CN111626323A - Electric meter state detection and evaluation method based on deep learning - Google Patents
Electric meter state detection and evaluation method based on deep learning Download PDFInfo
- Publication number
- CN111626323A CN111626323A CN202010280318.5A CN202010280318A CN111626323A CN 111626323 A CN111626323 A CN 111626323A CN 202010280318 A CN202010280318 A CN 202010280318A CN 111626323 A CN111626323 A CN 111626323A
- Authority
- CN
- China
- Prior art keywords
- photo
- ammeter
- deep learning
- detected
- electric meter
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/02—Recognising information on displays, dials, clocks
Abstract
The invention relates to the technical field of photo collection, in particular to an ammeter state detection and evaluation method based on deep learning, which comprises the following steps: acquiring a sample photo of the ammeter, and preprocessing the sample photo; dividing the preprocessed ammeter photos into a training set and a test set, training the training set for multiple times by adopting a deep learning algorithm to obtain parameters of a model, verifying the parameters on the test set, and selecting model parameters with optimal effects; inputting the photo of the ammeter to be detected into the trained model, and outputting a detection result; and evaluating the state of the electric meter according to the detection result. The method is used for detecting the ammeter based on the model obtained by deep learning algorithm training, and has higher identification accuracy, positioning accuracy and detection efficiency.
Description
Technical Field
The invention relates to the technical field of photo collection, in particular to an ammeter state detection and evaluation method based on deep learning.
Background
The ammeter is an important electric power metering device, and it is very important to know its safe state, and the safe power consumption of relation numerous households. The intelligent power grid introduces a computer technology, the movable end is used for shooting, the detection and evaluation of the condition of the electric meter are carried out, and a large amount of manpower and material resources can be saved. Evaluating the safety condition of an electricity meter relies on the accurate detection of the meter and its components, including the accurate identification and accurate positioning of the meter and the various components.
The traditional detection method relies on manual detection, and the identification accuracy, the positioning accuracy and the detection efficiency are low.
Disclosure of Invention
In order to solve the problems, the invention provides an ammeter state detection and evaluation method based on deep learning.
A deep learning-based electric meter state detection and evaluation method comprises the following steps:
acquiring a sample photo of the ammeter, and preprocessing the sample photo;
dividing the preprocessed ammeter photos into a training set and a test set, training the training set for multiple times by adopting a deep learning algorithm to obtain parameters of a model, verifying the parameters on the test set, and selecting model parameters with optimal effects;
inputting the photo of the ammeter to be detected into the trained model, and outputting a detection result;
and evaluating the state of the electric meter according to the detection result.
Preferably, the step of obtaining a sample photo of the electricity meter and preprocessing the sample photo comprises the following steps:
carrying out histogram equalization processing on the sample photo to increase the contrast of the photo;
applying operations of rotation, scaling, translation and affine change to the sample photo with the probability of k to achieve data expansion, wherein k is the probability of data expansion;
and (3) carrying out data annotation on the photo of each electric meter by using an annotation tool, and respectively annotating the position and the size of each part in the electric meter by using a rectangular frame.
Preferably, each component in the electricity meter comprises: ammeter main part, collector, seal, inlet wire switch of being qualified for the next round of competitions, pilot lamp.
Preferably, the step of inputting the photo of the electric meter to be detected into the trained model and outputting the detection result comprises the following steps:
if the seal cannot be detected, judging that the defect of seal missing exists;
if the collector cannot be detected, judging that the defect of collector loss exists;
if the damaged incoming and outgoing switch is detected but the normal incoming and outgoing switch is not detected, judging that the incoming and outgoing switch is damaged;
and detecting the color of the signal of the indicator light, and if the color is a warning color, judging that the signal light is abnormal.
Preferably, the evaluating the state of the electric meter according to the detection result comprises the following steps:
setting corresponding deduction coefficients aiming at different defects of the ammeter;
calculating to obtain a corresponding score according to a detection result of the ammeter photo to be detected;
and evaluating the electric meter to be detected according to the scores.
The invention has the beneficial effects that: obtaining a sample photo of an ammeter, and preprocessing the sample photo; dividing the preprocessed ammeter photos into a training set and a test set, training the training set for multiple times by adopting a deep learning algorithm to obtain parameters of a model, verifying the parameters on the test set, and selecting model parameters with optimal effects; inputting the photo of the ammeter to be detected into the trained model, and outputting a detection result; and evaluating the state of the electric meter according to the detection result. The method is used for detecting the ammeter based on the model obtained by deep learning algorithm training, and has higher identification accuracy, positioning accuracy and detection efficiency.
