Artificial intelligence submarine topography inversion system and method using gravitational field spatial spectrum data

Abstract

The invention provides an artificial intelligence submarine topography inversion system and method using gravitational field spatial spectrum data, and the system comprises a gravitational field spatial spectrum calculation module which uses a plurality of band-pass filters to intercept gravity data under different wavelengths; and the gravity-water depth mapping relation establishing module is used for processing gravity and water depth data and establishing a correlation between the gravity and water depth data. A gravitational field spatial spectrum of a sparse ship water depth measuring position is obtained and is paired with water depth data. Dividing the data set into a training set and a test set in proportion; and designing an artificial intelligence model suitable for a regression task for training and predicting the water depth. And the artificial intelligence submarine topography inversion module is used for predicting a water depth value by using the gravitational field spatial spectrum of the trained model input test set, and evaluating a prediction effect through a precision index. The method has the advantages that the characteristics of the gravity data are enriched through the calculation of the gravity field spatial spectrum, and different degrees of influences of gravity on terrain inversion are considered.

Description

Technical Field

[0001] The present application belongs to the field of ocean remote sensing mapping, and specifically relates to an artificial intelligence inversion system and method for seafloor topography using gravity field spatial spectrum data. Background Art

[0002] Seabed topography mapping can help humans better understand and develop marine resources. In scientific research, seabed topography is closely related to important studies such as ocean circulation, seabed geology, and tsunamis; in the field of economic development, seabed topography mapping provides key support and guarantee for the construction of marine facilities and navigation safety.

[0003] Satellite altimetry technology can quickly and cheaply obtain high-precision gravity field data for global oceans. Gravity anomalies and seafloor topography have a strong correlation within a certain band. This method can quickly and cheaply obtain seafloor topography in global regions. At present, the main methods for inverting seafloor topography include gravity geological method and SAS method. Dixon et al. (1983) proposed the method of admittance function, considering the flexural equilibrium theory on the basis of Parker's formula. Smith and Sandwell (1994) proposed the S&S method to construct a functional relationship between gravity and seafloor topography in the relevant bands, and obtain the proportional factor between gravity data and seafloor topography through the water depth and gravity of known sparse points. Tozer et al. (2019) updated the overall process of the S&S method and established a 15-second global seafloor topography model STRM15+V2.0. Abukatijiang (2019) adjusted the filtering and slope estimation of the S&S method for estimation in the Arctic region. Fan Diao (2020) used the Oldenburg perturbation method to invert the fourth-order seafloor topography based on the Parker formula. In order to simplify the calculation process and improve the practical application value of the inversion process, the physical parameters considered by these traditional methods are relatively simple, and only the most important variables affecting the relationship between gravity and water depth are considered, ignoring the influence of other details, which makes the inversion accuracy reach a bottleneck. The gravity geological method requires the use of ship survey data points to calculate shortwave information. If the ship survey water depth data points are insufficient, the shortwave information will also be reduced. The SAS method relies on ship survey water depth data points to calculate the proportional factor, and then generates a grid through the proportional factor to obtain the spatial mapping relationship between gravity and water depth. If the ship survey data points are insufficient, it will also affect the accuracy of the proportional factor. Therefore, both methods require sufficient ship survey water depth data, and cannot use the rich ship survey water depth data in other areas to assist in inversion.

[0004] With the development of artificial intelligence technology, artificial intelligence has been widely used in data processing algorithms in geosciences such as seismology, geophysics, and geochemistry, and has achieved remarkable results. Many scholars have tried to apply artificial intelligence technology to the inversion of seafloor topography based on ocean gravity data based on altimetry, and have achieved initial success. For example, Annan and Wan (2022) and Wan Xiaoyun et al. (2023) used neural networks with different structures to predict absolute water depth, such as inferring vertical deflection and gravity gradient deflection from gravity and its related quantities. Moran et al. (2022) tested a variety of machine learning algorithms, hoping to predict water depth through a series of geophysical and oceanographic characteristics. Yang Lei (2023) used shortwave gravity anomalies and other geophysical information as input, used fully connected networks and convolutional networks for prediction, and evaluated the performance of different input and model combinations. Sun Yongjin (2023) improved the accuracy of water depth measurement based on altimetry data by combining neural networks and gravity wavelet decomposition methods, and the accuracy was improved by 12.45% compared with the use of neural network methods alone. The powerful computing power in the field of artificial intelligence can mine the characteristic relationships between spatial data. Compared with traditional physical models, the complexity of its models can consider disturbances other than the main influencing variables. Artificial intelligence methods can establish a mapping between gravity and water depth in areas with dense ship-measured water depth data, and use a small amount of sparse ship-measured water depth data for fine-tuning in areas where ship-measured water depth data is scarce, thereby improving the efficiency of using ship-measured water depth data.

