Method for retrieving cold spring plume leakage flux based on multi-beam water body image

Abstract

The application discloses a cold spring plume leakage flux inversion method based on multi-beam water body images, and comprises the following steps: multi-beam water body data analysis and point cloud homing; the multi-beam water body backscattering intensity is corrected, and the average value of the backscattering intensity is calculated; according to the calculated average value of the backscattering intensity, the bubble size distribution parameters are inversed for multiple times, and statistical analysis is conducted on the multiple inversion results to determine the estimated values of the parameters, the parameters comprising the total number of bubbles in a volume element, the average bubble radius and the bubble radius standard deviation; finally, the plume flux is calculated according to the estimated values of the parameters. The scheme can realize stable and reliable inversion of the cold spring plume flux by using single-frequency multi-beam water body data, and avoids the ill-posed problem existing in the inversion process; the stable estimation of the plume flux is realized by using limited data, the observation cost can be significantly saved, and technical support is provided for natural gas hydrate exploration and evaluation.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of ocean exploration, and particularly relates to a deep-sea cold spring plume flux inversion method based on a multi-beam water body image. BACKGROUND

[0002] Cold spring activity has been an international research hotspot. It is not only an important part of modern seabed extreme environment system, but also an important way and even a central link for the transfer and exchange of matter from the lithosphere to the external sphere. It is one of the current deep-sea exploration fields and related scientific research frontiers. Cold spring activity is accompanied by fluid upward migration to form plumes, and the plume is usually the main object of cold spring exploration. Cold spring plume is mainly composed of a large number of bubbles of different sizes and different speeds, which overflow from the seabed under the action of the environment and continuously rise until they break and dissolve. Under the action of the ocean current field, the plume bubbles will show certain morphological characteristics. Due to the difference in medium inside and outside the bubbles, there is a strong acoustic impedance difference. When the bubbles are irradiated by sound waves, strong scattering will occur. Therefore, the plume can be detected by water acoustic image.

[0003] At present, the traditional method for detecting the flux of cold spring plume by using shipborne acoustic means is mostly based on single-beam sonar system and split-beam sonar system. With the development of sonar technology, the detection accuracy and imaging resolution of acoustic systems represented by multi-beam depth measurement systems have gradually improved. Multi-beam depth measurement systems are widely used in water depth measurement and water body target detection fields due to their wide range, high efficiency and full-sea-depth advantages. Multi-beam sonar system is one of the most effective technical means for detecting deep-sea cold spring plume. However, there is no systematic research on the cold spring plume flux inversion method based on multi-beam depth measurement system in the world. There are the following problems in the flux inversion of plume using acoustic data: (1) The rising speed and size distribution parameters of the bubble group in the plume are obtained by combining in-situ observation optical data, but the in-situ observation data is limited and cannot be widely used, and in most data collection processes, there is a lack of in-situ optical observation data support; (2) Multi-frequency acoustic data based on split-beam can easily invert the bubble size distribution and estimate the flux, but multi-beam acoustic data is basically single-frequency data, which makes the inversion process have a serious ill-posed problem, and the result is unstable and the solution is not unique. Therefore, it is urgent to develop a cold spring plume flux inversion method based on multi-beam depth measurement system to realize the quantitative evaluation of plume gas flux and provide technical support for natural gas hydrate exploration and evaluation. SUMMARY

[0004] The present application aims to provide a plume flux inversion method based on multi-beam water body images, to solve the problem of flux inversion in the absence of in-situ observation of bubble size distribution, and to expand the range of quantitative application of multi-beam water bodies.

[0005] The present application is implemented by adopting the following technical scheme: a cold spring plume leakage flux inversion method based on multi-beam water body images, comprising the following steps:

[0006] Step A, multi-beam water body data analysis and point cloud homing: analyze the original multi-beam water body data to obtain the backscattering intensity, ship position and attitude information of the multi-beam water body sampling points, and perform point position homing on the sampling points in the water body data to obtain the three-dimensional coordinates of each sampling point;

[0007] Step B, correct the multi-beam water body backscattering intensity, and calculate the average backscattering intensity TS obser ;

[0008] (1) When correcting the multi-beam water body backscattering intensity, the backscattering intensity of each Ping water body data is corrected, and the plume in each Ping water body data is extracted, the soft threshold denoising method is used to preliminarily suppress the noise of each Ping multi-beam water body data, and on this basis, the remaining noise points are manually removed to realize fine segmentation and extraction of plume point cloud;

[0009] (2) Calculate the average backscattering intensity TS obser in the following way:

[0010] After extracting the plume in each Ping water body data, the three-dimensional point cloud data of the plume is obtained, from the bottom of the plume, the point cloud data in a certain range is selected as the data for flux estimation, and the backscattering intensity of the point cloud in the range is averaged to obtain the average backscattering intensity of the plume at that depth:

[0011]

[0012] In the formula, TS obser is the average value of backscattering intensity, n represents the number of plume sampling points, and TS i represents the backscattering intensity value of the i-th sampling point;

[0013] Step C, according to the calculation result of step B, the bubble size distribution parameters are inversed multiple times, and the multiple inversion results are statistically analyzed to determine the estimated value of each parameter and estimate the bubble rising speed, the parameters include the total number of bubbles in the volume element, the average bubble radius and the bubble radius standard deviation;

[0014] (1) The size distribution N of bubble group refers to the distribution of the number of bubbles of different radii in a certain volume element. Assuming that the bubble size distribution N obeys a lognormal distribution, we have:

[0015]

[0016] wherein:

[0017]

[0018]

[0019] In the formula, N(r) represents a bubble size distribution function related to bubble radius r, N b represents the total number of bubbles in a volume element, r b represents the average radius of bubbles, and σ b represents the standard deviation of bubble radius.

[0020] (2) The above three parameters N b , r b , and σ b are initialized according to the genetic algorithm GA.

[0021] First, set the three parameters N b , r b , and σ b to have solution ranges of 0-10000, 0.1-10, and 0.1-10, respectively. Then, randomly initialize the population size within the parameter range. Each individual has three chromosomes, i.e., the three parameters Nb, rb, and σb to be solved. The random initialization method is as follows:

[0022]

[0023] In the formula, min() and max() represent the upper and lower limits of the solution range interval of the parameter, and rand() is used to generate a random number between 0 and 1.

[0024] (3) Estimate the backscattering intensity TS model of bubbles under the size distribution parameters according to the acoustic scattering model of a single bubble and the initialized bubble distribution parameters.

[0025] TS model = 10 log σ v

[0026]

[0027] wherein r represents the bubble radius, σ v represents the total scattering cross section of the bubble, f represents the working frequency of the sonar, f0 represents the resonance frequency of the bubble, and k represents the wave number.

[0028] (4) The average backscatter intensity TS of the multibeam water data calculated in step B obser As a constraint, the fitness is calculated according to the following formula, which measures the difference between the estimated and measured values ​​of the bubble backscattering intensity:

[0029] f = TS model -TS obser

[0030] For different individuals, the calculated TS model The values ​​of f and f are different; the closer f is to 0, the stronger the TS. model With TS obser The closer they are, that is, the closer the parameters Nb, rb, and σb are to their true values;

[0031] (5) Perform selection operations according to the size of f, and perform chromosome crossover and chromosome mutation operations. Repeat (4)-(5) until the set number of iterations is completed to obtain the first inversion result. Take the individual with the lowest fitness in the population as the optimal individual for this iteration and obtain the parameter N. b r b σ b A set of parameter solutions;

[0032] (6) Repeat (2)-(5) several times to obtain multiple inversion results. Then, perform statistical analysis on the multiple inversion results to determine the estimated values ​​of each parameter. When estimating the parameters of the multiple inversion results:

[0033] 1) For N in the results of multiple inversions b The parameters are used to fit a straight line using a random sampling consensus algorithm: a point is randomly selected, and its N... b We obtain a straight line equation, then calculate the Euclidean distance from each of the remaining points to this line. We set a distance threshold and count the number of points whose distance is less than the threshold; these are called interior points. After several iterations, we take the line equation value that results in the largest number of interior points as N. b The estimated value;

[0034] 2) Using cubic spline interpolation, N b The distribution range is divided into n intervals, and the point pairs (N) within each interval are used. b r b Fit a cubic spline function, and then use the N determined in the previous step. b Substitute the value into the fitted cubic spline function to obtain r. b The estimated value;

[0035] 3) Obtaining parameter N b and r b Then, fix the parameter N.b with r b , the range of the setting parameter sigma b is 0.1-10, according to the backscattering intensity TS obser , the estimated value of sigma b is obtained by inversion principle.

[0036] Step D, calculating the plume flux according to the estimated value of each parameter in step C:

[0037] The estimated parameters r b , sigma b are used to calculate the bubble distribution f(r) according to the bubble size distribution function N(r), the bubble size distribution parameters obtained by inversion are brought into the following flux estimation formula to complete the flux estimation:

[0038]

[0039] In the formula, r represents the radius of the bubble; f(r) represents the obtained bubble size distribution; D represents the sound intensity sampling unit, c represents the sound velocity, and tau represents the sound pulse length.

[0040] Compared with the prior art, the advantages and positive effects of the present application are that:

[0041] The inversion method proposed in the present application realizes stable and reliable inversion of the flux of cold spring plume by using single-frequency multi-beam water body data without external data support, avoids the ill-posed problem existing in the inversion process, realizes stable estimation of the plume flux by using limited data, can significantly save observation cost, can be applied to the investigation task of natural gas hydrate resources, and provides technical support for natural gas hydrate exploration and evaluation. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 It is a flowchart of the gas flux inversion method of the embodiment of the present application;

[0043] Figure 2 It is a schematic diagram of the flux inversion principle of the embodiment of the present application;

[0044] Figure 3 It is the backscattering intensity correction and soft threshold denoising result of the embodiment of the present application;

[0045] Figure 4 It is the three-dimensional point cloud of the plume after beam homing of the embodiment of the present application;

[0046] Figure 5 It is a schematic diagram of the RANSAC algorithm estimating parameter N b of the embodiment of the present application;

[0047] Figure 6A schematic diagram of estimating the parameter rb by cubic spline interpolation for the third embodiment of the present application;

[0048] Figure 7 Flux inversion results of the third embodiment of the present application. DETAILED DESCRIPTION

[0049] In order to more clearly understand the above-mentioned purposes, features and advantages of the present application, the present application will be further described below in conjunction with the accompanying drawings and embodiments. In the following description, a large number of specific details are set forth in order to facilitate a thorough understanding of the present application, however, the present application can also be implemented in other ways different from those described herein, and therefore, the present application is not limited to the specific embodiments disclosed below.

[0050] In order to realize the flux inversion of the plume, the present embodiment proposes a plume flux inversion method based on multi-beam water body image, which comprises the following steps:

[0051] Step A, multi-beam water body data analysis and point cloud homing;

[0052] Step B, multi-beam water body backscattering intensity correction, and calculation of backscattering intensity average value;

[0053] Step C, multiple inversions are performed on the bubble size distribution parameters, and statistical analysis is performed on the multiple inversion results to determine the estimated values of the parameters;

[0054] Step D, calculating the plume flux according to the estimated values of the parameters.

[0055] As shown in Figure 1 and Figure 2 , specifically:

[0056] In step A, the original multi-beam water body data is analyzed to obtain the backscattering intensity, ship position and attitude and other information of the multi-beam water body sampling points, and the point position of the water body data is calculated to obtain the three-dimensional coordinates of each sampling point.

[0057] Step B specifically comprises the following steps:

[0058] Step B1, correcting the backscattering intensity of each Ping water body data, and extracting the plume target in each Ping water body data;

[0059] The method for correcting the backscattering intensity is as follows:

[0060] TS r =A+(X-40)lgR-C (1)

[0061] In the formula, TS rFor the corrected relative backscattering intensity, A is the amplitude value of the sampling point recorded in the multi-beam water body data collection, X represents the propagation loss factor, R represents the sound wave propagation distance; C represents the time variable gain (TVG) offset, and the results before and after correction are as shown in Figure 3

[0062] In addition, when extracting the plume, a soft threshold denoising method is used to preliminarily suppress the noise of each Ping multi-beam water body data. On this basis, by manually removing the remaining noise points, fine segmentation and extraction of the plume point cloud are realized, as shown in Figure 4

[0063] Step B2, after extracting the plume from each Ping water body data, three-dimensional point cloud data of the plume is obtained. From the bottom of the plume, the point cloud data within a range of 20 meters in height is taken as the data for flux estimation. The backscattering intensity of the point cloud in this range is averaged to obtain the average backscattering intensity of the plume at this depth.

[0064]

[0065] In the formula, TS obser is the average value of the backscattering intensity, n represents the number of plume sampling points, and TS i represents the backscattering intensity value of the i-th sampling point.

[0066] In step C, when the bubble size distribution parameters are inversed and the estimated values of each parameter are determined, the following steps are included:

[0067] Step C1, multiple inversions of bubble size distribution parameters: the backscattering intensity TS obser is used to inverse the plume bubble group size distribution, which specifically includes:

[0068] 1) The size distribution N of the bubble group refers to the distribution of the number of bubbles of different radii in a certain volume element, which is a function of the bubble radius. It is assumed that the bubble size distribution N obeys a lognormal distribution.

[0069]

[0070] Wherein:

[0071]

[0072]

[0073] In the formula, N(r) represents the bubble size distribution function, which is related to the bubble radius r, N b represents the total number of bubbles in the volume element, and r b ​​represents the average radius of the bubbles, σ b represents the standard deviation of the bubble radius.

[0074] For the plume of flux to be inverted, when the parameters N b , r b , σ b are unknown. According to equations (1)-(3), when the three parameters are determined, the bubble size distribution function N(r) is determined.

[0075] 2) The above three parameters are initialized according to the genetic algorithm (Genetic Algorithm, GA).

[0076] First, the solution ranges of the three parameters are set to 0-10000; 0.1-10; 0.1-10, respectively. Then, the population size is randomly initialized to 600 within the parameter range, i.e., the population contains 600 individuals, each individual has three chromosomes, i.e., the three parameters Nb, rb, σb to be solved, and the random initialization method is:

[0077]

[0078] In the formula, min() and max() represent the upper and lower limits of the solution range interval of the parameters, and rand() is used to generate a random number between 0 and 1;

[0079] 3) According to the acoustic scattering model of a single bubble and the bubble distribution parameters obtained by initialization, the backscattering intensity TS model of the bubble under the size distribution parameters can be estimated according to the following formula.

[0080]

[0081] TS model = 10 log σ v (8)

[0082] Where r represents the bubble radius, σv represents the total scattering cross section of the bubble, f represents the working frequency of the sonar, f0 represents the resonance frequency of the bubble, and k represents the wave number.

[0083] 4) Take the backscattering intensity TS obser of the multi-beam water body data as a constraint, and calculate the fitness according to the following formula to measure the difference between the estimated value and the measured value of the bubble backscattering intensity.

[0084] f = |TS model -TS obser | (9)

[0085] In theory, for different individuals, the calculated TS modelDifferent, so f is also different. The closer f is to 0, the closer TSmodel is to TSobser, that is, the closer the parameters Nb, rb, σb are to the true values.

[0086] 5) Chromosome selection operation and crossover operation:

[0087] Selection of individuals: There are differences in the fitness of 600 individuals, and individuals with smaller fitness need to be selected for preservation. Therefore, the individuals are sorted according to the size of f, the selection rate is set to 0.1, and the first 60 individuals with smaller f are selected as the better individuals of this generation, and the remaining 540 individuals are used for crossover operation to generate new individuals.

[0088] From the remaining 540 individuals, two individuals p1 and p2 are randomly selected, and the crossover point is randomly determined according to the following formula:

[0089] p = ceil(rand*3) (10)

[0090] In the formula, ceil() is the ceiling function, rand is used to generate a random number between 0 and 1, and p is the calculated crossover point position, which can take values of 0, 1, 2, and 3.

[0091] According to the crossover point position p, the chromosomes (parameters) of the two individuals are divided into two parts and exchanged. Specifically as follows:

[0092]

[0093] After the chromosome crossover operation is completed, new individuals p 1new and p 2new are generated. This step is operated 270 times, and a total of 540 new individuals are generated. Combine with the above unchanged 60 individuals to form a population of 600 individuals.

[0094] 6) Chromosome mutation operation:

[0095] Set the mutation rate to 0.1, calculate the total number of mutated chromosomes from the newly generated 600 individuals: 600*0.1*3 = 180, and calculate the position of the mutated chromosome according to the following formula:

[0096] pm = ceil(600*3*rand(1,180)) (12)

[0097] In the formula, ceil() is the ceiling function. rand(1,180) is used to generate 180 random numbers between 0 and 1. After determining the mutation position, the mutation value is recalculated according to the initialization method of formula (6).

[0098] 7) Repeat steps 4) to 6) above until the set number of iterations is completed, and obtain the first inversion result. Select the individual with the lowest fitness in the population as the optimal individual (optimal solution) for this iteration.

[0099] Step C2: Estimation of bubble size distribution parameters

[0100] After the above iterations, a set of parameter solutions can be obtained, namely: N b r b σ b However, since this process is ill-posed, the inversion process is unstable and unreliable, and the solution is not unique, failing to meet the requirements of flux estimation. To address this issue, a parameter estimation method based on statistical characteristics and the correlation between parameters is proposed. First, the bubble size distribution is inverted multiple times. Then, parameter estimation is performed based on the statistical characteristics of the inversion results, avoiding the ill-posedness problem of a single inversion and the instability of the solution.

[0101] First, repeat step C1 100 times to obtain multiple inversion results. Then, perform statistical analysis on the multiple inversion results to determine the estimated values ​​of each parameter. The main process is as follows:

[0102] 1) For N in the results of multiple inversions b The parameters show good concentrated distribution characteristics, such as Figure 5 As shown. The Random Sample Consensus (RANSAC) algorithm is used for line fitting. The main steps are as follows:

[0103] Randomly select a point, and its N b We obtain the equation of a straight line. Then, we calculate the Euclidean distance from each of the remaining points to this line. We set a distance threshold of 50 and count the number of points whose distance is less than the threshold; these are called interior points. After 10,000 iterations, we take the equation of the line that results in the largest number of interior points as N. b The estimated value.

[0104] 2) According to the results of multiple inversions, N b With r b There is a significant monotonic relationship, based on N determined in step 1) above. b The value is obtained by cubic spline interpolation, and the result is as follows: Figure 6 The main steps include:

[0105] Divide the distribution range of Nb into n intervals, and fit a cubic spline function using the point pair (Nb, rb) within each interval. Then, based on the Nb value determined in the previous step, substitute it into the fitted cubic spline function to estimate the rb value.

[0106] 3) Obtaining parameter Nb and r b After that, the fixed parameter N b and r b , the parameter σ b is set to be in the range of 0.1-10, and the estimated value of σ b is obtained by inversion according to step C1.

[0107] In this embodiment, according to the distribution characteristics of the inversion result of Nb, the RANSAC algorithm is used to stably estimate Nb in the parameter inversion result, and according to the correlation between Nb and rb, the cubic spline interpolation is used to estimate rb in the inversion result, thereby avoiding the problem of unstable estimation of rb.

