CN109871726A - A kind of similar instrument registration method for early warning based on QR code and image recognition - Google Patents
A kind of similar instrument registration method for early warning based on QR code and image recognition Download PDFInfo
- Publication number
- CN109871726A CN109871726A CN201910158794.7A CN201910158794A CN109871726A CN 109871726 A CN109871726 A CN 109871726A CN 201910158794 A CN201910158794 A CN 201910158794A CN 109871726 A CN109871726 A CN 109871726A
- Authority
- CN
- China
- Prior art keywords
- code
- early warning
- image
- image recognition
- similar instrument
- 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
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a kind of similar instrument registration method for early warning based on QR code and image recognition, including S1, preparation QR code；S2, QR code identification model is generated；S3, decoding stage；S4, early warning determine, establish convolutional neural networks algorithm model by the above four-stage and scan QR code, identify threshold value of warning；The present invention is compared to common two-dimensional bar code, QR code can store more information, Chinese character at most can store 1800 characters, it also can be to text, the address URL and other kinds of data encryption, adjusting to a line scanner is needed in scanning without as ordinary two dimensional bar code simultaneously, using the above feature of QR code, QR code is combined with thermal power plant's water station instrument registration early warning, the reliability of energy effective guarantee early warning system.
Description
Technical field
The present invention relates to a kind of method for early warning more particularly to a kind of similar instrument number early warning based on QR code and image recognition
Method belongs to robot application technology field.
Background technique
Thermal power plant water station, is the place for carrying out water chemical treatment, and water chemical treatment can guarantee therrmodynamic system each section
With good water and steam quality, to prevent the fouling, corrosion and salification of heat power equipment, to safe, the economic fortune for guaranteeing power plant
Row has a very important significance.Changing includes a large amount of instrument in water station, for monitoring working medium and equipment during water chemical treatment
The important parameters such as temperature, pressure, electric current, the normal of parameters are of great significance to water chemical treatment process normal operation, institute
To need timing to check whether interpretation exceeds threshold value to instrument registration in water station is changed.It is consumed in traditional manual inspection time-consuming
Under the inefficient background of power, electric inspection process robot comes into being.Image recognition has had mature in conjunction with crusing robot
Research and application case.
Instrument primary categories have pressure gauge, thermometer, flowmeter in power plant's water station, wherein being no lack of has appearance similar or complete
Exactly the same instrument, and the threshold value of the instrument in water chemical treatment not homologous ray may be different, therefore use traditional image
Identification technology is likely to occur wrong report and fails to report situation when carrying out classification and early warning to instrument, threaten the operation peace of thermal power plant's water station
Full property and economy, or even influence the safety in operation and economy of entire fired power generating unit.
Therefore how on the basis of existing intellectualized technology realize the accurate early warning of instrument registration be those skilled in the art
Urgent problem to be solved.
Summary of the invention
Goal of the invention: in order to overcome the deficiencies in the prior art, the present invention, which provides one kind, can be realized instrument registration
The similar instrument registration method for early warning based on QR code and image recognition of accurate early warning.Technical solution: it is asked to solve above-mentioned technology
Topic, a kind of similar instrument registration method for early warning based on QR code and image recognition of the invention include the following steps:
S1, QR code is prepared: to each preparing its distinctive QR code to the similar instrument of early warning and the QR code is attached to it
In corresponding instrument, information that the QR code includes are as follows: meter number and threshold value of warning；
S2, it generates QR code identification model: acquiring the QR code image prepared, input in convolutional neural networks algorithm model and instruct
Practice, obtains the QR code identification model that can accurately identify QR code image；This QR code identification model is embedded in crusing robot system；
S3, decoding stage: crusing robot acquisition instrument during inspection corresponding QR code generates in starting step S3
QR code identification model the QR code of acquisition is decoded, obtain the number and threshold value of warning information of the instrument；
S4, early warning determine: crusing robot acquisition instrument reading information and with the early warning threshold that is carried in the QR code that parses
Value information comparison, if reading information, within the scope of threshold value of warning, crusing robot continues to execute inspection operation；If reading information
Not within the scope of threshold value of warning, then early warning is carried out.
