Acoustic-optical fusion cold spring methane detection method, system, device and storage medium

Abstract

The application discloses a kind of acoustic-optical fusion's cold spring methane detection method, system, device and storage medium, comprising: obtaining first preprocessed acoustic measurement data and second preprocessed optical measurement data, obtains environmental measurement data;According to the acoustic measurement data, the optical measurement data and the environmental measurement data, form joint data;According to the joint data extraction joint feature, according to the joint feature determines current methane total concentration, according to the joint data determines current methane total leakage flux.The embodiment of the application can improve the accuracy of cold spring methane detection, and can be widely applied in gas detection technical field.

Description

Technical Field

[0001] This invention relates to the field of gas detection technology, and in particular to a method, system, device, and storage medium for detecting methane from cold seeps using a combination of acoustic and optical methods. Background Technology

[0002] Cold seeps are seepage activities that occur continuously or intermittently on the seabed, originating below the seafloor sedimentary interface and containing fluids (mainly hydrocarbons) such as methane and hydrogen sulfide, which are widely developed along continental margins. These fluid seepage processes can induce submarine geological hazards (such as landslides and mud volcanoes) and pose a significant threat to global climate change. Therefore, accurately quantifying cold seep fluxes and analyzing their activity patterns has significant scientific and engineering implications.

[0003] The plumes formed by the upward movement of fluids during submarine cold seep activity are the main targets of cold seep detection. Among the related technologies, the accuracy of cold seep methane detection needs to be further improved. Summary of the Invention

[0004] In view of this, the purpose of this invention is to provide an acoustic-optical fusion method, system, device, and storage medium for cold seep methane detection, which can improve the accuracy of cold seep methane detection.

[0005] On one hand, embodiments of the present invention provide an acoustic-optical fusion method for detecting methane in cold seeps, comprising:

[0006] Acquire acoustic measurement data after the first preprocessing and optical measurement data after the second preprocessing, and acquire environmental measurement data;

[0007] Based on the acoustic measurement data, the optical measurement data, and the environmental measurement data, joint data is formed;

[0008] Based on the combined data, joint features are extracted, the current total methane concentration is determined based on the joint features, and the current total methane leakage flux is determined based on the combined data.

[0009] In some embodiments, forming joint data based on the acoustic measurement data, the optical measurement data, and the environmental measurement data includes:

[0010] The acoustic measurement data is used to determine the sound source location of the methane bubbles, and the optical measurement data is mapped to a preset coordinate system.

[0011] Based on the location of the sound source, obtain optical measurement data of the region of interest in the preset coordinate system;

[0012] Spatiotemporal matching is performed on the geometry data, acoustic data, and environmental data of the region of interest to establish a multimodal dataset and form joint data.

[0013] In some embodiments, the step of extracting joint features from the joint data and determining the current total methane concentration based on the joint features includes:

[0014] Joint features are extracted from the joint data, and model source data is constructed based on the joint features; the model source data includes acoustic source data and optical source data.

[0015] The concentration of methane in the bubble phase is calculated based on the acoustic source data, and the concentration of methane in the dissolved phase is calculated based on the optical source data.

[0016] The current total methane concentration is determined based on the methane concentration in the bubble phase and the methane concentration in the dissolved phase.

[0017] In some embodiments, determining the current total methane leakage flux based on the combined data includes:

[0018] The combined data were used to determine the methane bubble-state population flux and the methane dissolved-state flux.

[0019] The current total methane leakage flux is determined based on the methane bubble-state population flux and the methane dissolved-state flux.

[0020] In some embodiments, the acousto-optic fusion method for detecting methane in cold seeps further includes:

[0021] Predict the methane conversion concentration based on the current total methane leakage flux;

[0022] The future total methane concentration is predicted based on the methane conversion concentration and the current total methane concentration.

[0023] In some embodiments, the acousto-optic fusion method for detecting methane in cold seeps further includes:

[0024] Construct a deep learning model and training sample data; the deep learning model includes a Transformer module and an LSTM module, and the training sample data includes diffusion coefficients, environmental measurement data, and methane concentration fields;

[0025] The deep learning model is trained based on the training sample data, and the methane plume concentration field distribution is predicted based on the trained deep learning model.

[0026] On the other hand, embodiments of the present invention provide an acoustic-optical fusion cold seep methane detection system, comprising:

[0027] The acquisition module is used to acquire the first pre-processed acoustic measurement data and the second pre-processed optical measurement data, and to acquire environmental measurement data.

[0028] The data processing module is used to generate joint data based on the acoustic measurement data, the optical measurement data, and the environmental measurement data;

[0029] The methane calculation module is used to extract joint features based on the joint data, determine the current total methane concentration based on the joint features, and determine the current total methane leakage flux based on the joint data.

[0030] On the other hand, embodiments of the present invention provide an acoustic-optical fusion cold seep methane detection device, comprising:

[0031] At least one processor;

[0032] At least one memory for storing at least one program;

[0033] When the at least one program is executed by the at least one processor, the at least one processor performs the method described above.

[0034] On the other hand, embodiments of the present invention provide a computer-readable storage medium storing a processor-executable program, which, when executed by a processor, is used to perform the above-described method.

[0035] On the other hand, embodiments of the present invention provide an acoustic-optical fusion cold seep methane detection system, including a computer device and an acoustic detector, an optical detector, and an environmental detector connected to the computer device; wherein,

[0036] The acoustic detector is used to collect acoustic measurement data;

[0037] The optical detector is used to collect optical measurement data;

[0038] The environmental detector is used to collect environmental measurement data;

[0039] The computer device includes:

[0040] At least one processor;

[0041] At least one memory for storing at least one program;

[0042] When the at least one program is executed by the at least one processor, the at least one processor performs the method described above.

[0043] The implementation of this invention includes the following beneficial effects: This embodiment first acquires acoustic measurement data after first preprocessing and optical measurement data after second preprocessing, and acquires environmental measurement data. The acoustic measurement data can characterize the bubble phase methane data, and the optical measurement data can characterize the dissolved phase methane data. Then, based on the acoustic measurement data, optical measurement data, and environmental measurement data, joint data is formed. The joint data integrates bubble phase methane data and dissolved phase methane data. Finally, joint features are extracted based on the joint data, and the current total methane concentration is determined based on the joint features. The current total methane leakage flux is determined based on the joint data. In the process of calculating the methane concentration and leakage flux, multiple phases of methane data are considered, which improves the accuracy of cold seep methane detection. Attached Figure Description

[0044] Figure 1 This is a schematic flowchart of a method for detecting methane in cold seeps using a combination of acoustic and optical methods, provided in an embodiment of the present invention.

