CN116976837A - Ground disaster situation analysis method and system applied to data sharing - Google Patents
Ground disaster situation analysis method and system applied to data sharing Download PDFInfo
- Publication number
- CN116976837A CN116976837A CN202311235261.7A CN202311235261A CN116976837A CN 116976837 A CN116976837 A CN 116976837A CN 202311235261 A CN202311235261 A CN 202311235261A CN 116976837 A CN116976837 A CN 116976837A
- Authority
- CN
- China
- Prior art keywords
- ground disaster
- state
- ground
- disaster state
- vector
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 189
- 238000012544 monitoring process Methods 0.000 claims abstract description 75
- 238000012545 processing Methods 0.000 claims abstract description 61
- 238000009826 distribution Methods 0.000 claims abstract description 54
- 239000013598 vector Substances 0.000 claims description 360
- 238000005065 mining Methods 0.000 claims description 74
- 238000013507 mapping Methods 0.000 claims description 66
- 238000000034 method Methods 0.000 claims description 32
- 238000012549 training Methods 0.000 claims description 16
- 230000002776 aggregation Effects 0.000 claims description 14
- 238000004220 aggregation Methods 0.000 claims description 14
- 238000012512 characterization method Methods 0.000 claims description 11
- 230000009467 reduction Effects 0.000 claims description 10
- 238000006243 chemical reaction Methods 0.000 claims description 9
- 238000004590 computer program Methods 0.000 claims description 6
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 230000000875 corresponding effect Effects 0.000 description 230
- 230000008569 process Effects 0.000 description 23
- 238000004891 communication Methods 0.000 description 12
- 239000011159 matrix material Substances 0.000 description 9
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 8
- 239000011435 rock Substances 0.000 description 8
- 238000013528 artificial neural network Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 5
- 239000003673 groundwater Substances 0.000 description 4
- 229910052742 iron Inorganic materials 0.000 description 4
- 238000006424 Flood reaction Methods 0.000 description 3
- 238000001125 extrusion Methods 0.000 description 3
- 238000011946 reduction process Methods 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 238000005336 cracking Methods 0.000 description 2
- 230000001066 destructive effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000000630 rising effect Effects 0.000 description 2
- 239000002689 soil Substances 0.000 description 2
- 230000003213 activating effect Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013506 data mapping Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000009969 flowable effect Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 229910052500 inorganic mineral Inorganic materials 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 239000003550 marker Substances 0.000 description 1
- 238000002844 melting Methods 0.000 description 1
- 230000008018 melting Effects 0.000 description 1
- 239000011707 mineral Substances 0.000 description 1
- 210000003205 muscle Anatomy 0.000 description 1
- 238000003012 network analysis Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 239000013049 sediment Substances 0.000 description 1
- 239000004575 stone Substances 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
- G06Q10/103—Workflow collaboration or project management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/23—Updating
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
- Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping
Abstract
The invention provides a ground disaster situation analysis method and a ground disaster situation analysis system applied to data sharing, and relates to the technical field of artificial intelligence. In the invention, based on historical ground disaster monitoring data acquired from a ground disaster monitoring data sharing system, a sample ground disaster state mark chain is formed, and the sample ground disaster state mark chain is loaded into a candidate ground disaster situation analysis network for network updating processing to form an updated ground disaster situation analysis network; forming a current ground disaster state mark chain based on current ground disaster monitoring data acquired from a ground disaster monitoring data sharing system, and loading the current ground disaster state mark chain into an updated ground disaster situation analysis network; analyzing the estimated probability parameter distribution corresponding to the current ground disaster state sign chain by using an updated ground disaster situation analysis network; and determining the target ground disaster state based on the estimated probability parameter distribution. Based on the above, the reliability of ground disaster situation analysis can be improved to a certain extent.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a ground disaster situation analysis method and system applied to data sharing.
Background
The ground disaster generally affects and damages more application objects, for example, the communication tower is damaged, so that the data transmission function of the communication tower is limited. The following are some geological disasters that may affect the communication tower:
earthquake: earthquake may cause the communication tower to collapse or be damaged, especially when the earthquake intensity is large; debris flow and landslide: mud-rock flow and landslide can cause the communication tower to be washed out or buried; ground subsidence: ground subsidence is caused by over exploitation of underground water, exploitation of underground mineral deposits, geological structure change and other reasons, and the foundation of the communication tower can be unstable; storm and typhoon: strong storms and typhoons may generate strong winds and hurricanes, which may cause damage to the communication tower; storms and floods: storms and floods may cause the pylon to be submerged or subject to water intrusion; the ground water level rises: the ground water level rise may cause the foundation of the pylon to be immersed in water, thereby damaging the structure.
Therefore, in order to ensure the safety of the communication iron tower, the ground disaster situation of the area where the communication iron tower is located is generally analyzed and estimated, so that the generation of an emergency plan can be performed on the communication iron tower based on the analysis and estimated result, and the effective work of the communication iron tower is ensured. However, in the prior art, the monitored data is typically analyzed manually, and thus, the reliability thereof is relatively low.
Disclosure of Invention
Therefore, the invention aims to provide a ground disaster situation analysis method and a ground disaster situation analysis system applied to data sharing, so as to improve the reliability of ground disaster situation analysis to a certain extent.
In order to achieve the above purpose, the embodiment of the present invention adopts the following technical scheme:
a ground disaster situation analysis method applied to data sharing comprises the following steps:
forming a sample ground disaster state mark chain based on historical ground disaster monitoring data acquired from a ground disaster monitoring data sharing system, loading the sample ground disaster state mark chain into a candidate ground disaster situation analysis network to perform network updating processing to form a corresponding updated ground disaster situation analysis network, wherein the sample ground disaster state mark chain comprises ground disaster state marks which have correlation and are sequentially arranged, and the ground disaster state marks in the sample ground disaster state mark chain comprise at least two of state marks corresponding to various ground disaster events and state marks corresponding to non-ground disaster events;
forming a current ground disaster state mark chain based on the current ground disaster monitoring data acquired from the ground disaster monitoring data sharing system, and loading the current ground disaster state mark chain into the updated ground disaster situation analysis network, wherein the current ground disaster monitoring data is specific to a target monitoring area;
Analyzing estimated probability parameter distribution corresponding to the current ground disaster state sign chain by using the updated ground disaster situation analysis network, wherein each estimated probability parameter in the estimated probability parameter distribution is used for reflecting the probability of the ground disaster state corresponding to one ground disaster state sign;
and determining a target ground disaster state corresponding to the current ground disaster state mark chain based on the estimated probability parameter distribution, wherein the target ground disaster state is used for reflecting ground disaster events or non-ground disaster events which occur in the target monitoring area.
In some preferred embodiments, in the above ground disaster situation analysis method applied to data sharing, the step of forming a sample ground disaster state flag chain based on historical ground disaster monitoring data acquired from a ground disaster monitoring data sharing system, and loading the sample ground disaster state flag chain into a candidate ground disaster situation analysis network to perform network update processing, to form a corresponding updated ground disaster situation analysis network includes:
forming a sample ground disaster state mark chain based on historical ground disaster monitoring data acquired from a ground disaster monitoring data sharing system, and loading the sample ground disaster state mark chain into a candidate ground disaster situation analysis network;
Matching ground disaster state semantic vectors corresponding to ground disaster state marks positioned in front of coordinates to be analyzed in the sample ground disaster state mark chain from a reference ground disaster state semantic vector cluster by utilizing the candidate ground disaster situation analysis network, wherein the coordinates to be analyzed belong to the ground disaster state mark coordinates corresponding to the sample ground disaster state mark chain;
vector analysis processing is carried out on the ground disaster state semantic vectors corresponding to the ground disaster state marks before the coordinates to be analyzed, and the ground disaster state estimated vectors corresponding to the coordinates to be analyzed are output;
according to the ground disaster state prediction vector corresponding to the coordinate to be analyzed, a prediction possibility parameter corresponding to a ground disaster state sign on the coordinate to be analyzed is analyzed, wherein the prediction possibility parameter is used for reflecting a ground disaster state corresponding to the ground disaster state sign on the coordinate to be analyzed and a prediction correlation coefficient of a previous ground disaster state chain corresponding to the coordinate to be analyzed, and the previous ground disaster state chain comprises ground disaster states corresponding to each ground disaster state sign before the coordinate to be analyzed;
according to the estimated probability parameters corresponding to the ground disaster state marks on each coordinate to be analyzed in the sample ground disaster state mark chain, analyzing corresponding ground disaster state estimated errors;
And updating and optimizing network parameters of the candidate ground disaster situation analysis network according to the ground disaster state estimation error, and forming a corresponding updated ground disaster situation analysis network when the network parameters are matched with a preset reference network training rule.
In some preferred embodiments, in the above ground disaster situation analysis method applied to data sharing, the step of forming a sample ground disaster state flag chain based on historical ground disaster monitoring data acquired from a ground disaster monitoring data sharing system, loading the sample ground disaster state flag chain into a candidate ground disaster situation analysis network to perform network update processing, and forming a corresponding updated ground disaster situation analysis network further includes:
determining an original space mapping vector corresponding to each ground disaster state in a reference ground disaster state cluster;
loading the original space mapping vectors corresponding to each ground disaster state into an update vector mining model, wherein the update vector mining model is formed based on corresponding network update processing;
for the original space mapping vector of any one of the original space mapping vectors, performing deep mining on the original space mapping vector corresponding to the ground disaster state by utilizing the updated vector mining model, outputting the ground disaster state mining vector corresponding to the ground disaster state, performing dimension reduction processing on the ground disaster state mining vector corresponding to the ground disaster state, and outputting the ground disaster state semantic vector corresponding to the ground disaster state;
And combining the ground disaster state semantic vectors corresponding to the ground disaster states to form corresponding reference ground disaster state semantic vector clusters.