Further, the convolutional neural networks algorithm model in S3 passes through error backpropagation algorithm training.
Further, error backpropagation algorithm includes propagation stage and back-propagation stage forward；
The algorithm of propagation stage forward an are as follows: sample (X, Y) is chosen from QR code image obtained in S2, wherein X is
QR code image, Y are the corresponding device numberings of QR code, and X is inputted to obtain reality output O in minor function Ap；
Op=Fn(…(F2(F1(X·W(1))W(2))…)W(n))；
Wherein F1For the 1st layer network activation primitive, F2For layer 2 network activation primitive, FnFor n-th layer network activation function,
W1For the 1st layer network weight matrix, W2For layer 2 network weight matrix, WnFor n-th layer network weight matrix；
The algorithm in back-propagation stage are as follows: calculate the reality output O that propagation stage obtains forwardpWith corresponding ideal output Y
Difference；The as error of model calculation value and true value；Difference is the error of model calculation value and true value, by the error
Coefficient as each layer network of feedback modifiers.
Then according to the method for minimization error by error back propagation, weight matrix is adjusted using gradient descent method.
Further, the operating method of gray processing is to convert grayscale image for original RGB color image using mean value method
Picture.
Further, the operating method of image denoising is to effectively remove the random noise in image using median filtering method.
Further, the operating method of binaryzation is that image is carried out binary conversion treatment using Otsu algorithm, will be had originally
There is the image of 256 gray levels to be converted into the black white image that gray level only has 2, reduces memory space and operation complexity.
Workflow of the invention are as follows: the instrumentation with same model in statistics thermal power plant's water station first, to every
A instrument is numbered, and individually sets threshold value of warning to each instrument, using QR code generator, production comprising meter number and
Threshold value of warning information QR code.QR code printing and PVC protective case are mounted below corresponding instrument at 5cm.
QR code image is acquired, image is pre-processed, image inputs convolutional neural networks algorithm model by treated
Middle training obtains the model that can accurately identify QR code image, and the QR code identification model completed will have been trained to be embedded in crusing robot
In system.
Crusing robot is run near the meter location, is shot using high-definition camera to ambient enviroment, is worked as QR
After code identification model detects in content of shooting and QR code occurs, scanning QR code is decoded information therein, obtains the instrument
Number and threshold value of warning.Then image recognition reading is carried out to instrument, the instrument registration is judged whether in threshold value, if not
In threshold value, pre-warning signal is sent to background monitoring room；If not executing operation without departing from threshold range.Instrument is compiled
Number and registration typing background data base, then proceed to execute above-mentioned inspection operation, until completing this patrol task.
The utility model has the advantages that (1), compared to common two-dimensional bar code, QR code can store more information, and Chinese character at most may be used
It, also can be to text, the address URL and other kinds of data encryption, while without as common two to store 1800 characters
Dimension bar code is the same to need adjusting to a line scanner in scanning, using the above feature of QR code, by QR code and thermal power plant's water station instrument
Indicate that number early warning combines, the reliability of energy effective guarantee early warning system；(2) this method is implemented simple, at low cost, improves inspection
Robot Meter recognition precision, to effectively improve the safety of thermal power plant's water station operation.
Detailed description of the invention
Fig. 1 is flow chart of the invention.
Specific embodiment
The present invention will be further explained with reference to the accompanying drawing.
Embodiment
The problem of in order to realize the accurate early warning of the instrument registration pointed out in background technique, the application introduces QR code, and QR code is
One kind of two-dimensional bar code can store additional information than common bar code, also without needing adjusting to a line in scanning as common bar code
Scanner, and the coding and decoding technology of QR code book body is well-known technique, the application makees the instrument carried out just with this technology
Therefore the identification of table no longer specifically repeats.
