Gas leakage detection and rapid traceability device and method

Abstract

The invention discloses a gas leakage detection and rapid traceability device and method, relates to gas leakage detection and traceability, and aims to apply visible light imaging to gas leakage traceability and realize rapid traceability of gas leakage. The system comprises a visible light imaging module, a gas tunable semiconductor laser absorption spectrum detection module, a signal synchronous acquisition module and a data processing unit, the visible light imaging module is used for imaging a gas leakage scene; the gas tunable semiconductor laser absorption spectrum detection module is used for collecting the gas leakage concentration of a specified point in a scene; the signal synchronous acquisition module is used for synchronously acquiring data; the data processing unit is used for realizing gas leakage model modeling, concentration field fitting and projection; an included angle # imgabs 0 #: # imgabs 1 # exists between the optical axis of the visible light imaging module and the optical axis of the gas tunable semiconductor laser absorption spectrum detection module; and # imgabs2 #. The positioning time of the gas leakage source can be greatly shortened, and the leakage source can be quickly positioned.

Description

Technical Field

[0001] The present invention belongs to the technical field of gas detection, and relates to the detection and tracing of gas leakage, and in particular to a gas leakage detection and rapid tracing device and method. Background Art

[0002] Gas leak detection technology has broad and important applications in areas such as air pollution control, petrochemicals, and industrial process monitoring. For example, in industries like petrochemicals, there are numerous and complex methane gas storage or transportation equipment. Effectively monitoring leaks from these devices is a crucial prerequisite for ensuring safe production.

[0003] Contact leak detection equipment is currently widely used for gas leaks. The device's sensor needs to come into contact with the target gas to detect a leak. However, some leaks may be leaking, but most potential leaks have not yet occurred. Therefore, operators must personally inspect and detect each potential leak, which is inefficient and poses a health risk to inspectors. Furthermore, traditional gas leak detection technology can only measure concentration at a single point. While infrared imaging technology, developed specifically for methane gas leak detection, utilizes an infrared focal plane detector to image leaking gas, it offers the advantages of large-scale monitoring and rapid source tracing. However, it cannot quantitatively measure leak concentrations, and therefore cannot provide accurate situation analysis for subsequent accident assessment and remedial measures.

[0004] The invention patent application with application number 202010443190X discloses an infrared imaging and concentration detection device and method for methane gas leaks. The device includes an infrared thermal imager and a laser methane detection module. The infrared thermal imager is used to perform infrared imaging of the desired monitoring area to detect methane gas leaks. The laser methane detection module is aligned according to the leak points detected by infrared imaging and measures the methane concentration. The laser methane detection module is embedded in the infrared thermal imager, and the laser beam emission direction of the laser methane detection module is adjusted to be consistent with the optical axis direction of the infrared thermal imager. The two communicate with each other to achieve data exchange. The infrared thermal imager can detect a large area window to determine whether there is a natural gas leak in the window. The laser methane detection module can perform long-distance non-contact measurement of the methane leak point and obtain accurate methane gas concentration information. The combination of infrared thermal imaging technology for methane gas leak detection and laser methane detection technology simultaneously realizes non-contact, large-area rapid search for methane leak points and quantitative detection of methane gas concentration, achieving more efficient and accurate quantitative comprehensive detection than existing single detection devices.

[0005] In the above-mentioned invention patent application for infrared imaging and concentration detection device and method, although it achieves quantitative detection of methane leakage points and methane gas concentration. However, firstly, like traditional gas leak detection technology, it can only achieve single-point concentration measurement and cannot obtain the step-by-step concentration field; secondly, traditional gas leak tracing mostly uses infrared sensors for gas imaging, but infrared sensors are relatively expensive, which is not conducive to popularization and application; thirdly, existing visible light images only provide on-site evidence collection functions and do not have the ability to quickly trace the source. Therefore, in order to reduce costs and promote application, how to use visible light imaging for gas leak tracing and achieve rapid tracing of gas leaks is particularly important. It is necessary to provide a gas leak detection and rapid tracing device and method based on visible light imaging. Summary of the Invention

[0006] The purpose of the present invention is to solve the technical problem of how to use visible light imaging for gas leak tracing and achieve rapid tracing of gas leaks in the prior art, and to provide a gas leak detection and rapid tracing device and method. By combining multi-point TDLAS concentration value measurement with visible light imaging, the gas concentration field is fitted and the evidence of the leakage scene is obtained.