[0045] Figure 2 This is a schematic diagram of a process for forming joint data according to an embodiment of the present invention;

[0046] Figure 3 This is a schematic diagram of a process for determining the current total methane concentration according to an embodiment of the present invention;

[0047] Figure 4 This is a schematic diagram of a process for determining the current total methane leakage flux provided by an embodiment of the present invention;

[0048] Figure 5 This is a schematic diagram of a process for predicting future methane concentration provided by an embodiment of the present invention;

[0049] Figure 6 This is a schematic flowchart of a step-by-step method for predicting the concentration field distribution of a methane plume, provided by an embodiment of the present invention.

[0050] Figure 7 This is a structural block diagram of an acoustic-optical fusion cold seep methane detection system provided in an embodiment of the present invention;

[0051] Figure 8 This is a structural block diagram of an acoustic-optical fusion cold seep methane detection device provided in an embodiment of the present invention;

[0052] Figure 9 This is another structural block diagram of an acoustic-optical fusion cold seep methane detection system provided in an embodiment of the present invention. Detailed Implementation

[0053] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. The step numbers in the following embodiments are only for ease of explanation and do not limit the order of the steps. The execution order of each step in the embodiments can be adapted according to the understanding of those skilled in the art.

[0054] It should be noted that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., used in the specification, claims, and the foregoing drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatuses.

[0055] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0056] like Figure 1 As shown, this embodiment of the invention provides a method for detecting methane in cold seeps using a combination of sound and light, including steps S100 to S300.

[0057] S100: Acquire the first pre-processed acoustic measurement data and the second pre-processed optical measurement data, and acquire environmental measurement data.

[0058] The preprocessing procedure for acoustic measurement data is as follows:

[0059] Bubble Resonance Frequency Band Extraction: This step involves separating environmental noise using bandpass filtering to extract the characteristic frequency band of the bubble's acoustic radiation signal. The specific steps are as follows: 1) Environmental Noise Analysis: Using the raw sound pressure signal collected by a hydrophone, analyze the spectral characteristics of the background noise and identify the main frequency band range. Sound sources include biological noise (e.g., fish noise), mechanical noise (e.g., ship engine noise), and natural noise (e.g., ocean waves, ocean currents). 2) Bandpass Filtering: Based on the theoretical range of the bubble's resonance frequency (typically 1kHz to 50kHz), design a bandpass filter to filter out low-frequency environmental noise (e.g., ship noise, ocean wave noise) and high-frequency interference signals (e.g., electronic noise). 3) Signal Enhancement: Perform wavelet transform on the filtered signal to enhance the characteristic frequency band of the bubble's acoustic radiation signal. 4) Feature Extraction: Extract the time-domain waveform, spectrum, and phase information of the bubble's resonance frequency band to generate a timestamp-aligned sound pressure matrix P(t), providing a data foundation for subsequent bubble size distribution inversion.

[0060] The preprocessing of optical measurement data is as follows: 1) The Gordon model is used to remove the influence of water scattering. The Gordon model can be used to separate water radiation L. w (λ) and scattered radiation L s (λ). Its core idea is to model the scattering of water bodies to obtain the total radiation L. t Subtracting the scattered radiation from (λ) yields the pure water radiation L. w (λ), the formula is as follows:

[0061]

[0062] Scattered radiation It can be calculated using the following formula:

[0063]

[0064] In the formula: For sky radiation; Scattering reflectance is a quantitative indicator of the scattering characteristics of a water surface, and can be calculated using the following formula:

[0065]

[0066] In the formula: The absorption coefficient of water describes the ability of water to absorb light, and its unit is m. -1 It depends on the dissolved substances in the water (such as dissolved organic matter, pigments, etc.) and the absorption characteristics of pure water, and can be obtained by photometry. The scattering coefficient of a water body describes its ability to scatter light, and its unit is m. -1It depends on the scattering characteristics of suspended particles (such as silt, plankton, bubbles, etc.) and pure water in the water body, and can be obtained by measuring with a backscatter meter.

[0067] Scattering reflectivity matrix R s The number of rows and columns and the total radiation matrix L t The number of rows and columns are the same, and they are all matrices based on wavelength λ and spatial coordinates (x, y). Assuming that the water body is homogeneous within a small spatial area during in-situ observation (i.e., a(λ) and b(λ) are spatially invariant), then R... s Each element in the array has the same value.

[0068] The water radiation after scattering removal is a three-dimensional matrix with dimensions (λ, x, y). The water radiation L... w (λ) is converted to radiance for easier subsequent analysis. Radiance is a parameter describing the radiant energy of a water surface, representing the radiant energy per unit area, unit solid angle, and unit wavelength. The calculation formula is as follows:

[0069]

[0070] In the formula: θ The zenith angle is the solar zenith angle, which is the angle between the sun's rays and the normal to the Earth's surface, measured in degrees. At that time, sunlight is perpendicular to the Earth's surface (the sun is directly overhead); when... At that time, the sun's rays are parallel to the Earth's surface (the sun is on the horizon).

[0071] 2) Methane characteristic index:

[0072]

[0073] The numerical values ​​need to be calibrated in the laboratory. A standard methane solution can be used for spectral measurement to determine the core absorption peak band and the near-zero value band, and to establish a quantitative relationship model between methane concentration and characteristic index.

[0074] It should be noted that the environmental measurement data is determined based on the actual application, and this embodiment does not impose specific limitations. Examples include temperature, salinity, or depth. In a specific embodiment, the acquired measurement data includes: a timestamp-aligned sound pressure matrix P(t) (containing time-domain waveform, spectrum, and phase information), a hyperspectral cube I(λ,x,y), an environmental vector E=[T (temperature), S (salinity), D (depth), Turbine (turbidity), v (flow velocity)], and location information: GPS coordinates.

[0075] S200. Combined data is generated based on acoustic measurement data, optical measurement data, and environmental measurement data.

[0076] Specifically, acoustic measurement data, optical measurement data, and environmental measurement data are spatiotemporally aligned to form joint data. First, a mapping relationship is established between acoustic and optical measurement data, and then environmental measurement data is associated with and spatiotemporally matched to construct multimodal data, thus forming joint data.

[0077] S300. Extract joint features from the joint data, determine the current total methane concentration based on the joint features, and determine the current total methane leakage flux based on the joint data.

[0078] Joint features are extracted from the joint data for multiple modalities. Based on these joint features, methane concentrations for multiple phases are inverted, and the current total methane concentration is determined based on these phase concentrations. The current total methane leakage flux for multiple phases is calculated using the joint features and partial data from the joint data, and the current total methane leakage flux for multiple phases is then determined.

[0079] The implementation of this invention includes the following beneficial effects: This embodiment first acquires acoustic measurement data after first preprocessing and optical measurement data after second preprocessing, and acquires environmental measurement data. The acoustic measurement data can characterize the bubble phase methane data, and the optical measurement data can characterize the dissolved phase methane data. Then, based on the acoustic measurement data, optical measurement data, and environmental measurement data, joint data is formed. The joint data integrates bubble phase methane data and dissolved phase methane data. Finally, joint features are extracted based on the joint data, and the current total methane concentration is determined based on the joint features. The current total methane leakage flux is determined based on the joint data. In the process of calculating the methane concentration and leakage flux, multiple phases of methane data are considered, which improves the accuracy of cold seep methane detection.

