A microseismic detection of shear wave velocity structure real-time imaging method, system and storage medium

Abstract

The application discloses a kind of microseismic detection transverse wave velocity structure real-time imaging method, system and storage medium, belong to geophysical prospecting technical field, method includes: with array center point as edge computing node, with other nodes except array center point as data acquisition node;Real-time microseismic data is obtained using the data acquisition node and sent to the edge computing node;Using the edge computing node, the transverse wave velocity structure is calculated according to the microseismic data and real-time imaging is completed;Wherein, the data acquisition node and edge computing node are according to the optimal task processing scheme obtained by solving pre-constructed real-time imaging task processing model to carry out task processing, and according to the load state prediction result of data acquisition node output by pre-constructed computing resource quantity search algorithm to carry out computing resource allocation scheduling.The method can be transverse wave velocity structure real-time imaging, and quickly and efficiently obtain underground velocity structure information.

Description

Technical Field

[0001] This invention relates to a method, system, and storage medium for real-time imaging of shear wave velocity structures by micro-motion detection, belonging to the field of geophysical exploration technology. Background Technology

[0002] Due to the dense urban buildings, electromagnetic interference and vibration interference are severe, making geological disaster investigation, engineering geological survey, and detection of hidden urban faults a persistent challenge for urban geophysical exploration.

[0003] Micro-motion detection technology based on environmental noise imaging can adapt to the complex interference environment of cities and has been widely used. Micro-motion detection technology is a new geophysical exploration technique that collects signals from natural sources and calculates the shear wave velocity structure in the underground medium to explore geological structures. This technology is characterized by its immunity to electromagnetic interference, environmental friendliness, high resolution, large detection range, and cost-effectiveness, offering unparalleled advantages over traditional geophysical exploration methods in densely populated areas such as towns and cities.

[0004] The microseismic detection method employs a "distributed acquisition-centralized retrieval" model. After the distributed seismographs complete the raw data acquisition, the data from each acquisition station is centrally retrieved and sent to a central server for post-processing and shear wave velocity-structure imaging. However, this approach faces numerous challenges in complex real-world applications. For example, exploration cycles can last for several weeks, and the acquisition of effective information lags significantly behind the acquisition time, resulting in low exploration efficiency.

[0005] Unlike traditional exploration methods, wireless seismic sensor networks interconnect data acquisition nodes, allowing control centers or other handheld terminals to remotely access all wireless nodes. This enables real-time monitoring of seismic data quality and improves exploration efficiency. However, using wireless seismic sensor networks also introduces several challenges, such as network bandwidth, transmission distance, and node lifespan. Furthermore, it still cannot overcome the lag in acquiring shear wave velocity and structural information.

[0006] With technological advancements, IoT devices have gradually acquired certain storage and computing capabilities, enabling them to store, analyze, and process more data. Edge computing, characterized by low latency, high bandwidth, real-time computing, and security, has emerged as a result. Edge computing consists of a large number of computing devices deployed at the edge of a network. These devices can collaborate to enable tasks to be processed closer to the terminal. Edge computing deploys lightweight computing close to the acquisition node, providing highly flexible and fast-response computing services to meet the real-time requirements of latency-sensitive tasks, and holds promise for achieving instantaneous transverse wave velocity structural imaging in micro-motion detection.

[0007] However, the application of edge computing to real-time shear wave velocity structure imaging still faces challenges such as latency-sensitive load allocation and computational resource optimization. First, micro-motion shear wave velocity structure imaging methods, exemplified by spatial autocorrelation, involve multiple tasks, including raw data preprocessing, cross-correlation coefficient acquisition, dispersion curve extraction, Bessel function fitting, and interpolation. These tasks occur randomly, leading to varying loads as tasks are generated and system runtime increases. However, the computational resources, computing power, and network transmission capabilities of edge computing nodes are limited, posing new challenges to task processing in edge computing systems. Second, in practical applications of micro-motion detection, different types of tasks from different nodes may be simultaneously offloaded to edge computing nodes. These tasks typically have heterogeneous computational resource requirements. Due to the limited computational resources and capabilities of edge computing nodes, they cannot simultaneously meet the resource allocation needs of these diverse task types. Summary of the Invention

[0008] The purpose of this invention is to provide a method, system, and storage medium for real-time imaging of shear wave velocity structure during micro-motion detection. During the micro-motion detection process, there is no need to transmit the original data back to the data center, and the shear wave velocity structure can be imaged in real time, thereby quickly and efficiently obtaining underground velocity structure information.

