Large model-based DC converter valve equipment image detection method, device, equipment and storage medium

Abstract

The present application discloses a method, device, equipment and storage medium for image detection of DC professional converter valve equipment based on a large model, and relates to the technical field of converter valve detection. The present application comprises: acquiring an imaging image of a DC professional converter valve equipment, performing image segmentation on the imaging image, and obtaining a local imaging image of the DC professional converter valve equipment; performing feature extraction on the local imaging image, and obtaining an electric field distribution and a thermal field distribution; filtering the electric field distribution and the thermal field distribution respectively, and obtaining filtering results, and determining an insulation distance according to the filtering results; when the insulation distance is less than a safety distance, performing fault identification on the filtering results, determining a fault point, and outputting the fault point, thereby realizing image-based detection of the converter valve equipment, determining the operating status of the converter valve equipment, and outputting the fault when an abnormal point is determined.

Description

Technical Field

[0001] The present application relates to the field of converter valve detection technology, and in particular to a method, device, equipment and storage medium for image detection of DC professional converter valve equipment based on a large model. Background Art

[0002] The converter valve equipment is the core equipment of the DC transmission system. It operates in a high voltage and strong electromagnetic environment for a long time. Its safe and stable operation is crucial to the reliability of the entire power system. The failure of the converter valve equipment will not only cause the shutdown of DC transmission, but may even cause major safety accidents. Real-time monitoring of the converter valve equipment can ensure the reliability, availability and operation and maintenance efficiency of the system. For problems that may occur in the converter valve equipment, such as abnormal temperature and discharge, the traditional converter valve inspection method is manual inspection. Manual inspection has the problems of harsh environment, high labor intensity, high risk and low efficiency. Therefore, the current converter valve detection method has the disadvantage of low detection efficiency.

[0003] The above contents are only used to assist in understanding the technical solution of the present application and do not constitute an admission that the above contents are prior art. Summary of the invention

[0004] The main purpose of the present application is to provide a method, device, equipment and storage medium for image detection of DC professional converter valve equipment based on a large model, aiming to solve the technical problem of low detection efficiency of manual inspection methods in the prior art.

[0005] To achieve the above objectives, the present application proposes a large-model-based DC professional converter valve equipment image detection method, the large-model-based DC professional converter valve equipment image detection method comprising:

[0006] Acquire an imaging image of a DC professional converter valve device, perform image segmentation on the imaging image, and obtain a local imaging image of the DC professional converter valve device;

[0007] Extracting features from the local imaging image to obtain electric field distribution and thermal field distribution;

[0008] filtering the electric field distribution and the thermal field distribution respectively to obtain filtering results, and determining the insulation distance according to the filtering results;

[0009] When the insulation distance is less than the safety distance, fault identification is performed on the filtering result, a fault point is determined, and the fault point is output.

[0010] In one embodiment, the step of acquiring an image of a DC professional converter valve device, performing image segmentation on the image, and obtaining a local image of the DC professional converter valve device comprises:

[0011] Acquire an imaging image of a DC professional converter valve device, input the imaging image into a convolutional layer, and obtain an initial feature map based on the imaging image;

[0012] Performing a deconvolution upsampling process on the initial feature map to obtain a target feature map with the same size as the initial feature map;

[0013] The target feature map is segmented to obtain a local imaging image of the DC professional converter valve equipment.

[0014] In one embodiment, the step of extracting features from the local imaging image to obtain electric field distribution and thermal field distribution includes:

[0015] Preprocessing the local imaging image to obtain a preprocessed image, and randomly generating an initial image set from the preprocessed image;

[0016] Performing linear interpolation on the initial image set to obtain a real number vector of the initial image set;

[0017] Determining the fitness in the initial image set according to the feature weight and the real number vector;

[0018] Determine a selection probability according to the fitness, and perform mutation processing on the real number vector based on the selection probability and the fitness to obtain electric field characteristics and thermal field characteristics of a local imaging image;

[0019] Obtaining electric field distribution according to the electric field characteristics of the local imaging image;

[0020] The thermal field distribution is obtained according to the thermal field characteristics of the local imaging image.

[0021] In one embodiment, the step of determining the selection probability according to the fitness, performing mutation processing on the real number vector based on the selection probability and the fitness, and obtaining the electric field characteristics and thermal field characteristics of the local imaging image includes:

[0022] Determining an imaging type of the initial image set, and dividing the initial image set into an infrared image set and an ultraviolet image set according to the imaging type;

[0023] Determining the fitness of the infrared image set and the ultraviolet image set, wherein the fitness includes individual fitness and average fitness;

[0024] Determining a selection probability according to the individual fitness and the average fitness;

[0025] Determine the real number code of the local image in the infrared image set or the ultraviolet image set, and cross-process the fitness with the real number vector to obtain a cross-image set;

[0026] A critical value in the individual fitness is determined, and a variation process is performed based on the cross image set and the critical value to obtain the electric field characteristics and thermal field characteristics of the local imaging image.

[0027] In one embodiment, the steps of filtering the electric field distribution and the thermal field distribution respectively to obtain filtering results, and determining the insulation distance according to the filtering results include:

[0028] Determine a local heating point in the thermal field distribution, determine a discharge position in the electric field distribution, and determine a distance measurement error between the discharge position and a true position;

[0029] Determining a model reference of a filtering space based on the ranging error;

[0030] Filtering is performed according to the model benchmark and the electric field distribution to obtain a covariance matrix;

[0031] Performing gain processing on the covariance matrix to obtain a gain matrix;

[0032] The gain matrix is ​​updated to obtain a filtering result;

[0033] The electric field strength is determined according to the filtering result, and the insulation distance is determined according to the discharge position and the local heating point.

