Method, apparatus, and system for monitoring public road traffic and road conditions using earthquake data

Abstract

This application provides a method, apparatus, and system for monitoring traffic and road conditions using earthquake data. Specifically, a method for monitoring traffic volume using earthquake data may include acquiring earthquake data from earthquake recording equipment, acquiring an earthquake data deformation diagram that includes multiple earthquake data deformation curves from each earthquake recording equipment based on the earthquake data, and monitoring traffic volume based on the earthquake data deformation diagram.

Description

[Technical Field]

[0001] The present invention, as a whole, relates to the field of data processing technology, and more specifically to methods, systems, electronic devices, non-temporary computer-readable storage media and computer program products including memory instructions for traffic monitoring and road condition management. The present invention can improve the accuracy of predictions and response capabilities by using collected data to monitor and predict dynamic changes in traffic and traffic accidents in real time, and by performing comprehensive analysis in combination with external information (such as seismic wave data). [Background technology]

[0002] Road traffic accidents are one of the leading causes of death worldwide. According to estimates by the World Health Organization (WHO), approximately 1.3 million people die in road accidents each year. The Centers for Disease Control and Prevention (CDC) reported that in 2019, road accident fatalities were 3 to 10 times higher in rural areas than in urban areas, accounting for about 45% of all road accident deaths nationwide. Because accidents on these roads take a long time to report, it is beneficial to monitor these roads using traffic detection systems and methods.

[0003] Monitoring traffic volume across multiple roads and entire cities is crucial for traffic management departments to recognize any traffic problems in real time. The advent of autonomous vehicles has further increased the need for real-time traffic information. This information provides opportunities for lane traffic density management systems and methods that can automatically optimize the spacing between autonomous vehicles on the road, and between other autonomous vehicles and non-autonomous vehicles. More specifically, it allows for the automatic control of autonomous vehicle behavior to maintain optimal distances from vehicles ahead and behind.

[0004] Currently, multiple methods are used for vehicle detection, including video image processing, radar, and ultrasonic sensors. Installing numerous camera systems along roadsides can provide detailed traffic volume and individual vehicle information. However, monitoring traffic and road conditions using a large number of cameras presents several problems, such as the following:

[0005] 1. High cost: Because rural roads cover longer distances, it is difficult to actually implement these systems. The cost per traffic camera is typically tens of thousands of dollars, and covering vast distances requires a large number of cameras or sensors, making it extremely expensive.

[0006] 2. Real-time data processing requirements: To quickly grasp current traffic conditions, it is necessary to access large amounts of video data and extract useful information. This requires not only significant real-time computing power but also expensive equipment, increasing the overall system cost.

[0007] 3. Privacy Regulations: Many regions and countries have introduced privacy regulations that restrict or prohibit the use of video recording in public places. This creates legal and ethical challenges to the widespread use of cameras.

[0008] Furthermore, the mobile phone location services currently used in many navigation systems may not provide accurate traffic information. Therefore, new traffic and road monitoring methods need to be developed to solve the aforementioned technical problems. [Overview of the Initiative] [Means for solving the problem]

[0009] The first aspect of this application provides a method for monitoring traffic volume using earthquake data, the method comprising the following steps: ● Obtain earthquake data from earthquake recording equipment. ● Based on the acquired earthquake data, an earthquake data deformation map is obtained that includes multiple earthquake data deformation curves from each earthquake recording device. ● Monitor traffic volume based on earthquake data fluctuation maps.

[0010] In this method, seismic recording equipment can be placed at fixed or variable intervals on one side of the road, both sides, in the middle, or any combination of these positions.

[0011] Furthermore, generating earthquake data deformation maps includes the following: ● Apply signal enhancement to earthquake data. ● After signal amplification, separate or attenuate the bidirectional traffic wave field. ● Use a correction curve to balance the seismic data and generate a seismic data deformation map based on the balanced seismic data. This method further includes the following: ● To visually display the vehicle's motion, the vehicle speed spectrum is acquired based on seismic data deformation diagrams. ● By analyzing the similarity of seismic data between different seismic recording devices, the vehicle's speed and / or trajectory can be determined. ● Determine whether a vehicle is speeding based on its speed. ●The type and weight of the vehicle are determined by analyzing the peak values ​​of the vehicle in multiple seismic data fluctuation curves. ● Determine whether a vehicle is stopped or not based on its movement path. ● Determine whether pedestrians are walking on the road based on earthquake data. ● Based on earthquake data, it detects whether an object fell from the vehicle.

[0012] A second aspect of this invention application provides a system for monitoring traffic volume using earthquake data, and this system is ● Earthquake recording equipment configured to acquire earthquake data, ● A data processing module configured to generate a seismic data variation diagram including a plurality of seismic data variation curves of each seismic recording device based on seismic data; ● An analysis module configured to analyze traffic volume based on the seismic data variation diagram, and includes.

[0013] According to a second aspect, the analysis module can execute the method of the first aspect. A third aspect of the present application provides a system for monitoring traffic volume using seismic data. This system includes: ● A seismic recording device configured to acquire seismic data; ● A processor; ● A memory connected to the processor and storing instructions, When the instructions are executed, the processor ○ Acquires a seismic data variation diagram including a plurality of seismic data variation curves of each seismic recording device based on seismic data; ○ Monitors traffic volume based on the seismic data variation diagram Is executed as follows.

[0014] A fourth aspect of the present application provides a method for training a vehicle recognition model. This method includes: ● Acquiring a first image including a vehicle from a video; ● Converting the first image into a second image representing the vehicle and its lane and used as a training label; ● Acquiring first seismic data corresponding to the first image from seismic data; ● Training a vehicle recognition model based on the first seismic data and the second image.

[0015] A fifth aspect of the present application provides a vehicle recognition method. This method includes: ● Inputting seismic data to be recognized into a vehicle recognition model to determine the vehicle and its driving lane. Here, the vehicle recognition model is trained using the method of the fourth aspect.

[0016] A sixth aspect of this invention application provides a method for training a vehicle recognition model, the method being: ● Picking up earthquake data as labels in multiple time windows, ● Recognize the maximum peak or Gaussian distribution peak for each label, ● Determining the vehicle position corresponding to the maximum peak or Gaussian distribution peak, This may include training a vehicle recognition model based on location and the corresponding maximum peak or Gaussian-distributed peak.

[0017] A seventh aspect of this invention application provides a vehicle recognition method, which includes, ● Inputting earthquake data to be recognized into a vehicle recognition model to determine the vehicle's position, ● The vehicle recognition model is trained using a sixth method, which includes determining the vehicle's speed and / or travel path based on its position.

[0018] The eighth aspect of this invention application provides a method for monitoring geological conditions beneath a road using seismic data, and this method is: ● Obtain target earthquake data and reference earthquake data from earthquake recording equipment, ●Generating a target Green's function based on target earthquake data, ●Generating a reference Green's function based on reference earthquake data, ●This may include generating near-surface relative velocity changes based on a target Green's function and a reference Green's function to monitor the geological conditions beneath the road.

[0019] According to the eighth aspect, this method is, ●This may further include performing signal preprocessing, denoising, resampling, data offset correction, and filtering on the target earthquake data and reference earthquake data.

[0020] In the eighth direction, filtering is performed across a multi-frequency range (f 1i ,f 2i) Includes bandpass filtering, where f0 <f 1i <f 1、 f0 <f 2i <f 1であり、 f0 and f1 are the lower and upper limits of the frequency range of the earthquake recording device.

[0021] According to the eighth aspect, this method is, ● Extract one or more of the following from the target earthquake data and the reference earthquake data: bulk waves, elastic P waves, S waves, SH waves, surface waves, coda waves, Rayleigh waves, and Love waves. ●This may further include generating a reference Green's function and a target Green's function based on the extracted waves.

[0022] The eighth direction generates a change in relative velocity near the Earth's surface. ● This may include generating near-surface relative velocity changes based on a target Green's function and a reference Green's function using environmental noise imaging.

[0023] The ninth aspect of this application provides a method for monitoring road surface conditions using seismic data, and this method is ● Obtain target earthquake data and reference earthquake data from earthquake recording equipment, ● Based on reference earthquake data, calculate the total reference energy of vehicles passing through a certain frequency range at a reference time, ● Based on target earthquake data, calculate the target total energy of vehicles passing through the same frequency range at the target time, ● This may include monitoring road surface conditions based on reference total energy and target total energy.

[0024] According to the ninth principle, the calculation of the reference total energy and the target total energy are performed in the frequency domain or the time domain.

[0025] According to the ninth principle, the calculation of the reference total energy and the target total energy are performed across multiple frequency bands.

[0026] In the ninth area, monitoring road surface conditions based on reference total energy and target total energy is, ●Calculate the reference ratio between the total reference energies of multiple frequency bands, and calculate the target ratio between the total target energies of multiple frequency bands. ●This may include comparing the reference ratio with the target ratio and performing weighted addition or subtraction.

[0027] The tenth aspect of this invention application provides a method for predicting traffic accidents, and this method is ● Acquiring earthquake data from earthquake recording equipment, ● Obtaining vehicle information based on earthquake data, ●This may include predicting traffic accidents based on vehicle information, current and past road information, driver information, and weather information.

[0028] The eleventh aspect of this invention application provides a method for training a traffic accident prediction model, and this method is ● To acquire earthquake data, vehicle information, road information, driver information, and weather information at the time of the accident, ●This may include training a traffic accident prediction model based on the aforementioned earthquake data and information.

[0029] The twelfth aspect of this invention application provides a method for predicting traffic accidents, and this method is: ●This may include inputting earthquake data to be recognized into a traffic accident prediction model to predict traffic accidents and thereby obtaining a traffic accident score, wherein the traffic accident prediction model is trained using the eleventh-plane method.