In some preferred embodiments, in the above ground disaster situation analysis method applied to data sharing, the step of determining an original spatial mapping vector corresponding to each ground disaster state in the reference ground disaster state cluster includes:
determining a ground disaster state description data cluster corresponding to each ground disaster state in a reference ground disaster state cluster, wherein each ground disaster state description data in the ground disaster state description data cluster is used for describing attribute information of the ground disaster state and environment information of a time interval in which the ground disaster state occurs;
for any one of the ground disaster state description data clusters, performing vector space mapping on each ground disaster state description data in the ground disaster state description data clusters, outputting ground disaster state description data vectors corresponding to each ground disaster state description data respectively, and performing cascade combination on the ground disaster state description data vectors corresponding to each ground disaster state description data respectively to form an original space mapping vector corresponding to the ground disaster state description data clusters.
In some preferred embodiments, in the above ground disaster situation analysis method applied to data sharing, the step of forming a sample ground disaster state flag chain based on historical ground disaster monitoring data acquired from a ground disaster monitoring data sharing system, loading the sample ground disaster state flag chain into a candidate ground disaster situation analysis network to perform network update processing, and forming a corresponding updated ground disaster situation analysis network further includes:
determining a sample ground disaster state combination;
loading a first ground disaster state in the sample ground disaster state combination so as to load the first vector mining model to be updated, mining a ground disaster state semantic vector corresponding to the first ground disaster state, and loading a second ground disaster state in the sample ground disaster state combination so as to load the second vector mining model to be updated, and mining a ground disaster state semantic vector corresponding to the second ground disaster state;
according to vector matching parameters between a ground disaster state semantic vector corresponding to the first ground disaster state and a ground disaster state semantic vector corresponding to the second ground disaster state, analyzing ground disaster state relation analysis data corresponding to the sample ground disaster state combination, wherein the ground disaster state relation analysis data are data which are analyzed and used for reflecting whether a correlation exists between two sample ground disaster states in the sample ground disaster state combination;
According to the distinguishing information between the ground disaster state relation real data corresponding to the sample ground disaster state combination and the ground disaster state relation analysis data corresponding to the sample ground disaster state combination, updating and optimizing parameters of the first vector mining model to be updated and the second vector mining model to be updated, and forming an updated first vector mining model and an updated second vector mining model when the parameters are matched with a predetermined first configuration network training rule, wherein the ground disaster state relation real data are real data used for reflecting whether correlation exists between two sample ground disaster states in the sample ground disaster state combination;
and determining a vector mining model from the updated first vector mining model and the updated second vector mining model as the updated vector mining model.
In some preferred embodiments, in the above ground disaster situation analysis method applied to data sharing, the step of forming a sample ground disaster state flag chain based on historical ground disaster monitoring data acquired from a ground disaster monitoring data sharing system, loading the sample ground disaster state flag chain into a candidate ground disaster situation analysis network to perform network update processing, and forming a corresponding updated ground disaster situation analysis network further includes:
Determining coordinate semantic vectors corresponding to each ground disaster state sign coordinate before the coordinate to be analyzed by utilizing the candidate ground disaster situation analysis network;
vector aggregation processing is carried out on the ground disaster state semantic vector corresponding to the ground disaster state sign and the coordinate semantic vector corresponding to the ground disaster state sign coordinate corresponding to the ground disaster state sign, and the state coordinate aggregation semantic vector corresponding to each ground disaster state sign before the coordinate to be analyzed in the sample ground disaster state sign chain is output;
the step of carrying out vector analysis processing on the ground disaster state semantic vectors corresponding to the ground disaster state marks before the coordinates to be analyzed and outputting the ground disaster state estimated vectors corresponding to the coordinates to be analyzed comprises the following steps:
and carrying out vector analysis processing on the state coordinate aggregation semantic vectors corresponding to the ground disaster state marks positioned in front of the coordinates to be analyzed, and outputting the ground disaster state estimated vector corresponding to the coordinates to be analyzed.
In some preferred embodiments, in the above ground disaster situation analysis method applied to data sharing, the step of analyzing the estimated likelihood parameter corresponding to the ground disaster state flag on the coordinate to be analyzed according to the ground disaster state estimated vector corresponding to the coordinate to be analyzed includes:
Vector space conversion processing is carried out on the ground disaster state estimated vector corresponding to the coordinate to be analyzed, and candidate ground disaster state parameter distribution corresponding to the coordinate to be analyzed is output, wherein the candidate ground disaster state parameter distribution comprises characterization parameters respectively corresponding to each ground disaster state in a reference ground disaster state cluster;
performing characterization parameter reduction processing on the candidate ground disaster state parameter distribution, and outputting target ground disaster state parameter distribution corresponding to the coordinates to be analyzed, wherein the target ground disaster state parameter distribution comprises estimated possibility parameters corresponding to each ground disaster state in the reference ground disaster state cluster, and the reference ground disaster state cluster comprises ground disaster states corresponding to each ground disaster state mark in the sample ground disaster state mark chain;
and analyzing estimated probability parameters corresponding to the ground disaster state marks on the coordinates to be analyzed based on the target ground disaster state parameter distribution.
In some preferred embodiments, in the above ground disaster situation analysis method applied to data sharing, the candidate ground disaster situation analysis network includes a vector space mapping unit, a vector analysis unit, and a vector estimation unit, where the vector space mapping unit includes a first vector space mapping layer and a second vector space mapping layer, the first vector space mapping layer is used to determine a ground disaster state semantic vector, the second vector space mapping layer is used to determine a coordinate semantic vector, the vector analysis unit is used to perform vector analysis processing, and the vector estimation unit is used to analyze an estimated likelihood parameter;
The step of updating and optimizing the network parameters of the candidate ground disaster situation analysis network according to the ground disaster state estimation error, and forming a corresponding updated ground disaster situation analysis network when the network parameters are matched with a pre-configured reference network training rule comprises the following steps:
and updating and optimizing parameters of the second vector space mapping layer, the vector analysis unit and the vector estimation unit in the candidate ground disaster situation analysis network according to the ground disaster state estimation error, and forming a corresponding updated ground disaster situation analysis network when the parameters are matched with a pre-configured reference network training rule.
In some preferred embodiments, in the above ground disaster situation analysis method applied to data sharing, the vector estimation unit is configured to analyze, according to an output vector of the vector analysis unit, estimated likelihood parameters corresponding to each ground disaster state in a reference ground disaster state cluster;
the step of forming a sample ground disaster state mark chain based on the historical ground disaster monitoring data acquired from the ground disaster monitoring data sharing system, loading the sample ground disaster state mark chain into a candidate ground disaster situation analysis network for network updating processing to form a corresponding updated ground disaster situation analysis network, and further comprises the steps of:
After the updated ground disaster situation analysis network is formed, adding a ground disaster state semantic vector corresponding to the extended ground disaster state from the reference ground disaster state semantic vector cluster under the condition that the extended ground disaster state is distributed to the reference ground disaster state cluster;
loading a ground disaster state comparison mark chain into the updated ground disaster situation analysis network to form estimated probability parameters of each ground disaster state in the expanded reference ground disaster state cluster relative to each coordinate to be analyzed;
analyzing estimated probability parameters corresponding to the ground disaster state marks on the coordinates to be analyzed in the compared ground disaster state mark chain from the estimated probability parameters of the ground disaster states in the expanded reference ground disaster state cluster relative to the coordinates to be analyzed;
calculating a corresponding ground disaster state expansion error according to the estimated probability parameters corresponding to the ground disaster state marks on each coordinate to be analyzed in the ground disaster state mark comparison chain;
and updating and optimizing parameters of the vector estimation unit in the updated ground disaster situation analysis network according to the ground disaster state expansion error, and forming the updated ground disaster situation analysis network adapted to the expanded reference ground disaster state cluster when the parameters are matched with a second preset configuration network training rule.
The embodiment of the invention also provides a ground disaster situation analysis system applied to data sharing, which comprises a processor and a memory, wherein the memory is used for storing a computer program, and the processor is used for executing the computer program so as to realize the ground disaster situation analysis method applied to data sharing.
The ground disaster situation analysis method and the ground disaster situation analysis system for data sharing provided by the embodiment of the invention can utilize the neural network to analyze and predict the ground disaster state, so that the reliability of the ground disaster situation analysis is higher compared with the conventional technical scheme based on manual analysis; in addition, in the process of network updating, the sample ground disaster state mark chains are aimed at, and the ground disaster state marks included in the sample ground disaster state mark chains are related and arranged in sequence, so that the related relationship between various ground disaster events and non-ground disaster events can be fully learned in the process of network updating, and the related relationship is considered when the current ground disaster state mark chains are analyzed and estimated, so that the reliability of ground disaster situation analysis can be improved.
In order to make the above objects, features and advantages of the present invention more comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
Fig. 1 is a block diagram of a ground disaster situation analysis system applied to data sharing according to an embodiment of the present invention.
Fig. 2 is a schematic flow chart of steps included in a ground disaster situation analysis method applied to data sharing according to an embodiment of the present invention.
Fig. 3 is a schematic diagram of each module included in a ground disaster situation analysis device applied to data sharing according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, but not all embodiments of the present invention. The components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.
As shown in fig. 1, the embodiment of the invention provides a ground disaster situation analysis system applied to data sharing. The ground disaster situation analysis system applied to data sharing can comprise a memory and a processor. For example, in some specific applications, the memory and the processor are electrically connected directly or indirectly to enable transmission or interaction of data. For example, electrical connection may be made to each other via one or more communication buses or signal lines. The memory may store at least one software functional module (computer program) that may exist in the form of software or firmware. The processor may be configured to execute the executable computer program stored in the memory, so as to implement the ground disaster situation analysis method applied to data sharing provided by the embodiment of the present invention.