[0108] Step D, plume flux calculation

[0109] The estimated parameters r b , σ b are used to calculate the bubble distribution f(r) according to formulas (3), (4) and (5). The bubble size distribution parameters obtained by inversion are brought into the following flux estimation formula to complete the flux estimation, and the results are shown in Table 1: Figure 7

[0110]

[0111] In the formula, r represents the radius of the bubble; f(r) represents the obtained bubble size distribution; D represents the sound intensity sampling unit, which is calculated according to the following formula:

[0112]

[0113] In the formula, c represents the sound velocity, and τ represents the sound pulse length.

[0114] U(r) represents the bubble rising velocity model, which is calculated according to the following model:

[0115]

[0116] In the formula, r represents the bubble radius, which is obtained from the inversion bubble size distribution, ηs represents the shear viscosity, g represents the gravitational acceleration, ρ represents the seawater density, and τ represents the surface tension.

[0117] The above is only a preferred embodiment of the present application, and is not intended to limit the present application in other forms. Any person skilled in the art can modify or change the above disclosed technical content to equivalent embodiments applied to other fields, but any simple modification, equivalent change and modification made according to the technical essence of the present application to the above embodiments still falls within the protection scope of the present application.​

Claims

1. A method for inverting cold seep plume seepage flux based on multibeam bathymetry images, characterized in that, Includes the following steps: Step A: Multibeam bathymetry data analysis and point cloud localization: Step B: Correct the backscattering intensity of the multibeam bathywater and calculate the average backscattering intensity TS. obser ; Step C: Based on the calculation results of step B, perform multiple inversions on the bubble size distribution parameters, and perform statistical analysis on the multiple inversion results to determine the estimated values ​​of each parameter and estimate the bubble rising speed. The parameters include the total number of bubbles in the volume element, the average bubble radius, and the standard deviation of the bubble radius. Step D: Calculate the feather flux based on the estimated values ​​of each parameter from Step C: In step D, the estimated parameter r b σ b Calculate the bubble distribution f(r) according to N(r) representing the bubble size distribution function. Substitute the inverted bubble size distribution parameters into the following flux estimation formula to complete the flux estimation: In the formula, r represents the radius of the bubble; f(r) represents the obtained bubble size distribution; D represents the sound intensity sampling unit. c represents the speed of sound, and τ represents the length of the sound pulse. U(r) represents the bubble's upward velocity model, calculated according to the following model: In the formula, r represents the bubble radius, which is obtained from the inverted bubble size distribution, and η s ρ represents shear viscosity; g represents gravitational acceleration; w This indicates the density of seawater.

2. The method for inverting cold seep plume seepage flux based on multibeam bathymetry images according to claim 1, characterized in that: In step A, the original multibeam water data is analyzed to obtain the backscatter intensity, ship position and attitude information of the multibeam water sampling points, and the sampling points in the water data are recalculated to obtain the three-dimensional coordinates of each sampling point.

3. The method for inverting cold seep plume seepage flux based on multibeam bathymetry as described in claim 2, characterized in that: In step B, the backscatter intensity is corrected for the water data of each Ping, and the plume in the water data of each Ping is extracted as a target. The method for correcting the backscatter intensity is as follows: In the formula, TS r The corrected relative backscattering intensity is represented by A, which is the amplitude value of the sampling point recorded in the multibeam water data acquisition. X represents the propagation loss factor, R represents the sound wave propagation distance, and C represents the time-varying gain offset.

4. The method for inverting cold seep plume seepage flux based on multibeam bathymetry as described in claim 3, characterized in that: In step B, when extracting the feather flow, a soft threshold denoising method is used to initially suppress noise in the multibeam water data of each Ping. Based on this, the remaining noise points are manually removed to achieve fine segmentation and extraction of the feather flow point cloud.