[0080] In some embodiments, see Figure 2 Based on acoustic measurement data, optical measurement data, and environmental measurement data, joint data is formed, including:

[0081] S210. Determine the sound source location of the methane bubbles based on the acoustic measurement data, and map the optical measurement data to the preset coordinate system;

[0082] S220. Obtain optical measurement data of the region of interest in the preset coordinate system based on the location of the sound source;

[0083] S230. Based on the geometry data, acoustic data, and environmental data of the region of interest, perform spatiotemporal matching to establish a multimodal dataset and form joint data.

[0084] Acoustic localization: The accuracy of acoustic localization is affected by hydrophone position errors, sound velocity calculation errors, and timestamp errors. To reduce these errors, this application uses high-precision positioning equipment to calibrate the hydrophone position, employs a high-precision CTD (Conductivity, Temperature, Depth) sensor to measure temperature, salinity, and depth data, and calculates the sound velocity using the Mackenzie sound velocity formula. GPS clock synchronization technology is also introduced to ensure the accuracy of the timestamp. Furthermore, multi-hydrophone joint localization and least squares optimization of sound source location estimation further improve localization accuracy.

[0085] Calculation of the bubble swarm center position Z based on the time difference method bubble (t), the formula is as follows:

[0086]

[0087] In the formula: t1, t2, and t3 are the timestamps of the bubble sound signals received by the three hydrophones; the sound velocity c is calculated in real time from the temperature, salinity, and depth data collected by the CTD. The accuracy of the sound velocity calculation can be improved by using the Mackenzie sound velocity formula, and its unit is meters per second (m / s). The empirical formula is as follows:

[0088]

[0089] In the formula: S is the salinity of seawater, in parts per thousand (‰), T is the temperature of seawater, in degrees Celsius, and D is the depth, in meters.

[0090] The three-dimensional spatial location of the bubble swarm is determined by solving the sound wave propagation equation.

[0091] The propagation equations of sound waves describe the propagation behavior of sound waves in a medium. In the deep-sea environment, the propagation of sound waves can be simplified to the following equation:

[0092]

[0093] In the formula: p is the sound pressure, in Pascals (Pa), and t is the time, in seconds.

[0094] The time-difference method calculates the three-dimensional spatial location of the sound source by measuring the time difference between the arrival times of sound waves at different hydrophones and combining this with the speed of sound. Assuming the positions of the three hydrophones are r1, r2, and r3, and the sound source is located at r... s The relationship between the time difference and distance difference of the sound waves arriving at each hydrophone is as follows:

[0095]

[0096] In the formula: The time difference between hydrophone i and hydrophone j Let be the distance from the sound source to the hydrophone i.

[0097] Based on the time difference, the following distance difference equation can be established:

[0098]

[0099] For three hydrophones, two independent distance difference equations can be obtained. Solving these distance difference equations using the least squares method yields the three-dimensional spatial location r of the sound source. s =(x s ,y s ,z s ).

[0100] For three hydrophones, two independent distance difference equations can be obtained:

[0101]

[0102] Rewrite the distance difference equation as an error function:

[0103]

[0104] Define the sum of squared errors:

[0105]

[0106] Minimize the sum of squared errors using gradient descent. The location r of the sound source is obtained. s .

[0107] Optical projection: Based on inertial navigation system (INS) data, the pixel coordinates of the hyperspectral image are mapped to the seabed absolute coordinate system. The specific steps are as follows:

[0108] 1) Obtain device attitude information (such as pitch angle) provided by INS. θ (roll angle φ, yaw angle Ψ) and position information t=(t x ,t y ,t z ) T t can also be called the translation vector. Pitch angle θ The x-axis represents the angle of rotation around the x-axis, describing the equipment's head-up or head-down attitude. The roll angle φ represents the angle of rotation around the y-axis, describing the equipment's left-right tilt attitude. The yaw angle Ψ represents the angle of rotation around the z-axis, describing the equipment's left-right turning attitude.

[0109] 2) Combine with camera parameters, such as focal length f Pixel size , Based on the camera's installation position and angle, a geometric mapping relationship between image pixels and absolute seabed coordinates is established. The camera intrinsic parameter matrix K describes the influence of focal length and pixel size on imaging, and its form is:

[0110]

[0111] In the formula: f x , f y for The focal length in a direction converts the physical focal length into a focal length in pixels, used to describe the scaling relationship between image pixels and physical space. c x and c y These are the coordinates of the image center point.

[0112] The camera extrinsic parameter matrix T describes the camera's mounting position and orientation, and its form is:

[0113]

[0114] The extrinsic parameter matrix T, also known as the transformation matrix, is composed of the rotation matrix R and the translation vector t.

[0115] The rotation matrix R is determined by the pitch angle. θ The roll angle φ and yaw angle Ψ are calculated and their forms are as follows:

[0116]

[0117] In the formula:

[0118]

[0119]

[0120] Using the camera intrinsic parameter matrix K and extrinsic parameter matrix T, establish the absolute seabed coordinates X in homogeneous coordinates. world =(X,Y,Z,1) T Image pixel coordinates u=(u,v,1) T Mapping relationship between them:

[0121]

[0122] In the formula: This is the abbreviated form of the camera extrinsic parameter matrix T. K is the inverse of the camera extrinsic matrix. -1 It is the inverse of the camera intrinsic parameter matrix.

[0123] 3) Perform geometric correction on the image to eliminate distortions caused by device posture and motion.

[0124] Data correlation: Bubble cluster center Z located acoustically bubbleUsing (t) as the center, the region of interest (ROI) is selected as the hyperspectral analysis area. Only the ROI near the sound source is analyzed, saving computational resources and improving analysis efficiency.

[0125] The selection of the Region of Interest (ROI) is based on acoustic positioning accuracy and optical resolution. The radius of the ROI is set to 3 times the acoustic positioning error, while considering optical resolution, the radius is set to 10-20 times the optical resolution. Therefore, the radius of the ROI is set to 2 meters to ensure coverage of the sound source location and capture of details in the hyperspectral image, thus balancing data accuracy and computational efficiency.

[0126] The specific steps are as follows:

[0127] 1) Determine the spatial range of the ROI based on the acoustic positioning results, acoustic positioning accuracy, and optical resolution.

[0128] 2) Extract spectral data within the ROI from the hyperspectral image for concentration inversion of dissolved methane.

[0129] 3) Spatiotemporal matching of spectral data within the ROI with acoustic data (such as bubble size distribution) and environmental data (such as temperature, salinity, and flow velocity) is performed to construct a multimodal dataset, as follows:

[0130] a) Time alignment: Timestamp all data based on GPS-synchronized clocks to ensure data correlation within the same time window.

