A GNSS-R-based method and system for locating maritime targets

Abstract

This invention provides a GNSS-R-based method and system for locating maritime targets, belonging to the field of remote sensing technology. The method includes: collecting observation data from multiple global navigation satellite systems in the target area and extracting a time-delay Doppler shift map; performing a stationarity test on the time-delay Doppler shift map and training a pre-set neural network model to obtain a predicted sea clutter time-delay Doppler shift map; subtracting the predicted sea clutter time-delay Doppler shift map from the predicted target time-delay Doppler shift map to obtain a sea clutter-free time-delay Doppler shift map; selecting an initial target position and converting the initial target position into a spatial position based on a unified coordinate system; and calculating the converted spatial position using geometric semantic constraints to obtain the maritime target position. This invention enables long-term continuous monitoring of maritime targets in a specific area using satellite reflection signals, allowing for real-time and rapid precise location of maritime targets. This is of great significance for ensuring maritime shipping safety, preventing marine disasters, and conducting maritime rescue operations.

Description

Technical Field

[0001] This invention relates to the field of remote sensing technology, and in particular to a method and system for locating maritime targets based on GNSS-R. Background Technology

[0002] Target location at sea is a crucial means of ensuring maritime safety, preventing marine disasters, and conducting maritime rescue operations. This is especially important for immobile targets that require tugboat towing, such as drilling platforms, which require precise location to guarantee navigational safety.

[0003] There are many traditional methods for locating maritime targets. One method uses shore-based radar, which, while technologically mature, is generally fixed in location and has a limited detection range, only able to detect near-shore waters. Another method uses spaceborne synthetic aperture radar (SAR), providing all-weather, 24 / 7 Earth observation capabilities and enabling rapid target location. However, spaceborne SAR is limited by satellite orbits, resulting in long reentry periods. Finally, airborne SAR is used to detect maritime targets. Compared to spaceborne SAR, airborne SAR offers advantages such as operational flexibility, high timeliness, and high resolution, but also suffers from shorter loiter times and higher costs.

[0004] Building upon this foundation, Global Navigation Satellite System-Reflectometry (GNSS-R) is an emerging microwave remote sensing technology, essentially based on the bistatic radar principle. When a target appears on the sea surface, its Fresnel scattering coefficient and radar cross-section are much larger than those of the sea surface due to the target's metallic surface. This characteristic is used to distinguish between the sea surface and other maritime targets, thus detecting the sea surface target. By measuring the delay (time delay or phase delay) between the direct signal and the signal reflected from the surface using GNSS, and then determining the target's location based on the geometric positional relationship between the GNSS satellite, the onboard receiver, and the specular reflection point, the target can be located. Some methods utilize TDS-1 satellite data to reconstruct the sea surface scattering region for maritime target location, employing an adaptive threshold sea clutter compensation algorithm. However, due to the long reentry period of TDS-1, continuous tracking of the same target is not possible. Furthermore, there are methods that utilize multi-satellite observations from the CYGNSS global navigation satellite system, taking advantage of the short reentry period of the CYGNSS constellation to achieve rapid positioning of sea surface targets. However, the accuracy of this positioning method is still somewhat lacking because it only uses two satellites, failing to leverage the advantages of multiple satellites, and it does not combine the motion characteristics of sea targets for positioning. Summary of the Invention

[0005] This invention provides a GNSS-R-based method and system for locating maritime targets, which addresses the shortcomings of existing technologies that use radar for maritime target location due to high implementation costs, or use satellite positioning due to overly limited data and long implementation cycles, and which do not consider the motion characteristics of the maritime targets themselves.

[0006] In a first aspect, the present invention provides a GNSS-R-based method for locating maritime targets, comprising:

[0007] Collect observation data from multiple global navigation satellite systems in the target area, and extract the time delay Doppler frequency shift map from the observation data of the multiple global navigation satellite systems;

[0008] The stationarity of the time-delay Doppler frequency shift map is tested and a preset neural network model is trained to obtain the predicted sea clutter time-delay Doppler frequency shift map;

[0009] Obtain the target delay Doppler frequency shift map of the target area to be predicted, and subtract the predicted sea clutter delay Doppler frequency shift map from the target delay Doppler frequency shift map to obtain the sea clutter-free delay Doppler frequency shift map;

[0010] The initial target position in the sea clutter-eliminating delay Doppler frequency shift map is selected, and the initial target position is converted into a spatial position based on a unified coordinate system;

[0011] The spatial position is transformed using geometric semantic constraints to obtain the position of the maritime target.