Drawings
The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
FIG. 1 is a schematic flow chart of a deep learning-based electricity meter status detection and evaluation method according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart of step S1 in the deep learning-based meter status detection and evaluation method according to an embodiment of the present invention;
FIG. 3 is a schematic flow chart of step S3 in the deep learning-based meter status detection and evaluation method according to one embodiment of the present invention;
fig. 4 is a schematic flow chart of step S4 in the deep learning-based electricity meter status detection and evaluation method according to an embodiment of the present invention.
Detailed Description
The technical solutions of the present invention will be further described below with reference to the accompanying drawings, but the present invention is not limited to these embodiments.
The method comprises the following steps of obtaining a sample photo of the ammeter, and preprocessing the sample photo; dividing the preprocessed ammeter photos into a training set and a test set, training the training set for multiple times by adopting a deep learning algorithm to obtain parameters of a model, verifying the parameters on the test set, and selecting model parameters with optimal effects; inputting the photo of the ammeter to be detected into the trained model, and outputting a detection result; and evaluating the state of the electric meter according to the detection result. The method is used for detecting the electric meter based on the model obtained by deep learning algorithm training, and has higher identification accuracy, positioning accuracy and detection efficiency compared with manual detection.
Based on the above thought, the invention provides an electric meter state detection and evaluation method based on deep learning, as shown in fig. 1, comprising the following steps:
s1: and acquiring a sample photo of the ammeter, and preprocessing the sample photo.
Firstly, the photo collecting device collects photos of the electric meter, and approximately 10000 photos are collected. The photo collecting device can be a mobile phone and the like.
In order to train to obtain a more accurate model, as shown in fig. 2, in this embodiment, the preprocessing of the sample photo specifically includes the following steps:
s11: carrying out histogram equalization processing on the sample photo to increase the contrast of the photo;
s12: applying operations of rotation, scaling, translation and affine change to the sample photo with the probability of k to achieve data expansion, wherein k is the probability of data expansion;
s13: and (3) carrying out data annotation on the photo of each electric meter by using an annotation tool, and respectively annotating the position and the size of each part in the electric meter by using a rectangular frame.
Firstly, histogram equalization processing is carried out on all photos, the contrast of the photos is increased, subsequent target detection is facilitated, then rotation, scaling, translation and affine change operations are applied to the photos in a data set according to the probability of 0.5, the purpose of expanding the data set is achieved, finally, a marking tool is used for carrying out data marking on the photos of each ammeter, and the positions and the sizes of an ammeter main body, a collector, a seal, an incoming line and outgoing line switch and an indicator light are marked by rectangular frames respectively.
S2: dividing the preprocessed ammeter photos into a training set and a testing set, training the training set for multiple times by adopting a deep learning algorithm to obtain parameters of the model, verifying the parameters on the testing set, and selecting the model parameters with the optimal effect.
In this embodiment, all the electric meter photos after the labeling is completed are divided into two parts, wherein 8000 are training sets, and the rest 2000 are testing sets.
And (3) training the training set for multiple times by using a deep learning method to obtain parameters of the DCNN model, verifying the effect on the test set, respectively recording the detection accuracy of the ammeter, the collector, the seal, the incoming and outgoing line switch and the indicator light and the whole detection time, and selecting the model parameters with the optimal effect for application.
The concept of deep learning is derived from the research of an artificial neural network, and a multi-layer perceptron with multiple hidden layers is a deep learning structure. Deep learning forms a more abstract class or feature of high-level representation properties by combining low-level features to discover a distributed feature representation of the data. Deep learning is a method based on characterization learning of data in machine learning. Observations can be represented in a number of ways, such as a vector of intensity values for each pixel, or more abstractly as a series of edges, specially shaped regions, etc., while tasks are more easily learned from instances using some specific representation methods.
S3: and inputting the photo of the ammeter to be detected into the trained model, and outputting a detection result.
In this embodiment, as shown in fig. 3, the inputting the photo of the electric meter to be detected into the trained model and outputting the detection result includes the following steps:
s31: if the seal cannot be detected, judging that the defect of seal missing exists;
s32: if the collector cannot be detected, judging that the defect of collector loss exists;
s33: if the damaged incoming and outgoing switch is detected but the normal incoming and outgoing switch is not detected, judging that the incoming and outgoing switch is damaged;
s34: and detecting the color of the signal of the indicator light, and if the color is a warning color, judging that the signal light is abnormal.