[0131] Suppose the set of timestamps is a synchronous time series:

[0132]

[0133] in, Δt Data acquisition period (default 1s), t0 is the GPS synchronization start time, for any modal data D m (m represents acoustics / optics / environment), time labels are assigned according to the nearest neighbor principle:

[0134]

[0135] b) Coordinate system transformation: The acoustic positioning points, hyperspectral pixels, and environmental sampling points are mapped to the seabed absolute coordinate system through coordinate transformation.

[0136] The preprocessed sound pressure matrix P(t) is a three-dimensional matrix containing time-domain waveform, spectrum, and phase information, and its form is:

[0137]

[0138] In the formula: This represents a single sound pressure value in the sound pressure matrix, corresponding to time t. i,frequency f i and phase φ i .

[0139] Acoustic positioning point location The location of the bubble cluster center is calculated using the time difference method and the least squares method, and its form is as follows:

[0140]

[0141] In the formula: These represent the three-dimensional coordinates of the bubble swarm center in the absolute coordinate system of the seabed.

[0142] Generating a spatially aligned sound pressure matrix: This requires aligning each sound pressure value in the sound pressure matrix P(t). Acoustic positioning point Perform the association. First, based on the time-aligned timestamp t k Find the sound pressure value Corresponding acoustic positioning points Then the sound pressure value Mapped to acoustic localization point Its three-dimensional spatial position.

[0143]

[0144] Associate all sound pressure values ​​with acoustic localization points to generate a spatially aligned sound pressure matrix P. aligned (t).

[0145] Hyperspectral pixel coordinates:

[0146]

[0147] In the formula: (u,v) are the pixel coordinates within the ROI, and K is the camera intrinsic parameter matrix. This is the camera extrinsic parameter matrix.

[0148] Environmental parameter location: Updated in real time with GPS time, as shown in the following formula:

[0149]

[0150] c) Spatial gridding processing:

[0151] The grid G ​​of the absolute coordinate system on the seabed is a three-dimensional network, and its form is as follows:

[0152]

[0153] In the formula: These represent the three-dimensional coordinates of the grid points in the absolute seabed coordinate system. The horizontal grid size should be consistent with the optical pixel size, and the vertical grid size should be consistent with the depth interval of the CTD data.

[0154] For the environmental data T(x,y,z) of non-grid points, trilinear interpolation is used, and the calculation method is as follows:

[0155]

[0156] In the formula: W q As the weights of environmental parameters, These are the eight vertices in a 3D mesh. Trilinear interpolation calculates the value of interior points by weighted averaging the values ​​of these eight vertices. The weights are calculated inversely proportional to the Euclidean distance from each mesh vertex to the target point, as shown in the following formula:

[0157]

[0158] Unify the meshing of the sound pressure matrix with the three-dimensional mesh:

[0159] For each sound pressure value Find its nearest neighbor grid point in the 3D grid G. ,in:

[0160]

[0161]

[0162]

[0163] All sound pressure values Mapped into a 3D mesh G, generating a meshed sound pressure matrix. It can be represented as .

[0164] d) Construction of multimodal data foundation tensors: This ultimately generates multidimensional tensors to store the spatiotemporally matched foundational data.

[0165]

[0166] In the formula: t k The observation time (timestamp) has been discretized. x i and y j The planar grid positions of the ROI region; z l The vertical coordinates are measured from CTD data; the feature dimension M includes spectral vectors I(λ1), ..., I(λ). 120), where λ is the center wavelength of the nth band; the acoustic data is a gridded sound pressure matrix. The environment vector is defined as [T, S, D, v]. T The grid coordinates have been aligned using trilinear interpolation.

[0167] In some embodiments, see Figure 3 Based on the joint data, joint features are extracted, and the current total methane concentration is determined based on the joint features, including:

[0168] S310. Extract joint features from the joint data, and construct model source data based on the joint features; the model source data includes acoustic source data and optical source data.

[0169] S320. Calculate the methane concentration in the bubble phase based on acoustic source data, and calculate the methane concentration in the dissolved phase based on optical source data.

[0170] S330. Determine the current total methane concentration based on the methane concentration in the bubble phase and the methane concentration in the dissolved phase.

[0171] The feature extraction process for acoustic measurement data in the joint data is as follows:

[0172] 1) Bubble density: The methane bubble density is calculated using sound pressure signals, in kilograms per cubic meter. The calculation formula is as follows:

[0173]

[0174] In the formula, P rms : Sound pressure effective value, in Pascals, is the root mean square value of the sound pressure signal over one period, used to quantify the energy intensity of the sound wave; ρ: Seawater density, measured from CTD data, in kilograms per cubic meter; c: Sound speed, in meters per second, can be calculated using the Mackenzie sound speed formula mentioned above.

[0175] 2) Resonance frequency: A characteristic of bubble vibration, measured in Hertz (Hz). A Fast Fourier Transform (FFT) can be performed on P(t) to obtain the spectrum S(f), and the frequency corresponding to the maximum amplitude in the 1-10 kHz band can be found. f res .

[0176] f res = argmax(FFT(P(t)))

[0177] 3) Bubble radius: Since the bubble undergoes Mie resonance in the sound field, the resonant frequency f needs to be determined. res The formula for calculating the bubble radius r in meters is as follows:

[0178]

[0179] In the formula: f res ρ is the resonant frequency mentioned above; ρ is the density of seawater, in kilograms per cubic meter; γ is the specific heat ratio of methane, which is the specific heat of methane at constant pressure / specific heat at constant volume (1.31), which can be obtained from the table of physical constants and is a dimensionless physical constant; P0 is the hydrostatic pressure, which can be calculated from CTD depth data, in Pascals.

[0180] The feature extraction process for optical measurement data in the joint data is as follows:

[0181] Methane saturation: The maximum dissolved concentration of methane in seawater, expressed in moles per cubic meter. The calculation formula is:

[0182]

[0183] In the formula: k H P is the Henry's constant, with units of moles per cubic meter per pascal, which can be obtained from a table; CH4 The partial pressure of methane is expressed in Pascals and can be measured by the acoustic methane bubble density ρ. bubble The calculation formula is:

[0184]

[0185] In the formula: R is the ideal gas constant, which is 8.314 J / (mol·K), M is the molar mass of methane, which is 16 g / mol, and T is the temperature in Kelvin.

[0186] The feature extraction process for environmental measurement data in the joint data is as follows:

[0187] Rise velocity: Reflects the mass of methane transported by the bubble per unit time, and is a key parameter for flux calculation, measured in meters per second. The rise velocity of the bubble is constrained by both viscous forces (small bubbles) and inertial forces (large bubbles). Based on the theory of bubble motion dynamics and the core idea of ​​the classic Clift equation, this paper simplifies and improves the equation, selecting the smaller value of the viscous force and the inertial force to ensure the physical rationality of the calculation results.

