Geological disaster early warning method and system

Abstract

According to the geological disaster early warning method and system provided by the invention, each piece of geological disaster hidden danger point description information in an example geological disaster hidden danger point description information set has a corresponding stage type mark, and the stage type mark can correspond to the early warning object in the geological disaster hidden danger point description information. The method is used for distinguishing the stage information of the early warning object in the description information of the geological disaster hidden danger point. According to the embodiment of the invention, each piece of geological disaster hidden danger point description information in the example geological disaster hidden danger point description information set can have the current division label corresponding to the early warning object corresponding to the geological disaster hidden danger point description information, and the reliability of the barrel dividing result of the geological disaster early warning analysis network can be compared according to the current division label. If so, the corresponding loss can be determined. By accurately determining the loss, the accuracy and the confidence coefficient of the early warning result of the potential hazard point of the ground disaster can be guaranteed.

Description

Technical Field

[0001] This application relates to the technical field of geological disaster warning, and more specifically, to a geological disaster warning method and system. Background Art

[0002] Geological disaster warning technology is a three-dimensional solution for geological disaster warning, prediction, prevention and control, which is centered on computer software technology and combines professional geological disaster detection equipment and data analysis technology. The software is mainly implemented using J2EE technology based on the JAVA6.0 language, the ExtJS technology is used for some page presentation layers, the rain gauge synchronization function is implemented using the Webservice technology, and the presentation of hidden danger points, monitoring points, and meteorological distribution information is implemented using the innovative technology FakeGIS based on the concepts of Flex and GIS.

[0003] In the actual operation process, due to various interference factors in mountainous areas, for example, the terrain in mountainous areas is complex, and data may not be accurately obtained at some locations. In areas with lush forests in mountainous areas, the information on the ground may be blocked and cannot be obtained, etc. These interference factors lead to low accuracy of the obtained data, and there are also errors in the calculation of geological disaster warning technology. How to reduce interference and correct errors through calculation to improve the accuracy of early warning is an urgent problem to be solved at present. Summary of the Invention

[0004] To improve the technical problems existing in the related art, this application provides a geological disaster warning method and system.

[0005] First aspect, a geological disaster warning method is provided, which is applied to an accident tracing system. The method includes: obtaining a set of description information of exemplary geological disaster hazard points, where the set of description information of exemplary geological disaster hazard points includes pairs of description information of geological disaster hazard points composed of description information of geological disaster hazard points with the same warning object, and pairs of description information of geological disaster hazard points composed of description information of geological disaster hazard points with different warning objects; obtaining a first knowledge vector and a second knowledge vector of each piece of description information of geological disaster hazard points in the set of description information of exemplary geological disaster hazard points, and obtaining a first bucketing result by using the first knowledge vector of each piece of description information of geological disaster hazard points, where the first knowledge vector includes a stage-type knowledge vector and the second knowledge vector includes an attribute knowledge vector; performing knowledge vector update processing on each pair of description information of geological disaster hazard points in the set of description information of exemplary geological disaster hazard points to obtain a new pair of description information of geological disaster hazard points, where the knowledge vector update processing is to generate new first description information of a geological disaster hazard point by using the first knowledge vector of the first piece of description information of a geological disaster hazard point and the second knowledge vector of the second piece of description information of a geological disaster hazard point within the pair of description information of a geological disaster hazard point, and generating new second description information of a geological disaster hazard point by using the second knowledge vector of the first piece of description information of a geological disaster hazard point and the first knowledge vector of the second piece of description information of a geological disaster hazard point; using a specified geological disaster warning analysis network to obtain a first loss of the first bucketing result, a second loss of the new pair of description information of geological disaster hazard points, and a third loss of the first knowledge vector and the second knowledge vector of the new pair of description information of geological disaster hazard points; at least combining the first loss, the second loss, and the third loss to debug the geological disaster hazard analysis parameters of the geological disaster warning analysis network until a preset requirement is met, and parsing the hazard information corresponding to the preset requirement being met to obtain a geological disaster hazard point warning result.

