CN114998271A - Transformer substation high tension switchgear secondary bin panel image recognition system - Google Patents
Transformer substation high tension switchgear secondary bin panel image recognition system Download PDFInfo
- Publication number
- CN114998271A CN114998271A CN202210651684.6A CN202210651684A CN114998271A CN 114998271 A CN114998271 A CN 114998271A CN 202210651684 A CN202210651684 A CN 202210651684A CN 114998271 A CN114998271 A CN 114998271A
- Authority
- CN
- China
- Prior art keywords
- template
- picture
- image
- voltage switch
- switch cabinet
- 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
- 238000000034 method Methods 0.000 claims abstract description 20
- 238000012549 training Methods 0.000 claims abstract description 17
- 238000012545 processing Methods 0.000 claims abstract description 11
- 230000002159 abnormal effect Effects 0.000 claims abstract description 6
- 238000003825 pressing Methods 0.000 claims abstract description 6
- 238000005520 cutting process Methods 0.000 claims description 36
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 claims description 22
- 239000011159 matrix material Substances 0.000 claims description 22
- 230000009466 transformation Effects 0.000 claims description 21
- 238000004422 calculation algorithm Methods 0.000 claims description 19
- 238000013507 mapping Methods 0.000 claims description 11
- 238000007635 classification algorithm Methods 0.000 claims description 8
- 238000012360 testing method Methods 0.000 claims description 8
- 238000013145 classification model Methods 0.000 claims description 7
- 238000002372 labelling Methods 0.000 claims description 6
- 238000012163 sequencing technique Methods 0.000 claims description 6
- 238000012795 verification Methods 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 4
- 238000013434 data augmentation Methods 0.000 claims description 4
- 238000009826 distribution Methods 0.000 claims description 4
- 238000003860 storage Methods 0.000 claims description 4
- 238000004519 manufacturing process Methods 0.000 claims description 3
- 230000001131 transforming effect Effects 0.000 claims description 3
- 238000007689 inspection Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 3
- 238000013136 deep learning model Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Images
Classifications
-
- 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
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/02—Affine transformations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
-
- 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
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
-
- 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
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Quality & Reliability (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of image recognition of a high-voltage switch cabinet of a transformer substation, in particular to an image recognition system for a secondary cabin panel of the high-voltage switch cabinet of the transformer substation, which comprises the steps of firstly, posting an AprilTag label with unique ID on each secondary cabin panel of the high-voltage switch cabinet, collecting pictures of the secondary cabin panel, and selecting 1 clear picture as a processing template picture for each high-voltage switch cabinet; secondly, dividing each template drawing into an indicator light area, a pressing plate switch area and a meter area, and marking each part on the cut area image by using a part rectangular frame; then, AprilTag labels in the template picture and the non-template picture are detected, and an indicator light, a pressure plate switch data set and a digital table data set are obtained; training the ResNet50 classification network again, obtaining the states of the indicator light and the pressure plate switch, designing an automatic reading method based on template matching, and obtaining the reading of the digital meter; and finally, logically judging whether the components are abnormal or not according to the standard threshold value in the normal running state preset by each component, and giving an alarm in time.
Description
Technical Field
The invention relates to the technical field of image recognition of a high-voltage switch cabinet of a transformer substation, in particular to an image recognition system for a secondary bin panel of the high-voltage switch cabinet of the transformer substation.
Background
For daily inspection of a high-voltage switch cabinet of a transformer substation, a field worker records the state of each component by looking over a secondary bin panel of the high-voltage switch cabinet, so that the state of switch equipment in a primary bin of the high-voltage switch cabinet is obtained. However, at present, inspection of most transformer substation high-voltage switch cabinets mainly depends on manual work, the states of components on each high-voltage switch cabinet are recorded through regular manual inspection, and once a switch equipment fault cannot be found in time, serious loss can be caused. And because contain the part kind and numerous in quantity on the high tension switchgear secondary storehouse panel, need discern pilot lamp state, clamp plate on-off state and digital table reading simultaneously, be difficult to directly realize through single degree of deep learning model, and degree of deep learning model just can probably obtain good recognition effect on a large amount of mark data basis training. Meanwhile, the image recognition algorithm based on deep learning has the problems of component missing detection and non-component false detection as components in the actual application, so that the recognition result cannot correspond to the components in the actual position. Therefore, the reliable application of the image recognition of the secondary bin panel of the high-voltage switch cabinet cannot be realized only by adopting an algorithm based on deep learning, and an image recognition system which is general and high in recognition accuracy rate needs to be designed urgently, so that different types of components can be recognized simultaneously, and the high-voltage switch cabinet has popularization and application values.
Disclosure of Invention
The invention aims to solve the defects in the prior art, and simultaneously solves the problems that a high-voltage switch cabinet depends on manual routing inspection and inspection, and hands of workers are liberated, so that the secondary bin panel image identification system of the high-voltage switch cabinet of the transformer substation is provided.