[0034] In one embodiment, when the insulation distance is less than the safety distance, the steps of performing fault identification on the filtering result, determining the fault point, and outputting the fault point include:

[0035] When the insulation distance is less than the safety distance, analyzing the filtering result, determining the discharge position and the electric field strength, and determining the abnormal discharge position according to the discharge position and the electric field strength;

[0036] Determine the temperature of a local heating point, and when the temperature of the local heating point is greater than a preset temperature, determine the local heating point as an abnormal heating position;

[0037] The abnormal discharge position and the abnormal heating position are determined as fault points, and the fault points are output visually.

[0038] In one embodiment, the step of filtering the electric field distribution and the thermal field distribution respectively to obtain filtering results, and determining the insulation distance according to the filtering results further includes:

[0039] When the insulation distance is greater than the safety distance, determining a redundant distance between the insulation distance and the safety distance;

[0040] Sorting the redundant distances, and assigning monitoring weights according to the sorting results, setting the minimum monitoring weight for the maximum redundant distance among the redundant distances, and setting the maximum monitoring weight for the minimum redundant distance among the redundant distances;

[0041] The insulation distance is detected in real time based on the monitoring weight.

[0042] In addition, to achieve the above-mentioned purpose, the present application also proposes a large-model-based DC professional converter valve equipment image detection device, the large-model-based DC professional converter valve equipment image detection device comprising:

[0043] An image segmentation module is used to obtain an imaging image of a DC professional converter valve device, and perform image segmentation on the imaging image to obtain a local imaging image of the DC professional converter valve device;

[0044] A feature extraction module is used to extract features from the local imaging image to obtain electric field distribution and thermal field distribution;

[0045] A field filtering module, used to filter the electric field distribution and the thermal field distribution respectively to obtain filtering results, and determine the insulation distance according to the filtering results;

[0046] The fault output module is used to perform fault identification on the filtering result, determine the fault point, and output the fault point when the insulation distance is less than the safety distance.

[0047] In addition, to achieve the above-mentioned objectives, the present application also proposes a large-model-based DC professional converter valve equipment image detection device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the large-model-based DC professional converter valve equipment image detection method as described above.

[0048] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by the processor, the steps of the large model-based DC professional converter valve equipment image detection method are implemented as described above.

[0049] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the steps of the large model-based DC professional converter valve equipment image detection method as described above.

[0050] One or more technical solutions proposed in the present application have at least the following technical effects: acquiring an imaging image of a DC professional converter valve device, performing image segmentation on the imaging image, and obtaining a local imaging image of the DC professional converter valve device; performing feature extraction on the local imaging image to obtain electric field distribution and thermal field distribution; filtering the electric field distribution and thermal field distribution respectively to obtain filtering results, and determining the insulation distance according to the filtering results; when the insulation distance is less than the safety distance, performing fault identification on the filtering results, determining the fault point, and outputting the fault point, thereby realizing image-based detection of the converter valve device, determining the operating status of the converter valve device, and outputting the fault when the abnormal point is determined. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0053] Figure 1 A flow chart of the first embodiment of the large-model-based DC professional converter valve equipment image detection method provided in this application;

[0054] Figure 2 A schematic diagram of a local imaging image generation process provided in an embodiment of a method for image detection of DC professional converter valve equipment based on a large model of the present application;

[0055] Figure 3 A schematic diagram of an electric power device obtained by ultraviolet imaging provided in an embodiment of a method for detecting an image of a DC professional converter valve device based on a large model of the present application;

[0056] Figure 4 A schematic diagram of an electric power device obtained by infrared imaging provided in an embodiment of a method for detecting an image of a DC professional converter valve device based on a large model of the present application;

[0057] Figure 5 This is a schematic diagram of the module structure of a large-scale model-based image detection device for DC professional converter valve equipment according to an embodiment of the present application;

[0058] Figure 6 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the large model-based DC professional converter valve equipment image detection method in the embodiment of the present application.

[0059] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0060] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0061] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0062] The main solution of the embodiment of the present application is: obtaining an imaging image of a DC professional converter valve device, performing image segmentation on the imaging image, and obtaining a local imaging image of the DC professional converter valve device; performing feature extraction on the local imaging image to obtain electric field distribution and thermal field distribution; filtering the electric field distribution and thermal field distribution respectively to obtain filtering results, and determining the insulation distance according to the filtering results; when the insulation distance is less than the safety distance, performing fault identification on the filtering result, determining the fault point, and outputting the fault point.

[0063] In this embodiment, for ease of description, the following description is made by taking the identification of a large-model-based DC professional converter valve equipment image detection device as the execution subject.

[0064] Since the traditional inspection method of converter valves in the prior art is manual inspection, manual inspection has the problems of harsh environment, high labor intensity, high risk and low efficiency, the current converter valve detection method has the disadvantage of low detection efficiency, which is not conducive to the special working environment of converter valve inspection.

[0065] The present application provides a solution that can capture images of the converter valve equipment in a large model, replacing manual labor with machines. Firstly, it can reduce the operational risks caused by manual inspections, and secondly, it can have a more comprehensive and accurate inspection efficiency compared to manual inspections.