[0005] The advantage of artificial intelligence methods over traditional methods is that the model complexity can improve accuracy and can be transferred to areas where ship-measured water depth data is lacking. However, there are currently the following problems with artificial intelligence in inverting seabed topography from the ocean gravity field: ① Artificial intelligence methods use a large number of parameters to explore the characteristics between gravity and sea depth and fit the relationship between gravity and sea depth, but insufficient data will limit the effect of feature extraction. ② There are many types of gravity-related parameters, and the physical environment of different seabeds is affected differently by different parameters. It is impossible to accurately assign corresponding weights to different parameters, which affects the accuracy of the generated seabed topography. Summary of the invention

[0006] The purpose of this application is to overcome the defect of low accuracy of seabed topography prediction in the prior art.

[0007] In order to achieve the above objectives, the present application proposes an artificial intelligence inversion system for seafloor topography using gravity field spatial spectrum data, the system comprising:

[0008] The gravity field spatial spectrum calculation module is used to use several bandpass filters to intercept gravity data at different wavelengths to obtain the gravity field spatial spectrum; each band represents the gravity information at the corresponding wavelength;

[0009] The artificial intelligence inversion seabed topography module is used to input the spatial spectrum of the gravity field into the trained gravity-water depth mapping model to generate water depth prediction values.

[0010] As an improvement of the above system, the processing process of the gravity field spatial spectrum calculation module includes:

[0011] Calculate the gravity signal in the frequency domain;

[0012] Set wavelength nodes for bandpass filtering; establish k-1 bandpass filters, each filter processes the wavelength range between two nodes; k is the number of wavelength nodes;

[0013] Multiply the gravity signal with k-1 bandpass filters and perform inverse Fourier transform on each filtered frequency band;

[0014] The two-dimensional gravity signals processed by each filter are superimposed layer by layer to obtain a gravity spatial spectrum with k-1 bands in three dimensions.

[0015] As an improvement of the above system, the gravity signal is gravity anomaly or vertical gravity gradient.

[0016] As an improvement of the above system, the system also includes a gravity-water depth mapping relationship establishment module, which is used to train a machine learning model or a neural network model using the gravity field spatial spectrum obtained by the gravity field spatial spectrum calculation module to obtain a trained gravity-water depth mapping model.

[0017] As an improvement of the above system, the processing process of the gravity water depth mapping relationship establishment module includes:

[0018] Obtain the gravity field spatial spectrum of the position corresponding to the sparse ship water depth data, and the gravity field spatial spectrum and water depth at each position form a data pair;

[0019] Divide all data pairs into training sets and test sets according to the set ratio;

[0020] According to the model's requirements for the input data format, choose whether to convert the training set and test set into tensors; normalize the data and water depth of each band of the gravity field spatial spectrum respectively;

[0021] The normalized training set data is input into a machine learning model or a neural network model for training to obtain a trained gravity-water depth mapping model.

[0022] As an improvement of the above system, the processing process of the artificial intelligence inversion seabed topography module includes:

[0023] According to the requirements of the artificial intelligence model for the input data format, choose whether to convert the gravity field spatial spectrum data into tensors to normalize the data of each band of the gravity field spatial spectrum;

[0024] The normalized gravity spatial spectrum data is input into the trained gravity-water depth mapping model to obtain the seabed topography. The prediction results of the model are compared with the actual ship-measured water depth for accuracy evaluation, and the prediction accuracy is obtained according to the accuracy evaluation index.

