Big data-based power transmission line typical defect image big data analysis system and method

Abstract

The application discloses a power transmission line typical defect image big data analysis system and method based on big data, belongs to the technical field of power transmission line image analysis, and comprises an image feature extraction unit, a multiscale anomaly analysis unit, a defect image parameter analysis processing unit, an image feature storage and retrieval interaction unit, a defect diagnosis unit and a monitoring feedback regulation and analysis unit. The image feature extraction unit uses an optimized power transmission line image SIFT algorithm to perform feature extraction and matching operation on the input power transmission line image. The multiscale anomaly analysis unit performs multiscale anomaly analysis processing on the image feature data transmitted based on a multiscale power transmission line image anomaly analysis model. The application realizes efficient and accurate analysis of power transmission line defect images.

Description

Technical Field

[0001] This invention relates to the field of transmission line image analysis technology, specifically to a big data analysis system and method for typical defect images of transmission lines based on big data. Background Technology

[0002] With the continuous expansion of power systems, transmission lines, as critical infrastructure for power transmission, are of paramount importance for their safe and stable operation. Transmission lines are constantly exposed to complex natural environments and operating conditions, making them highly susceptible to various defects. Once a fault occurs, it will severely impact the reliability of power supply. Traditional manual inspection methods are not only inefficient but also suffer from limited inspection range and insufficient detection accuracy, making them unsuitable for the operation and maintenance needs of large-scale transmission lines. Against this backdrop, a transmission line defect image analysis system based on big data technology has emerged, aiming to achieve efficient and accurate detection of transmission line defects using advanced image analysis and processing technologies.

[0003] Existing technologies have several shortcomings in the analysis of transmission line defect images. Firstly, in the image feature extraction and analysis stage, traditional algorithms struggle to effectively handle the complex and variable characteristics of transmission line images, failing to accurately extract defect features. They are prone to false positives or false negatives when faced with changing lighting conditions and background interference. Secondly, at the multi-scale analysis and diagnostic decision-making level, existing systems lack the capability for comprehensive multi-scale analysis of transmission line images. This prevents the full extraction of defect information at different scales, leading to inaccurate assessments of defect severity. Consequently, the scientific validity and effectiveness of decision-making schemes are affected, making it difficult to ensure the safe and stable operation of transmission lines. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by this invention is: firstly, how to address the shortcomings of existing methods in the image feature extraction and analysis stage. Traditional algorithms struggle to effectively handle the complex and variable characteristics of transmission line images, failing to accurately extract defect features and prone to false positives or false negatives when faced with changes in lighting conditions and background interference. Secondly, in the multi-scale analysis and diagnostic decision-making stage, existing systems lack the ability to perform comprehensive multi-scale analysis of transmission line images, failing to fully extract defect information at different scales. This results in inaccurate assessments of defect severity, affecting the scientific validity and effectiveness of decision-making schemes and hindering the safe and stable operation of transmission lines.

[0006] To address the aforementioned technical problems, this invention provides the following technical solution: a big data analysis system for typical defect images of transmission lines based on big data, comprising an image feature extraction unit, a multi-scale anomaly analysis unit, a defect image parameter parsing and processing unit, an image feature storage and retrieval interaction unit, a defect diagnosis unit, and a monitoring feedback control analysis unit; the image feature extraction unit is connected to the multi-scale anomaly analysis unit via a data transmission channel, and uses an optimized transmission line image SIFT algorithm to perform feature extraction and matching operations on the input transmission line image, transmitting the extracted and matched image feature data to the multi-scale anomaly analysis unit; the multi-scale anomaly analysis unit is connected to the defect image parameter parsing and processing unit via a data interaction link, and performs multi-scale anomaly analysis processing on the transmitted image feature data based on a multi-scale transmission line image anomaly analysis model, and transmits the anomaly analysis result data; the defect image parameter parsing and processing unit and the image... The feature storage and retrieval interaction unit is connected via a bidirectional data path. Based on the defect parameter association model and the defect type identification parameter mapping model, it performs defect parameter parsing and extraction on anomaly analysis results and image features, completing the storage and retrieval of defect image parameter data. The image feature storage and retrieval interaction unit is connected to the defect diagnosis unit through a dedicated data interface. Based on the feature storage index association model, it realizes structured storage and efficient retrieval interaction of image features and defect parameter data, providing data support for defect diagnosis. The defect diagnosis unit is connected to the monitoring feedback control analysis unit through a monitoring feedback data link. Based on defect diagnosis rule matching and scheme optimization, it performs diagnostic analysis on the parsed defect image parameters and generates decision schemes, enabling the monitoring feedback control analysis unit to obtain defect diagnosis result data. Based on the monitoring feedback data of system operation status and decision execution results, the monitoring feedback control analysis unit performs feedback control analysis on each unit of the system.

[0007] As a preferred embodiment of the big data analysis system for typical defect images of transmission lines based on big data as described in this invention, the image feature extraction unit includes an optimized transmission line image SIFT algorithm to perform feature extraction and matching operations on the input transmission line image; the feature matching model constructed using the optimized transmission line image SIFT algorithm is expressed as follows:

[0008] ;

[0009] in, Indicates the first Image and the first Feature matching degree of the image For the first The first image 1 eigenvector For the first The mean of the feature vectors of the images. For the first The mean of the feature vectors of the images. For the first The first image 1 eigenvector is the dimension of the feature vector.

[0010] As a preferred embodiment of the big data analysis system for typical defect images of transmission lines based on big data as described in this invention, the image feature extraction unit further includes a feature descriptor enhancement model for constructing an optimized SIFT algorithm for transmission line images, expressed as follows:

[0011] ;

[0012] in, For the enhanced feature descriptor, For the original feature descriptor, To enhance the coefficient, This represents the scale value of the current image.

[0013] This invention employs an optimized SIFT algorithm for transmission line images. In the image feature extraction unit, it enhances the feature extraction capability under complex conditions such as illumination changes and background interference through a feature matching model and a descriptor enhancement model, thereby reducing the false detection and false negative rates.

[0014] As a preferred embodiment of the big data analysis system for typical defect images of transmission lines based on big data as described in this invention, the multi-scale anomaly analysis unit includes: performing multi-scale anomaly analysis processing on the transmitted image feature data based on a multi-scale transmission line image anomaly analysis model; the scale space construction formula of the multi-scale transmission line image anomaly analysis model is expressed as:

[0015] ;

[0016] in, Represents scale-space images It is a two-dimensional Gaussian function. The original input is a transmission line image; the two-dimensional Gaussian function is expressed as:

[0017] ;

[0018] in, As a scale factor, and This represents the coordinates of the image on a two-dimensional plane.

