CN113606502B - Method for judging whether operator performs pipeline air leakage detection based on machine vision - Google Patents
Method for judging whether operator performs pipeline air leakage detection based on machine vision Download PDFInfo
- Publication number
- CN113606502B CN113606502B CN202110803882.5A CN202110803882A CN113606502B CN 113606502 B CN113606502 B CN 113606502B CN 202110803882 A CN202110803882 A CN 202110803882A CN 113606502 B CN113606502 B CN 113606502B
- Authority
- CN
- China
- Prior art keywords
- detection model
- detection
- pipeline
- model
- machine vision
- 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.)
- Active
Links
Images
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F17—STORING OR DISTRIBUTING GASES OR LIQUIDS
- F17D—PIPE-LINE SYSTEMS; PIPE-LINES
- F17D5/00—Protection or supervision of installations
- F17D5/02—Preventing, monitoring, or locating loss
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F17—STORING OR DISTRIBUTING GASES OR LIQUIDS
- F17D—PIPE-LINE SYSTEMS; PIPE-LINES
- F17D3/00—Arrangements for supervising or controlling working operations
- F17D3/01—Arrangements for supervising or controlling working operations for controlling, signalling, or supervising the conveyance of a product
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Examining Or Testing Airtightness (AREA)
Abstract
The invention provides a method for judging whether an operator executes pipeline air leakage detection based on machine vision, and relates to the technical field of pipeline transportation safety and intelligent identification. The method comprises the following steps: acquiring field picture data, and preprocessing the field picture data, including zooming, clipping, contrast adjustment, brightness adjustment, noise addition, picture splicing and the like; constructing a pipeline connection detection model, and training the model; constructing a personnel detection model, and training the model; constructing a gas monitor detection model, and training the model; and (3) connecting the trained model into a monitoring camera, sequentially connecting the detection model, the personnel detection model and the gas monitor detection model through pipelines for processing, and judging whether a behavior for executing gas leakage detection exists or not through post-processing and judging. The detection model is constructed by utilizing a convolution layer, a batch normalization layer, an activation layer, a pooling layer and a residual block. The method has the advantages of high response speed, safe operation and high efficiency.
Description
Technical Field
The invention relates to the technical field of pipeline transportation safety and intelligent identification, in particular to the technical field of pipeline transportation safety and intelligent identification.
Background
With the economic development, the demand for energy is increased, and the use of natural gas as a new energy source is also increasingly widespread. Loading and unloading using LNG tankers is also becoming increasingly frequent as an important means of transporting natural gas. The air leak detection must occur after the pipe connection and before the gas is delivered. Due to the fact that the LNG loading and unloading procedures are multiple, operation is prone to leakage only through operators, if leakage occurs, serious safety accidents such as explosion and the like can be caused, and therefore monitoring measures are needed.
In the prior art, a monitoring means is generally adopted, such as on-site supervision by a specially-assigned person, more manpower is consumed, and the monitoring personnel are single in work and cannot reasonably utilize the manpower. And uploading the photo which is executing the gas detection action to a server, and randomly extracting the picture for auditing by inspectors, but the method has insufficient timeliness, can not prevent the violation action in advance, and can not realize efficient and accurate control. The automatic real-time judgment of whether an operator executes the pipeline air leakage detection behavior is possible along with the appearance of technologies such as artificial intelligence and deep learning.
Disclosure of Invention
In order to ensure the unloading safety of the LNG tank car, determine the detection of the air leakage of the pipeline, ensure the response speed and the operation safety and improve the monitoring efficiency, the invention provides a method for judging whether an operator executes the detection of the air leakage of the pipeline based on machine vision, and the specific technical scheme is as follows.
A method for judging that an operator performs pipeline air leakage detection based on machine vision comprises the following steps:
s1, collecting field picture data;
s2, preprocessing the picture data, wherein the preprocessing comprises zooming, cutting, contrast adjustment, brightness adjustment, noise reduction and picture splicing operation;
s3, constructing a pipeline connection detection model, and training the pipeline connection detection model;
s4, constructing a personnel detection model, and training the personnel detection model;
s5, constructing a gas monitor detection model, and training the gas monitor detection model;
s6, connecting the trained model into a monitoring camera, and sequentially connecting a detection model, a personnel detection model and a gas monitor detection model through pipelines for processing;
and S7, judging whether the behavior of executing the air leakage detection exists or not through post-processing, and judging.
Preferably, the pipeline connection detection model, the personnel detection model and the gas monitor detection model are constructed by utilizing a convolution layer, a batch normalization layer, an activation layer, a pooling layer and a residual block.
Preferably, the batch normalization layer normalizes each batch of data, and normalizes at any layer in the network.
It is also preferable that the activation layer performs an activation operation on the input data, and the sizes of the input and output data are equal during the operation.
More preferably, the pooling layer is provided between successive convolutional layers, and the image is compressed to remove part of the information and maintain the image scale.
Further preferably, the residual block acquires the attention feature map from the given intermediate feature map along two dimensions, adjusts the input intermediate feature map, and reinforces the useful information to suppress the useless information from keeping the sizes of the input and the output consistent.
It is further preferred that the field picture data comprises an LNG unloading field picture.
Further preferably, the behavior of detecting the air leakage is judged by first performing detection judgment of hose connection, and then performing personnel detection judgment and gas monitor detection judgment respectively after detecting the hose connection.
The method for judging whether the operator executes the detection of the air leakage of the pipeline based on the machine vision has the advantages that: the method can effectively detect the air leakage detection condition of the pipeline in the LNG unloading process, and has the advantages of quick response, good real-time performance, timely return of the behavior judgment result and high efficiency; in addition, the method utilizes the camera to carry out detection and judgment, thereby ensuring the objectivity and timeliness of monitoring, fully automatically processing the video stream, needing no manual supervision and having simple and convenient operation. By using the method, when the operator does not perform the air leakage detection behavior, the operator can be prevented from carrying air, and accidents are reduced.
Drawings
In order to more clearly illustrate the embodiments or technical solutions of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained based on these drawings without creative efforts.
FIG. 1 is a flow chart illustrating a method for determining that an operator performs detection of air leakage in a pipeline based on machine vision;
fig. 2 is a block diagram of the detection step.
Detailed Description
With reference to fig. 1 and fig. 2, a specific embodiment of a method for determining that an operator performs detection of a pipe leakage based on machine vision according to the present invention is described.
When LNG unloads the car, in order to judge whether the staff carries out the problem of pipeline gas leakage detection action, provide a method that whether to carry out pipeline gas leakage detection based on machine vision judgement operating personnel, solve many problems that exist among the artificial detection process, the concrete operating procedure of this method includes:
s1, collecting field picture data.
The method specifically comprises the steps that image data of an LNG unloading site can be collected to obtain an original data set, and a database is established; the scheme should include the picture of the pipeline connection at each angle and the picture of personnel detecting the pipeline leakage.
And S2, preprocessing the picture data, wherein the preprocessing comprises operations of zooming, cutting, contrast adjustment, brightness adjustment, noise reduction and picture splicing.
The picture for detecting the pipeline leakage comprises the zooming processing of the image in the video, the cutting, contrast adjustment and brightness adjustment of the image, image splicing operation and other required image processing modes.
And S3, constructing a pipeline connection detection model, and training the pipeline connection detection model.
And S4, constructing a personnel detection model, and training the personnel detection model.
And S5, constructing a gas monitor detection model, and training the gas monitor detection model.
The detection model in the step is specifically to carry out detection and judgment by operating the following steps, wherein unloading personnel sequentially remove a hose blind plate, a liquid loading and discharging pipe, a pressurized gas phase pipe and a liquid phase pipe, open a pressurized liquid phase valve, open a diffusing pipe, pressurize and pre-cool, observe whether the LNG pressure of the tank car is 0.5-0.6MPa, and open a liquid inlet valve of the unloading car; and finally, opening the liquid outlet valve by the unloader to unload the liquid. The unloading personnel sequentially close the liquid outlet valve of the tank car, close the liquid inlet pipeline valve, release and discharge pressure to the liquid outlet pipeline of the tank car, open the BOG pipeline valve, start the BOG compressor to recover the residual pressure of the tank car, observe whether the pressure of the tank car is reduced to about 0.3MPa, and disassemble the liquid outlet pipe of the tank car and the pressurized gas-phase and liquid-phase pipe; then the unloading personnel place the liquid outlet hose into the containing box.
Whether personnel utilize the gas monitoring appearance to detect this process again after the pipe connection is accomplished to guarantee the safety of the in-process pipeline of unloading, avoid the problem that artificial omission, false positive etc. brought.
And S6, connecting the trained model into a monitoring camera, and sequentially connecting the detection model, the personnel detection model and the gas monitor detection model through pipelines for processing.
Specifically, pictures acquired by a monitoring camera are directly sent into a pipeline connection detection model without being preprocessed, and if the hose is detected to be connected and the position of the hose is obtained, the pictures are respectively sent into a personnel detection model and a gas monitor detection model to detect whether personnel and a gas detector exist or not and the position of the personnel and the gas detector.
And S7, judging whether a behavior of executing the gas leakage detection exists by post-processing, and judging. If the hose connection, the personnel and the gas detector exist, performing post-processing to judge whether the personnel performs the action of gas leakage detection according to the positions of the personnel, the gas detector and the pipeline connection part obtained by detection; otherwise there is no leak detection behavior.
The pipeline connection detection model, the personnel detection model and the gas monitor detection model are constructed by utilizing a convolution layer, a batch normalization layer, an activation layer, a pooling layer and a residual block. When the convolutional neural network is used for image recognition, the image data subjected to conversion is input, and the depth is the number of bits used for storing each pixel in the image, such as a color image. The batch normalization layer normalizes each batch of data and normalizes at any layer in the network. The activation layer performs activation operation on input data, and the sizes of the input data and the output data are equal in the operation process. The pooling layer is arranged in the middle of the continuous convolution layer, and compresses the image to remove part of information and keep the image dimension. And the residual block acquires the attention characteristic diagram from the given intermediate characteristic diagram along two dimensions, adjusts the input intermediate characteristic diagram, strengthens useful information, inhibits useless information and keeps the sizes of input and output consistent.
The field picture data comprises an LNG unloading field picture. And (3) judging the behavior of air leakage detection, namely firstly detecting and judging hose connection, and then respectively detecting and judging personnel and a gas monitor after detecting the hose connection.
The method can effectively detect the pipeline air leakage detection condition in the LNG unloading process, and has the advantages of quick response, good real-time performance, timely return of behavior judgment results and high efficiency; in addition, the method utilizes the camera to carry out detection and judgment, thereby ensuring the objectivity and timeliness of monitoring, fully automatically processing the video stream, needing no manual supervision and having simple and convenient operation. By using the method, when the operator does not perform the air leakage detection behavior, the operator can be prevented from carrying air, and accidents are reduced.
It is to be understood that the above description is not intended to limit the present invention, and the present invention is not limited to the above examples, and those skilled in the art may make modifications, alterations, additions or substitutions within the spirit and scope of the present invention.
Claims (8)
1. A method for judging whether an operator performs pipeline air leakage detection based on machine vision is characterized by comprising the following steps:
s1, collecting field picture data;
s2, preprocessing the picture data, wherein the preprocessing comprises operations of zooming, cutting, contrast adjustment, brightness adjustment, noise reduction and picture splicing;
s3, constructing a pipeline connection detection model, and training the pipeline connection detection model;
s4, constructing a personnel detection model, and training the personnel detection model;
s5, constructing a gas monitor detection model, and training the gas monitor detection model;
s6, connecting the trained pipeline connection detection model, personnel detection model and gas monitor detection model into a monitoring camera, and sequentially processing through the pipeline connection detection model, the personnel detection model and the gas monitor detection model;
and S7, judging whether the behavior of executing the air leakage detection exists or not through post-processing, and judging.
2. The method for determining whether an operator performs pipeline leakage detection based on machine vision of claim 1, wherein the pipeline connection detection model, the personnel detection model and the gas monitor detection model are constructed by using a convolutional layer, a batch normalization layer, an activation layer, a pooling layer and a residual block.
3. The method for determining whether the operator performs the pipeline air leakage detection based on the machine vision according to claim 2, wherein the batch normalization layer normalizes each batch of data and normalizes any layer in a network.
4. The method for judging whether an operator performs pipeline air leakage detection based on machine vision according to claim 2, wherein the activation layer performs activation operation on input data, and the input data and the output data are equal in size in the operation process.
5. The method for determining whether an operator performs pipeline air leakage detection based on machine vision according to claim 2, wherein the pooling layer is arranged among the continuous convolution layers, and the image is compressed to remove part of information, so that the image scale is maintained.
6. The method for determining whether the operator performs the detection of the air leakage of the pipeline based on the machine vision as claimed in claim 2, wherein the residual block acquires the attention feature map from the given intermediate feature map along two dimensions, and adjusts the input intermediate feature map to reinforce the useful information and suppress the useless information to keep the sizes of the input and the output consistent.
7. The method for determining whether an operator performs a pipeline leakage test based on machine vision according to claim 1, wherein the field picture data comprises an LNG unloading field picture.
8. The method for machine vision-based determination of the performance of a duct leak detection by an operator according to claim 1, wherein the determination of the performance of a leak detection, firstly, hose connection detection and judgment are carried out, and personnel detection and judgment and gas monitor detection and judgment are respectively carried out after hose connection is detected.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110803882.5A CN113606502B (en) | 2021-07-16 | 2021-07-16 | Method for judging whether operator performs pipeline air leakage detection based on machine vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110803882.5A CN113606502B (en) | 2021-07-16 | 2021-07-16 | Method for judging whether operator performs pipeline air leakage detection based on machine vision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113606502A CN113606502A (en) | 2021-11-05 |
| CN113606502B true CN113606502B (en) | 2023-03-24 |
Family
ID=78337668
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110803882.5A Active CN113606502B (en) | 2021-07-16 | 2021-07-16 | Method for judging whether operator performs pipeline air leakage detection based on machine vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113606502B (en) |
Family Cites Families (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB201711412D0 (en) * | 2016-12-30 | 2017-08-30 | Maxu Tech Inc | Early entry |
| CN106845437A (en) * | 2017-02-13 | 2017-06-13 | 常州大学 | Leakage of City Gas Pipeline localization method based on Support vector regression |
| CN107013811B (en) * | 2017-04-12 | 2018-10-09 | 武汉科技大学 | A kind of pipeline liquid leakage monitoring method based on image procossing |
| CN108470143A (en) * | 2018-01-31 | 2018-08-31 | 衡阳泰豪通信车辆有限公司 | A kind of pipeline unmanned plane patrolling method and system |
| CN109373190B (en) * | 2018-10-28 | 2020-07-03 | 北京工业大学 | Buried steel pipeline damage full-tensor geomagnetic detection system and implementation method |
| CN109240311B (en) * | 2018-11-19 | 2021-06-25 | 国网四川省电力公司电力科学研究院 | Outdoor electric power field construction operation supervision method based on intelligent robot |
| CN109931506A (en) * | 2019-03-14 | 2019-06-25 | 三川智慧科技股份有限公司 | Pipeline leakage detection method and device |
| CN109945076A (en) * | 2019-04-11 | 2019-06-28 | 南京中禹智慧水利研究院有限公司 | A kind of pipeline silting water detection system based on Machine Vision Detection |
| CN110145692B (en) * | 2019-05-23 | 2021-03-26 | 昆山华新建设工程有限公司 | Sewage pipeline CCTV detection system and method |
| CN111553265B (en) * | 2020-04-27 | 2021-10-29 | 河北天元地理信息科技工程有限公司 | Method and system for detecting internal defects of drainage pipeline |
| CN112329588B (en) * | 2020-10-30 | 2024-01-05 | 中海石油(中国)有限公司 | Pipeline fault detection method based on Faster R-CNN |
| CN112762362A (en) * | 2021-01-15 | 2021-05-07 | 中国海洋石油集团有限公司 | Underwater pipeline leakage acoustic emission detection method based on convolutional neural network |
-
2021
- 2021-07-16 CN CN202110803882.5A patent/CN113606502B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN113606502A (en) | 2021-11-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110456723A (en) | A kind of emptying area, gas station security management and control system based on deep learning | |
| CN110264440B (en) | Large-scale train displacement fault detection method and system based on deep learning | |
| CN107782733A (en) | Image recognition the cannot-harm-detection device and method of cracks of metal surface | |
| CN111932709A (en) | Method for realizing violation safety supervision of inspection operation of gas station based on AI identification | |
| CN113606502B (en) | Method for judging whether operator performs pipeline air leakage detection based on machine vision | |
| CN111126196A (en) | Equipment oil leakage detection method | |
| CN111178198B (en) | Automatic monitoring method for potential safety hazards of laboratory dangerous goods based on machine vision | |
| CN114298384A (en) | Safe operation and maintenance prediction system and method suitable for ship loading and unloading arm | |
| CN110779077A (en) | Intelligent geothermal operation and maintenance system | |
| CN117686033A (en) | Intelligent recognition system for coal mine robot | |
| CN118197027A (en) | Unmanned aerial vehicle scheduling checking method based on pipeline early warning and related device | |
| CN112146909A (en) | Fault diagnosis method and device for refrigeration equipment, readable storage medium and computing equipment | |
| CN116432078A (en) | Building electromechanical equipment monitoring system | |
| CN116363773A (en) | Vision-based ship cabin itinerant monitoring system and method | |
| CN115909191A (en) | Analysis method for gas leakage detection behavior of gas station operator | |
| CN116164237A (en) | Long oil and gas pipeline management system based on data management | |
| CN116310953A (en) | Pontoon off-grid intelligent detection method of cold source interception net | |
| CN112483904B (en) | Pipeline monitoring equipment and monitoring method for hydrogen system | |
| CN213838865U (en) | CNG compressor remote diagnosis system | |
| CN115082650A (en) | Implementation method of automatic pipeline defect labeling tool based on convolutional neural network | |
| CN111986051A (en) | Method for realizing safety supervision of gas station equipment | |
| CN112295947A (en) | Detection system for flow valve machining and detection method thereof | |
| CN216253056U (en) | Hydraulic station monitoring system | |
| CN213397065U (en) | Visual detection device for vehicle-mounted air conditioner | |
| RU2786513C1 (en) | Automated system for control and regulation of flow rate of fuel gas of gas-pumping unit |
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 |