Method and device for classifying and identifying vehicles with magnetic signal characteristics

Abstract

The present invention provides a method and device for classifying and identifying vehicles with magnetic signal characteristics. The method includes: obtaining magnetic signal data of a first vehicle and preprocessing it; extracting features from the frequency spectrum of the three-component signal of the preprocessed magnetic signal data to obtain a first eigenvector of the magnetic signal data; determining the drive type and vehicle type of the first vehicle and generating corresponding labels; training a first initial network model and a second initial network model based on the first eigenvector and label to form a vehicle drive type detection model and a vehicle type detection model, respectively; processing the magnetic signal data of a second vehicle to be detected to generate a second eigenvector, and using the vehicle drive type detection model and the vehicle drive type detection model to process the second eigenvector, respectively, to achieve classification detection of the second vehicle. The method provided by the present invention can quickly and accurately classify and identify vehicles by vehicle type and vehicle drive type.

Description

Technical Field

[0001] The present invention belongs to the technical field of geomagnetic classification and identification, and particularly relates to a method and device for classifying and identifying vehicles with magnetic signal characteristics. Background Art

[0002] The introduction of intelligent transportation system (ITS) has enabled the current transportation system to comprehensively apply information technology, sensor technology, computer technology, etc. to the traffic management system, which has greatly improved transportation efficiency, alleviated traffic congestion, reduced traffic accidents, reduced energy consumption, and alleviated environmental pollution.

[0003] For effective traffic planning, vehicle detection and classification technologies are crucial. This requires effectively acquiring vehicle classification information in a dynamic traffic environment, a crucial requirement for sensor technology. Current traffic management systems primarily utilize acoustic, magnetic, and imaging sensors to acquire vehicle information. Magnetic sensors are widely used because they are unaffected by natural environmental factors like weather, offer high sensitivity, and are easy to install.

[0004] Existing solutions for collecting vehicle information utilize roadside magnetic sensors to collect large amounts of magnetic signal data. This data includes information such as vehicle type and speed. However, in real-world scenarios, vehicles are categorized in many different ways, such as by size, type, and drive type. Current methods are ineffective at processing magnetic signals and accurately classifying vehicles in a timely manner. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method and device for classifying and identifying vehicles with magnetic signal characteristics, which are used to quickly and accurately classify and identify the vehicle type and vehicle drive type.

[0006] The present invention includes a method for classifying and identifying vehicles with magnetic signal characteristics, comprising:

[0007] obtaining magnetic signal data of a first vehicle;

[0008] Preprocessing the magnetic signal data, wherein the preprocessing includes at least one of the following: magnetic compensation, noise reduction, data enhancement, and data expansion;

[0009] Performing a first feature extraction on the preprocessed magnetic signal data to obtain a frequency spectrum of a three-component signal, and performing a second feature extraction based on the frequency spectrum of the three-component signal to obtain a first eigenvector of the magnetic signal data;

[0010] determining a drive type and a vehicle type of the first vehicle and generating a corresponding label;

[0011] Using the first feature vector as input data and the label as output data to train a first initial network model and a second initial network model to form a vehicle drive type detection model and a vehicle type detection model respectively;

[0012] The magnetic signal data of the second vehicle to be detected is processed to generate a second feature vector, and the second feature vector is processed using the vehicle drive type detection model and the vehicle drive type detection model respectively to achieve classification detection of the drive type and vehicle type of the second vehicle.

[0013] In some embodiments, the magnetic signal data includes multiple segments of magnetic signals, and the preprocessing of the magnetic signal data includes:

[0014] Extracting the first 20 signal detection values of each signal segment in the magnetic signal data, and calculating the mean of all signal detection values as the background magnetic field strength value;

[0015] Performing magnetic compensation on the magnetic signal data based on the background magnetic field strength value to eliminate the background magnetic field of the signal;

[0016] Filtering and eliminating noise and interference signals in the magnetic signal data of the background magnetic field;

[0017] Gaussian white noise is added to each signal segment of the filtered magnetic signal according to a preset signal-to-noise ratio to perform data enhancement, so as to expand the magnetic signal data.

[0018] In some embodiments, the filtering and eliminating noise and interference signals in the magnetic signal data of the background magnetic field includes:

[0019] The high-frequency noise above 40 Hz and the power frequency interference signal of 50 Hz in the magnetic signal data of the background magnetic field are filtered out.

