CN116379359A - Natural gas leakage detection method and multi-mode natural gas leakage detection system - Google Patents

Natural gas leakage detection method and multi-mode natural gas leakage detection system Download PDF

Info

Publication number
CN116379359A
CN116379359A CN202310367612.3A CN202310367612A CN116379359A CN 116379359 A CN116379359 A CN 116379359A CN 202310367612 A CN202310367612 A CN 202310367612A CN 116379359 A CN116379359 A CN 116379359A
Authority
CN
China
Prior art keywords
natural gas
detection
optical fiber
monitoring
gas leakage
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202310367612.3A
Other languages
Chinese (zh)
Other versions
CN116379359B (en
Inventor
李逸梅
闫印强
姚兴仁
杨利达
李敏
孙俊虎
姜海昆
范宇
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Changyang Technology Beijing Co ltd
Original Assignee
Changyang Technology Beijing Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Changyang Technology Beijing Co ltd filed Critical Changyang Technology Beijing Co ltd
Priority to CN202310367612.3A priority Critical patent/CN116379359B/en
Publication of CN116379359A publication Critical patent/CN116379359A/en
Application granted granted Critical
Publication of CN116379359B publication Critical patent/CN116379359B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
    • G08B21/02—Alarms for ensuring the safety of persons
    • G08B21/12—Alarms for ensuring the safety of persons responsive to undesired emission of substances, e.g. pollution alarms
    • G08B21/16—Combustible gas alarms
    • 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
    • F17D5/005—Protection or supervision of installations of gas pipelines, e.g. alarm
    • 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
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
    • G01M3/00—Investigating fluid-tightness of structures
    • G01M3/002—Investigating fluid-tightness of structures by using thermal means
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
    • G01M3/00—Investigating fluid-tightness of structures
    • G01M3/02—Investigating fluid-tightness of structures by using fluid or vacuum
    • G01M3/04—Investigating fluid-tightness of structures by using fluid or vacuum by detecting the presence of fluid at the leakage point
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
    • G01M3/00—Investigating fluid-tightness of structures
    • G01M3/02—Investigating fluid-tightness of structures by using fluid or vacuum
    • G01M3/04—Investigating fluid-tightness of structures by using fluid or vacuum by detecting the presence of fluid at the leakage point
    • G01M3/24—Investigating fluid-tightness of structures by using fluid or vacuum by detecting the presence of fluid at the leakage point using infrasonic, sonic, or ultrasonic vibrations
    • G01M3/243—Investigating fluid-tightness of structures by using fluid or vacuum by detecting the presence of fluid at the leakage point using infrasonic, sonic, or ultrasonic vibrations for pipes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/776—Validation; Performance evaluation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content
    • G06V20/44—Event detection

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Multimedia (AREA)
  • Computing Systems (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Mechanical Engineering (AREA)
  • Combustion & Propulsion (AREA)
  • Business, Economics & Management (AREA)
  • Chemical & Material Sciences (AREA)
  • Environmental & Geological Engineering (AREA)
  • Toxicology (AREA)
  • Emergency Management (AREA)
  • Examining Or Testing Airtightness (AREA)

Abstract

The invention relates to the technical field of natural gas leakage detection, in particular to a natural gas leakage detection method and a multi-mode natural gas leakage detection system, wherein the method comprises the following steps: acquiring infrared video images of at least two continuous frames of areas to be detected; after the obtained two continuous frames of infrared video images are converted to a preset size, inputting the two continuous frames of infrared video images into a trained natural gas detection model for detection, and respectively obtaining corresponding detection results; the natural gas detection model is constructed by taking a yolov5 model frame as a main body, a C2f network structure is used for replacing a C3 network structure, and a software operation is used for replacing convolution operation for 2 times of downsampling to perform characteristic downsampling; based on the detection results corresponding to the two continuous frames of infrared video images, judging whether the detection results are false results, otherwise, judging that the natural gas leakage is detected, and outputting the detection results. The invention can effectively solve the problems of low intelligent degree, low detection efficiency and difficult positioning of the traditional natural gas leakage detection technology.