[0025] The present application also provides an artificial intelligence method for inverting seafloor topography using gravity field spatial spectrum data, which is implemented by the above system and includes:

[0026] The gravity field spatial spectrum calculation module intercepts gravity data at different wavelengths to obtain the gravity field spatial spectrum;

[0027] The spatial spectrum of the gravity field is input into the trained gravity-water depth mapping model in the artificial intelligence inversion seabed topography module to generate the water depth prediction value.

[0028] As an improvement of the above method, the method further comprises:

[0029] The gravity field spatial spectrum obtained by the gravity field spatial spectrum calculation module is used to train a machine learning model or a neural network model to obtain a trained gravity-water depth mapping model.

[0030] As an improvement of the above method, the method further comprises:

[0031] The prediction results of the model are compared with the actual ship-measured water depth for accuracy evaluation, and the prediction accuracy is obtained based on the accuracy evaluation index.

[0032] Compared with the prior art, the advantages of this application are:

[0033] 1. Gravity field spatial spectrum, compared with direct use of unprocessed full-band gravity anomaly data for water depth prediction, can explore the characteristics of seabed landforms more deeply from multiple scales and angles. This method not only improves the ability to identify small-scale terrain features, but also makes the analysis of the complex relationship between gravity and topography more detailed and accurate.

[0034] 2. The spatial spectrum of the gravity field also enhances the model's ability to handle the complex relationship between terrain and gravity information. The response relationship between gravity and terrain at different wavelengths is different, enabling machine learning and deep learning technologies to better play their "non-linear" role. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 Shown is the structure diagram of the seafloor topography inversion system using artificial intelligence using gravity field spatial spectrum data;

[0036] Figure 2 Shown is a flow chart of the artificial intelligence inversion method for seafloor topography using gravity field spatial spectrum data. DETAILED DESCRIPTION

[0037] The technical solution of the present application is described in detail below with reference to the accompanying drawings.

[0038] Since the existing artificial intelligence gravity inversion technology for seafloor topography is difficult to mine feature relationships when the amount of data is insufficient, especially for terrains in different physical environments, which are affected differently by different gravity parameters, simply stacking the number of parameters cannot adapt to the work of various types of terrain inversion. The applicant found through research that terrains of different wavelengths will have different degrees of influence on gravity. Therefore, it is proposed to enrich the characteristics of gravity data by calculating the spatial spectrum of the gravity field to improve the effect of artificial intelligence inversion of seafloor topography. Gravity of different wavelengths can be adaptively weighted, considering their different degrees of influence on terrain inversion, thereby improving the accuracy of gravity inversion of seafloor topography.

[0039] This application proposes an artificial intelligence inversion system for seafloor topography using gravity field spatial spectrum data, the system comprising:

[0040] Gravity field spatial spectrum calculation module: Use several bandpass filters to intercept gravity data at different wavelengths (such as gravity anomalies or vertical gravity gradients, etc.), and each band represents the gravity information at the corresponding wavelength.

[0041] Gravity-water-depth mapping relationship establishment module: used to process gravity and water depth data, and establish their correlation. This module obtains the spatial spectrum of the gravity field at the corresponding position of the sparse ship water depth data, and pairs the spatial spectrum of the gravity field at each position with the water depth data. Next, divide the data set into training and test sets in appropriate proportions. According to the requirements of the artificial intelligence model for the input data format, choose whether to convert it into a tensor and perform normalization. After that, design an artificial intelligence model suitable for regression tasks. The model can be a machine learning method (such as random forest or support vector machine) or a neural network method (such as fully connected network, convolutional neural network, Transformer, etc.) for training and predicting water depth.

[0042] Artificial intelligence inversion seabed topography module: used to evaluate the accuracy of seabed topography and invert regional seabed topography. Take the gravity field spatial spectrum in the test data, input the trained gravity-water depth mapping relationship model, generate water depth prediction values, and evaluate the prediction performance of the model by calculating accuracy indicators (such as standard deviation, root mean square error, correlation coefficient, etc.). Then, according to the requirements of the artificial intelligence model for the input data format, choose whether to convert the gravity field spatial spectrum data in the study area into a tensor and normalize each band. Finally, input the normalized gravity field spatial spectrum data into the model to generate a seabed topography map of the entire area.

[0043] The processing of the gravity field spatial spectrum calculation module includes:

[0044] Step A1: 2D Fourier transform:

[0045]

[0046] f x and f y are the coordinates in the frequency domain, ranging from 0 to M-1 and 0 to N-1 respectively. It is a complex exponential term that describes the transformation of the signal in the frequency domain.

[0047] Step A2: Set the band node and apply bandpass filtering. Bandpass filter the gravity data according to the set band node. Define filter W (Band Notes) i

[0048] W (Band Notes) i ={1, if Wavelenth∈(Band Notes i , Band Notes i+1 ], 0, otherwise}

[0049] Band Notes is the wavelength at the set band node. The number of Band Notes nodes can be any number, and the band nodes can be adjusted according to the characteristics of the study area. For example, suppose the set band nodes are:

[0050] Band Notes = [0, 3, 6, 10, 15, 30, 60, 100, 140, 180, 300, 1000] (unit: km)

[0051] Step A3: Calculate the gravity for each band

[0052] G0Spe{i}(f x , f y )=G0(f x , f y )·W(Band Notes{i})

[0053] Inverse Fourier transform back to the spatial domain:

[0054]

[0055] Step A4: Splicing gravity space spectrum bands

[0056] Concatenate all processed gravity space spectrum bands along the third dimension:

[0057] g0Spe(f x , f y )=stack[g0Spe{i}(f x , f y ), ..., g0Spe{n}(f x , f y )]

[0058] The processing of the gravity-water-depth mapping relationship establishment module includes:

[0059] Step B1: Obtain the gravity field spatial spectrum at the location corresponding to the sparse ship-measured water depth data. For the gravity field spatial spectrum at each location, extract the corresponding water depth data to form a data pair, that is, the gravity field spatial spectrum at each location is paired with the corresponding water depth data for subsequent model training and prediction.

[0060] Step B2: Divide all acquired data pairs into training sets and test sets according to a certain ratio (such as 7:3, 8:2 or 9:1). The training set is used to train the model, and the test set is used to evaluate the performance of the model. The data can be divided randomly or reasonably grouped according to characteristics such as spatial location to ensure that the training and test data are representative in distribution.

[0061] Step B3: According to the AI ​​model's requirements for input data format, choose whether to convert the training set and test set data into tensors to facilitate parallel computing and model training. Next, normalize the data of each band of the gravity field spatial spectrum and the corresponding water depth data, and scale them to the interval of [0,1] or [-1,1] to prevent the magnitude difference of different features from adversely affecting model training and improve the training efficiency and stability of the model.

[0062] Step B4: Design an AI model suitable for the regression task. The model can be a traditional machine learning method, such as random forest, support vector machine, etc., or a deep learning method based on neural network, such as fully connected neural network, convolutional neural network, or Transformer, etc.

[0063] Step B5: Define a suitable optimizer (such as SGD or Adam, etc.) to guide the model parameter update. Define a suitable loss function (such as mean square error loss function MSE Loss) to measure the difference between the model prediction value and the true value.

[0064] Step B6: Set several epochs. In each epoch, the model loads the training set data in batches through the data loader, calculates the loss batch by batch, and updates the model parameters through the optimizer. The training will continue until the loss function converges, that is, the model error is reduced to a low level, or the set training rounds are reached. After the training is completed, the final regression model is saved to obtain the mapping relationship model between the gravity field and the water depth, which is used for subsequent water depth prediction tasks.

[0065] The processing of the artificial intelligence inversion seabed topography module includes:

[0066] Step C1: The space of the gravity field extracted from the test data obtained in step B3 is input into the gravity-water depth mapping model to generate the corresponding water depth prediction value. The predicted water depth value is compared with the shipborne bathymetric data, and the prediction performance of the model is quantified by calculating a series of accuracy evaluation indicators (such as standard deviation, root mean square error RMSE, correlation coefficient, etc.) to evaluate the accuracy and reliability of the model.

[0067] Step C2: According to the AI ​​model's requirements for input data format, choose whether to integrate and convert all the spatial spectrum data of the gravity field in the study area into tensor form for further data processing. For each band of the gravity field spatial spectrum, perform data normalization to ensure the scale consistency of the data and prepare for subsequent model input.