[0020] In some embodiments, the step of adding Gaussian white noise to each segment of the filtered magnetic signal according to a preset signal-to-noise ratio to perform data enhancement includes:

[0021] Gaussian white noise is added to each signal segment of the filtered magnetic signal in sequence according to preset signal-to-noise ratios of 5dB, 10dB, 15dB, 20dB, and 25dB for data enhancement.

[0022] In some embodiments, the magnetic signal data includes multiple segments of magnetic signals, and the three-component signal includes three-component signals of each segment of magnetic signal;

[0023] The performing a second feature extraction based on the frequency spectrum of the three-component signal to obtain a first feature vector of the magnetic signal data includes:

[0024] Extracting a unilateral spectrum from the frequency spectrum of the three-component signal, and intercepting a low-frequency portion of the first n frequency points in the unilateral spectrum to form a spectrum feature;

[0025] extracting wavelet transform energy features from the three-component signals;

[0026] The spectrum features and wavelet transform energy features corresponding to the same segment of magnetic signal are serially fused to obtain a feature vector corresponding to each segment of magnetic signal.

[0027] In some embodiments, the first initial network model and the second initial network model are both composed of a support vector machine model and a convolutional neural network model.

[0028] In some embodiments, the first feature vector is used as input data and the label is used as output data to train a first initial network model to form a vehicle type detection model, including:

[0029] Using the SMOTE algorithm to perform data balancing on the first feature vector and label according to vehicle category;

[0030] A first initial network model is trained based on the processed first feature vector as input data and the label as output data to form a vehicle type detection model.

[0031] In some embodiments, the loss function of the support vector machine model is:

[0032]

[0033] The w and b are the parameters of the support vector machine model, T is the vector transpose symbol, x i is the model variable, y i is the actual value of the input vehicle type and vehicle driving type, are the predicted values of vehicle type and vehicle driving type.

[0034] In some embodiments, the convolutional neural network is a one-dimensional convolutional neural network consisting of two convolutional layers, two pooling layers, and two fully connected layers.

[0035] Another embodiment of the present invention also provides a device for classifying and identifying vehicles with magnetic signal characteristics, comprising:

[0036] an acquisition module, configured to obtain magnetic signal data of a first vehicle;

[0037] A preprocessing module, configured to preprocess the magnetic signal data, wherein the preprocessing includes at least one of the following: magnetic compensation, noise reduction, data enhancement, and data expansion;

[0038] an extraction module, configured to perform a first feature extraction on the preprocessed magnetic signal data to obtain a frequency spectrum of a three-component signal, and perform a second feature extraction based on the frequency spectrum of the three-component signal to obtain a first eigenvector of the magnetic signal data;

[0039] a first determining module, configured to determine a driving type and a vehicle type of the first vehicle and generate a corresponding label;

[0040] a training module, configured to train a first initial network model and a second initial network model using the first feature vector as input data and the label as output data to form a vehicle drive type detection model and a vehicle type detection model, respectively;

[0041] The detection module is used to process the magnetic signal data of the second vehicle to be detected to generate a second feature vector, and use the vehicle drive type detection model and the vehicle drive type detection model to process the second feature vector respectively to realize the classification detection of the drive type and vehicle type of the second vehicle.

[0042] The beneficial effects of the present invention include extracting the frequency domain features and time-frequency domain features of the vehicle's magnetic signal, and using support vector machines and convolutional neural networks as initial models of the vehicle type classification model and the vehicle drive type classification model. By processing the vehicle's magnetic signal to form feature vectors and classification labels for training, models for classifying and identifying the vehicle type and drive type are obtained respectively. Based on the two classification models, the classification and identification of the vehicle in two different classification situations can be realized, thereby achieving high-efficiency and high-accuracy vehicle identification and classification. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0044] Figure 1 Schematic diagram of the flow of a method for classifying and identifying vehicles with magnetic signal characteristics in an embodiment of the present invention.

[0045] Figure 2 2 is a flow chart of a method for classifying and identifying vehicles with magnetic signal characteristics in another embodiment of the present invention.

[0046] Figure 3 2 is a flow chart of a method for classifying and identifying vehicles with magnetic signal characteristics in another embodiment of the present invention.

[0047] Figure 44 is a structural block diagram of a device for classifying and identifying vehicles with magnetic signal characteristics in an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings, but are not intended to limit the present invention.

