CN116168019A - Power grid fault detection method and system based on machine vision technology - Google Patents
Power grid fault detection method and system based on machine vision technology Download PDFInfo
- Publication number
- CN116168019A CN116168019A CN202310421381.XA CN202310421381A CN116168019A CN 116168019 A CN116168019 A CN 116168019A CN 202310421381 A CN202310421381 A CN 202310421381A CN 116168019 A CN116168019 A CN 116168019A
- Authority
- CN
- China
- Prior art keywords
- data
- power grid
- image
- grid line
- unit
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 73
- 238000005516 engineering process Methods 0.000 title claims abstract description 28
- 238000000034 method Methods 0.000 claims abstract description 66
- 238000004458 analytical method Methods 0.000 claims abstract description 51
- 238000012937 correction Methods 0.000 claims abstract description 32
- 230000008569 process Effects 0.000 claims abstract description 30
- 230000002159 abnormal effect Effects 0.000 claims abstract description 26
- 238000012545 processing Methods 0.000 claims abstract description 22
- 238000007689 inspection Methods 0.000 claims abstract description 6
- 230000006870 function Effects 0.000 claims description 124
- 238000004364 calculation method Methods 0.000 claims description 48
- 238000001914 filtration Methods 0.000 claims description 41
- 238000011156 evaluation Methods 0.000 claims description 36
- 238000000605 extraction Methods 0.000 claims description 20
- 230000009466 transformation Effects 0.000 claims description 19
- 238000013500 data storage Methods 0.000 claims description 15
- 230000000007 visual effect Effects 0.000 claims description 13
- 238000004821 distillation Methods 0.000 claims description 10
- 238000006243 chemical reaction Methods 0.000 claims description 9
- 125000003275 alpha amino acid group Chemical group 0.000 claims description 8
- 230000000694 effects Effects 0.000 claims description 8
- 150000001875 compounds Chemical class 0.000 claims description 7
- 230000004927 fusion Effects 0.000 claims description 7
- 238000007726 management method Methods 0.000 claims description 7
- 230000009467 reduction Effects 0.000 claims description 7
- 238000000354 decomposition reaction Methods 0.000 claims description 6
- 230000008447 perception Effects 0.000 claims description 6
- 238000007405 data analysis Methods 0.000 claims description 5
- 239000000284 extract Substances 0.000 claims description 5
- 238000009499 grossing Methods 0.000 claims description 4
- 238000003709 image segmentation Methods 0.000 claims description 4
- 230000006872 improvement Effects 0.000 claims description 4
- 238000010801 machine learning Methods 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 4
- 230000001133 acceleration Effects 0.000 claims description 3
- 230000006835 compression Effects 0.000 claims description 3
- 238000007906 compression Methods 0.000 claims description 3
- 230000010354 integration Effects 0.000 claims description 3
- 238000007477 logistic regression Methods 0.000 claims description 3
- 238000003860 storage Methods 0.000 claims description 3
- 238000000547 structure data Methods 0.000 claims description 3
- 241000220479 Acacia Species 0.000 claims description 2
- 235000010643 Leucaena leucocephala Nutrition 0.000 claims description 2
- 239000012634 fragment Substances 0.000 claims description 2
- 230000000737 periodic effect Effects 0.000 abstract description 4
- 238000011161 development Methods 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 6
- 230000008859 change Effects 0.000 description 5
- 238000004891 communication Methods 0.000 description 5
- 230000001276 controlling effect Effects 0.000 description 5
- 238000012544 monitoring process Methods 0.000 description 5
- 230000005540 biological transmission Effects 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 230000010365 information processing Effects 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 238000003708 edge detection Methods 0.000 description 2
- 230000001629 suppression Effects 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 241000196324 Embryophyta Species 0.000 description 1
- 241000607479 Yersinia pestis Species 0.000 description 1
- 239000008186 active pharmaceutical agent Substances 0.000 description 1
- 150000001413 amino acids Chemical class 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000010485 coping Effects 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000005684 electric field Effects 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 230000010363 phase shift Effects 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012731 temporal analysis Methods 0.000 description 1
- 238000000700 time series analysis Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/08—Locating faults in cables, transmission lines, or networks
- G01R31/088—Aspects of digital computing
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/08—Locating faults in cables, transmission lines, or networks
- G01R31/081—Locating faults in cables, transmission lines, or networks according to type of conductors
- G01R31/086—Locating faults in cables, transmission lines, or networks according to type of conductors in power transmission or distribution networks, i.e. with interconnected conductors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration using histogram techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/73—Deblurring; Sharpening
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/77—Retouching; Inpainting; Scratch removal
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
- Y04S10/52—Outage or fault management, e.g. fault detection or location
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Medical Informatics (AREA)
- Marketing (AREA)
- General Health & Medical Sciences (AREA)
- General Business, Economics & Management (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Evolutionary Computation (AREA)
- Strategic Management (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- General Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Databases & Information Systems (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Primary Health Care (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a power grid fault detection method and a system based on a machine vision technology, which relate to the technical field of data identification and processing and solve the problems that power grid fault data information is identified and processed, a scanning acquisition module is carried by an unmanned aerial vehicle to patrol and detect a power grid line, then an image is corrected by a vision correction module, an analysis alarm module analyzes image data and alarms the identified abnormal data, image characteristic data filtered by a data recording module is recorded, backed up and cleared regularly, and a detection control module controls a power grid fault detection process. The invention can replace manual development of periodic comprehensive inspection, alarm the detected fault existing points, and greatly improve the identification and processing capacity of the power grid fault data information.
Description
Technical Field
The invention relates to the technical field of data identification and processing, in particular to a power grid fault detection method and system based on a machine vision technology.
Background
The power system is one of the infrastructures in the modern society, but as the power grid scale is continuously enlarged, the power grid fault rate is continuously increased, and great challenges are brought to the operation and maintenance of the power system. The traditional power grid fault detection mode mainly realizes local on-line monitoring through manual inspection or by utilizing a network system, and the mode can discover power grid faults, but needs to spend a large amount of manpower and material resources to develop periodic inspection, and based on the intelligent development of the power grid, the labor capacity of workers is increased continuously, and the problem that power equipment is damaged and large-area power failure is caused because the inspection is not in place or potential faults are not discovered in time exists.
The machine vision can realize the rapid processing of related data by utilizing an image information technology, so that the power grid fault detection efficiency can be greatly improved, and the construction of an automatic and intelligent detection system by means of the machine vision is important. However, with the generation of the power grid fault data information, how to process the power grid fault data information becomes a problem to be solved. The power grid fault data information comprises various data information such as current abnormal data information, voltage abnormal data information, temperature abnormal information, noise abnormal data information, internal faults of the transformer, mechanical damage and the like. These data information can reflect grid fault information to some extent, but the processing and mining capabilities of these data information are critical to improving grid fault detection.
The patent number CN115471796A discloses a power grid engineering supervision system and a method based on machine vision, wherein the system comprises a supervision route data acquisition module, an abnormal representation judgment module, an early warning mode analysis module, an abnormal representation quantity monitoring module and a deviation adjustment module, and is used for acquiring a planned erection supervision route of the power grid engineering, historical monitoring data and real-time monitoring data on a corresponding route, judging whether abnormal representation exists on the supervision route, analyzing an early warning mode of the corresponding abnormal representation based on the abnormal representation combined with the historical monitoring data, feeding back a first feedback time length and carrying out deviation adjustment on a supervision result; the method improves the power grid fault data information computing capability to a certain extent. But only consider the bird pest condition in the electric wire netting engineering, it is incomplete to electric wire netting fault detection, when meetting the complicated data information processing of electric wire netting, it is difficult to improve electric wire netting route data information processing through data information processing.