Description

Natural gas leakage detection method and multi-mode natural gas leakage detection system
Technical Field
The embodiment of the invention relates to the technical field of natural gas leakage detection, in particular to a natural gas leakage detection method, a multi-mode natural gas leakage detection system, electronic equipment and a storage medium.
Background
Natural gas is a flammable mixture, the main component being methane gas. The leakage of natural gas can cause explosion, so that the important property loss is caused, and the resource waste is also caused. Therefore, the method has great significance in detecting the leakage of the natural gas and locating the leakage points.
At present, the traditional natural gas leakage detection method generally uses a sniffer technology and a probe to detect, needs to closely contact a detected area, and has low safety when a detector is exposed to an environment with potential explosion risk, which cannot be seen by naked eyes. Meanwhile, in the area where natural gas leakage is found, the traditional leakage detection and investigation means are difficult to quickly locate specific leakage points, and are common pain points in the industry.
Disclosure of Invention
Based on the problem that the positioning of the position of the natural gas leakage is difficult to detect in the prior art, the embodiment of the invention provides a natural gas leakage detection method, which can efficiently and quickly position a gas leakage area based on infrared video images of two continuous frames of areas to be detected.
In a first aspect, an embodiment of the present invention provides a method for detecting leakage of natural gas, including:
acquiring infrared video images of at least two continuous frames of areas to be detected;
After the obtained two continuous frames of infrared video images are converted to a preset size, inputting the two continuous frames of infrared video images into a trained natural gas detection model for detection, and respectively obtaining corresponding detection results; the natural gas detection model is constructed by taking a yolov5 model frame as a main body, a C2f network structure is used for replacing a C3 network structure, a software operation is used for replacing convolution operation for 2 times of downsampling to perform feature downsampling, and the natural gas detection model is obtained by taking a sample natural gas infrared image with a preset size as input and a corresponding gas leakage area as output training;
based on the detection results corresponding to the two continuous frames of infrared video images, judging whether the detection results are false results, otherwise, judging that the natural gas leakage is detected, and outputting the detection results.
Optionally, the natural gas detection model is trained by:
acquiring a natural gas infrared image data set with a label, and carrying out data enhancement to obtain a data set formed by a sample natural gas infrared image; the tag is used for correspondingly marking a gas leakage area in the image;
the enhanced data set is then processed according to 6:2:2, dividing the proportion to obtain a training set, a verification set and a test set;
Constructing a yolov5 model frame, replacing a C3 network structure in the model frame with a C2f network structure, and replacing convolution operation for 2 times downsampling with SoftPool operation to obtain a natural gas detection model to be trained;
and training, verifying and testing the constructed natural gas detection model by using the obtained training set, verification set and test set to obtain a trained natural gas detection model.
Optionally, the determining whether the detection result is a false result based on the detection results corresponding to the two continuous frames of infrared video images includes:
judging whether detection results corresponding to two continuous frames of infrared video images both comprise a gas leakage area or not;
if not, the natural gas leakage is not detected;
if yes, ioU between the gas leakage areas corresponding to the two continuous frames of infrared video images is calculated; if IoU is greater than 0.5 but less than 1.0, then determining that natural gas leakage is detected, otherwise determining that a false result is obtained; wherein, ioU has the expression:
Figure BDA0004167393740000021
(x1 l ，y1 u ) And (x 1) r ，y1 d ) Respectively an upper left corner coordinate and a lower right corner coordinate of a frame of infrared video corresponding to the gas leakage area, (x 2) l ，y2 u ) And (x 2) r ，y2 d ) And the coordinates of the upper left corner and the lower right corner of the gas leakage area corresponding to the other frame of infrared video are respectively.
In a second aspect, an embodiment of the present invention further provides a multi-mode natural gas leakage detection system, including:
the system comprises an optical fiber sound wave monitoring subsystem, a photoelectric signal monitoring subsystem, a movable optical image detection subsystem and a multi-mode analysis platform; wherein, the liquid crystal display device comprises a liquid crystal display device,
the optical fiber acoustic wave monitoring subsystem comprises at least one group of optical fiber monitoring modules, and each group of optical fiber monitoring modules corresponds to one section of natural gas pipe respectively; the optical fiber monitoring module comprises two sections of auxiliary optical fibers, at least one section of monitoring optical fiber, an optical fiber pulse emitter and an optical fiber sensor, wherein the two sections of auxiliary optical fibers are tightly wound at two ends of a natural gas pipe section respectively, the monitoring optical fibers are spirally wound on a main body of the natural gas pipe section at preset pitches, each section of monitoring optical fiber is wound on the natural gas pipe section in a non-overlapping manner, one optical fiber pulse emitter is connected to one end of each section of auxiliary optical fiber and one end of each monitoring optical fiber, and the other end of each section of auxiliary optical fiber is connected to the corresponding optical fiber sensor; each section of monitoring optical fiber is also provided with a corresponding positioner which is arranged on the natural gas pipe section of the monitoring optical fiber winding area; the optical fiber monitoring module further comprises a signal transceiver, and the signal transceiver is connected with each optical fiber pulse emitter, each optical fiber sensor and each positioner which are included in the optical fiber monitoring module;
The photoelectric signal monitoring subsystem comprises at least one group of photoelectric monitoring modules, and each group of photoelectric monitoring modules corresponds to one pipeline valve; the photoelectric monitoring module comprises a laser emitter and a photoelectric sensor which are respectively arranged at two sides of the pipeline valve, and a laser beam emitted by the laser emitter passes through a space where a leakage point of the pipeline valve is easily located and then enters the photoelectric sensor; each group of photoelectric monitoring modules is also provided with a corresponding locator and a signal transceiver, the corresponding locator and the signal transceiver are arranged at the corresponding pipeline valve, and the signal transceiver is connected with the laser transmitter, the photoelectric sensor and the locator which are included by the photoelectric monitoring modules;
the movable optical image detection subsystem includes at least one set of movement detection modules; each group of movement detection modules comprises a movement device, an optical gas camera arranged on the movement device, a positioner and a signal transceiver; the movable device is used for driving the optical gas camera to move, the optical gas camera is used for shooting an infrared image so as to detect a gas leakage area through environmental temperature change, and the signal transceiver is connected with the movable device, the optical gas camera and the positioner which are included in the movable optical image detection subsystem;
The multimode analysis platform is connected with the optical fiber sound wave monitoring subsystem, the photoelectric signal monitoring subsystem and the movable optical image detection subsystem through signal transceivers;
the multimode analysis platform is used for acquiring detection signals of all the optical fiber sensors in the optical fiber acoustic wave monitoring subsystem and monitoring whether natural gas leakage occurs in a natural gas pipe section or not, the monitoring mode is that the average value of detection signals corresponding to the two ends of the auxiliary optical fibers is subtracted from detection signals corresponding to one section of the monitoring optical fiber in each optical fiber monitoring module, so that a denoised monitoring signal is obtained, demodulation is carried out to determine whether a high-frequency vibration sound source exists or not, and if the high-frequency vibration sound source exists, the natural gas leakage occurs in the corresponding natural gas pipe section;
the multi-mode analysis platform is further used for acquiring detection signals of the photoelectric sensors in the photoelectric signal monitoring subsystem and monitoring whether a natural gas leakage event occurs to a pipeline valve or not, the monitoring mode is that the detection signals corresponding to the photoelectric sensors in the photoelectric monitoring modules are demodulated to determine whether gas component changes occur, and if the gas component changes occur, the natural gas leakage event occurs to the corresponding pipeline valve is judged;
The multi-mode analysis platform is further used for acquiring corresponding locator data after judging that a natural gas leakage event occurs at any natural gas pipe section or pipeline valve, enabling any optical gas camera to move to a corresponding area according to the acquired locator data, carrying out natural gas leakage event rechecking through the natural gas leakage detection method according to any one of the above results based on the result shot by the optical gas camera, judging that false alarm occurs if the gas leakage area is not detected, and determining natural gas leakage and recording the position of the gas leakage area if the gas leakage area is detected.
Optionally, the multi-mode natural gas leakage detection system further comprises a plurality of alarms;
each group of optical fiber monitoring modules, each group of photoelectric monitoring modules and each group of mobile detection modules are respectively provided with at least one alarm;
the multi-mode analysis platform is used for generating an alarm instruction and sending the alarm instruction to an alarm in the corresponding optical fiber monitoring module after judging that the natural gas pipe section has a natural gas leakage event, generating an alarm instruction and sending the alarm instruction to an alarm in the corresponding photoelectric monitoring module after judging that the natural gas pipe section has a natural gas leakage event, generating an alarm instruction and sending the alarm instruction to an alarm in the corresponding mobile detection module after rechecking to confirm that the natural gas leakage occurs, and generating a cancel alarm instruction and sending the cancel alarm instruction to the alarm in the corresponding optical fiber monitoring module and/or the photoelectric monitoring module after rechecking to judge the false alarm.