[0012] Secondly, the present invention also provides a GNSS-R-based maritime target positioning system, comprising:

[0013] The acquisition module is used to acquire observation data from multiple global navigation satellite systems in the target area and extract the time delay Doppler frequency shift map of the observation data from the multiple global navigation satellite systems;

[0014] The training module is used to perform stationarity checks and train a preset neural network model on the time delay Doppler frequency shift map to obtain the predicted sea clutter time delay Doppler frequency shift map;

[0015] The elimination module is used to obtain the target delay Doppler frequency shift map of the target area to be predicted, and to subtract the predicted sea clutter delay Doppler frequency shift map from the target delay Doppler frequency shift map to obtain the eliminated sea clutter delay Doppler frequency shift map;

[0016] A unified module is used to filter the initial target position in the sea clutter-eliminating delay Doppler frequency shift map and convert the initial target position into a spatial position based on a unified coordinate system;

[0017] The transformation module is used to calculate and transform the spatial position using geometric semantic constraints to obtain the position of the target at sea.

[0018] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the GNSS-R-based maritime target positioning method as described above.

[0019] The GNSS-R-based maritime target positioning method and system provided by this invention can continuously monitor maritime targets in a specific area over a long period of time by using multiple satellite reflection signals. It can quickly and accurately locate the precise position of maritime targets in real time and has the advantages of short reentry cycle, high accuracy and low cost. It is of great significance for ensuring maritime shipping safety, preventing marine disasters and carrying out maritime rescue. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is one of the flowcharts of the GNSS-R-based maritime target positioning method provided by the present invention;

[0022] Figure 2 This is the second flowchart illustrating the GNSS-R-based maritime target positioning method provided by this invention.

[0023] Figure 3 This is a schematic diagram of the multi-satellite orbit observation method provided by the present invention;

[0024] Figure 4 This is a schematic diagram of the geometric constraints provided by the present invention;

[0025] Figure 5 This is a schematic diagram of the structure of the GNSS-R-based maritime target positioning system provided by the present invention;

[0026] Figure 6 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0028] Figure 1 This is one of the flowcharts illustrating the GNSS-R-based maritime target positioning method provided in this embodiment of the invention, such as... Figure 1 As shown, it includes:

[0029] Step 100: Collect observation data from multiple global navigation satellite systems in the target area, and extract the time delay Doppler frequency shift map from the observation data of the multiple global navigation satellite systems;

[0030] Step 200: Perform a stationarity test and train a preset neural network model on the time delay Doppler frequency shift map to obtain the predicted sea clutter time delay Doppler frequency shift map;

[0031] Step 300: Obtain the target delay Doppler frequency shift map of the target area to be predicted, and subtract the predicted sea clutter delay Doppler frequency shift map from the target delay Doppler frequency shift map to obtain the sea clutter-free delay Doppler frequency shift map;

[0032] Step 400: Filter the initial target position in the sea clutter-eliminating delay Doppler frequency shift map, and convert the initial target position into a spatial position based on a unified coordinate system;

[0033] Step 500: Calculate and transform the spatial position using geometric semantic constraints to obtain the position of the target at sea.

[0034] This invention targets maritime targets, particularly drilling platforms. It utilizes GNSS measurements of the delay (time delay or phase delay) between the direct signal and the signal reflected from the Earth's surface. Based on the geometric relationship between the GNSS satellite, the onboard receiver, and the specular reflection point, the maritime target can be located. Sea clutter sequences are predicted using ST-LSTM, and the original signal is then cleared of sea clutter. Since each Doppler Delay Map (DDM) establishes a coordinate system with its sub-satellite point (SP) as the origin, the coordinates need to be unified into a single coordinate system. Considering the generally linear motion of ships, geometric semantic constraints, specifically linear constraints, are added to the traditional Range-Doppler Equation to improve the accuracy of target detection.