[0007] In order to achieve the above-mentioned purpose, the present invention specifically adopts the following technical solutions: A gas leak detection and rapid tracing device, comprising a visible light imaging module, a gas tunable semiconductor laser absorption spectrum detection module, a signal synchronization acquisition module and a data processing unit; The visible light imaging module is used to image the leaking gas scene and provide the background and basic coordinate system for leak positioning; The gas tunable semiconductor laser absorption spectroscopy detection module is used to collect the gas leakage concentration at the leakage point in the scene; The signal synchronization acquisition module is used to synchronize data acquisition of the visible light imaging module and the gas tunable semiconductor laser absorption spectrum detection module; The data processing unit is used to realize gas leakage modeling, concentration field fitting and projection; There is an angle between the optical axis of the visible light imaging module and the optical axis of the gas tunable semiconductor laser absorption spectrum detection module , and the angle Satisfaction relationship: ; ; in, is the upper limit of the angle, is the distance between the optical center of the visible light imaging module and the optical center of the gas tunable semiconductor laser absorption spectrum detection module, is the nominal working distance.

[0008] Furthermore, the frame rate of the visible light imaging module and the sampling frequency of the gas tunable semiconductor laser absorption spectroscopy detection module The following conditions are met: ; ; ; ; in, is the lower limit of the visible light image sampling frame rate, is the lower limit of TDLAS data sampling rate, The moving angular velocity of the gas leak detection and rapid tracing device; It is an experience value, take 1.

[0009] Furthermore, when the visible light imaging module and the gas tunable semiconductor laser absorption spectrum detection module are installed, the distance between the optical center of the visible light imaging module and the optical center of the gas tunable semiconductor laser absorption spectrum detection module is The following conditions are met: ; in, is the laser beam waist width, is the laser Rayleigh distance; It is the experience value, take 10.

[0010] Furthermore, when the signal synchronization acquisition module is synchronously acquiring data, the deviation between the time when the visible light imaging module acquires the image and the time when the gas tunable semiconductor laser absorption spectrum detection module acquires the concentration is controlled within the range Inside, and the Expressed as: ; in, The moving angular velocity of the gas leak detection and rapid tracing device; It is an experience value, take 1.

[0011] A gas leak detection and rapid source tracing method comprises the following steps: Step S1, performing leak detection using a gas leak detection and rapid source tracing device, wherein the leak judgment criterion is the concentration of leaked gas detected by a tunable semiconductor laser absorption spectroscopy detection module; Step S2: The visible light imaging module captures the image of the scene to obtain a basic image. , and will take the base image The coordinate system of the visible light imaging module is defined as the basic imaging coordinate system ; Step S3, establishing a gas leakage concentration field model using fluid mechanics constraints; Step S4: Use the signal synchronization acquisition module to control the tunable semiconductor laser absorption spectrum detection module and the visible light imaging module to synchronously acquire the leakage concentration at time t. and the current image ; Step S5: In the same scene and at a similar time, the position of the gas leak detection and rapid tracing device is moved so that the detection point of the tunable semiconductor laser absorption spectroscopy detection module moves along a certain trajectory and the leakage concentration along the path is obtained. and the current image , get n data pairs on this trajectory ( )、( )、( ),…,( ); Step S6: perform image registration based on n data pairs and project all images into the basic imaging coordinate system. In the corresponding basic imaging coordinate system Multi-point concentration value ( )、( )、( ),…,( ), where x is the two-dimensional coordinate of the image; Step S7: Fit the gas leakage concentration field model obtained in steps S2 and S3, combined with the multi-point concentration values ​​obtained in step S6. ; Step S8: Based on the gas leakage concentration field model obtained in steps S2 and S3, combined with the basic image , calculate the current image and the base image The registration relationship between ; Using the registration relationship , the concentration field Align to current image and locate the leakage source based on the aligned images.