[0019] As a preferred embodiment of the big data analysis system for typical defect images of transmission lines based on big data as described in this invention, the multi-scale anomaly analysis unit includes an anomaly threshold dynamic adjustment model, represented as follows:

[0020] ;

[0021] ;

[0022] in, The updated anomaly threshold, The original abnormal threshold, To adjust the coefficient, The average anomaly level of the image at the current multi-scale; the multi-scale feature fusion model constructed in the multi-scale anomaly analysis unit is represented as:

[0023] ;

[0024] ;

[0025] in, The fused feature vector For the total number of scales, For the first Weights of each scale feature For the first Image feature vectors at various scales; a multi-scale anomaly trend prediction model is constructed in the multi-scale anomaly analysis unit, represented as follows:

[0026] ;

[0027] ;

[0028] in, To predict the degree of anomaly, The number of historical outliers. For the first The weight of each historical outlier data point For the first A historical anomaly degree value; a multi-scale anomaly region localization model is constructed in the multi-scale anomaly analysis unit, represented as:

[0029] ;

[0030] in, A collection of abnormal regions. Image pixel coordinates, Image pixel coordinates The degree of abnormality, This is the threshold for local anomalies.

[0031] The multi-scale anomaly analysis unit of this invention is based on a multi-scale transmission line image anomaly analysis model. It constructs a scale space and integrates multi-scale features to mine defect information from multiple dimensions, thus solving the problem of insufficient multi-scale analysis capability of existing systems.

[0032] As a preferred embodiment of the big data analysis system for typical defect images of transmission lines based on big data according to the present invention, the defect image parameter parsing and processing unit includes a defect size parameter parsing module, a defect shape parameter parsing module, a defect texture parameter parsing module, and a defect location parameter parsing module; the defect size parameter parsing module is used to analyze and extract the length and width-related parameters of the defect in the defect image; the defect shape parameter parsing module extracts and quantifies the shape feature parameters of the defect in the defect image through geometric feature analysis; the defect texture parameter parsing module uses a texture analysis algorithm to analyze the texture feature parameters of the defect in the defect image; the defect location parameter parsing module determines the specific coordinate location parameters of the defect in the transmission line image; based on the optimized transmission line image SIFT algorithm and the multi-scale transmission line image anomaly analysis model, a defect parameter association model is constructed as follows:

[0033] ;

[0034] ;

[0035] in, A comprehensive parameter index representing a defective image. These are the weighting coefficients. The feature matching degree is obtained by the optimized SIFT algorithm for transmission line images. This is the anomaly level value output by the multi-scale transmission line image anomaly analysis model.

[0036] As a preferred embodiment of the big data analysis system for typical defect images of transmission lines based on big data as described in this invention, the defect image parameter parsing and processing unit further includes constructing a defect type identification parameter mapping model, represented as follows:

[0037] ;

[0038] ;

[0039] in, A comprehensive quantitative score representing the type of defect. The number of defect image parameters, For the first The weights of each defective image parameter, For the first Each defect image parameter.

[0040] As a preferred embodiment of the big data analysis system for typical defect images of transmission lines based on big data as described in this invention, the image feature storage and retrieval interaction unit includes a feature data storage management module, a feature data index construction module, a feature data retrieval execution module, and a feature data update and maintenance module. The feature data storage management module is responsible for storing and managing the storage structure of transmission line image feature data and defect image parameter data. The feature data index construction module constructs an efficient data index based on image features and defect parameters for rapid retrieval. The feature data retrieval execution module quickly searches for matching feature data in the stored data according to the retrieval request. The feature data update and maintenance module updates and maintains the stored feature data to ensure data accuracy and timeliness. Based on the optimized transmission line image SIFT algorithm and the multi-scale transmission line image anomaly analysis model, the feature storage index association model is constructed as follows:

[0041] ;

[0042] ;

[0043] in, This indicates the feature storage index value. For index correlation coefficients, The key feature values ​​obtained by the SIFT algorithm for optimized transmission line images These are the key anomaly values ​​output by the multi-scale transmission line image anomaly analysis model.

[0044] As a preferred embodiment of the big data analysis system for typical defect images of transmission lines based on big data as described in this invention, the defect diagnosis unit includes a defect diagnosis rule matching module, a defect severity assessment module, a decision scheme generation module, and a decision scheme optimization module. The defect diagnosis rule matching module matches the parsed defect image parameters with preset diagnosis rules to preliminarily determine the defect type. The defect severity assessment module assesses the severity level of the defect based on the defect image parameters and the diagnosis rules. The decision scheme generation module generates corresponding processing decision schemes based on the defect diagnosis results and severity. The decision scheme optimization module optimizes and adjusts the generated decision schemes to form the final decision.

[0045] This invention provides a big data analysis method for typical defects in transmission lines based on big data images.

[0046] To address the aforementioned technical problems, this invention provides the following technical solution: a big data analysis method for typical defect images of transmission lines based on big data, comprising: using an optimized SIFT algorithm for transmission line images to extract features from input transmission line images and obtain image feature data; transmitting the extracted image feature data to a multi-scale anomaly analysis unit, performing multi-scale anomaly analysis on the image feature data based on a multi-scale transmission line image anomaly analysis model to obtain anomaly analysis results; transmitting the anomaly analysis results to a defect image parameter parsing and processing unit, parsing and extracting various parameters of the defect image to obtain defect image parameter data; transmitting the defect image parameter data to an image feature storage and retrieval interaction unit for storage and indexing; providing the stored data to a defect diagnosis unit, performing defect diagnosis based on diagnostic rules and parameter data, and generating a decision scheme; and a monitoring feedback control analysis unit monitoring the system operation status and decision execution, and controlling each unit of the system based on the monitoring results.