[0009] To achieve the above objectives, the present invention provides the following technical solution:

[0010] In a first aspect, the present invention provides a method for real-time imaging of transverse wave velocity structures by micro-motion detection, comprising:

[0011] The center point of the array is used as the edge computing node, and the other nodes are used as data acquisition nodes.

[0012] The data acquisition node acquires real-time microseismic data and sends it to the edge computing node;

[0013] The edge computing node is used to calculate and obtain the shear wave velocity structure based on the microseismic data and complete real-time imaging.

[0014] The data acquisition nodes and edge computing nodes perform task processing based on the optimal task processing scheme obtained by solving the pre-built real-time imaging task processing model, and allocate and schedule computing resources based on the load status prediction results of the data acquisition nodes output by the pre-built computing resource quantity search algorithm.

[0015] In conjunction with the first aspect, the tasks of the data acquisition nodes and edge computing nodes further include:

[0016] The data acquisition node performs a fast Fourier transform on the acquired raw data to obtain the power spectrum, and sends the power spectrum, the raw data, and the distance and angle between the data acquisition node and the edge computing node to the edge computing node.

[0017] The spatial autocorrelation coefficient is obtained by calculating the cross power spectrum based on the self-power spectrum, the original data, and the distance and angle between the data acquisition node and the edge computing node using the edge computing node;

[0018] The first type of zero-order Bessel function is fitted using the edge computing node based on the spatial autocorrelation coefficient, and the argument of the first type of zero-order Bessel function is solved.

[0019] The dispersion curve of the edge computing node is calculated and obtained based on the argument of the first type of zero-order Bessel function using the edge computing node;

[0020] Using the edge computing nodes, the shear wave velocity structure is obtained by inversion based on the dispersion curve, and real-time imaging is completed based on the shear wave velocity structure.

[0021] The edge computing node internally stores a data point table of the first type of zero-order Bessel function.

[0022] In conjunction with the first aspect, the formula for calculating the spatial autocorrelation coefficient is as follows:

[0023]

[0024] Where ρ is the spatial autocorrelation coefficient, r is the distance between the data acquisition node and the edge computing node, ω is the angular frequency, θ is the angle between the data acquisition node and the edge computing node, S(r,θ,ω) is the cross-power spectrum between the center point of the observation array and any node, S0(0,ω) is the autopower spectrum of the edge computing node, and S r (r,ω) represents the self-power spectrum of any data acquisition node in the observation array, and Re[·] represents the real part.

[0025] In conjunction with the first aspect, the formula for calculating the argument of the first kind of zeroth-order Bessel function is as follows:

[0026] x=2πf(i)r / Vr(f(i))

[0027] Where x is the argument of the zeroth-order Bessel function of the first kind, f(i) is the i-th frequency, r is the distance between the data acquisition node and the edge computing node, and Vr(f(i)) is the Rayleigh wave phase velocity with the i-th frequency f(i).

[0028] In conjunction with the first aspect, the calculation formula for the shear wave velocity structure is further as follows:

[0029]

[0030] Among them, V x,i Let V be the transverse wave velocity in the i-th period. r,i Let V be the Rayleigh wave speed of the i-th period. r,i-1 Let t be the Rayleigh wave speed of the (i-1)th period. i For the i-th period, t i-1 This is the (i-1)th cycle.

[0031] In conjunction with the first aspect, the real-time imaging task processing model further includes an objective function and constraints, wherein the objective function is:

[0032]

[0033] Where, f(i,j)(t) * The minimum latency for edge computing node j to process task i in time slot t is the optimal solution to the objective function, i.e., the optimal task processing scheme. The total latency for edge computing node j to process task i. For task i, calculate the queuing delay of node j at the edge in time slot t. Q j (t+1) represents the length of the task queue for edge computing node j in time slot t+1. Q j (t) represents the task queue length of edge computing node j in time slot t, Q j (0) = 0, μ represents the number of tasks offloaded to edge computing node j in time slot t. j (t) represents the number of tasks processed by edge computing node j in time slot t. κ is the instruction length of task n in time slot t. (i,m) Let P be the set of tasks entering edge computing node j in time slot t, m be the total number of tasks entering edge computing node j in time slot t, and P be the total number of tasks entering edge computing node j in time slot t. j Let j be the instruction length of the task processed by edge computing node j in each time slot. The total latency for task i to be transmitted from edge computing node j to data acquisition node k in time slot t. The network congestion delay for task i to be transmitted from edge computing node j to data acquisition node k in time slot t. Let be the number of remaining tasks that edge computing node j will transmit to data acquisition node k in time slot t, where is the network congestion status in time slot t. If there is no network congestion in time slot t, then there is no network congestion in time slot t; otherwise, there is network congestion in time slot t. Let n be the data size of task n. Let j be the set of tasks that edge computing node j has already transmitted to data acquisition node k in time slot t. The network bandwidth from edge computing node j to data acquisition node k. The transmission delay of task i from edge computing node j to data acquisition node k in time slot t. Let i be the data size of task i. Let $\mathbf{j}$ be the propagation delay from edge computing node $j$ to data acquisition node $k$.

[0034] In conjunction with the first aspect, the real-time imaging task processing model further includes an objective function and constraints, wherein the constraints are:

[0035]

[0036] Among them, Q j (t) represents the task queue length of edge computing node j in time slot t, R is the total number of edge computing nodes, E[·] is the mathematical expectation, T is the total number of time slots, and Qw j (t) represents the load queue length of edge computing node j in time slot t.

[0037] In conjunction with the first aspect, further, the data acquisition nodes and edge computing nodes perform computing resource allocation and scheduling based on the load status prediction results of the data acquisition nodes output by the pre-built computing resource quantity search algorithm, including:

[0038] Initially, the data acquisition node and the edge computing node operate in the self-power spectrum calculation stage, and the data acquisition node sends its load status to the edge computing node;

[0039] Within each time slot, the edge computing node executes a pre-built algorithm to search for the number of computing resources based on the load status of the previously received data acquisition node, and obtains the load status prediction result of the data acquisition node.

[0040] Within each time slot, the edge computing node allocates and schedules computing resources based on the load status prediction results of the data acquisition node.

[0041] In a second aspect, the present invention provides a real-time imaging system for detecting transverse wave velocity structures by micro-motion, including a processor and a storage medium;

[0042] The storage medium is used to store instructions;

[0043] The processor is configured to operate according to the instructions to perform the steps of the method according to any one of the first aspects.

[0044] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in any of the first aspects.

[0045] Compared with the prior art, the beneficial effects of the present invention are:

[0046] The real-time imaging method for shear wave velocity structure in micro-motion detection provided by this invention utilizes the low latency, high mobility, and distributed computing characteristics of edge computing to achieve integrated real-time imaging of shear wave velocity structure in micro-motion detection, encompassing sensing, processing, computation, and imaging. This approach balances the advantages of micro-motion detection, such as safety, environmental friendliness, ease of construction, and high signal-to-noise ratio. The real-time imaging method for shear wave velocity structure in micro-motion detection provided by this invention can acquire shear wave velocity structure information online in real time using only data acquisition nodes and edge computing nodes within a wireless seismic sensor network, offering a new approach to intelligent passive source seismic detection. Attached Figure Description

[0047] Figure 1 This is a schematic diagram of the structure of a data acquisition node provided in an embodiment of the present invention;

[0048] Figure 2 This is a schematic diagram of the structure of an edge computing node provided in an embodiment of the present invention;

[0049] Figure 3 This is a flowchart of the real-time imaging method for transverse wave velocity structures provided in an embodiment of the present invention;

[0050] Figure 4 This is a schematic diagram of system delay analysis in a single micro-motion detection area provided in an embodiment of the present invention;

[0051] Figure 5 This is a schematic diagram of the real-time parallel queuing model provided in an embodiment of the present invention;

[0052] Figure 6 This is a flowchart of a method for data processing using edge computing nodes provided in an embodiment of the present invention. Detailed Implementation

[0053] The technical solution of this application will be further described in detail below with reference to specific embodiments.

[0054] The embodiments of this application are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application. Unless otherwise specified, the embodiments of this application and the technical features within them can be combined with each other.

[0055] Example 1:

[0056] The real-time imaging method for micro-motion detection of shear wave velocity structure provided in this embodiment specifically includes the following steps:

[0057] Step 1: Use the center point of the array as the edge computing node, and use the other nodes as data acquisition nodes.

[0058] In this embodiment, based on the microseismic detection system which is mainly based on the spatial autocorrelation method, according to the layout structure of the microseismic detection array, the center point of the array is used as the edge computing node, and the other nodes are used as data acquisition nodes. The data acquisition nodes are used to acquire real-time microseismic data, and the edge computing nodes are used to complete the real-time imaging of the shear wave velocity structure and the imaging quality control.

[0059] Figure 1 The schematic diagram of the data acquisition node provided in this embodiment includes a connected detection unit, a high-precision data acquisition unit, a data recording unit, a power management unit, and a wireless data quality monitoring unit. The detection unit includes a connected detector and a tilt detection module. The high-precision data acquisition unit integrates a single-chip test signal generator, including a connected passive filter network, impedance matching circuit, ADC, and DAC, enabling data acquisition, filtering, analog-to-digital conversion, and channel testing of weak seismic signals. The CPU of the data recording unit is equipped with a GPS timing unit, an Ethernet unit, an SD card storage unit, and a status indication unit, enabling data storage and retrieval. Each data acquisition node in the detection area needs to communicate with the control center; therefore, each data acquisition node is equipped with a wireless ZigBee module for wireless transmission of status data with edge computing nodes. This embodiment uses low-power ZigBee to construct the wireless data quality monitoring unit. The power management unit uses a rechargeable lithium battery and power management and high-performance DC-DC conversion circuits to power the entire system.

[0060] Figure 2 This is a schematic diagram of the edge computing node provided in this embodiment. The edge computing node adopts the ZYNQ7020SOC from the XILINX series, which has high-speed data processing capabilities and low-latency computing and communication performance. It uses ARM+FPGASOC technology to integrate a dual-core ARM Cortex-A9 and FPGA programmable logic on a single chip. The two communicate with each other through an AXI interface. In addition, the core board contains two high-speed DDR3 SDRAM chips with a total of 1GB and one 256MB QSPI FLASH chip, providing the hardware foundation for data preprocessing, computing, load prediction, computing resource scheduling and real-time imaging functions in the edge computing process.

[0061] Edge computing nodes not only need to perform data acquisition, but also need to handle information exchange with multiple data acquisition nodes and meet real-time imaging requirements. This real-time imaging system uses the AXI VDMA IP core to cache and read real-time image data and display it on a VGA monitor, forming a complete real-time imaging processing platform.

[0062] Step 2: Acquire real-time microseismic data using data acquisition nodes and send it to edge computing nodes;

[0063] Step 3: Utilize edge computing nodes to calculate and obtain the shear wave velocity structure based on microseismic data and complete real-time imaging.

[0064] In this embodiment, as Figure 3 As shown, the process of acquiring real-time microseismic data using data acquisition nodes and sending it to edge computing nodes, and then using edge computing nodes to calculate shear wave velocity structures based on the microseismic data and complete real-time imaging, specifically includes the following steps:

[0065] Step 1: Use the data acquisition node to perform a fast Fourier transform on the acquired raw data to obtain the self-power spectrum, and send the self-power spectrum, raw data, and distance and angle between the data acquisition node and the edge computing node to the edge computing node;

[0066] After the data acquisition node completes the real-time microseismic data acquisition task, within the same time period, the self-power spectrum S is calculated using Fast Fourier Transform within the data acquisition node. r (r,ω), and then send some of the original data, the distance r between the data acquisition node and the edge computing node, the angle θ and other parameters to the edge computing node.

[0067] Step 2: Calculate the cross-power spectrum using edge computing nodes based on the auto-power spectrum, raw data, and the distance and angle between the data acquisition node and the edge computing node to obtain the spatial autocorrelation coefficient;

[0068] The formula for calculating the spatial autocorrelation coefficient is:

[0069]

[0070] Where ρ is the spatial autocorrelation coefficient, r is the distance between the data acquisition node and the edge computing node, ω is the angular frequency, θ is the angle between the data acquisition node and the edge computing node, S(r,θ,ω) is the cross-power spectrum between the center point of the observation array and any node, S0(0,ω) is the autopower spectrum of the edge computing node, and S r (r,ω) represents the self-power spectrum of any data acquisition node in the observation array, and Re[·] represents the real part.

[0071] Step 3: Use edge computing nodes to fit the first kind of zero-order Bessel function based on the spatial autocorrelation coefficient, and solve for the argument of the first kind of zero-order Bessel function;

[0072] The edge computing node stores a data point table JM(x) of the first kind of zero-order Bessel function. The edge computing node fits the r ~ ρ(r) curve of the first kind of zero-order Bessel function with the spatial autocorrelation coefficient to fit the edge computing node and the data acquisition node at different distances, and solves the argument x of the first kind of zero-order Bessel function.

[0073] The formula for calculating the argument of a zeroth-order Bessel function of the first kind is:

[0074] x=2πf(i)r / Vr(f(i))

[0075] Where x is the argument of the zeroth-order Bessel function of the first kind, f(i) is the i-th frequency, r is the distance between the data acquisition node and the edge computing node, and Vr(f(i)) is the Rayleigh wave phase velocity with the i-th frequency f(i).

[0076] Step 4: Calculate and obtain the dispersion curve of the edge computing node based on the arguments of the first-order zero-order Bessel function using the edge computing node;

[0077] Step 5: Using edge computing nodes, the shear wave velocity structure is obtained by inversion based on the dispersion curve, and real-time imaging is completed based on the shear wave velocity structure.

[0078] Based on the given initial model parameters (number of layers, shear wave velocity of each layer, layer thickness), the shear wave velocity structure below the survey area is obtained by inversion using dispersion curves. According to the obtained dispersion curves of the subsurface medium, the calculation formula for the shear wave velocity structure is as follows:

[0079]

[0080] Among them, V x,i V is the transverse wave velocity in the i-th period. r,i Let V be the Rayleigh wave speed of the i-th period. r,i-1 Let t be the Rayleigh wave speed of the (i-1)th period. i For the i-th period, t i-1 This is the (i-1)th cycle.

[0081] Steps ① to ⑤ are completed in the data acquisition node and the edge computing node. Steps ① to ⑤ are also the tasks of the data acquisition node and the edge computing node.

[0082] In this embodiment, the data acquisition node and the edge computing node perform task processing according to the optimal task processing scheme obtained by solving the pre-built real-time imaging task processing model, and perform computing resource allocation and scheduling according to the load status prediction results of the data acquisition node output by the pre-built computing resource quantity search algorithm.

[0083] The real-time imaging task processing model includes an objective function and constraints. The objective function is:

[0084]

[0085] Where, f(i,j)(t) * The minimum latency for edge computing node j to process task i in time slot t is the optimal solution to the objective function, i.e., the optimal task processing scheme. The total latency for edge computing node j to process task i. For task i, calculate the queuing delay of node j at the edge in time slot t. Q j (t+1) represents the length of the task queue for edge computing node j in time slot t+1. Q j (t) represents the task queue length of edge computing node j in time slot t, Q j (0) = 0, N j e (t) represents the number of tasks unloaded to edge computing node j in time slot t, μ j (t) represents the number of tasks processed by edge computing node j in time slot t. κ is the instruction length of task n in time slot t. (i,m) Let P be the set of tasks entering edge computing node j in time slot t, m be the total number of tasks entering edge computing node j in time slot t, and P be the total number of tasks entering edge computing node j in time slot t. j Let j be the instruction length of the task processed by edge computing node j in each time slot. The total latency for task i to be transmitted from edge computing node j to data acquisition node k in time slot t. The network congestion delay for task i to be transmitted from edge computing node j to data acquisition node k in time slot t. Let be the number of remaining tasks that edge computing node j will transmit to data acquisition node k in time slot t, where is the network congestion status in time slot t. If there is no network congestion in time slot t, then there is no network congestion in time slot t; otherwise, there is network congestion in time slot t. Let n be the data size of task n. Let j be the set of tasks that edge computing node j has already transmitted to data acquisition node k in time slot t. The network bandwidth from edge computing node j to data acquisition node k. The transmission delay of task i from edge computing node j to data acquisition node k in time slot t. Let i be the data size of task i. Let $\mathbf{j}$ be the propagation delay from edge computing node $j$ to data acquisition node $k$.

[0086] The real-time imaging task processing model includes an objective function and constraints, the constraints being:

[0087]

[0088] Among them, Q j (t) represents the task queue length of edge computing node j in time slot t, R is the total number of edge computing nodes, E[·] is the mathematical expectation, T is the total number of time slots, and Qw j (t) represents the load queue length of edge computing node j in time slot t.

[0089] Based on the definition of a discrete-time system, the task processing of the edge computing system for micro-motion detection using the spatial autocorrelation method in this embodiment can be abstracted into a discrete-time system.

[0090] To achieve real-time imaging of shear wave velocity structure in steps ① to ⑤ of edge computing nodes, a real-time imaging task processing model is constructed for each task of real-time imaging of shear wave velocity structure, transforming the real-time imaging task processing problem of shear wave velocity structure into a task processing problem of edge computing nodes.

[0091] In the micro-motion detection region, the task processing problem of real-time imaging of shear wave velocity structure can be divided into reducing system latency and allocating computing resources.

[0092] In the micro-motion monitoring network constructed in this embodiment, since the network transmission distance between nodes is relatively short and they are generally not connected through a backbone network, the network is single-hop and has limited network bandwidth.

[0093] Figure 4 This is a schematic diagram of system latency analysis in a single micro-motion detection area provided in this embodiment. The system latency includes network latency and computing service latency. The network latency includes network congestion latency, information transmission latency, and propagation latency, while the imaging computing service latency includes computing latency and task service latency.

[0094] Network congestion occurs when the amount of data being transmitted exceeds the network bandwidth. This represents the propagation delay from edge computing node j to data acquisition node k, in the form of... This represents the number of remaining tasks that edge computing node j will transmit to data acquisition node k in time slot t, where t represents the network congestion status of time slot t. If there is no network congestion in time slot t, then there is no network congestion in time slot t; otherwise, there is network congestion in time slot t. Therefore, the network congestion delay for task i to be transmitted from edge computing node j to data acquisition node k in time slot t is... The calculation formula is:

[0095]

[0096] in, Let n be the data size of task n. Let j be the set of tasks that edge computing node j has already transmitted to data acquisition node k in time slot t. Let $\mathbfl$ be the network bandwidth from edge computing node j to data acquisition node k.

[0097] The transmission delay of task i from edge computing node j to data acquisition node k in time slot t. The calculation formula is:

[0098]

[0099] in, Let be the data size for task i.

[0100] Therefore, the total delay of task i being transmitted from edge computing node j to data acquisition node k in time slot t is as follows. The calculation formula is:

[0101]

[0102] In the real-time imaging mission of transverse wave velocity structure, after the mission enters the computing device, it is first buffered into the task queue of the edge computing node to wait. The edge computing node then allocates computing resources to the tasks in the queue according to the queuing order. Once the task obtains computing resources, it leaves the queue and is processed.

[0103] Since each sensing node is generally composed of lightweight computing devices, it only needs to manage its own task scheduling and computing resource allocation. The task scheduling and computing resource allocation process of the sensing node can meet the requirements of real-time performance.

[0104] Edge nodes need to interact and perform calculations with each sensing node within the micro-motion array. The allocation and scheduling of their computing resources is highly complex, resulting in a long delay in adjusting the computing resources of edge nodes.

[0105] To address the high complexity of task scheduling and computing resource allocation algorithms for edge computing nodes, this embodiment uses a real-time parallel queuing model to model the task processing of edge computing nodes.

[0106] Figure 5 This is a schematic diagram of the real-time parallel queuing model provided in this embodiment. The real-time parallel queuing model uses multiple parallel task queues to queue and buffer tasks with different computing resource configuration requirements. The task sorting and buffering scheduling algorithm can be executed in parallel in each task queue, so that the edge computing node task scheduling and computing resource allocation algorithms can be executed in parallel in queues of different types of tasks, reducing the complexity of the edge computing node task scheduling and computing resource allocation algorithms.

[0107] Q j (t) represents the task queue length of edge computing node j in time slot t, Q j (0) = 0. Therefore, the task queue length Q of edge computing node j in time slot t+1 is... j The formula for calculating (t+1) is:

[0108]

[0109] Where, N j e (t) represents the number of tasks unloaded to edge computing node j in time slot t, μ j (t) represents the number of tasks processed by edge computing node j in time slot t.

[0110] by κ represents the instruction length of task n in time slot t. (i,m) P represents the set of tasks entering edge computing node j in time slot t. j Let be the instruction length of the task processed by edge computing node j in each time slot, and be the queuing delay of task i at edge computing node j in time slot t when it is offloaded to edge computing node j. The calculation formula is:

[0111]

[0112] Therefore, the total latency of edge computing node j processing task i The calculation formula is:

[0113]

[0114] Considering that the imaging system is a task-non-loss system, this embodiment solves the problem of minimizing task service latency at edge computing nodes by constructing a real-time imaging task processing model that performs parallel task ordering. The objective function of the real-time imaging task processing model is:

[0115]

[0116] Where, f(i,j)(t)* The minimum delay for edge computing node j to process task i in time slot t is the optimal solution of the objective function, i.e., the optimal task processing scheme.

[0117] To minimize latency while ensuring stable data transmission, the task latency issue of edge computing nodes requires additional constraints. Specifically, the task queue length of edge computing node j in time slot t must remain stable, and simultaneously, the load queue length of edge computing node j in time slot t must also remain stable for the system to remain stable. The constraints for the real-time imaging task processing model are:

[0118]

[0119] Among them, Q j (t) represents the task queue length of edge computing node j in time slot t, R is the total number of edge computing nodes, E[·] is the mathematical expectation, T is the total number of time slots, and Qw j (t) represents the load queue length of edge computing node j in time slot t.

[0120] To avoid severe queuing delays on edge computing nodes due to overloading, the task queues and load queues of edge computing nodes should remain stable. Edge computing nodes should adaptively allocate and schedule computing resources based on load changes, thereby ensuring the quality of service while avoiding oversupply or undersupply of computing resources.

[0121] This embodiment constructs an adaptive real-time imaging load prediction method based on the LSTM model. The method classifies the network load data according to the changing trend and characteristics of the input network load data, and adaptively schedules the input load data to the LSTM load prediction model that matches the type for prediction based on the classification results.

[0122] This embodiment designs a computational resource quantity search algorithm based on the load prediction results of the real-time imaging task of shear wave velocity structure. After all operations within a time slot are completed, each edge computing node immediately broadcasts its current load status information to all data acquisition nodes. Based on this operation, each node in the detection area needs to follow the following operation procedure:

[0123] (1) At the initial moment, the data acquisition node and the edge computing node run to the self-power spectrum calculation stage, and the data acquisition node sends its load status to the edge computing node;

[0124] (2) In each time slot, the edge computing node executes a pre-built computing resource quantity search algorithm based on the load status of the data acquisition node received last time to obtain the load status prediction result of the data acquisition node.

[0125] (3) Within each time slot, the edge computing node allocates and schedules computing resources based on the load status prediction results of the data acquisition node.

[0126] Once an edge computing node receives a new load status, it can update and restart the prediction of the load status of each data acquisition node within the detection area.

[0127] Within each time slot, each edge computing node manages its own load allocation based on its own load status and the predicted load status of each data acquisition node within the detection area.

[0128] This computing resource quantity search algorithm, while ensuring task service quality and system stability, obtains the required computing resources in advance based on the load status prediction results of data acquisition nodes, thereby realizing on-demand allocation and elastic scheduling of computing resources.

[0129] Figure 6 This is a flowchart of a method for data processing using edge computing nodes provided in this embodiment. By acquiring the imaging load, matching computing resources, and displaying the acquired shear wave velocity structure, real-time imaging is achieved at the edge computing node. The data of each sensor is visualized while monitoring the sensor status. The sensor nodes form a network, and any device in the network can access the raw data. Autonomous real-time imaging of the shear wave velocity structure is performed within the edge computing node.

[0130] The first layer displays the storage of data. A MySQL database is deployed in each sensor node. The database records the raw streaming data, cross-correlation results, and generated velocity maps. At the same time, it saves the initial model parameters (number of layers, wave velocity of each layer, and layer thickness) of the detection area for use in the imaging stage.

[0131] The second layer, serving as the data correlation stage, involves the continuous cross-correlation operation of the sensor node data streams and the transmission of the resulting data to the edge computing nodes. This ensures two important characteristics of the imaging system: first, the continuity of system operation; and second, the ability to adjust tomographic parameters and generate velocity maps at any time if needed.

[0132] The third layer is the imaging stage, which is executed after the initial setting time, or it can be defined in the configuration file.

[0133] Based on this data processing platform, an improved load prediction algorithm and computing resource scheduling algorithm are implemented. By acquiring the imaging task load in real time and continuously matching computing resources, the obtained shear wave velocity structure is displayed in real time, realizing a passive source seismic noise real-time imaging system based on edge computing nodes.

[0134] The real-time imaging method for shear wave velocity structure detection provided in this embodiment utilizes the low latency, high mobility, and distributed computing characteristics of edge computing to achieve integrated real-time imaging of shear wave velocity structure detection based on "sensing-processing-computation-imaging," thus balancing the advantages of micro-motion detection, such as safety, environmental friendliness, convenient construction, and high signal-to-noise ratio. The real-time imaging method for shear wave velocity structure detection provided by this invention can acquire shear wave velocity structure information online in real time using only data acquisition nodes and edge computing nodes within a wireless seismic sensor network, providing a new approach for intelligent passive source seismic detection.

[0135] Example 2:

[0136] This embodiment provides a system, including a processor and a storage medium;

[0137] Storage media are used to store instructions;

[0138] The processor is used to perform operations according to instructions to execute the steps of the method in Embodiment 1.

[0139] Example 3:

[0140] This embodiment provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method in Embodiment 1.

[0141] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0142] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0143] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0144] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0145] The above are merely preferred embodiments of this application. It should be noted that those skilled in the art can make several improvements and modifications without departing from the technical principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for real-time imaging of transverse wave velocity structures by micro-motion detection, characterized in that, include: The center point of the array is used as the edge computing node, and the other nodes are used as data acquisition nodes. Real-time microseismic data is acquired using data acquisition nodes and sent to edge computing nodes; Edge computing nodes are used to calculate and obtain shear wave velocity structures based on microseismic data and complete real-time imaging. Among them, the data acquisition nodes and edge computing nodes perform task processing according to the optimal task processing scheme obtained by solving the pre-built real-time imaging task processing model, and perform computing resource allocation and scheduling according to the load status prediction results of the data acquisition nodes output by the pre-built computing resource quantity search algorithm. The tasks of data acquisition nodes and edge computing nodes include: The data acquisition node performs a fast Fourier transform on the raw data it acquires to obtain the self-power spectrum, and then sends the self-power spectrum, the raw data, and the distance and angle between the data acquisition node and the edge computing node to the edge computing node. The spatial autocorrelation coefficient is obtained by calculating the cross power spectrum based on the self power spectrum, the original data, and the distance and angle between the data acquisition node and the edge computing node using the edge computing node; By using edge computing nodes to fit the first-order zero Bessel function based on the spatial autocorrelation coefficient, the argument of the first-order zero Bessel function is solved. The dispersion curve of the edge computing node is calculated and obtained by using the arguments of the first-order zero-order Bessel function of the edge computing node. Using edge computing nodes, the shear wave velocity structure is obtained by inversion based on the dispersion curve, and real-time imaging is completed based on the shear wave velocity structure. The edge computing nodes internally store a data point table of type I zero-order Bessel functions.

2. The real-time imaging method for detecting transverse wave velocity structures by micro-motion as described in claim 1, characterized in that, The formula for calculating the spatial autocorrelation coefficient is as follows: ； in, This is the spatial autocorrelation coefficient. This refers to the distance between the data acquisition node and the edge computing node. Angular frequency, The angle between the data acquisition node and the edge computing node. The cross-power spectrum between the center point of the observation array and any node. The self-power spectrum of the edge computing node. The power spectrum of any data acquisition node in the observation array. To take the real part.

3. The real-time imaging method for detecting transverse wave velocity structures by micro-motion as described in claim 1, characterized in that, The formula for calculating the argument of the first kind of zeroth-order Bessel function is: ； in, Let be the argument of a zeroth-order Bessel function of the first kind. For the first One frequency, This refers to the distance between the data acquisition node and the edge computing node. For the first The frequency is Rayleigh wave phase velocity.

4. The real-time imaging method for detecting transverse wave velocity structures by micro-motion as described in claim 1, characterized in that, The formula for calculating the shear wave velocity structure is as follows: ； in, For the first The transverse wave velocity per cycle, For the first Rayleigh wave speed per cycle For the first Rayleigh wave speed per cycle For the first One cycle, For the first One cycle.

5. The real-time imaging method for detecting transverse wave velocity structures by micro-motion as described in claim 1, characterized in that, The real-time imaging task processing model includes an objective function and constraints. The objective function is: ； in, For edge computing nodes In the time slot Processing tasks The minimum delay, i.e., the optimal solution to the objective function, is the optimal task processing scheme. For edge computing nodes Processing tasks Total delay, For the task In the time slot At edge computing nodes Queuing delays, , For edge computing nodes In the time slot Task queue length, , For edge computing nodes In the time slot Task queue length, , For time slots Offload to edge computing nodes The number of tasks, For time slots Edge computing nodes Number of tasks processed For time slots Task Instruction length, For time slots Entering the edge computing node The task set, For time slots Entering the edge computing node The total number of tasks For edge computing nodes The instruction length of the task processed in each time slot. For the task In the time slot From edge computing nodes Transmitted to data acquisition node Total delay, , For the task In the time slot From edge computing nodes Transmitted to data acquisition node Network congestion and delays, , For edge computing nodes In the time slot To data acquisition nodes The number of remaining tasks to be transmitted, i.e., time slots. The network congestion status, if Then time slot No network congestion, otherwise time slot There is network congestion. For the task Data size, For edge computing nodes In the time slot Data acquisition nodes have been sent The set of tasks to be transmitted For edge computing nodes To data acquisition node network bandwidth, For the task In the time slot From edge computing nodes Transmitted to data acquisition node Transmission delay, , For the task Data size, For edge computing nodes To data acquisition node The spread is delayed.

6. The real-time imaging method for detecting transverse wave velocity structures by micro-motion as described in claim 1, characterized in that, The instant imaging task processing model includes an objective function and constraints, wherein the constraints are as follows: ； in, For edge computing nodes In the time slot Task queue length, The total number of nodes is calculated at the edge. For mathematical expectation, For total timeslots, For edge computing nodes In the time slot The length of the load queue.

7. The real-time imaging method for detecting transverse wave velocity structures by micro-motion as described in claim 1, characterized in that, The data acquisition nodes and edge computing nodes perform computing resource allocation and scheduling based on the load status prediction results of the data acquisition nodes output by the pre-built computing resource quantity search algorithm, including: Initially, the data acquisition node and the edge computing node operate in the self-power spectrum calculation stage, and the data acquisition node sends its load status to the edge computing node; Within each time slot, the edge computing node executes a pre-built algorithm to search for the number of computing resources based on the load status of the previously received data acquisition node, and obtains the load status prediction result of the data acquisition node. Within each time slot, the edge computing node allocates and schedules computing resources based on the load status prediction results of the data acquisition node.

8. A real-time imaging system for detecting transverse wave velocity structures using micro-motion detection, characterized in that, Including processor and storage media; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1 to 7.

Images