Disclosure of Invention
Aiming at the defects of the technology, the invention discloses a power grid fault detection method and a system based on a machine vision technology, which can replace manual development of periodic comprehensive fault detection of a power grid, process different data information in the running process of the power grid and rapidly analyze the power grid fault data information. The automatic focusing control platform regulates and controls the position of the high-power fixed focus digital camera in real time to realize automatic focusing, so that the definition of the acquired power grid line image is ensured, the vision correction module corrects the restored image, the accuracy of power grid fault detection is improved, the similar image characteristic data is filtered by the similar filtering unit, the data analysis and calculation speed is improved, the power grid fault detection efficiency is improved, and the intelligent degree and the automation degree are high.
The invention adopts the following technical scheme:
a power grid fault detection method based on a machine vision technology comprises the following steps:
step one, carrying a scanning acquisition module to inspect and detect a power grid line through an unmanned aerial vehicle;
in the first step, the scanning acquisition module acquires a power grid line image through the high-power fixed focus digital camera, and automatically acquires the power grid line image at fixed time and transmits the power grid line image to the vision correction module for processing, and the high-power fixed focus digital camera regulates and controls the position of the high-power fixed focus digital camera in real time through the automatic focusing control platform to realize automatic focusing so as to ensure the definition of the acquired power grid line image;
Step two, filtering the clutter image and correcting the image by the power grid line image through a visual correction module, and correcting the restored image by the visual correction module through a gray level correction method;
analyzing the image data through an analysis alarm module and alarming the identified abnormal data;
in the third step, the corrected power grid line image extracts image feature data through a feature extraction unit and transmits the image feature data to a similar filtering unit, the analysis alarm module filters the similar image feature data through the similar filtering unit, an analysis calculation unit sets non-fault parameters based on multi-feature power grid line structure parameters, and if the image feature data is different from the non-fault parameters, faults are judged to occur and the analysis alarm module alarms through an alarm unit;
recording, backing up and regularly clearing the filtered image characteristic data through a data recording module;
in the fourth step, the data recording module records the image characteristic data of the power grid line through the data storage unit, and the data recording module realizes redundant backup, archiving control and useless data clearing of the image characteristic data of the power grid line through the data backup unit, the archiving control unit and the data clearing unit;
And fifthly, controlling a power grid fault detection process through a detection control module.
As a further technical scheme of the invention, the automatic focusing control platform comprises a host, a motion control card, a driver and a motor, wherein the host evaluates and analyzes the image focus performance of the power grid line through a definition evaluation function, and controls the motion control card, the driver and the motor to adjust the position of the high-power fixed-focus digital camera based on the analysis result so as to realize automatic focusing.
As a further technical scheme of the invention, the definition evaluation function judges the definition of the image through the variance of the adjacent pixel values of the power grid line image, and the output function formula of the definition evaluation function is as follows:
in the case of the formula (1),
for the variance of adjacent pixel values of the grid line image, < >>
Pixel coordinates of the grid line image, < +.>
The coordinates of adjacent pixels of the power grid line image are x, x is the abscissa of the pixels of the power grid line image, and y is the ordinate of the pixels of the power grid line image;
the definition evaluation function realizes the parallel judgment and integration evaluation of the grid line image definition in blocks by adopting a decomposition and combination operation mode, so as to improve the image definition judgment efficiency, wherein the decomposition and combination operation mode divides the pixel point coding arrangement of the grid line image into blocks 
Image sequence block, said sharpness evaluation function being based on +.>
And the image sequence blocks realize block parallel operation to obtain an M multiplied by M image sequence block definition data set, and the M multiplied by M image sequence block definition data set is integrated and compared for evaluation.
As a further technical scheme of the invention, the gray level correction method enhances the image deviation comparison effect of the power grid line through a gray level histogram, and accelerates the image deviation comparison process by adopting a Ten thousand sciences image processing accelerator;
the abscissa of the gray level histogram is gray level r, the ordinate is probability P of gray level occurrence, and the output function formula of probability P of gray level occurrence is:
in equation (2), N is the total number of pixels of the grid line image，
Is the pixel of the kth gray level of the gray level histogram,
for the kth gray level of the gray histogram, +.>
Representing a probability of occurrence of a kth gray level of the gray histogram;
the Vanken image processing accelerator adopts a memory mapping method, an outer core operation method and a delay calculation method to accelerate the image deviation comparison process.
As a further technical solution of the present invention, the implementation of the improved canny model includes the following steps:
step 3, calculating the amplitude and the direction of the image gradient through a maximum value calculation unit, and searching the local maximum value of the pixel point, wherein the output function formula of the gradient of the image point A of the power grid line in the gradient of the x and y directions is as follows:
in the formula (3), G is the gradient magnitude value of the grid line image point A, G x For the gradient value of the grid line image point A in the x-axis direction, G y For the gradient value of the grid line image point A in the y-axis direction, the angle output function formula of the gradient of the grid line image point A is as follows:
in the formula (4), θ is the gradient direction of the grid line image point a;
compared with the pixels along the corresponding gradient directions, the pixels of the central point B in the power grid line image field are reserved, if the pixels of the central point B are maximum, the pixels of the central point B are not maximum, the central gradient is set to zero, so that the point with the maximum local gradient is reserved, and the grid line image thinning edge is reserved;
step 4, selecting a high-low hysteresis threshold value through a threshold value setting unit, wherein the threshold value setting unit realizes self-adaptive determination of the threshold value through a maximum inter-class variance algorithm, the pixel number of a power grid line image is N, the gray level of the power grid line image is [0, L-1], the power grid line image is divided into a target area A and a background area B, and an output function formula of the maximum inter-class variance is as follows:
In the formula (5) of the present invention,
maximum variance of class AB +.>
Pixel weight of a +.>
The pixel weight value for B is given by,
for class a mean gray value +.>
For class B mean gray value +.>
For the overall average gray value of the grid line image, +.>
When the gray value T is the maximum value, the gray value T is the optimal estimation threshold value, and the hysteresis threshold value h is high 2 Low hysteresis threshold h =t 1 =h 2 2, reserving pixel points between the high hysteresis threshold and the low hysteresis threshold;
step 5, realizing compression acceleration of the Canni model through a data distillation module, compressing the Canni model by the data distillation module to obtain an improved Canni model, and transferring Canni model data to the improved Canni model to realize efficient operation of complex structure data, wherein the working method of the data distillation module is as follows:
and measuring the information influence quantity of the running state data of the power grid through a hyperbolic S transformation function, wherein the expression of the improved hyperbolic S transformation function is as follows:
in the formula (6), M represents an influence factor variable function,
representing the magnitude of the data, +.>
Representing the curvature of the variable function; d represents a matching function, alpha represents a condition conversion criterion factor, beta represents a fault data information conversion factor, t represents an abnormal parameter variation of the power grid, and tau represents an abnormal data fusion variation of the power grid;
Wherein the data amplitude variation function:
in the formula (7), the amino acid sequence of the compound,
the method is characterized in that an S transformation function is improved as a data amplitude variation function, the detected fault data reaction is realized by adding an asymmetric variable and a hyperbolic variable, and an output function formula of an improved S transformation function formula is as follows:
in (8)，
Representing an improvement S transformation function formula, f representing hyperbolic function variables, g representing asymmetric function variables;
measuring the power grid fault data information classification calculation through the unmatched function; the numerical calculation function of attribute mismatch is:
in formula (9), D a The dissimilarity of the two classified attribute power grid perception data is represented, a and b respectively represent dissimilarity parameters, and the parameters range is [0, q]The method comprises the steps of carrying out a first treatment on the surface of the i represents the number of data; δ（a i ,b i ) The value function of (2) is:
the formula for measuring the classification dissimilarity of the power grid data information is as follows:
in the formula (11) of the present invention,D 1 (a i ,b j ) For the classification dissimilarity of the power grid data information, q is the number of classification properties, a ip ，a jp Values of i and j in the P-th dimension attribute data; delta (a) ip ,a jp ) The value function of (2) is:
the collected samples of the power grid perception data comprise m pieces of data of classified properties and n pieces of data items, and a dissimilarity formula for measuring the classified property data is as follows:
in formula (13), D 2 (a i ,a j ) For dissimilarity of classified property data, n ip ，n jp Representing the number of occurrences of i and j in the data, n being the total number of occurrences of the property data; the numerical attribute data dissimilarity measure formula is:
in formula (14), D 3 (a i ,a j ) Is the dissimilarity of the numerical property data, q is the number of the numerical property data, a im ，a jm The values of i and j in the mth dimension attribute data.