In order to achieve the purpose, the invention adopts the following technical scheme: a transformer substation high tension switchgear secondary bin panel image recognition system comprises the following steps:
1) pasting an AprilTag label with unique ID on a horizontal axis of a secondary bin panel of each high-voltage switch cabinet, and recording a list of one-to-one correspondence of the names and the label IDs of the high-voltage switch cabinets;
2) collecting sample pictures of the high-voltage switch cabinet to form a sample library, wherein each sample picture comprises a secondary bin panel of the high-voltage switch cabinet of the transformer substation and an Apriltag;
3) selecting 1 clear picture as a template picture for each high-voltage switch cabinet from a sample library;
4) detecting all label IDs and four corner point pixel coordinates of labels in the template picture by calling an Apriltag algorithm, and calculating the difference value of the horizontal pixel coordinate of the center point of the label and the horizontal central axis pixel coordinate of the sample picture, wherein the minimum difference value is the corresponding label on the current high-voltage switch cabinet;
5) searching the name and ID list in the step 1) according to the corresponding label ID acquired in the step 4), acquiring the name of the high-voltage switch cabinet acquired by the template picture, and storing the template picture name, the label ID and the four corner pixel coordinates into a database;
6) dividing each template picture into an indicator light area, a pressing plate switch area and a digital meter area according to the types and distribution characteristics of components on a secondary bin panel of the high-voltage switch cabinet, cutting the area pictures and naming;
7) marking components in the area image, naming the names of the components, cutting the component images one by one according to the component marking frame, and setting the standard state of each component;
8) sequentially reading non-template pictures from the sample library, and repeating the step 4) to obtain a label ID and four corner pixel coordinates in each non-template picture;
9) calculating an affine transformation matrix between the template picture and the non-template picture according to the pixel coordinates of the label corner points in the step 8) and by combining the pixel coordinates of the label corner points in the template picture in the step 4);
10) mapping the area marking frame in the template picture to the non-template picture according to the affine transformation matrix, cutting the area picture, naming and storing, and mapping all the component marking frames in the template picture to the non-template picture to realize the positioning of the non-template picture components;
11) cutting each part image of the non-template picture in the sample library according to the positioned part marking frame, and naming each part image to obtain a part image data set;
12) dividing the data set into an indicator light data set, a pressure plate switch data set and a digital table data set according to the names of the component images;
13) based on the indicator light and pressure plate switch data set obtained in the step 12), data augmentation is carried out, the data set is divided into a training set, a verification set and a test set, the data set and a ResNet50 classification model are imported into an algorithm server, an indicator light and pressure plate switch classification algorithm model based on ResNet50 is trained, and the 'on' or 'off' state of the indicator light and the 'on' or 'on' state of the pressure plate switch are obtained;
14) designing an automatic reading algorithm based on template matching based on the digital table data set obtained in the step 12), making 0-9 LED digital binary template pictures, calculating the pixel matching degrees of single digital binary images of the LED digital table and the digital binary template pictures in the new pictures one by one, sequencing the matching degrees to obtain single digital readings with the highest matching degree, sequencing the digital positions according to horizontal coordinates, and calculating the actual readings of the digital table according to the digits;
15) inputting a newly acquired high-voltage switch cabinet picture into a high-voltage switch cabinet image identification system, repeating the step 4) and the step 5), acquiring four corner pixel coordinates of a label in the picture, and repeating the step 6) to acquire a region image;
16) calculating affine transformation matrixes of a newly acquired picture and a template picture, mapping all component marking frames in the template picture onto the newly acquired picture, cutting and naming the component images, calling a classification algorithm model based on ResNet50 or an automatic reading algorithm based on template matching according to component types in the component image names, and acquiring an identification result;
17) and reading the preset standard states of each indicator light, the pressure plate switch and the digital meter, logically judging whether the state of the component is abnormal or not, and timely giving an alarm.
As a further description of the above technical solution:
the tag ID of step 1) is specifically a natural number other than 0.
As a further description of the above technical solution:
the types of the parts on the secondary bin panel in the step 6) mainly comprise an LED indicator light, a pressure plate SWITCH and an LED digital METER which are sequentially named as LED, SWITCH and METER; the areas divided in the step 6) are divided by adopting an area rectangular frame, each secondary bin panel image of the high-voltage switch cabinet is divided into 3 areas at most, and the areas are sequentially named as 'P-PSx', wherein 'P' represents the high-voltage switch cabinet for short, PSx represents the areas for short, and the values of 'x' are 1, 2 and 3 and are not repeated; the region rectangular frame is required to be framed to all parts to be identified according to part types, the region name is used as a key value, pixel coordinates of upper left corner points and lower right corner points of the region rectangular frame are used as values, and the pixel coordinates are stored in a Redis database; cutting the image of the cutting region in the step 6) according to a region rectangular frame, naming the region image as 'P-PSx-time.jpg' according to the region name after cutting, wherein the 'time' is the current millisecond number of the computer, and storing the region image in a local server; the local server and the operating system are Linux Ubuntu.
As a further description of the above technical solution:
the marking component in the step 7) is marked by a component rectangular frame, each component is named, the standard state of each component in the normal operation state of the high-voltage switch cabinet is set, the storage component marks the frame corner point coordinate and the component name into a database, the component is named by combining the region name, the component type and the serial number, the specific naming rules of the three components are P-PSx-LEDy ', "P-PSx-SWITCH'," P-PSx-METERy "and" y "represent non-zero natural numbers, the y value is unique in a single region image, the storage component marks the frame corner point coordinate and the component name, the component name is a key value, and the component marking frame and the component standard state in the region image are value and are stored into a Redis database; cutting the part images of the step 7), naming each cut part image, naming the part images as ". P-PSx-LEDy-time. jpg" or ". P-PSx-SWITCHy-time. jpg" or ". P-PSx-measure-time. jpg" according to the part names, and saving the part images in the local server.
As a further description of the above technical solution:
the affine transformation matrix calculated in the step 9) specifically includes:
wherein M is 2×3 Is an affine transformation matrix and represents the homogeneous coordinate (x) of pixels in a template picture through linear change 0 ,y 0 And 1) transforming to homogeneous coordinates (x) of pixels at corresponding positions in the non-template picture 1 ,y 1 1), obtaining M through solving corresponding three pairs of pixel coordinates in the template picture and the non-template picture;
the three pairs of pixel coordinates are selected as the pixel coordinates of the upper left corner point, the upper right corner point and the lower right corner point of the AprilTag label, and the three pairs of pixel coordinates are substituted into the formula, so that the affine transformation matrix can be solved.