[0006] In this application, obtaining the first knowledge vector and the second knowledge vector of each piece of description information of geological disaster hazard points in the set of description information of exemplary geological disaster hazard points includes: inputting the M pieces of description information of geological disaster hazard points in the pair of description information of geological disaster hazard points into the stage-type feature extraction layer and the attribute feature extraction layer of the geological disaster warning analysis network, where M is equal to 2; using the stage-type feature extraction layer to obtain the first knowledge vector of the M pieces of description information of geological disaster hazard points within the pair of description information of a geological disaster hazard point, and using the attribute feature extraction layer to obtain the second knowledge vector of the M pieces of description information of geological disaster hazard points within the pair of description information of a geological disaster hazard point.

[0007] In this application, obtaining the first loss of the first bucketing result, the second loss of the new geological disaster hidden danger point description information pair, and the third loss of the first knowledge vector and the second knowledge vector of the new geological disaster hidden danger point description information pair by using the specified geological disaster warning analysis network includes: obtaining the first bucketing result of the first knowledge vector obtained through the stage-type feature extraction layer; using the first specified geological disaster warning analysis network, combining the first bucketing result with the current bucketing result corresponding to the input geological disaster hidden danger point description information, to obtain the first loss.

[0008] In this application, before inputting the M geological disaster hidden danger point descriptions of the geological disaster hidden danger point description information pair into the stage-type feature extraction layer, the method further includes: preprocessing the geological disaster hidden danger point description information set of the warning object in the M geological disaster hidden danger point descriptions of the geological disaster hidden danger point description information pair.

[0009] In this application, performing knowledge vector update processing on each geological disaster hidden danger point description information pair in the example geological disaster hidden danger point description information set to obtain a new geological disaster hidden danger point description information pair includes: inputting the first knowledge vector and the second knowledge vector of each geological disaster hidden danger point description in the geological disaster hidden danger point description information pair of the example geological disaster hidden danger point description information set into the generation data analysis layer of the geological disaster warning analysis network; performing the knowledge vector update processing on each geological disaster hidden danger point description information pair in the example geological disaster hidden danger point description information set through the generation data analysis layer to obtain the new geological disaster hidden danger point description information pair.

[0010] In this application, on the premise that the input geological disaster hidden danger point description information pair is the geological disaster hidden danger point description information of the same warning object, performing knowledge vector update processing on each geological disaster hidden danger point description information pair in the example geological disaster hidden danger point description information set to obtain a new geological disaster hidden danger point description information pair includes: performing knowledge vector update processing on the geological disaster hidden danger point description information in the geological disaster hidden danger point description information pair once to obtain the new geological disaster hidden danger point description information pair, which includes: generating a new first geological disaster hidden danger point description by using the first knowledge vector of the first geological disaster hidden danger point description and the second knowledge vector of the second geological disaster hidden danger point description in the geological disaster hidden danger point description information pair, and generating a new second geological disaster hidden danger point description by using the second knowledge vector of the first geological disaster hidden danger point description and the first knowledge vector of the second geological disaster hidden danger point description.

[0011] In this application, on the premise that the input geological disaster hidden danger point description information pairs are the geological disaster hidden danger point description information for different early warning objects, performing knowledge vector update processing on each geological disaster hidden danger point description information pair in the example geological disaster hidden danger point description information set to obtain new geological disaster hidden danger point description information pairs includes: performing two knowledge vector update processes on the geological disaster hidden danger point description information within the geological disaster hidden danger point description information pair to obtain new geological disaster hidden danger point description information pairs, which includes: generating a new first intermediate geological disaster hidden danger point description information using the first knowledge vector of the first geological disaster hidden danger point description information and the second knowledge vector of the second geological disaster hidden danger point description information in the geological disaster hidden danger point description information pair, and generating a new second intermediate geological disaster hidden danger point description information using the second knowledge vector of the first geological disaster hidden danger point description information and the first knowledge vector of the second geological disaster hidden danger point description information; generating the first geological disaster hidden danger point description information of the first using the first knowledge vector of the first intermediate geological disaster hidden danger point description information and the second knowledge vector of the second intermediate geological disaster hidden danger point description information, and generating a new second geological disaster hidden danger point description information using the second knowledge vector of the first intermediate geological disaster hidden danger point description information and the first knowledge vector of the second intermediate geological disaster hidden danger point description information.