As a further technical scheme of the invention, the data recording module realizes classified storage and archiving management of the image characteristic data of the power grid line through the data resource manager Azure so as to improve the extraction speed of the image characteristic data during inquiry.
As a further technical scheme of the invention, the output function formula of the non-fault parameters is as follows:
in the case of the formula (15),
for non-fault parameters of the power grid line, n is the number of structural parameters of the power grid line with multiple characteristics, and +.>
For the multi-element characteristic power grid line structure parameter +.>
For the identification of the image characteristics of the network lines,/->
。
The invention also adopts the following technical scheme:
a machine vision technology-based grid fault detection system, the grid fault detection system comprising the following modules:
the scanning acquisition module is used for inspecting and detecting the power grid line, the scanning acquisition module acquires the power grid line image through the high-power fixed-focus digital camera, timely and automatically captures the power grid line image and transmits the power grid line image to the vision correction module for processing, and the high-power fixed-focus digital camera regulates and controls the position of the high-power fixed-focus digital camera in real time through the automatic focusing control platform to realize automatic focusing so as to ensure the definition of the acquired image;
The visual correction module is used for filtering the clutter image and correcting the image, and correcting the restored image through a gray correction method;
the analysis alarm module is used for analyzing the image data and alarming the identified abnormal data, and comprises a feature extraction unit, a similar filtering unit, an analysis calculation unit and an alarm unit, wherein the output end of the feature extraction unit is connected with the input end of the analysis calculation unit, the output end of the analysis calculation unit is connected with the input end of the similar filtering unit, the output end of the similar filtering unit is connected with the input end of the analysis calculation unit, and the output end of the analysis calculation unit is connected with the input end of the alarm unit;
the data recording module is used for recording, backing up and clearing mass acquired image data and comprises a data storage unit, a data backup unit, an archiving control unit and a data clearing unit, wherein the output end of the data storage unit is connected with the input end of the data backup unit, the output end of the data storage unit is connected with the input end of the archiving control unit, and the output end of the data storage unit is connected with the input end of the data clearing unit;
The detection control module is used for controlling the power grid fault detection process, and the detection control module realizes the decentralized control and the centralized management of the power grid fault detection process through a distributed control system DCS;
the output end of the scanning acquisition module is connected with the input end of the data recording module, the output end of the scanning acquisition module is connected with the input end of the vision correction module, the output end of the vision correction module is connected with the input end of the analysis alarm module, and the output end of the analysis alarm module is connected with the input end of the detection control module.
As a further technical scheme of the invention, the similarity filtering unit realizes the high-speed filtering of the missing data through the large data collecting and editing NLPIR-IFCA system, and the large data collecting and editing NLPIR-IFCA system realizes the similar classification and filtering of the power grid line image data through machine learning so as to remove redundant data and improve the data analysis and calculation speed.
As a further technical scheme of the invention, the feature extraction unit realizes fragment type feature information extraction by circuit control based on a CS5463 chip;
the similar filtering unit is used for filtering similar data information of the power grid through an AMIS-49587 chip circuit;
The analysis and calculation unit also comprises a Clara algorithm model and a logistic regression model;
the alarm unit is an audible and visual information alarm unit.
Has the positive beneficial effects that:
according to the invention, periodic comprehensive fault detection of the power grid can be carried out instead of manual work, the existence of faults is analyzed, the automatic focusing is realized by regulating and controlling the position of the high-power fixed focus digital camera in real time through the automatic focusing control platform, the definition of the collected power grid line image is ensured, the restored image is corrected through the vision correction module, the accuracy of the fault detection of the power grid is improved, the characteristic data of the similar image is filtered through the similar filtering unit, the data analysis and calculation speed is improved, the fault detection efficiency of the power grid is improved, and the intelligent degree and the automation degree are high.
According to the invention, the power grid fault data information is processed and identified to improve the power grid fault detection capability, and the power grid data information identification and processing are converted into artificial intelligent information calculation and function information calculation to improve the power grid fault data information calculation capability, so that the power grid fault data information detection and application capability is improved.
Drawings
For a clearer description of an embodiment of the invention or of a technical solution in the prior art, the drawings that are necessary for the description of the embodiment or of the prior art will be briefly described, it being obvious that the drawings in the description below are only some embodiments of the invention, from which, without inventive faculty, other drawings are obtained for a person skilled in the art, in which:
FIG. 1 is a schematic overall flow diagram of a power grid fault detection method based on machine vision technology;
FIG. 2 is a schematic flow chart of the first and second steps in the power grid fault detection method based on the machine vision technology;
FIG. 3 is a schematic flow chart of a third step in the power grid fault detection method based on the machine vision technology;
FIG. 4 is a schematic diagram of the overall architecture of a power grid fault detection system based on machine vision technology according to the present invention;
FIG. 5 is a schematic diagram of an autofocus control platform in a machine vision based power grid fault detection system according to the present invention;
FIG. 6 is a schematic diagram of a CS5463 chip circuit in a power grid fault detection system based on machine vision technology according to the present invention;
FIG. 7 is a schematic diagram of an AMIS-49587 chip operating circuit in a power grid fault detection system based on machine vision technology;
fig. 8 is a schematic diagram of an embodiment of a power grid fault detection system based on machine vision technology according to the present invention.
Detailed Description
The preferred embodiments of the present invention will be described below with reference to the accompanying drawings, it being understood that the embodiments described herein are for illustration and explanation of the present invention only, and are not intended to limit the present invention.
As shown in fig. 1-5, a power grid fault detection method based on a machine vision technology includes the following steps:
step one, carrying a scanning acquisition module to inspect and detect a power grid line through an unmanned aerial vehicle;
in the first step, the scanning acquisition module acquires a power grid line image through the high-power fixed focus digital camera, and automatically acquires the power grid line image at fixed time and transmits the power grid line image to the vision correction module for processing, and the high-power fixed focus digital camera regulates and controls the position of the high-power fixed focus digital camera in real time through the automatic focusing control platform to realize automatic focusing so as to ensure the definition of the acquired power grid line image;
step two, filtering the clutter image and correcting the image by the power grid line image through a visual correction module, and correcting the restored image by the visual correction module through a gray level correction method;
analyzing the image data through an analysis alarm module and alarming the identified abnormal data;
in the third step, the corrected power grid line image extracts image feature data through a feature extraction unit and transmits the image feature data to a similar filtering unit, the analysis alarm module filters the similar image feature data through the similar filtering unit, an analysis calculation unit sets non-fault parameters based on multi-feature power grid line structure parameters, and if the image feature data is different from the non-fault parameters, faults are judged to occur and the analysis alarm module alarms through an alarm unit;
Recording, backing up and regularly clearing the filtered image characteristic data through a data recording module;
in the fourth step, the data recording module records the image characteristic data of the power grid line through the data storage unit, and the data recording module realizes redundant backup, archiving control and useless data clearing of the image characteristic data of the power grid line through the data backup unit, the archiving control unit and the data clearing unit;
and fifthly, controlling a power grid fault detection process through a detection control module.