As a further description of the above technical solution:
the non-template drawing part positioning in the step 10) specifically comprises the following steps:
a1: reading all data with key of 'P' from a Redis database, traversing and comparing the tag ID value in the value with the tag ID value detected in the current non-template picture, and obtaining the name of the high-voltage switch cabinet corresponding to the current non-template picture when the IDs are equal;
a2: reading all data with keys of 'P-PSx' from a Redis database, wherein 'P' is the name of a high-voltage switch cabinet corresponding to the current non-template picture, substituting pixel coordinates of upper left corner points and lower right corner points of a rectangular frame of a template picture region into the following formula to obtain region rectangular frame coordinates of the non-template picture, and cutting the non-template picture according to the transformed region rectangular frame;
a3: reading all data with key of 'P-PSx-LEDy', 'P-PSx-SWITCH' and 'P-PSx-Meter' from a Redis database, substituting the pixel coordinates of the upper left corner point and the lower right corner point of the rectangular frame of the template picture into the formula to obtain the coordinates of the rectangular frame of the non-template picture, and positioning the part of the non-template picture area image according to the transformed rectangular frame of the part.
As a further description of the above technical solution:
the step 14) of manufacturing the 0-9 LED digital binarization template drawing specifically comprises the following steps:
b1: collecting a plurality of digital table component images containing 0-9 digits, and performing image gray scale processing and binarization processing, wherein a binarization threshold value is set to be 150, namely the gray scale value of a pixel point is larger than 150, the pixel point is set to be 255, otherwise, the pixel point is set to be 0;
b2: searching the contour of the binary image by a cv2.findContours method in OpenCV to obtain a contour point [ (x) 2 ,y 2 ),...,(x n ,y n )]Traverse comparison x 1 ～x n And y 1 ～y n Get the pixel coordinate (x) at the top left corner of the smallest rectangle surrounding the digital LED min, y min ) And the lower right corner pixel coordinate (x) max ,y max )；
B3: and (3) cutting the digital table component image according to four parameters (x, y, w, h), wherein the parameter values are as follows:
b4: resetting the size of the digital binarization template image, uniformly setting the size as (42,51), storing the digital template image, and naming the digital binarization template image as num.jpg according to the actual meaning of the number, wherein the num value range is a natural number of 0-9.
As a further description of the above technical solution:
calculating the pixel matching degree in the step 14), sequentially reading a single digital binarization template image from the number of 0-9 according to the picture name, and calculating the matching degree of the current single digital binarization image and the template image:
wherein, b 0 A binarized pixel matrix for the current single digital image, b 1 Is the binary pixel matrix of the read-in template picture.
As a further description of the above technical solution:
and (3) sorting the matching degrees in the step 14), obtaining the maximum matching degree by adopting a bubble sorting method, wherein the number corresponding to the binary template graph is the reading result of the current single number, and is inserted into the result list.
As a further description of the above technical solution:
calculating the actual reading of the digital table in the step 14), and calculating the reading according to the meaning of the size of the digit:
where l is the length of the result list, i.e. the number of detected digits, and result (k) is the kth element of the result list.
The invention has the following beneficial effects:
compared with the prior art, the image recognition system for the secondary bin panel of the high-voltage switch cabinet of the transformer substation has the advantages that the method is high in recognition accuracy and universal for other types of high-voltage switch cabinets, only one ResNet50 classification model consumes GPU computing power, and therefore computing power consumption is low; according to the method, for each secondary bin panel of the high-voltage switch cabinet, only 5-10 images are required to be collected, algorithm design and training can be achieved, and the problem of dependence on a large amount of data is solved; according to the method, each high-voltage switch cabinet secondary bin panel image only needs to mark all parts for 1 time, so that the parts in each image can be accurately positioned, and the problem of complicated data marking is solved; according to the characteristics of the secondary bin panel of the high-voltage switch cabinet of the transformer substation, the method selects a mode of combining traditional image processing and a deep learning model, simultaneously identifies the problems of the indicator lamp, the pressure plate switch and the digital meter by the whole image, simplifies the problems of classification of the indicator lamp and the pressure plate switch and reading of the digital meter, and overcomes the defect of a single image processing mode.
Drawings
FIG. 1 is a drawing of a selected template for a 12-sided high voltage switchgear of the present invention;
FIG. 2 is a diagram of the present invention for dividing image areas of 4P and 6P high voltage switch cabinets;
FIG. 3 is a graph of the result of the invention cutting the 14P high-voltage switch cabinet image area;
FIG. 4 is a diagram of the result of labeling components in a 14P high voltage switchgear area image according to the present invention;
FIG. 5 is a diagram of an indicator light data set and classification obtained by the present invention;
FIG. 6 is a graph of platen switch data sets and classifications obtained by the present invention;
FIG. 7 is a loss curve of the model training of the ResNet50 classification algorithm of the present invention;
FIG. 8 is a block diagram of the digital meter auto-reading algorithm process flow of the present invention.
