A vector magnetic image matching method, device, equipment and medium

Abstract

The present application relates to the field of geomagnetic adaptive navigation technology, and discloses a vector magnetic map matching method, device, equipment and medium, wherein the method comprises: obtaining an inertial navigation track point set at the current moment; wherein the inertial navigation track point set comprises the inertial navigation track point coordinate position and the vector magnetic field strength corresponding to the inertial navigation track point; determining a candidate track point set corresponding to the inertial navigation track point from a preset vector magnetic map based on the vector magnetic field strength; screening the candidate track points in the candidate track point set, and performing image conversion according to the screening results to obtain a target sequence image; generating a predicted track point at the next moment based on the target sequence image and the inertial navigation track coordinate position. The technical solution provided by the present application can improve the real-time performance and accuracy of the predicted position in an environment dealing with complex inertial navigation errors.

Description

Technical Field

[0001] The present application relates to the field of geomagnetic adaptive navigation technology, and in particular to a vector magnetic map matching method, device, equipment and medium. Background Art

[0002] Geomagnetic navigation technology measures the Earth's magnetic field and compares it with a pre-built geomagnetic reference map to achieve precise positioning without external signals, especially in environments where signals cannot reach, such as underground or in urban canyons. Existing magnetic map matching methods rely on a combination of inertial navigation and geomagnetic information. However, traditional algorithms such as VICCP have problems such as poor real-time performance, local optimal solutions, and sensitivity to magnetic map errors, making it difficult to cope with the challenges of complex inertial navigation errors and dynamic environments. Summary of the invention

[0003] The present application provides a vector magnetic map matching method, device, equipment and medium, which achieve the technical effect of improving the real-time performance and accuracy of the predicted position in an environment dealing with complex inertial navigation errors.

[0004] In order to achieve the above objectives, the main technical solutions adopted in this application include:

[0005] In a first aspect, an embodiment of the present application provides a vector magnetic map matching method, the method comprising:

[0006] Acquire a set of inertial navigation trajectory points at the current moment; wherein the set of inertial navigation trajectory points includes the inertial navigation trajectory coordinate positions of the inertial navigation trajectory points and the vector magnetic field strength corresponding to the inertial navigation trajectory points;

[0007] Determining a set of candidate trajectory points corresponding to the inertial navigation trajectory point from a preset vector magnetic map based on the vector magnetic field strength;

[0008] Screening the candidate trajectory points in the candidate trajectory point set, and performing image conversion according to the screening results to obtain a target sequence image;

[0009] Based on the target sequence images and the inertial navigation trajectory coordinate position, a predicted trajectory point at the next moment is generated.

[0010] A vector magnetic map matching method provided in this embodiment can fully perceive the current position and provide important environmental data for subsequent trajectory prediction by acquiring the inertial navigation trajectory point set at the current moment, including position coordinates and vector magnetic field strength. Based on the current vector magnetic field strength, a set of candidate trajectory points is screened out from the preset magnetic field map, which effectively improves the matching accuracy of the trajectory points, especially in complex environments, where magnetic field information can compensate for the errors of inertial sensors. Combining the target sequence image and the inertial navigation trajectory coordinates, the predicted trajectory points at the next moment are generated, and the multimodal fusion of visual and inertial data is used to improve the accuracy and stability of trajectory prediction. Overall, by fusing inertial navigation, magnetic field data and visual information, not only the robustness of trajectory prediction is enhanced, but also the error accumulation is effectively reduced.

[0011] In one embodiment, determining a set of candidate trajectory points corresponding to the inertial navigation trajectory point from a preset vector magnetic map based on the vector magnetic field strength includes:

[0012] For any inertial navigation track point, determining the difference between the vector magnetic field strength and the magnetic field strength of each pixel point in the vector magnetic map;

[0013] Determine the target pixel point corresponding to the difference less than the preset difference threshold as the candidate trajectory point corresponding to any inertial navigation trajectory point, and obtain a set of candidate trajectory points corresponding to any inertial navigation trajectory point;

[0014] Each of the inertial navigation trajectory points is traversed until the candidate trajectory point set corresponding to all the inertial navigation trajectory points is obtained.

[0015] This embodiment achieves accurate trajectory point matching by calculating the difference in magnetic field strength between the inertial navigation trajectory point and each pixel point in the vector magnetic map. This process can efficiently identify the magnetic field characteristics that match the actual trajectory point, thereby enhancing the perception of the environment. By setting the difference threshold, the candidate trajectory points are effectively screened, the possibility of mismatching is reduced, and the high reliability of the candidate trajectory points is ensured. By traversing all inertial navigation trajectory points and establishing a complete set of candidate trajectory points, sufficient information is provided for subsequent trajectory point matching, thereby improving navigation accuracy and system robustness. In summary, combining vector magnetic field data for trajectory point matching not only improves positioning accuracy and enhances anti-interference ability, but also improves computing efficiency and real-time response capabilities, and is suitable for dynamic and complex navigation environments, such as unmanned driving and aircraft navigation.

[0016] In one embodiment, the screening of candidate trajectory points in the candidate trajectory point set includes:

[0017] For any candidate trajectory point in any candidate trajectory point set, determining the shortest distance between the any candidate trajectory point and a candidate trajectory point in an adjacent candidate trajectory point set;

[0018] determining a minimum value of the shortest distance;

[0019] When the minimum value is greater than a preset interval distance, determining any candidate trajectory point as an outlier trajectory point;

[0020] Each of the candidate trajectory point sets is traversed until outlier trajectory points corresponding to all the candidate trajectory point sets are obtained, and the outlier trajectory points are deleted.