[0035] Specifically, such as Figure 2 As shown, the first step is data preparation. In this embodiment of the invention, observation data from three CYGNSS satellites passing through the target area are prepared, and their time delay-Doppler shift maps (DDMs) are extracted. The stationarity of the DDMs is then checked, and the power time series of a fixed specific delay τ and Doppler fd cells are used. Autocorrelation calculations are performed to check for stationarity, and sequence data that meets the requirements are extracted as the pre-training set for the neural network model. Then, a large number of sea clutter DDMs that pass the stationarity test and are located in targetless sea areas are used as the pre-training set to train an improved long short-term memory network ST-LSTM. The trained model predicts sea clutter DDMs by inputting the sea clutter DDMs into the ST-LSTM and predicting future sea clutter DDM sequences based on previous sea clutter signals. The predicted sea clutter DDM sequences are then subtracted from the DDM sequences of possible targets to obtain the DDM sequences with eliminated sea clutter information. Finally, the position coordinates (τ, fd) of the maritime target in the DDM map are determined by the threshold method. Next, a unified coordinate system is established. Since each DDM map establishes a coordinate system with its nadir point (SP) as the origin, the coordinates need to be unified into the same coordinate system. Because CYGNSS data is in the earth-centered earth-fixed (ECEF) coordinate system, coordinate system transformation is required. The data after unifying the coordinate system is combined with geometric semantic constraints to obtain the position of the maritime target. Since maritime targets generally move approximately in a straight line in open sea areas, a straight line constraint is added to the traditional Range-Doppler Equation to obtain a more accurate position of the maritime target. Finally, the obtained target position is used to obtain the coordinates of the maritime target in the WGS-84 coordinate system through the latitude and longitude distance conversion formula.

[0036] This invention enables long-term continuous monitoring of maritime targets in a specific area using multiple satellite reflection signals. It can quickly and accurately locate the precise position of maritime targets in real time, and has the advantages of short reentry cycle, high accuracy and low cost. It is of great significance for ensuring maritime shipping safety, preventing marine disasters and conducting maritime rescue.

[0037] Based on the above embodiments, observation data from multiple Global Navigation Satellite Systems (GNSS) are collected in the target area, and time-delay Doppler frequency shift maps of the observation data from the multiple GNSS systems are extracted, including:

[0038] The system determines which satellites in a global navigation satellite system collect observation data from the multiple global navigation satellite systems.

[0039] The observation data of the multiple global navigation satellite systems collected by each satellite are used as a single time delay Doppler shift map sequence in the time delay Doppler shift map.

[0040] Understandably, in order to overcome the problem that the results of existing satellite-based maritime target positioning are too one-sided due to insufficient satellite data, this embodiment of the invention considers the accuracy and timeliness of target positioning in the data preparation stage. It uses data from three satellites to calculate the target position. By preparing observation data from three CYGNSS satellites passing through the target area, the satellite observation data is read and its time delay-Doppler shift map (DDM) is extracted.

[0041] Based on the above embodiments, a stationarity test is performed on the time-delay Doppler frequency shift map, including:

[0042] A single time-delay Doppler frequency shift map in a single time-delay Doppler frequency shift map sequence is converted into a one-dimensional row vector, and the one-dimensional row vectors corresponding to multiple time-delay Doppler frequency shift maps in the single time-delay Doppler frequency shift map sequence are combined to form a matrix;

[0043] Each column vector of the matrix is ​​determined as a power time series of any delay and Doppler shift unit. Autocorrelation is performed on all power time series based on the sequence lag number and the sequence average value. Power time series that are uncorrelated and exhibit tailing in the noise region are extracted as pre-trained sea clutter sequences.

[0044] Specifically, stationarity is tested on the DDM sequence of each satellite. Each DDM is converted into a 1×187 row vector. Then, the N DDMs are combined to obtain an N×187 matrix, where each column represents the power time series of a specific delay τ and Doppler fd cell. n∈[1,187]. The autocorrelation operation on this DDM sequence is expressed as:

[0045]

[0046] k is the sequence lag number. Let N be the mean of the sequence and N be the number of DDM sequences. By performing autocorrelation calculations on the DDM sequences, only sequences that exhibit no correlation in the noise region and whose pure sea clutter responses display a tailing characteristic can be used as pre-trained sea clutter sequences.