[0066] It can be seen from the above embodiments that the present application obtains an imaging image of a DC professional converter valve device, performs image segmentation on the imaging image, and obtains a local imaging image of the DC professional converter valve device; performs feature extraction on the local imaging image to obtain an electric field distribution and a thermal field distribution; respectively filters the electric field distribution and the thermal field distribution to obtain filtering results, and determines the insulation distance according to the filtering results; when the insulation distance is less than the safety distance, performs fault identification on the filtering result, determines the fault point, and outputs the fault point, thereby realizing image-based detection of the converter valve device, determining the operating status of the converter valve device, and outputting the fault when the abnormal point is determined.

[0067] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device capable of realizing the above functions, a DC professional converter valve equipment image detection device based on a large model, etc. The following takes the DC professional converter valve equipment image detection device based on a large model as an example to illustrate this embodiment and the following embodiments.

[0068] Based on this, the embodiment of the present application provides a method for detecting images of DC professional converter valve equipment based on a large model, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the large model-based DC professional converter valve equipment image detection method of the present application.

[0069] In this embodiment, the large model-based DC professional converter valve equipment image detection method includes steps S10 to S40:

[0070] Step S10, acquiring an imaging image of the DC professional converter valve device, performing image segmentation on the imaging image, and obtaining a local imaging image of the DC professional converter valve device.

[0071] It should be noted that the converter valve is the core equipment in the DC transmission project. Its main function is to convert three-phase AC voltage into DC voltage and realize precise control of power. The converter valve is composed of multiple thyristor devices (usually thyristors). By controlling the trigger angle of these devices, the current is conducted and cut off, thus playing a key role in the high-voltage DC transmission system (HVDC). The DC professional converter valve equipment is composed of converter valve tower, valve tower hardware, valve tower suspension insulator, valve tower mother pipe, equalizing ball and other components. The imaging image of the DC professional converter valve equipment refers to the imaging image of the entire DC professional converter valve equipment as a whole, which is generated by the image generation device performing image scanning, image acquisition and other processes on the DC professional converter valve equipment, and the local imaging image is the image information of the components of the DC professional converter valve equipment, such as the local imaging image of the converter valve tower, the local imaging image of the valve tower hardware, etc.

[0072] In the specific implementation, since the working position of the DC professional converter valve equipment is relatively special and the equipment itself is highly dangerous when working, when performing image acquisition or image modeling of the DC professional converter valve equipment, the DC professional converter valve equipment can be acquired in all directions by means of a drone, wherein the drone is equipped with an infrared imager and an ultraviolet imager. The image formed by the infrared imager can describe the working temperature of each component of the DC professional converter valve, and the image formed by the ultraviolet imager can describe the working electric field of each component of the DC professional converter valve. After obtaining the imaging image of the DC professional converter valve equipment, the image can be segmented by a deep learning-based method, and segmented based on the convolutional neural network CNN and the fully convolutional network FCN, and the imaging image can be divided into multiple local imaging images according to the components of the converter valve equipment.

[0073] In a feasible implementation manner, the steps of acquiring an imaging image of a DC professional converter valve device, performing image segmentation on the imaging image, and obtaining a local imaging image of the DC professional converter valve device include:

[0074] Acquire an imaging image of a DC professional converter valve device, input the imaging image into a convolutional layer, and obtain an initial feature map based on the imaging image;

[0075] Performing a deconvolution upsampling process on the initial feature map to obtain a target feature map with the same size as the initial feature map;

[0076] The target feature map is segmented to obtain a local imaging image of the DC professional converter valve equipment.

[0077] It should be noted that the initial feature map refers to the feature map obtained after the imaging image is convolved by the convolution layer, which is obtained by processing the input image with the convolution kernel of the convolution layer. The target feature map refers to the feature map obtained after deconvolution and upsampling of the initial feature map.

[0078] In the specific implementation, first, images of DC professional converter valve equipment are obtained through imaging equipment. These images are then input into the convolution layer of the convolutional neural network. In the convolution layer, the image is convolved by a filter (convolution kernel) to extract the preliminary feature map of the image. This step is the beginning of feature extraction. The convolution kernel slides on the image and calculates the weighted sum of the local area to generate feature maps, and the initial feature map is deconvolved and upsampled. Deconvolution is a common technique used to increase the spatial resolution of feature maps, that is, to restore from low resolution to high resolution. The deconvolution operation can be performed in a learning manner so that the network can better adapt to specific upsampling tasks. The key to the deconvolution operation is that it can learn how to perform upsampling optimally through training. The parameters of the deconvolution, such as the convolution kernel size, stride, padding, and output padding, will affect the final output size. After the deconvolution upsampling process, a target feature map with the same size as the initial feature map is obtained. This target feature map has a higher spatial resolution and is more suitable for pixel-level image segmentation tasks. Finally, these high-resolution feature maps are used for image segmentation to segment the local imaging images of the DC professional converter valve equipment. Figure 2 When performing image segmentation, the size of the imaged image can be unlimited, and image input of any scale can be supported. When performing convolution, the input image can be reduced in size, and the input image can be convolved to obtain features in the imaged image, thereby obtaining an initial feature map, which is then enlarged through deconvolution upsampling, and the reduced initial feature map is enlarged to obtain a feature map with the same size as the input image, namely, a target feature map. Image segmentation is performed on the target feature map to obtain a local imaging image of the DC professional converter valve equipment.