[0049] It should be understood that various modifications may be made to the embodiments disclosed herein. Therefore, the following description should not be considered as limiting, but merely as an example of the embodiments. Other modifications within the scope and spirit of the present application will occur to those skilled in the art.

[0050] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the present application and, together with the general description of the present application given above and the detailed description of the embodiments given below, serve to explain the principles of the present application.

[0051] It should also be understood that although the present application has been described with reference to certain specific examples, those skilled in the art will be able to implement many other equivalent forms of the present application that have the features described in the claims and are therefore within the scope of protection defined thereby.

[0052] The above and other aspects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings.

[0053] Specific embodiments of the present application will be described hereinafter with reference to the accompanying drawings; however, it should be understood that the embodiments described are merely examples of the present application and that the present application may be implemented in a variety of ways. Familiar and / or repetitive functions and structures are not described in detail to avoid obscuring the present application with unnecessary or redundant details. Therefore, the specific structural and functional details disclosed herein are not intended to be limiting, but rather serve merely as a basis and representative basis for the claims to teach those skilled in the art to variously utilize the present application with substantially any suitable detailed structure.

[0054] This specification may use the phrases "in one embodiment," "in another embodiment," "in yet another embodiment," or "in other embodiments," which may all refer to one or more of the same or different embodiments according to the present application.

[0055] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

[0056] like Figure 1 As shown, an embodiment of the present invention provides a method for classifying and identifying vehicles with magnetic signal characteristics, comprising:

[0057] S1: Obtain magnetic signal data of a first vehicle;

[0058] S2: Preprocessing the magnetic signal data, wherein the preprocessing includes at least one of the following: magnetic compensation, noise reduction, data enhancement, and data expansion;

[0059] S3: performing a first feature extraction on the preprocessed magnetic signal data to obtain a frequency spectrum of a three-component signal, and performing a second feature extraction based on the frequency spectrum of the three-component signal to obtain a first feature vector of the magnetic signal data;

[0060] S4: Determine the driving type and vehicle type of the first vehicle and generate a corresponding label;

[0061] S5: using the first feature vector as input data and the label as output data to train a first initial network model and a second initial network model to form a vehicle drive type detection model and a vehicle type detection model, respectively;

[0062] S6: Processing the magnetic signal data of the second vehicle to be detected to generate a second feature vector, and using the vehicle drive type detection model and the vehicle drive type detection model to respectively process the second feature vector to achieve classification detection of the drive type and vehicle type of the second vehicle.

[0063] For example, magnetic signal data from vehicles passing on a highway can be obtained. This magnetic signal data is collected by magnetic signal sensors installed on the highway. After collecting the magnetic signal data of a first vehicle, the magnetic signal data can be preprocessed and then feature extracted from the preprocessed magnetic signal data to form a first feature vector. The vehicle type and drive type of the first vehicle can then be determined. Finally, two initial models are trained based on the first feature vector and the determined type label to obtain classification models for vehicle type detection and vehicle drive type detection, respectively. Based on these two classification models, traffic system personnel can quickly and accurately classify and identify vehicles traveling on the highway according to different classification situations.

[0064] Based on the disclosure of the above embodiments, it can be known that this embodiment actually extracts the frequency domain features and time-frequency domain features of the vehicle's magnetic signal, and adopts support vector machines and convolutional neural networks as the initial models of the vehicle type classification model and the vehicle drive type classification model. The feature vectors and classification labels formed by processing the vehicle's magnetic signal are trained to obtain models for classifying and identifying the vehicle type and drive type respectively. Based on the two classification models, the classification and identification of the vehicle in two different classification situations can be realized, thereby achieving high-efficiency and high-accuracy vehicle identification and classification.

[0065] The magnetic signal data described in this embodiment includes multiple segments of magnetic signals, such as Figure 2 As shown, the preprocessing of the magnetic signal data includes:

[0066] S7: extracting the first 20 signal detection values of each signal segment in the magnetic signal data, and calculating the average of all signal detection values as the background magnetic field strength value;

[0067] S8: performing magnetic compensation on the magnetic signal data based on the background magnetic field strength value to eliminate the background magnetic field of the signal;

[0068] S9: filtering and eliminating noise and interference signals in the magnetic signal data of the background magnetic field;

[0069] S10: Adding Gaussian white noise to each signal segment in the filtered magnetic signal according to a preset signal-to-noise ratio to perform data enhancement, so as to expand the magnetic signal data.