[0012] Furthermore, the gas leakage detection and rapid tracing device adopts the above-mentioned gas leakage detection and rapid tracing device.

[0013] Furthermore, in step S2, the basic image Perform spatial perception to identify candidate targets where leakage may occur and use them as spatial constraints for the gas leakage concentration field model; When spatial perception is used, semantic segmentation is used to classify the base image Perform pixel-level classification and classify the basic image It is divided into flanges, pipelines, gas tanks, and background; among them, flanges, pipelines, and gas tanks are targets where leakage may occur.

[0014] Furthermore, in step S3, when establishing a gas leakage concentration field model, the model is: ; and They are Direction and The center of the gas cloud in the direction of Relevant, namely: ; ; in, The positive direction of the axis points to the downwind direction. The positive axis points in the direction of sight. The positive direction of the axis is vertical and the horizon is upward. is the gas leakage rate, for Wind speed in direction, and They are Direction and Variance of direction; represents the transfer factor of the gas cloud center along the y direction, represents the movement factor of the gas cloud center along the z direction, represents the offset factor of the gas cloud center along the y direction, represents the offset factor of the gas cloud center along the z direction. Furthermore, in step S7, the concentration field of the gas leakage is fitted. When , the fitting steps are: Step S7-1, based on the multi-point concentration values ​​in the basic imaging coordinate system The distribution in , calculates the possible foreground intersection points of the leakage area; Step S7-2, taking the foreground intersection as the origin, fitting the gas leakage concentration field model established in step S3 using the multi-point data, and calculating the fitting error; Step S7-3: If the error is too large, update the origin until the origin with smaller fitting error is found, and then fit to establish the final concentration field. .

[0015] Further, in step S8, the concentration field Align to current image When , the alignment method is: Step S8-1, calculate the basic image To the current image Feature point matching: Step S8-2: Calculate the basic image based on the matched feature points To the current image The transformation matrix ; Step S8-3, using the transformation matrix , the concentration field Align to current image middle.

[0016] The beneficial effects of the present invention are as follows: 1. In the present invention, the established gas leakage concentration field model is subject to spatial physical constraints and fluid mechanics physical constraints, and its spatial distribution has certain rules. Through scene understanding and fluid mechanics models, mathematical models such as Gaussian feathering models can be used to simplify the spatial distribution of gas concentration; by combining the concentration values ​​of multiple spatial positions collected by the tunable semiconductor laser absorption spectrum detection module, the spatial distribution of gas concentration is fitted, and the spatial distribution is projected into the visible light image coordinate system to achieve rapid positioning of the gas leakage source with high positioning accuracy, thereby achieving the rapid tracing capability of gas leakage that cannot be achieved by visible light wireless in traditional methods, and using lower-cost visible light imaging modules to replace higher-cost infrared sensors, effectively reducing the cost of gas leakage and tracing, which is conducive to the large-scale promotion and application of the gas leakage and tracing technology.

[0017] 2. In the present invention, spatial perception is performed on the basic image, and semantic segmentation is used to perform pixel-level classification of the scene. The image is divided into categories such as flanges, pipes, gas tanks, and background. Except for the background, the remaining foreground targets are all targets where leakage may occur, and they are used as spatial constraints for the gas leakage concentration field model. The gas leakage concentration field must be connected to an instance of the foreground target, and the concentration is highest near the connection point, which can greatly improve the time of locating the source of the gas leakage and achieve rapid positioning of the leakage source.