[0188] The improved Clift equation for calculating the rate of ascent is:

[0189]

[0190] Where: 0.71 is the Stokes coefficient of a spherical microbubble in a viscous fluid; 0.23 is the inertial term coefficient of a large bubble; μ is the viscosity of seawater, in Pascal-seconds (Pa·s); g is the acceleration due to gravity, in meters per second squared; r is the bubble radius, in meters; ρ waterρ is the density of seawater, which can be measured by CTD; Δρ is the density difference between the density of methane bubbles and the density of seawater.

[0191] However, during the ascent of cold seep bubbles, marine organic matter (dissolved organic carbon, DOC, proteins, lipids) adsorbs onto the bubble surface, forming an elastic film. This alters the interfacial tension, hinders gas exchange, and consequently reduces the ascent velocity. Therefore, the equation for calculating the actual ascent velocity is:

[0192]

[0193] In the formula: K surf The pollution factor is calculated using the following formula:

[0194]

[0195] In the formula: Δγ is the change in seawater surface tension caused by the adsorption of dissolved organic matter, in Newtons per meter; γ0 is the surface tension of pure seawater, in Newtons per meter; C DOC The concentration of dissolved organic matter, expressed in Newtons per meter, can be obtained by consulting relevant literature.

[0196] Constructing a derived feature dataset: A derived tensor is generated by the mapping function g(), which can be used as a direct input to the flux inversion model to ensure the spatiotemporal consistency of physical features.

[0197]

[0198] The acoustic inversion process is as follows:

[0199] The hydrophone inverts the concentration by capturing the acoustic scattering signal of methane bubbles. The bubbles undergo Mie scattering in the sound field, and the signal intensity is positively correlated with the number / size of the bubbles. The core formula is as follows, which is the analytical solution for a spherical scatterer:

[0200] 1) Bubble sound scattering cross section: A physical quantity describing the ability of a bubble to scatter sound waves, measured in square meters. Based on Mie scattering theory, the calculation formula is as follows:

[0201]

[0202] In the formula: r denoted by , where is the bubble radius in meters, which can be calculated using the bubble radius inversion formula described above; 'k' is a parameter describing the spatial frequency of the sound wave, representing the number of sound cycles per unit length. Its calculation formula is: k = 2π f / c, f The frequency of the sound wave measured by the hydrophone is in Hertz, and c is the speed of sound in meters per second, which can be calculated using the formula mentioned above.

[0203] 2) Signal-to-noise ratio (SNR): This is the ratio of methane scattered acoustic power to background noise power. It quantifies the intensity of the methane bubble acoustic signal and is primarily used to verify the reliability of acoustic inversion. SNR is positively correlated with the total scattering cross section of the bubbles, and consequently, with the methane concentration in the bubble phase. The calculation formula is as follows:

[0204]

[0205] In the formula: P scat P represents the acoustic power scattered by methane, measured in watts. It can be calculated by extracting the acoustic scattering event segments of methane bubbles from the sound pressure signal recorded by a hydrophone. noise Background noise power, measured in watts, obtained without acoustic events; C bubble The concentration of methane in the gaseous phase is expressed in moles per cubic meter. It is related to the total scattering cross section of the bubbles through a proportionality constant k. This proportionality constant needs to be experimentally calibrated, and the calculation formula is as follows:

[0206]

[0207] In the formula: k is the proportionality coefficient, which needs to be calibrated experimentally, and the unit is moles per square meter; r The radius of the bubble is in meters. It can be calculated using the sound scattering cross section of the bubble.

[0208] The optical inversion process is as follows:

[0209] Optical inversion captures the absorption spectral characteristics of water bodies using a hyperspectral analyzer to invert methane concentration, avoiding atmospheric interference.

[0210] Methane characteristic absorption depth model: This model describes the absorption characteristics of methane to light of a specific wavelength. The dissolved methane concentration is retrieved by measuring the ratio of the radiance received by the sensor to the background radiance. The calculation results based on the improved Beer-Lambert formula are as follows:

[0211]

[0212] In the formula: I λ Radiance received by the sensor, measured in watts per square meter per solid angle per nanometer; I 0,λ Background radiance of methane-free water bodies, measured in watts per square meter per solid angle per nanometer; both can be extracted using hyperspectral cubes; α λ,CH4 λ is the absorption coefficient of methane at wavelength λ, expressed in meters per mole, obtained through laboratory calibration; L is the optical path length, expressed in meters, calculated using the formula: L = z / cos θ z represents the water depth in meters. θ β is the incident angle of the light source; λ,H2OThe absorption coefficient of seawater at wavelength λ, expressed in units of per meter, can be obtained from reference materials; γ λ,CDOM : Absorption coefficient of the colored dissolved substance at wavelength λ, in units of per meter, obtained from laboratory calibration. C diss The concentration of methane in the dissolved phase, expressed in moles per cubic meter, can be calculated using the formula described above, as follows:

[0213]

[0214] Acoustic-optical combined model: This model combines the bubble concentration distribution obtained through acoustic inversion with the dissolved state concentration obtained through optical inversion to address the underestimation problem of single phases and construct a complete methane phase model. The formula is as follows:

[0215]

[0216] In the formula: C bubble C represents the acoustically inverted bubble phase concentration, expressed in moles per cubic meter, obtained through acoustic inversion. diss The concentration of dissolved phase obtained through optical inversion is expressed in moles per cubic meter and is derived from optical inversion. k t The phase transition weighting coefficient, controlled by depth D, can be obtained through experimental data calibration. In the high-pressure zone of the deep sea (D greater than 500 meters), bubble dissolution is slow. k t When the pressure approaches zero, the contribution of dissolved state is relatively weak; however, in shallow, low-pressure areas (D less than 100 meters), bubbles dissolve more rapidly. k t The methane phase distribution tends to be 1, dominated by the dissolved state. A pressure chamber simulation experiment was conducted to measure the methane phase distribution at different depths (D).

[0217] In some embodiments, see Figure 4 The current total methane leakage flux was determined based on joint data, including:

[0218] S340. Determine the methane bubble-state population flux and the methane dissolved-state flux based on the combined data;

[0219] S350. Determine the current total methane leakage flux based on the methane bubble-state mass flux and the methane dissolved-state flux.

[0220] Bubble-phase population flux: Reflects the mass of methane in the bubble phase passing through a unit area per unit time, and is the main component of leakage flux. The unit is moles per square meter per second, and the calculation formula is as follows:

[0221]

[0222] In the formula: ρ i Let be the bubble density of the i-th unit, in kilograms per cubic meter. ri The local average radius is expressed in meters. The actual rising speed of the bubbles, measured in meters per second, can be obtained through feature extraction in step three. A seep The area of ​​the leakage unit segmented from the hyperspectral image is expressed in square meters and can be obtained through hyperspectral image segmentation.