In the above embodiment, the automatic focusing control platform includes a host, a motion control card, a driver and a motor, where the host evaluates and analyzes the image focal performance of the power grid line through a definition evaluation function, and controls the motion control card, the driver and the motor to adjust the position of the high-power fixed-focus digital camera based on the analysis result to realize automatic focusing.
In the above embodiment, the data information acquisition capability is improved by applying the unmanned aerial vehicle technology, the running environment of the power grid is complex and changeable, and the high-altitude high-risk data information can be acquired by carrying the scanning acquisition module through the unmanned aerial vehicle.
In the above embodiment, the sharpness evaluation function determines the sharpness of the image according to the variance of the adjacent pixel values of the power grid line image, and the output function formula of the sharpness evaluation function is:
In the case of the formula (1),
for the variance of adjacent pixel values of the grid line image, < >>
Pixel coordinates of the grid line image, < +.>
The coordinates of adjacent pixels of the power grid line image are x, x is the abscissa of the pixels of the power grid line image, and y is the ordinate of the pixels of the power grid line image;
the definition evaluation function realizes the parallel judgment and integration evaluation of the grid line image definition in blocks by adopting a decomposition and combination operation mode, so as to improve the image definition judgment efficiency, wherein the decomposition and combination operation mode divides the pixel point coding arrangement of the grid line image into blocks
Image sequence block, said sharpness evaluation function being based on +.>
And the image sequence blocks realize block parallel operation to obtain an M multiplied by M image sequence block definition data set, and the M multiplied by M image sequence block definition data set is integrated and compared for evaluation.
In a specific embodiment, when the automatic focusing control platform is used, the camera is used for collecting the object image and transmitting the object image to the PC end in real time, then the PC end is used for firstly cutting the image into areas, then carrying out gray processing and DFT filtering, and then calculating the definition of the preprocessed image by using the definition evaluation function. To compare the effect of different evaluation functions, the values were normalized.
In the formula (2), STD is standard image definition,
the function value is evaluated for the sharpness of the current image,
is->
Mean value of->
Indicating a normalized value, the larger the value, the higher the sharpness of the image. In the process of collecting the image definition value, the transmission device continuously drives the camera to move towards the direction close to the focusing object, and when the image definition value is collected by the computer, +.>
When the value changes from small to large to small obviously, the whole automatic focusing process is completed, and the transmission device drives the camera to return to +.>
At maximum, i.e. at auto-focus. And normalizing values of camera positions and evaluation functions after focusing when different object heights are achieved.
As the camera moves, the sharpness of the image undergoes a blur-sharpness-blur process. In order to evaluate the definition of the images in the process, the data obtained by processing the image sequences with different evaluation functions are normalized to form a definition curve.
The abscissa represents the image sequence acquired by the upper graph, and the ordinate represents the normalized evaluation function value. The figure shows that the extreme point positions of the 3 evaluation functions are the same, and the curve has only one maximum value, namely has unimodal property, so that the automatic focusing function of the system is accurate and reliable. At the same time, the clearest image is located in the image sequence 6, and the position where the camera is stopped is just the focal position of the camera, which indicates that each evaluation function has no bias. By comparing different evaluation functions, the method has unimodal property and unbiased property, can obviously distinguish focusing and defocusing states, and can be used as an image definition evaluation criterion of an automatic focusing system. The difference is that the half-width of the curve obtained by the Laplacian gradient evaluation function is lower than that of other evaluation functions, which shows that the sensitivity of the evaluation function is higher than that of other evaluation functions, the evaluation function is more suitable to be used as the image definition evaluation function of the system, and the focusing method realizes automatic focusing of the camera under the condition of no reference gallery in industrial production and has better practical application effect.
In the above embodiment, the gray level correction method enhances the image deviation comparison effect of the power grid line through the gray level histogram, and accelerates the image deviation comparison process by adopting the universal image processing accelerator;
the abscissa of the gray level histogram is gray level r, the ordinate is probability P of gray level occurrence, and the output function formula of probability P of gray level occurrence is:
in equation (3), N is the total number of pixels of the grid line image,
is the pixel of the kth gray level of the gray level histogram,
for the kth gray level of the gray histogram, +.>
Straight for representing gray scaleFang Tudi k gray scale occurrences;
the Vanken image processing accelerator adopts a memory mapping method, an outer core operation method and a delay calculation method to accelerate the image deviation comparison process.
In particular embodiments, we can review some properties of an image by the state of the histogram: the histogram of the bright image tends to be on the side of the gray level; the histogram of the low contrast image is narrow and concentrated in the middle of the gray level, the histogram component of the high contrast image covers a very wide gray level and the distribution of pixels is not so uniform, only a small number of vertical lines are much higher than others. Intuitively, it is: an image has high contrast and variable gray tone if its pixels occupy all possible gray levels and are uniformly distributed.
In the above embodiment, the implementation of the modified canny model includes the following steps:
step 3, calculating the amplitude and the direction of the image gradient through a maximum value calculation unit, and searching the local maximum value of the pixel point, wherein the output function formula of the gradient of the image point A of the power grid line in the gradient of the x and y directions is as follows:
in the formula (4), G is the gradient magnitude value of the grid line image point A, G x For the gradient value of the grid line image point A in the x-axis direction, G y For the gradient value of the grid line image point A in the y-axis direction, the angle output function formula of the gradient of the grid line image point A is as follows:
in the formula (5), θ is the gradient direction of the grid line image point a;
compared with the pixels along the corresponding gradient directions, the pixels of the central point B in the power grid line image field are reserved, if the pixels of the central point B are maximum, the pixels of the central point B are not maximum, the central gradient is set to zero, so that the point with the maximum local gradient is reserved, and the grid line image thinning edge is reserved;
Step 4, selecting a high-low hysteresis threshold value through a threshold value setting unit, wherein the threshold value setting unit realizes self-adaptive determination of the threshold value through a maximum inter-class variance algorithm, the pixel number of a power grid line image is N, the gray level of the power grid line image is [0, L-1], the power grid line image is divided into a target area A and a background area B, and an output function formula of the maximum inter-class variance is as follows:
in the formula (6) of the present invention,
maximum variance of class AB +.>
Pixel weight of a +.>
The pixel weight value for B is given by,
for class a mean gray value +.>
For class B mean gray value +.>
For the overall average gray value of the grid line image, +.>
The gray value T is the optimal estimate when the gray value T is the maximum valueCount threshold, high hysteresis threshold h 2 Low hysteresis threshold h =t 1 =h 2 2, reserving pixel points between the high hysteresis threshold and the low hysteresis threshold;
step 5, realizing compression acceleration of the Canni model through a data distillation module, compressing the Canni model by the data distillation module to obtain an improved Canni model, and transferring Canni model data to the improved Canni model to realize efficient operation of complex structure data, wherein the working method of the data distillation module is as follows:
and measuring the information influence quantity of the running state data of the power grid through a hyperbolic S transformation function, wherein the expression of the improved hyperbolic S transformation function is as follows:
In the formula (7), M represents an influence factor variable function,
representing the magnitude of the data, +.>
Representing the curvature of the variable function; d represents a matching function, alpha represents a condition conversion criterion factor, beta represents a fault data information conversion factor, t represents an abnormal parameter variation of the power grid, and tau represents an abnormal data fusion variation of the power grid;
wherein the data amplitude variation function:
in the formula (8), the amino acid sequence of the compound,
the method is characterized in that an S transformation function is improved as a data amplitude variation function, the detected fault data reaction is realized by adding an asymmetric variable and a hyperbolic variable, and an output function formula of an improved S transformation function formula is as follows:
in the formula (9), the amino acid sequence of the compound,
representing an improvement S transformation function formula, f representing hyperbolic function variables, g representing asymmetric function variables;
measuring the power grid fault data information classification calculation through the unmatched function; the numerical calculation function of attribute mismatch is:
in the formula (10), D a The dissimilarity of the two classified attribute power grid perception data is represented, a and b respectively represent dissimilarity parameters, and the parameters range is [0, q]The method comprises the steps of carrying out a first treatment on the surface of the i represents the number of data; δ（a i ,b i ) The value function of (2) is:
the formula for measuring the classification dissimilarity of the power grid data information is as follows:
in the formula (12) of the present invention,D 1 (a i ,b j ) For the classification dissimilarity of the power grid data information, q is the number of classification properties, a ip ，a jp Values of i and j in the P-th dimension attribute data; delta (a) ip ,a jp ) The value function of (2) is:
the collected samples of the power grid perception data comprise m pieces of data of classified properties and n pieces of data items, and a dissimilarity formula for measuring the classified property data is as follows:
in formula (14), D 2 (a i ,a j ) For dissimilarity of classified property data, n ip ，n jp Representing the number of occurrences of i and j in the data, n being the total number of occurrences of the property data; the numerical attribute data dissimilarity measure formula is:
in formula (15), D 3 (a i ,a j ) Is the dissimilarity of the numerical property data, q is the number of the numerical property data, a im ，a jm The values of i and j in the mth dimension attribute data.