[0018] 3. In the present invention, a specific method is used to convert the concentration field Align to current image In the present invention, by pseudo-coloring the concentration field P and superimposing it on the current image I, it is convenient to quickly locate the leakage source, greatly shorten the time required to locate the leakage source, greatly improve the efficiency of locating the leakage source, and realize the rapid location of the leakage source. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a structural schematic diagram of the present invention; Among them, the dotted box is an optional functional module; Figure 2Schematic diagram of the scanning trajectory in the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments.

[0021] Therefore, based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative work shall fall within the scope of protection of the present invention.

[0022] Example 1 This embodiment provides a gas leak detection and rapid source tracing device that tightly integrates visible light imaging and a gas tunable semiconductor laser absorption spectroscopy detection module (i.e., TDLAS). It designs a TDLAS leakage concentration acquisition trajectory and combines the spatial perception of visible light images with the fluid mechanics constraints of gas leaks to achieve large-scale leakage concentration field estimation and rapid source tracing. Figure 1 As shown, it includes a visible light imaging module, a gas tunable semiconductor laser absorption spectrum detection module, a signal synchronization acquisition module and a data processing unit.

[0023] Among them, the visible light imaging module is used to image the leaking gas scene and provide the background and basic coordinate system for leak positioning.

[0024] The gas tunable semiconductor laser absorption spectroscopy detection module is used to collect the gas leakage concentration at the leakage point in the scene.

[0025] The signal synchronization acquisition module is used to synchronize the data acquisition of the visible light imaging module and the gas tunable semiconductor laser absorption spectrum detection module.

[0026] The data processing unit is used to realize gas leakage model building, concentration field fitting and projection.

[0027] In addition, a display screen, battery, and motion actuator can be optionally configured as needed. The display screen is used to display the detection and traceability process and results, the battery is used to provide power to the entire device, and the motion actuator is used to control imaging and the detection direction of TDLAS, including horizontal and vertical rotation. The display screen, battery, and motion actuator can directly adopt existing technologies without the need for creative work.

[0028] There is an angle between the optical axis of the visible light imaging module and the optical axis of the gas tunable semiconductor laser absorption spectrum detection module , and the angle Satisfaction relationship: ; angle It can be solved by the following formula: ; in, is the upper limit of the angle, is the distance between the optical center of the visible light imaging module and the optical center of the gas tunable semiconductor laser absorption spectroscopy detection module, is the nominal working distance.

[0029] The angle The calibration method can be used for calibration and fixation, and no secondary calibration is required during use. The calibration can be performed using an existing angle calibration method or an innovative calibration method (the innovative calibration method will be described and protected in another patent filed along with this application). In this embodiment, the existing calibration method can be used.

[0030] Frame rate of visible light imaging module and the sampling frequency of the gas tunable semiconductor laser absorption spectroscopy detection module The following conditions are met: ; ; Should and All with nominal working distance And it is related to the angular velocity of the device, specifically: ; ; in, is the lower limit of the visible light image sampling frame rate, is the lower limit of TDLAS data sampling rate, The moving angular velocity of the gas leak detection and rapid tracing device; It is an experience value, take 1.

[0031] When installing the visible light imaging module and the gas tunable semiconductor laser absorption spectrum detection module, the distance between the optical centers of the two is required to be as small as possible. In this embodiment, the specific standards for the distance between the optical centers of the visible light imaging module and the gas tunable semiconductor laser absorption spectrum detection module are: The distance between the optical center of the visible light imaging module and the optical center of the gas tunable semiconductor laser absorption spectroscopy detection module Should be less than the nominal working distance of the gas tunable semiconductor laser absorption spectroscopy detection module The radius of the light spot generated, the distance The following conditions should be met: ; in, is the laser beam waist width, is the laser Rayleigh distance; It is the experience value, take 10.

[0032] The signal synchronization acquisition module has software synchronization and hardware synchronization acquisition functions. When the signal synchronization acquisition module is used to synchronize data acquisition, the deviation between the time when the visible light imaging module collects images and the time when the gas tunable semiconductor laser absorption spectrum detection module collects concentrations should be controlled within an acceptable range. Inside, and the With nominal working distance , the angular velocity of the device It can be expressed as: ; in, The moving angular velocity of the gas leak detection and rapid tracing device; It is an experience value, take 1.