[0223] Dissolved flux: Dissolved flux reflects the mass of dissolved methane passing through a unit area per unit time. It is an important supplement to leakage flux, and the unit is moles per square meter per second. The calculation formula is as follows:

[0224]

[0225] In the formula: k dis The gas-liquid mass transfer coefficient, expressed in meters per second, is calculated based on the Wanninkhof equation, as shown in the following formula:

[0226]

[0227] In the formula: The turbulence intensity is calculated from the standard deviation of the ADCP velocity, in m / s; S C The value is the Schmidt number, which can be obtained from a table. 0.31, 0.5, and -0.5 are empirical coefficients obtained through fitting experimental data. C sat The saturated concentration of methane, calculated using Henry's Law in moles per cubic meter, can be obtained from the feature extraction step in step three; C diss The concentration of solution obtained from optical inversion is expressed in moles per cubic meter and can be obtained from the fourth step of optical inversion. A seep The leakage area was obtained through hyperspectral image segmentation.

[0228] Instantaneous leakage flux: Instantaneous leakage flux is the sum of the gaseous population flux and the dissolved flux, characterizing the total flux of cold seep methane leakage. It is the core model for fully quantifying leakage intensity, with units of moles per square meter per second. The calculation formula is as follows:

[0229]

[0230] In some embodiments, see Figure 5 The acoustic-optical fusion method for detecting methane in cold seeps also includes:

[0231] S411. Predict the methane conversion concentration based on the current total methane leakage flux;

[0232] S412. Predict the future total methane concentration based on the methane conversion concentration and the current total methane concentration.

[0233] The three-dimensional advection-diffusion model describes the diffusion process of methane in seawater (specifically for cold seep methane plumes), and the equations are as follows:

[0234]

[0235] That is, the rate of change of the leakage methane concentration C over time is equal to x Directional diffusion y The sum of directional diffusion, vertical methane transport by ocean currents, and methane release at the leakage point. The boundary condition is the methane concentration at the seabed. That is, the rate of change of methane concentration at the seabed is equal to the release intensity, and the concentration at the sea surface is... This means that the methane completely dissipates from the sea surface.

[0236] In the formula: C is the concentration of methane in seawater, in moles per cubic meter; D x D y U is the horizontal diffusion coefficient, measured in square meters per second, representing the ability of methane to diffuse in both horizontal and vertical directions. It is obtained by inversion from acoustic turbulence intensity. z The vertical flow velocity is measured in meters per second (m / s) and obtained by ADCP. η(t) is the leakage intensity function, representing the methane release intensity at the leakage point. It is a function of time t and is measured in moles per cubic meter per second (mol / s). Based on the above calculations, the formula is: η(t) = φ total (t) / A seep ; For the Dirac function, only when z=z s The value is 1 at time and 0 at other positions. It is dimensionless, and z is the depth of the seawater. s The depth of the leak point is shown in meters.

[0237] The methane concentration C is a function of space (x, y, z) and time (t), and its value is determined by the instantaneous leakage flux. and methane phase model concentration Jointly determined. Flux inversion results. This represents the mass of methane passing through a unit area per unit time, and is the direct source of concentration contribution; it needs to be converted into concentration contribution. The formula is:

[0238]

[0239] In the formula: C flux The concentration contribution is calculated from the flux inversion results, in moles per cubic meter; A seep U represents the leakage area; z The vertical flow velocity is [value missing].

[0240] methane phase model concentration Characterizing the total methane concentration in seawater, it is the basis for concentration calculations, and the concentration contribution C from flux inversion results is used. flux Concentration of methane phase model By combining these methods, the total methane concentration C, expressed in moles per cubic meter, was calculated, marking the first time that dynamic coupling between leakage flux and water concentration was achieved. The formula is as follows:

[0241]

[0242] In some embodiments, see Figure 6 The acoustic-optical fusion method for detecting methane in cold seeps also includes:

[0243] S421. Construct a deep learning model and training sample data; the deep learning model includes a Transformer module and an LSTM module, and the training sample data includes diffusion coefficients, environmental measurement data, and methane concentration fields.

[0244] S422. Train a deep learning model based on the training sample data, and predict the methane plume concentration field distribution based on the trained deep learning model.

[0245] T-LSTM-based plume concentration field prediction: Constructing a spatiotemporal fusion deep learning model to predict the dynamic diffusion process of methane plumes in seawater (utilizing the advantages of Transformer to overcome the limitations of traditional LSTM in capturing complex spatiotemporal correlations).

[0246] 1) Data preprocessing

[0247] Input features include: the three-dimensional advection-diffusion equation, and the horizontal diffusion coefficient D over the observation period. x D y Vertical flow velocity U z The leakage intensity function η(t) is included, as well as environmental parameters (temperature T, salinity S, depth D) as auxiliary inputs, and finally the methane concentration field C(x,y,z,t) is output.

[0248] The 3D ocean area is discretized into a grid with a resolution of 20 grids in the x-direction, 20 grids in the y-direction, and 100 grids in the z-direction. All parameters are resampled with a time step of 1 hour Δt. Therefore, the input parameters are aligned and then normalized, resulting in:

[0249]

[0250] To ensure that the data are uniformly distributed and thus improve the model training results, the methane concentration field C(x,y,z,t) is also normalized.

[0251] 2) Feature selection

[0252] The Pearson correlation coefficient measures the linear correlation between two variables. Its value ranges from -1 to 1, where 1 indicates a perfect positive correlation, -1 indicates a perfect negative correlation, and 0 indicates no linear correlation. The formula for the Pearson correlation coefficient is:

[0253]

[0254] In the formula: r c,xk Let C be the Pearson correlation coefficient for the k-th feature. t Let be the methane concentration at time t. X represents the average methane concentration. k,t Let be the value of the k-th feature at time t. It is the average value of the k-th feature.

[0255] However, the limitation of the Pearson correlation coefficient is that it is only sensitive to linear relationships. If the relationship is non-linear, even if there is a one-to-one correspondence between the two variables, the Pearson correlation coefficient may approach 0. Therefore, the distance correlation coefficient is used to overcome the shortcomings of the Pearson correlation coefficient. Thus, by combining these two coefficients, features strongly correlated with the evolution of methane concentration can be screened. The expression for the distance correlation coefficient is:

[0256]

[0257] In the formula: X represents k The distance covariance between and C, X represents k The distance variance, This represents the distance variance of C.

[0258] A composite coefficient based on Pearson correlation coefficient and distance correlation coefficient was selected. Greater than 0.7 and A value greater than 0.5 indicates a high correlation with methane concentration. This reduces computational complexity.

[0259] 3) Model Building

[0260] Long Short-Term Memory (LSTM) networks have garnered significant attention due to their superior performance in time series forecasting, while the Transformer architecture has demonstrated outstanding capabilities in handling complex tasks in recent years. Given the highly temporal nature of the dynamic diffusion process of methane plumes in seawater, a hybrid model combining a Transformer core structure with an LSTM network (T-LSTM) is employed. This model integrates the Transformer's embedding and encoder layers with LSTM layers to extract temporal features. This hybrid approach enhances the extraction of feature correlations, overcoming the limitations of traditional LSTM models in capturing such correlations.

[0261] The T-LSTM model proposed in this embodiment of the invention consists of a Transformer module and an LSTM module, and its structure is as follows: Transformer encoder (embedding layer + MHA + TD-FC) → Transformer decoder (two MHA layers + TD-FC) → LSTM layer → global average pooling → fully connected transition layer → fully connected output layer → Reshape layer. Specifically, as follows:

[0262] a) Transformer encoder

[0263] First, input the dataset. In the encoder, the low-dimensional input feature X'(t) is transformed into a high-dimensional feature E(t) through the embedding layer, as shown in the following formula:

[0264]

[0265] In the formula: X' is the input feature with a dimension of 1×4, W e This is the weight matrix of the embedding layer, with dimensions 4×64, b e is the bias term of the embedding layer, with a dimension of 1×64, and E is the high-dimensional feature representation of the embedding layer output, with a dimension of 1×64.

[0266] Multi-Head Attention (MHA) is used to extract the correlation between features, configured with 8 heads. The formula for calculating MHA is as follows:

[0267]

[0268] In the formula: Q, K, and V are the query, key, and value matrices, respectively, which are obtained by linear transformation of the input feature E(t). dk is the dimension of the key vector, and the softmax function is used to calculate the attention weights.

[0269] The output dimension of the MHA layer was reduced by using a temporally distributed fully connected layer (TD-FC) configured to have 64 neurons.

[0270] b) Transformer decoder

[0271] The feature sequence output from the encoder is input into the decoder. Eight heads are configured in the first MHA layer, and its calculation formula is the same as that of the MHA in the encoder stage. Eight heads are configured in the second MHA layer, matching the output dimension of the previous MHA layer. Its calculation formula is also the same as that of the MHA in the encoder stage. 64 neurons are configured in the TD-FC layer.

[0272] c) LSTM module

[0273] The feature sequence output from the Transformer decoder is fed into an LSTM layer to capture long-term dependencies in the time series. It is configured with 64 neurons to recover the data dimensionality. The computation formula for an LSTM unit is as follows:

[0274] The input gate controls the flow of new information into the cell state:

[0275]

[0276]

[0277] The forget gate selectively discards or filters out past information deemed unimportant from the cell state:

[0278]

[0279] Therefore, the cell state is affected by the input gate and the forget gate, as shown below:

[0280]

[0281] The output gate determines the output state and result of the LSTM cell:

[0282]

[0283]

[0284] In the formula: , , These are the output vectors of the input gate, forget gate, and output gate at time t, respectively. , These represent the current cell state and the candidate cell state, respectively. σ represents the hidden state at the current time step; σ is the Sigmoid activation function, and tanh is the hyperbolic tangent function. , , , These are the input gate, forget gate, and output gate, as well as the weight matrix related to the connection with the previous hidden state; , , , These are the input gate, forget gate, and output gate, as well as the bias term related to the connection with the previous hidden state; This indicates element-wise multiplication.

[0285] Global Average Pooling Layer (1D): Compresses the time dimension, converting the 3D tensor output by the LSTM into a 2D tensor, outputting 64-dimensional features. Fully Connected Transition Layer: Configured with 256 neurons, using the ReLU activation function. Fully Connected Output Layer: Configured with 40,000 neurons, using the Linear activation function, corresponding to a 20×20×100 3D grid. Reshape Layer: Converts the 40,000-dimensional vector into a 20×20×100 3D grid, used to generate the predicted methane concentration field C(x,y,z,t).

[0286] 4) Model Training

[0287] To optimize model performance, the following training strategy was adopted: 70% of the data was used as the training set, 15% as the validation set for parameter tuning, and the last 15% as the test set for final evaluation. Mean squared error (MSE) was used as the loss function, as shown in the following formula:

[0288]

[0289] In the formula: The methane concentration value predicted by the model. The value represents the actual methane concentration, and N represents the number of samples.

[0290] During training, the Adam optimizer is used to update model parameters with a learning rate of 0.001; batch training is used to improve training efficiency with a batch size of 32; an early stopping strategy is adopted to prevent overfitting with a patience value of 5. If the loss on the validation set does not decrease for 5 consecutive epochs, the training is stopped to prevent overfitting and save training time (an existing technique).

[0291] 5) Model performance evaluation

[0292] The root mean square error (RMSE) and relative absolute error (RAE) were used to evaluate the model's predictive performance for methane concentration. RMSE calculates the average difference between predicted and actual values, providing a comprehensive measure of prediction accuracy; while RAE expresses the error as a percentage of the baseline error, providing a relative measure of accuracy. A decrease in both RMSE and RAE indicates improved predictive effectiveness, reflecting an approximate agreement between the predicted and actual results.

[0293] 6) Predicted output

[0294] Once the model is trained, it can be used to predict the methane plume concentration field distribution in future time steps. The model can be built using existing data, and the prediction results can be presented in the form of a heatmap (generating the predicted methane concentration of the leak for the next week in 1-hour increments), visually demonstrating the spatiotemporal distribution of methane concentration in the three-dimensional ocean.

[0295] A 3D heatmap can be generated using Matplotlib or Seaborn libraries, with color depth representing methane concentration levels. The predicted concentration C(x,y,z,t) is mapped to the RGB color space; for example, blue can represent low concentration and red can represent high concentration. A horizontal heatmap is generated every 10 meters of depth, marking the location of the leak point. Vertical cross-sections are generated along the center point of the leak in east-west and north-south directions, also marking the leak point location. Finally, a sequence of images is generated in one-hour increments, which are then combined to create a spatiotemporal distribution map of the predicted leaked methane concentration, and finally, an MP4 animation is created.

[0296] This invention constructs a multimodal in-situ observation system integrating acoustics (hydrophone), optics (hyperspectrology), and environment (temperature, salinity, and depth sensor). The hydrophone captures the acoustic radiation spectrum of gaseous methane, and its strong penetrating power allows for the identification of the size distribution of methane bubbles, thereby identifying the location and intensity of leaks and compensating for the limitations of optical signals in harsh environments such as turbidity, low light, and great depth. The hyperspectral camera accurately detects the absorption characteristics of dissolved methane, confirming its presence and concentration. The temperature, salinity, and depth (CTD) sensor measures temperature / salinity / pressure gradients to obtain seawater density profiles, while the acoustic Doppler current profiler (ADCP) outputs a three-dimensional velocity field to calculate the diffusion coefficient, providing relevant parameters for the physical constraint model. As bubbles rise, they continuously dissolve, forming a dynamic transformation from "bubble state to dissolved state." A single sensor can only capture a portion of the phases, leading to an underestimation of the current total methane leakage flux. Finally, this invention combines a physical-driven model to realize a full-chain intelligent detection method from data acquisition, fusion, modeling to flux inversion and plume prediction, overcoming the challenge of real-time monitoring of intermittent leakage.