[0079] Step S20, extracting features from the local imaging image to obtain electric field distribution and thermal field distribution.

[0080] It should be noted that the electric field distribution describes the electric field formed by the existence of charges in space. The electric field distribution is an electric field distribution cloud map, which is used to describe the distribution of the electric field. The thermal field distribution refers to the collection of temperatures at each point in the material system.

[0081] In a specific implementation, when extracting features from a local imaging image, feature extraction can be performed through a genetic algorithm, which can preprocess the local imaging image, clean the local imaging image, filter out the influence of noise, obtain a preprocessed image set, generate a population based on the preprocessed image set, perform crossover and mutation on the individuals in the population, determine the fitness of each individual, and determine the optimal individual in the population based on the fitness, so as to output an image database and obtain the electric field distribution and thermal field distribution corresponding to the electric field and thermal field.

[0082] In a feasible implementation manner, the step of extracting features from the local imaging image to obtain electric field distribution and thermal field distribution includes:

[0083] Preprocessing the local imaging image to obtain a preprocessed image, and randomly generating an initial image set from the preprocessed image;

[0084] Performing linear interpolation on the initial image set to obtain a real number vector of the initial image set;

[0085] Determining the fitness in the initial image set according to the feature weight and the real number vector;

[0086] Determine a selection probability according to the fitness, and perform mutation processing on the real number vector based on the selection probability and the fitness to obtain electric field characteristics and thermal field characteristics of a local imaging image;

[0087] Obtaining electric field distribution according to the electric field characteristics of the local imaging image;

[0088] The thermal field distribution is obtained according to the thermal field characteristics of the local imaging image.

[0089] It should be noted that the real number vector is obtained by image conversion and is used to describe the chromosomes of the individuals in the population formed by the initial image set, and the fitness refers to the performance used to describe each image.

[0090] In the specific implementation, the local imaging image is preprocessed, the local imaging image is normalized, and the local imaging image is preprocessed by Gaussian denoising and other preprocessing operations to obtain a preprocessed image, and the initial image set is randomly generated with the preprocessed image, and a population is generated based on the initial image set, and the population size is , randomly generated The initial population of individuals Given a data selection range, since the determination of the initial population has a great influence on the global optimization of the (genetic algorithm) GA, a linear interpolation function is used to generate the individuals in the population A real vector of As a chromosome of the genetic algorithm, the chromosome length is .in, is the number of input layer nodes, is the number of hidden layer nodes, is the number of nodes in the output layer. Determine each individual in the population Represents the initial value of a BP neural network, individual A gene value in It represents a connection weight of a neural network. In order to obtain high-precision weights and shorten the length of the chromosome string, a floating-point encoding method is used.

[0091] Determine the evaluation function of the individual. Given a BP neural network evolution parameter, assign the chromosome obtained in step 1 to the BP neural network weights and thresholds, input the training sample to train the neural network, and obtain a network training output value when the set accuracy is achieved. , then the population Medium The fitness value of and the average fitness value .

[0092]

[0093]

[0094] in, is the training output value, Output expected value for training, is the number of phase points in the reconstructed phase space, For population size.

[0095] In the step of determining the selection probability according to the fitness, performing mutation processing on the real number vector based on the selection probability and the fitness, and obtaining the electric field characteristics and thermal field characteristics of the local imaging image, the method includes:

[0096] Determining an imaging type of the initial image set, and dividing the initial image set into an infrared image set and an ultraviolet image set according to the imaging type;

[0097] Determining the fitness of the infrared image set and the ultraviolet image set, wherein the fitness includes individual fitness and average fitness;

[0098] Determining a selection probability according to the individual fitness and the average fitness;

[0099] Determine the real number code of the local image in the infrared image set or the ultraviolet image set, and cross-process the fitness with the real number vector to obtain a cross-image set;

[0100] A critical value in the individual fitness is determined, and a variation process is performed based on the cross image set and the critical value to obtain the electric field characteristics and thermal field characteristics of the local imaging image.

[0101] The roulette wheel selection operator is used, that is, the selection strategy based on the fitness ratio is used to select chromosomes in each generation of the population. The selection probability is .

[0102]

[0103]

[0104] Since individuals are encoded using entity codes, real number crossover is used when performing crossover. and the first gene The crossover operation at position j is:

[0105]

[0106] Where b is a random number in [0,1].

[0107] Select the jth gene of the i-th individual for mutation operation:

[0108]

[0109]

[0110] in, and Gene The upper and lower limits of the value, r is a random number in [0,1], is a random number, g is the current iteration number, is the maximum number of evolutions.