[0021] This embodiment provides a basis for subsequent outlier identification by calculating the shortest distance between a candidate trajectory point and its adjacent trajectory points. The minimum value of the shortest distances of all candidate trajectory points is determined, so as to centrally process the relative positional relationship between the trajectory points. By comparing the minimum distance with the preset interval distance threshold, outliers that are far away from the surrounding trajectory points are automatically identified. All candidate trajectory points are traversed, and trajectory points determined to be outliers are deleted to optimize the data set. The entire process can effectively eliminate abnormal data, improve the accuracy and reliability of trajectory data, reduce interference, and improve the effect of subsequent analysis, while having high automation and flexibility.

[0022] In one embodiment, the screening of candidate trajectory points in the candidate trajectory point set includes:

[0023] For any candidate trajectory point in any candidate trajectory point set, determine the distance between the any candidate trajectory point and the end position of the trajectory point at the previous moment;

[0024] When the distance is greater than a preset distance threshold, determining any candidate trajectory point as a remote location trajectory point;

[0025] Each of the candidate trajectory point sets is traversed until all remote location trajectory points corresponding to the candidate trajectory point sets are obtained, and the remote location trajectory points are deleted.

[0026] This embodiment provides an accurate spatial measurement for subsequent analysis by calculating the distance between the candidate trajectory point and the end position of the trajectory point at the previous moment. By determining whether the distance exceeds the preset distance threshold, possible remote location trajectory points are identified. These points may be outliers caused by data anomalies or excessive trajectory fluctuations. The candidate trajectory point set is traversed and the trajectory points in these remote locations are deleted to purify the data set and remove noise that does not conform to the normal trajectory rules. Through this series of operations, the trajectory data can be automatically cleaned, the consistency and accuracy of the data can be improved, and a more reliable foundation can be laid for subsequent trajectory analysis, path planning and other tasks. At the same time, the processing efficiency and adaptability are improved, and it is suitable for different types of trajectory data scenarios.

[0027] In one embodiment, performing image conversion according to the screening results to obtain a target sequence image includes:

[0028] According to the screening results, a set of screened candidate trajectory points is obtained;

[0029] The selected candidate trajectory point set is subjected to binary image conversion to obtain the target serialized image.

[0030] This embodiment removes noise and errors by screening the original trajectory point data, ensuring that the data processed subsequently is more accurate and reliable. This screening process filters out inconsistent or outlier points by setting conditions, retaining only representative valid trajectory points, thereby improving the quality of the data and reducing redundant information. The screened candidate trajectory point sets are converted into binary images, with each trajectory point corresponding to a pixel in the image. Through this conversion, the trajectory data becomes concise and easy to process, providing a clear input form for subsequent trajectory analysis.

[0031] In one embodiment, the generating of the predicted trajectory point at the next moment based on the target sequence image and the inertial navigation trajectory coordinate position includes:

[0032] Inputting the target sequence image and the inertial navigation trajectory coordinate position into a trajectory point prediction model, wherein the trajectory point prediction model includes an image encoder, a trajectory encoder, a feature fusion module and an output decoder;

[0033] Generate an image feature vector corresponding to the target sequence image by the image encoder;

[0034] Generate a trajectory feature vector corresponding to the inertial navigation trajectory coordinate position by means of the trajectory encoder;

[0035] The image feature vector and the trajectory feature vector are fused by the feature fusion module to obtain a fused feature vector;

[0036] The fused feature vector is input into the output decoder to generate the predicted trajectory point at the next moment.

[0037] This embodiment provides multiple perspectives of visual information and motion trajectory by inputting the target sequence image and inertial navigation trajectory coordinate position into the trajectory point prediction model. The image encoder extracts spatial feature vectors from the image to capture the position information of the target; at the same time, the trajectory encoder processes the inertial navigation trajectory data to generate a trajectory feature vector that reflects the target's motion dynamics. Then, the image and trajectory feature vectors are combined through the feature fusion module to generate a comprehensive feature vector, which further enhances the model's ability to understand the target's spatiotemporal behavior. Finally, the output decoder predicts the target's future trajectory points based on the fused feature vector, thereby achieving accurate trajectory prediction, which can significantly improve the accuracy of path planning and decision-making in complex dynamic environments.

[0038] In one embodiment, the method further comprises:

[0039] Fitting the predicted trajectory points to obtain a target corrected trajectory;

[0040] The end position of the target corrected trajectory is determined as the end position of the trajectory point at the previous moment, so as to be used for screening the candidate trajectory points in the candidate trajectory point set.

[0041] This embodiment generates a target correction trajectory by fitting the predicted trajectory points, thereby eliminating the effects of noise, sensor errors and data fluctuations, making the trajectory smoother and more accurate. This process improves the accuracy and stability of trajectory prediction, avoids the accumulation of errors, and ensures that more reliable motion prediction can be provided in complex environments. The end position of the corrected trajectory is used as a reference for the trajectory point at the previous moment to filter suitable trajectory points in the candidate trajectory point set. This screening mechanism ensures that points that deviate far from the target trajectory can be excluded, and only those candidate trajectory points that are more consistent with the corrected trajectory are retained, thereby improving the accuracy and consistency of subsequent predictions.

[0042] In a second aspect, an embodiment of the present application provides a vector magnetic map matching device, the device comprising:

[0043] An inertial navigation track point acquisition unit, used to acquire an inertial navigation track point set at a current moment; wherein the inertial navigation track point set includes the inertial navigation track coordinate position of the inertial navigation track point and the vector magnetic field strength corresponding to the inertial navigation track point;

[0044] A candidate trajectory point determination unit, configured to determine a set of candidate trajectory points corresponding to the inertial navigation trajectory point from a preset vector magnetic map based on the vector magnetic field strength;

[0045] A trajectory point screening unit, used for screening candidate trajectory points in the candidate trajectory point set, and performing image conversion according to the screening result to obtain a target sequence image;

[0046] The trajectory point prediction unit is used to generate a predicted trajectory point at the next moment based on the target sequence image and the inertial navigation trajectory coordinate position.

[0047] In a third aspect, an embodiment of the present application provides a computer device, including:

[0048] A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the above-mentioned vector magnetic map matching method by executing the computer instructions.

