Aeromagnetic compensation method based on support vector machine regression algorithm

Abstract

The invention discloses an aeromagnetic compensation method based on a support vector machine regression algorithm, and relates to the technical field of geophysical exploration, and the method comprises the steps: collecting original magnetic method data and aircraft attitude data in a flight process through employing an aeromagnetic method measurement instrument; constructing a feature matrix based on the preprocessed aircraft attitude data, taking the deviation between the preprocessed original magnetic method data and the reference geomagnetic field reference value as a target value, and training and generating a support vector machine regression model in combination with the feature matrix and the target value; and predicting aircraft attitude data acquired in real time by using the support vector machine regression model to obtain a magnetic method error compensation value, and compensating the original magnetic method data by using the magnetic method error compensation value to obtain compensated magnetic method data. According to the method, the multi-collinearity problem can be effectively solved through the support vector machine regression model, so that the precision and robustness of data compensation of the aeromagnetic method are improved, and the method is suitable for high-precision geophysical exploration tasks.

Description

Technical Field

[0001] The invention relates to the technical field of geophysical exploration, and in particular to an aeromagnetic compensation method based on a support vector machine regression algorithm. Background Art

[0002] Aeromagnetic Survey is a geophysical exploration technology that uses a magnetometer mounted on an airplane or helicopter to measure changes in the magnetic field on the earth's surface and underground to infer underground geological structures, mineral resource distribution or geological anomaly areas. This aeromagnetic exploration technology mainly uses the magnetic difference between the earth's magnetic field and underground rocks and minerals to conduct non-invasive geological surveys over large areas quickly and efficiently.

[0003] In aeromagnetic exploration, traditional aeromagnetic data compensation methods mainly rely on the Tolles-Lawson compensation model, which uses linear regression methods to estimate errors based on parameters such as the flight attitude of the aircraft and the ground magnetic field. However, in practical applications, due to the small amplitude of flight attitude changes, the coefficient matrix in the compensation model often has serious multicollinearity problems, resulting in unsatisfactory compensation effects. Multicollinearity makes it difficult for the regression model to accurately estimate the impact of each parameter, which in turn affects the compensation accuracy of magnetic data.

[0004] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0005] In view of the problems in the related art, the present invention proposes an aeromagnetic compensation method based on a support vector machine regression algorithm to overcome the above technical problems existing in the existing related art.

[0006] To this end, the specific technical solution adopted by the present invention is as follows: According to one aspect of the present invention, an aeromagnetic compensation method based on a support vector machine regression algorithm is provided, the aeromagnetic compensation method comprising: Use aeromagnetic measuring instruments to collect raw magnetic data and aircraft attitude data during flight, and pre-process the raw magnetic data and aircraft attitude data; Based on the preprocessed aircraft attitude data, a feature matrix is ​​constructed, and the deviation between the preprocessed raw magnetic data and the reference value of the benchmark geomagnetic field is used as the target value. The support vector machine regression model is generated by combining the feature matrix and the target value training; The support vector machine regression model is used to predict the aircraft attitude data collected in real time to obtain the magnetic error compensation value, and the magnetic error compensation value is used to compensate the original magnetic data to obtain the compensated magnetic data.

[0007] Furthermore, the aircraft attitude data includes attitude parameters and environmental parameters of the aircraft; Among them, attitude parameters include the pitch angle, roll angle and yaw angle of the aircraft; environmental parameters include the flight altitude, speed and reference value of the reference geomagnetic field of the aircraft.

[0008] Furthermore, the preprocessing of the original magnetic data and the aircraft attitude data includes: The high-frequency noise and instantaneous interference in the original magnetic data are removed in sequence by using filters to obtain the pre-processed original magnetic data; The aircraft attitude data is normalized to the same scale to obtain the preprocessed aircraft attitude data.

[0009] Furthermore, a feature matrix is ​​constructed based on the preprocessed aircraft attitude data, and the deviation between the preprocessed original magnetic data and the reference value of the benchmark geomagnetic field is used as the target value. The support vector machine regression model generated by combining the feature matrix and the target value training includes: A feature matrix is ​​constructed based on the preprocessed attitude parameters and environmental parameters of the aircraft, and each row in the feature matrix represents a multi-dimensional attitude and environmental feature; The deviation between the original magnetic data and the reference value of the benchmark geomagnetic field is taken as the target value, and the support vector machine regression model is trained and generated using historical sampling data.