In a specific embodiment, the modified Canni model is an optimal image edge detection model. The basic idea of the model is to traverse all gradient amplitude values, find out the maximum value, respectively process the strong edge and the weak edge by using two threshold value ranges, and connect the strong edge and the weak edge, so that the edge effect is completely presented, and the effect of the algorithm on noise suppression is very good. The steps of the improved Canni model are as follows:
(1) smoothing the image with gaussian filtering;
(2) calculating the gradient amplitude and gradient direction by using a first-order differential operator;
(3) Carrying out a non-maximum suppression algorithm on the obtained gradient amplitude to find out local maximum points in the image gradient, and then setting other maximum points to zero to refine the edge of the image;
(4) setting a double threshold of the Canny operator. The criteria for edge judgment are: edges are the ones that are greater than the high threshold; sobel is less than the low threshold and is not an edge; the size is between the two, and whether the edge is judged according to the relation between the value of the adjacent pixel and the high threshold value and the low threshold value;
in order to more intuitively verify the effectiveness of the method, an image acquired in the inspection process is selected and processed by various algorithms, and the advantages and disadvantages of the various algorithms are observed and compared. The power line is the only detection target, however, ground flowers and plants, a telegraph pole and a very obvious road form a complex background, the gray level difference between the target and the background is not large, and the extraction of the power line is difficult. The quality and the resolution of the image are relatively improved through the filtering deblurring in the earlier stage, and the effect of the first-order differential gradient operator is basically similar through the simple analysis of the various algorithms, so that the selected image is subjected to edge detection by only selecting a Roberts model, a LOG model, a Canny model and an improved Canny model, wherein the Canny model adopts a threshold value automatically selected by a system.
In the above embodiment, the data recording module realizes classified storage and archiving management of the image feature data of the power grid line through the data resource manager Azure, so as to improve the extraction speed of the image feature data during query.
In a specific embodiment, the Azure data resource manager is a fully hosted high-performance and large data platform, so that you can easily analyze and record a large amount of data in real time. The Azure data resource manager toolkit provides an end-to-end solution for data import, query, visualization and management, the Azure data resource manager aims to provide interactive and thermal path analysis through APIs for massive data workloads, by analyzing structured, semi-structured and unstructured data across time sequences, and by using machine learning, the Azure data resource manager allows you to easily extract key insights, find patterns and trends, and create predictive models. The Azure data resource manager is scalable, secure, reliable, and enterprise-ready, and is very useful for log analysis, time series analysis, ioT, and universal heuristics analysis.
In the above embodiment, the output function formula of the non-fault parameter is:
in the formula (16) of the present invention, 
For electric networkThe line non-fault parameters, n is the number of the line structure parameters of the multi-element characteristic power grid,
for the multi-element characteristic power grid line structure parameter +.>
For the identification of the image characteristics of the network lines,/->
。
In a specific embodiment, the data information is converted into a hyperbolic S transformation function so as to vividly represent the power grid fault data information.
In a specific embodiment, the variable function of the influencing factor is specifically expressed as data information influencing accuracy in the power grid detection process, such as the influence degree of the unmanned aerial vehicle by external data information in the flight process. Such as by external data information such as magnetic fields, electric fields, detection environments, etc. The invention converts different data information influence quantity into function expression or curve expression to intuitively express the function expression.
The data amplitude variation measuring tool is expressed as the data information variation amplitude, namely the variation is increased or decreased, the data information variation increment is carried out on the highest value and the lowest value as a result of addition and subtraction, so that the variation difference value of different parameter amplitudes in the calculation process of the power grid fault data information parameters is improved.
The curvature of the variable function is specifically expressed in a mode of changing different parameters through S curvature so as to improve the computing capability of the data information.
The matching function is specifically expressed as calculating the standard number data information and the matching module data information so as to improve the matching calculation capability of the data information. Thereby improving the data information computing capability.
The condition conversion criterion factor is specifically expressed as comparing a set of real data (such as actually detected data information) under the condition of reference of a set of power grid equipment operation, power grid fault information detection or power grid communication node data information standard with another set of data under a new condition, so that a specific unit can be deduced to measure the size of a certain attribute. To improve data information computing power.
The fault data information conversion factor is specifically expressed as for example the ability to convert analog data information into digital data information,
the grid anomaly parameter variation is specifically represented as a value of one anomaly data message changing from one data state to another.
The power grid abnormal data fusion change measuring tool is expressed as a difference quantity which changes from one data fusion state to another data fusion state in the data fusion state.
And through the function calculation, the acquired data information is presented in a curve mode, so that the visual display capability of the data information is improved.
In a specific embodiment, a variety of data information contained in the power grid running state data information and the fault data information change range is between 0 and 1, and the data amplitude change quantity function is as follows:
in the formula (17), the amino acid sequence of the compound,
in the formula (17), for the abnormal data fluctuation curve of the power grid, if the change amplitude is 0, the function curve is stable; if the change amplitude is 0, it indicates that the function curve has fluctuation. />
Improving the S transformation function, and realizing the detected fault data reaction by adding an asymmetric variable and a hyperbolic variable, wherein the fault data reaction is shown as a formula (18):
in the formula (18), the amino acid sequence of the compound,
representing an improvement S transformation function formula, f representing hyperbolic function variables, g representing asymmetric function variables;
according to the analysis curve in the formula (18), mapping can be carried out on any detection data by improving hyperbolic S transformation, so that factors affecting abnormal data can be completely analyzed, meanwhile, the whole rule analysis of different factors is better carried out, and the data assurance is provided for subsequent abnormal data reason evidence.
The dissimilarity method for measuring the classified property data comprises the following steps: the method is set in the acquired power grid sensing data samples. Each data contains q categorical properties, and a, b is used to represent parameters of the categorical properties (e.g., voltage, current, temperature, jitter, etc.), the dissimilarity of a, b is measured to determine the value of the mismatch of the properties. The smaller the value after the determination, the higher the similarity of the two parameters a and b.
The numerical calculation function of attribute mismatch is:
in the formula (19), D a Representing the dissimilarity of two classified attribute power grid awareness data, ranging from [0, q]The method comprises the steps of carrying out a first treatment on the surface of the i is the number; δ（a i ,b i ) The value function of (2) is:
the formula for measuring the classification dissimilarity of the power grid data information is as follows:
in the formula (21), the amino acid sequence of the amino acid,D 1 (a i ,b j ) For the classification dissimilarity of the power grid data information, q is the number of classification properties, a ip ，a jp Values of i and j in the P-th dimension attribute data; delta (a) ip ,a jp ) Is taken from (a)The value function is:
the collected samples of the power grid sensing data comprise m pieces of data with classified properties and n pieces of data, and the dissimilarity between the data and the frequency of the data under the same property have a direct relation. The dissimilarity formula for measuring the classified property data is:
（23）
in the formula (23), D 2 (a i ,a j ) For dissimilarity of classified property data, n ip ，n jp The number of occurrences of i and j in the data is represented, and n is the total number of occurrences of the property data.