[0047] The beneficial effects of this invention are as follows: This invention proposes a big data analysis system for typical defects in transmission lines based on big data. This system employs an optimized SIFT algorithm for transmission line images. In the image feature extraction unit, feature matching and descriptor enhancement models are used to enhance feature extraction capabilities under complex conditions such as illumination changes and background interference, reducing false positives and false negatives. Simultaneously, the multi-scale anomaly analysis unit, based on a multi-scale transmission line image anomaly analysis model, constructs a scale space and integrates multi-scale features to mine defect information from multiple dimensions, solving the problem of insufficient multi-scale analysis capabilities in existing systems. At the defect diagnosis and decision-making level, the system comprehensively analyzes parameters such as size, shape, and texture of defect images through a defect image parameter parsing and processing unit, and accurately determines the defect type and severity using a defect parameter association model and a type recognition parameter mapping model. Combined with the efficient data storage and retrieval mechanism of the image feature storage and retrieval interaction unit, and the scientific diagnostic rule matching and decision scheme generation optimization process in the defect diagnosis unit, the generated decision schemes are ensured to be more scientific and effective. In addition, the monitoring, feedback, control and analysis unit monitors the system's operating status and decision execution in real time, dynamically adjusts system parameters, ensures the efficiency and safety of transmission line operation and maintenance, and provides reliable technical support for the intelligent operation and maintenance of transmission lines. Attached Figure Description

[0048] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1This is a system unit composition diagram of a big data analysis system for typical defects in transmission lines based on big data, provided as an embodiment of the present invention.

[0050] Figure 2 The flowchart illustrates the system operation of a big data analysis method for typical defect images of power transmission lines based on big data, as provided in one embodiment of the present invention. Detailed Implementation

[0051] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0052] Example 1, referring to Figure 1 This is one embodiment of the present invention, which provides a big data analysis system for typical defects in transmission lines based on big data, including:

[0053] The system includes an image feature extraction unit, a multi-scale anomaly analysis unit, a defect image parameter parsing and processing unit, an image feature storage and retrieval interaction unit, a defect diagnosis unit, and a monitoring feedback control analysis unit.

[0054] The image feature extraction unit is connected to the multi-scale anomaly analysis unit through a data transmission channel. It uses an optimized SIFT algorithm for transmission line images to perform feature extraction and matching operations on the input transmission line images and transmits the extracted and matched image feature data to the multi-scale anomaly analysis unit.

[0055] The multi-scale anomaly analysis unit is connected to the defect image parameter parsing and processing unit via a data interaction link. Based on the multi-scale transmission line image anomaly analysis model, it performs multi-scale anomaly analysis processing on the transmitted image feature data and transmits the anomaly analysis result data.

[0056] The defect image parameter parsing and processing unit and the image feature storage and retrieval interaction unit are connected through a bidirectional data path. Based on the defect parameter association model and the defect type identification parameter mapping model, the defect parameters are parsed and extracted from the anomaly analysis results and image features, and the storage and retrieval of defect image parameter data are completed.

[0057] The image feature storage and retrieval interaction unit is connected to the defect diagnosis unit through a dedicated data interface. Based on the feature storage index association model, it realizes the structured storage and efficient retrieval interaction of image feature and defect parameter data, providing data support for defect diagnosis.

[0058] The defect diagnosis unit and the monitoring feedback control analysis unit are connected through the monitoring feedback data link. Based on defect diagnosis rule matching and scheme optimization, the unit performs diagnostic analysis on the parsed defect image parameters and generates a decision scheme, enabling the monitoring feedback control analysis unit to obtain defect diagnosis result data.

[0059] The monitoring, feedback, control, and analysis unit performs feedback control and analysis on each unit of the system based on the monitoring feedback data of the operating status and decision execution results.

[0060] The image feature extraction unit uses an optimized scale-invariant feature transform (SIFT) algorithm to perform feature extraction and matching operations on the input transmission line image.

[0061] The multi-scale anomaly analysis unit is based on a multi-scale transmission line image anomaly analysis model to perform multi-scale anomaly analysis processing on the transmitted image feature data.

[0062] It should be noted that this invention achieves intelligent recognition, defect diagnosis, and closed-loop control of transmission line images through the coordinated processing of six functional units, and features high image processing accuracy, timely anomaly recognition, intelligent diagnostic decision-making, and strong system operation controllability.

[0063] Example 2, an embodiment of the present invention, provides a big data analysis system for typical defects in transmission lines based on big data, based on the previous embodiment, comprising:

[0064] The image feature extraction unit is connected to the multi-scale anomaly analysis unit through a data transmission channel, and is used to transmit the extracted and matched image feature data to the multi-scale anomaly analysis unit.

[0065] Specifically, the image feature extraction unit employs an optimized SIFT algorithm for transmission line images. Its core lies in targeted improvements to the traditional SIFT algorithm to adapt to the complex and variable characteristics of transmission line images. Regarding technical parameters, by adjusting the scale space construction parameters for keypoint detection and the dimensionality parameters for feature descriptor generation, the algorithm can stably and accurately extract feature points from images under different lighting conditions, weather conditions, and shooting angles. The image feature extraction unit can transform the original transmission line image into feature data containing crucial information, which serves as an important basis for subsequent anomaly analysis and defect diagnosis.

[0066] In its implementation, the image feature extraction unit first receives the input transmission line image, then performs scale-space extremum detection sequentially according to the optimized SIFT algorithm to determine the locations of key points in the image; next, it precisely locates the key points and removes unstable edge points; then, it assigns orientation parameters to each key point to ensure the rotation invariance of the feature description; finally, it generates unique feature descriptors to achieve effective extraction and matching of image features. Through this series of steps, the unit can efficiently extract key features from the transmission line image, providing reliable data support for subsequent units.

[0067] The image feature extraction unit uses an optimized SIFT algorithm for transmission line images to perform feature extraction and matching operations on the input transmission line images.

[0068] Preferably, in the image feature extraction unit, the feature matching model formula constructed by the optimized transmission line image SIFT algorithm is as follows:

[0069] ;

[0070] in, Indicates the first Image and the first Feature matching degree of the image For the first The first image 1 eigenvector For the first The mean of the feature vectors of the images. For the first The mean of the feature vectors of the images. For the first The first image 1 eigenvector is the dimension of the feature vector.

[0071] The optimized feature descriptor enhancement model for the SIFT algorithm in transmission line images is represented as follows:

[0072] ;

[0073] in, For the enhanced feature descriptor, For the original feature descriptor, To enhance the coefficient, This represents the scale value of the current image.

[0074] Specifically, the optimized SIFT algorithm feature matching model in the image feature extraction unit determines the image feature matching degree by calculating the similarity between feature vectors. Its technical parameters include feature vector dimension and mean calculation range, which directly affect the matching accuracy. The image feature extraction unit provides a reliable data foundation for subsequent anomaly analysis by quantifying the feature matching degree. In implementation, the system first extracts image feature vectors, normalizes them, calculates the mean, and then substitutes them into the model to calculate the feature matching degree. Finally, the results are transmitted to the multi-scale anomaly analysis unit to construct Gaussian kernel functions at different scales, enabling multi-scale representation and analysis of transmission line images.