[0049] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which computer instructions are stored, and the computer instructions are used to enable a computer to execute the vector magnetic map matching method described above. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0051] Figure 1 A flowchart of a vector magnetic map matching method provided in an embodiment of the present application;

[0052] Figure 2 A flowchart of step S3 provided in an embodiment of the present application;

[0053] Figure 3 A flowchart for screening candidate trajectory points in a candidate trajectory point set provided in an embodiment of the present application;

[0054] Figure 4 Another flowchart of screening candidate trajectory points in a candidate trajectory point set provided in an embodiment of the present application;

[0055] Figure 5 A flowchart for obtaining a target sequence image provided in an embodiment of the present application;

[0056] Figure 6 A schematic diagram of a group of target serialized images provided in an embodiment of the present application;

[0057] Figure 7A flowchart of step S7 provided in an embodiment of the present application;

[0058] Figure 8 A schematic diagram of a LOSS curve provided in an embodiment of the present application;

[0059] Fig. 9 The flowchart after step S7 provided for the embodiment of the present application;

[0060] Fig.10 A schematic diagram of a target correction trajectory output on a set of test data provided in an embodiment of the present application;

[0061] Fig.11 A block diagram of a vector magnetic image matching device provided in an embodiment of the present application;

[0062] Fig.12 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0063] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0064] Geomagnetic navigation technology uses the earth's inherent magnetic field as a natural resource. It measures the magnetic field characteristics of the carrier's location and compares it with a pre-built geomagnetic reference map to achieve positioning. The advantage of this technology is that it does not rely on external navigation signals and can work normally in environments where external signals cannot be received or are interfered with. It has strong autonomy and concealment. Compared with the inertial navigation system (INS), geomagnetic navigation can effectively correct the accumulated errors of INS, especially in environments such as underground and urban canyons where external navigation signals are difficult to cover. Geomagnetic navigation shows obvious advantages. Through real-time comparison and correction with INS output, geomagnetic navigation can improve positioning accuracy, play an important supplementary and corrective role, and enhance the robustness of the system.

[0065] Existing magnetic map matching navigation methods mainly rely on traditional feature matching and nonlinear optimization techniques. Among them, a common solution is a combined navigation method of high-speed aircraft inertial / geomagnetic matching search area adaptation. This method determines the position of the carrier by comparing the real-time measured geomagnetic field information with the pre-established geomagnetic reference map. Its specific steps include judging whether the aircraft has entered the geomagnetic map area, calculating the forward distance, estimating the inertial positioning error, determining the matching search area and step size, and finally performing the best correction through the minimum distance metric. However, this method has certain limitations. Since it is based on the difference comparison between the geomagnetic sensor measurement sequence and the sequence group of the inertial navigation translation search, the best correction obtained is usually a rigid transformation of the trajectory, which makes this method unable to adapt to more complex inertial navigation error situations. In addition, the magnetic map matching algorithm based on traditional methods, especially VICCP, has a cumbersome iterative process and cannot perform large-scale real-time iterative search, so it can often only find a local optimal solution, which limits the ability of real-time application.

[0066] In order to solve the above technical problems, according to an embodiment of the present application, a vector magnetic map matching method embodiment is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0067] In this embodiment, a vector magnetic map matching method is provided. Figure 1 A flowchart of a vector magnetic map matching method provided in an embodiment of the present application is shown in FIG. Figure 1 As shown, the process includes the following steps:

[0068] Step S1, obtaining a set of inertial navigation trajectory points at a current moment; wherein the set of inertial navigation trajectory points includes the inertial navigation trajectory coordinate positions of the inertial navigation trajectory points and the vector magnetic field strengths corresponding to the inertial navigation trajectory points.

[0069] Specifically, data can be collected using an inertial measurement unit (IMU), and the current position of the track point can be calculated using an inertial navigation algorithm (such as a Kalman filter). These position information usually include longitude, latitude, and altitude. A magnetometer is used to measure the geomagnetic field strength at the current moment to obtain the vector magnetic field strength corresponding to each inertial navigation track point. The vector magnetic field strength usually includes three components: an x-component, a y-component, and a z-component, which correspond to the strength of the geomagnetic field in three spatial axes, respectively. The position information and vector magnetic field strength information of the inertial navigation track point are combined to form a complete set of inertial navigation track points. Each point contains not only the position coordinates, but also the magnetic field strength information.

[0070] Step S3, determining a set of candidate trajectory points corresponding to the inertial navigation trajectory point from a preset vector magnetic map based on the vector magnetic field strength.

[0071] Specifically, for each inertial navigation track point, the measured vector magnetic field strength is compared with the magnetic field strength recorded in the vector magnetic map. The difference between the vector magnetic field strength of the inertial navigation track point and the vector magnetic field strength of each pixel in the vector magnetic map is calculated. This is usually achieved by calculating the infinite norm (i.e., the maximum absolute difference) between the two vectors. A deviation threshold ε is determined, and all vector magnetic map pixels whose magnetic field strength difference is less than this threshold are selected as candidate trajectory points to form a candidate trajectory point set for the inertial navigation track point. For each inertial navigation track point, the above process is repeated to construct the corresponding candidate trajectory point set.

[0072] Step S5, screening the candidate trajectory points in the candidate trajectory point set, and performing image conversion according to the screening result to obtain the target sequence image.

[0073] Specifically, outlier points and remote points are screened for the candidate trajectory points in the candidate trajectory point set, and the screened results are converted into a binary image sequence through imaging.

[0074] Step S7, generating the predicted trajectory point at the next moment based on the target sequence image and the inertial navigation trajectory coordinate position.

[0075] Specifically, the predicted trajectory point of the next moment is generated by the trajectory point prediction model. The trajectory point prediction model includes an image encoder, a trajectory encoder, a feature fusion module and an output decoder. The image encoder converts the target sequence image into an image feature vector. The trajectory encoder converts the inertial navigation trajectory coordinate position into a trajectory feature vector. The feature fusion module fuses the image feature vector and the trajectory feature vector to obtain a fused feature vector. The output decoder converts the fused feature vector into the final trajectory point prediction value.