For example, in some specific applications, the Memory may be, but is not limited to, random access Memory (Random Access Memory, RAM), read Only Memory (ROM), programmable Read Only Memory (Programmable Read-Only Memory, PROM), erasable Read Only Memory (Erasable Programmable Read-Only Memory, EPROM), electrically erasable Read Only Memory (Electric Erasable Programmable Read-Only Memory, EEPROM), and the like. The processor may be a general purpose processor including a central processing unit (Central Processing Unit, CPU), a network processor (Network Processor, NP), a System on Chip (SoC), etc.; but also Digital Signal Processors (DSPs), application Specific Integrated Circuits (ASICs), field Programmable Gate Arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.
For example, in some specific applications, the ground disaster situation analysis system applied to data sharing may be a server with data processing capability.
With reference to fig. 2, the embodiment of the invention also provides a ground disaster situation analysis method applied to data sharing, which can be applied to the ground disaster situation analysis system applied to data sharing. The method steps defined by the flow related to the ground disaster situation analysis method applied to data sharing can be realized by the ground disaster situation analysis system applied to data sharing.
The specific flow shown in fig. 2 will be described in detail.
Step S110, forming a sample ground disaster state mark chain based on historical ground disaster monitoring data acquired from a ground disaster monitoring data sharing system, and loading the sample ground disaster state mark chain into a candidate ground disaster situation analysis network to perform network updating processing to form a corresponding updated ground disaster situation analysis network.
In the embodiment of the invention, the ground disaster situation analysis system applied to data sharing can form a sample ground disaster state mark chain based on historical ground disaster monitoring data acquired from the ground disaster monitoring data sharing system, and load the sample ground disaster state mark chain into a candidate ground disaster situation analysis network for network updating processing to form a corresponding updated ground disaster situation analysis network. The ground disaster state mark chain comprises ground disaster state marks which have correlation and are sequentially arranged, and the ground disaster state marks in the sample ground disaster state mark chain can be sequentially arranged according to the sequence from the early to the late of occurrence time, and comprise at least two of state marks corresponding to various ground disaster events and state marks corresponding to non-ground disaster events, such as state mark 1 corresponding to ground disaster event 1-state mark 3 corresponding to ground disaster event 2-state mark 2 corresponding to ground disaster event 3-state mark 3 corresponding to ground disaster event 4-state mark 0 corresponding to non-ground disaster event 5-state mark 6 corresponding to ground disaster event 6; it is to be noted here that, for example, a ground disaster event (instead of a ground disaster event) corresponds to only one status flag, if the same ground disaster event occurs at the first time and the second time, the corresponding status flags are the same, that is, the ground disaster event 2 and the ground disaster event 4 in the chain actually occur two times of the same ground disaster event and are both represented by the status flag 3, for example, for two times of occurrence of the earthquake event, the two times of occurrence of the earthquake event and the two times of occurrence of the debris flow and the landslide event are both represented by the same status flag; various ground disaster events can be earthquakes, mud-rock flows, landslides, ground subsidence, storms, typhoons, storm and floods, ground water level rising, fire disasters and the like, and the ground disaster event does not occur any ground disaster. In addition, one sample ground disaster state mark chain is formed aiming at historical ground disaster monitoring data of one monitoring area, and the ground disaster monitoring data sharing system can share the historical ground disaster monitoring data of a plurality of monitoring areas, so that a plurality of sample ground disaster state mark chains can be formed, the sample richness of network updating processing is improved, the formed updated ground disaster situation analysis network can learn more information, and the ground disaster situation analysis accuracy is guaranteed. In addition, the ground disaster state mark refers to information for uniquely identifying a corresponding ground disaster state, such as names of ground disaster states, such as earthquake, debris flow, landslide, ground settlement, storm and typhoon, storm and flood, ground water level rising, fire disaster, no ground disaster and the like.
Step S120, forming a current ground disaster status flag chain based on the current ground disaster monitoring data acquired from the ground disaster monitoring data sharing system, and loading the current ground disaster status flag chain into the updated ground disaster situation analysis network.
In the embodiment of the invention, the ground disaster situation analysis system applied to data sharing can form a current ground disaster state mark chain based on the current ground disaster monitoring data acquired from the ground disaster monitoring data sharing system, and load the current ground disaster state mark chain into the updated ground disaster situation analysis network. The current ground disaster monitoring data are aimed at a target monitoring area, and the ground disaster state marks in the current ground disaster state mark chain comprise at least two of state marks corresponding to various ground disaster events and state marks corresponding to non-ground disaster events. The updated ground disaster situation analysis network is a neural network and can form a ground disaster situation analysis function based on the study of samples.
And step S130, analyzing the estimated possibility parameter distribution corresponding to the current ground disaster state sign chain by using the updated ground disaster situation analysis network.
In the embodiment of the invention, the ground disaster situation analysis system applied to data sharing can utilize the updated ground disaster situation analysis network to analyze the estimated possibility parameter distribution corresponding to the current ground disaster state sign chain. Each of the estimated likelihood parameter distributions is configured to reflect a likelihood that a ground disaster state corresponding to a ground disaster state flag occurs, for example, the estimated likelihood parameter distribution may include an estimated likelihood parameter 1, an estimated likelihood parameter 2, an estimated likelihood parameter 3, an estimated likelihood parameter 4, and an estimated likelihood parameter 5, the estimated likelihood parameter 1 is configured to reflect a likelihood that a ground disaster event 1 will occur, the estimated likelihood parameter 2 is configured to reflect a likelihood that a ground disaster event 2 will occur, the estimated likelihood parameter 3 is configured to reflect a likelihood that a ground disaster event 3 will occur, the estimated likelihood parameter 4 is configured to reflect a likelihood that a ground disaster event 4 will occur, and the estimated likelihood parameter 5 is configured to reflect a likelihood that a non-ground disaster event will occur.
Step S140, determining the target ground disaster state corresponding to the current ground disaster state sign chain based on the estimated probability parameter distribution.
In the embodiment of the invention, the ground disaster situation analysis system applied to data sharing can determine the target ground disaster state corresponding to the current ground disaster state sign chain based on the estimated probability parameter distribution. The target ground disaster state is used for reflecting ground disaster events or non-ground disaster events which occur in the target monitoring area. For example, the ground disaster event or non-ground disaster event corresponding to the largest estimated likelihood parameter among the estimated likelihood parameter 1, the estimated likelihood parameter 2, the estimated likelihood parameter 3, the estimated likelihood parameter 4 and the estimated likelihood parameter 5 included in the estimated likelihood parameter distribution may be determined as a target ground disaster state, such as an earthquake, no ground disaster occurring, or the like.
Illustrating:
the ground disaster monitoring data sharing system is assumed to provide monitoring data of ground disaster events such as earthquakes, debris flows, landslides, ground subsidence and the like for the past few years. From these data, one or more sample ground disaster status flag chains (determined according to the number of corresponding monitoring areas) may be formed, arranged in chronological order, for example: status markers 1- > corresponding to earthquake event 1, status markers 2- > corresponding to debris flow and landslide event 2, status markers 1- > corresponding to earthquake event 3, status markers 0- > corresponding to no ground disaster event, and status markers 1- > corresponding to earthquake event 4. Then, the sample ground disaster state mark chain is loaded into a candidate ground disaster situation analysis network to carry out network updating processing, and an updated ground disaster situation analysis network is generated. The ground disaster monitoring data sharing system is assumed to provide the latest monitoring data of ground disaster events such as earthquakes, debris flows, landslides and the like of the target monitoring area. From these data, a chain of current ground disaster status flags may be formed, including status flags corresponding to various ground disaster events and status flags corresponding to non-ground disaster events, such as: status flag 1- > status flag 2- > status flag 0 corresponding to no ground disaster event corresponding to debris flow and landslide event 2 corresponding to seismic event 1. And then, loading the current ground disaster state sign chain into an updated ground disaster situation analysis network. And analyzing the current ground disaster state sign chain by using the updated ground disaster situation analysis network to obtain the estimated possibility parameter distribution. Assuming that the estimated likelihood parameter distribution includes: the estimated likelihood parameter 1 represents the likelihood of the seismic event 1, the estimated likelihood parameter 2 represents the likelihood of the debris flow and landslide event 2, and the estimated likelihood parameter 3 represents the likelihood of a non-ground disaster event (non-ground disaster). From the network analysis results, specific values of each parameter can be obtained, for example: estimated likelihood parameter 1=0.8, estimated likelihood parameter 2=0.6, estimated likelihood parameter 3=0.2. Ge Nu estimates the maximum value in the probability parameter distribution and determines the target ground disaster state corresponding to the current ground disaster state flag chain. In the above example, the maximum estimated likelihood parameter is the estimated likelihood parameter 1 (the likelihood of the seismic event 1 is 0.8), and therefore, the target ground disaster state can be determined as the occurrence of the earthquake, that is, the occurrence of the earthquake is predicted.
Based on the above, the analysis and the prediction of the ground disaster state can be performed by using the neural network, so that the reliability of the ground disaster situation analysis can be higher compared with the conventional technical scheme based on manual analysis; in addition, in the process of network updating, the correlation relationship between the ground disaster state marks included in the sample ground disaster state mark chain is arranged in sequence, so that the correlation relationship between various ground disaster events and non-ground disaster events can be fully learned in the process of network updating, and the correlation relationship is considered when the current ground disaster state mark chain is analyzed and estimated, so that the reliability of ground disaster situation analysis can be improved, and the problem of low reliability in the prior art is solved.