[0075] Preferably, the optimized feature descriptor enhancement model formula for the SIFT algorithm in the image feature extraction unit for transmission line images is as follows:

[0076] ;

[0077] in, For the enhanced feature descriptor, For the original feature descriptor, To enhance the coefficient, This represents the scale value of the current image.

[0078] In the multi-scale anomaly analysis unit, the formula for the multi-scale feature fusion model is:

[0079] ;

[0080] ;

[0081] in, The fused feature vector For the total number of scales, For the first Weights of each scale feature For the first Image feature vectors at various scales.

[0082] Specifically, the system employs a feature descriptor enhancement model and a multi-scale feature fusion model. The enhancement model strengthens the expressive power of feature descriptors by introducing a scale factor, with technical parameters including enhancement coefficients and scale values, thereby improving the stability of features at different scales. The fusion model combines multi-scale feature vectors through weighted aggregation, with parameters involving the weight allocation for each scale. This unit improves the comprehensiveness and accuracy of feature representation. In implementation, the system first performs scale enhancement on the original descriptors, then fuses feature vectors from different scales according to their weights to form a more representative feature representation, providing richer information for defect identification.

[0083] Furthermore, the multi-scale anomaly analysis unit is connected to the defect image parameter parsing and processing unit via a data interaction link to transmit anomaly analysis result data.

[0084] Specifically, the multi-scale anomaly analysis unit is based on a multi-scale transmission line image anomaly analysis model, processing image feature data at multiple different scales. In terms of technical parameter settings, by determining appropriate scale ranges and scale intervals, it ensures coverage of various possible defect scales in transmission line images. The multi-scale anomaly analysis unit reflects the different characteristics of defects at both macroscopic and microscopic levels based on image information at different scales. Through multi-scale analysis, potential anomalies can be identified more accurately, avoiding missed detections or misjudgments caused by single-scale analysis.

[0085] During implementation, the multi-scale anomaly analysis unit receives feature data from the image feature extraction unit and first constructs image representations at multiple different scales based on preset scale parameters. Then, for each scale image, a preset analysis method is used to process and analyze the image features to detect anomalies. By comprehensively comparing and judging the analysis results at multiple scales, the system ultimately determines whether anomalies exist in the image, as well as the specific location and extent of the anomalies, providing accurate anomaly analysis results data for subsequent parameterized analysis of defective images.

[0086] The multi-scale anomaly analysis unit is based on a multi-scale transmission line image anomaly analysis model to perform multi-scale anomaly analysis processing on the transmitted image feature data.

[0087] The preferred formula for constructing the scale space of the multi-scale transmission line image anomaly analysis model is expressed as follows:

[0088] ;

[0089] in, Represents scale-space images It is a two-dimensional Gaussian function. The image is the original input of the power transmission line.

[0090] The two-dimensional Gaussian function is represented as:

[0091] ;

[0092] in, The scaling factor controls the smoothness of the Gaussian kernel; different scaling factors... The values ​​correspond to images at different scales; and This represents the coordinates of the image on a two-dimensional plane.

[0093] Furthermore, the defect image parameter parsing and processing unit and the image feature storage and retrieval interaction unit are connected through a bidirectional data path to store and retrieve defect image parameter data.

[0094] Specifically, the defect image parameter analysis and processing unit is primarily responsible for performing detailed parameter analysis on images identified as having anomalies after multi-scale anomaly analysis. In terms of technical parameters, it covers the quantitative setting of various parameters in the defect image, such as the measurement accuracy standard for defect size parameters, geometric description parameters for shape parameters, and texture feature extraction dimensions for texture parameters. The defect image parameter analysis and processing unit transforms abstract defect image information into specific, quantifiable parameter data. These parameters can intuitively reflect the characteristics and properties of the defect, providing a data foundation for accurately determining the defect type and severity.

[0095] In terms of implementation, after receiving the anomaly analysis results data transmitted by the multi-scale anomaly analysis unit, the defect image parameter parsing and processing unit analyzes various parameters of the defect for the image region containing anomalies using corresponding algorithms and techniques. For defect size, its actual size is determined through pixel measurement and proportional conversion; for shape, a geometric shape recognition algorithm is used to extract contour and geometric feature parameters; for texture, a texture analysis algorithm is used to extract texture patterns and statistical feature parameters. Through precise analysis of these parameters, parameterized processing of the defect image is achieved, providing detailed and accurate defect information for subsequent big data storage and diagnostic decision-making.

[0096] Preferably, the defect image parameter analysis and processing unit includes a defect size parameter analysis module, a defect shape parameter analysis module, a defect texture parameter analysis module, and a defect location parameter analysis module.

[0097] The defect size parameter analysis module is used to analyze and extract the length and width parameters of the defect in the defect image; the defect shape parameter analysis module extracts and quantifies the shape feature parameters of the defect in the defect image through geometric feature analysis; the defect texture parameter analysis module uses texture analysis algorithms to analyze the texture feature parameters of the defect in the defect image; and the defect location parameter analysis module determines the specific coordinate location parameters of the defect in the transmission line image.

[0098] Specifically, the defect image parameter analysis and processing unit consists of a defect size parameter analysis module, a defect shape parameter analysis module, a defect texture parameter analysis module, and a defect location parameter analysis module.

[0099] The defect size parameter analysis module extracts the geometric size parameters of the defect through pixel analysis technology; the defect shape parameter analysis module uses contour detection and geometric feature calculation methods to quantify the shape characteristics of the defect; the defect texture parameter analysis module uses algorithms such as gray-level co-occurrence matrix to analyze the texture characteristics of the defect; the contour detection and geometric feature calculation method first uses image processing technology to find the edge contour of the defect from the defect image and clarify the boundary shape of the defect; then, based on the contour, it calculates geometric feature values ​​such as perimeter, area, roundness (the ratio of the square of the perimeter to the area reflects the degree of near-circleness), and rectangularity (the ratio of the defect area to the area of ​​the smallest bounding rectangle) to quantify the shape characteristics of the defect, so that the shape of the defect can be presented with specific parameters, which is convenient for analysis and judgment.

[0100] Preferably, the defect image parameter parsing and processing unit, based on the optimized transmission line image SIFT algorithm and the multi-scale transmission line image anomaly analysis model, constructs the defect parameter correlation model formula as follows:

[0101] ;

[0102] ;

[0103] in, The comprehensive parameter index representing the defective image is obtained by combining the feature matching degree obtained from the optimized SIFT algorithm for transmission line images with the anomaly degree value output by the multi-scale transmission line image anomaly analysis model, and then multiplying the feature matching degree and the anomaly degree value by their respective weight coefficients (which sum to 1) according to two set weight coefficients, and then summing them up. These are the weighting coefficients. The feature matching degree is obtained by the optimized SIFT algorithm for transmission line images. This is the anomaly level value output by the multi-scale transmission line image anomaly analysis model. The anomaly level value is directly output by the multi-scale transmission line image anomaly analysis model after image analysis.

[0104] The multi-scale anomaly analysis unit also includes a dynamic adjustment model formula for the anomaly threshold:

[0105] ;

[0106] ;

[0107] in, The updated anomaly threshold, The original abnormal threshold, To adjust the coefficient, The average anomaly level of the image at the current multi-scale is obtained by statistically analyzing the anomaly level of each part of the image at the current scale and then calculating the average of these anomaly levels during multi-scale analysis.

[0108] Specifically, a defect parameter correlation model and an anomaly threshold dynamic adjustment model are constructed. The former generates a comprehensive parameter index by fusing feature matching degree and anomaly degree values. Its technical parameters involve setting weighting coefficients, which need to be dynamically adjusted according to different defect types. The latter adjusts the threshold based on historical anomaly data, with parameters including adjustment coefficients and average anomaly degree. The significance of these two models lies in achieving accurate quantification of defect parameters and adaptive adjustment of thresholds. In implementation, the system first obtains the feature matching degree and anomaly degree values, calculates the comprehensive parameter index through weighted average, and dynamically updates the threshold based on historical data to improve the system's adaptability to different environments.

[0109] Preferably, the defect image parameter parsing and processing unit has the following defect type identification parameter mapping model formula:

[0110] ;

[0111] ;

[0112] in, A comprehensive quantitative score representing the type of defect. The number of defect image parameters, For the first The weights of each defective image parameter, For the first Each defect image parameter.

[0113] Continuous values The comprehensive quantitative score for defect types is obtained by discretizing the continuous value through a preset classification threshold to correspond to the specific defect category. These are the image parameters of each defect. The weights (and the sum of the weights is 1) are used in the formula, which is actually a weighted fusion of multi-dimensional defect features. The resulting continuous value is used to characterize the confidence level of belonging to a certain type of defect. Then, a classification threshold is used to achieve discrete category label output. The classification threshold is set based on the statistical characteristics of a large amount of historical transmission line defect image data, the parameter distribution characteristics of different defect types, and operation and maintenance experience. Then, combined with the hazard level and handling priority of different defect types in actual operation and maintenance, the judgment boundary value of each type of defect is determined, thus forming the initial classification threshold. The classification threshold can be continuously adjusted through sample verification and iterative optimization, so that the continuous quantitative score can be accurately mapped to discrete defect categories, achieving reliable identification and scientific classification of defect types.

[0114] In the multi-scale anomaly analysis unit, the formula for the multi-scale anomaly trend prediction model is:

[0115] ;

[0116] ;

[0117] in, To predict the degree of anomaly, The number of historical outliers. For the first The weight of each historical outlier data point For the first A historical anomaly value.

[0118] Specifically, the system employs a defect type identification parameter mapping model and a multi-scale anomaly trend prediction model. The mapping model identifies defect types by weighting multiple defect parameters, with the technical parameters representing the weights of each parameter, which need to be determined through training with a large number of samples. The prediction model forecasts future trends based on historical anomaly data, with parameters including the amount of historical data and time weights. Its function is to automatically identify defect types and predict their development trends. In implementation, the system first extracts defect parameters and calculates their weights, substitutes them into the mapping model to determine the type, and then combines historical data to predict anomaly development trends, providing forward-looking support for operational and maintenance decisions.

[0119] Furthermore, the image feature storage and retrieval interaction unit is connected to the defect diagnosis unit through a dedicated data interface to provide data support for defect diagnosis;

[0120] Specifically, the image feature storage and retrieval interaction unit acts as a data management and interaction hub in the system. Its technical parameter settings involve data storage capacity planning, storage structure design parameters, and data retrieval index construction parameters and retrieval efficiency optimization parameters. This unit can efficiently store and quickly retrieve the large amounts of transmission line image feature data and defect image parameter data generated during system operation, ensuring data integrity and availability, and providing timely and accurate data support for defect diagnosis and decision generation.

[0121] During implementation, the image feature storage and retrieval interaction unit first classifies and organizes the defect image parameter data from the defect image parameter parsing and processing unit, as well as other relevant image feature data, and stores them in the corresponding storage media according to the preset data storage structure. Simultaneously, it constructs an efficient data index based on the data's characteristics and attributes to improve the speed and accuracy of data retrieval. When the defect diagnosis unit needs to access data, the image feature storage and retrieval interaction unit quickly locates and extracts the required data through the index according to the retrieval request, and transmits it to the corresponding unit, achieving efficient data interaction and ensuring the overall operational efficiency of the system and the scientific nature of decision-making.

[0122] The image feature storage and retrieval interaction unit includes a feature data storage and management module, a feature data index construction module, a feature data retrieval execution module, and a feature data update and maintenance module.

[0123] The feature data storage management module is responsible for storing and managing the storage structure of transmission line image feature data and defect image parameter data; the feature data index construction module builds an efficient data index based on image features and defect parameters for fast retrieval; the feature data retrieval execution module quickly finds matching feature data in the stored data according to the retrieval request; and the feature data update and maintenance module updates and maintains the stored feature data to ensure the accuracy and timeliness of the data.

[0124] Specifically, the feature data storage and management module adopts a distributed storage architecture to achieve efficient storage of massive image and parameter data; the feature data indexing construction module builds a multi-level index structure based on feature hashing and anomaly clustering techniques; the feature data retrieval execution module quickly locates target data through parallel query algorithms; and the feature data update and maintenance module adopts an incremental update strategy to ensure data timeliness. By optimizing the storage structure and retrieval algorithm, the efficiency bottleneck problem of traditional systems when processing large-scale transmission line data has been solved.