[0012] In this application, the use of the specified geological disaster early warning analysis network to obtain the first loss of the first bucketing result, the second loss of the new geological disaster hidden danger point description information pair, and the third loss of the first knowledge vector and the second knowledge vector of the new geological disaster hidden danger point description information pair includes: using the second specified geological disaster early warning analysis network to obtain the second loss of the new geological disaster hidden danger point description information pair relative to the initial geological disaster hidden danger point description information pair obtained through the generation data analysis layer of the geological disaster early warning analysis network.

[0013] In this application, the use of the specified geological disaster early warning analysis network to obtain the first loss of the first bucketing result, the second loss of the new geological disaster hidden danger point description information pair, and the third loss of the first knowledge vector and the second knowledge vector of the new geological disaster hidden danger point description information pair includes: according to the third specified geological disaster early warning analysis network, based on the first knowledge vector and the second knowledge vector of the new geological disaster hidden danger point description information pair and the first knowledge vector and the second knowledge vector of the corresponding initial geological disaster hidden danger point description information pair, obtaining the third loss of the first knowledge vector and the second knowledge vector of the new geological disaster hidden danger point description information pair.

[0014] In this application, after performing knowledge vector update processing on each geological disaster hidden danger point description information pair in the set of example geological disaster hidden danger point description information to obtain a new geological disaster hidden danger point description information pair, the method further includes: inputting the generated new geological disaster hidden danger point description information pair into the hidden danger information recognition layer of the geological disaster early warning analysis network to obtain a label knowledge vector representing the current degree of the new geological disaster hidden danger point description information pair; using a fourth specified geological disaster early warning analysis network to obtain a fourth loss of the new geological disaster hidden danger point description information pair based on the label knowledge vector.

[0015] In this application, debugging the geological disaster hidden danger analysis parameters of the geological disaster early warning analysis network by combining at least the first loss, the second loss, and the third loss until the preset requirements are met, includes: obtaining the loss of the geological disaster early warning analysis network by using the first loss, the second loss, the third loss, and the fourth loss; debugging the geological disaster hidden danger analysis parameters of the geological disaster early warning analysis network by using the loss of the geological disaster early warning analysis network until the preset requirements are met.

[0016] In this application, obtaining the loss of the geological disaster early warning analysis network by using the first loss, the second loss, the third loss, and the fourth loss includes: when the set of example geological disaster hidden danger point description information input to the geological disaster early warning analysis network is a geological disaster hidden danger point description information pair of the same early warning object, using a fifth specified geological disaster early warning analysis network to obtain a first network loss of the geological disaster early warning analysis network based on the first loss, the second loss, the third loss, and the fourth loss; when the set of example geological disaster hidden danger point description information input to the geological disaster early warning analysis network is a geological disaster hidden danger point description information pair of different early warning objects, using a sixth specified geological disaster early warning analysis network to obtain a second network loss of the geological disaster early warning analysis network based on the first loss, the second loss, the third loss, and the fourth loss; obtaining the loss of the geological disaster early warning analysis network based on the fusion result of the first network loss and the second network loss.

[0017] In a second aspect, a geological disaster early warning system is provided, including a processor and a memory that communicate with each other, where the processor is configured to read and execute a computer program from the memory to implement the above method.