The present invention provides a kind of similar instrument based on QR code and image recognition that can be realized the accurate early warning of instrument registration
Registration method for early warning, includes the following steps:
S1, QR code is prepared: to each preparing its distinctive QR code to the similar instrument of early warning and the QR code is attached to it
In corresponding instrument, information that the QR code includes are as follows: meter number and threshold value of warning；
S2, it generates QR code identification model: acquiring the QR code image prepared, input in convolutional neural networks algorithm model and instruct
Practice, obtains the QR code identification model that can accurately identify QR code image；This QR code identification model is embedded in crusing robot system；
S3, decoding stage: crusing robot acquisition instrument during inspection corresponding QR code generates in starting step S3
QR code identification model the QR code of acquisition is decoded, obtain the number and threshold value of warning information of the instrument；
S4, early warning determine: crusing robot acquisition instrument reading information and with the early warning threshold that is carried in the QR code that parses
Value information comparison, if reading information, within the scope of threshold value of warning, crusing robot continues to execute inspection operation；If reading information
Not within the scope of threshold value of warning, then early warning is carried out.
Convolutional neural networks algorithm model in S3 passes through error backpropagation algorithm training.
Error backpropagation algorithm includes propagation stage and back-propagation stage forward；
The algorithm of propagation stage forward an are as follows: sample (X, Y) is chosen from QR code image obtained in S2, wherein X is
QR code image, Y are the corresponding device numberings of QR code, and X is inputted to obtain reality output O in minor function Ap；
Op=Fn(…(F2(F1(X·W(1))W(2))…)W(n))；
Wherein F1For the 1st layer network activation primitive, F2For layer 2 network activation primitive, FnFor n-th layer network activation function,
W1For the 1st layer network weight matrix, W2For layer 2 network weight matrix, WnFor n-th layer network weight matrix；
The algorithm in back-propagation stage are as follows: calculate the reality output O that propagation stage obtains forwardpWith corresponding ideal output Y
Difference；The as error of model calculation value and true value；Difference is the error of model calculation value and true value, by the error
Coefficient as each layer network of feedback modifiers.
Then according to the method for minimization error by error back propagation, weight matrix is adjusted using gradient descent method.
The operating method of gray processing is to convert gray level image for original RGB color image using mean value method.Image is gone
The operating method made an uproar is to effectively remove the random noise in image using median filtering method.The operating method of binaryzation is to use
Image is carried out binary conversion treatment by Otsu algorithm, by originally with 256 gray levels image be converted into gray level only have 2 it is black
White image reduces memory space and operation complexity.
The basic principles, main features and advantages of the invention have been shown and described above.The technical staff of the industry should
Understand, the above embodiments do not limit the invention in any form, all obtained by the way of equivalent substitution or equivalent transformation
Technical solution is fallen within the scope of protection of the present invention.
Claims (7)
1. a kind of similar instrument registration method for early warning based on QR code and image recognition, it is characterised in that include the following steps:
S1, QR code is prepared: to each preparing its distinctive QR code to the similar instrument of early warning and the QR code is attached to its correspondence
Instrument on, information that the QR code includes are as follows: meter number and threshold value of warning；
S2, it generates QR code identification model: acquiring the QR code image prepared, input training in convolutional neural networks algorithm model,
Obtain to accurately identify the QR code identification model of QR code image；This QR code identification model is embedded in crusing robot system；
S3, decoding stage: the corresponding QR code of crusing robot acquisition instrument during inspection, the QR generated in starting step S3
Code identification model is decoded the QR code of acquisition, obtains the number and threshold value of warning information of the instrument；
S4, early warning determine: crusing robot acquisition instrument reading information and with the threshold value of warning letter that is carried in the QR code that parses
Breath comparison, if reading information, within the scope of threshold value of warning, crusing robot continues to execute inspection operation；If reading information does not exist
Within the scope of threshold value of warning, then early warning is carried out.
2. a kind of similar instrument registration method for early warning based on QR code and image recognition according to claim 1, feature
Be: the convolutional neural networks algorithm model in the S2 passes through error backpropagation algorithm training.