[0033] The data processing unit is composed of an ARM or DSP. The function of the data processing unit is to complete scene modeling, concentration field fitting and projection, which is achieved by implanting the algorithms and programs described in this application.

[0034] Example 2 This embodiment provides a gas leak detection and rapid source tracing method, which uses the above-mentioned gas leak detection and rapid source tracing device to quickly locate and trace the gas leak, and specifically includes the following steps: Step S1 : performing leakage detection using a gas leakage detection and rapid tracing device, wherein the leakage judgment criterion is the concentration of leaked gas detected by a tunable semiconductor laser absorption spectroscopy detection module.

[0035] Step S2: The visible light imaging module captures the image of the scene to obtain a basic image. , and will take the base image The coordinate system of the visible light imaging module is defined as the basic imaging coordinate system .

[0036] For the base image Perform spatial perception to identify candidate targets where leakage may occur and use them as spatial constraints for the gas leakage concentration field model.

[0037] When performing spatial perception, semantic segmentation is used to classify the base image Perform pixel-level classification and classify the basic image The model is divided into categories such as flanges, pipes, gas tanks, and background. In addition to the background, foreground objects such as flanges, pipes, and gas tanks are all potential leak targets. The spatial constraint of the leakage model is that the gas leakage concentration field must be connected to an instance of the foreground object, flange, pipe, gas tank, or background, and the concentration is highest near the connection.

[0038] When performing semantic segmentation, this example uses SAM, which is well known in the art, for semantic segmentation, that is, existing technology can be used. Those skilled in the art can choose and apply it as needed without having to make any creative efforts.

[0039] Step S3: establishing a gas leakage concentration field model using fluid mechanics constraints.

[0040] In this embodiment, when establishing the gas leakage concentration field model, a simplified model, namely the Gaussian feathering model, is adopted. The model is: ; and They are Direction and The center of the gas cloud in the direction of The correlation model between the two is: ; ; in, The positive direction of the axis points to the downwind direction. The positive axis points in the direction of sight. The positive direction of the axis is vertical and the horizon is upward. is the gas leakage rate, for Wind speed in direction, and They are Direction and Variance of direction; represents the transfer factor of the gas cloud center along the y direction, represents the movement factor of the gas cloud center along the z direction, represents the offset factor of the gas cloud center along the y direction, represents the offset factor of the gas cloud center along the z direction.

[0041] Step S4: Use the signal synchronization acquisition module to control the tunable semiconductor laser absorption spectrum detection module and the visible light imaging module to synchronously acquire the leakage concentration at time t. and the current image .

[0042] Step S5: In the same scene and at a similar time, the position of the mobile gas leak detection and rapid tracing device is adjusted so that the point detected by the tunable semiconductor laser absorption spectrum detection module moves along a certain trajectory. The trajectory is related to the wind direction at the current moment. The trajectory along the wind direction is longer, and the trajectory perpendicular to the wind direction is shorter, such as Figure 2 During use, the trajectory can be automatically determined by the device after collecting the wind speed and direction on site, and then drawn on the display screen, making it convenient for the operator to scan the trajectory. and the current image , get n data pairs on this trajectory ( )、( )、( ),…,( In this embodiment, 100 data pairs located on the trajectory are captured ( )、( )、( ),…,( ).

[0043] Step S6: perform image registration based on n data pairs and project all images into the basic imaging coordinate system. In the corresponding basic imaging coordinate system Multi-point concentration value ( )、( )、( ),…,( ), where x is the two-dimensional coordinate of the image.

[0044] According to the 100 data pairs collected in step S5, image registration is performed. The registration process is as follows: the current image and the base image The feature point pairs that match between the two images are then calculated. and the base image The affine matrix projects all images into the basic imaging coordinate system In the image, the image corresponding to the basic imaging coordinate system is the basic image Finally, the multi-point concentration values ​​in the basic imaging coordinate system are obtained ( )、( )、( ),…,( ), where x is the two-dimensional coordinate of the image.