For example, in some specific applications, the step S110 may further include the following detailed implementation details:
forming a sample ground disaster state mark chain based on historical ground disaster monitoring data acquired from a ground disaster monitoring data sharing system, and loading the sample ground disaster state mark chain into a candidate ground disaster situation analysis network, namely, taking the sample ground disaster state mark chain as a network updating basis of the candidate ground disaster situation analysis network;
By using the candidate ground disaster situation analysis network, matching ground disaster state semantic vectors corresponding to ground disaster state marks positioned before the to-be-analyzed coordinates in the sample ground disaster state mark chain from a reference ground disaster state semantic vector cluster, wherein the to-be-analyzed coordinates belong to each ground disaster state mark coordinate corresponding to the sample ground disaster state mark chain, and each ground disaster state mark coordinate can be used as the to-be-analyzed coordinates in turn, for example, a first ground disaster state mark coordinate is used as the to-be-analyzed coordinates, a second ground disaster state mark coordinate is used as the to-be-analyzed coordinates, a third ground disaster state mark coordinate is used as the to-be-analyzed coordinates, and a fourth ground disaster state mark coordinate is used as the to-be-analyzed coordinates, and each ground disaster state semantic vector included in the reference ground disaster state semantic vector cluster can be configured for various ground disasters in advance;
vector analysis processing is carried out on the ground disaster state semantic vectors corresponding to the ground disaster state marks before the coordinates to be analyzed, and the ground disaster state estimated vectors corresponding to the coordinates to be analyzed are output;
according to the ground disaster state prediction vector corresponding to the coordinate to be analyzed, a prediction possibility parameter corresponding to a ground disaster state sign on the coordinate to be analyzed is analyzed, wherein the prediction possibility parameter is used for reflecting a ground disaster state corresponding to the ground disaster state sign on the coordinate to be analyzed and a prediction correlation coefficient of a previous ground disaster state chain corresponding to the coordinate to be analyzed, and the previous ground disaster state chain comprises ground disaster states corresponding to each ground disaster state sign before the coordinate to be analyzed;
According to the estimated probability parameters corresponding to the ground disaster state marks on each coordinate to be analyzed in the sample ground disaster state mark chain, analyzing corresponding ground disaster state estimated errors; the ground disaster state prediction error can be obtained based on the estimated likelihood parameters respectively corresponding to the ground disaster state marks on the coordinates to be analyzed in the sample ground disaster state mark chain after the estimated likelihood parameters respectively corresponding to the ground disaster state marks on the coordinates to be analyzed in the sample ground disaster state mark chain are obtained, for example, the ground disaster state prediction error can be obtained based on the average value of the estimated likelihood parameters; the ground disaster state estimated error can be obtained based on the weighted average value of each estimated likelihood parameter; the ground disaster state prediction error can be obtained based on the median value in each prediction possibility parameter. It can be understood that the training objective of the network is to make the prediction possibility parameter corresponding to the ground disaster state mark on the coordinate to be analyzed larger and better, the prediction possibility parameter corresponding to the ground disaster state mark on the coordinate to be analyzed larger, which means that the network can accurately predict the ground disaster state mark on the coordinate to be analyzed based on the relevant features of each ground disaster state mark before the coordinate to be analyzed in the sample ground disaster state mark chain, which means that the network can more accurately predict the next ground disaster state based on the historical ground disaster state, and the network gradually has the ground disaster state prediction capability;
According to the ground disaster state estimation error, updating and optimizing the network parameters of the candidate ground disaster situation analysis network, and forming a corresponding updated ground disaster situation analysis network when the network parameters are matched with a pre-configured reference network training rule, for example, the network parameters of the candidate ground disaster situation analysis network can be updated and optimized along the direction of reducing the ground disaster state estimation error, so that the ground disaster state estimation error converges to a target value.
For example, the foregoing examples follow:
a local disaster state flag chain can be formed, and the local disaster state flag chains are arranged in time sequence as follows:
a status flag 1 corresponding to the seismic event 1;
a state mark 2 corresponding to the debris flow and landslide event 2;
a state flag 1 corresponding to the seismic event 3;
a state flag 0 corresponding to a ground disaster event;
a status flag 1 corresponding to the seismic event 4;
in the candidate ground disaster situation analysis network, a semantic vector cluster of a reference ground disaster state is arranged, wherein semantic vectors of different ground disaster states, such as earthquake and mud-rock flow/landslide, are contained. And (3) at the coordinate to be analyzed, matching the ground disaster state semantic vector corresponding to each ground disaster state mark positioned in front of the coordinate. Assuming that a state mark 1 corresponding to a seismic event 3 at a coordinate to be analyzed is selected, so that each disaster state mark positioned in front of the coordinate is a state mark 1 corresponding to the seismic event 1 and a state mark 2 corresponding to a debris flow and landslide event 2; and carrying out vector analysis processing on the ground disaster state semantic vectors corresponding to the two ground disaster state marks to obtain a ground disaster state estimated vector corresponding to the coordinates to be analyzed. Based on the ground disaster state prediction vector corresponding to the coordinate to be analyzed, the prediction possibility parameter of the ground disaster state sign (namely, the state sign 1 corresponding to the seismic event 3) on the coordinate can be analyzed. This parameter may reflect the likelihood of occurrence of the seismic event 3. Since the seismic event 3 has historically occurred, it is desirable to make the prediction probability parameter larger and better, i.e., the prediction capability of the characterization network, during the network update process.
For example, in some specific applications, the step of analyzing the corresponding ground disaster state prediction error according to the prediction likelihood parameters corresponding to the ground disaster state markers on each coordinate to be analyzed in the sample ground disaster state marker chain may further include the following detailed implementation details:
carrying out logarithmic operation on estimated probability parameters corresponding to the ground disaster state marks on each coordinate to be analyzed in the sample ground disaster state mark chain respectively to obtain corresponding logarithmic values;
and carrying out sum value calculation on the logarithmic values corresponding to the estimated probability parameters corresponding to the ground disaster state marks on each coordinate to be analyzed in the sample ground disaster state mark chain to obtain a target value, and determining a corresponding ground disaster state estimated error based on the target value, wherein the ground disaster state estimated error and the target value have a corresponding relationship with negative correlation, for example, the sum value between the ground disaster state estimated error and the target value is a fixed value, such as a value of 0 or 1.
For example, in some specific applications, the above step S110 may further include the following detailed implementation details:
Determining an original space mapping vector corresponding to each ground disaster state in a reference ground disaster state cluster, wherein the original space mapping vector can be formed by carrying out vector space mapping processing on related description data of the ground disaster state, so that the original space mapping vector corresponding to each ground disaster state can be obtained; in addition, the reference ground disaster state cluster may be formed by a manual configuration operation;
loading the original space mapping vectors corresponding to each ground disaster state into an update vector mining model, wherein the update vector mining model is formed based on corresponding network update processing;
for an original space mapping vector of any one of the original space mapping vectors, using the updated vector mining model, performing depth mining on the original space mapping vector corresponding to the ground disaster state, outputting the ground disaster state mining vector corresponding to the ground disaster state, performing dimension reduction processing on the ground disaster state mining vector corresponding to the ground disaster state, and outputting a ground disaster state semantic vector corresponding to the ground disaster state, wherein the depth mining can be that the original space mapping vector is subjected to low-dimension mapping, so that the original space mapping vector has better interpretability and stronger expression capability, that is, the ground disaster state mining vector has stronger semantic expression capability than the original space mapping vector, for example, the depth mining of the original space mapping vector can be realized through a coding neural network; the dimension reduction process is used for compressing the ground disaster state mining vector to a vector with uniform length, for example, the dimension reduction process is used for compressing the ground disaster state mining vector with several hundred dimensions to a ground disaster state semantic vector with several tens dimensions;
Combining to form a corresponding reference ground disaster state semantic vector cluster according to the ground disaster state semantic vectors respectively corresponding to the ground disaster states; that is, the reference ground disaster state semantic vector cluster includes ground disaster state semantic vectors corresponding to the respective ground disaster states.
For example, in some specific applications, the step of determining the original spatial mapping vector corresponding to each disaster state in the reference disaster state cluster may further include the following detailed implementation details:
determining a ground disaster state description data cluster corresponding to each ground disaster state in a reference ground disaster state cluster, wherein each ground disaster state description data in the ground disaster state description data cluster is used for describing attribute information of the ground disaster state (such as time, position, source and other specific information of an earthquake for an earthquake event, and information explaining the earthquake, namely, the earthquake, also known as earthquake and ground vibration, is vibration caused in the process of rapidly releasing energy of a crust, a natural phenomenon of earthquake waves can be generated during the process, the plate blocks on the earth are mutually extruded and collided, the edge of the plate block and the inner part of the plate block are caused to generate dislocation and rupture, the main reason for causing the earthquake is that the place where the earthquake starts to occur is called the source, the ground above the source is called the earthquake center, the place where the ground vibration of the destructive earthquake is most intense is called as the polar region, the area is often the area where the earthquake exists in the earthquake center), and the environmental information (such as the movement information of the crust, the information, such as the weather, the temperature, the illumination and the like) of the time interval where the ground disaster state occurs, the specific length of the earthquake is not limited by 10 days before;
For a ground disaster state description data cluster of any one of the ground disaster state description data clusters, performing vector space mapping on each ground disaster state description data in the ground disaster state description data cluster, outputting ground disaster state description data vectors corresponding to each ground disaster state description data respectively, for example, performing an emmbedding operation on each ground disaster state description data, and performing cascade combination on the ground disaster state description data vectors corresponding to each ground disaster state description data respectively to form an original space mapping vector corresponding to a ground disaster state corresponding to the ground disaster state description data cluster, such as { ground disaster state description data vector 1, ground disaster state description data vector 2 and ground disaster state description data vector 3}.
For example:
it is assumed that there are two ground disaster state description data clusters corresponding to the two ground disaster states of the earthquake and the debris flow. For a seismic state description data cluster, seismic event description data may be acquired that includes information:
specific information such as time, position, depth of a seismic source and the like of the earthquake;
the explanation of an earthquake, such as "earthquake, also called earth movement, earth vibration, is a natural phenomenon during which earthquake waves are generated during the process of rapidly releasing energy from the crust. The mutual extrusion collision between the earth upper plate and the plate causes the dislocation and cracking of the edge of the plate and the inside of the plate, which is the main cause of the earthquake. The place where the earthquake starts is called a seismic source, the ground just above the seismic source is called a epicenter, the place where the ground of the destructive earthquake vibrates most strongly is called a polar region, and the polar region is often the region where the epicenter is located.