[0076] A vector magnetic map matching method provided in this embodiment can fully perceive the current position and provide important environmental data for subsequent trajectory prediction by acquiring the inertial navigation trajectory point set at the current moment, including position coordinates and vector magnetic field strength. Based on the current vector magnetic field strength, a set of candidate trajectory points is screened out from the preset magnetic field map, which effectively improves the matching accuracy of the trajectory points, especially in complex environments, where magnetic field information can compensate for the errors of inertial sensors. Combining the target sequence image and the inertial navigation trajectory coordinates, the predicted trajectory points at the next moment are generated, and the multimodal fusion of visual and inertial data is used to improve the accuracy and stability of trajectory prediction. Overall, by fusing inertial navigation, magnetic field data and visual information, not only the robustness of trajectory prediction is enhanced, but also the error accumulation is effectively reduced.

[0077] Figure 2The flowchart of step S3 provided in the embodiment of the present application may include the following steps:

[0078] Step S31, for any inertial navigation track point, determining the difference between the vector magnetic field strength and the magnetic field strength of each pixel point in the vector magnetic map.

[0079] Step S33: determine the target pixel points corresponding to the difference less than the preset difference threshold as candidate trajectory points corresponding to any inertial navigation trajectory point, and obtain a set of candidate trajectory points corresponding to any inertial navigation trajectory point.

[0080] Step S35, traverse each inertial navigation trajectory point until a set of candidate trajectory points corresponding to all inertial navigation trajectory points is obtained.

[0081] Among them, inertial navigation track points refer to track points generated by the inertial navigation system (INS). These points contain information such as position and speed, which are used to determine the path and position of the aircraft. The vector magnetic field strength refers to the magnitude and direction of the magnetic field at a certain point in the magnetic field. It is usually represented by a vector and contains three components: x, y, and z components. The vector magnetic map is similar to a geographic information system (GIS) map, which records the vector magnetic field strength of each point in a certain area. It contains magnetic field information for each point in space, usually a three-dimensional matrix, representing the magnetic field strength at each location in the area.

[0082] Specifically, for each inertial navigation track point generated by the inertial navigation system, the vector magnetic field strength of the point is first extracted. This strength represents the magnetic field characteristics of the position. Then, the vector magnetic field strength of the point is compared with the magnetic field strength of each pixel in the vector magnetic map. Each pixel has a corresponding magnetic field strength value in the vector magnetic map. The difference obtained after comparison is determined by calculating the infinite norm between the two vectors, that is, the maximum absolute difference. If the difference between the magnetic field strength of a pixel point in a vector magnetic map and the vector magnetic field strength of the current inertial navigation track point is less than a preset difference threshold ε, then the pixel point is regarded as a candidate track point. In this way, for each inertial navigation track point, a set of candidate track points can be obtained, and the magnetic field strength of these candidate track points is very close to that of the inertial navigation track point, which may represent the location of the inertial navigation track point. By traversing all inertial navigation track points, the corresponding candidate track point sets are calculated one by one. Finally, a set of candidate track points corresponding to all inertial navigation track points is obtained, and these candidate track points represent the spatial positions that match the magnetic field strength of the inertial navigation track points.

[0083] The specific formula is as follows:

[0084]

[0085] Among them, P iis the set of candidate trajectory points corresponding to the i-th inertial navigation trajectory point, and M(x,y,z) is the magnetic field intensity at the position (x,y,z) in the vector magnetic map. i is the vector magnetic field intensity corresponding to the i-th inertial navigation trajectory point, ε is a preset difference threshold, and preferably, ε is 1 nanosecond.

[0086] This embodiment achieves accurate trajectory point matching by calculating the difference in magnetic field strength between the inertial navigation trajectory point and each pixel point in the vector magnetic map. This process can efficiently identify the magnetic field characteristics that match the actual trajectory point, thereby enhancing the perception of the environment. By setting the difference threshold, the candidate trajectory points are effectively screened, the possibility of mismatching is reduced, and the high reliability of the candidate trajectory points is ensured. By traversing all inertial navigation trajectory points and establishing a complete set of candidate trajectory points, sufficient information is provided for subsequent trajectory point matching, thereby improving navigation accuracy and system robustness. In summary, combining vector magnetic field data for trajectory point matching not only improves positioning accuracy and enhances anti-interference ability, but also improves computing efficiency and real-time response capabilities, and is suitable for dynamic and complex navigation environments, such as unmanned driving and aircraft navigation.

[0087] Figure 3 A flowchart of screening candidate trajectory points in a candidate trajectory point set provided in an embodiment of the present application may include the following steps:

[0088] Step S511 : for any candidate trajectory point in any candidate trajectory point set, determine the shortest distance between any candidate trajectory point and a candidate trajectory point in an adjacent candidate trajectory point set.

[0089] Specifically, for any candidate trajectory point in any candidate trajectory point set, calculate its relationship with the adjacent candidate trajectory point set (the previous candidate trajectory point set P i-1 And the next candidate trajectory point set P i+1 ). In other words, assuming that the set of candidate trajectory points is P i ={p1,p2,…,p n}, where p i is the set P i A candidate trajectory point in . Each candidate trajectory point p i The previous candidate trajectory point set P i-1 And the next candidate trajectory point set P i+1 is a set of adjacent trajectory points. Calculate the current candidate trajectory point p i Its adjacent candidate trajectory point set (i.e., P i-1 and P i+1 The shortest distance between all candidate trajectory points in . This shortest distance is d(p i ,P i+1) and d(p i ,P i-1 ), indicating the minimum value of p i The degree of proximity to the previous and next neighboring points. By calculating the shortest distance, we can preliminarily determine the degree of isolation of the candidate trajectory point. If the distance between a candidate trajectory point and its neighboring points is too large, it may be an outlier.