[0010] Furthermore, the support vector machine regression model generated by training the historical sampling data includes: Building a training set based on historical sampling data, where the historical sampling data includes historical attitude parameters and historical magnetic method errors; The support vector machine regression model is trained using the training set, the radial basis function is selected as the kernel function, and the support vector machine regression model parameters are optimized using cross-validation technology.

[0011] Furthermore, the support vector machine regression model parameters include penalty coefficients and kernel function parameters; Among them, the range of penalty coefficient is [1, 1000]; the range of kernel function parameter is [0.01, 1].

[0012] Furthermore, the support vector machine regression model is used to predict the real-time collected aircraft attitude data to obtain the magnetic method error compensation value, and the magnetic method error compensation value is used to compensate the original magnetic method data. The compensated magnetic method data includes: The trained support vector machine regression model is used to predict the attitude parameters and environmental parameters collected in real time to obtain the magnetic method error compensation value; The magnetic method error compensation value is combined with the original magnetic method data to calculate the magnetic method data compensation value, and the original magnetic method data is compensated based on the magnetic method data compensation value to obtain compensated magnetic method data; Verify the compensated magnetic data and evaluate the compensation effect of the magnetic data.

[0013] Furthermore, the calculation formula of the magnetic data compensation value is: ; Where D compensated Indicates the compensated magnetic data; D raw Represents the original magnetic data; y error Represents the magnetic error compensation value predicted by the support vector machine regression model.

[0014] Furthermore, the compensated magnetic data is verified and the compensation effect of the magnetic data is evaluated, including: The mean, standard deviation and deviation range between the original magnetic data before compensation and the magnetic data after compensation are calculated in turn, and the compensation effect of the magnetic data is verified based on the calculation results; A three-dimensional magnetic anomaly map is generated based on the compensated magnetic data, and the data quality of the compensated magnetic data is evaluated from the three-dimensional magnetic anomaly map.

[0015] According to another aspect of the present invention, an aeromagnetic compensation system based on a support vector machine regression algorithm is also provided, the aeromagnetic compensation system comprising: a data acquisition module, a compensation model generation module and a magnetic method data compensation module; A data acquisition module is used to collect raw magnetic data and aircraft attitude data during flight using an aeromagnetic measuring instrument, and to pre-process the raw magnetic data and aircraft attitude data; The compensation model generation module is used to construct a feature matrix based on the preprocessed aircraft attitude data, and use the deviation between the preprocessed original magnetic data and the reference value of the benchmark geomagnetic field as the target value, and combine the feature matrix and the target value to train and generate a support vector machine regression model; The magnetic data compensation module is used to predict the aircraft attitude data collected in real time by using the support vector machine regression model to obtain the magnetic error compensation value, and compensate the original magnetic data with the magnetic error compensation value to obtain the compensated magnetic data.

[0016] The beneficial effects of the present invention are: 1. The present invention can compensate for the deviation of magnetic anomaly data caused by factors such as flight attitude changes that may occur during the flight, thereby improving the accuracy and reliability of magnetic data, and further providing more accurate geophysical data support for applications such as geological exploration, mineral resource investigation and environmental monitoring. The present invention does not rely on complex physical modeling, but automatically identifies and corrects error patterns through the learning of historical data, effectively improving the compensation effect, especially in the case of small flight attitude changes or complex environments, which can significantly improve the accuracy of magnetic data.

[0017] 2. The present invention can effectively deal with the problem of multicollinearity by introducing a support vector machine regression model, thereby improving the accuracy and robustness of aeromagnetic data compensation, and is suitable for high-precision geophysical exploration tasks. It not only solves the calculation problem caused by multicollinearity in the traditional aeromagnetic compensation model, but also improves the compensation accuracy and robustness of the algorithm through machine learning technology, and is particularly suitable for high-precision aeromagnetic exploration tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 is a flow chart of an aeromagnetic compensation method based on a support vector machine regression algorithm according to an embodiment of the present invention; Figure 2 is a principle block diagram of an aeromagnetic compensation system based on a support vector machine regression algorithm according to an embodiment of the present invention; Figure 3 A three-dimensional magnetic anomaly map of original magnetic data in an aeromagnetic compensation method based on a support vector machine regression algorithm according to an embodiment of the present invention; Figure 4 It is a three-dimensional magnetic anomaly map of magnetic data after compensation by a support vector machine regression model in an aeromagnetic compensation method based on a support vector machine regression algorithm according to an embodiment of the present invention.