[0125] Preferably, the image feature storage and retrieval interaction unit, based on the optimized transmission line image SIFT algorithm and the multi-scale transmission line image anomaly analysis model, constructs the feature storage index association model formula as follows:

[0126] ;

[0127] ;

[0128] in, This indicates the feature storage index value. For index correlation coefficients, The key feature values ​​obtained by the SIFT algorithm for optimized transmission line images These are the key anomaly values ​​output by the multi-scale transmission line image anomaly analysis model. Feature key values ​​are calculated by extracting key information related to features from the transmission line image using an optimized SIFT algorithm. Anomaly key values ​​are calculated by analyzing the image using the multi-scale transmission line image anomaly analysis model, identifying key information related to anomalies, and then calculating them. Both are used together to construct a feature storage index association model. Key information refers to core parameter information that characterizes the local structural features and state change characteristics of the transmission line image. This includes key point locations, scales, orientations, local gradient distributions, and feature descriptors extracted by the optimized SIFT algorithm, reflecting the structural stability characteristics of normal components; and parameters such as brightness abrupt changes, texture damage, edge morphology anomalies, and multi-scale response differences identified by the multi-scale anomaly analysis model, reflecting potential defects or abnormal states.

[0129] In the multi-scale anomaly analysis unit, the formula for the multi-scale anomaly region localization model is:

[0130] ;

[0131] in, A collection of abnormal regions. Image pixel coordinates, Image pixel coordinates The degree of abnormality, The local anomaly threshold is a numerical limit used in transmission line defect image analysis to determine whether abnormal features exist in a local area of ​​the image. This threshold is usually set based on the statistical characteristics of historical transmission line image data.

[0132] Specifically, a feature storage index association model and a multi-scale anomaly region localization model are constructed. The association model generates an index by fusing feature key values ​​and anomaly key values, with the association coefficient as the technical parameter to balance the influence of the two key values. The localization model determines anomaly regions based on local thresholds, with parameters including the local threshold and the degree of pixel anomaly. Its purpose is to achieve efficient data indexing and accurate anomaly localization. In implementation, the system first calculates the feature and anomaly key values, generates an index after weighting, and stores it. Then, it filters out anomaly pixels through local thresholds to form a complete set of anomaly regions.

[0133] The enhancement factor of this application The enhancement effect of feature descriptors at different scales was tested, and the result was iteratively determined in conjunction with the defect identification accuracy; among the weight parameters... Initial values ​​are set by domain experts based on feature matching degree and anomaly degree, and the relative importance of feature key values ​​and anomaly key values, and then fine-tuned through sample verification; Then, by using machine learning algorithms (such as gradient descent and feature importance evaluation) and training and optimizing with a large number of labeled samples, the results of multi-scale feature fusion, defect type identification, and abnormal trend prediction are most consistent with the actual defect features. Based on the stability requirements of threshold adjustment under different scenarios, the influence of historical thresholds and the current average anomaly level is balanced, and the specific values ​​are ultimately determined with the goal of improving the accuracy of defect analysis.

[0134] Furthermore, the defect diagnosis unit and the monitoring feedback control and analysis unit are connected through a monitoring feedback data link, enabling the monitoring feedback control and analysis unit to obtain defect diagnosis result data;

[0135] Specifically, the defect diagnosis unit is a key component of the system for diagnosing and making maintenance decisions regarding transmission line defects. In terms of technical parameters, this includes preset defect diagnosis rule parameters, severity assessment grading standards, and strategy parameters for generating decision-making schemes. The defect diagnosis unit analyzes and judges the defect image parameter data obtained from the previous units, accurately identifying defect types and assessing defect severity. Based on the actual situation, it generates reasonable processing decision schemes, providing scientific guidance for the maintenance of transmission lines and ensuring their safe and stable operation. The preset defect diagnosis rule parameters refer to the basic parameter system set by the system before diagnosing transmission line defect images, used to constrain the defect identification process and judgment logic. These parameters include defect feature threshold parameters, morphological structure matching parameters, texture anomaly recognition parameters, brightness change judgment parameters, and quantification thresholds for multi-scale response differences. These parameters are used to determine the judgment criteria for different types of defects in terms of image structural features, texture features, changes in bright and dark areas, edge integrity, and local abnormal responses, providing a quantitative basis for the automated identification of defect types.

[0136] In terms of implementation, after receiving defect image parameter data from the image feature storage and retrieval interaction unit, the defect diagnosis unit first uses the defect diagnosis rule matching module to compare and match the parsed parameters with preset diagnosis rules one by one to preliminarily determine the type of defect. Then, through the defect severity assessment module, the severity of the defect is quantitatively assessed based on the defect parameters and preset assessment standards to determine its severity level. Finally, the decision scheme generation module combines the defect type, severity, and other relevant factors to generate a corresponding processing decision scheme according to the preset decision strategy. The decision scheme optimization module then adjusts and improves the scheme to form a final executable decision, providing specific action guidance for the maintenance and repair of transmission lines. The preset diagnosis rules are standardized judgment logic based on the common defect morphologies and visual characteristics of transmission lines, including structural feature matching rules, texture anomaly recognition rules, brightness and color abrupt change rules, edge shape destruction rules, and regional statistical feature judgment rules for defects such as rust, cracks, damage, burns, loosening, and dirt. The diagnosis rules achieve rapid identification of defect types by comparing the input defect image parameters with the corresponding feature templates or threshold ranges. When the image parameters meet a certain set of diagnostic conditions, the category of the defect can be determined, thus forming a structured defect diagnosis result, which provides a clear basis for subsequent severity assessment and decision-making.

[0137] The pre-defined decision-making strategy refers to the strategy model followed by the system when generating operation and maintenance solutions for transmission lines. Based on factors such as defect type, severity level, equipment importance, operating environment conditions, and historical operation and maintenance experience, it constructs rules for task priority ranking, inspection method selection, power outage maintenance condition determination, emergency response triggering, and long-term governance planning. The decision-making strategy comprehensively calculates the diagnostic results with the above multi-factor model to determine whether immediate action, scheduled maintenance, or periodic monitoring is required, and generates corresponding action suggestions or processing procedures.

[0138] Preferably, the defect diagnosis unit includes a defect diagnosis rule matching module, a defect severity assessment module, a decision scheme generation module, and a decision scheme optimization module.

[0139] The defect diagnosis rule matching module matches the parsed defect image parameters with preset diagnosis rules to preliminarily determine the defect type; the defect severity assessment module assesses the severity level of the defect based on the defect image parameters and diagnosis rules; the decision scheme generation module generates corresponding processing decision schemes based on the defect diagnosis results and severity; and the decision scheme optimization module optimizes and adjusts the generated decision schemes to form the final decision.