The dissimilarity process of the measured numerical property data is as follows: for the numerical property data in the collected power grid sensing data sample, in a specific embodiment, the wavelet analysis method can be mainly adopted to carry out quantization processing on the converted power grid sensing data signal, so as to measure the dissimilarity between the data. The numerical attribute data dissimilarity measure formula is:
In the formula (24), D 3 (a i ,a j ) Is the dissimilarity of the numerical property data, q is the number of the numerical property data, a im ，a jm The values of i and j in the mth dimension attribute data. Through the above process, two kinds of property data contained in the samples of the power grid sensing data signals can be obtained: the dissimilarity between the classified property data and the numerical property data supports the similarity relation of the power grid sensing data signals.
In the above embodiment, as shown in fig. 6 to 8, the following technical solutions are further adopted in the present invention: a machine vision technology-based grid fault detection system, the grid fault detection system comprising the following modules:
the scanning acquisition module is used for inspecting and detecting the power grid line, the scanning acquisition module acquires the power grid line image through the high-power fixed-focus digital camera, timely and automatically captures the power grid line image and transmits the power grid line image to the vision correction module for processing, and the high-power fixed-focus digital camera regulates and controls the position of the high-power fixed-focus digital camera in real time through the automatic focusing control platform to realize automatic focusing so as to ensure the definition of the acquired image;
the visual correction module is used for filtering the clutter image and correcting the image, and correcting the restored image through a gray correction method;
The analysis alarm module is used for analyzing the image data and alarming the identified abnormal data, and comprises a feature extraction unit, a similar filtering unit, an analysis calculation unit and an alarm unit, wherein the output end of the feature extraction unit is connected with the input end of the analysis calculation unit, the output end of the analysis calculation unit is connected with the input end of the similar filtering unit, the output end of the similar filtering unit is connected with the input end of the analysis calculation unit, and the output end of the analysis calculation unit is connected with the input end of the alarm unit;
the data recording module is used for recording, backing up and clearing mass acquired image data and comprises a data storage unit, a data backup unit, an archiving control unit and a data clearing unit, wherein the output end of the data storage unit is connected with the input end of the data backup unit, the output end of the data storage unit is connected with the input end of the archiving control unit, and the output end of the data storage unit is connected with the input end of the data clearing unit;
the detection control module is used for controlling the power grid fault detection process, and the detection control module realizes the decentralized control and the centralized management of the power grid fault detection process through a distributed control system DCS;
The output end of the scanning acquisition module is connected with the input end of the data recording module, the output end of the scanning acquisition module is connected with the input end of the vision correction module, the output end of the vision correction module is connected with the input end of the analysis alarm module, and the output end of the analysis alarm module is connected with the input end of the detection control module.
In the above embodiment, the similarity filtering unit realizes high-speed filtering of the acacia data through the large data collecting and editing NLPIR-IFCA system, and the large data collecting and editing NLPIR-IFCA system realizes similar classification and filtering of the power grid line image data through machine learning so as to remove redundant data and improve the data analysis and calculation speed.
In a further embodiment, the feature extraction unit implements fragmented feature information extraction by CS5463 based chip circuit control;
the similar filtering unit is used for filtering similar data information of the power grid through an AMIS-49587 chip circuit;
the analysis and calculation unit also comprises a Clara algorithm model and a logistic regression model;
the alarm unit is an audible and visual information alarm unit.
In a further embodiment, in order to improve the accuracy of the existing electric energy data acquisition system and the speed of acquiring and transmitting data, the electric energy data acquisition system is designed based on the chip AMIS-49587 and the chip CS5463, and the system can accurately measure real-time voltage, current, instantaneous power and other data and improve the communication anti-interference capability during data transmission. The chip used in the electric energy data acquisition system of the research design adopts a CS5463 chip and an AMIS ⁃ 49587 chip, each chip is responsible for different functional parts, the chip CS5463 is responsible for integral electric energy data acquisition, and the chip AMIS-49587 is responsible for the communication part of data
First, the CS5463 chip, which is responsible for the data sampling calculation processing part, includes an analog-to-digital converter and a power calculation function, and also includes an electric energy/frequency converter. The chip has a bidirectional serial port communicated with the controller, and has a programmable electric energy-pulse output function so as to calibrate errors, ensure the accuracy of the system and ensure stable operation
Vin+ Vin-written on the chip CS5463 is a sampling differential voltage data port and includes a multiplication amplifier, iin+ Iin-is a sampling differential current data port and includes a programmable amplifier, the PFMON interface has a power supply detecting portion in the chip, and Xin, xout and CPUCLK are clock generating modules in the chip. The CS, SDI, SDO, SCLK interface is a serial interface module therein. E1, E2 have power/frequency converter sections therein.
The second part of the power data acquisition system is the AMIS ⁃ 49587 modem chip responsible for the communication function. The modulation mode adopted by the intelligent electric energy data acquisition system is a frequency shift keying S-FSK modulation mode, and the modulation mode has the advantages of being capable of easily coping with common narrow-band interference in a power grid and being more suitable for being used in the intelligent electric energy data acquisition system. In contrast, the modulation technique of phase shift keying PSK is less reliable than the frequency shift keying modulation, although it is less costly than the frequency shift keying modulation used in the chip.
The AMIS-49587 chip running circuit consists of an AMIS-49587 chip, a power supply circuit and an NCS5650 chip. The NCS5650 chip functions to power amplify the signal and simultaneously low pass filter the signal. The AMIS ⁃ 49587 modem chip adopts a UART communication mode, has a very flexible modem mode and an excellent demodulation algorithm, and also has a very good interference resistance. The AMIS ⁃ 49587 modem chip is also internally provided with a protocol processing function, so that the software development cost can be greatly saved. The time required for design is reduced. The above is the hardware design part of the electric energy data acquisition system.
While specific embodiments of the present invention have been described above, it will be understood by those skilled in the art that the foregoing detailed description is given by way of example only, and that various omissions, substitutions and changes in the form of the details of the method and system illustrated may be made by those skilled in the art without departing from the spirit and scope of the invention; for example, it is within the scope of the present invention to combine the above-described method steps to perform substantially the same function in substantially the same way to achieve substantially the same result; accordingly, the scope of the invention is limited only by the following claims.
Claims (10)
1. A power grid fault detection method based on a machine vision technology is characterized by comprising the following steps of: comprises the following steps:
step one, carrying a scanning acquisition module to inspect and detect a power grid line through an unmanned aerial vehicle;
in the first step, the scanning acquisition module acquires a power grid line image through the high-power fixed focus digital camera, and automatically acquires the power grid line image at fixed time and transmits the power grid line image to the vision correction module for processing, and the high-power fixed focus digital camera regulates and controls the position of the high-power fixed focus digital camera in real time through the automatic focusing control platform to realize automatic focusing so as to ensure the definition of the acquired power grid line image;
step two, filtering the clutter image and correcting the image by the power grid line image through a visual correction module, and correcting the restored image by the visual correction module through a gray level correction method;
analyzing the image data through an analysis alarm module and alarming the identified abnormal data;
in the third step, the corrected power grid line image extracts image feature data through a feature extraction unit and transmits the image feature data to a similar filtering unit, the analysis alarm module filters the similar image feature data through the similar filtering unit, an analysis calculation unit sets non-fault parameters based on multi-feature power grid line structure parameters, and if the image feature data is different from the non-fault parameters, faults are judged to occur and the analysis alarm module alarms through an alarm unit;
Recording, backing up and regularly clearing the filtered image characteristic data through a data recording module;
the characteristic extraction unit extracts the power grid line characteristics through an improved Canni model, and the improved Canni model comprises a noise reduction unit, an image segmentation unit, a maximum value calculation unit, a threshold setting unit and a data distillation module;
in the fourth step, the data recording module records the image characteristic data of the power grid line through the data storage unit, and the data recording module realizes redundant backup, archiving control and useless data clearing of the image characteristic data of the power grid line through the data backup unit, the archiving control unit and the data clearing unit;
and fifthly, controlling a power grid fault detection process through a detection control module.