[0047] Based on the above embodiments, a preset neural network model is trained on the time delay Doppler frequency shift map to obtain a predicted sea clutter time delay Doppler frequency shift map, including:

[0048] The pre-trained sea clutter sequence is input into an improved long short-term memory ST-LSTM network for training to obtain a sea clutter sequence prediction model. The ST-LSTM network includes a four-layer ConvLSTM structure and a channel from top-level information to bottom-level information at adjacent time points.

[0049] The time-delay Doppler frequency shift map is input into the sea clutter sequence prediction model to obtain the predicted sea clutter time-delay Doppler frequency shift map.

[0050] Specifically, in this embodiment of the invention, a large number of sea clutter DDMs that have passed the stationarity test and are located in targetless sea areas are prepared as a pre-training set. Training an improved long short-term memory network, ST-LSTM.

[0051] Here, ST-LSTM is an improved LSTM network. It is a structural improvement on the standard LSTM network. Compared with the standard LSTM, ST-LSTM adds cell connections at corresponding positions between different layers. It is widely used in existing prediction scenarios where high accuracy is required for time variables and where time delays are prone to occur. In particular, because ST-LSTM has a four-layer ConvLSTM structure and adds channels for top-layer information to the bottom layers at adjacent time points, it can more completely preserve the time information of the upper layers and more accurately predict sea clutter sequences.

[0052] When it is necessary to detect a target, the sea clutter sequence preceding the target area needs to be obtained. The input is fed into a trained ST-LSTM network to obtain a predicted sea clutter sequence passing through the target area.

[0053] Based on the above embodiments, the sea clutter sequence predicted by ST-LSTM and the DDM sequence for detecting maritime targets will be compared. Subtracting the two yields the DDM sequence that retains only the target information.

[0054] Based on the above embodiments, the initial target position in the sea clutter time delay Doppler frequency shift map includes:

[0055] The sea clutter delay Doppler frequency shift map is filtered using a threshold method, and false alarm values ​​in the filtered sea clutter delay Doppler frequency shift map are removed to obtain the initial target position.

[0056] The initial target position includes the time delay and Doppler shift of each initial target position.

[0057] Specifically, after removing sea clutter from the DDM image, this embodiment of the invention requires filtering the results to obtain the time delay and Doppler frequency shift of the target location. A threshold method is used to filter the targets, removing false alarms, and obtaining the target location (τ, fd) in each DDM image.

[0058] Based on the above embodiments, converting the initial target position into a spatial position based on a unified coordinate system includes:

[0059] The origin of the coordinate system is determined by setting the nadir point of any Doppler shift map in the sea clutter time delay Doppler shift map as the origin of the coordinate system. The offset value between the nadir point of other Doppler shift maps in the sea clutter time delay Doppler shift map and the origin of the coordinate system is calculated. Based on the offset value, the other Doppler shift maps are mapped to the unified coordinate system.

[0060] By using a rotation matrix to transform the geocentric and geofixed coordinate system data of multiple global navigation satellites, satellite motion state data can be obtained.

[0061] By combining the unified coordinate system and the satellite motion state data, the initial target position is transformed to obtain the spatial position.

[0062] Specifically, in order to determine the true location of a target at sea, it is necessary to convert the target's position in the delay Doppler domain into a spatial position.

[0063] Based on the GNSS-R geometric model, a Local Reference Frame (LRF) is established. In the LRF, the nadir point (SP) of the image is used as the origin. Because multiple images exist, coordinate unification among target points is necessary when solving simultaneous equations. Since the LRF is not convenient for unifying coordinate systems, a station-centered coordinate system, also known as the local northeast-sky coordinate system (ENU), is introduced as an intermediate coordinate system. One DDM image is selected, and its nadir point is set as the origin to establish a unified coordinate system. The offsets of the nadir point in the unified coordinate system from those in other DDM images are calculated, and targets in other coordinate systems are mapped to this unified coordinate system.