[0140] Specifically, the defect diagnosis rule matching module builds a rule base based on expert knowledge and initially determines the defect type through pattern matching; the defect severity assessment module uses the analytic hierarchy process (AHP) to comprehensively evaluate the defect level using multiple parameters; the decision solution generation module generates corresponding processing solutions based on a predefined decision tree model; and the decision solution optimization module optimizes the parameters of the initial solution using a simulated annealing algorithm. Through standardized diagnostic processes and optimization algorithms, the entire process from defect identification to decision generation is automated, improving operational efficiency and scientific rigor.

[0141] Furthermore, the monitoring, feedback, control, and analysis unit is primarily used to ensure the stable operation and continuous optimization of the system. In terms of technical parameters, it involves indicators for monitoring the system's operational status, such as the operating efficiency of each unit, data processing accuracy, and system resource usage; as well as threshold parameters and control strategies for feedback control. The monitoring, feedback, control, and analysis unit monitors the system's operational status in real time, promptly identifies potential problems and performance bottlenecks, and dynamically adjusts each unit of the system through a feedback control mechanism to ensure the system maintains a consistently efficient and stable operating state, while continuously optimizing system performance to adapt to the complex and ever-changing needs of transmission line operation and maintenance.

[0142] During implementation, the monitoring, feedback, control, and analysis unit monitors the operational data of each unit in real time and compares it with preset monitoring indicators to determine whether the system is operating normally. If any operational anomalies or performance degradation are detected, such as a unit processing too slowly or an increased data processing error rate, the unit adjusts parameters, schedules tasks, or reallocates resources for the relevant unit based on preset feedback control strategies and threshold parameters. Furthermore, the monitoring, feedback, control, and analysis unit collects various feedback information during system operation, such as decision execution effect data, to optimize and improve the system's algorithm model and processing flow. This enables continuous optimization and adaptive adjustment of the system, ensuring long-term stable and efficient service for transmission line defect detection and maintenance.

[0143] To address the shortcomings of existing systems in multi-scale analysis capabilities and the lack of scientific rigor in diagnostic decision-making, the system's multi-scale anomaly analysis unit, based on a multi-scale transmission line image anomaly analysis model, constructs a scale space and integrates multi-scale features to deeply mine defect information from multiple dimensions, enabling the system to comprehensively perceive defect characteristics at different scales. Simultaneously, the defect image parameter analysis and processing unit utilizes a defect parameter association model and a type recognition parameter mapping model to systematically analyze parameters such as size, shape, and texture of defect images, accurately determining the defect type and severity. The image feature storage and retrieval interaction unit establishes an efficient data storage and retrieval mechanism, providing accurate data support for defect diagnosis. The defect diagnosis unit combines diagnostic rules and parameter data to scientifically diagnose defects and generate optimized decision-making schemes, ensuring the scientific rigor and effectiveness of operation and maintenance decisions.

[0144] Furthermore, the monitoring, feedback, control, and analysis unit monitors the system's operational status and decision execution in real time, dynamically adjusting each unit based on real-time data to ensure the system remains in a highly efficient and stable operating state. Through orderly collaboration via data transmission channels and interactive links, each unit forms a complete and intelligent transmission line defect analysis system, significantly improving the intelligence and reliability of transmission line operation and maintenance, and providing a solid guarantee for the safe and stable operation of the power system.

[0145] Example 3, referring to Figure 2 This is one embodiment of the present invention, which provides a big data analysis method for typical defect images of transmission lines based on big data, including:

[0146] S1. Using the optimized SIFT algorithm for transmission line images, feature extraction is performed on the input transmission line images to obtain image feature data.

[0147] S2. The extracted image feature data is transmitted to the multi-scale anomaly analysis unit. Based on the multi-scale transmission line image anomaly analysis model, multi-scale anomaly analysis is performed on the image feature data to obtain the anomaly analysis results.

[0148] S3. Transmit the anomaly analysis results to the defect image parameter parsing and processing unit, and analyze and extract various parameters of the defect image to obtain defect image parameter data.

[0149] S4. Transmit the defective image parameter data to the image feature storage and retrieval interaction unit for storage and build an index.

[0150] S5. Provide the stored data to the defect diagnosis unit, perform defect diagnosis based on the diagnosis rules and parameter data, and generate a decision plan.

[0151] S6, the monitoring, feedback, control and analysis unit monitors the system's operating status and decision execution, and controls each unit of the system based on the monitoring results.

[0152] In the image feature extraction unit, this invention employs an optimized SIFT algorithm for transmission line images. By leveraging feature matching and descriptor enhancement models, it enhances the ability to extract image features under complex lighting and interference backgrounds. This enables precise capture of defect details, significantly reduces the probability of false detection and missed detection, and ensures accurate extraction of defect features.

[0153] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A big data analysis system for typical defects in transmission lines based on big data, characterized in that: include, Image feature extraction unit, multi-scale anomaly analysis unit, defect image parameter parsing and processing unit, image feature storage and retrieval interaction unit, defect diagnosis unit, monitoring feedback control analysis unit; The image feature extraction unit is connected to the multi-scale anomaly analysis unit through a data transmission channel. It uses an optimized SIFT algorithm for transmission line images to perform feature extraction and matching operations on the input transmission line images and transmits the extracted and matched image feature data to the multi-scale anomaly analysis unit. The multi-scale anomaly analysis unit is connected to the defect image parameter parsing and processing unit via a data interaction link. Based on the multi-scale transmission line image anomaly analysis model, it performs multi-scale anomaly analysis processing on the transmitted image feature data and transmits the anomaly analysis result data. The defect image parameter parsing and processing unit and the image feature storage and retrieval interaction unit are connected through a bidirectional data path. Based on the defect parameter association model and the defect type identification parameter mapping model, the defect parameter parsing and extraction are performed on the anomaly analysis results and image features to complete the storage and retrieval of defect image parameter data. The image feature storage and retrieval interaction unit is connected to the defect diagnosis unit through a dedicated data interface. Based on the feature storage index association model, it realizes the structured storage and efficient retrieval interaction of image feature and defect parameter data, providing data support for defect diagnosis. The defect diagnosis unit and the monitoring feedback control analysis unit are connected through the monitoring feedback data link. Based on defect diagnosis rule matching and scheme optimization, the unit performs diagnostic analysis on the parsed defect image parameters and generates a decision scheme, enabling the monitoring feedback control analysis unit to obtain defect diagnosis result data. The monitoring, feedback, control, and analysis unit performs feedback control and analysis on each unit of the system based on the monitoring feedback data of the operating status and decision execution results. The image feature extraction unit, include, The image feature extraction unit uses an optimized SIFT algorithm for transmission line images to perform feature extraction and matching operations on the input transmission line images; The feature matching model constructed using the optimized SIFT algorithm for transmission line images is represented as follows: ； in, Indicates the first Image and the first Feature matching degree of the image For the first The first image 1 eigenvector For the first The mean of the feature vectors of the images, For the first The mean of the feature vectors of the images, For the first The first image 1 eigenvector is the dimension of the feature vector.