Detailed Description
Referring to fig. 1-8, the invention provides a transformer substation high tension switchgear secondary bin panel image recognition system, which comprises the following steps:
1) pasting an AprilTag label with unique ID on a horizontal axis of a secondary bin panel of each high-voltage switch cabinet, and recording a list of one-to-one correspondence of the names and the label IDs of the high-voltage switch cabinets;
2) collecting sample pictures of the high-voltage switch cabinet to form a sample library, wherein each sample picture comprises a secondary bin panel of the high-voltage switch cabinet of the transformer substation and an Apriltag;
3) selecting 1 clear picture as a template picture for each high-voltage switch cabinet from a sample library;
4) detecting all label IDs and four corner point pixel coordinates of labels in the template picture by calling an Apriltag algorithm, and calculating the difference value of the horizontal pixel coordinate of the center point of the label and the horizontal central axis pixel coordinate of the sample picture, wherein the minimum difference value is the corresponding label on the current high-voltage switch cabinet;
5) searching the name and ID list in the step 1) according to the corresponding label ID obtained in the step 4), obtaining the name of the high-voltage switch cabinet collected by the template picture, and storing the template picture name, the label ID and the four corner pixel coordinates into a database;
6) dividing each template picture into an indicator light area, a pressing plate switch area and a digital meter area according to the types and distribution characteristics of components on a secondary bin panel of the high-voltage switch cabinet, cutting the area pictures and naming;
7) marking components in the area image, naming the names of the components, cutting the component images one by one according to the component marking frame, and setting the standard state of each component;
8) sequentially reading non-template pictures from the sample library, and repeating the step 4) to obtain a label ID and four corner pixel coordinates in each non-template picture;
9) calculating an affine transformation matrix between the template picture and the non-template picture according to the pixel coordinates of the label corner points in the step 8) and by combining the pixel coordinates of the label corner points in the template picture in the step 4);
10) mapping the area marking frame in the template picture to the non-template picture according to the affine transformation matrix, cutting the area picture, naming and storing, and mapping all the component marking frames in the template picture to the non-template picture to realize the positioning of the non-template picture components;
11) cutting each part image of the non-template picture in the sample library according to the positioned part marking frame, and naming each part image to obtain a part image data set;
12) dividing the data set into an indicator light data set, a pressure plate switch data set and a digital table data set according to the names of the component images; the method specifically comprises the following steps: the data of the indicating lamp can be divided into two categories, namely, the on and off of the indicating lamp are respectively recorded as LED _ on and LED _ off, and a data set of a digital meter, namely an LED digital meter displays different readings;
13) based on the indicator light and pressure plate switch data set obtained in the step 12), data augmentation is carried out, the data set is divided into a training set, a verification set and a test set, the data set and a ResNet50 classification model are imported into an algorithm server, an indicator light and pressure plate switch classification algorithm model based on ResNet50 is trained, and the 'on' or 'off' state of the indicator light and the 'on' or 'on' state of the pressure plate switch are obtained; the method specifically comprises the following steps: the indicator light and the pressure plate SWITCH data set are a mixture of the indicator light data set and the pressure plate SWITCH data set and are divided into four component states of LED _ on, LED _ off, SWITCH _ on and SWITCH _ off;
14) designing an automatic reading algorithm based on template matching based on the digital table data set obtained in the step 12), making 0-9 LED digital binary template pictures, calculating the pixel matching degrees of single digital binary images of the LED digital tables and the digital binary template pictures in a new picture one by one, sequencing the matching degrees to obtain single digital readings with the highest matching degree, sequencing digital positions according to horizontal coordinates, and calculating actual readings of the digital tables according to digits;
15) inputting a newly acquired high-voltage switch cabinet picture into a high-voltage switch cabinet image identification system, repeating the step 4) and the step 5), acquiring pixel coordinates of four corner points of a label in the picture, and repeating the step 6) to acquire a region image;
16) calculating affine transformation matrixes of a newly acquired picture and a template picture, mapping all component marking frames in the template picture onto the newly acquired picture, cutting and naming the component images, calling a classification algorithm model based on ResNet50 or an automatic reading algorithm based on template matching according to component types in the component image names, and acquiring an identification result;
17) reading the preset standard states of each indicator light, the pressure plate switch and the digital meter, logically judging whether the state of the component is abnormal or not, and giving an alarm in time; the method specifically comprises the following steps:
for the indicator light and the pressure plate switch, when the standard state and the identification state are the same, the indicator light and the pressure plate switch are judged to be normal, and when the standard state and the identification state are different, the indicator light and the pressure plate switch are judged to be abnormal;
and for the digital table, judging the digital table to be normal when the identification reading is within the preset threshold range, and judging the digital table to be abnormal when the identification reading is larger than or smaller than the preset threshold range.
As a further embodiment of the above technical solution:
the tag ID of step 1) is specifically a natural number other than 0.
As a further embodiment of the above technical solution:
the types of the parts on the secondary bin panel in the step 6) mainly comprise an LED indicator light, a pressure plate SWITCH and an LED digital METER which are sequentially named as LED, SWITCH and METER; dividing the areas in the step 6) by adopting an area rectangular frame, dividing each high-voltage switch cabinet secondary bin panel image into 3 areas at most, and sequentially naming the areas as 'P-PSx', wherein 'P' represents a high-voltage switch cabinet for short, PSx represents an area for short, and 'x' takes values of 1, 2 and 3 without repetition; the region rectangular frame is required to be framed to all parts to be identified according to the part types, the region name is used as a key value, pixel coordinates of upper left corner points and lower right corner points of the region rectangular frame are used as values, and the pixel coordinates are stored in a Redis database; cutting the image of the region in the step 6), cutting the image according to a region rectangular frame, naming the image of the region as P-PSx-time.jpg according to the region name after cutting, wherein the time is the current millisecond of the computer, and storing the image of the region in a local server; and the local server and the operating system are Linux Ubuntu.
As a further embodiment of the above technical solution:
marking components in the step 7), namely marking each component by using a component rectangular frame, naming each component, setting a standard state of each component in a normal operation state of the high-voltage switch cabinet, storing component marking frame corner point coordinates and component names into a database, naming the components, combining region names, component types and numbers, wherein specific naming rules of the three components are 'P-PSx-LEDy', 'P-PSx-SWITCH', 'P-PSx-METERy' and 'y' represent non-zero natural numbers, in a single region image, y takes a unique value, storing the component marking frame corner point coordinates and the component names, and storing the component marking frame and the component standard state in the region image as value into a Redis database; cutting the part images of the step 7), naming each cut part image, naming the part images as ". P-PSx-LEDy-time. jpg" or ". P-PSx-SWITCHy-time. jpg" or ". P-PSx-measure-time. jpg" according to the part names, and saving the part images in the local server.
As a further embodiment of the above technical solution:
calculating an affine transformation matrix in the step 9), specifically:
wherein M is 2×3 Is an affine transformation matrix and represents the homogeneous coordinate (x) of pixels in a template picture through linear change 0 ,y 0 And 1) transforming to homogeneous coordinates (x) of pixels at corresponding positions in the non-template picture 1 ,y 1 1), obtaining M through solving corresponding three pairs of pixel coordinates in the template picture and the non-template picture;
and selecting three pairs of pixel coordinates, namely the pixel coordinates of the upper left corner point, the upper right corner point and the lower right corner point of the AprilTag label, and substituting the three pixel coordinates into the above formula to solve the affine transformation matrix.