[0070] The filtering and eliminating of noise and interference signals in the magnetic signal data of the background magnetic field includes:

[0071] S11: filtering and eliminating high-frequency noise above 40 Hz and 50 Hz power frequency interference signals in the magnetic signal data of the background magnetic field.

[0072] The step of adding Gaussian white noise to each segment of the filtered magnetic signal according to a preset signal-to-noise ratio for data enhancement includes:

[0073] S12: Gaussian white noise is added to each signal segment of the filtered magnetic signal in sequence according to preset signal-to-noise ratios of 5dB, 10dB, 15dB, 20dB, and 25dB for data enhancement.

[0074] For example, taking a certain type of vehicle as an example, its sample data set is A, that is, the magnetic signal data is A. After data enhancement, the sample data set of this type of vehicle is A. new for:

[0075] A new =A∪A add_5dB ∪A add_10dB ∪A add_15dB ∪A add_20dB ∪A add_25dB

[0076] Furthermore, when performing the first feature extraction on the pre-processed magnetic signal data to obtain the spectrum of the three-component signal, this embodiment uses fast Fourier transform on the three-component signal of each segment of the magnetic signal to obtain the spectrum of the signal.

[0077] like Figure 3 As shown, when the system performs a second feature extraction based on the frequency spectrum of the three-component signal to obtain a first feature vector of the magnetic signal data, the system includes:

[0078] S13: extracting a unilateral spectrum from the spectrum of the three-component signal, and intercepting a low-frequency portion of the first n frequency points in the unilateral spectrum to form a spectrum feature;

[0079] S14: extracting wavelet transform energy features from the three-component signal;

[0080] S15: performing serial feature fusion on the spectrum features and wavelet transform energy features corresponding to the same segment of magnetic signal to obtain a feature vector corresponding to each segment of magnetic signal.

[0081] For example, the spectrum of the three-component signal is extracted and a single-sided spread is performed, and the low-frequency portion corresponding to below 10Hz in the first 52 frequency points is intercepted to form a spectrum feature. Next, the wavelet transform energy feature is extracted for the three-component signal of each sample. The wavelet decomposition layer number is 10, and each signal component obtained is a 1*11 energy feature vector. The spectrum feature extracted from the three-component signal of each sample and the wavelet transform energy feature are then serially fused to obtain the characteristic vector of each sample:

[0082]

[0083] The beneficial effect of the above scheme is that after preprocessing the data, the spectral characteristics and energy characteristics of each magnetic signal are extracted using fast Fourier transform and wavelet transform, and the extracted features are fused to obtain a feature set. Model training based on this feature set can enable the model to better learn the nonlinear relationship between different types of vehicles and improve the recognition accuracy of the model.

[0084] In another embodiment, the first initial network model and the second initial network model are both composed of a support vector machine model and a convolutional neural network model. When training the model, magnetic signal data from multiple different types of vehicles can be collected and processed based on the above method to generate a first feature vector for model training. Specifically, the method in this embodiment further includes:

[0085] Train a vehicle drive type classification model:

[0086] The 2056 feature vectors and vehicle drive type label data are used as training input data and training label data; the 686 feature vectors and vehicle drive type label data are used as test input data and test label data;

[0087] Using the training input data and the training output label data to train a vehicle driving type classification and recognition support vector machine model and a convolutional neural network model, and inputting the test input data into the trained support vector machine model and the convolutional neural network model to obtain the label data output by the support vector machine model and the convolutional neural network model;

[0088] The support vector machine model and the convolutional neural network model are evaluated based on the output label data. When the accuracy of the output label data is qualified, the training of the support vector machine model and the convolutional neural network model is completed.

[0089] Train the vehicle type classification model:

[0090] The first feature vector is used as input data and the label is used as output data to train a first initial network model to form a vehicle type detection model, including:

[0091] S16: Using the SMOTE algorithm to perform data balancing on the first feature vectors and labels according to vehicle categories;

[0092] S17: Training a first initial network model based on the processed first feature vector as input data and the label as output data to form a vehicle type detection model.