2. The big data analysis system for typical defects in transmission lines based on big data as described in claim 1, characterized in that: The image feature extraction unit further includes constructing an optimized feature descriptor enhancement model for the SIFT algorithm of transmission line images, represented as follows: ； in, For the enhanced feature descriptor, For the original feature descriptor, To enhance the coefficient, This represents the scale value of the current image.

3. The big data analysis system for typical defects in transmission lines based on big data as described in claim 2, characterized in that: The multi-scale anomaly analysis unit includes, Based on the multi-scale transmission line image anomaly analysis model, multi-scale anomaly analysis processing is performed on the transmitted image feature data; The scale space construction formula for the multi-scale transmission line image anomaly analysis model is expressed as follows: ； in, Represents scale-space images It is a two-dimensional Gaussian function. The original input image of the power transmission line; The two-dimensional Gaussian function is expressed as follows: ； in, As a scale factor, and This represents the coordinates of the image on a two-dimensional plane.

4. The big data analysis system for typical defects in transmission lines based on big data as described in claim 3, characterized in that: The multi-scale anomaly analysis unit includes an anomaly threshold dynamic adjustment model, which is represented as follows: ； ； in, The updated anomaly threshold. The original abnormal threshold, To adjust the coefficient, This represents the average anomaly level of the image across multiple scales. The multi-scale feature fusion model constructed in the multi-scale anomaly analysis unit is represented as follows: ； ； in, The fused feature vector For the total number of scales, For the first Weights of each scale feature For the first Image feature vectors at various scales; The multi-scale anomaly trend prediction model constructed in the multi-scale anomaly analysis unit is represented as follows: ； ； in, To predict the degree of anomaly, The number of historical outliers. For the first The weight of each historical outlier data point For the first A historical anomaly value; The multi-scale anomaly region localization model constructed in the multi-scale anomaly analysis unit is represented as follows: ； in, A collection of abnormal regions. Image pixel coordinates, Image pixel coordinates The degree of abnormality, This is the threshold for local anomalies.

5. The big data image analysis system for typical defects in transmission lines based on big data as described in claim 4, characterized in that: The defect image parameter parsing and processing unit includes a defect size parameter parsing module, a defect shape parameter parsing module, a defect texture parameter parsing module, and a defect location parameter parsing module; The defect size parameter parsing module is used to parse and extract the length and width-related parameters of the defect in the defect image; The defect shape parameter analysis module extracts and quantifies the shape feature parameters of defects in the defect image through geometric feature analysis; The defect texture parameter parsing module uses a texture analysis algorithm to parse the texture feature parameters of defects in the defect image; The defect location parameter parsing module determines the specific coordinate location parameters of the defect in the transmission line image; Based on the optimized SIFT algorithm for transmission line images and the multi-scale transmission line image anomaly analysis model, a defect parameter correlation model is constructed as follows: ； ； in, A comprehensive parameter index representing a defective image. These are the weighting coefficients. The feature matching degree is obtained by the optimized SIFT algorithm for transmission line images. This is the anomaly level value output by the multi-scale transmission line image anomaly analysis model.

6. The big data image analysis system for typical defects in transmission lines based on big data as described in claim 5, characterized in that: The defect image parameter parsing and processing unit further includes constructing a defect type identification parameter mapping model, represented as follows: ； ； in, A comprehensive quantitative score representing the type of defect. The number of defect image parameters, For the first The weights of each defective image parameter, For the first Each defect image parameter.

7. The big data analysis system for typical defects in transmission lines based on big data as described in claim 6, characterized in that: The image feature storage and retrieval interaction unit includes a feature data storage management module, a feature data index construction module, a feature data retrieval execution module, and a feature data update and maintenance module. The feature data storage management module is responsible for storing and managing the storage structure of transmission line image feature data and defect image parameter data; The feature data indexing construction module constructs an efficient data index based on image features and defect parameters for rapid retrieval; The feature data retrieval and execution module quickly searches for matching feature data in the stored data according to the retrieval request; The feature data update and maintenance module updates and maintains the stored feature data to ensure the accuracy and timeliness of the data. Based on the optimized SIFT algorithm for transmission line images and the multi-scale transmission line image anomaly analysis model, a feature storage index association model is constructed as follows: ； ； in, This indicates the feature storage index value. For index correlation coefficients, The key feature values ​​obtained by the SIFT algorithm for optimized transmission line images These are the key anomaly values ​​output by the multi-scale transmission line image anomaly analysis model.

8. The big data analysis system for typical defects in transmission lines based on big data as described in claim 7, characterized in that: The defect diagnosis unit includes a defect diagnosis rule matching module, a defect severity assessment module, a decision scheme generation module, and a decision scheme optimization module. The defect diagnosis rule matching module matches the parsed defect image parameters with preset diagnosis rules to preliminarily determine the defect type. The defect severity assessment module evaluates the severity level of the defect based on the defect image parameters and diagnostic rules; The decision-making module generates corresponding processing decision-making schemes based on the defect diagnosis results and severity. The decision optimization module optimizes and adjusts the generated decision schemes to form the final decision.

9. A method for big data analysis of typical defect images of transmission lines based on big data, using the big data analysis system for typical defect images of transmission lines based on big data as described in any one of claims 1 to 8, characterized in that, include: An optimized SIFT algorithm for transmission line images is used to extract features from the input transmission line images and obtain image feature data. The extracted image feature data is transmitted to a multi-scale anomaly analysis unit. Based on the multi-scale transmission line image anomaly analysis model, multi-scale anomaly analysis is performed on the image feature data to obtain the anomaly analysis results. The anomaly analysis results are transmitted to the defect image parameter parsing and processing unit, where various parameters of the defect image are parsed and extracted to obtain defect image parameter data. The defective image parameter data is transmitted to the image feature storage and retrieval interaction unit for storage and indexing; The stored data is provided to the defect diagnosis unit, which performs defect diagnosis and generates a decision-making plan based on the diagnosis rules and parameter data. The monitoring, feedback, control, and analysis unit monitors the system's operating status and decision execution, and controls each unit of the system based on the monitoring results.

Images