[0068] Step C3: Input the gravity field spatial spectrum processed in step C2 into the gravity-water depth mapping relationship model, and generate the seabed topography data of the entire study area through the calculation of the model.

[0069] The present application also provides an artificial intelligence method for inverting seafloor topography using gravity field spatial spectrum data, which is implemented based on the above system and includes:

[0070] Step S1: First, calculate the gravity signal in the frequency domain, such as gravity anomaly or vertical gravity gradient. Then, set the band nodes of the gravity field spatial spectrum and perform truncated bandpass filtering to retain the information in a specific wavelength range. The node position can be adjusted according to the characteristics of the study area. Then, perform an inverse Fourier transform on the filtered frequency domain gravity of each band and convert it into spatial domain data. Finally, superimpose all bands in layers to form a gravity field spatial spectrum with multiple bands in three dimensions.

[0071] Step S2: First, obtain the spatial spectrum of the gravity field at the corresponding position of the sparse ship water depth data, and pair the spatial spectrum of the gravity field at each position with the water depth data. Next, divide the data set into a training set and a test set according to a certain ratio. According to the requirements of the artificial intelligence model for the input data format, choose whether to convert it into a tensor, and normalize the gravity field spatial spectrum and water depth data respectively. Subsequently, design and train an artificial intelligence model (such as random forest, support vector machine, fully connected network, etc.), and use an optimizer (such as SGD or Adam) and a loss function (such as MSE Loss) to update the model parameters until the loss function converges. After the training is completed, save the regression model, and finally obtain the mapping relationship model between the gravity field and the water depth.

[0072] Step S3: First, take the gravity field spatial spectrum in the test set obtained in step S2 and input it into the trained gravity-water depth mapping relationship model to obtain the water depth prediction value, and evaluate the accuracy of the prediction result and the actual ship-measured water depth, and calculate the accuracy evaluation index (for example: standard deviation, root mean square error and correlation coefficient, etc.) to measure the performance of the model. Then, according to the requirements of the artificial intelligence model for the input data format, choose whether to convert the gravity field spatial spectrum data in the study area into a tensor and normalize each band. Finally, input the normalized gravity spatial spectrum data into the gravity-water depth mapping relationship model to generate the seabed topography of the entire area.

[0073] Example 1

[0074] The present application proposes an artificial intelligence inversion system for seafloor topography using gravity field spatial spectrum data, including a gravity field spatial spectrum calculation module, a gravity-water depth mapping relationship establishment module and an artificial intelligence inversion module for seafloor topography.

[0075] The design of each module is as follows:

[0076] 1. Gravity field spatial spectrum calculation module

[0077] Gravity data (such as gravity anomalies or vertical gravity gradients) are intercepted from different wavelengths using multiple bandpass filters, and each band corresponds to gravity information at a specific wavelength.

[0078] The processing of the gravity field spatial spectrum calculation module includes:

[0079] Step 1: Calculate the gravity signal in the frequency domain (such as gravity anomaly or vertical gravity gradient).

[0080] Step 2: Set wavelength nodes, which are used for bandpass filtering. You can choose any number of nodes, and the specific location can be adjusted according to the characteristics of the study area. If there are k nodes, there will be k-1 bandpass filters, each of which processes the wavelength range between two nodes.

[0081] Step 3: Multiply the gravity signal with k-1 bandpass filters. Perform an inverse Fourier transform on each filtered frequency band.

[0082] Step 4: Superimpose the two-dimensional gravity signals processed by each filter layer by layer to obtain a gravity field spatial spectrum with k-1 bands in three dimensions.

[0083] 2. Gravity-water-depth mapping relationship establishment module:

[0084] This module is used to process gravity and water depth data and establish the relationship between the two. Obtain the spatial spectrum of the gravity field at the location corresponding to the sparse ship water depth data, and pair the spatial spectrum of the gravity field at that location with the corresponding water depth data. Divide the data set into training and test sets in an appropriate proportion, and then choose whether to convert the data into a tensor and perform normalization according to the requirements of the artificial intelligence model for the input data format. Design an artificial intelligence model suitable for regression tasks. This model can use machine learning methods (for example: random forest or support vector machine) or neural network methods (such as fully connected networks, convolutional neural networks or Transformers, etc.) to train data and predict water depth values.