As a further embodiment of the above technical solution:
step 10), positioning the non-template drawing part, specifically:
a1: reading all data with keys of 'P' from a Redis database, traversing and comparing the label ID value in the value with the label ID value detected in the current non-template picture, and obtaining the name of the high-voltage switch cabinet corresponding to the current non-template picture when the IDs are equal;
a2: reading all data with keys of 'P-PSx' from a Redis database, wherein 'P' is the name of a high-voltage switch cabinet corresponding to the current non-template picture, substituting pixel coordinates of upper left corner points and lower right corner points of a rectangular frame of a template picture region into the following formula to obtain region rectangular frame coordinates of the non-template picture, and cutting the non-template picture according to the transformed region rectangular frame;
a3: reading all data with key of 'P-PSx-LEDy', 'P-PSx-SWITCH' and 'P-PSx-Meter' from a Redis database, substituting the pixel coordinates of the upper left corner point and the lower right corner point of the rectangular frame of the template picture into the formula to obtain the coordinates of the rectangular frame of the non-template picture, and positioning the part of the non-template picture area image according to the transformed rectangular frame of the part.
As a further embodiment of the above technical solution:
step 14) of manufacturing 0-9 LED digital binarization template pictures, specifically:
b1: collecting a plurality of digital table component images containing 0-9 digits, and performing image gray scale processing and binarization processing, wherein a binarization threshold value is set to be 150, namely the gray scale value of a pixel point is larger than 150, the pixel point is set to be 255, otherwise, the pixel point is set to be 0;
b2: searching the contour of the binary image by a cv2.findContours method in OpenCV to obtain a contour point [ (x) 2 ,y 2 ),...,(x n ,y n )]Traverse comparison x 1 ～x n And y 1 ～y n Get the pixel coordinate (x) at the top left corner of the smallest rectangle surrounding the digital LED min, y min ) And the lower right corner pixel coordinate (x) max ,y max )；
B3: and (3) cutting the digital table component image according to four parameters (x, y, w, h), wherein the parameter values are as follows:
b4: resetting the size of the digital binarization template image, uniformly setting the size as (42,51), storing the digital template image, and naming the digital binarization template image as num.jpg according to the actual meaning of the number, wherein the num value range is a natural number of 0-9.
As a further embodiment of the above technical solution:
calculating the pixel matching degree in the step 14), sequentially reading a single digital binarization template image from the number of 0-9 according to the picture name, and calculating the matching degree of the current single digital binarization image and the template image:
wherein, b 0 A binarized pixel matrix for the current single digital image, b 1 Is the binary pixel matrix of the read-in template picture.
As a further embodiment of the above technical solution:
and step 14), sorting the matching degrees, namely obtaining the maximum value of the matching degrees by adopting a bubble sorting method, wherein the number corresponding to the binary template graph is the reading result of the current single number and is inserted into a result list.
As a further embodiment of the above technical solution:
calculating the actual reading of the digital table in the step 14), and calculating the reading according to the meaning of the size of the digit:
where l is the length of the result list, i.e. the number of detected digits, and result (k) is the kth element of the result list.
The working principle is as follows:
patrol and examine through unmanned aerial vehicle and gather high tension switchgear secondary storehouse panel picture, 12 high tension switchgear in total, in the task is patrolled and examined to the single, unmanned aerial vehicle respectively gathers 1 pictures to every high tension switchgear, carries out 10 times altogether at different moments and patrols and examines the task, and every high tension switchgear gathers 10 images promptly, founds the sample storehouse. The pixel of a single picture is 3000 multiplied by 4000, and 1 template picture is selected for each high-voltage switch cabinet.
The deviation of the pictures collected in different batches is large, the pixel deviation of the other 9 collected pictures and the template pictures is counted, and the average deviation reaches 132.46 pixel points, so that component positioning is needed.
According to the type and distribution characteristics of the high-voltage switch cabinet components, the secondary bin panel pictures of typical 4P and 6P high-voltage switch cabinets are divided into regions. All 12 template pictures were divided into regions as shown in table 1.
High-voltage switch cabinet with surface 112 surface full scale and area division
Dividing the areas according to table 1, performing area cutting on the 14P high-voltage switch cabinet secondary bin panel template picture, and cutting out a 14P-PS1 indicator light area, a 14P-PS2 indicator light area and a 14P-PS3 pressing plate switch area.
The region images 14P-PS1, 14P-PS2, 14P-PS3 of the 14P template picture are subjected to component labeling, resulting in a component type, a component labeling frame, and a component standard state.
Reading in a non-template picture of a 14P high-voltage switch cabinet, detecting a label corner point C1 in the non-template picture, reading a label corner point C0 in a template picture stored in a Redis database:
selecting the first three diagonal point pixel coordinates, and calculating an affine change matrix M:
and mapping the region rectangular frame in the 14P template picture into the 14P non-template picture through affine transformation, and cutting the region image.
And mapping the rectangular frame of the part in the region image of the 14P template picture to the region image of the 14P non-template picture through affine transformation, so as to realize accurate part positioning.
And repeating the process for the pictures of the secondary bin panels of the rest high-voltage switch cabinets in the sample library, and cutting out part images according to the affine transformation rear part labeling frame to obtain an indicator light data set, a pressure plate switch data set and a digital table data set.
After data augmentation processing such as rotation and overturning, the number of various data of the data sets of the indicator light and the pressing plate switch is obtained, as shown in table 2, and images of all kinds of components in the data sets are divided into a training set, a verification set and a test set.
TABLE 2 indicator light and pressure plate switch data set
An LED indicator lamp and pressure plate switch classification model based on ResNet50 is trained in an algorithm server, the operating system of the server is 64-bit Ubuntu16.04, the basic configuration is shown in Table 3, the classification model is realized based on a Pytroch 1.6.0 deep learning framework, and Python2.7 is used in a programming language.
TABLE 3 Algorithm Server configuration parameters
In the ResNet50 model training process, the ResNet50 model pre-trained on ImageNet is first used to initialize the network weights and set the main training parameters of the network as shown in table 4.