5. The method for inverting cold seep plume seepage flux based on multibeam bathymetry as described in claim 1, characterized in that: In step B, the average backscattering intensity TS is calculated. obser The following methods are used: After extracting the plume from each Ping water body data, three-dimensional point cloud data of the plume is obtained. Starting from the bottom of the plume, a certain range of point cloud data is selected upwards as flux estimation data. The backscattering intensity of the point cloud within this range is averaged to obtain the average backscattering intensity of the plume at that depth. In the formula, TS obser The average backscattering intensity is given by TS, where n represents the number of sampling points in the plume. i This represents the backscattering intensity value at the i-th sampling point.

6. The method for inverting cold seep plume seepage flux based on multibeam bathymetry as described in claim 1, characterized in that: In step C, the size distribution of the plume-like bubble swarm is inverted using the backscattering intensity: (1) The size distribution N of the bubble group refers to the distribution of the number of bubbles with different radii within a certain volume element. Assuming that the bubble size distribution N follows a log-normal distribution, then: in: In the formula, N(r) represents the bubble size distribution function, which is related to the bubble radius r. b The total number of bubbles in the volume element, r b Indicates the average radius of the bubble, σ b Indicates the standard deviation of bubble radius; (2) Apply the genetic algorithm (GA) to the above three parameters N. b r b σ b Perform initialization; (3) Based on the acoustic scattering model of a single bubble and the bubble distribution parameters obtained from initialization, estimate the backscattering intensity TS of the bubble under this size distribution parameter. model : Where r represents the bubble radius, σ v The total scattering cross section of the bubble is represented by f, the operating frequency of the sonar is represented by f0, the resonant frequency of the bubble is represented by k, and the wave number is represented by k. (4) The average backscatter intensity TS of the multibeam water data calculated in step B. obser As a constraint, the fitness is calculated according to the following formula, which measures the difference between the estimated and measured values ​​of the bubble backscattering intensity: For different individuals, the calculated TS model The values ​​of f and f are different; the closer f is to 0, the stronger the TS. model With TS obser The closer they are, that is, the closer the parameters Nb, rb, and σb are to their true values; (5) Perform selection operations according to the size of f, and perform chromosome crossover and chromosome mutation operations. Repeat (4)-(5) until the set number of iterations is completed to obtain the first inversion result. Take the individual with the lowest fitness in the population as the optimal individual for this iteration and obtain the parameter N. b r b σ b A set of parameter solutions; (6) Repeat (2)-(5) several times to obtain multiple inversion results, and then perform statistical analysis on the multiple inversion results to determine the estimated values ​​of each parameter.

7. The method for inverting cold seep plume seepage flux based on multibeam bathymetry as described in claim 6, characterized in that: In step C, when estimating the parameters of the multiple inversion results: 1) For N in the multiple inversion results b The parameters are used to fit a straight line using a random sampling consensus algorithm: a point is randomly selected, and its N... b We obtain a straight line equation, then calculate the Euclidean distance from each of the remaining points to this line. We set a distance threshold and count the number of points whose distance is less than the threshold; these are called interior points. After several iterations, we take the line equation value that results in the largest number of interior points as N. b The estimated value; 2) Using cubic spline interpolation, N b The distribution range is divided into n intervals, and the point pairs (N) within each interval are used. b r b Fit a cubic spline function, and then use the N determined in the previous step. b Substitute the value into the fitted cubic spline function to obtain r. b The estimated value; 3) Obtaining parameter N b and r b Then, fix the parameter N. b With r b Set parameter σ b The range is 0.1~10, according to the backscattering intensity TS obser Inversion principle to obtain σ b The estimated value.

8. The method for inverting cold seep plume seepage flux based on multibeam bathymetry as described in claim 6, characterized in that: In step C, the principle of initialization using the genetic algorithm (GA) is as follows: First, set three parameters N. b r b σ b The corresponding solution ranges are 0~10000; 0.1~10; and 0.1~10, respectively. Then, the population size is randomly initialized within each parameter range. Each individual has three chromosomes, which are the three parameters to be solved: Nb, rb, and σb. The random initialization method is as follows: (6) In the formula, min() and max() represent the upper and lower limits of the solution range of the parameters, respectively, and rand() is used to generate random numbers between 0 and 1.

Images