[0111] The optimal individual of the genetic algorithm is decomposed into the connection weights and thresholds of the BP neural network. The BP algorithm is used to train the BP neural network prediction model to obtain the optimal solution for chaotic time series prediction. The fuzzy neural network can be learned through a hybrid algorithm of the BP algorithm and the least squares estimation method to adjust the antecedent and consequent parameters of the system. The forward stage is calculated to the 4th layer, and the error signal is transmitted backwards in the reverse stage. The forward parameters are updated using the BP algorithm. Based on the intelligent image processing topology of GA and BP neural network, the weights and thresholds of the probability density function are first set, and deep learning training is performed in combination with the imaging images of the DC professional converter valve equipment to obtain the weights and thresholds after iterative adjustment of the probability density function. Finally, the image database is output to obtain the electric field and potential distribution cloud map. After that, the obtained electric field distribution and thermal field distribution component database, the obtained power equipment schematic diagram obtained by ultraviolet imaging, and the obtained power equipment schematic diagram obtained by infrared imaging can be obtained. Figure 3 and Figure 4 As shown. Among them, Figure 3 (a) is the ultraviolet imaging image of the converter valve tower. Figure 3 (b) is the UV imaging image of the valve tower fittings. Figure 3 (c) is the UV imaging image of the valve tower suspension insulator. Figure 3(d) in the figure is the ultraviolet imaging image of the valve tower mother pipe. Figure 4 (a) is the infrared imaging image of the conductor fittings. Figure 4 (b) is the infrared imaging image of the casing end. Figure 4 (c) is the infrared imaging image of the tube busbar. Figure 4 (d) is the infrared imaging image of the top of the valve tower. Figure 4 (e) is the infrared imaging image of the bottom of the valve tower. Figure 4 (f) in the figure is the infrared imaging image of the pressure equalizing ball. Then, the imaging images of each part and the electric field distribution and thermal field distribution of the corresponding position are obtained.

[0112] Step S30, filtering the electric field distribution and the thermal field distribution respectively to obtain filtering results, and determining the insulation distance according to the filtering results.

[0113] In a specific implementation, the electric field distribution and the thermal field distribution are filtered respectively to obtain filtering results, and the insulation distance is determined according to the filtering results, which specifically includes: determining a local heating point in the thermal field distribution, determining a discharge position in the electric field distribution, and determining a ranging error between the discharge position and a true position; determining a model reference of the filtering space based on the ranging error; filtering according to the model reference and the electric field distribution to obtain a covariance matrix; performing gain processing on the covariance matrix to obtain a gain matrix; updating the state of the gain matrix to obtain a filtering result; and determining the electric field strength, the discharge position, and the local heating point according to the filtering result to determine the insulation distance.

[0114] In the specific implementation, the electric field distribution and the thermal field distribution are filtered, and the Kalman filter is used to process the noise in the electric field distribution and the thermal field distribution. and , and predict the position at the kth moment based on the position at the k-1th moment, creating a dynamic description of the spatial model:

[0115]

[0116] in, is the state prediction at time k+1, is the state transfer matrix, which describes how to transfer from time k to time k+1.

[0117] Further predict the covariance matrix;

[0118]

[0119] in, is the covariance prediction matrix at time k+1, represents the gain of the process noise covariance matrix, and Q is the process noise covariance matrix.

[0120] Obtain the filter gain matrix to determine the local heating points in the thermal field distribution and the discharge positions in the electric field distribution:

[0121]

[0122] in, is the gain matrix, H is the measurement matrix, and R is the observation noise covariance matrix.

[0123] Filter with the model reference in the filter space and update the covariance matrix state:

[0124]

[0125]

[0126] in, is the residual at time k+1, It is the state update at time k+1, based on the observation information at time k+1.

[0127] The expression after updating the state is:

[0128]

[0129]

[0130] in, is the identity matrix, is the covariance update matrix at time k+1, is the initial state estimate, is the initial covariance matrix.

[0131] After obtaining the filtering results, the electric field strength, discharge position and local heating point are directly determined based on the filtering results to determine the insulation distance.

[0132] Step S40, when the insulation distance is less than the safety distance, fault identification is performed on the filtering result, a fault point is determined, and the fault point is output.

[0133] It should be noted that the safe distance refers to the limited distance within which the discharge range is within a controllable range and will not affect power transmission.

[0134] In a specific implementation, when the insulation distance is less than the safety distance, the filtering result is parsed to determine the discharge position and the electric field strength, and the abnormal discharge position is determined according to the discharge position and the electric field strength; the temperature of the local heating point is determined, and when the temperature of the local heating point is greater than the preset temperature, the local heating point is determined as the abnormal heating position; the abnormal discharge position and the abnormal heating position are determined as the fault point, and the fault point is output visually. Specifically: when it is determined that the insulation distance is less than the safety distance, the filtering result is parsed to determine the discharge position and the electric field strength, and the abnormal discharge position is determined in combination with the discharge position and the electric field strength, and when the electric field strength of the discharge position is greater than the preset electric field strength, the current discharge position is determined as the abnormal discharge position, and the current position is marked, and when the local heating point is detected, when the local heating point is greater than the preset temperature, the position of the abnormal heating point is marked, and both the abnormal discharge position and the abnormal heating position are determined as fault points, and are distinguished by different marks, and the fault point information is output in a visual form, and solutions are automatically generated when conditions permit.

[0135] In a feasible implementation manner, the step of filtering the electric field distribution and the thermal field distribution respectively to obtain filtering results, and determining the insulation distance according to the filtering results further includes:

[0136] When the insulation distance is greater than the safety distance, determining a redundant distance between the insulation distance and the safety distance;

[0137] Sorting the redundant distances, and assigning monitoring weights according to the sorting results, setting the minimum monitoring weight for the maximum redundant distance among the redundant distances, and setting the maximum monitoring weight for the minimum redundant distance among the redundant distances;

[0138] The insulation distance is detected in real time based on the monitoring weight.