[0018] A geological disaster warning method and system provided by an embodiment of the present application obtain a set of description information of exemplary geological disaster hidden danger points; obtain a first knowledge vector and a second knowledge vector of each piece of description information of geological disaster hidden danger points in the set of description information of exemplary geological disaster hidden danger points, and use the first knowledge vector of each piece of description information of geological disaster hidden danger points to obtain a first bucketing result, where the first knowledge vector includes a stage-type knowledge vector, and the second knowledge vector includes an attribute knowledge vector; perform knowledge vector update processing on each pair of description information of geological disaster hidden danger points in the set of description information of exemplary geological disaster hidden danger points to obtain a new pair of description information of geological disaster hidden danger points. The knowledge vector update processing is to generate new first description information of geological disaster hidden danger points by using the first knowledge vector of the first description information of geological disaster hidden danger points and the second knowledge vector of the second description information of geological disaster hidden danger points within the description information of geological disaster hidden danger points, and generate new second description information of geological disaster hidden danger points by using the second knowledge vector of the first description information of geological disaster hidden danger points and the first knowledge vector of the second description information of geological disaster hidden danger points; use a specified geological disaster warning analysis network to obtain a first loss of the first bucketing result, a second loss of the new pair of description information of geological disaster hidden danger points, and a third loss of the first knowledge vector and the second knowledge vector of the new pair of description information of geological disaster hidden danger points; at least combine the first loss, the second loss, and the third loss to debug the geological disaster hidden danger analysis parameters of the geological disaster warning analysis network until the preset requirements are met, and analyze the hidden danger information corresponding to the preset requirements to obtain a geological disaster hidden danger point warning result. Each piece of description information of geological disaster hidden danger points in the set of description information of exemplary geological disaster hidden danger points in the embodiment of the present disclosure has a corresponding stage-type label, and this stage-type label can correspond to the warning object in the description information of geological disaster hidden danger points and is used to distinguish the stage-type information of the warning object in the description information of geological disaster hidden danger points. In the embodiment of the present disclosure, each piece of description information of geological disaster hidden danger points in the set of description information of exemplary geological disaster hidden danger points may have a current division label corresponding to its corresponding warning object, and according to this current division label, the reliability of the bucketing result of the geological disaster warning analysis network can be compared, such as the corresponding loss can be determined. By accurately determining the loss, the accuracy and confidence of the geological disaster hidden danger point warning result can be guaranteed. Description of the Drawings

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present application and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0020] Figure 1 It is a flowchart of a geological disaster warning method provided by an embodiment of the present application.

[0021] Figure 2 It is a block diagram of a geological disaster early warning device provided by an embodiment of the present application. Specific implementation manners

[0022] To better understand the above technical solution, the technical solution of the present application will be described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical solution of the present application, rather than limitations on the technical solution of the present application. Without conflict, the technical features in the embodiments of the present application and the embodiments can be combined with each other.

[0023] Please refer to Figure 1 , which shows a geological disaster early warning method, and the technical solution described in steps 100-step 500 may be included in this method.

[0024] Step 100: Obtain a set of description information of exemplary geological disaster hidden danger points. The set of description information of exemplary geological disaster hidden danger points includes pairs of description information of geological disaster hidden danger points composed of description information of geological disaster hidden danger points with the same early warning object, and pairs of description information of geological disaster hidden danger points composed of description information of geological disaster hidden danger points with different early warning objects.

[0025] For example, the description information of geological disaster hidden danger points is the hidden danger information of possible geological disasters in each area, such as: information on loose mountain bodies and crustal movements. A pair of description information of geological disaster hidden danger points can be understood as a data pair composed of two pieces of description information of geological disaster hidden danger points.

[0026] Step 200: Obtain a first knowledge vector and a second knowledge vector of each piece of description information of geological disaster hidden danger points in the set of description information of exemplary geological disaster hidden danger points, and obtain a first bucketing result by using the first knowledge vector of each piece of description information of geological disaster hidden danger points. The first knowledge vector includes a stage-type knowledge vector, and the second knowledge vector includes an attribute knowledge vector.

[0027] For example, a knowledge vector can be understood as a feature, which is important information for triggering geological disasters obtained from the description information of geological disaster hidden danger points. Among them, the first knowledge vector can be understood as important information for geological disasters triggered by natural factors, and the second knowledge vector can be understood as important information for geological disasters triggered by human factors (such as: the vibration of large equipment triggers the occurrence of geological disasters).

[0028] Among them, the stage-type knowledge vector can be understood as a local knowledge vector, and this part of the content may be the key factor for the occurrence of geological disasters.