[0090] Step S513, determining the minimum value of the shortest distance.

[0091] Specifically, for each candidate trajectory point p i , take the minimum distance between it and the adjacent points as the representative distance of the point. This minimum value is used to measure the relative proximity between the candidate trajectory point and the surrounding points. If this minimum value is very large, it means that the candidate trajectory point may not belong to the main path of the original trajectory and may be an outlier.

[0092] Step S515: when the minimum value is greater than the preset interval distance, any candidate trajectory point is determined as an outlier trajectory point.

[0093] Specifically, if the minimum value is greater than the preset interval distance, it is considered that p i The distance between its neighboring trajectory points is far away from the normal trajectory interval range, so it is determined that p i is the outlier trajectory point.

[0094] Step S517, traverse each candidate trajectory point set until outlier trajectory points corresponding to all candidate trajectory point sets are obtained, and the outlier trajectory points are deleted.

[0095] Specifically, traverse each candidate trajectory point set, identify and delete all outlier trajectory points. The specific formula is as follows:

[0096]

[0097] Where d(p,S) is the shortest distance from element p to the elements in set S; p i is a candidate trajectory point; i+1 Relative to p i The adjacent previous candidate trajectory point set; P i-1 Relative to p i The next set of adjacent candidate trajectory points; δ represents the expected interval between adjacent candidate trajectory points; k is a multiplication factor, which can be adjusted according to actual needs to affect the strictness of outlier determination; kδ is a preset interval distance, preferably, k=3, δ=5.

[0098] This embodiment provides a basis for subsequent outlier identification by calculating the shortest distance between a candidate trajectory point and its adjacent trajectory points. The minimum value of the shortest distances of all candidate trajectory points is determined, so as to centrally process the relative positional relationship between the trajectory points. By comparing the minimum distance with the preset interval distance threshold, outliers that are far away from the surrounding trajectory points are automatically identified. All candidate trajectory points are traversed, and trajectory points determined to be outliers are deleted to optimize the data set. The entire process can effectively eliminate abnormal data, improve the accuracy and reliability of trajectory data, reduce interference, and improve the effect of subsequent analysis, while having high automation and flexibility.

[0099] Figure 4 Another flowchart for screening candidate trajectory points in a candidate trajectory point set provided in an embodiment of the present application may include the following steps:

[0100] Step S531: for any candidate trajectory point in any candidate trajectory point set, determine the distance between any candidate trajectory point and the end position of the trajectory point at the previous moment.

[0101] Specifically, for each candidate trajectory point in the candidate trajectory point set, the Euclidean distance between it and the end position of the trajectory point at the previous moment is calculated. The Euclidean distance measures the spatial proximity between the candidate trajectory point and the trajectory point at the previous moment. Assume that the candidate trajectory point set is P i ={p1,p2,…,p n}, where each candidate trajectory point p i is a candidate trajectory point in the data, and the end position of the trajectory point at the previous moment is Q last , for each candidate trajectory point p i , calculate its position Q with the end of the trajectory point at the previous moment last The Euclidean distance between: ||Q last -p i ||2.

[0102] Step S533: when the distance is greater than a preset distance threshold, any candidate trajectory point is determined as a remote location trajectory point.

[0103] Specifically, for each candidate trajectory point p i , if it is the same as the end position Q of the trajectory point at the previous moment last The Euclidean distance between them exceeds the preset distance threshold d th , then the candidate trajectory point is marked as a remote location trajectory point, that is, ||Q last -p i ||2>d th By setting the distance threshold d th , d thA displacement of, for example, 60 can effectively identify trajectory points that are far away from the previous trajectory point. These trajectory points may be outliers due to abnormal reasons (such as measurement errors, data loss, equipment failure, etc.), or intentional deviations from the trajectory.

[0104] Step S535 , traverse each candidate trajectory point set until all remote location trajectory points corresponding to the candidate trajectory point sets are obtained, and the remote location trajectory points are deleted.

[0105] Specifically, for each candidate trajectory point set P, traverse all candidate trajectory points p i , check whether the condition ||Q is satisfied last -p i ||2>d th , if the condition is met, the remote location trajectory point is deleted.

[0106] This embodiment provides an accurate spatial measurement for subsequent analysis by calculating the distance between the candidate trajectory point and the end position of the trajectory point at the previous moment. By determining whether the distance exceeds the preset distance threshold, possible remote location trajectory points are identified. These points may be outliers caused by data anomalies or excessive trajectory fluctuations. The candidate trajectory point set is traversed and the trajectory points in these remote locations are deleted to purify the data set and remove noise that does not conform to the normal trajectory rules. Through this series of operations, the trajectory data can be automatically cleaned, the consistency and accuracy of the data can be improved, and a more reliable foundation can be laid for subsequent trajectory analysis, path planning and other tasks. At the same time, the processing efficiency and adaptability are improved, and it is suitable for different types of trajectory data scenarios.

[0107] Figure 5 The flowchart for obtaining the target sequence image provided in the embodiment of the present application may include the following steps:

[0108] Step S551, obtaining a filtered candidate trajectory point set according to the filtering result.

[0109] Step S553, converting the screened candidate trajectory point set into a binary image to obtain a target serialized image.

[0110] Specifically, according to the screening results obtained after the screening step, a set of screened candidate trajectory points corresponding to each inertial navigation trajectory point is obtained. The screened candidate trajectory point set is converted into a binary image, in which the value of each pixel indicates whether the point belongs to the candidate trajectory point set. That is, each screened candidate trajectory point set P i Convert to the corresponding binary image G i , if the pixel (x, y) belongs to P i , then G i The value of (x,y) is 1, otherwise it is 0. See Figure 6 , Figure 6 A schematic diagram of a set of target serialized images provided in an embodiment of the present application.