Optionally, the multi-mode natural gas leakage detection system further comprises: a display platform;
the display platform is connected with the multi-mode analysis platform and is used for acquiring and displaying corresponding locator data after a natural gas leakage event occurs at any natural gas pipe section or pipeline valve, and acquiring and displaying the locator data corresponding to the mobile detection module and the image shot by the optical gas camera after the natural gas leakage event recheck begins.
Optionally, the mobile device in the mobile detection module is a remote control car, and the optical gas camera is mounted on the top end of the remote control car.
Optionally, the moving device in the movement detection module is a rotating pan-tilt, and the optical gas camera is disposed on the rotating pan-tilt.
In a third aspect, an embodiment of the present invention further provides an electronic device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the method for detecting natural gas leakage according to any embodiment of the present specification is implemented.
In a fourth aspect, embodiments of the present invention further provide a computer-readable storage medium having a computer program stored thereon, which when executed in a computer, causes the computer to perform the method for detecting natural gas leakage according to any of the embodiments of the present specification.
The embodiment of the invention provides a natural gas leakage detection method, electronic equipment and a storage medium, wherein the method is based on infrared video images of two continuous frames of areas to be detected, and the problems of low intelligent degree, low detection efficiency and difficult positioning of the traditional natural gas leakage detection technology can be effectively solved by carrying out the gas leakage area through a natural gas detection model constructed by taking a yolov5 model frame as a main body, so that the real-time, efficient and accurate gas leakage detection is realized.
The embodiment of the invention also provides a multi-mode natural gas leakage detection system, which can realize real-time monitoring of the whole natural gas pipeline, acquire corresponding position information after monitoring the natural gas leakage event of any natural gas pipe section or pipeline valve, then move an optical gas camera to a corresponding area to carry out natural gas leakage event recheck, avoid false alarm caused by an optical fiber acoustic wave monitoring subsystem and a photoelectric signal monitoring subsystem, and rapidly finish accurate positioning of the leakage area.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic diagram of steps of a method for detecting leakage of natural gas according to an embodiment of the present invention;
FIG. 2 (a) is a C3 network architecture;
FIG. 2 (b) is a C2f network structure;
FIG. 3 is a flow chart of a method for detecting natural gas leakage according to an embodiment of the present invention;
FIG. 4 is a schematic structural diagram of an optical fiber monitoring module and an optoelectronic monitoring module according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a mobile detection module according to an embodiment of the present invention;
fig. 6 is a flowchart of a multi-mode natural gas leakage detection method according to an embodiment of the present invention.
In the figure: 1: a natural gas pipe section; 11: an optical fiber pulse emitter; 12: an optical fiber sensor;
2: a pipeline valve; 21: a laser emitter; 22: a photoelectric sensor;
31: a signal transceiver; 32: a positioner; 33: an alarm;
4: a mobile device; 41: an optical gas camera.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments, and all other embodiments obtained by those skilled in the art without making any inventive effort based on the embodiments of the present invention are within the scope of protection of the present invention.
As described above, conventional natural gas leak detection methods generally use "sniffer" technology and probes to detect, require intimate contact with the area to be detected, and are less safe when the detector is exposed to an environment that is invisible to the naked eye and is at risk of a potential explosion. Meanwhile, in the area where natural gas leakage is found, the traditional leakage detection and investigation means are difficult to quickly locate specific leakage points, and are common pain points in the industry. In view of the above, the invention provides a natural gas leakage detection method, which realizes gas leakage area detection by using a natural gas detection model constructed by taking a yolov5 model frame as a main body and utilizing two continuous frames of infrared video images, solves the problems of low intelligent degree, low detection efficiency and difficult positioning of the traditional natural gas leakage detection technology, and realizes real-time, efficient and accurate gas leakage detection.
Specific implementations of the above concepts are described below.
Referring to fig. 1, an embodiment of the present invention provides a method for detecting natural gas leakage, including the following steps:
step 100, acquiring at least two continuous frames of infrared video images of a region to be detected;
the infrared video image may be captured by an optical gas camera; the optical gas phase can detect gas leakage according to environmental temperature change generated by the outside when the gas is exhausted, a scanned image shows a smoke in a gas leakage area, a target detection model is used for detecting the smoke in a picture, and the position of the gas leakage and the range of the gas leakage can be determined at the same time, namely the gas leakage area is determined;
102, after the obtained two continuous frames of infrared video images are transformed to a preset size, inputting the two continuous frames of infrared video images into a trained natural gas detection model for detection, namely, performing model reasoning on the two continuous frames of infrared video images transformed to the preset size by using the trained natural gas detection model to respectively obtain corresponding detection results, namely, respectively obtaining gas leakage areas corresponding to the two continuous frames of infrared video images; if the detection result is that the gas leakage area is not detected, the corresponding gas leakage area is indicated to be empty;
the natural gas detection model is constructed by taking a yolov5 model frame as a main body, a C2f network structure is used for replacing a C3 network structure in an original yolov5 model, a SoftPool operation is used for replacing convolution operation for 2 times of downsampling in the original yolov5 model to perform feature downsampling, and the natural gas detection model is obtained by taking a sample natural gas infrared image with a preset size as input and a corresponding gas leakage area as output training;
the yolov5 model is a target detection network model, and the specific structure of the model can refer to the prior art; transforming the obtained infrared video image to a preset size, and adopting a resize image processing mode; as the input of the natural gas detection model, the infrared images of the sample natural gas and the infrared video images with the same size after the conversion are set according to actual needs by specific numerical values of preset sizes;
And 104, judging whether the detection result is a false result or not based on the detection results corresponding to the two continuous frames of infrared video images, otherwise, judging that the natural gas leakage is detected, and outputting the detection result.
The embodiment of the invention provides a natural gas leakage detection method, which adopts a natural gas detection model constructed by taking a yolov5 model frame as a main body to realize automatic detection based on infrared video images, and under the original yolov5 model frame, a backbone network part modifies a C3 network structure used by the original yolov5 model, changes the C3 network structure into a C2f network structure with higher precision and smaller parameter quantity, and uses a new pooling mode SoftPool operation to perform feature downsampling. Fig. 2 (a) shows a C3 network structure, and fig. 2 (b) shows a C2f network structure, wherein Conv represents a convolution layer, bottleneck represents a Bottleneck layer, concat represents a splice layer, split represents a Split layer, h represents a high feature map, w represents a wide feature map, c_in represents an input feature channel number, c_out represents an output feature channel number, C represents a feature map channel number, k represents a convolution kernel size, s represents a downsampling rate, p represents a padding number, shortcut represents a residual connection layer, and n represents the number of batch training. The SoftPool operation has better characteristic polymerization capability, the parameter number can be reduced, and the SoftPool can be accumulated and activated in an exponential weighting mode. Compared with a series of other pooling methods (such as average pooling and maximum pooling), the SoftPool operation can retain more information in the downsampling activation mapping, can obtain better classification precision, and can bring about 1-2% improvement of consistency precision by replacing convolution operation for 2 times downsampling in the original yolov5 model with the SoftPool operation. The output value of the SoftPool operation is obtained by summing all the criteria for weighted activation in the kernel neighborhood R of the feature map, expressed as:
Figure BDA0004167393740000071
Wherein, the liquid crystal display device comprises a liquid crystal display device,
Figure BDA0004167393740000072
representing the output value, W, of a SoftPool operation i Representing the specific gravity, a, of each value in the kernel neighborhood R i A value representing a position i in the kernel neighborhood R, i representing a position i in the kernel neighborhood R.
The SoftPool operation uses softmax of the region to produce a normalized result with a probability distribution proportional to each activation value relative to the neighboring activation values of the kernel region. This is in direct contrast to selecting the maximum activation value (i.e., max pooling) or averaging all activation values of the kernel region (i.e., average pooling), where the output activation of the kernel region is not regularized. The SoftPool operation is therefore minimal.
Optionally, the natural gas detection model is trained by:
acquiring a natural gas infrared image data set with a label, and carrying out data enhancement to obtain a data set formed by a sample natural gas infrared image; the obtained natural gas infrared image data set is used for correspondingly marking a gas leakage area in an image; the image size of the sample natural gas infrared image is a preset size;
the enhanced data set is then processed according to 6:2:2, dividing the proportion to obtain a training set, a verification set and a test set;