[0085] Step 1: Obtain the gravity field spatial spectrum of the location corresponding to the sparse ship-derived water depth data. The gravity field spatial spectrum and water depth at each location form a data pair.

[0086] Step 2: Divide all data pairs into training sets and test sets. The data of the training sets and test sets are divided in a suitable ratio (eg, 7:3, 8:2 or 9:1) and manner.

[0087] Step 3: According to the AI ​​model's requirements for the input data format, choose whether to convert the training set and test set into tensors. Normalize the data and water depth of each band of the gravity field spatial spectrum.

[0088] Step 4: Design an artificial intelligence model, which can be machine learning (e.g., random forest, support vector machine, etc.) or neural network (e.g., fully connected network, convolutional neural network, Transformer, etc., which can be used for regression tasks).

[0089] Step 5: Define the optimizer (e.g. SGD or Adam, etc.) and loss function (e.g. MSE Loss).

[0090] Step 6: The model uses the data loader in step 2 to load data batches in several epochs, calculates the loss and updates the model parameters until the loss converges. The training ends, saves the regression model, and obtains the gravity-water depth mapping relationship model.

[0091] 3. Artificial intelligence inversion of seafloor topography module: used to evaluate the accuracy of seafloor topography and invert regional seafloor topography. First, the gravity field spatial spectrum in the test data is extracted and input into the trained gravity-water depth mapping model to generate the water depth prediction value. The model performance is evaluated by calculating the accuracy indicators such as standard deviation, root mean square error, correlation coefficient, etc. The normalized data is input into the model to generate the seafloor topography of the entire area.

[0092] Step 1: Take the gravity field spatial spectrum in the test data obtained in the previous module, input the gravity-water depth mapping relationship model to obtain the water depth prediction value, and calculate the accuracy index (for example: standard deviation, root mean square error, correlation coefficient, etc.) with the corresponding ship-measured water depth to evaluate the performance of predicting water depth.

[0093] Step 2: According to the requirements of the artificial intelligence model for the input data format, choose whether to convert the gravity spatial spectrum data of the entire study area into a tensor and normalize the data of each band of the gravity field spatial spectrum.

[0094] Step 3: Input the gravity spatial spectrum data normalized in the previous step into the gravity-water depth mapping relationship model to obtain the seabed topography of the entire area.

[0095] Example 2

[0096] like Figure 2 As shown, the present application also proposes an artificial intelligence inversion method for seafloor topography using gravity field spatial spectrum data, which is implemented based on the above system and includes:

[0097] Step 1: Calculate the gravity signal in the frequency domain (such as gravity anomaly or vertical gravity gradient).

[0098] Step 2: Set wavelength nodes, which are used for bandpass filtering. You can choose any number of nodes, and the specific location can be adjusted according to the characteristics of the study area. If there are k nodes, there will be k-1 bandpass filters, each of which processes the wavelength range between two nodes.

[0099] Step 3: Multiply the gravity signal with k-1 bandpass filters. Perform an inverse Fourier transform on each filtered frequency band.

[0100] Step 4: Superimpose the two-dimensional gravity signals processed by each filter layer by layer to obtain a gravity field spatial spectrum with k-1 bands in three dimensions.

[0101] Step 5: Obtain the spatial spectrum of the gravity field at the corresponding position of the sparse ship-derived water depth data, and pair the gravity field spectrum at each position with the water depth value to form a data pair.

[0102] Step 6: Divide all data pairs into training sets and test sets in an appropriate ratio (e.g., 7:3, 8:2, or 9:1) and perform the split in an appropriate manner.

[0103] Step 7: According to the AI ​​model's requirements for the input data format, choose whether to convert the training set and test set into tensor format. Normalize each band of the gravity field spectrum and the water depth data separately.

[0104] Step 8: Design an artificial intelligence model. The model can be a machine learning method (such as random forest, support vector machine, etc.) or a neural network (such as fully connected neural network, convolutional neural network, Transformer, etc. suitable for regression tasks).

[0105] Step 9: Select an optimizer (such as SGD or Adam, etc.) and a loss function (such as mean squared error loss function MSELoss).