[0297] See Figure 7 This invention provides an acoustic-optical fusion cold seep methane detection system, comprising:

[0298] The acquisition module is used to acquire the first pre-processed acoustic measurement data and the second pre-processed optical measurement data, and to acquire environmental measurement data.

[0299] The data processing module is used to generate combined data based on acoustic measurement data, optical measurement data, and environmental measurement data;

[0300] The methane calculation module is used to extract joint features from the joint data, determine the current total methane concentration based on the joint features, and determine the current total methane leakage flux based on the joint data.

[0301] It is evident that the content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0302] like Figure 8 As shown, this embodiment of the invention also provides an acoustic-optical fusion cold seep methane detection device, comprising:

[0303] At least one processor;

[0304] At least one memory for storing at least one program;

[0305] When the at least one program is executed by the at least one processor, the at least one processor implements the steps of the acousto-optic fusion cold seep methane detection method described in the above method embodiments.

[0306] The memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. The memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory may optionally include remote memory located remotely relative to the processor, which can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0307] It is evident that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented in this device embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0308] Furthermore, this application also discloses a computer program product or computer program stored in a computer-readable storage medium. A processor of a computer device can read the computer program from the computer-readable storage medium, and the processor executes the computer program, causing the computer device to perform the described method. Similarly, the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0309] This invention also provides a computer-readable storage medium storing a processor-executable program that, when executed by a processor, implements the above-described method.

[0310] It is understood that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0311] See Figure 9 This invention provides an acoustic-optical fusion cold seep methane detection system, including a computer device and acoustic detectors, optical detectors, and environmental detectors connected to the computer device; wherein,

[0312] Acoustic detectors are used to collect acoustic measurement data;

[0313] Optical detectors are used to collect optical measurement data;

[0314] Environmental detectors are used to collect environmental measurement data;

[0315] Computer equipment includes:

[0316] At least one processor;

[0317] At least one memory for storing at least one program;

[0318] When at least one program is executed by at least one processor, the at least one processor performs the above-described method.

[0319] Specifically, acoustic detection includes, but is not limited to, hydrophones; optical detectors include, but are not limited to, hyperspectral detectors; and environmental detectors include, but are not limited to, temperature, salinity, and depth sensors. As for the computer equipment, it can be different types of electronic devices, including, but not limited to, desktop computers, laptops, wearable devices, and other terminals.

[0320] It is evident that the content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0321] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0322] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0323] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0324] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A method for detecting methane in cold seeps using a combination of acoustic and optical methods, characterized in that, include: Acquire acoustic measurement data after the first preprocessing and optical measurement data after the second preprocessing, and acquire environmental measurement data; Based on the acoustic measurement data, the optical measurement data, and the environmental measurement data, joint data is formed; Based on the combined data, extract combined features, determine the current total methane concentration based on the combined features, and determine the current total methane leakage flux based on the combined data; The step of forming joint data based on the acoustic measurement data, the optical measurement data, and the environmental measurement data includes: The acoustic measurement data is used to determine the sound source location of the methane bubbles, and the optical measurement data is mapped to a preset coordinate system. Based on the location of the sound source, obtain optical measurement data of the region of interest in the preset coordinate system; Spatiotemporal matching is performed on the optical measurement data, acoustic measurement data, and environmental measurement data of the region of interest to establish a multimodal dataset and form joint data; The step of extracting joint features from the joint data and determining the current total methane concentration based on the joint features includes: Joint features are extracted from the joint data, and model source data is constructed based on the joint features; the model source data includes acoustic source data and optical source data. The concentration of methane in the bubble phase is calculated based on the acoustic source data, and the concentration of methane in the dissolved phase is calculated based on the optical source data. The current total methane concentration is determined based on the methane concentration in the bubble phase and the methane concentration in the dissolved phase.

2. The method according to claim 1, characterized in that, Determining the current total methane leakage flux based on the combined data includes: The combined data were used to determine the methane bubble-state population flux and the methane dissolved-state flux. The current total methane leakage flux is determined based on the methane bubble-state population flux and the methane dissolved-state flux.

3. The method according to any one of claims 1-2, characterized in that, The method further includes: Predict the methane conversion concentration based on the current total methane leakage flux; The future total methane concentration is predicted based on the methane conversion concentration and the current total methane concentration.

4. The method according to any one of claims 1-2, characterized in that, The method further includes: Construct a deep learning model and training sample data; the deep learning model includes a Transformer module and an LSTM module, and the training sample data includes diffusion coefficients, environmental measurement data, and methane concentration fields; The deep learning model is trained based on the training sample data, and the methane plume concentration field distribution is predicted based on the trained deep learning model.

5. A sound-optical fusion cold seep methane detection system, characterized in that, include: The acquisition module is used to acquire the first pre-processed acoustic measurement data and the second pre-processed optical measurement data, and to acquire environmental measurement data. The data processing module is used to generate joint data based on the acoustic measurement data, the optical measurement data, and the environmental measurement data; The methane calculation module is used to extract joint features based on the joint data, determine the current total methane concentration based on the joint features, and determine the current total methane leakage flux based on the joint data. The step of forming joint data based on the acoustic measurement data, the optical measurement data, and the environmental measurement data includes: The acoustic measurement data is used to determine the sound source location of the methane bubbles, and the optical measurement data is mapped to a preset coordinate system. Based on the location of the sound source, obtain optical measurement data of the region of interest in the preset coordinate system; Spatiotemporal matching is performed on the optical measurement data, acoustic measurement data, and environmental measurement data of the region of interest to establish a multimodal dataset and form joint data; The step of extracting joint features from the joint data and determining the current total methane concentration based on the joint features includes: Joint features are extracted from the joint data, and model source data is constructed based on the joint features; the model source data includes acoustic source data and optical source data. The concentration of methane in the bubble phase is calculated based on the acoustic source data, and the concentration of methane in the dissolved phase is calculated based on the optical source data. The current total methane concentration is determined based on the methane concentration in the bubble phase and the methane concentration in the dissolved phase.

6. A sound-optical fusion cold seep methane detection device, characterized in that, include: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor performs the method as described in any one of claims 1-4.

7. A computer-readable storage medium storing a processor-executable program, characterized in that, The processor-executable program, when executed by the processor, is used to perform the method as described in any one of claims 1-4.

8. A sound-optical fusion cold seep methane detection system, characterized in that, This includes computer equipment and acoustic detectors, optical detectors, and environmental detectors connected to the computer equipment; wherein, The acoustic detector is used to collect acoustic measurement data; The optical detector is used to collect optical measurement data; The environmental detector is used to collect environmental measurement data; The computer device includes: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor performs the method as described in any one of claims 1-4.

Images