For the mud-rock flow state description data cluster, mud-rock flow event description data including the following information may be collected:
specific information such as time, position, triggering reason and the like of the debris flow;
the explanation of the debris flow, such as "the debris flow is a natural phenomenon that a large amount of mixed substances such as soil and stones, sediment, water and the like rapidly flow due to mountain collapse, heavy rain, snow melting and the like. Debris flow typically occurs in steep slope areas where rocks and soil are washed and eroded by rain water to form flowable mixes. The mud-rock flow causes serious harm to the surrounding environment and the human living area, and can cause disasters such as land collapse, house collapse, road blockage and the like;
next, each ground disaster state description data is vector space mapped:
each ground disaster state description data may be subjected to an ebedding operation using Word embedding techniques (e.g., word2Vec, gloVe, etc.) to convert it into a fixed length vector representation.
And cascading and combining the vector representations of the ground disaster state description data to form an original space mapping vector corresponding to the ground disaster state description data cluster. For example, for a seismic state description data cluster, the following raw spatial mapping vectors may be derived: { seismic state description data vector 1, seismic state description data vector 2, seismic state description data vector 3,. }; for the mud-rock flow state description data cluster, the following original spatial mapping vector can be obtained: { debris flow state description data vector 1, debris flow state description data vector 2, debris flow state description data vector 3,.}.
Among these, the following seismic state description data is assumed:
seismic description 1: an earthquake, also called earth movement or earth vibration, is a vibration caused by the process of rapidly releasing energy from the crust. "
Seismic description 2: the mutual extrusion collision between the earth upper plate and the plate, which causes the dislocation and cracking of the plate edge and the plate inside, is the main cause of the earthquake. "
The description data can be segmented to obtain the following segmentation results:
seismic description 1 word segmentation results: "earthquake", "also known as" ground movement "," ground vibration "," is "," crust "," fast "," release "," energy "," process "," medium "," cause "," vibration "]; seismic description 2 word segmentation results: the terms "earth", "upper", "plate", "and", "between", "mutual", "extrusion", "impact", "causing", "plate", "edge", "and", "plate", "inner", "producing", "staggering", "and", "breaking", "causing", "earthquake", "main", "cause"
A pre-trained Word embedding model (e.g., word2Vec model) may then be used to convert each Word into a fixed length vector representation. These vector representations capture semantic and grammatical relations between words. Assuming that the Word2Vec model maps each Word into a 100-dimensional vector space, then the following embedded vectors of seismic state description data are derived:
Embedding vector of seismic description 1: [0.25, -0.12, 0.08, ], 0.92];
embedding vector of seismic description 2: [ -0.05, 0.36, -0.29,..., 0.76].
For example, in some specific applications, the above step S110 may further include the following detailed implementation details:
determining a sample ground disaster state combination, wherein the sample ground disaster state combination can comprise two ground disaster states, such as a first ground disaster state and a second ground disaster state, and the two ground disaster states can be selected arbitrarily; for example, two adjacent ground disaster states can be selected from the sample ground disaster state flag chain to form a sample ground disaster state combination, so that real ground disaster state relation data corresponding to the sample ground disaster state combination is that a correlation exists between the two sample ground disaster states;
loading a first ground disaster state in the sample ground disaster state combination so as to load the first vector mining model which needs to be updated, mining a ground disaster state semantic vector corresponding to the first ground disaster state, loading a second ground disaster state in the sample ground disaster state combination so as to load the second vector mining model which needs to be updated, mining a ground disaster state semantic vector corresponding to the second ground disaster state, and as mentioned above, performing vector space mapping on description data corresponding to the ground disaster state to obtain an original space mapping vector, then performing depth mining on the original space mapping vector corresponding to the ground disaster state to obtain a ground disaster state mining vector, and finally performing dimension reduction processing on the ground disaster state mining vector corresponding to the ground disaster state to obtain a ground disaster state semantic vector corresponding to the ground disaster state to obtain the ground disaster state semantic vector corresponding to the ground disaster state;
According to vector matching parameters, such as cosine similarity, between the ground disaster state semantic vectors corresponding to the first ground disaster state and the ground disaster state semantic vectors corresponding to the second ground disaster state, analyzing ground disaster state relation analysis data corresponding to the sample ground disaster state combination, wherein the ground disaster state relation analysis data are data which are analyzed and used for reflecting whether correlation exists between two sample ground disaster states in the sample ground disaster state combination, for example, the vector matching parameters are larger than or equal to a pre-configured reference vector matching parameter, such as 0.6, and can be considered as data with correlation between two sample ground disaster states in the sample ground disaster state combination, and the vector matching parameters are smaller than the pre-configured reference vector matching parameter, and can be considered as data without correlation between two sample ground disaster states in the sample ground disaster state combination;
according to distinguishing information between the ground disaster state relation real data corresponding to the sample ground disaster state combination and ground disaster state relation analysis data corresponding to the sample ground disaster state combination, updating and optimizing parameters of the first vector mining model to be updated and the second vector mining model to be updated, and forming an updated first vector mining model and an updated second vector mining model when the parameters are matched with a predetermined first configuration network training rule, wherein the ground disaster state relation real data are data which are truly used for reflecting whether correlation exists between two sample ground disaster states in the sample ground disaster state combination, namely determining errors positively correlated with the distinguishing information, then updating and optimizing parameters of the first vector mining model to be updated and the second vector mining model to be updated along a direction for reducing the errors, so that the errors are converged, and performing supervision on the first vector mining model to be updated and the second vector mining model to be updated based on the sample ground disaster state combination and the corresponding ground disaster state relation real data, and extracting the second vector mining model to be updated quickly and the second vector mining model to be updated can be used for extracting the corresponding ground disaster state quickly;
And determining a vector mining model from the updated first vector mining model and the updated second vector mining model, wherein any vector mining model can be used as the updated vector mining model.
For example, in some specific applications, the above step S110 may further include the following detailed implementation details:
determining coordinate semantic vectors corresponding to the ground disaster state mark coordinates before the coordinates to be analyzed by using the candidate ground disaster situation analysis network, for example, carrying out the ebedding operation on the ground disaster state mark coordinates to obtain the corresponding coordinate semantic vectors;
vector aggregation processing is carried out on the ground disaster state semantic vector corresponding to the ground disaster state sign and the coordinate semantic vector corresponding to the ground disaster state sign coordinate corresponding to the ground disaster state sign, and the state coordinate aggregation semantic vector corresponding to each ground disaster state sign before the coordinate to be analyzed in the sample ground disaster state sign chain is output; by way of example, a ground disaster state semantic vector corresponding to the ground disaster state flag and a coordinate semantic vector corresponding to a ground disaster state flag coordinate corresponding to the ground disaster state flag may be subjected to superposition operation, so as to obtain a state coordinate aggregation semantic vector.
For example, in some specific applications, in the case of the foregoing examples, the step of performing vector analysis processing on the ground disaster state semantic vector corresponding to each ground disaster state flag located before the coordinate to be analyzed and outputting the ground disaster state estimated vector corresponding to the coordinate to be analyzed may further include the following detailed implementation details:
and carrying out vector analysis processing on the state coordinate aggregation semantic vectors corresponding to the ground disaster state marks positioned in front of the coordinates to be analyzed, and outputting the ground disaster state estimated vector corresponding to the coordinates to be analyzed.
The state coordinate aggregation semantic vector is fused with a ground disaster state semantic vector and a coordinate semantic vector, and through the state coordinate aggregation semantic vector, the network can learn semantic information of a ground disaster state sign in input data, and can capture coordinate information of the ground disaster state sign (namely, a precedence relationship among the ground disaster state signs) in the input data, so that a correlation vector of a next ground disaster state related to a historical ground disaster state can be better understood and generated.
For example, in some specific applications, the step of analyzing the estimated likelihood parameter corresponding to the ground disaster status flag on the coordinate to be analyzed according to the ground disaster status estimated vector corresponding to the coordinate to be analyzed may further include the following detailed implementation contents:
Vector space conversion processing is carried out on the ground disaster state estimated vector corresponding to the coordinate to be analyzed, and candidate ground disaster state parameter distribution corresponding to the coordinate to be analyzed is output, wherein the candidate ground disaster state parameter distribution comprises characterization parameters respectively corresponding to each ground disaster state in a reference ground disaster state cluster; the vector space conversion process is used for mapping the ground disaster state estimated vector to a vector space with preset dimension, carrying out vector space conversion process on the ground disaster state estimated vector to obtain candidate ground disaster state parameter distribution, wherein the dimension of the candidate ground disaster state parameter distribution is the preset dimension, and the dimension of the candidate ground disaster state parameter distribution is equal to the size of the reference ground disaster state cluster, wherein the size of the reference ground disaster state cluster reflects the number of ground disaster states in the reference ground disaster state cluster;
performing characterization parameter reduction processing on the candidate ground disaster state parameter distribution, and outputting target ground disaster state parameter distribution corresponding to the coordinates to be analyzed, wherein the target ground disaster state parameter distribution comprises estimated possibility parameters corresponding to each ground disaster state in the reference ground disaster state cluster, and the reference ground disaster state cluster comprises ground disaster states corresponding to each ground disaster state mark in the sample ground disaster state mark chain; the characterization parameter reduction process may be performed by a Softmax function or the like, for example;
Analyzing estimated probability parameters corresponding to the ground disaster state marks on the coordinates to be analyzed based on the target ground disaster state parameter distribution; for example, the estimated likelihood parameter corresponding to the ground disaster state of the ground disaster state sign chain on the coordinate to be analyzed may be obtained from the target ground disaster state parameter distribution, and the estimated likelihood parameter corresponding to the ground disaster state sign on the coordinate to be analyzed may be used as the estimated likelihood parameter corresponding to the ground disaster state sign.