[0029] Step 300: Perform knowledge vector update processing on the description information of each geological disaster hidden danger point in the set of description information of the exemplary geological disaster hidden danger points to obtain a new pair of description information of geological disaster hidden danger points. The knowledge vector update processing is to generate new first description information of geological disaster hidden danger points by using the first knowledge vector of the first description information of geological disaster hidden danger points and the second knowledge vector of the second description information of geological disaster hidden danger points within the description information of geological disaster hidden danger points, and generate new second description information of geological disaster hidden danger points by using the second knowledge vector of the first description information of geological disaster hidden danger points and the first knowledge vector of the second description information of geological disaster hidden danger points.

[0030] Step 400: Use the specified geological disaster warning analysis network to obtain the first loss of the first bucketing result, the second loss of the new pair of description information of geological disaster hidden danger points, and the third loss of the first knowledge vector and the second knowledge vector of the new pair of description information of geological disaster hidden danger points.

[0031] For example, the loss can be understood as a loss value. Specifically, it can be understood as the data loss situation of the data processed by the specified geological disaster warning analysis network.

[0032] Step 500: At least in combination with the first loss, the second loss, and the third loss, debug the geological disaster hidden danger analysis parameters of the geological disaster warning analysis network until the preset requirements are met, and analyze the hidden danger information corresponding to the preset requirements being met to obtain the warning result of the geological disaster hidden danger point.

[0033] For example, the preset requirements can be understood as error tolerance conditions.

[0034] In the embodiments of the present disclosure, when training a geological disaster warning analysis network through the embodiments of the present disclosure, the description information set of exemplary geological disaster potential points can be input to the geological disaster warning analysis network first, and this description information set of exemplary geological disaster potential points serves as the description information of exemplary geological disaster potential points for training the geological disaster warning analysis network. Among them, in the embodiments of the present disclosure, the description information set of exemplary geological disaster potential points can include two types of description information of exemplary geological disaster potential points. The first type of example is a pair of description information of geological disaster potential points composed of different description information of geological disaster potential points of the same warning object. The second type of example is a pair of description information of geological disaster potential points composed of different description information of geological disaster potential points of different warning objects. That is, in the first type of example, the description information of geological disaster potential points within each pair of description information of geological disaster potential points is different description information of geological disaster potential points of the same warning object. In the second type of example, the description information of geological disaster potential points within each pair of description information of geological disaster potential points is different description information of geological disaster potential points of different warning objects. Each pair of description information of geological disaster potential points can include M pieces of description information of geological disaster potential points, such as the following first description information of geological disaster potential points and second description information of geological disaster potential points. In addition, the embodiments of the present disclosure can use these two types of description information of exemplary geological disaster potential points to train the geological disaster warning analysis network respectively.

[0035] Furthermore, each piece of description information of geological disaster potential points in the description information set of exemplary geological disaster potential points in the embodiments of the present disclosure has a corresponding stage type tag, and this stage type tag can correspond to the warning object in the description information of geological disaster potential points and is used to distinguish the stage type information of the warning object in the description information of geological disaster potential points. In the embodiments of the present disclosure, each piece of description information of geological disaster potential points in the description information set of exemplary geological disaster potential points can have a current division tag corresponding to its corresponding warning object. According to this current division tag, the reliability of the bucketing result of the geological disaster warning analysis network can be compared, such as corresponding losses can be determined; through the accurate determination of losses, the accuracy and confidence of the warning result of geological disaster potential points can be guaranteed.

[0036] After obtaining the description information set of exemplary geological disaster potential points, the specific update steps of the geological disaster warning analysis network can be executed. In step 200, first, the first knowledge vector and the second knowledge vector of the first description information of geological disaster potential points and the second description information of geological disaster potential points of each pair of description information of geological disaster potential points can be identified. The second knowledge vector can be a knowledge vector other than the first knowledge vector, such as an attribute knowledge vector. The following is an example of obtaining the first knowledge vector and the second knowledge vector for the geological disaster warning analysis network.

[0037] In step 200 of the network update method according to an embodiment of the present disclosure, obtaining a first knowledge vector and a second knowledge vector of each geological disaster hidden danger point description information in the set of exemplary geological disaster hidden danger point description information, and obtaining a first bucketing result by using the first knowledge vector of each geological disaster hidden danger point description information may include the following content.