3. a kind of similar instrument registration method for early warning based on QR code and image recognition according to claim 2, feature
Be: the error backpropagation algorithm includes propagation stage and back-propagation stage forward；
The algorithm of the propagation stage forward are as follows: the QR code image obtained in the S2 forms data set, Cong Zhongxuan with corresponding equipment
A sample (X, Y) is taken, wherein X is QR code image, and Y is the corresponding device numbering of QR code, and X is inputted to obtain in minor function A
To reality output Op；
Op=Fn(…(F2(F1(X·W(1)) W(2))…) W(n))；
Wherein F1For the 1st layer network activation primitive, F2For layer 2 network activation primitive, FnFor n-th layer network activation function, W1For
1st layer network weight matrix, W2For layer 2 network weight matrix, WnFor n-th layer network weight matrix;
The algorithm in the back-propagation stage are as follows: calculate the reality output O that propagation stage obtains forwardpWith corresponding ideal output Y's
The error of difference, as model calculation value and true value；
Then according to the method for minimization error by error back propagation, weight matrix is adjusted using gradient descent method.
4. a kind of similar instrument registration method for early warning based on QR code and image recognition according to claim 1, feature
It is: after acquiring the QR code image prepared in step S2, is pre-processed, the pre-treatment step includes gray processing, image
Denoising and binary conversion treatment.
5. a kind of similar instrument registration method for early warning based on QR code and image recognition according to claim 4, feature
Be: the operating method of the gray processing is to convert gray level image for original RGB color image using mean value method.
6. a kind of similar instrument registration method for early warning based on QR code and image recognition according to claim 4, feature
Be: the operating method of described image denoising is using median filtering method.
7. a kind of similar instrument registration method for early warning based on QR code and image recognition according to claim 4, feature
Be: the operating method of the binaryzation is that image is carried out binary conversion treatment using Otsu algorithm, will have 256 ashes originally
The image of degree grade is converted into the black white image that gray level only has 2.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2018116222471 | 2018-12-28 | ||
| CN201811622247 | 2018-12-28 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109871726A true CN109871726A (en) | 2019-06-11 |
Family
ID=66919600
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910158794.7A Pending CN109871726A (en) | 2018-12-28 | 2019-03-04 | A kind of similar instrument registration method for early warning based on QR code and image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109871726A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113962231A (en) * | 2021-10-13 | 2022-01-21 | 杭州胜铭纸业有限公司 | Optical identification comparison method and system for information codes of packing cases |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103955694A (en) * | 2014-04-09 | 2014-07-30 | 广州邦讯信息***有限公司 | Image recognition meter reading system and method |
| CN104700062A (en) * | 2015-03-20 | 2015-06-10 | 中国联合网络通信集团有限公司 | Method and equipment for identifying two-dimension code |
| CN106442553A (en) * | 2016-09-12 | 2017-02-22 | 佛山市南海区广工大数控装备协同创新研究院 | Detection and recognition device and method for sprayed codes on cylindrical surfaces of copper rings |
| US20180052053A1 (en) * | 2015-05-12 | 2018-02-22 | Don Atkinson | Universal color tag and tracing system |
| CN108269371A (en) * | 2017-09-27 | 2018-07-10 | 缤果可为（北京）科技有限公司 | Commodity automatic settlement method, device, self-service cashier |
| CN108520193A (en) * | 2018-03-27 | 2018-09-11 | 康体佳智能科技（深圳）有限公司 | Quick Response Code identifying system based on neural network and recognition methods |
| CN108890657A (en) * | 2018-07-11 | 2018-11-27 | 清华大学合肥公共安全研究院 | A kind of power distribution room intelligent inspection robot |
-
2019
- 2019-03-04 CN CN201910158794.7A patent/CN109871726A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103955694A (en) * | 2014-04-09 | 2014-07-30 | 广州邦讯信息***有限公司 | Image recognition meter reading system and method |