For example:
assume that the reference ground disaster state cluster comprises 5 ground disaster states including earthquake, debris flow, heavy rain, fire and no ground disaster. Performing ground disaster state estimation on a certain coordinate to be analyzed:
vector space conversion processing: vector space conversion processing is carried out on the ground disaster state estimated vector corresponding to the coordinates to be analyzed, the ground disaster state estimated vector is mapped to a vector space with preset dimensions, 5 ground disaster states are in the reference ground disaster state cluster, the vector space with the preset dimensions being 5 is selected, and a candidate ground disaster state parameter distribution is obtained through the conversion processing, wherein each dimension corresponds to one ground disaster state, and the characterization parameters represent the characteristics of the ground disaster state;
for example, after vector space conversion, the following candidate ground disaster state parameter distribution is obtained:
[0.2, 0.4, 0.1, 0.2, 0.1]；
In this example, the candidate ground disaster state parameter distribution is a vector with a length of 5, which corresponds to the earthquake, debris flow, heavy rain, fire and ground disaster free states, respectively. Each value represents a characteristic parameter of the corresponding disaster state;
characterization parameter reduction processing: and carrying out characterization parameter reduction processing on the candidate ground disaster state parameter distribution to obtain a target ground disaster state parameter distribution, and normalizing the candidate parameter by using a Softmax function to convert the candidate parameter into a probability distribution. Continuing to take the candidate ground disaster state parameter distribution as an example, performing reduction processing through a Softmax function to obtain the following target ground disaster state parameter distribution:
[0.174, 0.348, 0.087, 0.174, 0.087]
in this example, the target ground disaster state parameter distribution is still a vector of length 5, each value representing a predicted likelihood parameter of the corresponding ground disaster state, the parameters representing the relative weights or probabilities of the individual ground disaster states on the coordinates to be analyzed.
Analyzing and predicting the probability parameters: based on the target ground disaster state parameter distribution, the estimated probability parameters corresponding to the ground disaster state marks on the coordinates to be analyzed can be analyzed, and the estimated probability parameters corresponding to the earthquakes in the sample ground disaster state mark chains are supposed to be obtained. According to the distribution of the target ground disaster state parameters, the estimated probability parameter corresponding to the earthquake is 0.174.
For example, in some specific applications, the candidate ground disaster situation analysis network includes a vector space mapping unit, a vector analysis unit, and a vector estimation unit, where the vector space mapping unit includes a first vector space mapping layer and a second vector space mapping layer, the first vector space mapping layer is used to determine a ground disaster state semantic vector (specific determination process, as described in the foregoing related description), the second vector space mapping layer is used to determine a coordinate semantic vector (specific determination process, as described in the foregoing related description), and the vector analysis unit is used to perform a vector analysis process, and the vector estimation unit is used to analyze an estimated likelihood parameter (specific analysis process, as described in the foregoing related description).
Based on the above network architecture, the step of updating and optimizing the network parameters of the candidate ground disaster situation analysis network according to the ground disaster state estimation error, and forming a corresponding updated ground disaster situation analysis network when matching with a pre-configured reference network training rule may further include the following detailed implementation contents:
updating and optimizing parameters of the second vector space mapping layer, the vector analysis unit and the vector estimation unit in the candidate ground disaster situation analysis network according to the ground disaster state estimation error, and forming a corresponding updated ground disaster situation analysis network when the parameters are matched with a pre-configured reference network training rule; the first vector space mapping layer is used for matching the ground disaster state semantic vector, namely extracting from the reference ground disaster state semantic vector cluster, so that updating and optimizing of parameters are not needed, and the second vector space mapping layer, the vector analysis unit and the vector estimation unit all have data mapping processes and need updating and optimizing.
For example, in some specific applications, the vector estimation unit is configured to analyze, according to the output vector of the vector analysis unit (i.e. the ground disaster state estimation vector), estimation likelihood parameters corresponding to each ground disaster state in the reference ground disaster state cluster, and based on this, the above-mentioned step S110 may further include the following detailed implementation details:
after the updated ground disaster situation analysis network is formed, under the condition that the extended ground disaster state is allocated to the reference ground disaster state cluster, adding a ground disaster state semantic vector corresponding to the extended ground disaster state from the reference ground disaster state semantic vector cluster, wherein 5 ground disaster states are respectively earthquake, debris flow, storm, fire and ground disaster-free, such as ground disaster states with increased muscle ground subsidence, in the reference ground disaster state semantic vector cluster, and therefore, the ground disaster state semantic vector corresponding to the ground disaster state with increased ground subsidence is also required to be added in the reference ground disaster state semantic vector cluster;
loading a comparison ground disaster state mark chain into the updated ground disaster situation analysis network to form estimated probability parameters of each ground disaster state in the expanded reference ground disaster state cluster relative to each coordinate to be analyzed, wherein the comparison ground disaster state mark chain can be the sample ground disaster state mark chain or can be different from the sample ground disaster state mark chain, and as mentioned above, the estimated probability parameters of the coordinates to be analyzed are 5, and the estimated probability parameters of the coordinates to be analyzed are 6 after expansion;
Analyzing estimated probability parameters corresponding to the ground disaster state marks on the coordinates to be analyzed in the compared ground disaster state mark chain from the estimated probability parameters of the ground disaster states in the expanded reference ground disaster state cluster relative to the coordinates to be analyzed;
calculating a corresponding ground disaster state expansion error according to the estimated probability parameters corresponding to the ground disaster state marks on each coordinate to be analyzed in the comparative ground disaster state mark chain, wherein the specific determination process can be the same as the above-mentioned process of analyzing the corresponding ground disaster state estimated error according to the estimated probability parameters corresponding to the ground disaster state marks on each coordinate to be analyzed in the sample ground disaster state mark chain, and the detailed description is omitted herein;
according to the ground disaster state expansion error, updating and optimizing parameters of the vector estimation unit in the updated ground disaster situation analysis network, and forming an updated ground disaster situation analysis network matched with the expanded reference ground disaster state cluster when the parameters are matched with a second preset configuration network training rule, namely updating parameters along the direction of reducing the ground disaster state expansion error; in the process of updating parameters, only the parameters of the vector estimation unit are updated and optimized, and the updated ground disaster situation analysis network is finely tuned under the condition of ensuring the data processing efficiency, so that the updated ground disaster situation analysis network can adapt to the extended ground disaster state.
For example, in some specific applications, the specific analysis process of the vector analysis unit, that is, the vector analysis processing is performed on the state coordinate aggregate semantic vector corresponding to each disaster status flag located before the coordinate to be analyzed, and the step of outputting the disaster status prediction vector corresponding to the coordinate to be analyzed may further include details of implementation described below (the state coordinate aggregate semantic vector corresponding to each disaster status flag located before the coordinate to be analyzed includes a first state coordinate aggregate semantic vector and a second state coordinate aggregate semantic vector are illustrated):
performing linear mapping on the first state coordinate aggregation semantic vector to obtain a first query vector, a first key vector and a first value vector (the first query vector, the first key vector and the first value vector can be obtained by multiplying three weight matrixes, and the three weight matrixes can be used as parameters of the vector analysis unit for optimization updating);
linearly mapping the second state coordinate aggregation semantic vector to obtain an er-th query vector, a second key vector and a second value vector (as described above);
performing multiplication operation on the first query vector and the first key vector to obtain a corresponding first similarity matrix, and performing multiplication operation on the second query vector and the second key vector to obtain a corresponding second similarity matrix;
Respectively carrying out normalization processing on the row parameters on the first similarity matrix and the second similarity matrix to obtain a corresponding first attention weight matrix and a corresponding second attention weight matrix;
weighting (e.g., multiplying) the first value vector based on the first attention weight matrix, outputting a corresponding first weighted value vector, and weighting (e.g., multiplying) the second value vector based on the second attention weight matrix, outputting a corresponding second weighted value vector;
cascade-combining the first weight vector and the second weight vector, and outputting a corresponding focus mining vector, such as { the first weight vector, the second weight vector };
multiplying the focusing mining vector by a linear mapping weight matrix included in the vector analysis unit, adding the multiplied result and a linear mapping bias parameter included in the vector analysis unit, and then activating the added result to obtain a linear focusing mining vector;
and performing superposition operation on the linear focusing mining vector and the focusing mining vector, and performing normalization processing on a superposition operation result to obtain a ground disaster state estimated vector corresponding to the coordinate to be analyzed.
With reference to fig. 3, the embodiment of the invention also provides a ground disaster situation analysis device applied to data sharing, which can be applied to the ground disaster situation analysis system applied to data sharing. The ground disaster situation analysis device applied to data sharing may include:
the network updating processing module is used for forming a sample ground disaster state mark chain based on historical ground disaster monitoring data acquired from a ground disaster monitoring data sharing system, loading the sample ground disaster state mark chain into a candidate ground disaster situation analysis network to perform network updating processing to form a corresponding updated ground disaster situation analysis network, wherein the sample ground disaster state mark chain comprises ground disaster state marks which have correlation and are sequentially arranged, and the ground disaster state marks in the sample ground disaster state mark chain comprise at least two of state marks corresponding to various ground disaster events and state marks corresponding to non-ground disaster events;
the ground disaster monitoring data loading module is used for forming a current ground disaster state mark chain based on the current ground disaster monitoring data acquired from the ground disaster monitoring data sharing system, and loading the current ground disaster state mark chain into the updated ground disaster situation analysis network, wherein the current ground disaster monitoring data is aimed at a target monitoring area;
The ground disaster monitoring data analysis module is used for analyzing estimated probability parameter distribution corresponding to the current ground disaster state mark chain by utilizing the updated ground disaster situation analysis network, and each estimated probability parameter in the estimated probability parameter distribution is used for reflecting the probability of the ground disaster state corresponding to one ground disaster state mark;
the ground disaster state determining module is used for determining a target ground disaster state corresponding to the current ground disaster state mark chain based on the estimated possibility parameter distribution, and the target ground disaster state is used for reflecting ground disaster events or non-ground disaster events which occur in the target monitoring area.