[0030] A thirteenth aspect of the present invention application provides a method for training a vehicle weight model, the method being: ● Obtaining peak vehicle values ​​from earthquake data, ●This may include training a vehicle weight model based on the peak value and the vehicle's weight.

[0031] The fourteenth aspect of this invention application provides a method for obtaining vehicle weight, the method being, ●The earthquake data to be recognized is input into a vehicle weight model to obtain the vehicle's weight, and the vehicle weight model is trained using the thirteenth method.

[0032] The fifteenth aspect of this invention application provides an electronic device, the electronic device is ●At least one processor, ●Memory connected to at least one processor, ●The method may include a memory that, when executed by at least one processor, stores instructions that cause at least one processor to perform the method described above.

[0033] A sixteenth aspect of the present invention application provides a non-temporary computer-readable storage medium that, when executed by a computer, stores instructions causing the computer to perform the method described above.

[0034] The seventeenth aspect of the present invention application provides a computer program product that, when executed by a computer, includes a computer program having instructions that cause the computer to perform the method described above. [Brief explanation of the drawing]

[0035] [Figure 1] Figure 1 is a schematic diagram of a system architecture for implementing a traffic and road monitoring method according to an exemplary embodiment of the present disclosure. [Figure 2] Figure 2 is a schematic flowchart illustrating a method for monitoring traffic volume according to an exemplary embodiment of the present disclosure. [Figure 3] Figure 3 shows schematic diagrams illustrating several design examples for placing earthquake sensor recording locations along roads to monitor traffic. [Figure 4] Figure 4 shows a schematic flowchart for processing earthquake data. [Figure 5A] Figures 5A to 5C show graphs of surface-matched amplitude correction. [Figure 5B] Figures 5A to 5C show graphs of surface-matched amplitude correction. [Figure 5C] Figures 5A to 5C show graphs of surface-matched amplitude correction. [Figure 6] Figure 6 shows the graph after bandpass signal processing. [Figure 7] Figure 7 shows the vehicle speed spectrum generated from the bandpass signal processing data. [Figure 8] Figure 8 shows a graph of the vehicle speed spectrum generated from signal-enhanced traffic data. [Figure 9] Figure 9 shows how vehicle speed is extracted from the vehicle's path and speed spectrum. [Figure 10] Figure 10 shows a graph for determining whether a vehicle is speeding. [Figure 11] Figure 11 shows the interpretation of the relationship between vehicle type, speed, weight and earthquake amplitude. [Figure 12] Figure 12 shows a graph illustrating the stopping of a vehicle. [Figure 13] Figure 13 shows the path of a single vehicle moving along a road (shown as a curve). [Figure 14] Figure 14 shows a typical seismic response of a vehicle. [Figure 15] Figure 15 shows the seismic response of a person walking on a highway (indicated within the area). [Figure 16] Figure 16 shows the extrapolation of the velocity spectrum in the spatial domain. [Figure 17] Figure 17 is a diagram showing the configuration of system 1700, which monitors traffic volume using earthquake data according to the embodiment. [Figure 18]Figure 18 shows the process of converting a first video image to a second two-color image, where (a) displays a video recording of a road and vehicle detection is performed using image / video recognition technology, and (b) shows training labels generated using the vehicle detected from the video recording. [Figure 19] Figure 19 shows how machine learning is used to detect and interpret vehicles. [Figure 20] Figure 20 shows vehicle paths obtained directly from ML-spike (Figure 20(a)) or ML-curve (Figure 20(b)). [Figure 21] Figure 21 is a schematic flowchart illustrating a method for monitoring geological conditions beneath a road using seismic data, according to an exemplary embodiment of the present disclosure. [Figure 22] Figure 22 shows the reference Green's function between pairs of stations (earthquake sensors). [Figure 23] Figure 23 shows a comparison of Green's functions between reference time and target time. [Figure 24] Figure 24 shows an image illustrating the relative changes in seismic velocity beneath the road. [Figure 25] Figure 25 is a schematic flowchart illustrating a method for monitoring road surface conditions using seismic data according to an exemplary embodiment of the present disclosure. [Figure 26] Figure 26 shows the change in the road surface characteristic response over time. [Figure 27] Figure 27 is a schematic flowchart illustrating a method for predicting traffic accidents according to an exemplary embodiment of the present disclosure. [Figure 28] Figure 28 shows an electronic device according to an exemplary embodiment of the present disclosure. [Modes for carrying out the invention]

[0036] Various embodiments of this disclosure will be described below with reference to the accompanying drawings. In the following description, specific details such as detailed configurations and components are provided solely to aid in the overall understanding of these embodiments of this disclosure. Therefore, those skilled in the art will understand that various changes and modifications of the embodiments described herein can be carried out without departing from the scope and spirit of this disclosure. Furthermore, for clarity and brevity, descriptions of well-known functions and configurations have been omitted.

[0037] These exemplary embodiments and the terminology used herein are not intended to limit the technology disclosed herein to any particular form, but should be understood to include various modifications, equivalents, and / or substitutions to the corresponding embodiments. In the description of the drawings, similar reference numerals may be used to indicate similar components. Unless otherwise clearly indicated in context, singular expressions may include plural expressions.

[0038] Furthermore, the various embodiments described herein are not necessarily mutually exclusive, and some embodiments can be combined with one or more other embodiments to form new embodiments.

[0039] In this specification, the singular form may also include the plural form unless the context clearly indicates otherwise. The expressions “first,” “second,” “the first,” or “the second,” used in various embodiments of this disclosure may modify various elements, regardless of order and / or importance, but are not limited to the corresponding elements. When an element (e.g., the first element) is described as being “connected (functionally or communicatively)” or “directly coupled” to another element (e.g., the second element), that element may be directly connected to the other element or connected to the other element via another element (e.g., the third element).

[0040] As used in various embodiments of this disclosure, the expression “configured” may, depending on the context, be interchangeable with expressions such as “suitable,” “capable,” “designed,” “modified,” “made,” or “capable” with respect to hardware or software. In some cases, the expression “device configured to” may refer to a situation in which such device “can,” for example, work together with other devices or components. For example, the phrase “processor modified (or configured) to perform A, B, and C” refers to (but is not limited to) a dedicated processor (e.g., an embedded processor) that performs the corresponding operations, or a general-purpose processor (e.g., a central processing unit (CPU) or application processor (AP)) that performs the corresponding operations by running one or more software programs stored in a memory device.

[0041] Here, unless otherwise specified, the term "or" as used herein refers to a non-exclusive "or". The examples used herein are intended solely to facilitate understanding of the embodiments of this specification and to enable those skilled in the art to further implement the embodiments herein. Therefore, these examples should not be construed as limiting the scope of the embodiments herein.

[0042] As is customary in the art, embodiments are described and explained in terms of blocks that perform the functions described. These blocks may be referred to herein as units, engines, money, modules or similar names, and may be physically realized by analog and / or digital circuits such as logic gates, integrated circuits, microprocessors, microcontrollers, memory circuits, passive electronic components, active electronic components, optical components, and hardwired circuits, and may be selectively driven by firmware and / or software. For example, these circuits may be realized on one or more semiconductor chips or on a substrate support such as a printed circuit board. The circuits constituting a block may be realized by dedicated hardware, or by a processor (e.g., one or more programmed microprocessors and their associated circuits), or by a combination of dedicated hardware that performs some of the functions of the block and a processor that performs other functions of the block. Each block of an embodiment may be physically separated into two or more interacting separate blocks without departing from the scope of the disclosure. Similarly, blocks of an embodiment may be physically combined into more complex blocks without departing from the scope of the disclosure.

[0043] When expressions such as "at least one of A, B, and C" are used, they should generally be interpreted according to the meaning generally understood by those skilled in the art (for example, "a system comprising at least one of A, B, and C" includes, but is not limited to, systems comprising A, B, C, A and B, A and C, B and C, and / or A, B, C, etc.).

[0044] In earthquake monitoring, seismic sensors are installed to record elastic waves propagating from the earthquake source. These are typically designed to monitor wide areas of the Earth, and due to installation constraints, the sensor locations may be irregular. Each sensor records a time series of ground vibrations at uniform sampling intervals. The signal can be recorded in the form of ground vibration displacement, velocity (velocity sensor), or acceleration (accelerometer). Numerical calculations show that these three data formats are largely interchangeable. Due to hardware and electronic limitations, seismic sensors are designed to record ground vibrations within a specific frequency range. These recordings can be used to estimate the location of the earthquake's epicenter, seismic intensity, epicenter mechanism, and the velocity of the medium between the epicenter and the receiver. A sensor may include a single vertical component vibrating perpendicular to the ground, or it may record three components vibrating along three orthogonal directions. A single three-component sensor can, under simplified assumptions, determine the direction in which the earthquake's P-waves or S-waves are incident.

[0045] Earthquake research is not limited to earthquake problems; it is also widely used for the discovery of underground oil, natural gas, and other resources. In seismic exploration for minerals, coal, oil, and gas, seismography uses one or more controlled sources and one or more receivers to collect information about the phase and amplitude of seismic waves. The sources and receivers used in seismography can be placed regularly or irregularly on the Earth's surface, either along a line (for 2D images) or across a region (for 3D images), according to a designed geometry. In these investigations, the sources and receivers are placed at known locations. The velocity medium or rock interface is the target of imaging.

[0046] Seismic data may contain numerous seismic traces. Each sensor records either one vertical component trace or three traces of three components. A seismic trace consists of data on ground movement over time, recorded by receivers such as seismographs, hydrophones, micro-electromechanical systems (MEMS) sensors, distributed acoustic sensing (DAS), and other seismic monitoring devices. Receivers can be placed at approximately constant intervals along the surface. While ground movement can be collected passively, seismography uses epicenters to create seismic waves from known sources. An epicenter is the location where seismic waves originate. Primary excitation generates seismic waves through the release of energy at the epicenter, such as the explosion of a single dynamite or a vibrator. The location where the excitation occurs is called the epicenter. Primary excitation can be performed at the epicenter to generate seismic waves. Multiple excitations can be performed at different epicenters and at different times.

[0047] Seismometers are installed on the ground to record vibrations caused by earthquakes or controlled epicenters. Analyzing this data reveals epicenter information and information about the Earth's medium through which seismic waves propagated. Seismic sensors can also be installed along roadsides to monitor ground movement caused by passing vehicles acting as passive epicenters. Analyzing this data through designed algorithms and methods helps to reveal the speed, weight, spacing, position, and driving patterns of all vehicles across the road being monitored by the seismic sensors. This data can also be used to monitor daily changes in road and bridge conditions for risk assessment. These results help minimize the occurrence of accidents and disasters and promote traffic control and safe driving at low cost.

[0048] Designing a traffic volume reconstruction process using earthquake data involves one or more of the following steps:

[0049] 1) Receive traffic vibration data recorded on-site and transmitted to a computer in real time.

[0050] 2) Data signals are enhanced through digital processing. 3) Separation or amplitude attenuation processing of bidirectional traffic wave fields.

[0051] 4) Ground consistency amplitude correction. 5) Extract amplitude and frequency attributes to determine the vehicle's weight and type.

[0052] 6) Scan the vehicle speed to generate a speed spectrum. 7) Use machine learning methods to detect vehicles and output vehicle labels, which are assigned to the vehicle arrival time using pulses or Gaussian signals, to the time data.

[0053] 8) The pulse or Gaussian signal obtained in step (7) is correlated to generate a spatiotemporal trajectory curve that reflects the vehicle traveling along the road.

[0054] 9) Infer driving behavior from the spatiotemporal trajectory curve of the vehicle's movement and lane position. 10) Reconstruct real-time traffic volume including vehicle type, weight, speed, and location (position and lane).

[0055] One embodiment of this application provides a method for monitoring traffic volume using seismic data, thereby reducing system layout costs, reducing the amount of data that needs to be processed, protecting individual privacy, and providing high timeliness. This method may include acquiring seismic data from seismic recording equipment, acquiring a seismic data deformation diagram that includes multiple seismic data deformation curves for each seismic recording equipment based on the seismic data, and monitoring traffic volume based on the seismic data deformation diagram.

[0056] Extracting information such as the speed, weight, number of axles, position, lane position, and driving mode of all vehicles from ground vibration data is relatively easy, and the amount of data is orders of magnitude less than that of video data. The results of vibration data analysis allow for a comprehensive view of all vehicles on a road or in a city, and camera systems can focus on specific vehicles or roads with minimal cost and computation. Camera systems are like human "eyes," and hearing is like human "ears." The combination of "eyes" and "ears" enables more powerful monitoring capabilities in many applications.

[0057] The drawings, particularly Figures 1 through 28, are referenced here. In the figures, similar reference letters consistently indicate corresponding features throughout the multiple figures. Preferred embodiments are shown.

[0058] Figure 1 is a schematic diagram of a system architecture for implementing a traffic and road monitoring method according to an exemplary embodiment of the present disclosure.

[0059] Figure 1 is merely an example of a system architecture applicable to the implementation disclosed in this application and is intended to help those skilled in the art understand the technical content disclosed herein. It should be noted that this does not mean that the disclosed implementation cannot be used in other devices, systems, environments, or scenarios.

[0060] As shown in Figure 1, the system architecture 100 of this embodiment may include earthquake recording equipment such as earthquake sensors 101, 102, and 103, a network 104, and a server 105. The network 104 is a medium that provides a communication link between the earthquake sensors 101, 102, and 103 and the server 105. The network 104 may include various types of connections, such as wired, wireless, and fiber optic cables.

[0061] In the embodiment, earthquake data can be transmitted via a 5G / 4G cellular network, Wi-Fi, or fiber optic internet cable. Server 105 may be a server that provides various services. Server 105 may be a cloud server, also called a cloud computing server or cloud host. Server 105 may be a server for a distributed system, or a server combined with blockchain technology.

[0062] The traffic and road monitoring method of this embodiment is typically performed by server 105. Accordingly, a unit or module for performing the method of this embodiment may be provided to server 105. The traffic and road monitoring method of this embodiment may also be performed by a device, server, or server cluster that is different from server 105 but can communicate with server 105. Accordingly, a unit or module for performing the traffic and road monitoring method of this embodiment may be installed in a device, server, or server cluster that is different from server 105 but can communicate with server 105.

[0063] Specific Embodiments traffic monitoring Figure 2 is a schematic flowchart 200 illustrating a method for monitoring traffic volume according to an exemplary embodiment of the present disclosure. As shown in Figure 2, the flowchart 200 may include the following operations: ●In operation S210, earthquake data is acquired from the earthquake recording equipment. ●In operation S220, an earthquake data deformation map can be obtained based on earthquake data, and the earthquake data deformation map includes multiple earthquake data deformation curves from each earthquake recording device. ●Operation S230 allows for monitoring of traffic volume based on earthquake data deformation maps.

[0064] In operation S210 of the embodiment, the seismic recording device may be a seismograph, a seismic sensor, or a ground vibration recording device. In the embodiment, the seismic recording devices can be arranged at fixed or variable intervals on one side of the road, both sides of the road, in the middle of the road, or any combination thereof. This will be illustrated with reference to Figure 3. Detailed information on operation S220 will be explained in detail with reference to Figures 4 to 6. Seismic vibrations generated by vehicles traveling along the road are recorded by the seismic sensor. By using the recorded seismic data, traffic volume can be reconstructed and moving vehicle information can be estimated in real time. Traffic volume monitoring may include digital traffic volume reconstruction and detection of abnormal activity through operation S230.

[0065] Figure 3 shows schematic diagrams illustrating several design examples of roadside seismic sensor recording locations for monitoring traffic. As shown in Figure 3, examples of seismic sensor placements may include unilateral placement (e.g., on the right side of the road as shown in Figure 3(a), and on the left side as shown in Figure 3(b)), center lane placement (e.g., on a central island or median as shown in Figure 3(c)), bilateral placement such as alternating placement on both sides of the road as shown in Figure 3(d), and combinations thereof.

[0066] Vehicle Vibration Data Frequency Range - The frequency range of vibration data recorded by seismic sensors varies from 0.01 Hz to several kiloHz. Across the entire frequency spectrum of the recorded data, the frequency response of larger and heavier vehicles differs from that of smaller and lighter vehicles. Therefore, it is important to capture the entire frequency range of ground vibrations caused by vehicles. Furthermore, using the entire frequency range of recorded vehicle vibrations is important for determining the road surface conditions and the structure beneath them.

[0067] Figure 4 shows a schematic flowchart 400 for processing earthquake data. As shown in Figure 4, flowchart 400 may include the following steps. ●In step S410, earthquake data is signal-enhanced. ●In step S420, after signal amplification, the bidirectional traffic wave field can be separated or attenuated. ●In step S430, the seismic data is balanced using a correction curve. ●In step S440, a seismic data deformation map can be generated based on balanced seismic data.

[0068] In step S410, signal enhancement can be performed on the seismic data. Vehicle signal enhancement - Since earthquake sensors record vehicle signals along with noise in the same or different frequency bands, denoising or signal enhancement is required. This may include applying machine learning (ML), bandpass filtering, median filtering, root mean square (RMS) filtering, automatic gain control (AGC), trace balancing, wavelet transform, superimposing signals in travel time windows, extracting signal envelopes, transforming data using the STA / LTA ratio (ratio of short-time window to long-time window), and wavelet and convolution.

[0069] Step S410 may include one or more of the following processes when performing signal enhancement on the seismic data:

[0070] 1) Remove or attenuate noise. 2) Eliminate or attenuate irregular signals other than vehicle vibrations.

[0071] 3) Eliminate or dampen the secondary response of the vehicle's initial vibrations. 4) Apply mathematical logarithmic operations to the traffic data to normalize the amplitude.

[0072] 5) Obtain an arbitrary index (including the square root) of the traffic data and normalize the amplitude. 6) Apply automatic gain control (AGC) to the traffic data to normalize the amplitude.

[0073] 7) Apply the root mean square (RMS) to the traffic data to normalize the amplitude. 8) Apply median filtering to the traffic data to smooth the data and remove outliers.

[0074] 9) Apply mean filtering to the traffic data to smooth it out. 10) Apply integral filtering to traffic data to amplify weak signals.

[0075] 11) Apply global normalization to the traffic data. 12) Apply single-station (sensor) normalization to the traffic data.

[0076] 13) Apply local window normalization to the traffic data. 14) Apply items 4) to 13) above to the traffic data signal envelope.

[0077] 15) Apply bandpass filtering to traffic data or preprocessed data. 16) Apply linear motion correction (LMO) to the data and use a pre-set vehicle speed relative to the recording station.

[0078] 17) Apply any or all combinations of 1) to 16). In step S420, the bidirectional traffic wave field can be separated or attenuated after signal enhancement. Separation or attenuation of the bidirectional traffic wave field includes the following: ● Frequency-wavenumber domain filtering (FK filtering). ● Perform FK filtering on the pre-processed earthquake data. ● Apply attenuation to the pre-processed earthquake data. ● Apply machine learning (ML) to input bidirectional traffic data and output unidirectional traffic data.

[0079] In step S430, seismic data can be balanced using a ground surface consistent amplitude correction curve. Balancing seismic data between recording stations equipped with seismic sensors includes the following: ●Apply ground surface consistent amplitude correction to traffic data. ● Apply normalization of earthquake traces to traffic data.

[0080] Ground Surface Coherence Amplitude Correction - When the same vehicle passes different roadside receivers at the same speed, the amplitude response may change due to variations in road structure and the receiver's coupling to the ground. This amplitude amplification or attenuation effect is specific to the location and receiver, and all vehicles passing through the same location are affected in the same way. Therefore, this is consistent with the ground surface. Amplitude changes can be measured from existing datasets and the response curves can be interpolated to correct for these amplitude changes. To account for differences in response due to differences in vehicle weight, multiple correction curves can be generated for each group of different vehicle weights.

[0081] Because the road and roadbed conditions differ at each station, it is necessary to correct the earthquake data in which vehicle vibrations have been recorded.

[0082] Figures 5A-5C illustrate the correction process, in which the original recorded data is multiplied by the station's correction function / correction curve to obtain data with corrected seismic amplitude for further processing. The vertical axis represents the spatial position of the seismic sensor. The horizontal axis of the seismic data graphs (Figures 5A and 5C) represents the recording time, while the horizontal axis of Figure 5B represents the station correction value.

[0083] Figure 5A shows the original data, Figure 5B shows the correction curve, and Figure 5C shows the figure after surface-matched amplitude correction.

[0084] In step S440, a seismic data deformation map is obtained based on balanced seismic data.

[0085] To remove some of the background noise and highlight vehicle events, it is necessary to bandpass filter different frequency ranges. In one embodiment, bandpass signal processing can be performed after the seismic data has been balanced.

[0086] Figure 6 shows the application of bandpass filtering to amplitude-corrected data. Figure 6(a) shows a schematic diagram of ground-surface-matched amplitude correction, and Figure 6(b) shows a schematic diagram of bandpass signal processing. It can be seen that the seismic data deformation map is clearer after bandpass signal processing.

[0087] In one embodiment, a seismic data deformation diagram can be obtained based on seismic data after bandpass signal processing, and then a vehicle speed spectrum can be obtained based on the seismic data deformation diagram. The vehicle speed spectrum is used to intuitively display the vehicle's motion state.

[0088] A vehicle speed spectrum is a two-dimensional image spectral diagram with coordinates (time, speed) or (speed, distance). Bright points in the image spectrum correspond to a moving vehicle, and the amplitude of any point in the image spectrum is related to the vehicle's weight. To generate a vehicle speed spectrum, all receivers (locations where seismic traces are recorded) are traversed, and the time window within the trace is traversed. For each seismic trace recorded by a receiver within the time window, multiple adjacent receiver traces on one side of the current receiver and multiple adjacent receiver traces on the other side are selected. A linear motion correction (LMO) is applied to each selected receiver (location where a trace is recorded) on both sides of the current receiver. The correction time is calculated by dividing the distance offset between the two receivers by the specified vehicle speed. After subtracting the calculated correction time from all corresponding seismic traces, all corrected seismic traces can be superimposed to generate a single trace data associated with the specified vehicle speed. All speeds are cycled within the specified vehicle speed scanning range, and the above process is repeated. This generates the (time, speed) spectrum associated with the current receiver. The above process is performed for each receiver in a small time window centered on time t, and the rotated velocity spectrum (velocity, time) is placed at the corresponding receiver's position. The velocity spectra of all receivers at time t are combined within the region to create the final velocity spectrum (velocity, distance) along the road.

[0089] Figure 7 shows the vehicle speed spectrum generated from bandpass signal processing data. Linear motion overlay (LMO) and speed scanning are performed on the preprocessed data to obtain the vehicle speed spectrum. Figure 7(a) shows the bandpass signal processing data for LMO overlay. Figure 7(b) shows the time-domain vehicle speed spectrum at station locations, plotted with the signal. The vertical axis of the speed spectrum represents the vehicle speed. Individual bright spots (clouds) in the speed spectrum represent individual vehicles.

[0090] Generating a vehicle speed spectrum from signal-enhanced traffic data may include the following: ● Scans vehicle speeds within a certain range. ●Applies linear motion superposition to the specified scanning speed. ● Create a vehicle speed spectral image in the time domain.

[0091] Figure 8 shows a schematic diagram illustrating the generation of a vehicle speed spectrum from a single waveform of signal-enhanced traffic data.

[0092] As shown in Figure 8, the seismic data deformation curve is scanned with a virtual scan line. The intersections of the virtual scan line and the seismic data deformation curve are mapped to points in the vehicle speed spectrum.

[0093] In a specific case, the vehicle's speed and / or trajectory can be estimated by comparing the similarity of seismic data from different seismic recording devices. Similarity between seismic data from different seismic recording devices includes similarity between peak values ​​and / or speeds.

[0094] Vehicle association – Vehicle image points in different vehicle speed spectra are associated to reveal the movement of the same vehicle and derive a speed curve that follows the passage of time or distance for each vehicle. This is achieved by connecting all image points by applying a function to predict the position of the next vehicle, or by tracking the same image points in different vehicle speed spectra by applying machine learning methods. To accurately associate (or match) vehicles detected from vehicle speed spectra, the Hungarian method is used to recognize the best match and create a vehicle speed curve that follows the passage of time or distance. In addition to deriving vehicle speed curves from vehicle speed spectra, it is also possible to generate vehicle speed curves directly from recorded seismic data using machine learning methods.

[0095] The Hungarian algorithm is a combinatorial optimization algorithm that solves assignment problems in polynomial time. This algorithm finds the optimal pair of elements between two sets of elements to solve a linear assignment problem, while considering a specific cost or weight associated with each pair. For vehicle detection and association using traffic data, the Hungarian algorithm can be used to associate detected vehicles over time and / or at stations, optimizing the overall assignment of trajectories and vehicles and minimizing a certain cost function. This cost function includes factors such as differences in velocity and amplitude between vehicles.

[0096] In one embodiment, determining vehicle speed and / or vehicle travel path using a geophysical method may include the following: ● Utilize the time-domain vehicle speed spectrum of nearby stations. ● Find related information for the same vehicle. ●Connecting the same vehicles on a line (including the following): ○ Use sectioned linear lines. ○ Use piecewise polynomial curves. ○Smooth the curve at the recording station using a piecewise polynomial curve.

[0097] Figure 9 shows the process of extracting vehicle speed from vehicle travel paths and speed spectra. To obtain the speed and travel path of each individual vehicle, it is necessary to correlate the bright points (clouds) between stations.

[0098] As shown in Figure 9(b), the upward curve formed by connecting multiple points at different stations represents the vehicle's travel path. The vehicle's speed at a point on the upward curve can be determined from the slope at that point. Since the upward curve is not a straight line, the vehicle speed will differ at each station. Based on the slope at different points, a speed curve of the vehicle in motion can be obtained, as shown in Figure 9(c). Using the travel path and speed information of each individual vehicle, abnormal driving and traffic violations can be identified, and warnings for traffic accidents can be issued.

[0099] In one embodiment, determining vehicle speed from a vehicle speed spectrum and travel path may include the following: ● Use the vehicle speed in the speed spectrum of the recording station. ● Interpolates the speed of the same vehicle between recording stations. ● Differentiate the piecewise polynomial curve used and / or the single vehicle movement curve obtained using the piecewise polynomial curve, and smooth it at the recording station.

[0100] In one embodiment, it is possible to determine whether a vehicle is speeding by comparing its speed with a preset reference line. Figure 10 shows a schematic diagram for determining whether a vehicle is speeding. Specifically, if the vehicle speed curve is higher than the reference line, the vehicle is speeding. If the vehicle speed curve is lower than the reference line, the vehicle is not speeding.

[0101] In one embodiment, the vehicle type and weight are determined based on the vehicle's peak value in each of several seismic data fluctuation curves.

[0102] Vehicle Weight - The amplitude response related to a vehicle in seismic records is generated by the vehicle's weight, speed, number of axles, road conditions, soil conditions beneath the road, the sensor's instrument response, and the sensor-ground coupling. Vehicle weight plays a significant role in the signal amplitude. By calibrating or separating other factors, vehicle weight can be estimated from the amplitude response.

[0103] In the embodiment, determining the relative vehicle weight and vehicle type by extracting the vehicle's amplitude and frequency attributes (including data envelope and power spectrum) may include the following: ● Determine the vehicle type from the vibration signal. ● Vehicle weight is determined from earthquake amplitude and phase, vehicle speed, number of axles, data envelope, and power spectrum. ● Create a linear regression empirical relationship between vehicle weight, lane position, number of axles, and the vehicle's seismic response at the recording station. ● Create a nonlinear regression empirical relationship between vehicle weight, lane position, number of axles, and the vehicle's seismic response at the recording station.

[0104] Figure 11 shows the interpretation of the relationship between vehicle type, speed, weight and earthquake amplitude. Figure 11(a) shows the result of interpreting vehicle type from earthquake amplitude. Figure 11(b) shows the relationship between vehicle type and vehicle weight. Figure 11(c) shows the relationship between vehicle weight and earthquake amplitude obtained by linear regression.

[0105] In this embodiment, four types of vehicles (small, medium, medium-large, and large) are assumed. It can be seen that the amplitude increases as the vehicle size and weight increase.

[0106] Vehicle-generated seismic data includes information about vehicle speed, weight, type, and lane, allowing machine learning (ML) to extract this vehicle information. This disclosure introduces an intermediate ML step in applying machine learning, improving vehicle detection by converting seismic time-series data into a road heatmap, thereby generating real-time road condition information. This information can be used to detect vehicles and determine their location, lane, travel path, speed, and even type and weight.

[0107] In the embodiment, a recognition model for vehicle type, speed, and weight can be used to determine the vehicle type, speed, and weight.

[0108] In the example, a vehicle type, speed, and weight recognition model can be trained using multiple known vehicle types, speeds, weights, and corresponding earthquake data.

[0109] In this embodiment, the vehicle type, speed, and weight can be determined by inputting earthquake data into a vehicle type, speed, and weight recognition model.

[0110] Vehicles stopped on highways – Vehicles stopped on highways can pose significant risks and challenges to drivers and overall traffic. Sudden stops can occur due to emergencies, breakdowns, or unexpected circumstances, potentially creating dangerous situations. When a vehicle stops on a busy highway, it disrupts smooth traffic flow and increases the likelihood of rear-end collisions. In such situations, timely response from traffic management is crucial, and warnings to approaching drivers are even more important. When a vehicle stops on a highway, the continuity of vehicle vibration signals along the road being recorded is interrupted. Applying algorithms and machine learning methods to analyze the data can help detect sudden stops and notify traffic control departments. Reconstructing traffic volume in real time using seismic data can provide early warnings to police, enabling immediate action to avoid sudden stops on highways, thereby maintaining traffic safety and reducing the risk of dangerous collisions.

[0111] In this embodiment, it is possible to determine whether a vehicle is stopped based on its movement path. Figure 12 is a schematic diagram illustrating a vehicle stopping. Within the elliptical region, it can be seen that the seismic data at the center of the ellipse disappears at the next sensor position. This indicates that the vehicle stopped suddenly.

[0112] Figure 13 shows the travel path of a single vehicle along a road (indicated by a curve). This curve is tracked by recorded seismic data of the vehicle. From this selected vehicle's travel path, it is easy to obtain the vehicle's speed (curve slope) and analyze the driver's behavior to determine whether the driver was speeding.

[0113] Vehicle-corrected static corrections are small time corrections used to adjust for signal phase variations of the same vehicle in a particular set of traces, enabling high-quality superposition in speed spectrum calculations. Signal shifts in several adjacent sets of traces should be nearly linear. Variations may occur due to irregularities in receiver position, lane changes, and changes in vehicle speed. Such corrected static corrections may be calculated by correlating superimposed traces with the single traces before superposition, or they may be extracted directly from the data using machine learning methods.

[0114] Vehicle Warning System - Using traffic volume information obtained from earthquake sensor data, the system automatically issues a warning to all following vehicles when it detects a sudden drop in overall traffic speed or the speed of some or individual vehicles near an earthquake sensor. This warning can be displayed on roadside electronic information boards, as a warning light or siren, or on monitoring screens of traffic management departments. It can also send a warning message to nearby mobile phones or car navigation systems. The vehicle warning system is designed to avoid or minimize traffic accidents, especially in foggy weather or on curves with poor visibility.

[0115] As shown in Figure 13, the curve represents the vehicle's travel path. The speed of the corresponding vehicle can be determined from this travel path. A gradual decrease in speed indicates congestion ahead.

[0116] In this embodiment, pedestrians walking on the road can be detected based on earthquake data. Walking on highways is extremely dangerous due to high-speed traffic and the lack of suitable facilities for pedestrians. This highly dangerous act increases the likelihood of accidents and injuries for pedestrians and drivers. By analyzing earthquake data recorded along highways and using machine learning techniques, it becomes possible to issue early warnings about incidents and prevent potential, fatal accidents. When a person walks over an earthquake sensor, the sensor records their footprints, which are then recognized through real-time analysis.

[0117] Figure 14 shows a typical vehicle response during an earthquake. Figure 15 shows the seismic response when a person is walking on a highway (shown within the area). By analyzing the type of seismic signal, it is possible to distinguish between a person walking on the road and a vehicle driving in the lane.

[0118] Objects falling from moving vehicles – Objects falling from moving vehicles can pose a significant safety hazard on the road. Falling objects, such as unsecured cargo, debris, or accidentally dropped items, can cause accidents, property damage, or personal injury. Rapid detection of objects falling from moving vehicles is crucial for preventing accidents and ensuring road safety. By leveraging seismic data recorded along roads and employing advanced geophysical methods and machine learning algorithms, real-time traffic monitoring systems can improve road safety and reduce potential accidents by rapidly detecting objects falling from moving vehicles and notifying relevant authorities.

[0119] Natural Disaster Intensity Map - Earthquakes, storms, hurricanes, typhoons, and other natural disasters can cause ground shaking. By using the same roadside earthquake sensors, a real-time map of ground vibration intensity can be displayed and reported. This map allows authorities and the public to plan rescue operations and issue warnings.

[0120] Vibration-triggered video alerting – Information from earthquake sensors can complement video surveillance. When an earthquake sensor detects any unusual vibration, it can alert nearby video cameras in the same location. Such a triggering system avoids the need to process large amounts of camera and video data.

[0121] Similar to pedestrian detection, detecting objects that have fallen from moving vehicles can also be done using natural disaster intensity maps and video detection based on vibrations.

[0122] Figure 16 shows the velocity spectrum extrapolated from the time domain to the spatial domain. The vehicle velocity spectrum along the road shown in Figure 16(b) is obtained by extrapolating the time-domain vehicle velocity spectrum shown in Figure 16(a). Figure 16(b) intuitively illustrates the movement of a vehicle along the road. The image shown in Figure 16(b) provides a more intuitive and convenient view for traffic managers.

[0123] To obtain a vehicle speed spectrum along the road, it is necessary to extrapolate the vehicle speed spectrum generated by the superposition of LMOs in the time domain at each measurement station along the road.

[0124] To generate a vehicle speed spectrum along a road in the spatial domain, the following processes are involved: ● Use the time-domain vehicle speed spectrum of multiple recording stations. ●Extrapolate the velocity spectrum along the road in the spatial domain.

[0125] Predicting the next vehicle position - By utilizing the current vehicle speed spectrum (time, speed) of a receiver within a specific time window, it is possible to predict the vehicle speed spectrum of a virtual receiver near (forward or backward) the current receiver, assuming a constant vehicle speed. By using the current vehicle speed spectra (speed, distance) of multiple receivers at a specific point in time, it is possible to predict the vehicle speed spectrum at the next point in time (forward or backward) that is close to that point in time, assuming a constant vehicle speed. Both of the above predictions are achieved by calculating a time or distance offset using the speed values ​​within the speed spectrum.

[0126] Figure 16(b) shows the vehicle speed spectrum at a specific point in time across a series of receivers along the road. The vehicle's next position can be calculated from this graph.

[0127] One embodiment of this application provides a system for monitoring traffic volume through earthquake data. Figure 17 shows the structure of a system 1700 for monitoring traffic volume based on earthquake data, designed according to the embodiment. As shown in Figure 17, the system 1700 may include earthquake recording devices 1710-1 to 1710-n, a data processing module 1720, and an analysis module 1730.

[0128] In the embodiment, earthquake recording devices 1710-1 to 1710-n can be configured to acquire earthquake data. The data processing module 1720 is configured to acquire earthquake data deformation diagrams based on the earthquake data, and the earthquake data deformation diagrams include multiple earthquake data deformation curves from each earthquake recording device.

[0129] In the embodiment, the analysis module 1730 may be configured to analyze traffic volume based on seismic data deformation maps, and the analysis module 1730 may be configured to perform the operations described above.

[0130] Embodiments of this application provide a method for training a vehicle recognition model, the method of which may include the following: ● Obtain the first image, including the vehicle, from the video. ● The first image is converted into a second image, which represents the vehicle and the lane it is traveling in, and is used as a training label. ● Obtain the first earthquake data corresponding to the first image from the earthquake data. ●Train the vehicle recognition model based on the first earthquake data and the second image.

[0131] Figure 18 shows the process of converting a first video image to a second two-color image. Here, (a) shows a video recording of a road, with vehicles detected using image / video recognition technology, and (b) shows training labels generated using the vehicles detected from the video recording.

[0132] In the embodiment, the vehicle in the first video image may be captured using the YOLO algorithm.

[0133] The YOLO algorithm is a real-time target detection system that can simultaneously recognize, classify, and locate multiple objects within an image or video frame. For machine learning-based vehicle detection and its association with traffic data, the YOLO algorithm is used to generate labels, accurately capturing the actual situation of vehicles moving on roads from recorded video.

[0134] The most crucial part of training a machine learning model is obtaining labels. We want to record road video to obtain absolute truth information about passing vehicles (Figure 18(a)). To detect and capture vehicles in the video, we can use the YOLO algorithm for each frame of the video. This can then be converted into a bird's-eye view heatmap of the road (Figure 18(b)) for easier analysis.

[0135] In this example, the YOLO algorithm is used to detect and capture vehicles in recorded video, and the actual video recording is converted into a bird's-eye view heatmap cartoon of the road, which can then be used for machine learning-based vehicle detection.

[0136] In the examples, the visualization of vehicle data is used to generate machine learning training labels for a neural network, which includes the following: ● Generate labels from the intermediate steps of detecting and capturing vehicles in recorded video, and transform actual video recordings into a bird's-eye view heatmap cartoon of the road. ● Divide the earthquake data into small window frames and retrieve the earthquake data corresponding to the labels.

[0137] Using machine learning to generate heatmaps and detect vehicles includes the following: ● Generate heatmaps using any single-component data from a single-component sensor and / or multiple-component sensors to detect vehicles, types, weight, speed, and lanes. ● Generate road heatmaps using any combination of multiple component data to detect vehicles, types, weights, speeds, and lanes. Multiple seismic sensors are used to generate a heatmap, detecting vehicles, their type, weight, speed, and lane. ● Create a heatmap using any time series generated by the vehicle to detect the vehicle type, weight, speed, and lane. ● Use any combination of the above data to generate a heatmap and detect vehicles, type, weight, speed, and lane.

[0138] Figure 19 illustrates the use of machine learning for vehicle detection and interpretation. Here, (a) shows the earthquake data map used for vehicle detection and interpretation, and (b) shows the generated road heatmap indicating the predicted location of the vehicles.

[0139] To obtain vehicle seismic data corresponding to labels, the seismic data is divided into small window frames (as shown in Figure 19(a)). After preprocessing the seismic data (normalization, Fourier transform, etc.), a neural network is trained using the data-label pairs, and this network is used to generate a cartoonish visual representation of the road (see Figure 19(b)). Furthermore, the visual representation of the road is used to analyze vehicle positions and obtain relevant vehicle and road information.

[0140] Embodiments of this application provide a vehicle recognition method, which includes inputting seismic data to be recognized into a vehicle recognition model to determine the vehicle and the lane in which the vehicle is traveling, the vehicle recognition model being trained using the above method.

[0141] Another embodiment of this application provides a method for training a vehicle recognition model. This method includes: ● Select earthquake data from multiple time windows as labels. ● Recognize the maximum peak or Gaussian peak for each label. ● Determine the vehicle position corresponding to the maximum peak or Gaussian distribution peak. ● The vehicle recognition model is trained based on the location and its corresponding maximum peak or Gaussian distribution peak.

[0142] Figure 20 shows vehicle paths obtained directly from ML-spike (Figure 20(a)) or ML-curve (Figure 20(b)).

[0143] In addition to deriving vehicle trajectories solely from vehicle velocity spectra, another method is to directly extract vehicle trajectories from recorded vibration data, as shown in Figure 20. To obtain vehicle trajectories, it is crucial to generate ML-spike or ML-curve labels using the recorded vehicle vibration data. By training a neural network, either an ML-spike or ML-curve model is constructed, and these models can be used to detect vehicles in seismic data. In Figure 20(a), the bright dots represent vehicles recognized using the ML-spike model, and in Figure 20(b), the curves represent vehicles recognized by the ML-curve model. The Hungarian optimal matching method can be used to extract vehicle trajectories based on vehicles detected from the ML-spike or ML-curve model (represented as curves in both Figures 20(a) and (b)).

[0144] In the examples, detecting and associating vehicles with traffic data using machine learning and the Hungarian optimal matching algorithm involves one or more of the following steps:

[0145] 1) Preprocess the traffic data. 2) Create ML-spike labels (such as peaks) for the traffic data. Here, vehicle signals correspond to spikes at the same time on the recorded trajectory.

[0146] 3) Train an ML-spike model using traffic data as input and spikes as labels.

[0147] 4) Create ML Gaussian curve labels for the traffic data. Here, vehicle signals correspond to Gaussian signals at the same time on the recorded trajectory.

[0148] 5) Train an ML Gaussian curve model using traffic data as input and spikes as labels.

[0149] 6) Create ML curve labels for the user-defined curve function in the original or preprocessed data.

[0150] 7) Create an ML curve model for the original data or preprocessed data using a user-defined curve function.

[0151] 8) Run the model from step 3) to detect ML spikes and obtain a small data window around the corresponding spike in the traffic data or preprocessed traffic data.

[0152] 9) Run the ML model from step 5) to detect the ML curve and obtain a small data window around the peak corresponding to the ML curve in the traffic data or preprocessed traffic data.

[0153] 10) Run the ML model from step 7) to detect the ML curve, which is a user-defined curve, and obtain a small data window around the peak corresponding to the ML curve in the traffic data or preprocessed traffic data.

[0154] 11) Apply the Hungarian optimal matching algorithm to associate the vehicles detected in step 8), 9), or 10) with multiple adjacent stations and form vehicle movement curves at these stations.

[0155] One embodiment of this application provides a vehicle recognition method. This method includes the following: ●The earthquake data to be recognized is input into the vehicle recognition model to determine the vehicle's position. ●Based on its position, the vehicle's speed and / or vehicle movement path are determined, and the vehicle recognition model is trained using the method described above.

[0156] The neural network includes additional source data, such as station information. This source data, including temperature and road type, can improve vehicle detection results. Other domain adaptation methods and transfer learning techniques can also be used to improve inter-station results. Alternatively, if resources allow, one model can be created for each specific station.

[0157] road monitoring The recorded vehicle vibration data can be used not only to reconstruct traffic volume, but also to monitor geophysical conditions above and below the road, and to obtain images of underground seismic velocities and structures. By utilizing changes in geophysical conditions below the road, potential hazards can be warned about early and prevented from occurring.

[0158] Embodiments of this application provide a method for monitoring geological conditions beneath a road using seismic data, which includes the following: ● Obtain target earthquake data and reference earthquake data from earthquake recording equipment. ●Generate a target Green's function based on target earthquake data. ●Generate a reference Green's function based on reference earthquake data. ● To monitor the geological conditions beneath the road, a relative velocity change near the surface is generated based on a target Green's function and a reference Green's function.

[0159] Figure 21 is a schematic flowchart 2100 illustrating a method for monitoring geological conditions and structures beneath a road using seismic data, according to an exemplary embodiment of the present disclosure. As shown in Figure 21, the flowchart 2100 includes the following operations:

[0160] Operation 2110 allows you to obtain target earthquake data and reference earthquake data from earthquake recording equipment.

[0161] In operation 2120, a target Green's function can be generated based on target seismic data.

[0162] In operation 2130, a reference Green's function can be generated based on reference seismic data.

[0163] In operation 2140, a relative velocity change near the ground surface can be generated based on the target Green's function and the reference Green's function to monitor the geological conditions under the road.

[0164] In one embodiment, signal preprocessing, noise removal, resampling, data offset correction, and filtering can be performed on the target seismic data and the reference seismic data.

[0165] In one embodiment, a method for monitoring geological conditions under a road using seismic data includes one or more of the following. 1) The preprocessing workflow of traffic data includes the following.

[0166] a) Noise suppression; b) Resampling; c) Data drift correction; d) Multi-frequency range (f 1i , f 2i ) Bandpass filtering, provided that f0 < f 1i < f1; f0 < f 2i < f1; f0 and f1 are the lower and upper limits of the sensor frequency range; e) OneBit filtering; f) Median filtering; g) Root mean square (RMS) filtering; h) Smoothing filter; i) Fourier transform of traffic vibration data; j) Conjugate of the Fourier-transformed data; k) Inverse Fourier transform; l) Inverse Fourier transform of the conjugate Fourier-transformed data; m) Apply a tapered window to the data in the time domain, frequency domain, or both. 2) Extract bulk waves, elastic P-waves, S-waves, SH-waves, surface waves, coda waves, Rayleigh waves, and Love waves from traffic data, and generate a reference Green's function and a target Green's function. 3) Generating a Green function from traffic data includes the following:

[0167] a) Use the preprocessed data described in 1), or use any combination of a) through m) in 1).

[0168] b) Generate a reference Green's function by cross-correlating preprocessed data between sensor pairs in the time domain or frequency domain.

[0169] c) Cross-correlate preprocessed data between sensor pairs in the time domain or frequency domain to generate a target Green's function (or the current time Green's function).

[0170] d) Superimpose the cross-correlation results of (b) and (c) from (3). e) Weight the cross-correlation results of (b) and (c) from (3). 4) Use environmental noise imaging, a reference Green's function, and a target Green's function to create changes in relative velocity near the ground surface.

[0171] Background Noise Imaging - Earthquake sensors installed along roadsides can record vehicle movement and environmental noise. Environmental noise can be recorded particularly well when traffic is stopped, especially after midnight, and can be used to estimate the surface wave response between any two receivers. Correlative analysis of Green's functions or codas over time intervals yields changes in velocity in the near-ground region, which helps to image changes in road conditions.

[0172] Figure 22 shows the reference Green's function between pairs of stations (earthquake sensors). Figure 23 shows a comparison of Green's functions between reference time and target time.

[0173] Figure 24 shows an image of the relative velocity change of the underground medium beneath the road. According to Figure 24, changes in the geology beneath the road can be observed.

[0174] In addition to reconstructing real-time traffic volume and monitoring geophysical conditions beneath roads, seismic data recorded from vehicle vibrations is also used to monitor changes in road surface characteristics over time. By analyzing the response of road surface characteristics, it is possible to understand and predict future changes and provide real-time early warnings about road conditions. As a result, relevant departments can implement maintenance strategies and construction technologies that improve road safety and lifespan, ultimately benefiting all road users.

[0175] Road surface characteristics response refers to the dynamic behavior of a road under various conditions, including interaction with environmental factors, traffic load, and overall stress tolerance. Road surfaces are constantly subjected to vehicle loads, and it is crucial to understand how roads deteriorate, wear down, or experience stress over time due to environmental changes, temperature fluctuations, and extreme weather conditions (rain and snow). Seismic data recorded along roads can be used to monitor road surface conditions in real time and assess road integrity. By monitoring roads in real time and analyzing seismic data in the time and / or frequency domains, road service agencies can develop timely maintenance plans to ensure safe and durable infrastructure for travelers.

[0176] Rockfalls and Landslides on Slopes – Rockfalls and landslides on roads are always a challenge to safe traffic, especially in mountainous areas and geologically unstable regions. For example, sudden rockfalls and gradually progressing landslides can be serious hazards, potentially leading to vehicle damage, personal injury, or road closures. Installing seismic sensors along roads and analyzing the collected seismic data enables real-time monitoring and early warning, helping to manage these hazards, ensure safe passage, and minimize the impact of rockfalls and landslides on road users.

[0177] Avalanches – In the mountainous regions of the north, avalanches can rapidly block roads during winter, posing a risk of isolation, accidents, and even death or injury. Designing roadside infrastructure equipped with seismic sensors allows for continuous monitoring of roads. Analyzing seismic data collected from these sensors is crucial for minimizing avalanche risks and ensuring safer travel for everyone.

[0178] Embodiments of this application provide a method for monitoring road surface conditions using seismic data, which includes the following: ● Obtain target earthquake data and reference earthquake data from earthquake recording equipment. ● Based on reference earthquake data, calculate the first total energy of passing vehicles in the first frequency range within the reference time. ● Based on the target earthquake data, calculate the second total energy of passing vehicles in the first frequency range within the target time. ●The road surface conditions are monitored based on the first total energy and the second total energy.

[0179] Figure 25 is a schematic flowchart illustrating a method for monitoring road surface conditions using seismic data, according to an exemplary embodiment of the present disclosure.

[0180] As shown in Figure 25, flowchart 2500 can include the following operations. Operation 2501 allows the system to acquire target earthquake data and reference earthquake data from earthquake recording equipment.

[0181] Operation 2502 allows for the calculation of a first total energy of passing vehicles in a first frequency range within a reference time, based on reference earthquake data.

[0182] Operation 2503 allows for the calculation of a second total energy for vehicles passing through a first frequency range within a target time, based on target earthquake data.

[0183] In operation 2504, the road surface condition can be monitored based on the first total energy and the second total energy.

[0184] In one embodiment, a method of using seismic data to monitor road surface conditions can include one or more of the following steps. 1) The preprocessing of traffic data includes the following.

[0185] a. Noise suppression. b. Resampling.

[0186] c. Data drift correction. d. Multi-frequency range (f 1i , f 2i ) Band-pass filtering, provided that f0 < f 1i < f1; f0 < f 2i < f1; f0 and f1 are the lower and upper limits of the sensor frequency range.

[0187] e. OneBit filtering. f. Median filtering.

[0188] g. Root mean square (RMS) filtering. h. Smoothing filter.

[0189] i. Fourier transform of traffic vibration data. j. Conjugate of the Fourier-transformed data.

[0190] k. Inverse Fourier transform. l. Inverse Fourier transform of the conjugate Fourier-transformed data.

[0191] m. Apply a taper window to the data in the time domain, frequency domain, or both domains. 2) Use the preprocessed band-limited time-domain data to calculate the total energy of the passing vehicles in the selected frequency range within the reference time. 3) Using pre-processed, bandwidth-limited time-domain data, calculate the total energy of passing vehicles in a selected frequency range within the target (or current) time. 4) Using pre-processed, band-limited frequency domain data, calculate the total energy of passing vehicles in a selected frequency range within a reference time. 5) Using the pre-processed, band-limited frequency domain data, calculate the total energy of passing vehicles in the selected frequency range at the target (or current) time. 6) Calculate the ratio of the total energy of the different frequency bands in steps 2) and 3). 7) Calculate the ratio of the total energy of the different frequency bands in steps 4) and 5). 8) Perform weighted addition and / or subtraction on the results of steps 6) and 7).

[0192] Following the steps described above, Figure 26 shows the change in the road surface characteristic response over time.

[0193] According to one embodiment of this application, a method for predicting traffic accidents is provided, which may include the following: ● Obtain earthquake data from earthquake recording equipment. ● Obtain vehicle information based on earthquake data. ● Predicts traffic accidents based on vehicle information, current and past road information, driver information, and weather information.

[0194] Figure 27 shows a schematic flowchart illustrating a method for predicting traffic accidents according to an exemplary embodiment of the present disclosure.

[0195] As shown in Figure 27, flowchart 2700 can include the following operations. Operation 2701 allows you to acquire earthquake data from earthquake recording equipment.

[0196] Operation 2702 allows you to obtain information about vehicles based on earthquake data. Operation 2703 can predict traffic accidents based on vehicle information, current and past road information, driver information, and weather information.

[0197] According to one embodiment of this application, a method for training a traffic accident prediction model is provided, which may include the following: ● Acquire earthquake data, vehicle information, road information, driver information, and weather information at the time of the accident. ●Train a traffic accident prediction model based on earthquake data and this information.

[0198] According to one embodiment of this application, a method for predicting traffic accidents is provided, which may include the following: ●The earthquake data to be recognized is input into the traffic accident prediction model to predict traffic accidents and obtain a traffic accident score. Here, the traffic accident prediction model is trained using the method described above.

[0199] According to one embodiment of this application, a method for training a vehicle weight model is provided, which may include the following: ● Obtain peak vehicle values ​​from earthquake data. ● Train a traffic accident prediction model based on peak values ​​and vehicle weight.

[0200] According to one embodiment of this application, a method for obtaining vehicle weight is provided, which may include the following: ●The earthquake data to be recognized is input into the vehicle weight model to obtain the vehicle's weight, and this vehicle weight model is trained using the method described above.

[0201] Traffic accidents are predicted through machine learning or algorithms using reconstructed digital traffic volume information and historical data. This information includes one or more of the following: 1) The driver's driving experience or driving record. 2) The driver's accident history. 3) Abnormal driving records of the driver discovered from vibration data. 4) Abnormal road conditions. 5) Roads under construction or maintenance. 6) Roads with an accident history. 7) Adverse weather such as snow, storms, hail, rain, wind, fog, etc. 8) Poor visibility. 9) Traffic volume density. 10) The driving route of the vehicle. 11) Vehicle speed and speed change data. 12) The vehicle changes lanes. 13) Historical data of road vibrations at the time of accident occurrence.

[0202] By applying the above machine learning model and the detected vehicle data, real-time traffic volume can be monitored and driving can be assisted.

[0203] Vehicle movement curves can be directly generated from seismic data by manual pickup, automatic pickup, and machine learning pickup.

[0204] Using the directly picked up moving vehicle curve, the instantaneous vehicle speed can be calculated. The instantaneous vehicle speed is the slope of the vehicle's movement curve, that is, v = dx / dt.

[0205] The average speed of the vehicle can be calculated from the directly picked up moving vehicle curve. v = X / T, where X is the distance the vehicle moves along the road during the period T.

[0206] The average driving speed of the vehicle can be calculated from the directly picked up moving vehicle curve. The formula is v = X / T, where X represents the distance the vehicle moves along the road within the period T.

[0207] Providing traffic warnings using the reconstructed digital traffic information includes the following. 1) Analyze accident - traffic data and abnormal behaviors of vehicle movement curves or speed spectra to provide accident warnings. 2) Overweight vehicles - Determine the vehicle weight as described above to judge overweight vehicles. 3) The speed is calculated from the vehicle's movement path - vehicle-to-vehicle. v = d / t. Here, d is the distance between sensors, and t is the movement time of the vehicle between sensors. 4) Discriminate the speeding vehicle from the vehicle speed spectrum - Warn the vehicle that exceeds the speed limit based on the vehicle speed spectrum. 5) Discriminate the vehicle below the speed limit from the speed spectrum - Warn the vehicle that is moving very slowly or below the speed limit based on the vehicle speed spectrum. 6) The vehicle suddenly stops - Find the break point when the vehicle's movement curve disappears. From the vehicle speed spectrum, find a significant decrease in the speed spectrum (energy) value in the vehicle's moving direction or directly process the vibration data. 7) Vehicle falling objects - Use machine learning or analysis to detect objects that have fallen from a moving vehicle and collided with the ground through the signal pattern from the vibration data. 8) Irregular driving - Use the vehicle's movement curve, speed, and lane information to find irregular lane and speed changes within a time window. 9) Frequent lane changes - Analyze the lane information. 10) Detect pedestrians walking on the highway by signal matching between traffic signals recorded on the roadside and running signals. 11) Detect pedestrians walking on the highway using a machine learning model. 12) Detect abnormal road surface conditions and underground road surface conditions by post-processing the road surface feature response and underground structure changes. 13) Detect abnormal driving weather conditions by analyzing the response of road surface features and the change in the relative speed of the underground over time.

[0208] According to an embodiment of the present application, there is provided an electronic device including at least one processor and a memory communicably connected to the at least one processor, the memory storing instructions that, when executed by the at least one processor, enable the at least one processor to execute the above method.

[0209] Figure 28 shows an electronic device according to an exemplary embodiment of the present disclosure. As shown in Figure 28, the electronic device may include a memory 2801 and at least one processor 2802. The memory 2801 is communicated to the at least one processor 2802. The memory 2801 stores instructions that, when executed by the at least one processor, cause the at least one processor to perform the method described above.

[0210] According to one embodiment of the present application, a non-temporary computer-readable storage medium is provided that stores instructions and, when executed on a computer, causes the computer to perform the above method.

[0211] According to one embodiment of this application, a computer program product is provided which, when executed by a computer, includes a computer program having instructions that cause the computer to perform the above method.

[0212] In one embodiment of this application, an earthquake sensor can be calibrated using a calibration vehicle. Calibration vehicle – A vehicle used to calibrate roadside seismic sensors over time. This is the same or similar fixed-weight vehicle that passes the roadside seismic sensors at a specified speed. This process is repeated weekly, monthly, or at any other time interval, and the data is analyzed to detect any changes. If systemic amplitude changes occur, these changes are taken into account in the ground surface-compatible amplitude correction.

[0213] Road and Bridge Quality Assurance - Using a vehicle at a specified speed, seismic sensors are installed on roads or bridges, and any changes in complete waveform data recorded over a set period are detected. Large changes may indicate a change in the quality of the road or bridge.

[0214] Although the various block diagrams above show multiple elements, those skilled in the art will understand that the embodiments of this disclosure can be realized with or without one or more elements, or in combination with specific elements.

[0215] Although the steps described above are illustrated in order, those skilled in the art will understand that these steps may be performed in a different order, or that embodiments of the present disclosure may be realized without performing one or more of the above steps.

[0216] From the above, it can be understood that the electronic components of one or more systems or devices include, but are not limited to, at least one processing unit, memory, and a communication bus or communication device that connects various components, including memory, to the processing unit. This system or device may include, or have access to, various device-readable media. System memory includes device-readable storage media in the form of volatile memory and / or non-volatile memory (e.g., read-only memory (ROM) and / or random-access memory (RAM)). For example, system memory may also include, but are not limited to, operating systems, application programs, other program modules, and program data.

[0217] The embodiments of this application can be implemented as systems, methods, or program products. Accordingly, all embodiments can take the form of hardware embodiments or embodiments that include software (including firmware, resident software, microcode, etc.), which are collectively referred to herein as “circuits,” “modules,” or “systems.” Furthermore, embodiments can also take the form of program products embodied in at least one device-readable medium that includes device-readable program code.

[0218] This application allows the use of a combination of multiple device-readable storage media. In the context of this document, a device-readable storage medium ("storage medium") may be any tangible non-signaling medium that contains or can store a program consisting of program code configured to be used with or in combination with an instruction execution system, device, or device. For the purposes of this disclosure, storage media or devices should be interpreted as non-transient, i.e., signaling or propagating media are excluded.

[0219] Although the present invention has been described in relation to embodiments of the inventive concept shown in the drawings, those skilled in the art will understand that various changes and modifications are possible without departing from the technical spirit and essential features of the inventive concept. It will be obvious to those skilled in the art that various substitutions, modifications, and changes are possible without departing from the scope and spirit of the inventive concept. Accordingly, all such modifications shall be considered to fall within the scope of the present invention as defined in the claims.

Claims

1. A method of monitoring traffic volume using earthquake data, ○ Obtaining earthquake data from earthquake recording equipment, ○ Based on earthquake data, obtain earthquake data deformation diagrams that include multiple earthquake data deformation curves from each earthquake recording device, ○ Includes monitoring traffic volume based on earthquake data and / or earthquake data deformation maps. A method characterized by the following:

2. The seismic recording equipment is placed on one side, both sides, in the middle of the road, or any combination of these positions, and its spacing is constant or variable. The method according to claim 1.

3. The step of obtaining seismic data deformation maps based on seismic data includes signal enhancement for the seismic data. The method according to claim 1.

4. The step of obtaining seismic data deformation maps based on seismic data further includes, after signal enhancement, performing bidirectional traffic wave field separation or attenuation on the seismic data. The method according to claim 3.

5. The step of obtaining seismic data deformation maps based on earthquake data is: ○ After separating or attenuating the bidirectional traffic wave field, balance the seismic data using a correction curve, ○ Obtaining seismic data deformation maps based on balanced seismic data, Includes The method according to claim 4.

6. This further includes obtaining a vehicle speed spectrum to visually represent the vehicle's motion based on seismic data and / or seismic data deformation maps. The method according to claim 1.

7. This further includes determining the vehicle's speed and / or trajectory based on the similarity between seismic data from different seismic recording devices. The method according to claim 6.

8. This further includes determining whether a vehicle is speeding based on its speed. The method according to claim 7.

9. This includes determining the vehicle type and weight based on the vehicle's peak values ​​in multiple seismic data deformation curves, and further including The method according to claim 1.

10. This further includes determining whether a vehicle is stopped or not based on the vehicle's movement path. The method according to claim 7.

11. This further includes determining whether there are pedestrians on the road based on earthquake data. The method according to claim 1.

12. This further includes detecting whether an object fell from a vehicle based on earthquake data. The method according to claim 1.

13. A system that uses earthquake data to monitor traffic volume, ○ An earthquake recording device configured to acquire earthquake data, ○A data processing module that acquires earthquake data deformation maps, including multiple earthquake data deformation curves from each earthquake recording device, based on earthquake data. ○An analysis module configured to analyze traffic volume based on earthquake data deformation maps, including A system characterized by the following features.

14. The analysis module is further configured to perform the method according to any one of claims 2 to 11. The system according to claim 13.

15. A system that uses earthquake data to monitor traffic volume, ○ An earthquake recording device configured to acquire earthquake data, ○ Processor and, ○Includes memory connected to the processor and where instructions are stored, When the aforementioned instruction is executed, the processor, Based on earthquake data, we obtain earthquake data deformation diagrams that include multiple earthquake data deformation curves from each earthquake recording device. ○ Monitor traffic volume based on earthquake data deformation maps. A system characterized by the following features.

16. A method for training a vehicle recognition model, ○ Obtain a first image from the video that includes the vehicle, ○ Converting the first image into a second image representing a vehicle and its lane, which will be used as a training label, ○ Obtain the first group of earthquake data corresponding to the first image from the earthquake data, ○Train the vehicle recognition model based on the earthquake data from Group 1 and the images from Group 2, including Training methods for vehicle recognition models.

17. A vehicle recognition method, ○This includes inputting earthquake data to be recognized into a vehicle recognition model in order to determine the vehicle and its lane, Here, the vehicle recognition model is trained using the method described in claim 16. Vehicle recognition method.

18. A method for training a vehicle recognition model, ○ Picking up earthquake data as labels in multiple time windows, ○ Recognize the maximum peak or Gaussian distribution peak for each label, ○Determine the vehicle position corresponding to the maximum peak or Gaussian distribution peak, This includes training a vehicle recognition model based on location and the corresponding maximum peak or Gaussian-distributed peak. Training methods for vehicle recognition models.

19. A vehicle recognition method, ○In order to determine the vehicle's position, the earthquake data to be recognized is input into the vehicle recognition model, ○ This includes determining the vehicle's speed and / or path based on the said position, ○The vehicle recognition model is trained using the method of claim 18. Vehicle recognition method.

20. A method for monitoring the geological conditions beneath a road using seismic data, ○ Obtain target earthquake data and reference earthquake data from earthquake recording equipment, ○Generating a target Green's function based on target earthquake data, ○Generating a reference Green's function based on reference earthquake data, ○ Includes generating near-surface relative velocity changes based on a target Green's function and a reference Green's function in order to monitor the geological conditions beneath the road. A method characterized by the following:

21. ○Further includes performing signal preprocessing, denoising, resampling, data offset correction, and filtering on target earthquake data and reference earthquake data. The method according to claim 20.

22. Filtering includes multi-frequency range (f １ｉ , f ２ｉ ) band-pass filtering, where f ０ < f １ｉ < f １、 f ０ < f ２ｉ < f １ and f ０ and f １ are the lower and upper limits of the frequency range of the seismograph The method according to claim 21.

23. ○ Extract one or more of the following from the target earthquake data and the reference earthquake data: bulk waves, elastic P waves, S waves, SH waves, surface waves, coda waves, Rayleigh waves, and Love waves. ○ Further includes generating a reference Green's function and a target Green's function based on the extracted waves. The method according to claim 21.

24. Generating a change in relative velocity near the Earth's surface is ○ Includes generating near-surface relative velocity changes based on a target Green's function and a reference Green's function using an environmental noise imaging method. The method according to claim 21.

25. A method for monitoring road surface conditions using earthquake data, ○ Obtain target earthquake data and reference earthquake data from earthquake recording equipment, ○Calculate the first total energy of the first frequency range of passing vehicles at the reference time based on the reference earthquake data, ○ Based on the target earthquake data, calculate the second total energy of the first frequency range of passing vehicles at the target time, ○ Includes monitoring road surface conditions based on the first total energy and the second total energy. A method characterized by the following:

26. The calculation of the first total energy and the calculation of the second total energy are performed in the frequency domain or the time domain. The method according to claim 25.

27. The calculation of the first total energy and the calculation of the second total energy are performed in multiple frequency bands. The method according to claim 25.

28. Monitoring road surface conditions based on the first total energy and the second total energy is ○Calculate the ratio of the first total energy of multiple frequency bands, and calculate the ratio of the second total energy of multiple frequency bands, ○ Includes performing weighted addition and / or subtraction on the first ratio and the second ratio. The method according to claim 27.

29. A method for predicting traffic accidents, a. Obtaining earthquake data from earthquake recording equipment, b. Obtaining information about vehicles based on earthquake data, c. Predicting traffic accidents based on vehicle information, current road information, past road information, driver information, and weather information, including Methods for predicting traffic accidents.

30. A method for training a traffic accident prediction model, a. To acquire earthquake data, vehicle information, road information, driver information, and weather information at the time of the accident, b. Including earthquake data and training a traffic accident prediction model based on this information. Training methods for traffic accident prediction models.

31. A method for predicting traffic accidents, a. Inputting earthquake data to be recognized into a traffic accident prediction model trained using the method of claim 30 to predict traffic accidents and obtain a traffic accident score, including Methods for predicting traffic accidents.

32. A method for training a vehicle weight model, a. Obtaining peak vehicle values ​​from earthquake data, b. Training a vehicle weight model based on the peak value and the vehicle's weight, including Training methods for vehicle weight models.

33. A method for obtaining vehicle weight, a. Includes inputting earthquake data to be recognized into a vehicle weight model trained using the method of claim 32 to obtain the vehicle weight. How to obtain vehicle weight.

34. a. At least one processor, b. A memory connected to at least one processor, which, when executed by at least one processor, stores instructions causing at least one processor to perform the method according to any one of claims 1 to 11 or 16 to 33. electronic equipment.

35. When executed on a computer, it stores instructions that cause the computer to perform the method according to any one of claims 1 to 11 or 16 to 33. A non-temporary, computer-readable storage medium.

36. The computer program includes, when executed by a computer, instructions that cause the computer to perform the method described in any one of claims 1 to 11 or 16 to 33. Computer program products.

Images