[0093] For example, the SMOTE algorithm is used to perform data balancing on 2056 feature vectors according to vehicle categories; the 4234 feature vectors and vehicle category label data after data balancing are used as training input data and training label data; the 686 feature vectors and vehicle category label data are used as test input data and test label data; the training input data and training output label data are used to train the vehicle drive type classification and recognition support vector machine model and convolutional neural network model, and the test input data is input into the trained support vector machine model and convolutional neural network model to obtain the label data output by the support vector machine model and convolutional neural network model; the support vector machine model and convolutional neural network model are evaluated based on the obtained label data, and if their classification accuracy is qualified, the training of the support vector machine model and convolutional neural network model is completed.

[0094] Furthermore, in this embodiment, the loss function of the support vector machine model is:

[0095]

[0096] The w and b are the parameters of the support vector machine model, T is the vector transpose symbol, x i is the model variable, y i is the actual value of the input vehicle type and vehicle driving type, are the predicted values of vehicle type and vehicle driving type.

[0097] The convolutional neural network is a one-dimensional convolutional neural network consisting of two convolutional layers, two pooling layers and two fully connected layers. The size of the convolution kernel and the pooling kernel are both 3, the step size is 1, and no zero padding is performed. The number of filters in the first convolutional layer is 8, the number of filters in the second convolutional layer is 1, and the pooling layer uses the maximum pooling method. After the activation function of the convolutional layer and before the pooling layer, batch normalization of the data is performed. After convolution and pooling, the data is further processed through two fully connected layers. The number of neuron nodes in the first fully connected layer is 180, and the number of neuron nodes in the second fully connected layer is 90. The final classification output is achieved based on the two fully connected layers, and the number of neuron nodes in the output layer is given according to the number of classification categories.

[0098] Based on the method of this embodiment, the spectrum and time-frequency domain features related to the target category can be effectively extracted, so that the classification and recognition accuracy is high. At the same time, the support vector machine model adopted in this embodiment has the characteristic of automatically finding the optimal parameters. Based on this, both classification models can simultaneously classify and identify vehicles in two different classification situations, and have broad application prospects.

[0099] like Figure 4 As shown, another embodiment of the present invention also provides a classification and identification device 100 for vehicles with magnetic signal characteristics, comprising:

[0100] an acquisition module, configured to obtain magnetic signal data of a first vehicle;

[0101] A preprocessing module, configured to preprocess the magnetic signal data, wherein the preprocessing includes at least one of the following: magnetic compensation, noise reduction, data enhancement, and data expansion;

[0102] an extraction module, configured to perform a first feature extraction on the preprocessed magnetic signal data to obtain a frequency spectrum of a three-component signal, and perform a second feature extraction based on the frequency spectrum of the three-component signal to obtain a first eigenvector of the magnetic signal data;

[0103] a first determining module, configured to determine a driving type and a vehicle type of the first vehicle and generate a corresponding label;

[0104] a training module, configured to train a first initial network model and a second initial network model using the first feature vector as input data and the label as output data to form a vehicle drive type detection model and a vehicle type detection model, respectively;

[0105] The detection module is used to process the magnetic signal data of the second vehicle to be detected to generate a second feature vector, and use the vehicle drive type detection model and the vehicle drive type detection model to process the second feature vector respectively to realize the classification detection of the drive type and vehicle type of the second vehicle.

[0106] In some embodiments, the magnetic signal data includes multiple segments of magnetic signals, and the preprocessing of the magnetic signal data includes:

[0107] Extracting the first 20 signal detection values of each signal segment in the magnetic signal data, and calculating the mean of all signal detection values as the background magnetic field strength value;

[0108] Performing magnetic compensation on the magnetic signal data based on the background magnetic field strength value to eliminate the background magnetic field of the signal;

[0109] Filtering and eliminating noise and interference signals in the magnetic signal data of the background magnetic field;

[0110] Gaussian white noise is added to each signal segment of the filtered magnetic signal according to a preset signal-to-noise ratio to perform data enhancement, so as to expand the magnetic signal data.

[0111] In some embodiments, the filtering and eliminating noise and interference signals in the magnetic signal data of the background magnetic field includes:

[0112] The high-frequency noise above 40 Hz and the power frequency interference signal of 50 Hz in the magnetic signal data of the background magnetic field are filtered out.

[0113] In some embodiments, the step of adding Gaussian white noise to each segment of the filtered magnetic signal according to a preset signal-to-noise ratio to perform data enhancement includes:

[0114] Gaussian white noise is added to each signal segment of the filtered magnetic signal in sequence according to preset signal-to-noise ratios of 5dB, 10dB, 15dB, 20dB, and 25dB for data enhancement.

[0115] In some embodiments, the magnetic signal data includes multiple segments of magnetic signals, and the three-component signal includes three-component signals of each segment of magnetic signal;

[0116] The performing a second feature extraction based on the frequency spectrum of the three-component signal to obtain a first feature vector of the magnetic signal data includes:

[0117] Extracting a unilateral spectrum from the frequency spectrum of the three-component signal, and intercepting a low-frequency portion of the first n frequency points in the unilateral spectrum to form a spectrum feature;

[0118] extracting wavelet transform energy features from the three-component signals;

[0119] The spectrum features and wavelet transform energy features corresponding to the same segment of magnetic signal are serially fused to obtain a feature vector corresponding to each segment of magnetic signal.

[0120] In some embodiments, the first initial network model and the second initial network model are both composed of a support vector machine model and a convolutional neural network model.

[0121] In some embodiments, the first feature vector is used as input data and the label is used as output data to train a first initial network model to form a vehicle type detection model, including:

[0122] Using the SMOTE algorithm to perform data balancing on the first feature vector and label according to vehicle category;

[0123] A first initial network model is trained based on the processed first feature vector as input data and the label as output data to form a vehicle type detection model.

[0124] In some embodiments, the loss function of the support vector machine model is:

[0125]

[0126] The w and b are the parameters of the support vector machine model, y i is the actual value of the input vehicle type and vehicle driving type, are the predicted values of vehicle type and vehicle driving type.

[0127] In some embodiments, the convolutional neural network is a one-dimensional convolutional neural network consisting of two convolutional layers, two pooling layers, and two fully connected layers.

[0128] Another embodiment of the present invention also provides an electronic device, which includes: a memory and a processor, wherein the memory stores a computer program run by the processor, and when the computer program is run by the processor, the processor executes the method for classifying and identifying vehicles with magnetic signal characteristics as described in any of the embodiments above.

[0129] Another embodiment of the present application further provides a computer-readable storage medium, which includes a stored program, wherein when the program is run, a device including the storage medium is controlled to execute the method for classifying and identifying vehicles with magnetic signal characteristics as described in any of the embodiments above.

[0130] The present application also provides a computer program product tangibly stored on a computer-readable medium and comprising computer-readable instructions. When executed, the computer-executable instructions cause at least one processor to perform a method for classifying and identifying vehicles with magnetic signature characteristics, such as that described in the above-described embodiments. It should be understood that each solution in this embodiment has the corresponding technical effects of the above-described method embodiments and will not be further elaborated here.

[0131] It should be noted that the computer storage medium of the present application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. Computer-readable media can be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or devices, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage medium, a magnetic storage medium, or any suitable combination of the above. In the present application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device, or device. In the present application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. This propagated data signal can take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program configured for use by or in conjunction with an instruction execution system, apparatus, or device. Program code embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, antenna, optical cable, RF, or any suitable combination thereof.

[0132] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of protection of the present application is limited to these examples. In line with the present application, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of different aspects of one or more embodiments of the present application as described above, which are not provided in detail for the sake of simplicity.

[0133] The one or more embodiments of this application are intended to encompass all such substitutions, modifications, and variations that fall within the broad scope of this application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of one or more embodiments of this application should be included in the scope of protection of this application.

Claims

1. A method for classifying and identifying vehicles with magnetic signal characteristics, characterized in that: include: obtaining magnetic signal data of a first vehicle; Preprocessing the magnetic signal data, wherein the preprocessing includes at least one of the following: magnetic compensation, noise reduction, data enhancement, and data expansion; Performing a first feature extraction on the preprocessed magnetic signal data to obtain a frequency spectrum of a three-component signal, and performing a second feature extraction based on the frequency spectrum of the three-component signal to obtain a first eigenvector of the magnetic signal data; determining a drive type and a vehicle type of the first vehicle and generating a corresponding label; Using the first feature vector as input data and the label as output data to train a first initial network model and a second initial network model to form a vehicle drive type detection model and a vehicle type detection model respectively; The magnetic signal data of the second vehicle to be detected is processed to generate a second feature vector, and the second feature vector is processed using the vehicle drive type detection model and the vehicle drive type detection model respectively to achieve classification detection of the drive type and vehicle type of the second vehicle.

2. The method for classifying and identifying vehicles with magnetic signal characteristics according to claim 1, characterized in that: The magnetic signal data includes multiple segments of magnetic signals, and the preprocessing of the magnetic signal data includes: Extracting the first 20 signal detection values of each signal segment in the magnetic signal data, and calculating the mean of all signal detection values as the background magnetic field strength value; Performing magnetic compensation on the magnetic signal data based on the background magnetic field strength value to eliminate the background magnetic field of the signal; Filtering and eliminating noise and interference signals in the magnetic signal data of the background magnetic field; Gaussian white noise is added to each signal segment of the filtered magnetic signal according to a preset signal-to-noise ratio to perform data enhancement, so as to expand the magnetic signal data.

3. The method for classifying and identifying vehicles with magnetic signal characteristics according to claim 2, characterized in that: The filtering and eliminating noise and interference signals in the magnetic signal data of the background magnetic field includes: The high-frequency noise above 40 Hz and the power frequency interference signal of 50 Hz in the magnetic signal data of the background magnetic field are filtered out.

4. The method for classifying and identifying vehicles with magnetic signal characteristics according to claim 2, characterized in that: The step of adding Gaussian white noise to each segment of the filtered magnetic signal according to a preset signal-to-noise ratio for data enhancement includes: Gaussian white noise is added to each signal segment of the filtered magnetic signal in sequence according to preset signal-to-noise ratios of 5dB, 10dB, 15dB, 20dB, and 25dB for data enhancement.

5. The method for classifying and identifying vehicles with magnetic signal characteristics according to claim 1, characterized in that: The magnetic signal data includes multiple segments of magnetic signals, and the three-component signal includes three-component signals of each segment of magnetic signal; The performing a second feature extraction based on the frequency spectrum of the three-component signal to obtain a first feature vector of the magnetic signal data includes: Extracting a unilateral spectrum from the frequency spectrum of the three-component signal, and intercepting a low-frequency portion of the first n frequency points in the unilateral spectrum to form a spectrum feature; extracting wavelet transform energy features from the three-component signals; The spectrum features and wavelet transform energy features corresponding to the same segment of magnetic signal are serially fused to obtain a feature vector corresponding to each segment of magnetic signal.

6. The method for classifying and identifying vehicles with magnetic signal characteristics according to claim 1, characterized in that: The first initial network model and the second initial network model are both composed of a support vector machine model and a convolutional neural network model.

7. The method for classifying and identifying vehicles with magnetic signal characteristics according to claim 6, characterized in that: The first feature vector is used as input data and the label is used as output data to train a first initial network model to form a vehicle type detection model, including: Using the SMOTE algorithm to perform data balancing on the first feature vector and label according to vehicle category; A first initial network model is trained based on the processed first feature vector as input data and the label as output data to form a vehicle type detection model.

8. The method for classifying and identifying vehicles with magnetic signal characteristics according to claim 6, characterized in that: The loss function of the support vector machine model is: The w and b are the parameters of the support vector machine model, T is the vector transpose symbol, x i is the model variable, y i is the actual value of the input vehicle type and vehicle driving type, are the predicted values of vehicle type and vehicle driving type.

9. The method for classifying and identifying vehicles with magnetic signal characteristics according to claim 6, characterized in that: The convolutional neural network is a one-dimensional convolutional neural network consisting of two convolutional layers, two pooling layers and two fully connected layers.

10. A device for classifying and identifying vehicles with magnetic signal characteristics, characterized in that: include: an acquisition module, configured to obtain magnetic signal data of a first vehicle; A preprocessing module, configured to preprocess the magnetic signal data, wherein the preprocessing includes at least one of the following: magnetic compensation, noise reduction, data enhancement, and data expansion; an extraction module, configured to perform a first feature extraction on the preprocessed magnetic signal data to obtain a frequency spectrum of a three-component signal, and perform a second feature extraction based on the frequency spectrum of the three-component signal to obtain a first eigenvector of the magnetic signal data; a first determining module, configured to determine a driving type and a vehicle type of the first vehicle and generate a corresponding label; a training module, configured to train a first initial network model and a second initial network model using the first feature vector as input data and the label as output data to form a vehicle drive type detection model and a vehicle type detection model, respectively; The detection module is used to process the magnetic signal data of the second vehicle to be detected to generate a second feature vector, and use the vehicle drive type detection model and the vehicle drive type detection model to process the second feature vector respectively to realize the classification detection of the drive type and vehicle type of the second vehicle.

Images