[0064] Meanwhile, since CYGNSS data is in the earth-centered earth-fixed (ECEF) coordinate system, the original data needs to be transformed. The rotation matrix is ​​represented as follows:

[0065]

[0066] Where λ represents the longitude of the mirror reflection point. T represents the latitude of the mirror reflection point. x,ecef T y,ecef and T z,ecef T represents the x, y, and z coordinates in the Earth-centered Earth-fixed coordinate system. x,enu T y,enu and T z,enu These represent the x, y, and z coordinates in the station-centered coordinate system, respectively.

[0067] Based on the above embodiments, the spatial position is calculated and transformed using geometric semantic constraints to obtain the position of the maritime target, including:

[0068] Obtain the satellite speed, altitude, and serial number of multiple global navigation satellites, and obtain the elevation angle of the Global Positioning System satellites;

[0069] Based on the satellite velocity, satellite altitude, serial number, and GPS satellite elevation angle, the spatial position is calculated using the range-Doppler equation to obtain the initial position of the maritime target.

[0070] The observation sequence number of the maritime target at different positions, the slope and intercept of the target's motion line are determined. Based on the observation sequence number at different positions, the slope and the intercept, the initial maritime target position is further calculated using the geometric semantic constraints to obtain the maritime target position.

[0071] Specifically, the embodiment of the present invention combines geometric semantic constraints to obtain the target position: to solve for the target position, it can be obtained through the Range-Doppler Equation, and the target equation is as follows:

[0072]

[0073] Where i represents the sequence number of the CYGNSS satellite, with values ​​ranging from 1 to 3. V GNSS，i =(V GNSS，xi V GNSS，yi V GNSS，zi V represents the speed of the navigation satellite. CYGNSS，i =(V CYGNSS，xi V CYGNSS，yi V CYGNSS，zi ) represents the velocity of the CYGNSS satellite, h CYGNSS，i γ represents the altitude of the CYGNSS satellite. i This refers to the elevation angle of the GPS satellites. For example... Figure 3 As shown, joint observations of maritime targets were conducted using satellites CY01, CY02, and CY03, with points SP1, SP2, and SP3 in the figure. Solving the Range-Doppler Equation for only a single satellite cannot eliminate positional ambiguity and distinguish between real and false targets. However, by jointly solving the six Range-Doppler Equations listed by each satellite, the positional ambiguity problem can be resolved, directly yielding the target position resulting from the combined observations of all three satellites.

[0074] Furthermore, such as Figure 4As shown, maritime targets mostly move in straight lines to save time and fuel, such as when a drilling platform is towed by a ship. Based on this characteristic, geometric semantic constraints, namely straight-line constraints, are added to the traditional Range-Doppler Equation to improve the accuracy of target detection. The problem can be transformed into the following equation:

[0075]

[0076] Here, j is added to the above to represent the order in which the target is observed at different positions during its motion, with values ​​ranging from 1 to t. n Let k and b be the slope and intercept of the target's moving line, respectively. By solving the above equation, positional ambiguity can be eliminated, and the precise position and trajectory of the target at different times can be obtained. It can be seen that adding geometric semantic constraints can effectively reduce the impact of satellite detection errors on target positioning.

[0077] Finally, the position of the maritime target is converted to the WGS-84 coordinate system according to the latitude and longitude distance conversion formula.

[0078] The GNSS-R-based maritime target positioning system provided by this invention will be described below. The GNSS-R-based maritime target positioning system described below can be referred to in correspondence with the GNSS-R-based maritime target positioning method described above.

[0079] Figure 5 This is a schematic diagram of the structure of a GNSS-R-based maritime target positioning system provided in an embodiment of the present invention, as shown below. Figure 5 As shown, it includes: a data acquisition module 51, a training module 52, an elimination module 53, a unification module 54, and a conversion module 55, wherein:

[0080] The acquisition module 51 is used to acquire observation data from multiple global navigation satellite systems in the target area and extract the time-delay Doppler frequency shift map from the observation data of the multiple global navigation satellite systems; the training module 52 is used to perform stationarity checks and train a preset neural network model on the time-delay Doppler frequency shift map to obtain a predicted sea clutter time-delay Doppler frequency shift map; the elimination module 53 is used to obtain the target time-delay Doppler frequency shift map to be predicted in the target area, and subtract the predicted sea clutter time-delay Doppler frequency shift map from the target time-delay Doppler frequency shift map to obtain an eliminated sea clutter time-delay Doppler frequency shift map; the unification module 54 is used to filter the initial target position in the eliminated sea clutter time-delay Doppler frequency shift map and convert the initial target position into a spatial position based on a unified coordinate system; the transformation module 55 is used to calculate and transform the spatial position using geometric semantic constraints to obtain the position of the target at sea.

[0081] Figure 6An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6 As shown, the electronic device may include: a processor 610, a communication interface 620, a memory 630, and a communication bus 640, wherein the processor 610, the communication interface 620, and the memory 630 communicate with each other through the communication bus 640. The processor 610 can call logic instructions in the memory 630 to execute a GNSS-R-based maritime target positioning method. This method includes: acquiring observation data from multiple global navigation satellite systems (GNSS) in the target area; extracting time-delay Doppler shift maps from the multiple GNSS observation data; performing stationarity checks and training a preset neural network model on the time-delay Doppler shift maps to obtain a predicted sea clutter time-delay Doppler shift map; acquiring a predicted target time-delay Doppler shift map for the target area; subtracting the predicted sea clutter time-delay Doppler shift map from the predicted target time-delay Doppler shift map to obtain a de-cluttered time-delay Doppler shift map; filtering the initial target position in the de-cluttered time-delay Doppler shift map; converting the initial target position into a spatial position based on a unified coordinate system; and calculating the converted spatial position using geometric semantic constraints to obtain the maritime target position.

[0082] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0083] On the other hand, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program performs the GNSS-R-based maritime target positioning method provided by the above methods. The method includes: acquiring observation data from multiple global navigation satellite systems in a target area; extracting time-delay Doppler frequency shift maps from the observation data of the multiple global navigation satellite systems; performing stationarity checks and training a preset neural network model on the time-delay Doppler frequency shift maps to obtain a predicted sea clutter time-delay Doppler frequency shift map; obtaining a target time-delay Doppler frequency shift map to be predicted in the target area; subtracting the predicted sea clutter time-delay Doppler frequency shift map from the target time-delay Doppler frequency shift map to obtain a sea clutter-free time-delay Doppler frequency shift map; filtering the initial target position in the sea clutter-free time-delay Doppler frequency shift map; converting the initial target position into a spatial position based on a unified coordinate system; and calculating the converted spatial position using geometric semantic constraints to obtain the maritime target position.

[0084] The device embodiments described above are merely illustrative. 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 modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0085] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0086] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for locating maritime targets based on GNSS-R, characterized in that, include: Collect observation data from multiple global navigation satellite systems in the target area, and extract the time delay Doppler frequency shift map from the observation data of the multiple global navigation satellite systems; The stationarity of the time-delay Doppler frequency shift map is tested and a preset neural network model is trained to obtain the predicted sea clutter time-delay Doppler frequency shift map; Obtain the target delay Doppler frequency shift map of the target area to be predicted, and subtract the predicted sea clutter delay Doppler frequency shift map from the target delay Doppler frequency shift map to obtain the sea clutter-free delay Doppler frequency shift map; The initial target position in the sea clutter-eliminating delay Doppler frequency shift map is selected, and the initial target position is converted into a spatial position based on a unified coordinate system; The spatial position is transformed using geometric semantic constraints to obtain the position of the maritime target; Collect observation data from multiple Global Navigation Satellite Systems (GNSS) in the target area, and extract the time-delay Doppler frequency shift map from the observation data of the multiple GNSS systems, including: The system determines which satellites in a global navigation satellite system collect observation data from the multiple global navigation satellite systems, wherein the multiple satellites are three satellites; The observation data of the multiple global navigation satellite systems collected by each satellite are used as a single time delay Doppler shift map sequence in the time delay Doppler shift map.

2. The GNSS-R-based maritime target positioning method according to claim 1, characterized in that, The stationarity of the time-delay Doppler frequency shift map is checked, including: A single time-delay Doppler frequency shift map in a single time-delay Doppler frequency shift map sequence is converted into a one-dimensional row vector, and the one-dimensional row vectors corresponding to multiple time-delay Doppler frequency shift maps in the single time-delay Doppler frequency shift map sequence are combined to form a matrix; Each column vector of the matrix is ​​determined as a power time series of any delay and Doppler shift unit. Autocorrelation is performed on all power time series based on the sequence lag number and the sequence average value. Power time series that are uncorrelated and exhibit tailing in the noise region are extracted as pre-trained sea clutter sequences.