[0020] In the figure: 1. Data acquisition module; 2. Compensation model generation module; 3. Magnetic data compensation module. DETAILED DESCRIPTION

[0021] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments. They can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in this field should be able to understand other possible implementation methods and advantages of the present invention.

[0022] According to an embodiment of the present invention, an aeromagnetic compensation method based on a support vector machine regression algorithm is provided.

[0023] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, according to one embodiment of the present invention, an aeromagnetic compensation method based on a support vector machine regression algorithm is provided, and the aeromagnetic compensation method includes: S1. Use an aeromagnetic measuring instrument to collect raw magnetic data and aircraft attitude data during flight, and pre-process the raw magnetic data and aircraft attitude data.

[0024] Among them, the aircraft attitude data includes the aircraft's attitude parameters and environmental parameters that may affect the magnetic data; the attitude parameters include the aircraft's pitch angle, roll angle and yaw angle; the environmental parameters include the aircraft's flight altitude, speed and reference value of the benchmark geomagnetic field.

[0025] Among them, the preprocessing of the original magnetic data and aircraft attitude data includes: The high-frequency noise and instantaneous interference in the original magnetic data are removed in turn by using filters to obtain the preprocessed original magnetic data; the aircraft attitude data are normalized to the same scale to obtain the preprocessed aircraft attitude data.

[0026] It should be noted that the collected raw magnetic data are preprocessed, including denoising, outlier detection and attitude parameter normalization, to ensure the consistency and stability of the input data; denoising includes using filters to remove high-frequency noise and instantaneous interference; normalization includes normalizing the attitude parameters and auxiliary variables to the same scale to improve the stability of model training.

[0027] S2. Construct a feature matrix based on the preprocessed aircraft attitude data, and use the deviation between the preprocessed original magnetic data and the reference value of the benchmark geomagnetic field as the target value. Combine the feature matrix and the target value training to generate a support vector machine regression model.

[0028] Among them, the feature matrix is ​​constructed based on the preprocessed aircraft attitude data, and the deviation between the preprocessed original magnetic data and the reference value of the benchmark geomagnetic field is used as the target value. The support vector machine regression model generated by combining the feature matrix and the target value training includes: A feature matrix is ​​constructed based on the preprocessed attitude parameters and environmental parameters of the aircraft, and each row in the feature matrix represents multi-dimensional attitude and environmental characteristics. The deviation between the original magnetic data and the reference value of the benchmark geomagnetic field is taken as the target value, and the support vector machine regression model is trained and generated using historical sampling data.

[0029] Among them, using historical sampling data to train and generate a support vector machine regression model includes: A training set is constructed based on historical sampling data, and the historical sampling data includes historical attitude parameters and historical magnetic errors. The support vector machine regression model is trained using the training set, and the radial basis function is selected as the kernel function. The support vector machine regression model parameters are optimized through cross-validation technology.

[0030] It should be noted that the aircraft attitude data (pitch angle, roll angle, yaw angle) and other related variables (such as flight altitude and speed) are constructed as a feature matrix X as the input of the support vector machine regression model. Each row of the feature matrix represents the multidimensional attitude and environmental characteristics of a sample. At the same time, the deviation between the original magnetic data and the reference geomagnetic field is taken as the target value y, and a training set (X, y) is constructed based on historical flight data, including historical attitude parameters and corresponding historical observation magnetic errors. The support vector machine regression (SVR) algorithm is used for model training, and the radial basis function (RBF) is used as the kernel function. The model parameters (such as the penalty coefficient C and the kernel function parameter γ) are optimized by the cross-validation method to minimize the model error.