| CN104700062A (en) * | 2015-03-20 | 2015-06-10 | 中国联合网络通信集团有限公司 | Method and equipment for identifying two-dimension code |
| US20180052053A1 (en) * | 2015-05-12 | 2018-02-22 | Don Atkinson | Universal color tag and tracing system |
| CN106442553A (en) * | 2016-09-12 | 2017-02-22 | 佛山市南海区广工大数控装备协同创新研究院 | Detection and recognition device and method for sprayed codes on cylindrical surfaces of copper rings |
| CN108269371A (en) * | 2017-09-27 | 2018-07-10 | 缤果可为（北京）科技有限公司 | Commodity automatic settlement method, device, self-service cashier |
| CN108520193A (en) * | 2018-03-27 | 2018-09-11 | 康体佳智能科技（深圳）有限公司 | Quick Response Code identifying system based on neural network and recognition methods |
| CN108890657A (en) * | 2018-07-11 | 2018-11-27 | 清华大学合肥公共安全研究院 | A kind of power distribution room intelligent inspection robot |
Non-Patent Citations (3)
| Title |
|---|
| 樊重俊等: "《大数据分析与应用》", 31 January 2016, 立信会计出版社 * |
| 陈玲等: "《图像目标分割方法》", 30 November 2016, 哈尔滨工程大学出版社 * |
| 高宏伟等: "《电子封装工艺与装备技术基础教程》", 30 June 2017 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113962231A (en) * | 2021-10-13 | 2022-01-21 | 杭州胜铭纸业有限公司 | Optical identification comparison method and system for information codes of packing cases |
| CN113962231B (en) * | 2021-10-13 | 2024-03-26 | 杭州胜铭纸业有限公司 | Packaging box information code optical identification comparison method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113160123B (en) | Leather defect detection method, system and device based on YOLOv5 | |
| CN108399419A (en) | Chinese text recognition methods in natural scene image based on two-dimentional Recursive Networks | |
| CN115951584B (en) | Temperature control system and method for roller fixation machine | |
| CN109165376A (en) | Style character generating method based on a small amount of sample | |
| CN107492098A (en) | It is a kind of based on PCA and CNN high-temperature forging surface defect in position detecting method | |
| CN111652233B (en) | Text verification code automatic identification method aiming at complex background | |
| CN108403111A (en) | A kind of epileptic electroencephalogram (eeg) recognition methods and system based on convolutional neural networks | |
| CN112506797B (en) | Performance test method for medical image recognition system | |
| CN107749058A (en) | A kind of machine vision detection method and system of boiler tubing surface defect | |
| CN102194275A (en) | Automatic ticket checking method for train tickets | |
| CN110309512A (en) | A kind of Chinese grammer error correction method thereof based on generation confrontation network | |
| CN115302963B (en) | Bar code printing control method, system and medium based on machine vision | |
| WO2013043338A1 (en) | Processes and systems for optical character recognition using generated font templates | |
| CN109102029A (en) | Information, which maximizes, generates confrontation network model synthesis face sample quality appraisal procedure | |
| CN110830489A (en) | Method and system for detecting counterattack type fraud website based on content abstract representation | |
| CN113190846A (en) | Malicious domain name training data generation method based on generation countermeasure network model | |
| CN109871726A (en) | A kind of similar instrument registration method for early warning based on QR code and image recognition | |
| CN108647683A (en) | It is a kind of that the character made an uproar is added to fight verification code generation method and system based on frequency domain | |
| CN115497107B (en) | Zero-sample Chinese character recognition method based on stroke and radical decomposition | |
| Chaban et al. | Machine learning attack on copy detection patterns: are $1\times 1$ patterns cloneable? | |
| CN117373111A (en) | AutoHOINet-based human-object interaction detection method | |
| Zhang et al. | Automatic detection of prohibited items with small size in X-ray images | |
| CN109829527A (en) | A kind of the water sample tracking and tracking system of surface water environment monitoring | |
| CN104481508A (en) | Oilfield rod-pumping well fault diagnosis method by combining comentropy and gray level incidence matrix | |
| CN107886093A (en) | A kind of character detection method, system, equipment and computer-readable storage medium |
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: 20190611 |