In summary, the ground disaster situation analysis method and system for data sharing provided by the invention can utilize the neural network to analyze and predict the ground disaster situation, so that the reliability of the ground disaster situation analysis is higher compared with the conventional technical scheme based on manual analysis; in addition, in the process of network updating, the sample ground disaster state mark chains are aimed at, and the ground disaster state marks included in the sample ground disaster state mark chains are related and arranged in sequence, so that the related relationship between various ground disaster events and non-ground disaster events can be fully learned in the process of network updating, and the related relationship is considered when the current ground disaster state mark chains are analyzed and estimated, so that the reliability of ground disaster situation analysis can be improved.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. The ground disaster situation analysis method applied to data sharing is characterized by comprising the following steps:
forming a sample ground disaster state mark chain based on historical ground disaster monitoring data acquired from a ground disaster monitoring data sharing system, loading the sample ground disaster state mark chain into a candidate ground disaster situation analysis network to perform network updating processing to form a corresponding updated ground disaster situation analysis network, wherein the sample ground disaster state mark chain comprises ground disaster state marks which have correlation and are sequentially arranged, and the ground disaster state marks in the sample ground disaster state mark chain comprise at least two of state marks corresponding to various ground disaster events and state marks corresponding to non-ground disaster events;
forming a current ground disaster state mark chain based on the current ground disaster monitoring data acquired from the ground disaster monitoring data sharing system, and loading the current ground disaster state mark chain into the updated ground disaster situation analysis network, wherein the current ground disaster monitoring data is specific to a target monitoring area;
Analyzing estimated probability parameter distribution corresponding to the current ground disaster state sign chain by using the updated ground disaster situation analysis network, wherein each estimated probability parameter in the estimated probability parameter distribution is used for reflecting the probability of the ground disaster state corresponding to one ground disaster state sign;
and determining a target ground disaster state corresponding to the current ground disaster state mark chain based on the estimated probability parameter distribution, wherein the target ground disaster state is used for reflecting ground disaster events or non-ground disaster events which occur in the target monitoring area.
2. The ground disaster situation analyzing method for data sharing according to claim 1, wherein the step of forming a sample ground disaster state flag chain based on historical ground disaster monitoring data acquired from a ground disaster monitoring data sharing system, and loading the sample ground disaster state flag chain into a candidate ground disaster situation analyzing network for network updating processing to form a corresponding updated ground disaster situation analyzing network comprises:
forming a sample ground disaster state mark chain based on historical ground disaster monitoring data acquired from a ground disaster monitoring data sharing system, and loading the sample ground disaster state mark chain into a candidate ground disaster situation analysis network;
Matching ground disaster state semantic vectors corresponding to ground disaster state marks positioned in front of coordinates to be analyzed in the sample ground disaster state mark chain from a reference ground disaster state semantic vector cluster by utilizing the candidate ground disaster situation analysis network, wherein the coordinates to be analyzed belong to the ground disaster state mark coordinates corresponding to the sample ground disaster state mark chain;
vector analysis processing is carried out on the ground disaster state semantic vectors corresponding to the ground disaster state marks before the coordinates to be analyzed, and the ground disaster state estimated vectors corresponding to the coordinates to be analyzed are output;
according to the ground disaster state prediction vector corresponding to the coordinate to be analyzed, a prediction possibility parameter corresponding to a ground disaster state sign on the coordinate to be analyzed is analyzed, wherein the prediction possibility parameter is used for reflecting a ground disaster state corresponding to the ground disaster state sign on the coordinate to be analyzed and a prediction correlation coefficient of a previous ground disaster state chain corresponding to the coordinate to be analyzed, and the previous ground disaster state chain comprises ground disaster states corresponding to each ground disaster state sign before the coordinate to be analyzed;
according to the estimated probability parameters corresponding to the ground disaster state marks on each coordinate to be analyzed in the sample ground disaster state mark chain, analyzing corresponding ground disaster state estimated errors;
And updating and optimizing network parameters of the candidate ground disaster situation analysis network according to the ground disaster state estimation error, and forming a corresponding updated ground disaster situation analysis network when the network parameters are matched with a preset reference network training rule.
3. The ground disaster situation analyzing method for data sharing according to claim 2, wherein the step of forming a sample ground disaster state flag chain based on historical ground disaster monitoring data acquired from the ground disaster monitoring data sharing system, and loading the sample ground disaster state flag chain into a candidate ground disaster situation analyzing network for network updating processing to form a corresponding updated ground disaster situation analyzing network further comprises:
determining an original space mapping vector corresponding to each ground disaster state in a reference ground disaster state cluster;
loading the original space mapping vectors corresponding to each ground disaster state into an update vector mining model, wherein the update vector mining model is formed based on corresponding network update processing;
for the original space mapping vector of any one of the original space mapping vectors, performing deep mining on the original space mapping vector corresponding to the ground disaster state by utilizing the updated vector mining model, outputting the ground disaster state mining vector corresponding to the ground disaster state, performing dimension reduction processing on the ground disaster state mining vector corresponding to the ground disaster state, and outputting the ground disaster state semantic vector corresponding to the ground disaster state;
And combining the ground disaster state semantic vectors corresponding to the ground disaster states to form corresponding reference ground disaster state semantic vector clusters.
4. The method for analyzing a ground disaster situation for data sharing as set forth in claim 3, wherein the step of determining an original spatial mapping vector corresponding to each ground disaster state in the reference ground disaster state cluster includes:
determining a ground disaster state description data cluster corresponding to each ground disaster state in a reference ground disaster state cluster, wherein each ground disaster state description data in the ground disaster state description data cluster is used for describing attribute information of the ground disaster state and environment information of a time interval in which the ground disaster state occurs;
for any one of the ground disaster state description data clusters, performing vector space mapping on each ground disaster state description data in the ground disaster state description data clusters, outputting ground disaster state description data vectors corresponding to each ground disaster state description data respectively, and performing cascade combination on the ground disaster state description data vectors corresponding to each ground disaster state description data respectively to form an original space mapping vector corresponding to the ground disaster state description data clusters.
5. The ground disaster situation analyzing method for data sharing according to claim 3, wherein the step of forming a sample ground disaster state flag chain based on historical ground disaster monitoring data acquired from the ground disaster monitoring data sharing system, and loading the sample ground disaster state flag chain into a candidate ground disaster situation analyzing network for network updating processing, and forming a corresponding updated ground disaster situation analyzing network further comprises:
determining a sample ground disaster state combination;
loading a first ground disaster state in the sample ground disaster state combination so as to load the first vector mining model to be updated, mining a ground disaster state semantic vector corresponding to the first ground disaster state, and loading a second ground disaster state in the sample ground disaster state combination so as to load the second vector mining model to be updated, and mining a ground disaster state semantic vector corresponding to the second ground disaster state;
according to vector matching parameters between a ground disaster state semantic vector corresponding to the first ground disaster state and a ground disaster state semantic vector corresponding to the second ground disaster state, analyzing ground disaster state relation analysis data corresponding to the sample ground disaster state combination, wherein the ground disaster state relation analysis data are data which are analyzed and used for reflecting whether a correlation exists between two sample ground disaster states in the sample ground disaster state combination;
According to the distinguishing information between the ground disaster state relation real data corresponding to the sample ground disaster state combination and the ground disaster state relation analysis data corresponding to the sample ground disaster state combination, updating and optimizing parameters of the first vector mining model to be updated and the second vector mining model to be updated, and forming an updated first vector mining model and an updated second vector mining model when the parameters are matched with a predetermined first configuration network training rule, wherein the ground disaster state relation real data are real data used for reflecting whether correlation exists between two sample ground disaster states in the sample ground disaster state combination;
and determining a vector mining model from the updated first vector mining model and the updated second vector mining model as the updated vector mining model.
6. The ground disaster situation analyzing method for data sharing according to claim 2, wherein the step of forming a sample ground disaster state flag chain based on historical ground disaster monitoring data acquired from the ground disaster monitoring data sharing system, and loading the sample ground disaster state flag chain into a candidate ground disaster situation analyzing network for network updating processing to form a corresponding updated ground disaster situation analyzing network further comprises:
Determining coordinate semantic vectors corresponding to each ground disaster state sign coordinate before the coordinate to be analyzed by utilizing the candidate ground disaster situation analysis network;
vector aggregation processing is carried out on the ground disaster state semantic vector corresponding to the ground disaster state sign and the coordinate semantic vector corresponding to the ground disaster state sign coordinate corresponding to the ground disaster state sign, and the state coordinate aggregation semantic vector corresponding to each ground disaster state sign before the coordinate to be analyzed in the sample ground disaster state sign chain is output;
the step of carrying out vector analysis processing on the ground disaster state semantic vectors corresponding to the ground disaster state marks before the coordinates to be analyzed and outputting the ground disaster state estimated vectors corresponding to the coordinates to be analyzed comprises the following steps:
and carrying out vector analysis processing on the state coordinate aggregation semantic vectors corresponding to the ground disaster state marks positioned in front of the coordinates to be analyzed, and outputting the ground disaster state estimated vector corresponding to the coordinates to be analyzed.
7. The method for analyzing a ground disaster situation for data sharing according to claim 2, wherein the step of analyzing the estimated likelihood parameter corresponding to the ground disaster state flag on the coordinate to be analyzed according to the ground disaster state estimated vector corresponding to the coordinate to be analyzed comprises:
Vector space conversion processing is carried out on the ground disaster state estimated vector corresponding to the coordinate to be analyzed, and candidate ground disaster state parameter distribution corresponding to the coordinate to be analyzed is output, wherein the candidate ground disaster state parameter distribution comprises characterization parameters respectively corresponding to each ground disaster state in a reference ground disaster state cluster;
performing characterization parameter reduction processing on the candidate ground disaster state parameter distribution, and outputting target ground disaster state parameter distribution corresponding to the coordinates to be analyzed, wherein the target ground disaster state parameter distribution comprises estimated possibility parameters corresponding to each ground disaster state in the reference ground disaster state cluster, and the reference ground disaster state cluster comprises ground disaster states corresponding to each ground disaster state mark in the sample ground disaster state mark chain;
and analyzing estimated probability parameters corresponding to the ground disaster state marks on the coordinates to be analyzed based on the target ground disaster state parameter distribution.