Constructing a yolov5 model frame, replacing a C3 network structure in the model frame with a C2f network structure, and replacing convolution operation for 2 times downsampling with SoftPool operation to obtain a natural gas detection model to be trained;
and training, verifying and testing the constructed natural gas detection model by using the obtained training set, verification set and test set to obtain a trained natural gas detection model.
The above embodiment provides a specific training manner to obtain a trained natural gas detection model. The tagged natural gas infrared image dataset may be obtained by:
acquiring a batch of original natural gas infrared image data;
and marking the natural gas leakage area in the batch of data by using a labelmg marking tool to obtain a labeled natural gas infrared image data set.
Further, performing data enhancement may include performing data enhancement operations such as panning, rotating, cutting, scaling, etc., on the initial data set. In order to obtain the infrared image of the sample natural gas with the preset size, the image conversion can also be performed by adopting a size image processing mode, wherein the preset size is preferably 384×640 resolution, that is, the size of the image input into the natural gas detection model is preferably 384×640 resolution. Too high a resolution may affect the model inference speed, and too low a resolution may result in a decrease in detection accuracy.
Optionally, in step 104, "determining whether the image is a false result based on the detection results corresponding to the two consecutive frames of infrared video images" includes:
judging whether detection results corresponding to two continuous frames of infrared video images both comprise a gas leakage area or not;
if not, the natural gas leakage is not detected, that is, the gas leakage area corresponding to any frame of infrared video image is empty, and the natural gas leakage is not detected;
if yes, further calculating IoU between the gas leakage areas corresponding to the two continuous frames of infrared video images; if IoU is greater than 0.5 but less than 1.0, judging that the natural gas leakage is detected, otherwise judging that the false result is generated, and considering that the natural gas leakage is not detected; wherein, ioU has the expression:
Figure BDA0004167393740000091
(x1 l ，y1 u ) And (x 1) r ，y1 d ) Respectively the coordinates of the upper left corner point and the lower right corner point of the gas leakage area corresponding to one frame of infrared video in two continuous frames of infrared video images, (x 2) l ，y2 u ) And (x 2) r ，y2 d ) The left upper corner coordinates and the right lower corner coordinates of the gas leakage area corresponding to the other frame of infrared video in the two continuous frames of infrared video images are respectively.
In the above embodiment, ioU is adopted to determine whether the detected natural gas leakage event is a false alarm event, if the two frames of images have natural gas leakage detection results and the two detection results are not completely consistent, ioU of the two detection frames (i.e. corresponding gas leakage areas) satisfy more than 0.5 and less than 1.0, the natural gas leakage detection result is considered to be effective, and if the two frames of image detection results are completely consistent, the detection results are false alarm, because the gas is dynamically changed in the natural gas leakage process, the detection frames in the two continuous frames of infrared video images cannot be completely consistent.
Further, the "output detection result" in step 104 includes:
outputting the gas leakage area corresponding to the infrared video image of the next frame, or
And outputting gas leakage areas corresponding to two continuous frames of infrared video images respectively and marking corresponding shooting time.
The next frame of infrared video image is a frame which is the next frame of infrared video image in two continuous frames and is relatively time-delayed. Outputting the gas leakage areas of the two frames of images and marking the corresponding shooting time respectively is helpful for the user to know the gas leakage diffusion condition.
Further, the "output detection result" in step 104 further includes:
and marking a gas leakage area corresponding to the infrared video image on the infrared video image according to the detection result.
Further, the "output detection result" in step 104 further includes:
and reporting the infrared video image with the detection result and sending out an alarm.
According to the output detection result, a user can quickly acquire the position of gas leakage and the range of gas leakage, the mode is high in processing speed and detection precision, manpower is saved, and detection personnel do not need to be exposed in an environment with potential explosion risk.
The method for detecting the natural gas leakage provided by the invention can be applied to a vehicle-mounted optical gas camera detection device and a multi-mode information intelligent analyzer, wherein the vehicle-mounted optical gas camera detection device comprises an optical gas camera, a remotely controllable intelligent robot car, a GPS (global positioning system) positioner and a signal transceiver, an inspection route in a factory is set for the intelligent robot car, each leakage point which is easy to leak is shot by the optical gas camera loaded on the intelligent robot car, the shot picture is sent to the multi-mode information intelligent analyzer in a network mode, the shot picture is analyzed by the multi-mode information intelligent analyzer (namely, the natural gas leakage detection method is implemented, the natural gas leakage detection method is shown in fig. 3), abnormal information of gas leakage in a current picture is determined, an alarm signal is sent to the alarm through the signal transceiver, and the position information of the GPS positioner is sent to the multi-mode information intelligent analyzer. After the user sees the result sent by the multi-mode information intelligent analyzer, the gas leakage area in the image can be further verified according to the alarm signal and the position information of the GPS positioner.
Referring to fig. 4 and 5, the present invention further provides a multi-mode natural gas leakage detection system, and an embodiment of the present invention provides a multi-mode natural gas leakage detection system, which can be applied to a natural gas pipeline for transmitting natural gas, including: the system comprises an optical fiber sound wave monitoring subsystem, a photoelectric signal monitoring subsystem, a movable optical image detection subsystem and a multi-mode analysis platform;
the optical fiber acoustic wave monitoring subsystem comprises at least one group of optical fiber monitoring modules; each group of optical fiber monitoring modules corresponds to one section of natural gas pipe section 1 respectively; the optical fiber monitoring module comprises two sections of auxiliary optical fibers, at least one section of monitoring optical fiber, and a corresponding optical fiber pulse emitter 11 and optical fiber sensor 12, wherein the optical fiber pulse emitter 11 and the optical fiber sensor 12 in the optical fiber monitoring module are equal in number and equal to the sum of the auxiliary optical fibers and the monitoring optical fibers; two sections of auxiliary optical fibers are tightly wound at two ends of the natural gas pipe section 1 respectively, the monitoring optical fibers are spirally wound on the main body of the natural gas pipe section 1 at preset pitches, each section of the monitoring optical fibers is wound on the natural gas pipe section 1 in a non-overlapping manner, one optical fiber pulse emitter 11 is connected to one end of each section of the auxiliary optical fibers and one end of each monitoring optical fiber is connected to the corresponding optical fiber sensor 12; each section of monitoring optical fiber is also provided with a corresponding locator 32, the locator 32 is arranged on the natural gas pipe section 1 of the monitoring optical fiber winding area and is used for outputting locator data containing position information, and the position of the corresponding monitoring optical fiber can be determined through the locator 32; the optical fiber monitoring module further comprises a signal transceiver 31, and the signal transceiver 31 is connected with each optical fiber pulse emitter 11, each optical fiber sensor 12 and each positioner 32 included in the optical fiber monitoring module and is used for receiving and transmitting signals, namely transmitting data;
The photoelectric signal monitoring subsystem comprises at least one group of photoelectric monitoring modules; each group of photoelectric monitoring modules corresponds to one pipeline valve 2; the photoelectric monitoring module comprises a laser emitter 21 and a photoelectric sensor 22 which are respectively arranged at two sides of the pipeline valve 2, and a laser beam emitted by the laser emitter 21 passes through a space where a leakage point of the pipeline valve 2 is easily located and then enters the photoelectric sensor 22; each group of photoelectric monitoring modules is also provided with a corresponding positioner 32 and a signal transceiver 31, the corresponding positioner 32 and the signal transceiver 31 are arranged at the corresponding pipeline valve 2, the positioner 32 is used for outputting positioner data containing position information, the position of the corresponding photoelectric monitoring module can be determined through the positioner 32, and the signal transceiver 31 is connected with the laser transmitter 21, the photoelectric sensor 22 and the positioner 32 which are included in the photoelectric monitoring module and is used for transmitting data;
the movable optical image detection subsystem includes at least one set of movement detection modules; each group of movement detection modules comprises a movement device 4, an optical gas camera 41 arranged on the movement device 4, a positioner 32 and a signal transceiver 31; the moving device 4 is configured to drive the optical gas camera 41 to move, the optical gas camera 41 is configured to capture an infrared image to detect a gas leakage area through an environmental temperature change, the positioner 32 is configured to output positioner data including position information, and the position of the corresponding movement detection module can be determined by the positioner 32, and the signal transceiver 31 is connected to the moving device 4, the optical gas camera 41 and the positioner 32 included in the movable optical image detection subsystem;
The multimode analysis platform is connected with the optical fiber sound wave monitoring subsystem, the photoelectric signal monitoring subsystem and the movable optical image detection subsystem through signal transceivers;
the multimode analysis platform is used for acquiring detection signals of all the optical fiber sensors in the optical fiber acoustic wave monitoring subsystem and monitoring whether natural gas leakage events occur in a natural gas pipe section or not, and the monitoring mode is as follows: subtracting the average value of the optical fiber sensor detection signals corresponding to the auxiliary optical fibers at the two ends from the optical fiber sensor detection signal corresponding to one section of the monitoring optical fiber in each optical fiber monitoring module to obtain a denoised monitoring signal, and demodulating the denoised monitoring signal to determine whether a high-frequency vibration sound source exists in the corresponding natural gas pipe section, and judging that a natural gas leakage event occurs in the corresponding natural gas pipe section if the high-frequency vibration sound source exists;