[0111]

[0112] This embodiment removes noise and errors by screening the original trajectory point data, ensuring that the data processed subsequently is more accurate and reliable. This screening process filters out inconsistent or outlier points by setting conditions, retaining only representative valid trajectory points, thereby improving the quality of the data and reducing redundant information. The screened candidate trajectory point sets are converted into binary images, with each trajectory point corresponding to a pixel in the image. Through this conversion, the trajectory data becomes concise and easy to process, providing a clear input form for subsequent trajectory analysis.

[0113] Figure 7 The flowchart of step S7 provided in the embodiment of the present application may include the following steps:

[0114] Step S71, inputting the target sequence image and the inertial navigation trajectory coordinate position into a trajectory point prediction model, wherein the trajectory point prediction model includes an image encoder, a trajectory encoder, a feature fusion module and an output decoder.

[0115] Specifically, the trajectory point prediction model is a model based on Vision Transformer (ViT), in which the image encoder is M ViT Image encoder, which is an image feature extraction network based on Vision Transformer. Its input is an image sequence (image data input sequentially in time), and its output is the feature vector of these images. ViT Through the self-attention mechanism, the spatiotemporal information within the image is captured, and high-dimensional features are extracted for subsequent prediction tasks.

[0116] Track encoder is M Trj The trajectory encoder is used to process trajectory data. The input is the original inertial navigation trajectory coordinate position, which is a time series containing position points. The trajectory encoder maps the inertial navigation trajectory coordinate position to a high-dimensional feature representation through a fully connected layer, allowing the model to understand and learn trajectory patterns.

[0117] The feature fusion module is M fusion The feature fusion module is used to fuse the image feature vector and the trajectory feature vector. These two features are concatenated together and further fused through a fully connected layer (usually a multi-layer perceptron) to output a comprehensive feature representation.

[0118] The output decoder is M outputThe output decoder is used to pass the fused feature vector through the output decoder to generate the predicted trajectory point at the next moment. The decoder maps the high-dimensional feature representation back to the two-dimensional coordinate space and outputs the predicted value of the trajectory point.

[0119] Step S73, generating an image feature vector corresponding to the target sequence image through an image encoder.

[0120] Specifically, the image encoder M ViT The filtered target sequence image G i Mapped to the corresponding image feature vector f i :

[0121]

[0122]

[0123] Step S75, generating a trajectory feature vector corresponding to the inertial navigation trajectory coordinate position through a trajectory encoder.

[0124] Specifically, the track encoder M Trj The original inertial navigation trajectory coordinate position L i Mapped to trajectory feature vector F trajectory :

[0125]

[0126] Step S77, fusing the image feature vector and the trajectory feature vector through a feature fusion module to obtain a fused feature vector.

[0127] Specifically, M fusion The feature fusion module concatenates and fuses the image feature vector and the trajectory feature vector to form a comprehensive feature.

[0128] Step S79, input the fused feature vector to the output decoder to generate the predicted trajectory point at the next moment.

[0129] Specifically, the fused feature vector is obtained by M output The output decoder generates the predicted trajectory point R output :

[0130]

[0131] The predicted trajectory point shape is (S batch ,num points ×2), which means that for each sample in the batch, there is a length of num points ×2 one-dimensional array. For example, if num pointsis 8 (indicating that there are 8 trajectory points), then each sample will have an array of length 16, indicating that the x and y coordinates of the 8 points are arranged in sequence. Afterwards, the predicted trajectory points are reshaped into (S batch ,num points ,2) shape, that is, each length of num points ×2 array is split into num points An array of length 2 is generated, each array of length 2 represents the (x, y) coordinates of a trajectory point, and the final predicted trajectory point is obtained.

[0132] For example, input to M ViT The image encoder’s time series image data has the shape of (S batch ,seq len ,c,h,w). Among them, seq len =8 indicates the length of the time series, c=3 indicates that each frame has three channels (RGB), and h=w=224 indicates that the image resolution is 224×224 pixels. Trj The inertial navigation trajectory coordinate position data is in the form of (S batch ,num points ×2). Here num points Represents the number of trajectory points, each of which consists of two coordinates (x, y). ViT , the image data G at each time step i Convert to image feature vector f i By M Trj , convert the inertial navigation trajectory coordinate position data into trajectory feature vector F trajectory , the output shape is (S batch ,512). The image feature vector F image and trajectory feature vector F trajectory Splice to form a comprehensive feature vector. fusion The fusion fully connected layer maps the dimension of the comprehensive feature vector to 1024, further fusing the image and trajectory information. The fused feature vector is input to the output decoder M output , which is converted into the final trajectory point prediction value through the fully connected layer. The initial output shape is (S batch ,num points ×2). The predicted trajectory points are reshaped into (S batch ,num points ,2), represents the two-dimensional coordinates (x,y) of each trajectory point in each batch.

[0133] In addition, the data loading amount S for trajectory point prediction model training batch Set to 64, the trajectory length num pointsSet to 8, using 20,000 sets of data for simulation training. For the LOSS curve during training, see Figure 8 , Figure 8 A schematic diagram of the LOSS curve provided for the embodiment of the present application, wherein Epoch means that the entire training data set is completely passed through the model once. It can be understood that the model learns all training samples once. During the training process, the model continuously optimizes parameters through multiple epochs to reduce errors and improve the accuracy of the model.

[0134] This embodiment provides multiple perspectives of visual information and motion trajectory by inputting the target sequence image and inertial navigation trajectory coordinate position into the trajectory point prediction model. The image encoder extracts spatial feature vectors from the image to capture the target position information; at the same time, the trajectory encoder processes the inertial navigation trajectory data to generate a trajectory feature vector that reflects the target motion dynamics. Then, the image and trajectory feature vectors are combined through the feature fusion module to generate a comprehensive feature vector, which further enhances the model's ability to understand the target's spatiotemporal behavior. Finally, the output decoder predicts the target's future trajectory points based on the fused feature vector, thereby achieving accurate trajectory prediction, which can significantly improve the accuracy of path planning and decision-making in complex dynamic environments.