TABLE 4 network training parameters
| Parameter name | Parameter value |
| Number of iterations/ |
120 |
| Initial learning rate/learning | 1.0 |
| Weight decay factor/Weightdecay | 0.0001 |
| Momentum attenuation factor/Momentum | 0.9 |
| Batch size/ |
32 |
Training loss (training loss) and verification loss (validation loss) curves generated by network training at the end of training. As can be seen from the figure, the training loss and the verification loss of the network are synchronously converged, and the ResNet50 classification model has a good training effect.
And (3) performing classification test on the part images in the test set by using the trained classification algorithm model, wherein the statistics of test results are shown in tables 4-8, the average identification accuracy rate reaches 99%, and the classification method has a very good classification effect.
TABLE 4-8 Algorithm test results
For the digital table data set, 10 sheets in total were counted. According to the method provided by the invention, 0-9 single digital template pictures are manufactured. And then automatically reading the digital table component image through the steps of gray processing, binaryzation, contour searching and the like. The automatic reading accuracy of the digital meter reaches 100 percent when the digital meter is tested on a data set of the digital meter.
Therefore, the method can realize automatic identification of the secondary bin panel image of the high-voltage switch cabinet of the transformer substation, has higher accuracy, has the advantages of extremely small data volume of image acquisition, high universality and the like, and can be applied to an intelligent inspection system of the transformer substation.
Finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that modifications may be made to the embodiments or portions thereof without departing from the spirit and scope of the invention.
Claims (10)
1. A secondary bin panel image recognition system of a high-voltage switch cabinet of a transformer substation comprises the following steps:
1) pasting an AprilTag label with unique ID on a horizontal axis of a secondary bin panel of each high-voltage switch cabinet, and recording a list of one-to-one correspondence of the names and the label IDs of the high-voltage switch cabinets;
2) collecting sample pictures of the high-voltage switch cabinet to form a sample library, wherein each sample picture comprises a secondary bin panel of the high-voltage switch cabinet of the transformer substation and an Apriltag;
3) selecting 1 clear picture as a template picture for each high-voltage switch cabinet from a sample library;
4) detecting all label IDs and four corner point pixel coordinates of labels in the template picture by calling an Apriltag algorithm, and calculating the difference value of the horizontal pixel coordinate of the center point of the label and the horizontal central axis pixel coordinate of the sample picture, wherein the minimum difference value is the corresponding label on the current high-voltage switch cabinet;
5) searching the name and ID list in the step 1) according to the corresponding label ID acquired in the step 4), acquiring the name of the high-voltage switch cabinet acquired by the template picture, and storing the template picture name, the label ID and the four corner pixel coordinates into a database;
6) dividing each template picture into an indicator light area, a pressing plate switch area and a digital meter area according to the types and distribution characteristics of components on a secondary bin panel of the high-voltage switch cabinet, cutting the area pictures and naming;
7) marking components in the area image, naming the names of the components, cutting the component images one by one according to the component marking frame, and setting the standard state of each component;
8) sequentially reading non-template pictures from the sample library, and repeating the step 4) to obtain a label ID and four corner pixel coordinates in each non-template picture;
9) calculating an affine transformation matrix between the template picture and the non-template picture according to the pixel coordinates of the label corner points in the step 8) and by combining the pixel coordinates of the label corner points in the template picture in the step 4);
10) mapping the area marking frames in the template picture to the non-template picture according to the affine transformation matrix, cutting the area picture, naming and storing, and mapping all the component marking frames in the template picture to the non-template picture to realize the positioning of the non-template picture components;
11) cutting each part image of the non-template picture in the sample library according to the positioned part marking frame, and naming each part image to obtain a part image data set;
12) dividing the data set into an indicator light data set, a pressure plate switch data set and a digital table data set according to the names of the component images;
13) based on the indicator light and pressure plate switch data set obtained in the step 12), data augmentation is carried out, the data set is divided into a training set, a verification set and a test set, the data set and a ResNet50 classification model are imported into an algorithm server, an indicator light and pressure plate switch classification algorithm model based on ResNet50 is trained, and the 'on' or 'off' state of the indicator light and the 'on' or 'on' state of the pressure plate switch are obtained;
14) designing an automatic reading algorithm based on template matching based on the digital table data set obtained in the step 12), making 0-9 LED digital binary template pictures, calculating the pixel matching degrees of single digital binary images of the LED digital tables and the digital binary template pictures in a new picture one by one, sequencing the matching degrees to obtain single digital readings with the highest matching degree, sequencing digital positions according to horizontal coordinates, and calculating actual readings of the digital tables according to digits;
15) inputting a newly acquired high-voltage switch cabinet picture into a high-voltage switch cabinet image identification system, repeating the step 4) and the step 5), acquiring pixel coordinates of four corner points of a label in the picture, and repeating the step 6) to acquire a region image;
16) calculating affine transformation matrixes of a newly acquired picture and a template picture, mapping all component marking frames in the template picture onto the newly acquired picture, cutting and naming the component images, calling a classification algorithm model based on ResNet50 or an automatic reading algorithm based on template matching according to component types in the component image names, and acquiring an identification result;
17) and reading the preset standard states of each indicator light, the pressure plate switch and the digital meter, logically judging whether the state of the component is abnormal or not, and timely giving an alarm.
2. The transformer substation high-voltage switch cabinet secondary bin panel image recognition system according to claim 1, characterized in that: the tag ID of step 1) is specifically a natural number other than 0.