the multi-mode analysis platform is also used for acquiring detection signals of all the photoelectric sensors in the photoelectric signal monitoring subsystem and monitoring whether a pipeline valve has a natural gas leakage event or not, and the monitoring mode is as follows: demodulating detection signals corresponding to the photoelectric sensors in the photoelectric monitoring modules to determine whether the corresponding pipeline valves have gas component changes, and judging that natural gas leakage events occur at the corresponding pipeline valves if the gas component changes;
The multi-mode analysis platform is further configured to obtain corresponding locator data after judging that a natural gas leakage event occurs at any natural gas pipe section or pipeline valve, and move any one of the optical gas cameras to a corresponding area according to the obtained locator data, and based on a result photographed by the optical gas camera, perform natural gas leakage event recheck through the natural gas leakage detection method according to any one of the embodiments, if no gas leakage area is detected in the corresponding area (i.e., at the corresponding natural gas pipe section or pipeline valve), determine that a false alarm occurs, and if a gas leakage area is detected, determine natural gas leakage and record a position of the gas leakage area.
The embodiment of the invention provides a multi-mode natural gas leakage detection system. In this system, the optical fiber acoustic monitoring subsystem may include multiple sets of optical fiber monitoring modules, where each set of optical fiber monitoring modules is used to monitor different sections of the natural gas pipe section, as shown in fig. 4 (for convenience of illustration, the auxiliary optical fiber and the monitoring optical fiber are not distinguished in fig. 4). If the natural gas pipe section is not long, two sections of auxiliary optical fibers and one section of monitoring optical fiber can be adopted to wind the natural gas pipe section; if the natural gas pipe section is longer, a plurality of sections of monitoring optical fibers can be additionally arranged for monitoring each region of the natural gas pipe section. For the auxiliary optical fiber and the monitoring optical fiber, the optical fiber pulse transmitter is used for transmitting pulses from the head of the optical fiber, the optical fiber sensor is used for receiving pulse signals from the tail of the optical fiber to obtain detection signals, namely, ripple pulses are continuously transmitted to the auxiliary optical fiber and the monitoring optical fiber through the optical fiber pulse transmitter respectively, the corresponding optical fiber sensor is used for receiving the transmitted pulse ripples, the detection signals are obtained through sampling, and the detection signals corresponding to the optical fibers are transmitted to the multi-mode analysis platform through the signal transceiver. Monitoring the spiral winding of the optical fiber increases the fiber-to-pipe coverage ratio r=lf/Lp, where Lf is the length of the optical fiber used to cover a length Lp of natural gas pipe segment, as compared to straight fiber applications along the pipe, helping to improve the measurement sensitivity and spatial resolution of potentially weak signals. Considering that the arrangement of a single length of monitoring fiber on an excessively long natural gas pipe section may lead to a specific location where it may be difficult to locate a leak event, the fiber monitoring module may employ multiple lengths of monitoring fiber that are wound without overlapping to monitor different areas of the pipe section, respectively. Also, because the natural gas pipe is formed by connecting multiple segments of natural gas pipe, the presence of connection points (e.g., pipe flanges) makes continuous fiber applications impractical, and there is always a short free hanging fiber section between the natural gas pipe segments, which is prone to spurious signals. In order to reduce false alarms and avoid the situation of vibration of the pipe section and the like from interfering with monitoring results, the invention applies the tightly wound auxiliary optical fiber at the positions of the two ends of the pipe section close to the flanges for detecting the noise of the natural gas pipe section, and can improve the accuracy and the reliability of the optical fiber monitoring module by removing the noise. The invention monitors whether the natural gas pipe section leaks or not by utilizing the acoustic wave change of the bearing device during gas leakage, and generally when a leakage event of the natural gas pipe occurs, the flow of leakage objects at a leakage port can generate a high-frequency sound signal which can be detected. The optical fiber acoustic wave sensing technology utilizes the characteristic that signals transmitted by light in an optical fiber are sensitive to acoustic wave vibration characteristics to realize the vibration detection of the acoustic wave of the leakage point, and is efficient and accurate.
The photoelectric signal monitoring subsystem of the invention can comprise a plurality of groups of photoelectric monitoring modules, each photoelectric monitoring module is respectively arranged at each pipeline valve in the natural gas pipeline, laser beams emitted by the laser emitters are emitted from one side of the pipeline valve, pass through the space where the pipeline valve is easy to leak, enter photoelectric sensors at the other side of the pipeline valve, and detection signals corresponding to each photoelectric sensor are transmitted to the multi-modal analysis platform through the signal transceiver. The invention monitors whether the pipeline valve leaks or not by utilizing the physical property change of the gas, and when the light passes through the target gas (namely the natural gas) in the propagation process, the parameters such as the intensity, the phase, the polarization state, the frequency and the like of the light can be changed, and whether the pipeline valve area leaks or not can be judged by comparing the parameter change in the leakage state with the parameter change in the normal non-leakage state. The pipeline valve is used as an easy leakage point, the real-time monitoring cannot be carried out in a fiber winding mode, and the photoelectric monitoring module can monitor the pipeline valve area, so that the manpower is reduced, and the maintenance cost is reduced.
In consideration of the fact that whether natural gas leakage occurs in a natural gas pipeline or not can be monitored in real time and rapidly through the optical fiber sound wave monitoring subsystem and the photoelectric signal monitoring subsystem, false alarm (namely false alarm) possibly occurs due to high sensitivity, and natural gas leakage is accurately detected.
The multi-mode analysis platform transmits data through each signal transceiver, the optical fiber sound wave monitoring subsystem, the photoelectric signal monitoring subsystem and the movable optical image detection subsystem, and each optical fiber monitoring module, the photoelectric monitoring module and the movable detection module are in wireless connection with the multi-mode analysis platform, so that wiring can be reduced, and the movable detection module is small in movement limitation and flexible.
The multi-mode analysis platform acquires and demodulates detection signals of all optical fiber sensors in the optical fiber acoustic wave monitoring subsystem, and the detection signals depend on high-frequency acoustic wave signals: the pulse laser is sampled and received by the optical fiber sensor after being transmitted by the optical fiber, the detection signal is denoised, and then Fourier transformation is carried out, if the spectrum data has peak abnormality (high-frequency acoustic interference occurs), the leakage of the natural gas pipeline area can be determined.
The multi-mode analysis platform acquires and demodulates detection signals of all photoelectric sensors in the photoelectric signal monitoring subsystem, and the detection signals depend on gas component changes: the light emitted by the laser emitter interacts with the gas to be measured, the information of the light changes, the light with the measured information is converted into an electric signal by the photoelectric sensor, and the concentration data of the gas to be measured is obtained by detecting the light information.
The multi-mode analysis platform obtains the result of the shooting of the optical gas camera and carries out natural gas leakage incident recheck, the gas radiates infrared energy outwards from the leakage point and affects the surrounding background environment, and when the optical gas camera is used for shooting in a large area, the abnormal heat radiation position can be found through pictures or videos, and the found abnormal heat condition (such as being hotter or colder than the background environment) can be found. The optical gas camera can display the position and the diffusion direction of gas leakage in real time. The multi-mode analysis platform adopts the natural gas leakage detection method to detect the video shot by the optical gas camera, can automatically determine the position of the gas leakage area in the video, is real-time and rapid, and reduces the workload of manual scanning.
Optionally, the multi-modal natural gas leak detection system further includes a plurality of alarms 33;
each group of optical fiber monitoring modules, each group of photoelectric monitoring modules and each group of movement detection modules are respectively provided with at least one alarm 33;
the multi-mode analysis platform is used for generating an alarm instruction and sending the alarm instruction to an alarm in the corresponding optical fiber monitoring module after judging that the natural gas pipe section has a natural gas leakage event, generating an alarm instruction and sending the alarm instruction to an alarm in the corresponding photoelectric monitoring module after judging that the natural gas pipe section has a natural gas leakage event, generating an alarm instruction and sending the alarm instruction to an alarm in the corresponding mobile detection module after rechecking to confirm that the natural gas leakage occurs, and generating a cancel alarm instruction and sending the cancel alarm instruction to the alarm in the corresponding optical fiber monitoring module and/or the photoelectric monitoring module after rechecking to judge the false alarm.
The above embodiment adopts a plurality of alarms to send out alarm signals after suspected or determined natural gas leakage is detected, surrounding staff can be warned, corresponding areas of the staff are warned that danger possibly exists, and the staff can also go to a leakage point in time for maintenance according to the warning, so that the danger is eliminated rapidly.
Optionally, the multi-mode natural gas leakage detection system further comprises a display platform;
the display platform is connected with the multi-mode analysis platform and is used for acquiring and displaying corresponding locator data after a natural gas leakage event occurs at any natural gas pipe section or pipeline valve, and acquiring and displaying the locator data corresponding to the mobile detection module and the image shot by the optical gas camera after the natural gas leakage event recheck begins.
Further, the display platform can also display the locator data of each group of mobile detection modules in real time so as to monitor the real-time position of each mobile detection module and allocate the closest mobile detection module when required.
Further, the display platform is also used for acquiring and displaying alarm instructions.
Through the display platform, the staff can monitor the safe condition of whole natural gas line directly perceivedly. The display platform preferably displays the information on the basis of a natural gas pipeline panorama so that a worker can quickly determine the occurrence position of a natural gas leakage event by using the display platform and timely determine the specific gas leakage point.
Optionally, as shown in fig. 5, the moving device in the movement detection module is a remote control vehicle, and the optical gas camera is mounted on the top end of the remote control vehicle and can move along with the remote control vehicle.
The embodiment adopts the remote control car to realize the position movement of the optical gas camera, the vehicle-mounted optical gas camera is convenient to move and high in flexibility, and natural gas leakage incident rechecks can be carried out on different areas.
Optionally, the moving device in the movement detection module is a rotating cradle head, and the optical gas camera is arranged on the rotating cradle head and can rotate along with the rotating cradle head.
According to the embodiment, the rotary holder is used for moving the shooting position of the optical gas camera, the moving speed of the rotary holder is high, and the recheck of the region suspected to have the natural gas leakage event can be rapidly carried out.
Optionally, the multi-mode analysis platform is further configured to generate a corresponding setting instruction according to an input instruction, and correspondingly send the setting instruction to the optical fiber acoustic wave monitoring subsystem, the photoelectric signal monitoring subsystem and/or the movable optical image detection subsystem.
Further, the multi-mode analysis platform is used for setting working parameters of the optical fiber pulse emitter and the optical fiber sensor in each optical fiber monitoring module, wherein the working parameters comprise the frequency, the wavelength and the amplitude of the emitted pulse of the optical fiber pulse emitter, the sampling frequency and the space window of the optical fiber sensor; the multi-mode analysis platform is used for setting working parameters of the laser transmitters and the photoelectric sensors in the photoelectric monitoring modules, including the wavelength, the amplitude and the intensity of laser transmitted by the laser transmitters, and the sampling frequency and the space window of the photoelectric sensors; the multi-mode analysis platform is used for setting working parameters of each mobile detection module, including a moving route, a moving speed and the like of the mobile device, and image resolution, focal length, minimum detection level of target gas and the like of the optical gas camera. The multi-mode analysis platform is also used for setting working parameters of the signal transceiver, including a signal receiving source, a signal sending source, a signal encoding and decoding mode, a signal transmission bit rate and the like.
The multimode analysis platform is utilized for setting, so that unified management of the optical fiber acoustic wave monitoring subsystem, the photoelectric signal monitoring subsystem and the movable optical image detection subsystem can be realized, and the workload of manual setting of staff is reduced.
Optionally, the locator is a GPS locator, and the locator data is GPS locating information of a location where the locator is located.
The GPS localizer has the advantages of high precision and strong anti-interference capability.
Optionally, in the optical fiber monitoring module, the pulse length range of the pulse laser emitted by the optical fiber pulse emitter is 150 ns-250 ns, preferably 200ns, and the pulse frequency range is 70 kHz-90 kHz, preferably 80kHz.
The relatively large pulse length helps to increase the sensitivity of the system to potentially weak leakage-inducing signals.
Further, in the optical fiber monitoring module, the auxiliary optical fiber and the monitoring optical fiber are both standard single mode optical fibers without jackets; the length of each section of the auxiliary optical fiber ranges from 8m to 12m, preferably 10m; the preset pitch of each section of the monitoring optical fiber is 2 cm-3 cm, preferably 2.5cm, and the winding area is not more than 10m, preferably 5 m-6 m.
The standard single-mode fiber has the advantages of low cost and stable performance. The 10m auxiliary optical fiber which is closely wound is applied to the positions, close to the flanges, of the two ends of the pipe section, so that the interference of weak signals caused by pipeline noise on monitoring signals can be effectively reduced. Too small a pitch of monitoring fiber winding may cause increased false alarm events, and too large a pitch may reduce the resolution of monitoring. The monitoring optical fibers arranged in a segmented mode are beneficial to improving the positioning accuracy of leakage events, and false alarm events are easy to occur to overlong optical fibers.
Optionally, the multi-mode analysis platform monitors whether a natural gas pipe section has a natural gas leakage event, including performing the following steps for each of the optical fiber monitoring modules:
acquiring a section of detection signals corresponding to the monitoring optical fiber and the auxiliary optical fibers at the two ends according to the set time window;
denoising the detection signal corresponding to the monitoring optical fiber by taking the detection signal in the time window as an analysis object to obtain a monitoring signal; the denoising step comprises the step of subtracting the average value of the detection signals corresponding to the auxiliary optical fibers at the two ends from the detection signals corresponding to the monitoring optical fibers;
the monitoring signals in the time window are subjected to windowing treatment to realize short-time fast Fourier transform, so that corresponding frequency spectrums are obtained;
judging whether a high-frequency vibration sound source exists or not based on the frequency spectrum corresponding to the current time window and the historical average frequency spectrum; if the spectrum corresponding to the current time window has a peak exceeding a preset threshold value relative to the historical average spectrum, judging that the high-frequency vibration sound source exists, otherwise, updating the historical average spectrum based on the spectrum corresponding to the current time window.
According to the embodiment, whether the high-frequency vibration sound source exists or not is judged based on the frequency spectrum corresponding to the current time window and the historical average frequency spectrum, the historical average frequency spectrum is relative to a normal value serving as a reference, the historical average frequency spectrum is an average result obtained according to the frequency spectrum in a past period, if the frequency spectrum corresponding to the current time window has a peak exceeding a preset threshold value relative to the historical average frequency spectrum, the high-frequency vibration sound source is judged to exist, the natural gas pipe section corresponding to the high-frequency vibration sound source is considered to have a natural gas leakage event, and if the frequency spectrum corresponding to the current time window does not have a peak exceeding the preset threshold value relative to the historical average frequency spectrum, the historical average frequency spectrum is updated based on the frequency spectrum corresponding to the current time window so as to facilitate subsequent detection.
Optionally, as shown in fig. 6, the present invention further provides a method for detecting multi-mode natural gas leakage, which is implemented by using the multi-mode natural gas leakage detection system according to any one of the embodiments, and includes the following steps:
monitoring whether natural gas leakage occurs to each natural gas pipe section and a pipeline valve of the natural gas pipeline or not through the optical fiber sound wave monitoring subsystem and the photoelectric signal monitoring subsystem respectively;
if a natural gas leakage event occurs at any natural gas pipe section or pipeline valve, acquiring corresponding locator data;
and according to the acquired locator data, any optical gas camera in the movable optical image detection subsystem is moved to a corresponding area to carry out natural gas leakage event re-detection, if the gas leakage area is not detected, the false alarm is judged to appear, and if the gas leakage area is detected, the natural gas leakage is determined and the position of the gas leakage area is recorded.
According to the multi-mode natural gas leakage detection system and method provided by the invention, whether leakage occurs in the natural gas pipe section is monitored by utilizing the sound wave change of the bearing device when gas leaks, whether leakage occurs in the pipeline valve is monitored by utilizing the physical property change of the gas, the real-time monitoring of the whole natural gas pipeline is realized, the work such as manual inspection is replaced, after the natural gas leakage event occurs in any natural gas pipe section or pipeline valve is monitored, the corresponding position information is obtained, the optical gas camera is moved to the corresponding area for carrying out natural gas leakage event rechecking, the false alarm caused by the optical fiber sound wave monitoring subsystem and the photoelectric signal monitoring subsystem is avoided, the accurate positioning of the leakage point can be realized quickly, the problems of low intelligent degree, low detection efficiency and complex operation of the traditional manual inspection mode are solved, and the automatic, real-time, high-efficiency and accurate detection of the natural gas leakage is realized.
It is noted that relational terms such as first and second, and the like, are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
Those of ordinary skill in the art will appreciate that: all or part of the steps for implementing the above method embodiments may be implemented by hardware related to program instructions, and the foregoing program may be stored in a computer readable storage medium, where the program, when executed, performs steps including the above method embodiments; and the aforementioned storage medium includes: various media in which program code may be stored, such as ROM, RAM, magnetic or optical disks.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims (10)

1. A method of natural gas leak testing comprising:
acquiring infrared video images of at least two continuous frames of areas to be detected;
after the obtained two continuous frames of infrared video images are converted to a preset size, inputting the two continuous frames of infrared video images into a trained natural gas detection model for detection, and respectively obtaining corresponding detection results; the natural gas detection model is constructed by taking a yolov5 model frame as a main body, a C2f network structure is used for replacing a C3 network structure, a software operation is used for replacing convolution operation for 2 times of downsampling to perform feature downsampling, and the natural gas detection model is obtained by taking a sample natural gas infrared image with a preset size as input and a corresponding gas leakage area as output training;
based on the detection results corresponding to the two continuous frames of infrared video images, judging whether the detection results are false results, otherwise, judging that the natural gas leakage is detected, and outputting the detection results.
2. The method for detecting leakage of natural gas according to claim 1, wherein,
the natural gas detection model is obtained through training in the following mode:
acquiring a natural gas infrared image data set with a label, and carrying out data enhancement to obtain a data set formed by a sample natural gas infrared image; the tag is used for correspondingly marking a gas leakage area in the image;
The enhanced data set is then processed according to 6:2:2, dividing the proportion to obtain a training set, a verification set and a test set;
constructing a yolov5 model frame, replacing a C3 network structure in the model frame with a C2f network structure, and replacing convolution operation for 2 times downsampling with SoftPool operation to obtain a natural gas detection model to be trained;
and training, verifying and testing the constructed natural gas detection model by using the obtained training set, verification set and test set to obtain a trained natural gas detection model.
3. The method for detecting leakage of natural gas according to claim 1, wherein,
the step of judging whether the detection result is a false result based on the detection result corresponding to the continuous two frames of infrared video images comprises the following steps:
judging whether detection results corresponding to two continuous frames of infrared video images both comprise a gas leakage area or not;
if not, the natural gas leakage is not detected;
if yes, ioU between the gas leakage areas corresponding to the two continuous frames of infrared video images is calculated; if IoU is greater than 0.5 but less than 1.0, then determining that natural gas leakage is detected, otherwise determining that a false result is obtained; wherein, ioU has the expression:
Figure FDA0004167393730000021
(x1 l ，y1 u ) And (x 1) r ，y1 d ) Respectively an upper left corner coordinate and a lower right corner coordinate of a frame of infrared video corresponding to the gas leakage area, (x 2) l ，y2 u ) And (x 2) r ，y2 d ) And the coordinates of the upper left corner and the lower right corner of the gas leakage area corresponding to the other frame of infrared video are respectively.
4. A multi-modal natural gas leak detection system, for use with a natural gas pipeline, comprising: the system comprises an optical fiber sound wave monitoring subsystem, a photoelectric signal monitoring subsystem, a movable optical image detection subsystem and a multi-mode analysis platform; wherein, the liquid crystal display device comprises a liquid crystal display device,
the optical fiber acoustic wave monitoring subsystem comprises at least one group of optical fiber monitoring modules, and each group of optical fiber monitoring modules corresponds to one section of natural gas pipe respectively; the optical fiber monitoring module comprises two sections of auxiliary optical fibers, at least one section of monitoring optical fiber, an optical fiber pulse emitter and an optical fiber sensor, wherein the two sections of auxiliary optical fibers are tightly wound at two ends of a natural gas pipe section respectively, the monitoring optical fibers are spirally wound on a main body of the natural gas pipe section at preset pitches, each section of monitoring optical fiber is wound on the natural gas pipe section in a non-overlapping manner, one optical fiber pulse emitter is connected to one end of each section of auxiliary optical fiber and one end of each monitoring optical fiber, and the other end of each section of auxiliary optical fiber is connected to the corresponding optical fiber sensor; each section of monitoring optical fiber is also provided with a corresponding positioner which is arranged on the natural gas pipe section of the monitoring optical fiber winding area; the optical fiber monitoring module further comprises a signal transceiver, and the signal transceiver is connected with each optical fiber pulse emitter, each optical fiber sensor and each positioner which are included in the optical fiber monitoring module;
The photoelectric signal monitoring subsystem comprises at least one group of photoelectric monitoring modules, and each group of photoelectric monitoring modules corresponds to one pipeline valve; the photoelectric monitoring module comprises a laser emitter and a photoelectric sensor which are respectively arranged at two sides of the pipeline valve, and a laser beam emitted by the laser emitter passes through a space where a leakage point of the pipeline valve is easily located and then enters the photoelectric sensor; each group of photoelectric monitoring modules is also provided with a corresponding locator and a signal transceiver, the corresponding locator and the signal transceiver are arranged at the corresponding pipeline valve, and the signal transceiver is connected with the laser transmitter, the photoelectric sensor and the locator which are included by the photoelectric monitoring modules;
the movable optical image detection subsystem includes at least one set of movement detection modules; each group of movement detection modules comprises a movement device, an optical gas camera arranged on the movement device, a positioner and a signal transceiver; the movable device is used for driving the optical gas camera to move, the optical gas camera is used for shooting an infrared image so as to detect a gas leakage area through environmental temperature change, and the signal transceiver is connected with the movable device, the optical gas camera and the positioner which are included in the movable optical image detection subsystem;
The multimode analysis platform is connected with the optical fiber sound wave monitoring subsystem, the photoelectric signal monitoring subsystem and the movable optical image detection subsystem through signal transceivers;
the multimode analysis platform is used for acquiring detection signals of all the optical fiber sensors in the optical fiber acoustic wave monitoring subsystem and monitoring whether natural gas leakage occurs in a natural gas pipe section or not, the monitoring mode is that the average value of detection signals corresponding to the two ends of the auxiliary optical fibers is subtracted from detection signals corresponding to one section of the monitoring optical fiber in each optical fiber monitoring module, so that a denoised monitoring signal is obtained, demodulation is carried out to determine whether a high-frequency vibration sound source exists or not, and if the high-frequency vibration sound source exists, the natural gas leakage occurs in the corresponding natural gas pipe section;
the multi-mode analysis platform is further used for acquiring detection signals of the photoelectric sensors in the photoelectric signal monitoring subsystem and monitoring whether a natural gas leakage event occurs to a pipeline valve or not, the monitoring mode is that the detection signals corresponding to the photoelectric sensors in the photoelectric monitoring modules are demodulated to determine whether gas component changes occur, and if the gas component changes occur, the natural gas leakage event occurs to the corresponding pipeline valve is judged;
The multi-mode analysis platform is further used for acquiring corresponding locator data after judging that a natural gas leakage event occurs at any natural gas pipe section or pipeline valve, enabling any optical gas camera to move to a corresponding area according to the acquired locator data, carrying out natural gas leakage event rechecking through the natural gas leakage detection method according to any one of claims 1-3 based on the result shot by the optical gas camera, judging that a false alarm occurs if a gas leakage area is not detected, and determining natural gas leakage and recording the position of the gas leakage area if the gas leakage area is detected.
5. The multi-modal natural gas leak detection system of claim 4, further comprising a plurality of alarms;
each group of optical fiber monitoring modules, each group of photoelectric monitoring modules and each group of mobile detection modules are respectively provided with at least one alarm;
the multi-mode analysis platform is used for generating an alarm instruction and sending the alarm instruction to an alarm in the corresponding optical fiber monitoring module after judging that the natural gas pipe section has a natural gas leakage event, generating an alarm instruction and sending the alarm instruction to an alarm in the corresponding photoelectric monitoring module after judging that the natural gas pipe section has a natural gas leakage event, generating an alarm instruction and sending the alarm instruction to an alarm in the corresponding mobile detection module after rechecking to confirm that the natural gas leakage occurs, and generating a cancel alarm instruction and sending the cancel alarm instruction to the alarm in the corresponding optical fiber monitoring module and/or the photoelectric monitoring module after rechecking to judge the false alarm.
6. The multi-modal natural gas leak detection system of claim 4, further comprising: a display platform;
the display platform is connected with the multi-mode analysis platform and is used for acquiring and displaying corresponding locator data after a natural gas leakage event occurs at any natural gas pipe section or pipeline valve, and acquiring and displaying the locator data corresponding to the mobile detection module and the image shot by the optical gas camera after the natural gas leakage event recheck begins.
7. The multi-modality gas leak detection system of claim 4, wherein the mobile device in the mobile detection module is a remote control car, the optical gas camera being mounted on top of the remote control car.
8. The multi-modal natural gas leak detection system of claim 4, wherein the mobile device in the mobile detection module is a rotating head, the optical gas camera being disposed on the rotating head.
9. An electronic device comprising a memory and a processor, the memory having stored therein a computer program, characterized in that the processor, when executing the computer program, implements the method of any of claims 1-3.
10. A storage medium having stored thereon a computer program, which, when executed in a computer, causes the computer to perform the method of any of claims 1-3.
CN202310367612.3A 2023-04-07 2023-04-07 Natural gas leakage detection method and multi-mode natural gas leakage detection system Active CN116379359B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310367612.3A CN116379359B (en) 2023-04-07 2023-04-07 Natural gas leakage detection method and multi-mode natural gas leakage detection system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310367612.3A CN116379359B (en) 2023-04-07 2023-04-07 Natural gas leakage detection method and multi-mode natural gas leakage detection system