[0038] Step 201: Input the M geological disaster hidden danger point description information of the geological disaster hidden danger point description information pair into the stage-type feature extraction layer and the attribute feature extraction layer of the geological disaster early warning analysis network.

[0039] Step 202: Use the stage-type feature extraction layer to obtain a first knowledge vector of the M geological disaster hidden danger point description information in the geological disaster hidden danger point description information pair, and use the attribute feature extraction layer to obtain a second knowledge vector of the M geological disaster hidden danger point description information in the geological disaster hidden danger point description information pair.

[0040] Step 203: Use the classification layer of the geological disaster early warning analysis network to obtain a first bucketing result corresponding to the first knowledge vector.

[0041] Among them, bucketing means clustering in this application.

[0042] Among them, the geological disaster early warning analysis network according to the embodiment of the present disclosure may include a stage-type feature extraction layer and an attribute feature extraction layer. The stage-type feature extraction layer may be used to identify the stage-type knowledge vector of the warning object in the geological disaster hidden danger point description information, and the attribute feature extraction layer may be used to identify the attribute knowledge vector of the warning object in the geological disaster hidden danger point description information. Therefore, each geological disaster hidden danger point description information pair in the obtained set of exemplary geological disaster hidden danger point description information can be respectively input into the above-mentioned stage-type feature extraction layer and attribute feature extraction layer. Through the stage-type feature extraction layer, a first knowledge vector of the M geological disaster hidden danger point description information in the received geological disaster hidden danger point description information pair can be obtained, and a second knowledge vector of the M geological disaster hidden danger point description information in the received geological disaster hidden danger point description information pair can be obtained by using the attribute feature extraction layer. For example, if the M geological disaster hidden danger point description information in the input geological disaster hidden danger point description information pair is respectively represented by X and Y, then the first knowledge vector of X obtained through the stage-type feature extraction layer is X1, the first knowledge vector of Y obtained through the stage-type feature extraction layer is Y1, the second knowledge vector of X obtained through the attribute feature extraction layer is X2, and the second knowledge vector of Y obtained through the attribute feature extraction layer is Y2.

[0043] After extracting the first knowledge vector and the second knowledge vector of the description information of the M geological disaster hidden danger points in the geological disaster hidden danger point description information pair, the embodiments of the present disclosure can perform the operation of classification and recognition using the first knowledge vector, and at the same time, can also perform subsequent knowledge vector update processing.

[0044] In addition, after obtaining the first knowledge vector and the second knowledge vector of each geological disaster hidden danger point description information, knowledge vector update processing can be performed between every M geological disaster hidden danger point descriptions of each pair of geological disaster hidden danger point description information. Among them, the knowledge vector update processing can be to update the second knowledge vector of the first geological disaster hidden danger point description information and the second knowledge vector of the second geological disaster hidden danger point description information in the geological disaster hidden danger point description information pair, and obtain new geological disaster hidden danger point description information based on the first knowledge vector and the updated second knowledge vector.

[0045] Through the knowledge vector update processing, the first knowledge vector of one geological disaster hidden danger point description information can be fused with the second knowledge vector of another geological disaster hidden danger point description information to form new geological disaster hidden danger point description information. Using this new geological disaster hidden danger point description information for classification can effectively realize the recognition of stage-type information based on stage-type knowledge vectors, thereby reducing interference.

[0046] In step 300 of the method for processing geological disaster hidden danger point description information in the embodiments of the present disclosure, where performing knowledge vector update processing on each pair of geological disaster hidden danger point description information in the exemplary geological disaster hidden danger point description information set to obtain a new pair of geological disaster hidden danger point description information may include the following steps.

[0047] Step 301: Input the pair of geological disaster hidden danger point description information in the exemplary geological disaster hidden danger point description information set into the generation data analysis layer of the geological disaster early warning analysis network.

[0048] Step 302: Through the generation data analysis layer, perform the knowledge vector update processing on each pair of geological disaster hidden danger point description information in the exemplary geological disaster hidden danger point description information set to obtain the new pair of geological disaster hidden danger point description information.