And inputting the photo to be detected into the model, detecting the positions and sizes of the ammeter, the collector, the seal, the damaged incoming and outgoing line switch and the indicator lamp area, and recording the detection result returned by the model.
S4: and evaluating the state of the electric meter according to the detection result.
In this embodiment, as shown in fig. 4, the evaluating the state of the electric meter according to the detection result includes the following steps:
s41: setting corresponding deduction coefficients aiming at different defects of the ammeter;
s42: calculating to obtain a corresponding score according to a detection result of the ammeter photo to be detected;
s43: and evaluating the electric meter to be detected according to the scores.
And dividing the defect grade of the electric meter into five grades according to the final score: class I: the fraction is less than 60; and II: the score is more than or equal to 60 and less than 70; class III: the score is more than or equal to 70 and less than 80; and IV: the score is more than or equal to 80 and less than 90; and (3) the other: the score is more than or equal to 90.
Various modifications or additions may be made to the described embodiments or alternatives may be employed by those skilled in the art without departing from the spirit or ambit of the invention as defined in the appended claims.
Claims (5)
1. A method for detecting and evaluating the state of an electric meter based on deep learning is characterized by comprising the following steps:
acquiring a sample photo of the ammeter, and preprocessing the sample photo;
dividing the preprocessed ammeter photos into a training set and a test set, training the training set for multiple times by adopting a deep learning algorithm to obtain parameters of a model, verifying the parameters on the test set, and selecting model parameters with optimal effects;
inputting the photo of the ammeter to be detected into the trained model, and outputting a detection result;
and evaluating the state of the electric meter according to the detection result.
2. The deep learning-based electric meter state detection and evaluation method according to claim 1, wherein the step of obtaining a sample photo of the electric meter and preprocessing comprises the following steps:
carrying out histogram equalization processing on the sample photo to increase the contrast of the photo;
applying operations of rotation, scaling, translation and affine change to the sample photo with the probability of k to achieve data expansion, wherein k is the probability of data expansion;
and (3) carrying out data annotation on the photo of each electric meter by using an annotation tool, and respectively annotating the position and the size of each part in the electric meter by using a rectangular frame.
3. The deep learning based meter condition detection and assessment method according to claim 1, wherein said meter components include: ammeter main part, collector, seal, inlet wire switch of being qualified for the next round of competitions, pilot lamp.
4. The deep learning-based meter state detection and evaluation method according to claim 3, wherein the step of inputting the photo of the meter to be detected into the trained model and outputting the detection result comprises the following steps:
if the seal cannot be detected, judging that the defect of seal missing exists;
if the collector cannot be detected, judging that the defect of collector loss exists;
if the damaged incoming and outgoing switch is detected but the normal incoming and outgoing switch is not detected, judging that the incoming and outgoing switch is damaged;
and detecting the color of the signal of the indicator light, and if the color is a warning color, judging that the signal light is abnormal.