Publications (2)

Publication Number Publication Date
CN116379359A true CN116379359A (en) 2023-07-04
CN116379359B CN116379359B (en) 2023-11-28

Family

ID=86969004

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310367612.3A Active CN116379359B (en) 2023-04-07 2023-04-07 Natural gas leakage detection method and multi-mode natural gas leakage detection system

Country Status (1)

Country Link
CN (1) CN116379359B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116543241A (en) * 2023-07-07 2023-08-04 杭州海康威视数字技术股份有限公司 Detection method and device for leakage gas cloud, storage medium and electronic equipment
CN117670878A (en) * 2024-01-31 2024-03-08 天津市沛迪光电科技有限公司 VOCs gas detection method based on multi-mode data fusion

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0536586A1 (en) * 1991-10-08 1993-04-14 Osaka Gas Co., Ltd. Gas visualizing apparatus and gas visualizing method
CN104122054A (en) * 2014-07-27 2014-10-29 北京航星网讯技术股份有限公司 Infrared thermal imaging type laser gas leak detection method and apparatus
US20190339158A1 (en) * 2016-03-27 2019-11-07 Opgal Optronic Industrial Ltd. Combined gas leakage detection and quantification
CN114170552A (en) * 2021-12-09 2022-03-11 上海谱盟光电科技有限公司 Natural gas leakage real-time early warning method and system based on infrared thermal imaging
CN114298989A (en) * 2021-12-17 2022-04-08 洛阳热感科技有限公司 YOLOV 5-based thermal infrared gas leakage detection method, detection device and detection system
CN114964628A (en) * 2022-05-12 2022-08-30 昆明理工大学 Shuffle self-attention light-weight infrared detection method and system for ammonia gas leakage
CN115751203A (en) * 2022-12-05 2023-03-07 南通中油燃气有限责任公司 Natural gas pipeline leakage monitoring system based on thermal infrared imager

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0536586A1 (en) * 1991-10-08 1993-04-14 Osaka Gas Co., Ltd. Gas visualizing apparatus and gas visualizing method
CN104122054A (en) * 2014-07-27 2014-10-29 北京航星网讯技术股份有限公司 Infrared thermal imaging type laser gas leak detection method and apparatus
US20190339158A1 (en) * 2016-03-27 2019-11-07 Opgal Optronic Industrial Ltd. Combined gas leakage detection and quantification
CN114170552A (en) * 2021-12-09 2022-03-11 上海谱盟光电科技有限公司 Natural gas leakage real-time early warning method and system based on infrared thermal imaging
CN114298989A (en) * 2021-12-17 2022-04-08 洛阳热感科技有限公司 YOLOV 5-based thermal infrared gas leakage detection method, detection device and detection system
CN114964628A (en) * 2022-05-12 2022-08-30 昆明理工大学 Shuffle self-attention light-weight infrared detection method and system for ammonia gas leakage
CN115751203A (en) * 2022-12-05 2023-03-07 南通中油燃气有限责任公司 Natural gas pipeline leakage monitoring system based on thermal infrared imager

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116543241A (en) * 2023-07-07 2023-08-04 杭州海康威视数字技术股份有限公司 Detection method and device for leakage gas cloud, storage medium and electronic equipment
CN116543241B (en) * 2023-07-07 2023-09-15 杭州海康威视数字技术股份有限公司 Detection method and device for leakage gas cloud, storage medium and electronic equipment
CN117670878A (en) * 2024-01-31 2024-03-08 天津市沛迪光电科技有限公司 VOCs gas detection method based on multi-mode data fusion
CN117670878B (en) * 2024-01-31 2024-04-26 天津市沛迪光电科技有限公司 VOCs gas detection method based on multi-mode data fusion

Also Published As

Publication number Publication date
CN116379359B (en) 2023-11-28

Similar Documents

Publication Publication Date Title
CN116379359B (en) Natural gas leakage detection method and multi-mode natural gas leakage detection system
US10281912B2 (en) Autonomous inspection system
CN107965674B (en) A kind of scan-type gas leakage whole audience early warning system
US9335299B2 (en) Method and system for testing a bundle of tubular objects guided by a computing device
CN101886743B (en) Method and device for positioning submarine pipeline robot
CN113056665A (en) Detection and fault detection of corrosion under fire-resistant insulation using hybrid sensing system
JP2004514154A (en) Ultrasonic probe
CN110470669B (en) Leakage detection method and system for underwater pipeline and related device
CN115436759A (en) Inspection robot monitoring method and system based on acoustic imaging sensing technology
CN108507725B (en) Sulfur hexafluoride gas leakage automatic detection device
CN117092710B (en) Underground line detection system for construction engineering investigation
CN112837286A (en) Detection method and system for support deformation
CN116388858A (en) Fiber breakage detection method and device based on phase demodulation
CN116295788B (en) Multi-mode natural gas leakage detection system and method
CN114354744B (en) Power cable external damage prevention monitoring method, system and device
EP4273743A1 (en) Method and apparatus for detecting fire spots, electronic device, and storage medium
CN113189991B (en) Intelligent system capable of automatically planning safe inspection route and application method thereof
CN115938077A (en) Chemical industry park gas remote measuring alarm integration method and system
JP2024516568A (en) Mapping using optical fiber sensing
CN116131938B (en) Method and device for determining abnormal signals of equipment carrying optoelectronic module
JP2996323B2 (en) Gas leak position detection method
CN116679172B (en) Visual detection system and method for local discharge source of electric power site
CN114360221B (en) Multi-source real-time monitoring system for protecting submarine optical cable
KR20210003518A (en) Diagnostic apparatus for facilities of power transmission using unmaned aerial vehicle and method thereof
CN217469960U (en) Optical time domain reflectometer and distributed optical fiber sensor optical cable fault acquisition 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
GR01 Patent grant
GR01 Patent grant