[0139] In a specific implementation, when the insulation distance is greater than the safety distance, the redundant distance between the insulation distance and the safety distance is determined, that is, the danger level is determined, that is, the farther the insulation distance is from the safety distance, the higher the safety level is. Therefore, the detection attention to this part can be appropriately reduced. When determining the detection attention, the redundant distances can be sorted, and monitoring weights can be assigned according to the sorting results. The maximum redundant distance in the redundant distances is set to the minimum monitoring weight, and the minimum redundant distance in the redundant distances is set to the maximum monitoring weight. The insulation distance is detected in real time based on the monitoring weights.

[0140] The present embodiment provides an image detection method for DC professional converter valve equipment based on a large model. The method comprises the following steps: acquiring an imaging image of the DC professional converter valve equipment, performing image segmentation on the imaging image, and obtaining a local imaging image of the DC professional converter valve equipment; performing feature extraction on the local imaging image, and obtaining an electric field distribution and a thermal field distribution; filtering the electric field distribution and the thermal field distribution respectively to obtain filtering results, and determining an insulation distance according to the filtering results; when the insulation distance is less than a safety distance, performing fault identification on the filtering results, determining a fault point, and outputting the fault point, thereby realizing image-based detection of the converter valve equipment, determining the operating status of the converter valve equipment, and outputting the fault when an abnormal point is determined.

[0141] This application also provides a DC professional converter valve equipment image detection device based on a large model, please refer to Figure 5 , the large model-based DC professional converter valve equipment image detection device includes:

[0142] An image segmentation module 10 is used to obtain an image of a DC professional converter valve device, and perform image segmentation on the image to obtain a local image of the DC professional converter valve device;

[0143] A feature extraction module 20 is used to extract features from the local imaging image to obtain electric field distribution and thermal field distribution;

[0144] A field filtering module 30, for filtering the electric field distribution and the thermal field distribution respectively to obtain filtering results, and determining the insulation distance according to the filtering results;

[0145] The fault output module 40 is used to perform fault identification on the filtering result, determine the fault point, and output the fault point when the insulation distance is less than the safety distance.

[0146] In one embodiment, the image segmentation module 10 is also used to obtain an imaging image of a DC professional converter valve device, input the imaging image into a convolution layer, and obtain an initial feature map based on the imaging image; perform deconvolution upsampling processing on the initial feature map to obtain a target feature map of the same size as the initial feature map; perform image segmentation on the target feature map to obtain a local imaging image of the DC professional converter valve device.

[0147] In one embodiment, the feature extraction module 20 is also used to preprocess the local imaging image to obtain a preprocessed image, and randomly generate an initial image set from the preprocessed image; perform linear interpolation on the initial image set to obtain a real vector of the initial image set; determine the fitness of the initial image set based on the feature weight and the real vector; determine the selection probability based on the fitness, and perform mutation processing on the real vector based on the selection probability and the fitness to obtain electric field characteristics and thermal field characteristics of the local imaging image; obtain electric field distribution based on the electric field characteristics of the local imaging image; obtain thermal field distribution based on the thermal field characteristics of the local imaging image.

[0148] In one embodiment, the feature extraction module 20 is also used to determine the imaging type of the initial image set, and divide the initial image set into an infrared image set and an ultraviolet image set according to the imaging type; determine the fitness of the infrared image set and the ultraviolet image set, and the fitness includes individual fitness and average fitness; determine the selection probability according to the individual fitness and the average fitness; determine the real number encoding of the local image in the infrared image set or the ultraviolet image set, cross-process the fitness with the real number vector to obtain a cross-image set; determine the critical value in the individual fitness, perform mutation processing based on the cross-image set and the critical value, and obtain the electric field characteristics and thermal field characteristics of the local imaging image.

[0149] In one embodiment, the field filtering module 30 is also used to determine the local heating point in the thermal field distribution, determine the discharge position in the electric field distribution, and determine the ranging error between the discharge position and the actual position; determine the model benchmark of the filtering space based on the ranging error; filter according to the model benchmark and the electric field distribution to obtain a covariance matrix; perform gain processing on the covariance matrix to obtain a gain matrix; update the state of the gain matrix to obtain a filtering result; determine the electric field strength, the discharge position and the local heating point to determine the insulation distance based on the filtering result.

[0150] In one embodiment, the fault output module 40 is also used to analyze the filtering result when the insulation distance is less than the safety distance, determine the discharge position and the electric field strength, and determine the abnormal discharge position according to the discharge position and the electric field strength; determine the temperature of the local heating point, and when the temperature of the local heating point is greater than a preset temperature, determine the local heating point as an abnormal heating position; determine the abnormal discharge position and the abnormal heating position as fault points, and visually output the fault points.

[0151] In one embodiment, the fault output module 40 is also used to determine the redundant distances of the insulation distance and the safety distance when the insulation distance is greater than the safety distance; sort the redundant distances, and assign monitoring weights according to the sorting results, set the minimum monitoring weight for the maximum redundant distance among the redundant distances, and set the minimum redundant distance among the redundant distances as the maximum monitoring weight; and perform real-time detection of the insulation distance based on the monitoring weights.

[0152] The large-model-based image detection device for DC professional converter valve equipment provided by the present application adopts the large-model-based image detection method for DC professional converter valve equipment in the above-mentioned embodiment, which can solve the technical problem of low detection efficiency of manual inspection methods in the prior art. Compared with the prior art, the beneficial effects of the large-model-based image detection device for DC professional converter valve equipment provided by the present application are the same as the beneficial effects of the large-model-based image detection method for DC professional converter valve equipment provided by the above-mentioned embodiment, and the other technical features of the large-model-based image detection device for DC professional converter valve equipment are the same as the features disclosed in the above-mentioned embodiment method, which will not be repeated here.

[0153] The present application provides a large-model-based image detection device for a DC professional converter valve device. The large-model-based image detection device for a DC professional converter valve device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the large-model-based image detection method for a DC professional converter valve device in the above-mentioned embodiment one.

[0154] Reference below Figure 6 , which shows a schematic diagram of the structure of a large-model-based DC professional converter valve device image detection device suitable for implementing the embodiment of the present application. The large-model-based DC professional converter valve device image detection device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions: tablet computers), PMPs (Portable Media Players: portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 6 The large-model-based DC professional converter valve equipment image detection device shown is merely an example and should not bring any limitation to the functions and scope of application of the embodiments of the present application.

[0155] like Figure 6As shown, the image detection device for DC professional converter valve equipment based on a large model may include a processing device 1001 (such as a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM: Read Only Memory) 1002 or the program loaded from the storage device 1003 to the random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the image detection device for DC professional converter valve equipment based on a large model are also stored. The processing device 1001, ROM1002 and RAM1004 are connected to each other through a bus 1005. The input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems may be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 1009. The communication device 1009 may allow the large-model-based DC professional converter valve device image detection device to communicate wirelessly or wired with other devices to exchange data. Although the large-model-based DC professional converter valve device image detection device with various systems is shown in the figure, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or have alternatively.

[0156] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0157] The large-model-based image detection device for DC professional converter valve equipment provided by the present application adopts the large-model-based image detection method for DC professional converter valve equipment in the above-mentioned embodiment, which can solve the technical problem of low detection efficiency of manual inspection methods in the prior art. Compared with the prior art, the beneficial effects of the large-model-based image detection device for DC professional converter valve equipment provided by the present application are the same as the beneficial effects of the large-model-based image detection method for DC professional converter valve equipment provided by the above-mentioned embodiment, and the other technical features of the large-model-based image detection device for DC professional converter valve equipment are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0158] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0159] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0160] The present application provides a computer-readable storage medium having computer-readable program instructions (i.e., computer programs) stored thereon, and the computer-readable program instructions are used to execute the large model-based image detection method for DC professional converter valve equipment in the above-mentioned embodiment.

[0161] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read Only Memory), an erasable programmable read-only memory (EPROM: Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.

[0162] The above-mentioned computer-readable storage medium may be included in the image detection device for DC professional converter valve equipment based on a large model; or it may exist independently without being assembled into the image detection device for DC professional converter valve equipment based on a large model.

[0163] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the large-model-based DC professional converter valve device image detection device, the large-model-based DC professional converter valve device image detection device:

[0164] Acquire an imaging image of a DC professional converter valve device, perform image segmentation on the imaging image, and obtain a local imaging image of the DC professional converter valve device;

[0165] Extracting features from the local imaging image to obtain electric field distribution and thermal field distribution;

[0166] filtering the electric field distribution and the thermal field distribution respectively to obtain filtering results, and determining the insulation distance according to the filtering results;

[0167] When the insulation distance is less than the safety distance, fault identification is performed on the filtering result, a fault point is determined, and the fault point is output.

[0168] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0169] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0170] The modules involved in the embodiments of the present application may be implemented by software or hardware, wherein the name of the module does not limit the unit itself in some cases.

[0171] The readable storage medium provided in the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned large-model-based DC professional converter valve equipment image detection method, which can solve the technical problem of low detection efficiency of manual inspection methods in the prior art. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in the present application are the same as the beneficial effects of the large-model-based DC professional converter valve equipment image detection method provided in the above-mentioned embodiment, and will not be repeated here.

[0172] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the above-mentioned large-model-based DC professional converter valve equipment image detection method.

[0173] The computer program product provided by the present application can solve the technical problem of low detection efficiency of manual inspection methods in the prior art. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as the beneficial effects of the large model-based DC professional converter valve equipment image detection method provided in the above embodiment, which will not be repeated here.

[0174] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A method for image detection of DC professional converter valve equipment based on a large model, characterized in that: The large model-based DC professional converter valve equipment image detection method includes: Acquire an imaging image of a DC professional converter valve device, perform image segmentation on the imaging image, and obtain a local imaging image of the DC professional converter valve device; Extracting features from the local imaging image to obtain electric field distribution and thermal field distribution; filtering the electric field distribution and the thermal field distribution respectively to obtain filtering results, and determining the insulation distance according to the filtering results; When the insulation distance is less than the safety distance, performing fault identification on the filtering result, determining the fault point, and outputting the fault point; The steps of filtering the electric field distribution and the thermal field distribution respectively to obtain filtering results, and determining the insulation distance according to the filtering results include: Determine a local heating point in the thermal field distribution, determine a discharge position in the electric field distribution, and determine a distance measurement error between the discharge position and a true position; Determining a model reference of a filtering space based on the ranging error; Filtering is performed according to the model benchmark and the electric field distribution to obtain a covariance matrix; Performing gain processing on the covariance matrix to obtain a gain matrix; The gain matrix is ​​updated to obtain a filtering result; Determine the electric field strength according to the filtering result, and determine the insulation distance according to the discharge position and the local heating point; When the insulation distance is less than the safety distance, the steps of performing fault identification on the filtering result, determining the fault point, and outputting the fault point include: When the insulation distance is less than the safety distance, analyzing the filtering result, determining the discharge position and the electric field strength, and determining the abnormal discharge position according to the discharge position and the electric field strength; Determine the temperature of a local heating point, and when the temperature of the local heating point is greater than a preset temperature, determine the local heating point as an abnormal heating position; The abnormal discharge position and the abnormal heating position are determined as fault points, and the fault points are output visually.

2. The method according to claim 1, characterized in that The step of acquiring an imaging image of the DC professional converter valve device, performing image segmentation on the imaging image, and obtaining a local imaging image of the DC professional converter valve device comprises: Acquire an imaging image of a DC professional converter valve device, input the imaging image into a convolutional layer, and obtain an initial feature map based on the imaging image; Performing a deconvolution upsampling process on the initial feature map to obtain a target feature map having the same size as the initial feature map; The target feature map is segmented to obtain a local imaging image of the DC professional converter valve equipment.

3. The method according to claim 1, characterized in that The step of extracting features from the local imaging image to obtain electric field distribution and thermal field distribution comprises: Preprocessing the local imaging image to obtain a preprocessed image, and randomly generating an initial image set from the preprocessed image; Performing linear interpolation on the initial image set to obtain a real number vector of the initial image set; Determining the fitness in the initial image set according to the feature weight and the real number vector; Determine a selection probability according to the fitness, and perform mutation processing on the real number vector based on the selection probability and the fitness to obtain electric field characteristics and thermal field characteristics of a local imaging image; Obtaining electric field distribution according to the electric field characteristics of the local imaging image; The thermal field distribution is obtained according to the thermal field characteristics of the local imaging image.

4. The method according to claim 3, characterized in that The step of determining the selection probability according to the fitness, performing mutation processing on the real number vector based on the selection probability and the fitness, and obtaining the electric field characteristics and thermal field characteristics of the local imaging image comprises: Determining an imaging type of the initial image set, and dividing the initial image set into an infrared image set and an ultraviolet image set according to the imaging type; Determining the fitness of the infrared image set and the ultraviolet image set, wherein the fitness includes individual fitness and average fitness; Determine a selection probability according to the individual fitness and the average fitness; Determine the real number code of the local image in the infrared image set or the ultraviolet image set, and cross-process the fitness with the real number vector to obtain a cross-image set; A critical value in the individual fitness is determined, and a variation process is performed based on the cross image set and the critical value to obtain the electric field characteristics and thermal field characteristics of the local imaging image.

5. The method according to any one of claims 1 to 4, characterized in that After the step of filtering the electric field distribution and the thermal field distribution respectively to obtain filtering results, and determining the insulation distance according to the filtering results, the step further includes: When the insulation distance is greater than the safety distance, determining a redundant distance between the insulation distance and the safety distance; Sorting the redundant distances, and assigning monitoring weights according to the sorting results, setting the minimum monitoring weight for the maximum redundant distance among the redundant distances, and setting the maximum monitoring weight for the minimum redundant distance among the redundant distances; The insulation distance is detected in real time based on the monitoring weight.

6. A large-scale model-based image detection device for DC professional converter valve equipment, characterized in that: The device comprises: An image segmentation module is used to obtain an imaging image of a DC professional converter valve device, and perform image segmentation on the imaging image to obtain a local imaging image of the DC professional converter valve device; A feature extraction module is used to extract features from the local imaging image to obtain electric field distribution and thermal field distribution; A field filtering module, used to filter the electric field distribution and the thermal field distribution respectively to obtain filtering results, and determine the insulation distance according to the filtering results; A fault output module, used for performing fault identification on the filtering result, determining a fault point, and outputting the fault point when the insulation distance is less than the safety distance; The filtering of the electric field distribution and the thermal field distribution to obtain filtering results respectively and determining the insulation distance according to the filtering results comprises: Determine a local heating point in the thermal field distribution, determine a discharge position in the electric field distribution, and determine a distance measurement error between the discharge position and a true position; Determining a model reference of a filtering space based on the ranging error; Filtering is performed according to the model benchmark and the electric field distribution to obtain a covariance matrix; Performing gain processing on the covariance matrix to obtain a gain matrix; The gain matrix is ​​updated to obtain a filtering result; Determine the electric field strength according to the filtering result, and determine the insulation distance according to the discharge position and the local heating point; When the insulation distance is less than the safety distance, performing fault identification on the filtering result, determining the fault point, and outputting the fault point includes: When the insulation distance is less than the safety distance, analyzing the filtering result, determining the discharge position and the electric field strength, and determining the abnormal discharge position according to the discharge position and the electric field strength; Determine the temperature of a local heating point, and when the temperature of the local heating point is greater than a preset temperature, determine the local heating point as an abnormal heating position; The abnormal discharge position and the abnormal heating position are determined as fault points, and the fault points are output visually.

7. A large-model-based image detection device for DC professional converter valve equipment, characterized in that: The device comprises: a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the large model-based image detection method for DC professional converter valve equipment as described in any one of claims 1 to 5.

8. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the large-model-based DC professional converter valve equipment image detection method as described in any one of claims 1 to 5 are implemented.

Images