2. The machine vision technology-based power grid fault detection method as claimed in claim 1, wherein: the automatic focusing control platform comprises a host, a motion control card, a driver and a motor, wherein the host evaluates and analyzes the power grid line image focus performance through a definition evaluation function, and controls the motion control card, the driver and the motor to adjust the position of the high-power fixed-focus digital camera based on an analysis result so as to realize automatic focusing.
3. The machine vision technology-based power grid fault detection method as claimed in claim 2, wherein: the definition evaluation function judges the definition of the image through the variance of adjacent pixel values of the power grid line image, and an output function formula of the definition evaluation function is as follows:
in the case of the formula (1),
for the variance of adjacent pixel values of the grid line image, < >>
Pixel coordinates of the grid line image, < +.>
The coordinates of adjacent pixels of the power grid line image are x, x is the abscissa of the pixels of the power grid line image, and y is the ordinate of the pixels of the power grid line image;
the resolution evaluation function adopts a decomposition and combination operation mode to realize the parallel judgment and integration evaluation of the resolution of the power grid line image in blocks so as to improve the image resolution judgment efficiency, and the decomposition and combination operation mode divides the pixel point coding arrangement of the power grid line imageCutting to form
Image sequence block, said sharpness evaluation function being based on +.>
And the image sequence blocks realize block parallel operation to obtain an M multiplied by M image sequence block definition data set, and the M multiplied by M image sequence block definition data set is integrated and compared for evaluation.
4. The machine vision technology-based power grid fault detection method as claimed in claim 1, wherein: the gray level correction method enhances the image deviation comparison effect of the power grid line through a gray level histogram, and accelerates the image deviation comparison process by adopting a Ten thousand sciences image processing accelerator;
The abscissa of the gray level histogram is gray level r, the ordinate is probability P of gray level occurrence, and the output function formula of probability P of gray level occurrence is:
in equation (2), N is the total number of pixels of the grid line image,
pixels of the kth gray level of the gray level histogram, < >>
For the kth gray level of the gray histogram, +.>
Representing a probability of occurrence of a kth gray level of the gray histogram;
the Vanken image processing accelerator adopts a memory mapping method, an outer core operation method and a delay calculation method to accelerate the image deviation comparison process.
5. The machine vision technology-based power grid fault detection method as claimed in claim 1, wherein: the implementation of the improved Canni model comprises the following steps:
step 1, eliminating image noise through a noise reduction unit, wherein the noise reduction unit eliminates power grid line image noise through Gaussian smoothing filter convolution;
step 2, the image segmentation unit identifies the image block edge by detecting the gray level of the power grid line image, and segments the image based on the image block edge;
step 3, calculating the amplitude and the direction of the image gradient through a maximum value calculation unit, and searching the local maximum value of the pixel point, wherein the output function formula of the gradient of the image point A of the power grid line in the gradient of the x and y directions is as follows:
In the formula (3), G is the gradient magnitude value of the grid line image point A, G x For the gradient value of the grid line image point A in the x-axis direction, G y For the gradient value of the grid line image point A in the y-axis direction, the angle output function formula of the gradient of the grid line image point A is as follows:
in the formula (4), θ is the gradient direction of the grid line image point a;
compared with the pixels along the corresponding gradient directions, the pixels of the central point B in the power grid line image field are reserved, if the pixels of the central point B are maximum, the pixels of the central point B are not maximum, the central gradient is set to zero, so that the point with the maximum local gradient is reserved, and the grid line image thinning edge is reserved;
step 4, selecting a high-low hysteresis threshold value through a threshold value setting unit, wherein the threshold value setting unit realizes self-adaptive determination of the threshold value through a maximum inter-class variance algorithm, the pixel number of a power grid line image is N, the gray level of the power grid line image is [0, L-1], the power grid line image is divided into a target area A and a background area B, and an output function formula of the maximum inter-class variance is as follows:
in the formula (5) of the present invention,
maximum variance of class AB +.>
Pixel weight of a +.>
Pixel weight of B, +. >
For class a mean gray value +.>
For class B mean gray value +.>
For the overall average gray value of the grid line image, +.>
When the gray value T is the maximum value, the gray value T is the optimal estimation threshold value, and the hysteresis threshold value h is high 2 Low hysteresis threshold h =t 1 =h 2 2, reserving pixel points between the high hysteresis threshold and the low hysteresis threshold;
step 5, realizing compression acceleration of the Canni model through a data distillation module, compressing the Canni model by the data distillation module to obtain an improved Canni model, and transferring Canni model data to the improved Canni model to realize efficient operation of complex structure data, wherein the working method of the data distillation module is as follows: and measuring the information influence quantity of the running state data of the power grid through a hyperbolic S transformation function, wherein the expression of the improved hyperbolic S transformation function is as follows:
in the formula (6), M represents an influence factor variable function,
representing the magnitude of the data, +.>
Representing the curvature of the variable function; d represents a matching function, alpha represents a condition conversion criterion factor, beta represents a fault data information conversion factor, t represents an abnormal parameter variation of the power grid, and tau represents an abnormal data fusion variation of the power grid;
wherein the data amplitude variation function:
in the formula (7), the amino acid sequence of the compound,
the method is characterized in that an S transformation function is improved as a data amplitude variation function, the detected fault data reaction is realized by adding an asymmetric variable and a hyperbolic variable, and an output function formula of an improved S transformation function formula is as follows:
In the formula (8), the amino acid sequence of the compound,
representing an improvement S transformation function formula, f representing hyperbolic function variables, g representing asymmetric function variables;
measuring the power grid fault data information classification calculation through the unmatched function; the numerical calculation function of attribute mismatch is:
in formula (9), D a The dissimilarity of the two classified attribute power grid perception data is represented, a and b respectively represent dissimilarity parameters, and the parameters range is [0, q]The method comprises the steps of carrying out a first treatment on the surface of the i represents the number of data;δ（a i ,b i ) The value function of (2) is:
the formula for measuring the classification dissimilarity of the power grid data information is as follows:
in the formula (11) of the present invention,D 1 (a i ,b j ) For the classification dissimilarity of the power grid data information, q is the number of classification properties, a ip ，a jp Values of i and j in the P-th dimension attribute data; δ(a ip ,a jp ) The value function of (2) is:
the collected samples of the power grid perception data comprise m pieces of data of classified properties and n pieces of data items, and a dissimilarity formula for measuring the classified property data is as follows:
in formula (13), D 2 (a i ,a j ) For dissimilarity of classified property data, n ip ，n jp Representing the number of occurrences of i and j in the data, n being the total number of occurrences of the property data; the numerical attribute data dissimilarity measure formula is:
in formula (14), D 3 (a i ,a j ) Is the dissimilarity of the numerical property data, q is the number of the numerical property data, a im ，a jm The values of i and j in the mth dimension attribute data.
6. The machine vision technology-based power grid fault detection method as claimed in claim 1, wherein: the data recording module realizes classified storage and archiving management of the image characteristic data of the power grid line through a data resource manager Azure so as to improve the extraction speed of the image characteristic data during inquiry.