[0106] Step 10: During the model training process, use the data loader in step 2 to load data in batches in several epochs, calculate the loss and continuously update the model parameters until the loss function converges. After the training is completed, save the regression model, that is, the gravity-water depth mapping model.

[0107] Step 11: Apply the model trained in the previous step to the spatial spectrum of the gravity field in the test data to predict the water depth value. By comparing it with the ship-measured water depth, calculate the accuracy indicators such as standard deviation, root mean square error, correlation coefficient, etc. to evaluate the prediction effect of the model.

[0108] Step 12: According to the requirements of the artificial intelligence model for the input data format, choose whether to convert the gravity field spectrum data of the entire study area into a tensor and normalize the data of each spectral band.

[0109] Step 13: Input the gravity field spectrum data processed in step 12 into the trained gravity-water depth mapping model to generate a seabed topography map of the entire area.

[0110] Table 1 summarizes the verification results of ship-measured water depth data, and compares the accuracy of water depth inverted by artificial intelligence methods using gravity field spatial spectrum, artificial intelligence methods without gravity field spatial spectrum, and traditional methods. It can be seen that the root mean square error of gravity field spatial spectrum is greatly reduced, whether it is random forest or fully connected network. And the S&S of gravity field spatial spectrum fully connected network is reduced from 105 meters to 78 meters compared with the traditional method. This shows the superiority of gravity field spatial spectrum for seabed topography inversion.

[0111] Table 1 Verification results of ship depth measurement (unit: m)

[0112]

[0113] The present application may also provide a computer device, comprising: at least one processor, a memory, at least one network interface and a user interface. The various components in the device are coupled together through a bus system. It is understood that the bus system is used to achieve connection and communication between these components. In addition to the data bus, the bus system also includes a power bus, a control bus and a status signal bus.

[0114] The user interface may include a display, a keyboard or a pointing device, such as a mouse, a trackball, a touch pad or a touch screen.

[0115] It is understood that the memory in the embodiments disclosed in the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus random access memory (DRRAM). The memories described herein are intended to include, but are not limited to, these and any other suitable types of memories.

[0116] In some embodiments, the memory stores the following elements, executable modules or data structures, or a subset thereof, or an extended set thereof: an operating system and applications.

[0117] The operating system includes various system programs, such as a framework layer, a core library layer, a driver layer, etc., which are used to implement various basic services and process hardware-based tasks. The application includes various application programs, such as a media player (Media Player), a browser (Browser), etc., which are used to implement various application services. The program for implementing the method of the embodiment of the present disclosure can be included in the application.

[0118] In the above embodiment, the processor may also call a program or instruction stored in the memory, specifically, a program or instruction stored in an application program, and is used to:

[0119] Execute the steps of the above method.

[0120] The above method can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor or an instruction in the form of software. The above processor may be a general processor, a digital signal processor (Digital Signal Processor, DSP), an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The above-disclosed methods, steps and logic block diagrams can be implemented or executed. The general processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the above-disclosed method can be directly embodied as a hardware decoding processor to execute, or the hardware and software modules in the decoding processor are combined to execute. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0121] It is understood that the embodiments described in the present application can be implemented by hardware, software, firmware, middleware, microcode or a combination thereof. For hardware implementation, the processing unit can be implemented in one or more application specific integrated circuits (ASIC), digital signal processors (DSP), digital signal processing devices (DSPD), programmable logic devices (PLD), field programmable gate arrays (FPGA), general-purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described in the present application or a combination thereof.

[0122] For software implementation, the technology of the present application can be implemented by executing the functional modules (such as procedures, functions, etc.) of the present application. The software code can be stored in a memory and executed by a processor. The memory can be implemented in the processor or outside the processor.

[0123] The present application may also provide a non-volatile storage medium for storing a computer program. When the computer program is executed by a processor, each step in the above method embodiment can be implemented.

[0124] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present application and are not intended to limit it. Although the present application is described in detail with reference to the embodiments, a person skilled in the art should understand that any modification or equivalent replacement of the technical solution of the present application does not depart from the spirit and scope of the technical solution of the present application and should be included in the scope of the claims of the present application.