[0045] Step S7: Fit the gas leakage concentration field model obtained in steps S2 and S3, combined with the multi-point concentration values ​​obtained in step S6. .

[0046] Fitting the concentration field of gas leakage When , the fitting steps are: Step S7-1, based on the multi-point concentration values ​​in the basic imaging coordinate system The distribution in , calculates the possible foreground intersection points of the leakage area.

[0047] In step S7-2, the gas leakage concentration field model established in step S3 is fitted with the multi-point data using the foreground intersection as the origin, and the fitting error is calculated. When fitting, the existing fitting method can be used without creative work.

[0048] Step S7-3: If the error is too large, update the origin until the origin with smaller fitting error is found, and then fit to establish the final concentration field. .

[0049] Step S8: Based on the gas leakage concentration field model obtained in steps S2 and S3, combined with the basic image , calculate the current image and the base image The registration relationship between ; Using the registration relationship , the concentration field Align to current image and locate the leakage source based on the aligned images.

[0050] In the concentration field Align to current image When , the alignment method is: Step S8-1, calculate the basic image To the current image Feature point matching: This embodiment can use the SIFT method commonly used in the art to perform feature matching.

[0051] Step S8-2: Calculate the basic image based on the matched feature points To the current image The transformation matrix .

[0052] Step S8-3, using the transformation matrix , the concentration field Align to current image middle.

[0053] By converting the concentration field Perform pseudo color processing and superimpose it on the current image It is convenient for operators to quickly locate the source of the leak.

Claims

1. A gas leak detection and rapid source tracing device, comprising a gas tunable semiconductor laser absorption spectrum detection module, a signal synchronization acquisition module and a data processing unit; Features: Also included is a visible light imaging module; The visible light imaging module is used to image the leaking gas scene and provide the background and basic coordinate system for leak positioning; The gas tunable semiconductor laser absorption spectrum detection module is used to collect the gas leakage concentration at the leakage point in the scene; The signal synchronization acquisition module is used to synchronize the data acquisition of the visible light imaging module and the gas tunable semiconductor laser absorption spectrum detection module; The data processing unit is used to realize gas leakage modeling, concentration field fitting and projection; There is an angle between the optical axis of the visible light imaging module and the optical axis of the gas tunable semiconductor laser absorption spectrum detection module , and the angle Satisfaction relationship: ； ； in, is the upper limit of the angle, is the distance between the optical center of the visible light imaging module and the optical center of the gas tunable semiconductor laser absorption spectrum detection module, is the nominal working distance.

2. A gas leak detection and rapid source tracing device as claimed in claim 1, characterized in that: Frame rate of visible light imaging module and the sampling frequency of the gas tunable semiconductor laser absorption spectroscopy detection module The following conditions are met: ； ； ； ； in, is the lower limit of the visible light image sampling frame rate, is the lower limit of TDLAS data sampling rate, The moving angular velocity of the gas leak detection and rapid source tracing device; It is an experience value, take 1.

3. A gas leak detection and rapid source tracing device as claimed in claim 1, characterized in that: When installing the visible light imaging module and the gas tunable semiconductor laser absorption spectrum detection module, the distance between the optical center of the visible light imaging module and the optical center of the gas tunable semiconductor laser absorption spectrum detection module The following conditions are met: ； in, is the laser beam waist width, is the laser Rayleigh distance; It is the experience value, take 10.

4. A gas leak detection and rapid source tracing device as claimed in claim 1, characterized in that: When the signal synchronization acquisition module is synchronously acquiring data, the deviation between the time when the visible light imaging module acquires the image and the time when the gas tunable semiconductor laser absorption spectrum detection module acquires the concentration is controlled within the range within, and the It is expressed as: ； in, The moving angular velocity of the gas leak detection and rapid source tracing device; It is an experience value, take 1.

5. A method for detecting and quickly tracing a gas leak, characterized in that: The steps include: Step S1, using a gas leak detection and rapid source tracing device to perform leak detection, the leak judgment criterion being the concentration of leaked gas detected by a tunable semiconductor laser absorption spectroscopy detection module; Step S2: The visible light imaging module captures the image of the scene to obtain a basic image. , and will take the base image The coordinate system of the visible light imaging module is defined as the basic imaging coordinate system ; Step S3, establishing a gas leakage concentration field model using fluid mechanics constraints; Step S4, using the signal synchronization acquisition module to control the tunable semiconductor laser absorption spectrum detection module and the visible light imaging module to synchronously acquire the leakage concentration at time t and the current image ; Step S5, in the same scene and at a similar time, the position of the mobile gas leakage detection and rapid tracing device is adjusted so that the point detected by the tunable semiconductor laser absorption spectrum detection module moves along a certain trajectory, and the leakage concentration on the path is obtained. and the current image , get n data pairs on this trajectory ( )、( )、( )、…、( ); Step S6: perform image registration based on n data pairs and project all images into the basic imaging coordinate system. In the corresponding basic imaging coordinate system Multi-point concentration value ( )、( )、( )、…、( ), where x is the two-dimensional coordinate of the image; Step S7, fitting the gas leakage concentration field model obtained in steps S2 and S3 in combination with the multi-point concentration values ​​obtained in step S6 ; Step S8, based on the gas leakage concentration field model obtained in steps S2 and S3, combined with the basic image , calculate the current image and the base image The registration relationship between ; Using the registration relationship , the concentration field Align to current image and locate the leakage source based on the aligned images.

6. A gas leak detection and rapid source tracing method as claimed in claim 5, characterized in that: The gas leakage detection and rapid source tracing device adopts the gas leakage detection and rapid source tracing device described in any one of claims 1-4.

7. A gas leak detection and rapid source tracing method as claimed in claim 5, characterized in that: In step S2, the base image Perform spatial perception to identify candidate targets where leakage may occur and use them as spatial constraints for the gas leakage concentration field model; When spatially aware, semantic segmentation is used to segment the base image Perform pixel-level classification and classify the basic image It is divided into flange, pipeline, gas tank and background; among them, flange, pipeline and gas tank are the targets where leakage may occur.

8. A gas leak detection and rapid source tracing method as claimed in claim 5, characterized in that: In step S3, when establishing a gas leakage concentration field model, the model is: ； and They are Direction and The center of the gas cloud in the direction of Relevant, namely: ； ； in, The positive axis points downwind. The positive axis points in the direction of sight. The positive axis is vertical with the horizon pointing upwards. is the gas leakage rate, for Direction of wind speed, and They are Direction and Variance of direction; represents the transfer factor of the gas cloud center along the y direction, represents the movement factor of the gas cloud center along the z direction, represents the offset factor of the gas cloud center along the y direction, Represents the offset factor of the gas cloud center along the z direction.

9. A gas leak detection and rapid source tracing method as claimed in claim 5, characterized in that: In step S7, the concentration field of the gas leakage is fitted When , the fitting steps are: Step S7-1, based on the multi-point concentration values ​​in the basic imaging coordinate system The distribution in, calculates the possible foreground intersection points of the leakage area; Step S7-2, taking the foreground intersection point as the origin, fitting the gas leakage concentration field model established in step S3 using multi-point data, and calculating the fitting error; Step S7-3: If the error is too large, update the origin until an origin with a smaller fitting error is found, and fit to establish the final concentration field. .

10. A gas leakage detection and rapid source tracing method as claimed in claim 5, characterized in that: In step S8, the concentration field Align to current image When , the alignment method is: Step S8-1, calculate the basic image To current image Feature point matching: Step S8-2, calculate the basic image based on the matched feature points To current image The transformation matrix ; Step S8-3, using the transformation matrix , the concentration field Align to current image middle.

Images