7. The machine vision technology-based power grid fault detection method as claimed in claim 1, wherein: the output function formula of the non-fault parameters is:
in the case of the formula (15),
for non-fault parameters of the power grid line, n is the number of structural parameters of the power grid line with multiple characteristics, and +.>
For the multi-element characteristic power grid line structure parameter +.>
For the identification of the image characteristics of the network lines,/->
。
8. A power grid fault detection system based on a machine vision technology is characterized in that: a method of grid fault detection using machine vision technology as claimed in any one of claims 1 to 7, wherein the grid fault detection system comprises:
the scanning acquisition module is used for inspecting and detecting the power grid line, the scanning acquisition module acquires the power grid line image through the inspection and detection of the high-power fixed-focus digital camera, the automatic focusing of the high-power fixed-focus digital camera is realized through the automatic focusing control platform, so that the definition of acquired images is ensured, and the scanning acquisition module automatically intercepts the power grid line image at fixed time to realize image acquisition;
The visual correction module is used for filtering the disordered image and correcting the image, judging the properties of the image through the image gray histogram state and correcting the restored image through a gray correction method;
the analysis alarm module is used for analyzing the image data and alarming the identified abnormal data, and comprises a feature extraction unit, a similar filtering unit, an analysis calculation unit and an alarm unit, wherein the output end of the feature extraction unit is connected with the input end of the analysis calculation unit, the output end of the analysis calculation unit is connected with the input end of the similar filtering unit, the output end of the similar filtering unit is connected with the input end of the analysis calculation unit, and the output end of the analysis calculation unit is connected with the input end of the alarm unit;
the data recording module is used for recording, backing up and clearing mass acquired image data and comprises a data storage unit, a data backup unit, an archiving control unit and a data clearing unit, wherein the output end of the data storage unit is connected with the input end of the data backup unit, the output end of the data storage unit is connected with the input end of the archiving control unit, and the output end of the data storage unit is connected with the input end of the data clearing unit;
The detection control module is used for controlling the power grid fault detection process, and the detection control module realizes the decentralized control and the centralized management of the power grid fault detection process through a distributed control system DCS;
the output end of the scanning acquisition module is connected with the input end of the data recording module, the output end of the scanning acquisition module is connected with the input end of the vision correction module, the output end of the vision correction module is connected with the input end of the analysis alarm module, and the output end of the analysis alarm module is connected with the input end of the detection control module.
9. A grid fault detection system according to claim 8, wherein: the similarity filtering unit realizes high-speed filtering of the acacia data through a large data collecting and editing NLPIR-IFCA system, and the large data collecting and editing NLPIR-IFCA system realizes similar classification and filtering of the power grid line image data through machine learning so as to remove redundant data and improve data analysis and calculation speed.
10. A grid fault detection system according to claim 8, wherein:
the feature extraction unit is used for realizing fragment type feature information extraction through circuit control based on a CS5463 chip;
The similar filtering unit is used for filtering similar data information of the power grid through an AMIS-49587 chip circuit;
the analysis and calculation unit also comprises a Clara algorithm model and a logistic regression model;
the alarm unit is an audible and visual information alarm unit.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310421381.XA CN116168019B (en) | 2023-04-19 | 2023-04-19 | Power grid fault detection method and system based on machine vision technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310421381.XA CN116168019B (en) | 2023-04-19 | 2023-04-19 | Power grid fault detection method and system based on machine vision technology |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116168019A true CN116168019A (en) | 2023-05-26 |
| CN116168019B CN116168019B (en) | 2024-02-02 |
Family
ID=86416566
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310421381.XA Active CN116168019B (en) | 2023-04-19 | 2023-04-19 | Power grid fault detection method and system based on machine vision technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116168019B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117579814A (en) * | 2024-01-16 | 2024-02-20 | 苏州宣雄智能科技有限公司 | Quick lens detection method based on focusing detection |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103954897A (en) * | 2014-05-20 | 2014-07-30 | 电子科技大学 | Intelligent power grid high voltage insulation damage monitoring system and method based on ultraviolet imagery |
| US20210048487A1 (en) * | 2019-08-12 | 2021-02-18 | Wuhan University | Power transformer winding fault positioning method based on deep convolutional neural network integrated with visual identification |
| US20210382120A1 (en) * | 2020-06-08 | 2021-12-09 | Wuhan University | Failure diagnosis method for power transformer winding based on gsmallat-nin-cnn network |
| CN115909093A (en) * | 2022-10-21 | 2023-04-04 | 前郭富汇风能有限公司 | Power equipment fault detection method based on unmanned aerial vehicle inspection and infrared image semantic segmentation |
-
2023
- 2023-04-19 CN CN202310421381.XA patent/CN116168019B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103954897A (en) * | 2014-05-20 | 2014-07-30 | 电子科技大学 | Intelligent power grid high voltage insulation damage monitoring system and method based on ultraviolet imagery |
| US20210048487A1 (en) * | 2019-08-12 | 2021-02-18 | Wuhan University | Power transformer winding fault positioning method based on deep convolutional neural network integrated with visual identification |
| US20210382120A1 (en) * | 2020-06-08 | 2021-12-09 | Wuhan University | Failure diagnosis method for power transformer winding based on gsmallat-nin-cnn network |
| CN115909093A (en) * | 2022-10-21 | 2023-04-04 | 前郭富汇风能有限公司 | Power equipment fault detection method based on unmanned aerial vehicle inspection and infrared image semantic segmentation |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117579814A (en) * | 2024-01-16 | 2024-02-20 | 苏州宣雄智能科技有限公司 | Quick lens detection method based on focusing detection |
| CN117579814B (en) * | 2024-01-16 | 2024-03-22 | 苏州宣雄智能科技有限公司 | Quick lens detection method based on focusing detection |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116168019B (en) | 2024-02-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110930357B (en) | In-service steel wire rope surface defect detection method and system based on deep learning | |
| CN108416294B (en) | Fan blade fault intelligent identification method based on deep learning | |
| CN113283344B (en) | Mining conveyor belt deviation detection method based on semantic segmentation network | |
| CN109800697B (en) | Transformer target detection and appearance defect identification method based on VGG-net style migration | |
| CN116168019B (en) | Power grid fault detection method and system based on machine vision technology | |
| CN113406107B (en) | Fan blade defect detection system | |
| CN112697798B (en) | Infrared image-oriented diagnosis method and device for current-induced thermal defects of power transformation equipment | |
| CN116231504A (en) | Remote intelligent inspection method, device and system for booster station | |
| CN116388379A (en) | Remote infrared intelligent inspection method and system for transformer substation | |
| CN116986246A (en) | Intelligent inspection system and method for coal conveying belt | |
| CN111767826A (en) | Timing fixed-point scene abnormity detection method | |
| CN114663672A (en) | Method and system for detecting corrosion of steel member of power transmission line tower | |
| CN117422938B (en) | Dam slope concrete structure anomaly analysis method based on three-dimensional analysis platform | |
| CN114863311A (en) | Automatic tracking method and system for inspection target of transformer substation robot | |
| CN116147632B (en) | Dynamic inspection path planning method and system for newly-added power equipment | |
| CN117114420A (en) | Image recognition-based industrial and trade safety accident risk management and control system and method | |
| CN110569990A (en) | Operation and maintenance system and operation and maintenance method suitable for marine nuclear power platform | |
| CN116309407A (en) | Method for detecting abnormal state of railway contact net bolt | |
| CN104820818A (en) | Fast recognition method for moving object | |
| CN114233581A (en) | Intelligent patrol alarm system for fan engine room | |
| Hwang et al. | Soiling detection for photovoltaic modules based on an intelligent method with image processing | |
| CN113850144A (en) | Video reliability automatic inspection system based on image recognition | |
| Shen et al. | Transmission line safety early warning technology based on multi-source data perception | |
| CN110781758A (en) | Dynamic video monitoring method and device for abnormal pantograph structure | |
| CN110688919A (en) | Intelligent line patrol fault identification method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |