Search

CN-122017948-A - Underground magnetic induction positioning method and earthquake emergency disaster reduction intelligent body system

CN122017948ACN 122017948 ACN122017948 ACN 122017948ACN-122017948-A

Abstract

The invention discloses an underground magnetic induction positioning method and an earthquake emergency disaster reduction intelligent body system, which belong to the field of natural disaster intelligent emergency response and follow in a near field by utilizing magnetic induction signals The physical characteristics of intensity attenuation are used for inverting the target distance, wherein in order to adapt to complex buried medium, a layered medium model is introduced, a generalized transmission coefficient is obtained through interlayer superposition, attenuation and multipath effects caused by permeability, conductivity and thickness are accurately described, and high-precision penetrating type positioning can be realized on an underground target on the premise of overcoming strong interference and adapting to complex medium distribution. On the basis, by introducing a virtual multi-view collaborative sensing mechanism and designing a data alignment and feature complementation rule among the views of space-time blending, virtual-real interleaving and far-near alternation, the system can generate a comprehensive decision scheme which takes into consideration multiple targets of risk early warning, life rescue, damage evaluation and the like when dealing with the full-period disaster of an earthquake.

Inventors

  • LIU GUANGHUA
  • YUAN HAORAN
  • Jia Zhouhui

Assignees

  • 华中科技大学

Dates

Publication Date
20260512
Application Date
20260121

Claims (10)

  1. 1. A method for positioning underground magnetic induction is characterized by positioning an object carrying coil equipment capable of receiving and transmitting magnetic induction signals in an underground space, wherein M magnetic induction sensors are arranged at preset positions in the underground space, M is a positive integer, each magnetic induction sensor comprises three coils which are orthogonal to each other, a first coil and a second coil Coil device carried on individual object To the first First of the individual magnetic induction sensors Individual coils Between which a transmission link is formed ; D; ; D is the target number; The underground magnetic induction positioning method comprises the following steps: transmitting a plurality of transmitting signals S with different frequencies to an underground space respectively, and acquiring frequency response signals from each magnetic induction sensor; separately calculating each frequency Down transmission link Length of (2) ; Is a preset constant; R is preset impedance; For frequency The following generalized transmission coefficients are specifically: ; For frequency The corresponding angular frequency, N is the number of medium layers in the near field region, and the near field region is the distance from the ground in the underground space Is a region of (2); ; ; And The minimum frequency and the maximum frequency of the transmitted transmission signal S, respectively; c is the speed of light; And Respectively the first Layer (a) Interface between layers of medium at angular frequency The lower transmission coefficient and the equivalent Fresnel reflection coefficient; is the first Layer medium at angular frequency The complex propagation constant below; is the first The thickness of the layer medium; For frequencies extracted from frequency response signals Down transmission link The RSSI true value of the received signal of the upper magnetic induction sensor; Computing a transmission link Final length of (2) Wherein, the method comprises the steps of, A set of frequencies for the transmitted signal; For frequency The corresponding weight; Length of (C) 、 And As the weighted sum result of (a) Target and the first The distance of the magnetic induction sensors, thereby achieving positioning.
  2. 2. The method of claim 1, wherein the method comprises the steps of 、 And When the weighted sum is to be made, 、 And The corresponding weights are respectively normalized results of link quality factors of the corresponding transmission links; wherein the transmission link The link quality factor of (2) is: For the equivalent transmission coefficient magnitude value, ; For transmission links Shadow fading variance of (a).
  3. 3. A method for subsurface magnetic induction positioning according to claim 1 or 2, Is determined by the following means: The obtained orientation underground space has a transmitting frequency of Measured at the time of transmitting signal S Is connected with the receiving voltage and coil of the transformer And taking the ratio of the two as the frequency Down transmission link Is a near field coupled eigen term of (a); coupling the near field to the eigenvalue Is the product of (2) as the frequency Down transmission link Frequency domain gain of (2) to obtain frequency Down transmission link A predicted value of RSSI of a received signal of the upper magnetic induction sensor; Will be Subtracting the corresponding predicted value to obtain the frequency Down transmission link Shadow perturbation of (a) Further calculating to obtain the corresponding shadow fading weight ; Is the standard deviation of shadow fading; Constructing graph data, wherein nodes in the graph data comprise anchor nodes and target nodes, the anchor nodes represent magnetic induction sensors in underground space, the target nodes represent coil equipment carried on a target, edges in the graph data represent signal association among the nodes, shadow fading weights of transmission links between the anchor nodes and the target nodes are carried on the edges of the anchor nodes and the target nodes in the graph data, and RSSI true values of received signals on the anchor nodes, and for each anchor node in the graph data, when the mean value deviation of the RSSI true values of the received signals on any frequency and the RSSI true values of the received signals on one-hop neighbor anchor nodes exceeds a preset deviation, the RSSI true values of the received signals on the one-hop neighbor anchor nodes are updated to be the mean value of the RSSI true values of the received signals on the one-hop neighbor anchor nodes; inputting the graph data into a graph neural network, converging and fusing the characteristics of all transmission links according to frequencies through a message transmission mechanism of the graph neural network to obtain the fusion characteristics of each transmission link under different frequencies, wherein the transmission links are connected with each other through the message transmission mechanism of the graph neural network At the frequency of The following fusion characteristics are noted as ; Respectively calculating the scores corresponding to the frequencies, and normalizing to obtain weights corresponding to the frequencies, wherein the frequencies Corresponding score ; Representing a multi-layer perceptron.
  4. 4. The method of claim 1 or 2, further comprising infrared detecting the subsurface space to determine the orientation of the target.
  5. 5. The earthquake emergency disaster reduction intelligent agent system is characterized by being used for performing earthquake emergency response on an underground space where an earthquake area is located, wherein M magnetic induction sensors are deployed at preset positions in the underground space, M is a positive integer, and each magnetic induction sensor comprises three coils which are mutually orthogonal; The earthquake emergency disaster reduction intelligent system comprises an in-disaster emergency response module, a disaster emergency response module and a disaster emergency disaster reduction intelligent system, wherein the in-disaster emergency response module is used for executing the underground magnetic induction positioning method according to any one of claims 1-4 after an earthquake occurs so as to position an underground target.
  6. 6. The earthquake emergency disaster reduction agent system of claim 5, further comprising a pre-disaster risk early warning module for pre-disaster risk early warning before an earthquake occurs, comprising: acquiring three-component seismic wave data and corresponding metadata, which are acquired in real time by a seismic station corresponding to a seismic area, wherein the metadata comprise address information of the seismic station and recorded P wave theory arrival time; Preprocessing each component wave in the three-component seismic wave data, and then picking up the P wave arrival time of the component wave by adopting a long-short time window average ratio method Further intercepting the component wave The sub-band of the initial preset time period is taken as a corresponding P-band; and inputting the P wave bands corresponding to the metadata and the three-component seismic wave data into a pre-trained deep learning model, obtaining a prediction result of the earthquake magnitude, the earthquake source depth and the arrival time difference of the earthquake waves, and carrying out pre-disaster risk early warning.
  7. 7. The seismic emergency disaster reduction agent system of claim 6, wherein the pre-disaster risk pre-warning prior to the occurrence of the earthquake further comprises: The method comprises the steps of acquiring historical seismic catalogue data of a seismic area, and preprocessing, wherein the historical seismic catalogue data comprise the occurrence time, the occurrence place, the seismic source depth and the seismic level of the seismic area; space grid division is carried out on the seismic area, and data of the seismic occurrence places in the same space grid in the preprocessed historical seismic catalog data are aggregated to obtain historical seismic catalog data under different space grids; Performing feature extraction on historical seismic catalog data under each space grid to obtain corresponding seismic features, and performing PCA dimension reduction processing to obtain corresponding candidate features, wherein the seismic features comprise a value b for describing the slope of seismic magnitude distribution, recurrence period features, strain energy release rate, space concentration degree, fault distance and historical aftershock attenuation in a Gutenberg-Richter relation, wherein the recurrence period features are mean value, variance and quantile of adjacent event intervals in the historical seismic catalog data; Inputting candidate features corresponding to each space grid into a pre-trained machine learning model to obtain the probability of earthquake occurrence of the space grid; Based on the probability of each spatial grid for earthquake, a risk thermodynamic diagram and a trend curve of earthquake occurrence in the earthquake area are generated.
  8. 8. The seismic emergency disaster reduction agent system of claim 6, wherein the pre-disaster risk pre-warning prior to the occurrence of the earthquake further comprises: The method comprises the steps of carrying out weighted summation on video anomaly scores and audio anomaly scores of a seismic area to obtain environmental precursor scores, and sending out precursor early warning when the environmental precursor scores exceed preset scores; the video anomaly score of the seismic area is calculated by the following method: detecting animal targets in each frame of the obtained animal monitoring video stream of the seismic area, and calculating the centroid track of each animal target and the movement speed at each frame time; Taking a frame where a moving target with the moving speed exceeding a preset speed and the duration exceeding a first preset time in an animal monitoring video stream is located as an abnormal frame; Taking a frame in which the animal aggregation group with the duration exceeding the second preset time is positioned in the animal monitoring video stream as an abnormal frame, wherein the number of animal targets in the animal aggregation group exceeds the preset number, and the minimum value of the distance between every two animal targets does not exceed the preset distance; Calculating the ratio of the number of abnormal frames in the animal monitoring video stream to the total frame number as a video abnormal score; the audio anomaly score for a seismic area is calculated by: The method comprises the steps of acquiring the environment audio of an earthquake region by a sliding window, dividing the acquired environment audio of the earthquake region into a plurality of environmental sub-audios, extracting the audio characteristics of each environmental sub-audio, and splicing the audio characteristics and the triaxial acceleration of synchronously acquired audio acquisition equipment together into multi-modal characteristic vectors of the corresponding environment sub-audios; carrying out abnormal or non-abnormal classification on each environmental sub-audio based on the multi-modal feature vector; The ratio of the number of abnormal environmental sub-audios to the total number of environmental sub-audios is calculated as an audio abnormality score.
  9. 9. The seismic emergency disaster reduction agent system of claim 5, further configured to obtain a social media data stream related to the seismic area after the occurrence of the earthquake, comprising text data and video data; Performing BERT coding on the token sequence to obtain a context vector of each token, predicting whether each token is an earthquake-related entity based on the context vector by adopting a sequence marking head, and extracting the token of which the token sequence is classified as the earthquake-related entity as a target token; respectively calculating the probability that each target token belongs to a low public opinion emergency type, a medium public opinion emergency type and a high public opinion emergency type by adopting a classifier, and carrying out weighted summation to obtain public opinion emergency scores of the corresponding target tokens; taking the sum of the public opinion emergency scores of all the target token as the total public opinion emergency score of the earthquake area; Classifying whether an earthquake occurs in each frame of image in the video data aiming at the video data, and calculating the ratio of the number of frames of the earthquake to the total number of frames in the video data as the damage score of the earthquake area; taking each target token related to the location as a disaster-affected location, and constructing a disaster-affected point distribution map of the seismic area; extracting population density of each disaster point from a population density thermodynamic diagram of the earthquake region, and constructing a population density distribution diagram of the disaster point of the earthquake region after normalization processing; and calculating the weighted sum result of the population density, the total public opinion score and the damage score of the earthquake region aiming at each disaster-stricken point as a rescue requirement index of the disaster-stricken point, and further constructing a disaster-stricken point rescue requirement thermodynamic diagram of the earthquake region.
  10. 10. The seismic emergency disaster reduction agent system of claim 5, further comprising a post-disaster recovery reconstruction guidance module, wherein the post-disaster recovery reconstruction guidance module comprises: The macro change detection unit is used for acquiring a post-disaster satellite image map of an earthquake area acquired in real time, detecting a changed area in the post-disaster satellite image map of the earthquake area compared with a pre-disaster satellite image map of the earthquake area, and performing marking so as to guide post-disaster recovery reconstruction; Microcosmic retrieval damage detection unit for acquiring each building in earthquake area acquired by unmanned aerial vehicle in real time Respectively classifying the damage level of each building image by using a classifier to obtain each building Is a damage level quantification score of (2) ; The secondary disaster risk deduction unit is used for calculating each building in the earthquake region by adopting a Bayesian algorithm based on geological environment data and historical earthquake knowledge The probability of a secondary disaster currently occurring is used as the risk of the secondary disaster of the building ; Planning guidance unit for calculating each building Comprehensive priority score of (2) Wherein, the method comprises the steps of, 、 、 、 All are preset coefficients; Is a building Is a preset functional importance score of (2); Is a building To the predicted repair cost of (a) By maximising, as a constraint For a pair of Solving to generate a reconstruction planning scheme, wherein, , Representing no building The repair is carried out and the repair is carried out, Representation pair building And repairing.

Description

Underground magnetic induction positioning method and earthquake emergency disaster reduction intelligent body system Technical Field The invention belongs to the field of intelligent emergency response of natural disasters, and particularly relates to an underground magnetic induction positioning method and an earthquake emergency disaster reduction intelligent body system. Background The high concentration of population and economic elements in the rapid urban process further amplifies natural disasters, such as building collapse, lifeline engineering paralysis, secondary disasters and other chain reactions caused by earthquakes, and forms serious threat to the life and property safety of people, wherein after the natural disasters occur, the underground targets are subjected to magnetic induction positioning to form an indispensable part for building an efficient and accurate emergency rescue system. The existing underground disaster relief positioning method has the common problem of poor environment suitability that one type of method relies on far-field electromagnetic wave propagation, such as WiFi, zigBee and the like, is fast attenuated in conductive/high-loss media such as soil, concrete, reinforcing steel bars and the like due to strong absorption, multipath and non-line-of-sight propagation, so that a link is unstable and even fails in penetration, satellites such as GPS and the like cannot be positioned in underground scenes, UWB, RFID or radar short-distance schemes are easily affected by vortex interference and multipath in the reinforcing steel bar-concrete-rock soil hybrid environment, the error is often more than 0.8 m, the rescue precision requirement is difficult to meet, the acoustic and optical are blocked, the multipath and the media cannot penetrate, and the like, positioning ambiguity is easy to generate, and the real-time performance and reliability are insufficient. In addition, existing subsurface Magnetic Induction (MI) positioning has near field penetration advantages, but most still ignore layered non-uniform media and near field effective coupling boundaries based on channel assumptions of "uniform media/simplified scene", resulting in model and field mismatch. Therefore, development of an underground magnetic induction positioning technical scheme capable of overcoming strong interference, adapting to complex medium distribution and realizing high-precision penetrating positioning is needed. Disclosure of Invention Aiming at the defects or improvement demands of the prior art, the invention provides an underground magnetic induction positioning method and an earthquake emergency disaster reduction intelligent body system, which aim to realize high-precision penetrating type positioning on the premise of overcoming strong interference and adapting to complex medium distribution. In order to achieve the above object, in a first aspect, the present invention provides an underground magnetic induction positioning method for positioning an object carrying a coil device capable of transceiving magnetic induction signals in an underground space; wherein M magnetic induction sensors are arranged at preset positions in the underground space, M is a positive integer, and each magnetic induction sensor comprises three coils which are mutually orthogonal in pairs, and a first coilCoil device carried on individual objectTo the firstFirst of the individual magnetic induction sensorsIndividual coilsBetween which a transmission link is formed;D;;D is the target number; the underground magnetic induction positioning method comprises the following steps: transmitting a plurality of transmitting signals S with different frequencies to an underground space respectively, and acquiring frequency response signals from each magnetic induction sensor; separately calculating each frequency Down transmission linkLength of (2);Is a preset constant; R is preset impedance; For frequency The following generalized transmission coefficients are specifically:; For frequency The corresponding angular frequency, N is the number of medium layers in the near field region, and the near field region is the distance from the ground in the underground spaceIs a region of (2);;; And The minimum frequency and the maximum frequency of the transmitted transmission signal S, respectively; c is the speed of light; And Respectively the firstLayer (a)Interface between layers of medium at angular frequencyThe lower transmission coefficient and the equivalent Fresnel reflection coefficient; is the first Layer medium at angular frequencyThe complex propagation constant below; is the first The thickness of the layer medium; For frequencies extracted from frequency response signals Down transmission linkThe RSSI true value of the received signal of the upper magnetic induction sensor; Computing a transmission link Final length of (2)Wherein, the method comprises the steps of,A set of frequencies for the transmitted signal; For f