Claims

1. An artificial intelligence inversion system for seafloor topography using gravity field spatial spectrum data, characterized in that: The system comprises: A gravity field spatial spectrum calculation module is used to use a number of bandpass filters to intercept gravity data at different wavelengths to obtain a gravity field spatial spectrum; each band represents gravity information at a corresponding wavelength; and The artificial intelligence inversion seabed topography module is used to input the spatial spectrum of the gravity field into the trained gravity-water depth mapping model to generate water depth prediction values.

2. The artificial intelligence inversion system for seafloor topography using gravity field spatial spectrum data according to claim 1, characterized in that: The processing process of the gravity field spatial spectrum calculation module includes: Calculate the gravity signal in the frequency domain; Set wavelength nodes for bandpass filtering; establish k-1 bandpass filters, each filter processes the wavelength range between two nodes; k is the number of wavelength nodes; Multiply the gravity signal with k-1 bandpass filters and perform inverse Fourier transform on each filtered frequency band; The two-dimensional gravity signals processed by each filter are superimposed layer by layer to obtain a gravity spatial spectrum with k-1 bands in three dimensions.

3. The artificial intelligence inversion system for seafloor topography using gravity field spatial spectrum data according to claim 2, characterized in that: The gravity signal is a gravity anomaly or a vertical gravity gradient.

4. The artificial intelligence inversion system for seafloor topography using gravity field spatial spectrum data according to claim 1, characterized in that: The system further comprises: The gravity-water-depth mapping relationship establishment module is used to train a machine learning model or a neural network model using the gravity field spatial spectrum obtained by the gravity field spatial spectrum calculation module to obtain a trained gravity-water-depth mapping model.

5. The artificial intelligence inversion system for seafloor topography using gravity field spatial spectrum data according to claim 4, characterized in that: The processing process of the gravity water depth mapping relationship establishment module includes: Obtain the gravity field spatial spectrum of the position corresponding to the sparse ship water depth data, and the gravity field spatial spectrum and water depth at each position form a data pair; Divide all data pairs into training sets and test sets according to the set ratio; According to the model's requirements for the input data format, choose whether to convert the training set and test set into tensors; normalize the data and water depth of each band of the gravity field spatial spectrum respectively; The normalized training set data is input into a machine learning model or a neural network model for training to obtain a trained gravity-water depth mapping model.

6. The artificial intelligence inversion system for seafloor topography using gravity field spatial spectrum data according to claim 1, characterized in that: The processing process of the artificial intelligence inversion seabed topography module includes: According to the requirements of the artificial intelligence model for the input data format, choose whether to convert the gravity space spectrum data into tensors and normalize the data of each band of the gravity field space spectrum; The processed gravity spatial spectrum data are input into the trained gravity-bathymetry mapping model to obtain the seabed topography.

7. The artificial intelligence inversion system for seafloor topography using gravity field spatial spectrum data according to claim 6, characterized in that: The processing process of the artificial intelligence inversion seabed topography module also includes: The prediction results of the model are compared with the actual ship-measured water depth for accuracy evaluation, and the accuracy of the predicted seabed topography is obtained based on the accuracy evaluation index.

8. An artificial intelligence method for inverting seafloor topography using gravity field spatial spectrum data, implemented based on the system of any one of claims 1 to 7, the method comprising: The gravity field spatial spectrum calculation module intercepts gravity data at different wavelengths to obtain the gravity field spatial spectrum; The spatial spectrum of the gravity field is input into the trained gravity-water depth mapping model in the artificial intelligence inversion seabed topography module to generate the water depth prediction value.

9. The artificial intelligence inversion method for seafloor topography using gravity field spatial spectrum data according to claim 8, characterized in that: The method further comprises: The gravity field spatial spectrum obtained by the gravity field spatial spectrum calculation module is used to train a machine learning model or a neural network model to obtain a trained gravity-water depth mapping model.

10. The artificial intelligence inversion method for seafloor topography using gravity field spatial spectrum data according to claim 9, characterized in that: The method further comprises: The seabed topography was predicted using the gravity-bathymetry mapping model. The prediction results of the model were compared with the actual ship-measured water depth for accuracy evaluation, and the prediction accuracy was obtained based on the accuracy evaluation index.

Images