[0031] The support vector machine regression model parameters include the penalty coefficient and kernel function parameters. The penalty coefficient ranges from [1, 1000], and the recommended value is 100. The kernel function parameter ranges from [0.01, 1], and the recommended value is 0.1.

[0032] The sampling frequency of attitude parameters and raw magnetic data is 20 Hz. The training data of the support vector machine regression model is recommended to contain at least 10 hours of flight data to ensure the generalization ability of the model.

[0033] S3. Use the support vector machine regression model to predict the aircraft attitude data collected in real time to obtain a magnetic method error compensation value, and use the magnetic method error compensation value to compensate the original magnetic method data to obtain compensated magnetic method data.

[0034] Among them, the support vector machine regression model is used to predict the real-time collected aircraft attitude data to obtain the magnetic method error compensation value, and the magnetic method error compensation value is used to compensate the original magnetic method data. The compensated magnetic method data includes: The trained support vector machine regression model is used to predict the attitude parameters and environmental parameters collected in real time to obtain the magnetic error compensation value. The magnetic error compensation value is combined with the original magnetic data to calculate the magnetic data compensation value, and the original magnetic data is compensated based on the magnetic data compensation value to obtain the compensated magnetic data.

[0035] The calculation formula of magnetic data compensation value is: ; Where D compensated Indicates the compensated magnetic data; D raw Represents the original magnetic data; y error Represents the magnetic error compensation value predicted by the support vector machine regression model.

[0036] Verify the compensated magnetic data and evaluate the compensation effect of the magnetic data.

[0037] Among them, verifying the compensated magnetic data and evaluating the compensation effect of the magnetic data include: The mean, standard deviation and deviation range between the original magnetic data before compensation and the compensated magnetic data are calculated in turn, and the compensation effect of the magnetic data is verified based on the calculation results; a three-dimensional magnetic anomaly map is generated based on the compensated magnetic data, and the data quality of the compensated magnetic data is evaluated from the three-dimensional magnetic anomaly map.

[0038] Among them, Figure 3-Figure 4 As shown, Figure 3 The original magnetic data collected during the aeromagnetic survey are shown. The original magnetic data contain errors caused by uncorrected flight attitude changes and other interference factors, showing obvious abnormal characteristics and interference patterns. Figure 4 The magnetic data processed by the aeromagnetic compensation method based on the support vector machine regression model proposed in the present invention are demonstrated. Compared with the original magnetic data, the compensated magnetic data effectively removes the errors caused by attitude changes and interference factors, the abnormal features are clearer, and the data quality is significantly improved.

[0039] It should be noted that the trained SVR model is used to predict the attitude parameters collected during real-time flight, generate the corresponding magnetic error compensation value, and apply it to the original magnetic data to obtain the compensated magnetic data. The compensated magnetic data is verified to evaluate its compensation effect, including the error mean, standard deviation, and matching degree with the geological model. If the effect does not meet expectations, it can be further optimized by adjusting the model parameters or expanding the training data set.

[0040] The embodiments of the present invention can be used in the following application scenarios: Mineral resource exploration: Use compensated magnetic data to improve the positioning accuracy of ore bodies.

[0041] Basic geological research: Improve the accuracy of geophysical models and provide high-precision data support for geological research.

[0042] In summary, the present invention effectively overcomes the limitations of traditional compensation methods, and provides an efficient and reliable aeromagnetic compensation method for the field of aeromagnetic exploration by improving compensation accuracy and enhancing the robustness of the algorithm, which is mainly reflected in the following aspects: 1. Solve the problem of multicollinearity: The present invention adopts a support vector machine regression (SVR) algorithm to provide an effective solution to the multicollinearity problem of the coefficient matrix caused by the small change in flight attitude in the traditional Tolles-Lawson model (the Tolles-Lawson model is a classic error compensation model in aeromagnetic exploration, used to correct the magnetic field measurement errors caused by the aircraft attitude, such as roll, pitch and yaw). The nonlinear regression method is introduced to avoid the negative impact of multicollinearity on compensation accuracy and model stability.

[0043] 2. Improve compensation accuracy: Traditional compensation methods rely on linear models and have limited ability to compensate for nonlinear errors. The present invention utilizes the powerful nonlinear modeling capability of the SVR model to significantly improve the accuracy of aeromagnetic data compensation and provide more reliable data support for the precise positioning of geological targets.

[0044] 3. Enhance model robustness: The compensation method of the present invention does not rely on complex physical models or strict assumptions about flight attitude parameters, but realizes automatic identification and correction of error laws through learning from historical data. The method has strong generalization ability and can adapt to different flight conditions and instrument error characteristics.

[0045] 4. Wide adaptability: The present invention can perform adaptive compensation for different types of aircraft and measuring instruments, and can ensure the stability and accuracy of the compensation effect, especially under conditions of complex terrain and large changes in flight trajectory.

[0046] 5. Simplify the compensation model building process: The present invention does not need to rely on the complex modeling of parameters such as the ground magnetic field and flight attitude in traditional methods. High-precision compensation can be achieved only through machine learning algorithms, which significantly reduces the complexity and workload of model construction.

[0047] According to another embodiment of the present invention, Figure 2As shown, an aeromagnetic compensation system based on a support vector machine regression algorithm is also provided, the aeromagnetic compensation system comprising: a data acquisition module 1, a compensation model generation module 2 and a magnetic method data compensation module 3; The data acquisition module 1 is used to collect the original magnetic data and the aircraft attitude data during the flight by using the aeromagnetic measuring instrument, and to pre-process the original magnetic data and the aircraft attitude data; The compensation model generation module 2 is used to construct a feature matrix based on the preprocessed aircraft attitude data, and use the deviation between the preprocessed original magnetic method data and the reference value of the benchmark geomagnetic field as the target value, and combine the feature matrix and the target value to train and generate a support vector machine regression model; The magnetic data compensation module 3 is used to predict the aircraft attitude data collected in real time by using the support vector machine regression model to obtain the magnetic error compensation value, and compensate the original magnetic data with the magnetic error compensation value to obtain the compensated magnetic data.

[0048] In summary, with the help of the above-mentioned technical scheme of the present invention, the present invention can compensate for the deviation of magnetic anomaly data caused by factors such as flight attitude changes that may occur during the flight, thereby improving the accuracy and reliability of magnetic data, and further providing more accurate geophysical data support for applications such as geological exploration, mineral resource investigation and environmental monitoring. The present invention does not rely on complex physical modeling, but automatically identifies and corrects error rules through learning historical data, effectively improving the compensation effect, especially in the case of small flight attitude changes or complex environments, the accuracy of magnetic data can be significantly improved; the present invention can effectively deal with the problem of multicollinearity by introducing a support vector machine regression model, thereby improving the accuracy and robustness of aeromagnetic data compensation, and is suitable for high-precision geophysical exploration tasks, thereby not only solving the calculation problems caused by multicollinearity in traditional aeromagnetic compensation models, but also improving the compensation accuracy and robustness of the algorithm through machine learning technology, and is particularly suitable for high-precision aeromagnetic exploration tasks.

[0049] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An aeromagnetic compensation method based on support vector machine regression algorithm, characterized in that: The aeromagnetic compensation method includes: Use aeromagnetic measuring instruments to collect raw magnetic data and aircraft attitude data during flight, and pre-process the raw magnetic data and aircraft attitude data; Based on the preprocessed aircraft attitude data, a feature matrix is ​​constructed, and the deviation between the preprocessed raw magnetic data and the reference value of the benchmark geomagnetic field is used as the target value. The support vector machine regression model is generated by combining the feature matrix and the target value training; The support vector machine regression model is used to predict the aircraft attitude data collected in real time to obtain the magnetic error compensation value, and the magnetic error compensation value is used to compensate the original magnetic data to obtain the compensated magnetic data.

2. The aeromagnetic compensation method based on support vector machine regression algorithm according to claim 1 is characterized in that: The aircraft attitude data includes attitude parameters and environmental parameters of the aircraft; The attitude parameters include the pitch angle, roll angle and yaw angle of the aircraft; the environmental parameters include the flight altitude, speed and reference value of the reference geomagnetic field of the aircraft.

3. The aeromagnetic compensation method based on support vector machine regression algorithm according to claim 2 is characterized in that: The preprocessing of the original magnetic method data and the aircraft attitude data comprises: The high-frequency noise and instantaneous interference in the original magnetic data are removed in sequence by using filters to obtain the pre-processed original magnetic data; The aircraft attitude data is normalized to the same scale to obtain the preprocessed aircraft attitude data.

4. The aeromagnetic compensation method based on support vector machine regression algorithm according to claim 3 is characterized in that: The feature matrix is ​​constructed based on the pre-processed aircraft attitude data, and the deviation between the pre-processed raw magnetic method data and the reference value of the benchmark geomagnetic field is used as the target value. Combining the feature matrix and target value training to generate a support vector machine regression model includes: A feature matrix is ​​constructed based on the preprocessed attitude parameters and environmental parameters of the aircraft, and each row in the feature matrix represents a multi-dimensional attitude and environmental feature; The deviation between the original magnetic data and the reference value of the benchmark geomagnetic field is taken as the target value, and the support vector machine regression model is trained and generated using historical sampling data.

5. The aeromagnetic compensation method based on support vector machine regression algorithm according to claim 4 is characterized in that: The method of using historical sampling data to train and generate a support vector machine regression model includes: Building a training set based on historical sampling data, where the historical sampling data includes historical attitude parameters and historical magnetic method errors; The support vector machine regression model is trained using the training set, the radial basis function is selected as the kernel function, and the support vector machine regression model parameters are optimized using cross-validation technology.

6. The aeromagnetic compensation method based on support vector machine regression algorithm according to claim 5, characterized in that: The support vector machine regression model parameters include penalty coefficients and kernel function parameters; The range of the penalty coefficient is [1, 1000]; the range of the kernel function parameter is [0.01, 1].

7. The aeromagnetic compensation method based on support vector machine regression algorithm according to claim 6 is characterized in that: The support vector machine regression model is used to predict the real-time collected aircraft attitude data to obtain a magnetic method error compensation value, and the magnetic method error compensation value is used to compensate the original magnetic method data to obtain the compensated magnetic method data including: The trained support vector machine regression model is used to predict the attitude parameters and environmental parameters collected in real time to obtain the magnetic method error compensation value; The magnetic method error compensation value is combined with the original magnetic method data to calculate the magnetic method data compensation value, and the original magnetic method data is compensated based on the magnetic method data compensation value to obtain compensated magnetic method data; Verify the compensated magnetic data and evaluate the compensation effect of the magnetic data.

8. The aeromagnetic compensation method based on support vector machine regression algorithm according to claim 7, characterized in that: The calculation formula of the magnetic data compensation value is: ； Where D compensated Indicates the compensated magnetic data; D raw Represents the original magnetic data; y error Represents the magnetic error compensation value predicted by the support vector machine regression model.

9. The aeromagnetic compensation method based on support vector machine regression algorithm according to claim 8, characterized in that: The verifying of the compensated magnetic data and evaluating the compensation effect of the magnetic data includes: The mean, standard deviation and deviation range between the original magnetic data before compensation and the magnetic data after compensation are calculated in turn, and the compensation effect of the magnetic data is verified based on the calculation results; A three-dimensional magnetic anomaly map is generated based on the compensated magnetic data, and the data quality of the compensated magnetic data is evaluated from the three-dimensional magnetic anomaly map.

10. An aeromagnetic compensation system based on a support vector machine regression algorithm, used to implement the aeromagnetic compensation method based on a support vector machine regression algorithm as described in any one of claims 1 to 9, characterized in that: The aeromagnetic compensation system includes: a data acquisition module, a compensation model generation module and a magnetic method data compensation module; The data acquisition module is used to collect raw magnetic data and aircraft attitude data during flight using an aeromagnetic measuring instrument, and pre-process the raw magnetic data and aircraft attitude data; The compensation model generation module is used to construct a feature matrix based on the preprocessed aircraft attitude data, and use the deviation between the preprocessed original magnetic method data and the reference value of the benchmark geomagnetic field as the target value, and combine the feature matrix and the target value to train and generate a support vector machine regression model; The magnetic data compensation module is used to predict the aircraft attitude data collected in real time using a support vector machine regression model to obtain a magnetic error compensation value, and to compensate the original magnetic data with the magnetic error compensation value to obtain compensated magnetic data.

Images