[0135] Fig. 9 The flowchart after step S7 provided in the embodiment of the present application may include the following steps:

[0136] Step S91, fitting the predicted trajectory points to obtain the target corrected trajectory.

[0137] Step S93: determining the end position of the target corrected trajectory as the end position of the trajectory point at the previous moment, so as to be used for screening the candidate trajectory points in the candidate trajectory point set.

[0138] Specifically, the predicted trajectory points can be fitted by methods such as polynomial fitting and spline fitting to obtain a smoother and more accurate target correction trajectory. Fig.10 , Fig.10 A schematic diagram of a target correction trajectory output on a set of test data provided in an embodiment of the present application.

[0139] The end position of the target corrected trajectory is used to filter the candidate trajectory points in the candidate trajectory point set to improve the accuracy of subsequent trajectory prediction. That is, the end position of the target corrected trajectory is used as the end position of the trajectory point at the previous moment in step S513 to filter the remote location trajectory points in the candidate trajectory points. By using the end position of the target corrected trajectory as a standard, it is possible to effectively filter out those candidate trajectory points that are closer to the true trajectory and exclude those candidate trajectory points that are far away from the true trajectory.

[0140] This embodiment generates a target correction trajectory by fitting the predicted trajectory points, thereby eliminating the effects of noise, sensor errors and data fluctuations, making the trajectory smoother and more accurate. This process improves the accuracy and stability of trajectory prediction, avoids the accumulation of errors, and ensures that more reliable motion prediction can be provided in complex environments. The end position of the corrected trajectory is used as a reference for the trajectory point at the previous moment to filter suitable trajectory points in the candidate trajectory point set. This screening mechanism ensures that points that deviate far from the target trajectory can be excluded, and only those candidate trajectory points that are more consistent with the corrected trajectory are retained, thereby improving the accuracy and consistency of subsequent predictions.

[0141] Accordingly, please refer to Fig.11 A block diagram of a vector magnetic image matching device provided in an embodiment of the present application, the device comprising:

[0142] The inertial navigation track point acquisition unit 101 is used to acquire the inertial navigation track point set at the current moment; wherein the inertial navigation track point set includes the inertial navigation track coordinate position of the inertial navigation track point and the vector magnetic field strength corresponding to the inertial navigation track point;

[0143] A candidate trajectory point determination unit 103, configured to determine a set of candidate trajectory points corresponding to an inertial navigation trajectory point from a preset vector magnetic map based on the vector magnetic field strength;

[0144] A trajectory point screening unit 105 is used to screen candidate trajectory points in the candidate trajectory point set, and perform image conversion according to the screening result to obtain a target sequence image;

[0145] The trajectory point prediction unit 107 is used to generate the predicted trajectory point at the next moment based on the target sequence image and the inertial navigation trajectory coordinate position.

[0146] In some optional implementations, the candidate trajectory point determination unit 103 includes:

[0147] For any inertial navigation track point, determine the difference between the vector magnetic field intensity and the magnetic field intensity of each pixel point in the vector magnetic map;

[0148] The target pixel points corresponding to the difference less than the preset difference threshold are determined as candidate trajectory points corresponding to any inertial navigation trajectory point, and a set of candidate trajectory points corresponding to any inertial navigation trajectory point is obtained;

[0149] Traverse each inertial navigation trajectory point until a set of candidate trajectory points corresponding to all inertial navigation trajectory points is obtained.

[0150] In some optional implementations, the trajectory point screening unit 105 includes:

[0151] For any candidate trajectory point in any candidate trajectory point set, determine the shortest distance between any candidate trajectory point and a candidate trajectory point in an adjacent candidate trajectory point set;

[0152] Determine the minimum value of the shortest distance;

[0153] When the minimum value is greater than the preset interval distance, any candidate trajectory point is determined as an outlier trajectory point;

[0154] Traverse each candidate trajectory point set until the outlier trajectory points corresponding to all candidate trajectory point sets are obtained, and then delete the outlier trajectory points.

[0155] In some optional implementations, the trajectory point screening unit 105 includes:

[0156] For any candidate trajectory point in any candidate trajectory point set, determine the distance between any candidate trajectory point and the end position of the trajectory point at the previous moment;

[0157] When the distance is greater than a preset distance threshold, any candidate trajectory point is determined as a remote location trajectory point;

[0158] Traverse each candidate trajectory point set until all the remote location trajectory points corresponding to the candidate trajectory point sets are obtained, and delete the remote location trajectory points.

[0159] In some optional implementations, the trajectory point screening unit 105 includes:

[0160] According to the screening results, a set of screened candidate trajectory points is obtained;

[0161] The filtered candidate trajectory point set is converted into a binary image to obtain a target serialized image.

[0162] In some optional implementations, the trajectory point prediction unit 107 includes:

[0163] The target sequence image and the inertial navigation trajectory coordinate position are input into the trajectory point prediction model, which includes an image encoder, a trajectory encoder, a feature fusion module and an output decoder;

[0164] Generate an image feature vector corresponding to the target sequence image through an image encoder;

[0165] Generate a trajectory feature vector corresponding to the inertial navigation trajectory coordinate position through a trajectory encoder;

[0166] The image feature vector and the trajectory feature vector are fused through a feature fusion module to obtain a fused feature vector;

[0167] The fused feature vector is input into the output decoder to generate the predicted trajectory point at the next moment.

[0168] In some optional embodiments, the device further comprises:

[0169] Fit the predicted trajectory points to obtain the target corrected trajectory;

[0170] The end position of the target corrected trajectory is determined as the end position of the trajectory point at the previous moment, so as to be used for screening the candidate trajectory points in the candidate trajectory point set.

[0171] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0172] In this embodiment, a vector magnetic image matching device is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0173] See also Fig.12 , Fig.12 A schematic diagram of the structure of a computer device provided in an embodiment of the present application is shown in FIG. Fig.12 As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Fig.12 A processor 10 is taken as an example.

[0174] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0175] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0176] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0177] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0178] The computer device further comprises a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0179] The embodiment of the present application also provides a computer-readable storage medium. The above method according to the embodiment of the present application can be implemented in hardware, firmware, or implemented as a computer code that can be recorded in a storage medium, or is implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0180] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0181] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing the present application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0182] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0183] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0184] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0185] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0186] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0187] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0188] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

[0189] Although the embodiments of the present application have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present application, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A vector magnetic map matching method, characterized in that: The method comprises: Acquire a set of inertial navigation trajectory points at the current moment; wherein the set of inertial navigation trajectory points includes the inertial navigation trajectory coordinate positions of the inertial navigation trajectory points and the vector magnetic field strength corresponding to the inertial navigation trajectory points; Determining a set of candidate trajectory points corresponding to the inertial navigation trajectory point from a preset vector magnetic map based on the vector magnetic field strength; Wherein, determining a set of candidate trajectory points corresponding to the inertial navigation trajectory point from a preset vector magnetic map based on the vector magnetic field strength includes: For any inertial navigation track point, determining the difference between the vector magnetic field strength and the magnetic field strength of each pixel point in the vector magnetic map; Determine the target pixel point corresponding to the difference less than the preset difference threshold as the candidate trajectory point corresponding to any inertial navigation trajectory point, and obtain a set of candidate trajectory points corresponding to any inertial navigation trajectory point; Traversing each of the inertial navigation trajectory points until the candidate trajectory point set corresponding to all the inertial navigation trajectory points is obtained; Screening the candidate trajectory points in the candidate trajectory point set, and performing image conversion according to the screening results to obtain a target sequence image; Based on the target sequence images and the inertial navigation trajectory coordinate position, a predicted trajectory point at the next moment is generated.

2. The method according to claim 1, characterized in that The screening of the candidate trajectory points in the candidate trajectory point set includes: For any candidate trajectory point in any candidate trajectory point set, determining the shortest distance between the any candidate trajectory point and a candidate trajectory point in an adjacent candidate trajectory point set; determining a minimum value of the shortest distance; When the minimum value is greater than a preset interval distance, determining any candidate trajectory point as an outlier trajectory point; Each of the candidate trajectory point sets is traversed until outlier trajectory points corresponding to all the candidate trajectory point sets are obtained, and the outlier trajectory points are deleted.

3. The method according to claim 1, characterized in that The screening of the candidate trajectory points in the candidate trajectory point set includes: For any candidate trajectory point in any candidate trajectory point set, determine the distance between the any candidate trajectory point and the end position of the trajectory point at the previous moment; When the distance is greater than a preset distance threshold, determining any candidate trajectory point as a remote location trajectory point; Each of the candidate trajectory point sets is traversed until all remote location trajectory points corresponding to the candidate trajectory point sets are obtained, and the remote location trajectory points are deleted.

4. The method according to claim 1, characterized in that: The image conversion is performed according to the screening results to obtain the target sequence image, including: According to the screening results, a set of screened candidate trajectory points is obtained; The selected candidate trajectory point set is subjected to binary image conversion to obtain the target serialized image.

5. The method according to claim 1, characterized in that The generating of the predicted trajectory point at the next moment based on the target sequence image and the inertial navigation trajectory coordinate position comprises: Inputting the target sequence image and the inertial navigation trajectory coordinate position into a trajectory point prediction model, wherein the trajectory point prediction model includes an image encoder, a trajectory encoder, a feature fusion module and an output decoder; Generate an image feature vector corresponding to the target sequence image by the image encoder; Generate a trajectory feature vector corresponding to the inertial navigation trajectory coordinate position by means of the trajectory encoder; The image feature vector and the trajectory feature vector are fused by the feature fusion module to obtain a fused feature vector; The fused feature vector is input into the output decoder to generate the predicted trajectory point at the next moment.

6. The method according to claim 1, characterized in that The method further comprises: Fitting the predicted trajectory points to obtain a target corrected trajectory; The end position of the target corrected trajectory is determined as the end position of the trajectory point at the previous moment, so as to be used for screening the candidate trajectory points in the candidate trajectory point set.

7. A vector magnetic image matching device, characterized in that: The device comprises: An inertial navigation track point acquisition unit, used to acquire an inertial navigation track point set at a current moment; wherein the inertial navigation track point set includes the inertial navigation track coordinate position of the inertial navigation track point and the vector magnetic field strength corresponding to the inertial navigation track point; A candidate trajectory point determination unit, configured to determine a set of candidate trajectory points corresponding to the inertial navigation trajectory point from a preset vector magnetic map based on the vector magnetic field strength; Wherein, determining a set of candidate trajectory points corresponding to the inertial navigation trajectory point from a preset vector magnetic map based on the vector magnetic field strength includes: For any inertial navigation track point, determining the difference between the vector magnetic field strength and the magnetic field strength of each pixel point in the vector magnetic map; Determine the target pixel point corresponding to the difference less than the preset difference threshold as the candidate trajectory point corresponding to any inertial navigation trajectory point, and obtain a set of candidate trajectory points corresponding to any inertial navigation trajectory point; Traversing each of the inertial navigation trajectory points until the candidate trajectory point set corresponding to all the inertial navigation trajectory points is obtained; A trajectory point screening unit, used for screening candidate trajectory points in the candidate trajectory point set, and performing image conversion according to the screening result to obtain a target sequence image; The trajectory point prediction unit is used to generate a predicted trajectory point at the next moment based on the target sequence image and the inertial navigation trajectory coordinate position.

8. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the vector magnetic map matching method according to any one of claims 1 to 6 by executing the computer instructions.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the vector magnetic map matching method according to any one of claims 1 to 6.

Images