated with irrelevant signals and noise, which can overwhelm useful signals. Data preprocessing can clean up data, reduce noise, reduce data dimensionality, and address data type issues, providing a better data foundation for subsequent modeling and analysis. Wavelet denoising and convolution smoothing techniques are used to remove noise from image signals. Figure 8 This is a wavelet threshold denoising process. Based on the principle of wavelet transform, it decomposes the signal into multiple frequency sub-bands. Then, a threshold is applied to each sub-band signal to perform signal processing. This thresholding process reduces the impact of noise on the signal. Finally, the processed sub-band signal is subjected to an inverse wavelet transform to complete the signal denoising.

[0054] Convolution smoothing technology is an efficient data processing tool that uses the least squares method to process complex signals, thereby reducing random noise and retaining valuable information. It can effectively suppress the influence of "burrs", making the image smoother. The smoothing window width used in this method of the present invention is 2m+1, that is, the window contains the number of original data points mm (n=2m+1). Usually, the user needs to use an odd-numbered window in order to select a part of the measured original data set as a window instead of including it all. By performing SG convolution smoothing on each subset, the noise in the signal can be effectively reduced and a more accurate analysis result can be obtained. It is assumed that the original data points in the window can be fitted with a k-1 degree polynomial:

[0055] y i =a0+a1i+a2i 2 +…+a k-1 i k-1 . Specific embodiment three:

[0057] like Figure 1-9 As shown in the figure, if the image information noise is properly processed, the image processing system will perform refraction correction on the image. Using MSC for spectral analysis can effectively suppress the scattering phenomenon, thereby more accurately capturing the characteristics of each component in the "ideal spectrum" and improving the accuracy of the analysis. Therefore, in order to obtain more accurate results, it is necessary to first establish an "ideal spectrum" to better capture the characteristics of each component. This method can achieve ideal results by adjusting the NIRS of each sample, which can not only correct the baseline deviation, but also provide more accurate results. The algorithm analysis and implementation principle of MSC are shown in the following formula:

[0058] Compute the average spectrum:

[0059]

[0060] Univariate linear regression:

[0061]

[0062] Multi-source scattering correction:

[0063]

[0064] The matrix A is the average spectral vector based on the original spectra of all samples, which is a calibration spectrum matrix of n×p dimensions, where n is the number of samples, p is the number of wavelength points corresponding to the sample spectrum, and A i Represents the spectrum vector of a single sample, which is a 1×p-dimensional matrix. m i and b iIt represents the relative offset and translation between the NIRS of each sample and the average spectrum after univariate linear regression.

[0065] After correction, feature screening is performed. In the present invention, a competitive adaptive reweighting algorithm is selected for feature screening. Based on the idea of ​​survival of the fittest, a competitive adaptive reweighting sampling method is generated by respectively utilizing the feature screening method of the PLS-DA coefficient and Monte Carlo sampling.

[0066] The Monte Carlo sampling formula is as follows:

[0067] Specific embodiment four:

[0069] like Figure 1-9 As shown, a multispectral detection and analysis system is built into the image processing system. The multispectral detection and analysis system is a multispectral analysis system established by the image evaluation system based on the acquired hyperspectral image. The multispectral detection and analysis system includes the design of the imaging system, the establishment of the classification model and the development of software, and uses the support vector machine (SVM) to evaluate the contamination level by identifying the contamination attachment site in the image. Specific embodiment five:

[0071] like Figure 1-9 As shown, auxiliary detectors are installed on both sides of the multispectral camera on the front of the drone. The auxiliary detectors are temperature sensors, humidity sensors, pressure sensors and flammable gas concentration sensors.

[0072] The multispectral camera and the auxiliary detector jointly obtain the surface information of the insulator. The detection value obtained by the auxiliary detector is expressed by x1, x2, ..., x n Indicates that these test values ​​are all unbiased judgment results of the true value x, and at the same time, each test value is independent of each other. Indicates its corresponding variance, with X m (i) and w1, w2, …, w n Represent the detection value of the mth sensor at time i and the weights of different sensors, thereby determining the fused judgment value of the surrounding environment of the transmission line The sum weights must conform to the following formula.

[0073]

[0074] In the present invention, the assessment of insulator surface contamination is mainly carried out through multispectral image analysis and information fusion of auxiliary detectors. The insulator surface contamination is assessed based on the fused information. When evaluating the insulator surface contamination, it is mainly divided into three levels according to the degree of contamination:

[0075] Level I: The insulator surface is free of contamination or only slightly contaminated. The main contaminants at this level are fine dust and light scale, which can be wiped off with a fine sandpaper or cotton cloth.

[0076] Level II: There is moderate contamination on the surface of the insulator, such as wax, gum, or bird droppings. The main contaminants at this level are heavy scale and oil.

[0077] Level III: A large amount of dirt is deposited on the surface of the insulator. If not cleaned in time, it will cause a flashover accident. Some common pollutants include sea salt, sand, dust, and sludge. Specific embodiment six:

[0079] like Figure 1-9 As shown, the auxiliary detectors in the present invention are installed on both sides of the front of the drone with a dedicated power supply system. During the evaluation, the safety status of the insulator can be detected and evaluated through four auxiliary detectors. The threshold value of the external safety status of the insulator detected by the auxiliary detector is specifically set as follows:

[0080] Highly safe state: temperature not higher than 15°C, humidity not higher than 5%, pressure not higher than 30MPa, flammable gas concentration not higher than 80mg / m 3 ;

[0081] Moderate safety state: temperature not higher than 30℃, humidity not higher than 20%, pressure not higher than 60MPa, flammable gas concentration not higher than 120mg / m 3 ;

[0082] Critical safety state: temperature not higher than 45℃, humidity not higher than 35%, pressure not higher than 90MPa, flammable gas concentration not higher than 160mg / m 3 ;

[0083] Moderately dangerous state: temperature not higher than 60℃, humidity not higher than 50%, pressure not higher than 120MPa, flammable gas concentration not higher than 190mg / m 3 ;

[0084] Highly dangerous state: temperature not higher than 75℃, humidity not higher than 65%, pressure not higher than 150MPa, flammable gas concentration not higher than 220mg / m 3 .

[0085] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further restrictions, an element defined by the statement "comprising a reference structure" does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.

[0086] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. The insulator contamination level assessment method based on multispectral analysis is characterized by: The insulator pollution level assessment process includes precise positioning of the drone-mounted camera, multispectral camera image acquisition, image processing, and image evaluation. The specific steps are as follows: SP1: The main control of the drone body includes a flight trajectory control system, a flight attitude control system and an automatic navigation and positioning calculation system. The drone is equipped with a multispectral camera, which is installed in the center of the drone. At the same time, temperature detection probes and humidity detection probes are installed on both sides of the drone. The drone will obtain the height image of the insulator in advance, and then the drone calculates the vertical level difference from itself to the insulator through the automatic navigation and positioning calculation system, specifically calculating the image height of the high-altitude high-voltage line A. and the image height of insulator B The image distance between the high-altitude high-voltage line A and the insulator B, and the distance between the high-altitude high-voltage line A and the insulator B and the drone and And the angle between the high-voltage line A and the insulator B relative to the imaging center , the safe distance between the UAV and the transmission line plane is: ； SP2: When the drone's built-in computing system calculates the height of the insulator, it transmits the information to the drone's flight drive module, which then controls the drone's body to automatically ascend. SP3: When the drone is in flight, it will automatically search, detect, and monitor various equipment around the insulator. During this process, the station staff will calibrate and control the flight speed, flight position, and flight trajectory of the drone. During the control process, the roll angle and pitch angle control variables will be calculated in detail based on the attitude stability of the drone. The flight trajectory will be controlled and adjusted based on the calculation results to maximize the stability of the drone's inspection flight and make the multispectral camera more stable during operation. SP4: The front-end imaging element of the multispectral camera is a filter array obtained by splicing multiple independent filter substrates in sequence. The filter array is installed in front of the imaging device. When the multi-band light obtained by the light splitting by the filter array is dispersed to different imaging positions, multispectral images of corresponding bands in different strip areas are obtained. The imaging device is a planar array detector.

2. The insulator contamination level assessment method based on multispectral analysis according to claim 1, characterized in that: When the multispectral camera on the drone is close to the insulator, the light source is turned on and the light intensity and angle are adjusted according to the light reflection of the surface of the insulator to be evaluated. Parallel light reaches the surface of the insulator to be evaluated and generates reflected light. Multi-band images are used to collect multi-band images of the reflected light on the insulator surface.

3. The insulator contamination level assessment method based on multispectral analysis according to claim 1, characterized in that: The multispectral camera needs to be preheated for half an hour before being placed close to the insulator for image acquisition to maintain the stability of the light source. Parameters can be set after the preheating is completed. Due to the current of the device itself and the light source image in the environment, black and white correction is required to obtain hyperspectral images.

4. The insulator contamination level assessment method based on multispectral analysis according to claim 1, characterized in that: After acquiring images of the insulator surface, the multispectral camera transmits them to an image processing system, which performs principal component analysis (PCA) and minimum noise factoring (MNF) dimensionality reduction analysis on the full-band hyperspectral image to identify the most obvious contamination components on the insulator surface. Feature wavelengths are then screened based on the competitive adaptive reweighting (CARS) algorithm and the successive projections (SPA) algorithm, and the contamination adhesion degree is classified using a support vector machine (SVM).

5. The insulator contamination level assessment method based on multispectral analysis according to claim 4 is characterized in that: The image processing system performs denoising on the image data and removes noise in the image signal through wavelet denoising and convolution smoothing technology.

6. The insulator contamination level assessment method based on multispectral analysis according to claim 5, characterized in that: After the image information noise is processed, the image processing system will perform refraction correction on the image and use MSC for spectral analysis to effectively suppress scattering, thereby more accurately capturing the characteristics of each component in the "ideal spectrum" and improving the accuracy of the analysis.

7. The insulator contamination level assessment method based on multispectral analysis according to claim 4, characterized in that: A multispectral detection and analysis system is built into the image processing system. The multispectral detection and analysis system is a multispectral analysis system established by the image evaluation system based on the acquired hyperspectral images. The multispectral detection and analysis system includes the design of the imaging system, the establishment of the classification model and the development of software. It also uses the support vector machine (SVM) to evaluate the contamination level by identifying the location of contamination attachment in the image.

8. The insulator contamination level assessment method based on multispectral analysis according to claim 1, characterized in that: Auxiliary detectors are installed on both sides of the multi-spectral camera on the front of the drone. The auxiliary detectors are temperature sensors, humidity sensors, pressure sensors and flammable gas concentration sensors.

9. The method for evaluating insulator contamination levels based on multispectral analysis according to claim 8, characterized in that: The multi-spectral camera and the auxiliary detector jointly obtain the surface information of the insulator, and the detection value obtained by the auxiliary detector is obtained by , ,…, Indicates that these test values ​​are true values Unbiased judgment results, at the same time, each test value is independent of each other, Indicates its corresponding variance, , ,…, Respectively The detection value of the mth sensor at the moment and the weights of different sensors are used to determine the fused judgment value of the surrounding environment of the transmission line The sum weights must conform to the following formula: .

10. The insulator contamination level assessment method based on multispectral analysis according to claim 9, characterized in that: The auxiliary detector detects the insulator external safety state threshold value specifically set as follows: Highly safe state: temperature not higher than 15°C, humidity not higher than 5%, pressure not higher than 30MPa, flammable gas concentration not higher than 80mg / m³; Moderate safety state: temperature not higher than 30°C, humidity not higher than 20%, pressure not higher than 60MPa, flammable gas concentration not higher than 120mg / m³; Critical safety state: temperature not higher than 45°C, humidity not higher than 35%, pressure not higher than 90MPa, flammable gas concentration not higher than 160mg / m³; moderate danger state: temperature not higher than 60°C, humidity not higher than 50%, pressure not higher than 120MPa, flammable gas concentration not higher than 190mg / m³; Highly dangerous state: temperature not higher than 75 ℃, humidity not higher than 65%, pressure not higher than 150MPa, flammable gas concentration not higher than 220mg / m³.

Images