3. The transformer substation high-voltage switch cabinet secondary bin panel image recognition system according to claim 1, characterized in that: the types of the parts on the secondary bin panel in the step 6) mainly comprise an LED indicator light, a pressure plate SWITCH and an LED digital METER which are sequentially named as LED, SWITCH and METER; the areas divided in the step 6) are divided by adopting area rectangular frames, each secondary bin panel image of the high-voltage switch cabinet is divided into 3 areas at most, and the areas are named as ' P-PSx ', wherein ' P ' represents the high-voltage switch cabinet for short, PSx represents the areas for short, x ' takes values of 1, 2 and 3, and the values are not repeated; the region rectangular frame is required to be framed to all parts to be identified according to part types, the region name is used as a key value, pixel coordinates of upper left corner points and lower right corner points of the region rectangular frame are used as values, and the pixel coordinates are stored in a Redis database; cutting the image of the cut region in the step 6) according to a region rectangular frame, naming the image of the cut region as 'P-PSx-time.jpg' according to the region name, wherein 'time' is the current millisecond number of the computer, and storing the image of the cut region in a local server; the local server and the operating system are Linux Ubuntu.
4. The transformer substation high-voltage switch cabinet secondary bin panel image recognition system according to claim 1, characterized in that: the labeling component in the step 7) is labeled by a component rectangular frame, each component is named, the standard state of each component in the normal operation state of the high-voltage switch cabinet is set, the storage component is labeled with a frame corner point coordinate and a component name and is stored in a database, the component is named by combining an area name, a component type and a number, the specific naming rules of the three components are 'P-PSx-LEDy', 'P-PSx-SWITCH', 'P-PSx-METERy' and 'y' represent non-zero natural numbers, the y value is unique in a single area image, the storage component is labeled with the frame corner point coordinate and the component name, the component name is a key value, and the component labeling frame and the component standard state in the area image are value and are stored in a Redis database; cutting the part images of the step 7), naming each cut part image, naming the part images as ". P-PSx-LEDy-time. jpg" or ". P-PSx-SWITCHy-time. jpg" or ". P-PSx-measure-time. jpg" according to the part names, and saving the part images in the local server.
5. The transformer substation high-voltage switch cabinet secondary bin panel image recognition system according to claim 1, characterized in that: the affine transformation matrix calculated in the step 9) specifically includes:
wherein M is 2×3 Is an affine transformation matrix and represents the homogeneous coordinate (x) of pixels in a template picture through linear change 0 ,y 0 And 1) transforming to homogeneous coordinates (x) of pixels at corresponding positions in the non-template picture 1 ,y 1 1), obtaining M through solving corresponding three pairs of pixel coordinates in the template picture and the non-template picture;
and selecting three pairs of pixel coordinates, namely the pixel coordinates of the upper left corner point, the upper right corner point and the lower right corner point of the AprilTag label, and substituting the three pairs of pixel coordinates into the above formula to solve the affine transformation matrix.
6. The transformer substation high-voltage switch cabinet secondary bin panel image recognition system according to claim 1, characterized in that: the non-template drawing part positioning in the step 10) specifically comprises the following steps:
a1: reading all data with keys of 'P' from a Redis database, traversing and comparing the label ID value in the value with the label ID value detected in the current non-template picture, and obtaining the name of the high-voltage switch cabinet corresponding to the current non-template picture when the IDs are equal;
a2: reading all data with keys of 'P-PSx' from a Redis database, wherein 'P' is the name of a high-voltage switch cabinet corresponding to the current non-template picture, substituting pixel coordinates of upper left corner points and lower right corner points of a rectangular frame of a template picture region into the following formula to obtain region rectangular frame coordinates of the non-template picture, and cutting the non-template picture according to the transformed region rectangular frame;
a3: reading all data with key of 'P-PSx-LEDy', 'P-PSx-SWITCH' and 'P-PSx-Meter' from a Redis database, substituting the pixel coordinates of the upper left corner point and the lower right corner point of the rectangular frame of the template picture into the formula to obtain the coordinates of the rectangular frame of the non-template picture, and positioning the part of the non-template picture area image according to the transformed rectangular frame of the part.
7. The transformer substation high-voltage switch cabinet secondary bin panel image recognition system according to claim 1, characterized in that: the step 14) of manufacturing the 0-9 LED digital binarization template drawing specifically comprises the following steps:
b1: collecting a plurality of digital table component images containing 0-9 digits, and performing image gray scale processing and binarization processing, wherein a binarization threshold value is set to be 150, namely the gray scale value of a pixel point is larger than 150, the pixel point is set to be 255, otherwise, the pixel point is set to be 0;
b2: searching the contour of the binary image by a cv2.findContours method in OpenCV to obtain a contour point [ (x) 2 ,y 2 ),...,(x n ,y n )]Traverse comparison x 1 ～x n And y 1 ～y n Get the pixel coordinate (x) at the top left corner of the smallest rectangle surrounding the digital LED min, y min ) And the lower right corner pixel coordinate (x) max ,y max )；
B3: and (3) cutting the digital table part image according to four parameters (x, y, w, h), wherein the parameter values are as follows:
b4: resetting the size of the digital binarization template image, uniformly setting the size as (42,51), storing the digital template image, and naming the digital binarization template image as num.jpg according to the actual meaning of the number, wherein the num value range is a natural number of 0-9.
8. The transformer substation high-voltage switch cabinet secondary bin panel image recognition system according to claim 1, characterized in that: calculating the pixel matching degree in the step 14), sequentially reading a single digital binarization template image from the number of 0-9 according to the picture name, and calculating the matching degree of the current single digital binarization image and the template image:
wherein, b 0 A binarized pixel matrix for the current single digital image, b 1 Is a binary pixel matrix of the read-in template picture.
9. The transformer substation high-voltage switch cabinet secondary bin panel image recognition system according to claim 1, characterized in that: and (5) sorting the matching degrees in the step 14), obtaining the maximum matching degree by adopting a bubble sorting method, wherein the number corresponding to the binary template graph is the reading result of the current single number, and is inserted into a result list.
10. The transformer substation high-voltage switch cabinet secondary bin panel image recognition system according to claim 1, characterized in that: calculating the actual reading of the digital table in the step 14), and calculating the reading according to the meaning of the size of the digit:
where l is the length of the result list, i.e. the number of detected digits, and result (k) is the kth element of the result list.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210651684.6A CN114998271B (en) | 2022-06-10 | 2022-06-10 | Secondary bin panel image recognition system of transformer substation high-voltage switch cabinet |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210651684.6A CN114998271B (en) | 2022-06-10 | 2022-06-10 | Secondary bin panel image recognition system of transformer substation high-voltage switch cabinet |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114998271A true CN114998271A (en) | 2022-09-02 |
| CN114998271B CN114998271B (en) | 2023-06-13 |
Family
ID=83032379
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210651684.6A Active CN114998271B (en) | 2022-06-10 | 2022-06-10 | Secondary bin panel image recognition system of transformer substation high-voltage switch cabinet |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114998271B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107563396A (en) * | 2017-08-10 | 2018-01-09 | 南京大学 | The construction method of protection screen intelligent identifying system in a kind of electric inspection process |
| CN111386529A (en) * | 2017-06-22 | 2020-07-07 | 宾通机器人公司 | System and method for quickly identifying and processing image regions of interest |
| CN112115807A (en) * | 2020-08-28 | 2020-12-22 | 国网福建省电力有限公司厦门供电公司 | Transformer substation secondary equipment switching operation anti-error simulation method and system based on image recognition |
| CN112150366A (en) * | 2020-09-30 | 2020-12-29 | 广西亚像科技有限责任公司 | Method for identifying states of upper pressure plate and indicator lamp of transformer substation control cabinet |
| CN112183369A (en) * | 2020-09-29 | 2021-01-05 | 国网上海市电力公司 | Pointer instrument reading identification method for transformer substation unmanned inspection |
| WO2022037281A1 (en) * | 2020-08-19 | 2022-02-24 | 广西电网有限责任公司贺州供电局 | Real-time substation field operation monitoring and alarm system based on machine vision |
-
2022
- 2022-06-10 CN CN202210651684.6A patent/CN114998271B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111386529A (en) * | 2017-06-22 | 2020-07-07 | 宾通机器人公司 | System and method for quickly identifying and processing image regions of interest |
| CN107563396A (en) * | 2017-08-10 | 2018-01-09 | 南京大学 | The construction method of protection screen intelligent identifying system in a kind of electric inspection process |
| WO2022037281A1 (en) * | 2020-08-19 | 2022-02-24 | 广西电网有限责任公司贺州供电局 | Real-time substation field operation monitoring and alarm system based on machine vision |
| CN112115807A (en) * | 2020-08-28 | 2020-12-22 | 国网福建省电力有限公司厦门供电公司 | Transformer substation secondary equipment switching operation anti-error simulation method and system based on image recognition |
| CN112183369A (en) * | 2020-09-29 | 2021-01-05 | 国网上海市电力公司 | Pointer instrument reading identification method for transformer substation unmanned inspection |
| CN112150366A (en) * | 2020-09-30 | 2020-12-29 | 广西亚像科技有限责任公司 | Method for identifying states of upper pressure plate and indicator lamp of transformer substation control cabinet |
Non-Patent Citations (1)
| Title |
|---|
| 郑昌庭;王俊;郑克;: "基于图像识别的变电站巡检机器人仪表识别研究", no. 05, pages 61 - 65 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114998271B (en) | 2023-06-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110455822A (en) | A kind of detection method of pcb board defect | |
| CN108009515A (en) | A kind of power transmission line positioning identifying method of the unmanned plane image based on FCN | |
| CN106022345B (en) | A kind of high voltage isolator state identification method based on Hough forest | |
| CN109389109B (en) | Automatic testing method and device for OCR full-text recognition accuracy | |
| CN116188475B (en) | Intelligent control method, system and medium for automatic optical detection of appearance defects | |
| CN111598843A (en) | Power transformer respirator target defect detection method based on deep learning | |
| CN108764134A (en) | A kind of automatic positioning of polymorphic type instrument and recognition methods suitable for crusing robot | |
| CN114511718B (en) | Intelligent management method and system for materials for building construction | |
| CN112070135A (en) | Power equipment image detection method and device, power equipment and storage medium | |
| CN111914767A (en) | Scattered-pollution enterprise detection method and system based on multi-source remote sensing data | |
| CN110309843A (en) | The automatic identifying method of multiclass component in a kind of power equipment image | |
| CN110008877A (en) | A kind of substation isolating-switch detection recognition method based on Faster RCNN | |
| CN113822870A (en) | AI detection method for surface defects of electroluminescent semiconductor plate | |
| CN116993725B (en) | Intelligent patch information processing system of flexible circuit board | |
| CN114998271A (en) | Transformer substation high tension switchgear secondary bin panel image recognition system | |
| CN109598349A (en) | Overhead transmission line fault detection data sample batch processing training method based on classification stochastical sampling | |
| CN113076730A (en) | Test report intelligent management method, device, system and storage medium | |
| CN110909653B (en) | Method for automatically calibrating screen cabinet of distribution room by indoor robot | |
| CN115115609A (en) | Image analysis method and system for plant leaf positive phenotypic characters | |
| CN112699824B (en) | Method and device for detecting constant of electric energy meter and storage medium | |
| CN203149105U (en) | Digital universal meter automatic calibrating device | |
| CN117289667B (en) | Component size quality control system and method | |
| CN109975686A (en) | A kind of circuit board short circuit automatic identifying method based on infrared image processing | |
| He et al. | Multi-State Recognition Method of Substation Switchgear Based on Image Enhancement and Deep Learning | |
| Shan et al. | Research and Implementation of Image Recognition FOR Certificate Sortingand Stamping Robot |
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 | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 572024 area C310, third floor, phase II standard plant, yazhouwan science and Technology City, Yazhou District, Sanya City, Hainan Province Patentee after: Hainan Research Institute of Zhejiang University Patentee after: ZHEJIANG University Address before: 572024 area C310, third floor, phase II standard plant, yazhouwan science and Technology City, Yazhou District, Sanya City, Hainan Province Patentee before: Research Institute of Hainan Zhejiang University Patentee before: ZHEJIANG University |
|
| CP01 | Change in the name or title of a patent holder |