3. The GNSS-R-based maritime target positioning method according to claim 2, characterized in that, Training a preset neural network model on the time delay Doppler frequency shift map yields a predicted sea clutter time delay Doppler frequency shift map, including: The pre-trained sea clutter sequence is input into an improved long short-term memory ST-LSTM network for training to obtain a sea clutter sequence prediction model. The ST-LSTM network includes a four-layer ConvLSTM structure and a channel from top-level information to bottom-level information at adjacent time points. The time-delay Doppler frequency shift map is input into the sea clutter sequence prediction model to obtain the predicted sea clutter time-delay Doppler frequency shift map.

4. The GNSS-R-based maritime target positioning method according to claim 1, characterized in that, The initial target location in the sea clutter-eliminating time-delay Doppler frequency shift map includes: The sea clutter delay Doppler frequency shift map is filtered using a threshold method, and false alarm values ​​are removed from the filtered sea clutter delay Doppler frequency shift map to obtain the initial target position. The initial target position includes the time delay and Doppler shift of each initial target position.

5. The GNSS-R-based maritime target positioning method according to claim 4, characterized in that, Converting the initial target position into a spatial position based on a unified coordinate system includes: The origin of the coordinate system is determined by setting the nadir point of any Doppler shift map in the sea clutter time delay Doppler shift map as the origin of the coordinate system. The offset value between the nadir point of other Doppler shift maps in the sea clutter time delay Doppler shift map and the origin of the coordinate system is calculated. Based on the offset value, the other Doppler shift maps are mapped to the unified coordinate system. By using a rotation matrix to transform the geocentric and geofixed coordinate system data of multiple global navigation satellites, satellite motion state data can be obtained. By combining the unified coordinate system and the satellite motion state data, the initial target position is transformed to obtain the spatial position.

6. The GNSS-R-based maritime target positioning method according to claim 1, characterized in that, The spatial position is transformed using geometric semantic constraints to obtain the position of the maritime target, including: Obtain the satellite speed, altitude, and serial number of multiple global navigation satellites, and obtain the elevation angle of the Global Positioning System satellites; Based on the satellite velocity, satellite altitude, serial number, and GPS satellite elevation angle, the spatial position is calculated using the range-Doppler equation to obtain the initial position of the maritime target. The observation sequence number of the maritime target at different positions, the slope and intercept of the target's motion line are determined. Based on the observation sequence number at different positions, the slope and the intercept, the initial maritime target position is further calculated using the geometric semantic constraints to obtain the maritime target position.

7. The GNSS-R-based maritime target positioning method according to claim 1, characterized in that, After calculating and transforming the spatial position using geometric semantic constraints to obtain the position of the maritime target, the process further includes: The location of the maritime target is converted to latitude and longitude distance to obtain the WGS-84 coordinate system position value of the maritime target.

8. A GNSS-R-based maritime target positioning system, based on the GNSS-R-based maritime target positioning method according to any one of claims 1 to 7, characterized in that, include: The acquisition module is used to acquire observation data from multiple global navigation satellite systems in the target area and extract the time delay Doppler frequency shift map of the observation data from the multiple global navigation satellite systems; The training module is used to perform stationarity checks and train a preset neural network model on the time delay Doppler frequency shift map to obtain the predicted sea clutter time delay Doppler frequency shift map; The elimination module is used to obtain the target delay Doppler frequency shift map of the target area to be predicted, and to subtract the predicted sea clutter delay Doppler frequency shift map from the target delay Doppler frequency shift map to obtain the eliminated sea clutter delay Doppler frequency shift map; A unified module is used to filter the initial target position in the sea clutter-eliminating delay Doppler frequency shift map and convert the initial target position into a spatial position based on a unified coordinate system; The transformation module is used to calculate and transform the spatial position using geometric semantic constraints to obtain the position of the target at sea.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the GNSS-R-based maritime target positioning method as described in any one of claims 1 to 7.

Images