5. The deep learning-based meter state detection and evaluation method according to claim 4, wherein the evaluation of the meter state according to the detection result comprises the following steps:
setting corresponding deduction coefficients aiming at different defects of the ammeter;
calculating to obtain a corresponding score according to a detection result of the ammeter photo to be detected;
and evaluating the electric meter to be detected according to the scores.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010280318.5A CN111626323A (en) | 2020-04-10 | 2020-04-10 | Electric meter state detection and evaluation method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010280318.5A CN111626323A (en) | 2020-04-10 | 2020-04-10 | Electric meter state detection and evaluation method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111626323A true CN111626323A (en) | 2020-09-04 |
Family
ID=72259630
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010280318.5A Pending CN111626323A (en) | 2020-04-10 | 2020-04-10 | Electric meter state detection and evaluation method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111626323A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112633739A (en) * | 2020-12-30 | 2021-04-09 | 安徽广志电气有限公司 | Power distribution control cabinet energy loss assessment method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017214503A1 (en) * | 2016-06-10 | 2017-12-14 | Brain Corporation | Systems and methods for automatic detection of spills |
| CN108647677A (en) * | 2018-05-14 | 2018-10-12 | 山东师范大学 | The ammeter appearance and performance intelligent detecting method and device that view-based access control model calculates |
| CN110046617A (en) * | 2019-03-15 | 2019-07-23 | 西安交通大学 | A kind of digital electric meter reading self-adaptive identification method based on deep learning |
| WO2019208066A1 (en) * | 2018-04-25 | 2019-10-31 | 株式会社リンクジャパン | Image capture device and system |
| CN110503046A (en) * | 2019-08-26 | 2019-11-26 | 华北电力大学（保定） | A kind of lead sealing method of identification based on image recognition technology |
-
2020
- 2020-04-10 CN CN202010280318.5A patent/CN111626323A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017214503A1 (en) * | 2016-06-10 | 2017-12-14 | Brain Corporation | Systems and methods for automatic detection of spills |
| WO2019208066A1 (en) * | 2018-04-25 | 2019-10-31 | 株式会社リンクジャパン | Image capture device and system |
| CN108647677A (en) * | 2018-05-14 | 2018-10-12 | 山东师范大学 | The ammeter appearance and performance intelligent detecting method and device that view-based access control model calculates |
| CN110046617A (en) * | 2019-03-15 | 2019-07-23 | 西安交通大学 | A kind of digital electric meter reading self-adaptive identification method based on deep learning |
| CN110503046A (en) * | 2019-08-26 | 2019-11-26 | 华北电力大学（保定） | A kind of lead sealing method of identification based on image recognition technology |
Non-Patent Citations (5)
| Title |
|---|
| CHENG DAI 等: "Intelligent Ammeter Reading Recognition Method", 《2019 IEEE 8TH JOINT INTERNATIONAL INFORMATION TECHNOLOGY AND ARTIFICIAL INTELLIGENCE CONFERENCE (ITAIC)》 * |
| 商曦文 等: "智能电能表运行状态评估技术研究综述", 《电测与仪表》 * |
| 宋青松 等: "基于多尺度卷积神经网络的交通标志识别", 《湖南大学学报(自然科学版)》 * |
| 彭宇: "基于卷积神经网络的电能表液晶屏缺陷检测系统设计", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
| 朱天宇 等: "基于指标权重分析的电能表智能管理体系", 《国网技术学院学报》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112633739A (en) * | 2020-12-30 | 2021-04-09 | 安徽广志电气有限公司 | Power distribution control cabinet energy loss assessment method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103324937B (en) | The method and apparatus of label target | |
| CN108074231B (en) | Magnetic sheet surface defect detection method based on convolutional neural network | |
| CN111242123B (en) | Power equipment fault diagnosis method based on infrared image | |
| CN109344905A (en) | A kind of transmission facility automatic fault recognition methods based on integrated study | |
| CN110288032B (en) | Vehicle driving track type detection method and device | |
| CN111401419A (en) | Improved RetinaNet-based employee dressing specification detection method | |
| CN111667455A (en) | AI detection method for various defects of brush | |
| CN110782461A (en) | Mask-RCNN-based infrared image segmentation method for electric power equipment | |
| CN111368690A (en) | Deep learning-based video image ship detection method and system under influence of sea waves | |
| CN111852792B (en) | Fan blade defect self-diagnosis positioning method based on machine vision | |
| CN111127448B (en) | Method for detecting air spring fault based on isolated forest | |
| CN108734691A (en) | A kind of transmission line of electricity defect image recognition methods | |
| CN109946304A (en) | Surface defects of parts on-line detecting system and detection method based on characteristic matching | |
| CN114627082A (en) | Photovoltaic module EL defect detection method based on improved YOLOv5 network | |
| CN114066848A (en) | FPCA appearance defect visual inspection system | |
| CN113780484B (en) | Industrial product defect detection method and device | |
| CN115019294A (en) | Pointer instrument reading identification method and system | |
| CN111626323A (en) | Electric meter state detection and evaluation method based on deep learning | |
| CN109376738A (en) | Altering detecting method based on vehicle VIN code start-stop symbol type | |
| CN111291691B (en) | Deep learning-based substation secondary equipment instrument panel reading detection method | |
| CN112132088B (en) | Inspection point missing inspection identification method | |
| CN109829887A (en) | A kind of image quality measure method based on deep neural network | |
| CN112699824B (en) | Method and device for detecting constant of electric energy meter and storage medium | |
| CN112927222B (en) | Method for realizing multi-type photovoltaic array hot spot detection based on hybrid improved Faster R-CNN | |
| CN114494103A (en) | Defect detection method and device |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200904 |