8. The ground disaster situation analysis method for data sharing according to claim 2, wherein the candidate ground disaster situation analysis network comprises a vector space mapping unit, a vector analysis unit and a vector estimation unit, the vector space mapping unit comprises a first vector space mapping layer and a second vector space mapping layer, the first vector space mapping layer is used for determining a ground disaster state semantic vector, the second vector space mapping layer is used for determining a coordinate semantic vector, the vector analysis unit is used for carrying out vector analysis processing, and the vector estimation unit is used for analyzing an estimated likelihood parameter;
The step of updating and optimizing the network parameters of the candidate ground disaster situation analysis network according to the ground disaster state estimation error, and forming a corresponding updated ground disaster situation analysis network when the network parameters are matched with a pre-configured reference network training rule comprises the following steps:
and updating and optimizing parameters of the second vector space mapping layer, the vector analysis unit and the vector estimation unit in the candidate ground disaster situation analysis network according to the ground disaster state estimation error, and forming a corresponding updated ground disaster situation analysis network when the parameters are matched with a pre-configured reference network training rule.
9. The method for analyzing ground disaster situations for data sharing according to claim 8, wherein the vector estimation unit is configured to analyze estimated likelihood parameters corresponding to each ground disaster state in the reference ground disaster state cluster according to an output vector of the vector analysis unit;
the step of forming a sample ground disaster state mark chain based on the historical ground disaster monitoring data acquired from the ground disaster monitoring data sharing system, loading the sample ground disaster state mark chain into a candidate ground disaster situation analysis network for network updating processing to form a corresponding updated ground disaster situation analysis network, and further comprises the steps of:
After the updated ground disaster situation analysis network is formed, adding a ground disaster state semantic vector corresponding to the extended ground disaster state from the reference ground disaster state semantic vector cluster under the condition that the extended ground disaster state is distributed to the reference ground disaster state cluster;
loading a ground disaster state comparison mark chain into the updated ground disaster situation analysis network to form estimated probability parameters of each ground disaster state in the expanded reference ground disaster state cluster relative to each coordinate to be analyzed;
analyzing estimated probability parameters corresponding to the ground disaster state marks on the coordinates to be analyzed in the compared ground disaster state mark chain from the estimated probability parameters of the ground disaster states in the expanded reference ground disaster state cluster relative to the coordinates to be analyzed;
calculating a corresponding ground disaster state expansion error according to the estimated probability parameters corresponding to the ground disaster state marks on each coordinate to be analyzed in the ground disaster state mark comparison chain;
and updating and optimizing parameters of the vector estimation unit in the updated ground disaster situation analysis network according to the ground disaster state expansion error, and forming the updated ground disaster situation analysis network adapted to the expanded reference ground disaster state cluster when the parameters are matched with a second preset configuration network training rule.
10. A ground disaster situation analysis system for data sharing, comprising a processor and a memory, wherein the memory is configured to store a computer program, and the processor is configured to execute the computer program to implement the ground disaster situation analysis method for data sharing according to any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311235261.7A CN116976837B (en) | 2023-09-25 | 2023-09-25 | Ground disaster situation analysis method and system applied to data sharing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311235261.7A CN116976837B (en) | 2023-09-25 | 2023-09-25 | Ground disaster situation analysis method and system applied to data sharing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116976837A true CN116976837A (en) | 2023-10-31 |
| CN116976837B CN116976837B (en) | 2024-01-30 |
Family
ID=88481837
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311235261.7A Active CN116976837B (en) | 2023-09-25 | 2023-09-25 | Ground disaster situation analysis method and system applied to data sharing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116976837B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130132045A1 (en) * | 2011-11-21 | 2013-05-23 | International Business Machines Corporation | Natural Disaster Forecasting |
| CN108961688A (en) * | 2018-07-13 | 2018-12-07 | 福建特力惠信息科技股份有限公司 | A kind of big data support under Geological Hazards Monitoring and method for early warning |
| CN109872509A (en) * | 2019-04-02 | 2019-06-11 | 西安邮电大学 | Massif Geological Hazards Monitoring and early warning system and method based on the twin driving of number |
| CN111880210A (en) * | 2020-08-05 | 2020-11-03 | 中国南方电网有限责任公司 | Ground disaster monitoring and processing method and device for power transmission line, early warning system and equipment |
| CN113159431A (en) * | 2021-04-28 | 2021-07-23 | 宁夏回族自治区国土资源调查监测院 | Analysis early warning method and device based on ground disaster data |
| CN113393037A (en) * | 2021-06-16 | 2021-09-14 | 潍坊科技学院 | Regional geological disaster trend prediction method and system |
| CN113723446A (en) * | 2021-07-15 | 2021-11-30 | 杭州鲁尔物联科技有限公司 | Ground disaster monitoring and early warning method and device, computer equipment and storage medium |
| CN114118576A (en) * | 2021-11-25 | 2022-03-01 | 深圳市地质局 | Regional geological disaster trend prediction method and system |
-
2023
- 2023-09-25 CN CN202311235261.7A patent/CN116976837B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130132045A1 (en) * | 2011-11-21 | 2013-05-23 | International Business Machines Corporation | Natural Disaster Forecasting |
| CN108961688A (en) * | 2018-07-13 | 2018-12-07 | 福建特力惠信息科技股份有限公司 | A kind of big data support under Geological Hazards Monitoring and method for early warning |
| CN109872509A (en) * | 2019-04-02 | 2019-06-11 | 西安邮电大学 | Massif Geological Hazards Monitoring and early warning system and method based on the twin driving of number |
| CN111880210A (en) * | 2020-08-05 | 2020-11-03 | 中国南方电网有限责任公司 | Ground disaster monitoring and processing method and device for power transmission line, early warning system and equipment |
| CN113159431A (en) * | 2021-04-28 | 2021-07-23 | 宁夏回族自治区国土资源调查监测院 | Analysis early warning method and device based on ground disaster data |
| CN113393037A (en) * | 2021-06-16 | 2021-09-14 | 潍坊科技学院 | Regional geological disaster trend prediction method and system |
| CN113723446A (en) * | 2021-07-15 | 2021-11-30 | 杭州鲁尔物联科技有限公司 | Ground disaster monitoring and early warning method and device, computer equipment and storage medium |
| CN114118576A (en) * | 2021-11-25 | 2022-03-01 | 深圳市地质局 | Regional geological disaster trend prediction method and system |
Non-Patent Citations (1)
| Title |
|---|
| 何文熹 等: "基于多源遥感数据的矿区地质灾害变化态势预测系统", 《自动化与仪器仪表》, no. 8, pages 58 - 61 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116976837B (en) | 2024-01-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Dodangeh et al. | Novel hybrid intelligence models for flood-susceptibility prediction: Meta optimization of the GMDH and SVR models with the genetic algorithm and harmony search | |
| Martinotti et al. | Landslides, floods and sinkholes in a karst environment: the 1–6 September 2014 Gargano event, southern Italy | |
| Do Hoai et al. | Downscaling global weather forecast outputs using ANN for flood prediction | |
| Lin et al. | Development of a real-time regional-inundation forecasting model for the inundation warning system | |
| Anagnostou et al. | Using high‐resolution satellite precipitation for flood frequency analysis: Case study over the Connecticut River Basin | |
| Mustafa et al. | Extreme rainfalls as a cause of urban flash floods; a case study of the Erbil-Kurdistan region of Iraq | |
| Azam et al. | Artificial intelligence based techniques for earthquake prediction: a review | |
| Ahmad et al. | Prediction of slope stability using Tree Augmented Naive-Bayes classifier: Modeling and performance evaluation | |
| Yang et al. | Joint occurrence probability analysis of typhoon-induced storm surges and rainstorms using trivariate Archimedean copulas | |
| Wang et al. | Rapid prediction of urban flood based on disaster-breeding environment clustering and Bayesian optimized deep learning model in the coastal city | |
| CN116976837B (en) | Ground disaster situation analysis method and system applied to data sharing | |
| Gao et al. | Landslide risk evaluation in Shenzhen based on stacking ensemble learning and InSAR | |
| Mu et al. | Simulation, precursor analysis and targeted observation sensitive area identification for two types of ENSO using ENSO-MC v1. 0 | |
| Lyu et al. | Flood susceptibility prediction using tree-based machine learning models in the GBA | |
| Yan et al. | A multiobjective evolutionary optimization method based critical rainfall thresholds for debris flows initiation | |
| Jang et al. | A probabilistic model for real‐time flood warning based on deterministic flood inundation mapping | |
| Kim et al. | Probabilistic coastal storm surge analyses using synthesized tracks based on historical typhoon parameters | |
| Bárdossy et al. | Circulation patterns identified by spatial rainfall and ocean wave fields in Southern Africa | |
| Zhang et al. | Pixel-CRN: A new machine learning approach for convective storm nowcasting | |
| Huang et al. | Landslide susceptibility evaluation using different slope units based on BP neural network | |
| Makris et al. | Climate change impacts on the coastal sea level extremes of the east-central Mediterranean Sea | |
| Nofal et al. | BIM-GIS integration approach for high-fidelity wind hazard modeling at the community-level | |
| Ishibashi | Framework for risk assessment of economic loss from structures damaged by rainfall-induced landslides using machine learning | |
| Pasang et al. | Information value model based landslide susceptibility mapping at Phuentsholing, Bhutan | |
| Bulteau et al. | Spatial extreme value analysis of significant wave heights along the French coast |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |