CN109815250A - A kind of sectional members assembly error-preventing method and system - Google Patents
A kind of sectional members assembly error-preventing method and system Download PDFInfo
- Publication number
- CN109815250A CN109815250A CN201910089031.1A CN201910089031A CN109815250A CN 109815250 A CN109815250 A CN 109815250A CN 201910089031 A CN201910089031 A CN 201910089031A CN 109815250 A CN109815250 A CN 109815250A
- Authority
- CN
- China
- Prior art keywords
- sectional members
- assembled
- number information
- image
- module
- 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
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of sectional members assembly error-preventing method and systems.This method comprises: acquisition has the image of the sectional members to be assembled of artificial setting number from assembly environment;The method of the artificial setting number includes spraying, engraving and hand-written;Identify the number information in sectional members image to be assembled;The number information includes classification and the position of number;The classification includes number and letter;The position includes the position of the number and the position of the letter;It is retrieved from database according to the number information, determines matching number;Output assembling scheme is numbered according to the matching;The assembling scheme includes construction requirement, installation site and operating process;Sectional members to be assembled are assembled according to the assembling scheme.This method or system can reduce the dependence in Large-Scale Equipment assembling process to operator's subjectivity, reduce misloading and neglected loading situation in assembling process.
Description
Technical field
The present invention relates to sectional members to assemble field, assembles error-preventing method and system more particularly to a kind of sectional members.
Background technique
The misloading of sectional members and neglected loading are incident quality problems in Large-Scale Equipment manufacture assembling process, especially carefully
The neglected loading of small sectional members.These neglected loadings or wrongly installed sectional members are in the inside of system mostly, it is difficult to by simple
Sight check discovery.When assembly crewman is unfamiliar with assembling link, fails to understand or misunderstands drawing, not when stringent construction in line with the drawings, hold very much
Easily cause misloading or neglected loading.This mistake will affect the execution of subsequent technique, and even need to do over again modification.
Due to there is the number of one section of spraying in each sectional members of Large-Scale Equipment, to record the opposite position of the sectional members
It sets, the information such as operating process, quality requirement.Therefore, it mainly prevents from assembling using inspection table in Large-Scale Equipment manufacture at present
The misloading of journey sectional members and neglected loading, but the subjectivity of this error protection mode is stronger, the subjectivity vulnerable to operator
Property influence.
Summary of the invention
The object of the present invention is to provide a kind of sectional members assembly error-preventing method and systems, to information-based, automatic chemoprevention
There is mistake in sectional members assembly only.
To achieve the above object, the present invention provides following schemes:
A kind of sectional members assembly error-preventing method, which comprises
Acquisition has the image of the sectional members to be assembled of artificial setting number from assembly environment;The artificial setting is compiled
Number method include spraying, engraving and it is hand-written;
Identify the number information in sectional members image to be assembled;The number information includes classification and the position of number;
The classification includes number and letter;The position includes the position of the number and the position of the letter;
It is retrieved from database according to the number information, determines matching number;
Output assembling scheme is numbered according to the matching;The assembling scheme includes construction requirement, installation site and behaviour
Make process;
Sectional members to be assembled are assembled according to the assembling scheme.
Optionally, the number information in the identification sectional members image to be assembled, specifically includes:
Obtain identification model;
The sectional members image to be assembled is input to the identification model, obtains number information.
Optionally, between the acquisition identification model, further includes:
Training sample is obtained, the training sample is the image of 100 or more the sectional members with artificial setting number;
By 100 the above have the image of the sectional members of artificial setting number to convolutional neural networks model into
Row training is exported as a result, the output result is number information;
Judge whether the output result meets trained termination condition;The trained termination condition is in output result by just
The ratio of the total character quantity of character Zhan really identified reaches 95% or more;
If so, determining that the convolutional neural networks model is identification model;
If it is not, adjusting the parameter of the convolutional neural networks model by back-propagation algorithm, make to export result satisfaction instruction
Practice termination condition.
Optionally, described to be retrieved from database according to the number information, it determines matching number, specifically includes:
It is retrieved from database according to the number information, obtains the number completely the same with the number information,
It is determined as matching number.
The present invention also provides a kind of sectional members to assemble fail-safe system, the system comprises:
Image capture module, the figure for the sectional members to be assembled that there is artificial setting to number for acquisition from assembly environment
Picture;The method of the artificial setting number includes spraying, engraving and hand-written;
Identification module, for identification number information in sectional members image to be assembled;The number information includes number
Classification and position;The classification includes number and letter;The position include the number position and it is described letter
Position;
Retrieval module determines matching number for being retrieved from database according to the number information;
Output module, for numbering output assembling scheme according to the matching;The assembling scheme includes construction requirement, peace
Holding position and operating process;
Load module, for being assembled according to the assembling scheme to sectional members to be assembled.
Optionally, the identification module specifically includes:
Acquiring unit, for obtaining identification model;
Input unit obtains number information for the sectional members image to be assembled to be input to the identification model.
Optionally, the assembly fail-safe system further include:
Sample acquisition module, for obtaining training sample, the training sample, which is 100 or more, has artificial setting number
Sectional members image;
Training module, for having the image of the sectional members of artificial setting number to convolution by 100 the above
Neural network model is trained, and is exported as a result, the output result is number information;
Judgment module, for judging whether the output result meets trained termination condition;The trained termination condition is
The ratio for the total character quantity of character Zhan being correctly validated in output result reaches 95% or more;
As a result determining module, for determining the convolutional Neural net when the output result meets training termination condition
Network model is identification model;
Module is adjusted, for being adjusted by back-propagation algorithm when the output result is unsatisfactory for training termination condition
The parameter of the convolutional neural networks model makes to export result satisfaction training termination condition.
Optionally, the retrieval module specifically includes:
Retrieval unit obtains complete with the number information for being retrieved from database according to the number information
Complete consistent number is determined as matching number.
Compared with prior art, the present invention has following technical effect that the present invention identifies sectional members figure to be assembled first
Number information as in;Then it is retrieved from database according to the number information, determines matching number;According to described
Assembling scheme is exported with number;Sectional members to be assembled are assembled according to the assembling scheme.It can reduce large size in this way
The dependence in assembling process to operator's subjectivity is equipped, misloading and neglected loading situation in assembling process are reduced.
Detailed description of the invention
It in order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, below will be to institute in embodiment
Attached drawing to be used is needed to be briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations of the invention
Example, for those of ordinary skill in the art, without any creative labor, can also be according to these attached drawings
Obtain other attached drawings.
Fig. 1 is the flow chart that sectional members of the embodiment of the present invention assemble error-preventing method;
Fig. 2 is the structural block diagram that sectional members of the embodiment of the present invention assemble fail-safe system.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
The object of the present invention is to provide a kind of sectional members assembly error-preventing method and systems, to information-based, automatic chemoprevention
There is mistake in sectional members assembly only.
In order to make the foregoing objectives, features and advantages of the present invention clearer and more comprehensible, with reference to the accompanying drawing and specific real
Applying mode, the present invention is described in further detail.
As shown in Figure 1, a kind of sectional members assembly error-preventing method the following steps are included:
Step 101: acquisition has the image of the sectional members to be assembled of artificial setting number from assembly environment.Make to set
Standby is a kind of Portable acquiring equipment convenient for using in erecting yard.The image of acquisition includes tri- channel informations of RGB
Color image.Portable acquiring equipment can include but is not limited to: tablet computer, mobile phone, digital camera, IP Camera
Equal image capture devices.
Step 102: identifying the number information in sectional members image to be assembled;The number information includes the classification of number
The position and;The classification includes number and letter;The position includes the position of the number and the position of the letter.Tool
Body:
Obtain identification model;
The sectional members image to be assembled is input to the identification model, obtains number information.For from containing compile
The number of composition number information or position and the classification of letter are detected and identified in the image of number information, and according to these elements
Topological relation obtain the number information for being subordinated to sectional members to be assembled in conjunction with sectional members coding rule.
Between the acquisition identification model described in step 102, further includes:
Training sample is obtained, the training sample is the image of 100 or more the sectional members with artificial setting number;
By 100 the above have the image of the sectional members of artificial setting number to convolutional neural networks model into
Row training is exported as a result, the output result is number information;
Whether the output result meets trained termination condition;The trained termination condition is correctly known in output result
The ratio of other total character quantity of character Zhan reaches 95% or more;
If so, determining that the convolutional neural networks model is identification model;
If it is not, adjusting the parameter of the convolutional neural networks model by back-propagation algorithm, make to export result satisfaction instruction
Practice termination condition.
The training process of convolutional neural networks model specifically includes:
The input that image is trained as this is randomly choosed from training sample;
It is positive in convolutional neural networks to the data of input successively to calculate, using the calculated result of output layer as volume
Product neural network is to the position of character in input picture and the prediction of classification.It specifically includes: the apex coordinate of the encirclement frame of character
With classification confidence score;
By the legitimate reading of prediction result and mark, loss function is calculated;(note: loss function is one and is artificially arranged
Function, the difference capableing of between the prediction result and legitimate reading of quantitative measurement convolutional neural networks, the value of loss function is about
Small, then prediction result and legitimate reading difference are smaller.)
Backpropagation is carried out, the parameter in convolutional neural networks model is updated;(note: backpropagation refers to from convolutional Neural
The output layer of network successively asks forward each network parameter to the partial derivative of loss function, and the direction that parameter updates is exactly the parameter
To the opposite direction of loss function partial derivative.)
Repeat aforesaid operations, until cycle-index is more than preset number, such as 120000 times.
The output of convolutional neural networks model is the encirclement frame and classification of each character, is closed according to the topology that character surrounds frame
System is that character surrounds relative position of the frame in picture, and the encirclement frame with a line is the kinds of characters belonged in a number, according to
This, which gets up the corresponding Connection operator of encirclement frame from left to right, is completely numbered.Since plating numerals have certain name
Rule, and the characters such as " ", "-" in numbering area very little shared in picture, are difficult to be convolved neural network model detection
It arrives, and position of these characters in number is fixed, therefore can be according to the naming rule of number in the fixation position of number
Directly add these characters.
Step 103: being retrieved from database according to the number information, determine matching number.In database
Sectional members number is matched, and is exactly matched if numbering with sectional members a certain in database, and retrieval unit exports this point
The number of section component.
Step 104: output assembling scheme is numbered according to the matching;The assembling scheme includes construction requirement, installation position
It sets and operating process.If exporting the highest 3 sectional members number of matching degree, this 3 volumes are shown on a display screen
Number, it allows user to select correctly matching number, then shows the assembling scheme of the corresponding sectional members of number of user's selection;
If output is the unique number exactly matched, the assembling scheme of the number corresponding segments component is shown on a display screen.
Step 105: sectional members to be assembled being assembled according to the assembling scheme.
The specific embodiment provided according to the present invention, the invention discloses following technical effects: the present invention identify first to
Assemble the number information in sectional members image;Then it is retrieved from database according to the number information, determines matching
Number;Output assembling scheme is numbered according to the matching;Sectional members to be assembled are assembled according to the assembling scheme.This
Sample can reduce the dependence in Large-Scale Equipment assembling process to operator's subjectivity, reduce the misloading in assembling process and neglected loading
Situation.
As shown in Fig. 2, the present invention also provides a kind of sectional members to assemble fail-safe system, the system comprises:
Image capture module 201, for the sectional members to be assembled that there is artificial setting to number for acquisition from assembly environment
Image.
Identification module 202, for identification number information in sectional members image to be assembled;The number information includes compiling
Number classification and position;The classification includes number and letter;The position include the number position and the letter
Position.
The identification module 202 specifically includes:
Acquiring unit, for obtaining identification model;
Input unit obtains number information for the sectional members image to be assembled to be input to the identification model.
Retrieval module 203 determines matching number for being retrieved from database according to the number information.
The retrieval module 203 specifically includes:
Retrieval unit obtains complete with the number information for being retrieved from database according to the number information
Complete consistent number is determined as matching number.
Output module 204, for numbering output assembling scheme according to the matching;The assembling scheme includes that construction is wanted
It asks, installation site and operating process.
Load module 205, for being assembled according to the assembling scheme to sectional members to be assembled.
The assembly fail-safe system further includes further include:
Sample acquisition module, for obtaining training sample, the training sample, which is 100 or more, has artificial setting number
Sectional members image;
Training module, for having the image of the sectional members of artificial setting number to convolution by 100 the above
Neural network model is trained, and is exported as a result, the output result is number information;
Judgment module, for judging whether the output result meets trained termination condition;The trained termination condition is
The ratio for the total character quantity of character Zhan being correctly validated in output result reaches 95% or more;
As a result determining module, for determining the convolutional Neural net when the output result meets training termination condition
Network model is identification model;
Module is adjusted, for being adjusted by back-propagation algorithm when the output result is unsatisfactory for training termination condition
The parameter of the convolutional neural networks model makes to export result satisfaction training termination condition.
Each embodiment in this specification is described in a progressive manner, the highlights of each of the examples are with other
The difference of embodiment, the same or similar parts in each embodiment may refer to each other.For system disclosed in embodiment
For, since it is corresponded to the methods disclosed in the examples, so being described relatively simple, related place is said referring to method part
It is bright.
Used herein a specific example illustrates the principle and implementation of the invention, and above embodiments are said
It is bright to be merely used to help understand method and its core concept of the invention;At the same time, for those skilled in the art, foundation
Thought of the invention, there will be changes in the specific implementation manner and application range.In conclusion the content of the present specification is not
It is interpreted as limitation of the present invention.
Claims (8)
1. a kind of sectional members assemble error-preventing method, which is characterized in that the described method includes:
Acquisition has the image of the sectional members to be assembled of artificial setting number from assembly environment;The artificial setting number
Method include spraying, engraving and it is hand-written;
Identify the number information in sectional members image to be assembled;The number information includes classification and the position of number;It is described
Classification includes number and letter;The position includes the position of the number and the position of the letter;
It is retrieved from database according to the number information, determines matching number;
Output assembling scheme is numbered according to the matching;The assembling scheme includes construction requirement, installation site and operation stream
Journey;
Sectional members to be assembled are assembled according to the assembling scheme.
2. sectional members according to claim 1 assemble error-preventing method, which is characterized in that the identification segmentation structure to be assembled
Number information in part image, specifically includes:
Obtain identification model;
The sectional members image to be assembled is input to the identification model, obtains number information.
3. sectional members according to claim 2 assemble error-preventing method, which is characterized in that the acquisition identification model it
Between, further includes:
Training sample is obtained, the training sample is the image of 100 or more the sectional members with artificial setting number;
By 100 the above there is the image of the sectional members of artificial setting number to instruct to convolutional neural networks model
Practice, is exported as a result, the output result is number information;
Judge whether the output result meets trained termination condition;The trained termination condition is correctly known in output result
The ratio of other total character quantity of character Zhan reaches 95% or more;
If so, determining that the convolutional neural networks model is identification model;
If it is not, adjusting the parameter of the convolutional neural networks model by back-propagation algorithm, make to export the training of result satisfaction eventually
Only condition.
4. sectional members according to claim 1 assemble error-preventing method, which is characterized in that described according to the number information
It is retrieved from database, determines matching number, specifically include:
It is retrieved from database according to the number information, obtains the number completely the same with the number information, determined
For matching number.
5. a kind of sectional members assemble fail-safe system, which is characterized in that the system comprises:
Image capture module, the image for the sectional members to be assembled that there is artificial setting to number for acquisition from assembly environment;
The method of the artificial setting number includes spraying, engraving and hand-written;
Identification module, for identification number information in sectional members image to be assembled;The number information includes the class of number
Other and position;The classification includes number and letter;The position includes the position of the number and the position of the letter;
Retrieval module determines matching number for being retrieved from database according to the number information;
Output module, for numbering output assembling scheme according to the matching;The assembling scheme includes construction requirement, installation position
It sets and operating process;
Load module, for being assembled according to the assembling scheme to sectional members to be assembled.
6. sectional members according to claim 5 assemble fail-safe system, which is characterized in that the identification module specifically wraps
It includes:
Acquiring unit, for obtaining identification model;
Input unit obtains number information for the sectional members image to be assembled to be input to the identification model.
7. sectional members according to claim 5 assemble fail-safe system, which is characterized in that the assembly fail-safe system also wraps
It includes:
Sample acquisition module, for obtaining training sample, the training sample is 100 or more points with artificial setting number
The image of section component;
Training module, for having the image of the sectional members of artificial setting number to convolutional Neural by 100 the above
Network model is trained, and is exported as a result, the output result is number information;
Judgment module, for judging whether the output result meets trained termination condition;The trained termination condition is output
As a result the ratio for the total character quantity of character Zhan being correctly validated in reaches 95% or more;
As a result determining module, for determining the convolutional neural networks mould when the output result meets training termination condition
Type is identification model;
Adjust module, for when the output result is unsatisfactory for trained termination condition, adjusted by back-propagation algorithm described in
The parameter of convolutional neural networks model makes to export result satisfaction training termination condition.
8. sectional members according to claim 5 assemble fail-safe system, which is characterized in that the retrieval module is specifically wrapped
It includes:
Retrieval unit obtains and the number information complete one for being retrieved from database according to the number information
The number of cause is determined as matching number.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910089031.1A CN109815250A (en) | 2019-01-30 | 2019-01-30 | A kind of sectional members assembly error-preventing method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910089031.1A CN109815250A (en) | 2019-01-30 | 2019-01-30 | A kind of sectional members assembly error-preventing method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109815250A true CN109815250A (en) | 2019-05-28 |
Family
ID=66605793
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910089031.1A Pending CN109815250A (en) | 2019-01-30 | 2019-01-30 | A kind of sectional members assembly error-preventing method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109815250A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111604281A (en) * | 2020-04-13 | 2020-09-01 | 天津中车机辆装备有限公司 | Part sorting method and system for rail transit |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106709486A (en) * | 2016-11-11 | 2017-05-24 | 南京理工大学 | Automatic license plate identification method based on deep convolutional neural network |
US20180025256A1 (en) * | 2015-10-20 | 2018-01-25 | Tencent Technology (Shenzhen) Company Limited | Method and apparatus for recognizing character string in image |
-
2019
- 2019-01-30 CN CN201910089031.1A patent/CN109815250A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180025256A1 (en) * | 2015-10-20 | 2018-01-25 | Tencent Technology (Shenzhen) Company Limited | Method and apparatus for recognizing character string in image |
CN106709486A (en) * | 2016-11-11 | 2017-05-24 | 南京理工大学 | Automatic license plate identification method based on deep convolutional neural network |
Non-Patent Citations (1)
Title |
---|
孙静等: "《防差错技术在舰船建造过程的应用分析与建议》", 《质量与可靠性》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111604281A (en) * | 2020-04-13 | 2020-09-01 | 天津中车机辆装备有限公司 | Part sorting method and system for rail transit |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109408384B (en) | Software application testing method and device, processor and electronic device | |
CN110059694B (en) | Intelligent identification method for character data in complex scene of power industry | |
KR102220174B1 (en) | Learning-data enhancement device for machine learning model and method for learning-data enhancement | |
CN105808416B (en) | A kind of automated testing method and system of man-machine figure interactive interface | |
CN109472193A (en) | Method for detecting human face and device | |
CN110348522A (en) | A kind of image detection recognition methods and system, electronic equipment, image classification network optimized approach and system | |
CN106021101A (en) | Method and device for testing mobile terminal | |
CN111160432A (en) | Automatic classification method and system for panel production defects | |
CN109710528A (en) | A kind of test script generation method, device, equipment and medium | |
CN113723300A (en) | Artificial intelligence-based fire monitoring method and device and storage medium | |
CN105577472A (en) | Data acquisition test method and device | |
CN110866473B (en) | Target object tracking detection method and device, storage medium and electronic device | |
CN112527676A (en) | Model automation test method, device and storage medium | |
CN116597413A (en) | Real-time traffic sign detection method based on improved YOLOv5 | |
CN114724140A (en) | Strawberry maturity detection method and device based on YOLO V3 | |
CN116563762A (en) | Fire detection method, system, medium, equipment and terminal for oil and gas station | |
CN108415886A (en) | A kind of data label error correction method and device based on production process | |
CN112989768B (en) | Method and device for correcting connection questions, electronic equipment and storage medium | |
CN109815250A (en) | A kind of sectional members assembly error-preventing method and system | |
CN112529548B (en) | Online education platform cloud resource management system | |
CN117437227A (en) | Image generation and defect detection method, device, medium, equipment and product | |
CN106910207B (en) | Method and device for identifying local area of image and terminal equipment | |
CN117115614A (en) | Object identification method, device, equipment and storage medium for outdoor image | |
CN111539390A (en) | Small target image identification method, equipment and system based on Yolov3 | |
CN112149698A (en) | Method and device for screening difficult sample data |
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: 20190528 |