[0049] The geological disaster early warning analysis network in the embodiments of the present disclosure may further include a generation data analysis layer, which can perform knowledge vector update processing on the first knowledge vector and the second knowledge vector obtained by the stage-type feature extraction layer and the attribute feature extraction layer, and obtain new geological disaster hidden danger point description information according to the updated knowledge vectors.

[0050] For example, as described in the above embodiments, the exemplary geological disaster hidden danger point description information sets input in the embodiments of the present disclosure may include two types of exemplary geological disaster hidden danger point description information sets. Among them, the geological disaster hidden danger point description information pairs in the first exemplary are the geological disaster hidden danger point description information for the same early warning object. For the geological disaster hidden danger point description information pairs in the first exemplary, the embodiments of the present disclosure may perform a knowledge vector update process on the geological disaster hidden danger point description information within each geological disaster hidden danger point description information pair.

[0051] Among them, for the first exemplary, the process of performing a knowledge vector update process on the geological disaster hidden danger point description information within the exemplary geological disaster hidden danger point description information set to obtain new geological disaster hidden danger point description information pairs may include: performing a knowledge vector update process on the geological disaster hidden danger point description information within the geological disaster hidden danger point description information pair to obtain the new geological disaster hidden danger point description information pair. This process may include: generating new first geological disaster hidden danger point description information by using the first knowledge vector of the first geological disaster hidden danger point description information and the second knowledge vector of the second geological disaster hidden danger point description information in each geological disaster hidden danger point description information pair, and generating new second geological disaster hidden danger point description information by using the second knowledge vector of the first geological disaster hidden danger point description information and the first knowledge vector of the second geological disaster hidden danger point description information.

[0052] Among them, since the M geological disaster hidden danger point description information in the geological disaster hidden danger point description information pairs within the first exemplary are all different geological disaster hidden danger point description information for the same early warning object, the new geological disaster hidden danger point description information obtained after performing the knowledge vector update process is still the early warning object for the same early warning object. After completing the knowledge vector update process, the difference between the obtained new geological disaster hidden danger point description information and the corresponding initial geological disaster hidden danger point description information, and the difference between the first knowledge vector and the second knowledge vector of the new geological disaster hidden danger point description information and the first knowledge vector and the second knowledge vector of the corresponding initial geological disaster hidden danger point description information may be used to determine the loss of the geological disaster early warning analysis network, and the recognition and classification may be directly performed according to the generated new geological disaster hidden danger point description information. At this time, the generated new geological disaster hidden danger point description information pair may be input to the classification layer to perform classification to obtain the second bucketing result.

[0053] For example, the description information pair of geological disaster hidden danger points in the first example includes geological disaster hidden danger point description information X and geological disaster hidden danger point description information Y. Through the stage-type feature extraction layer, the first knowledge vector of X can be obtained as X1, the first knowledge vector of Y can be obtained as Y1 through the stage-type feature extraction layer, the second knowledge vector of X can be obtained as X2 through the attribute feature extraction layer, and the second knowledge vector of Y can be obtained as Y2 through the attribute feature extraction layer. M and Y are respectively the first geological disaster hidden danger point description information and the second geological disaster hidden danger point description information of the same warning object, and the first geological disaster hidden danger point description information and the second geological disaster hidden danger point description information are different. When performing the knowledge vector update process, the new first geological disaster hidden danger point description information Xa can be obtained by using the first knowledge vector X1 of X and the second knowledge vector Y2 of Y, and the new second geological disaster hidden danger point description information Ya can be obtained by using the first knowledge vector Y1 of Y and the second knowledge vector X2 of X.

[0054] Furthermore, the geological disaster warning analysis network of the embodiments of the present disclosure may include a generated data analysis layer, and the generated data analysis layer may be used to generate new geological disaster hidden danger point description information according to the received first knowledge vector and second knowledge vector. For example, the generated data analysis layer may include at least one convolutional layer, or may also include other processing layers, and the geological disaster hidden danger point description information corresponding to the first knowledge vector and the second knowledge vector can be obtained through the generated data analysis layer. That is, the above process of updating the second knowledge vector and generating the geological disaster hidden danger point description information based on the updated knowledge vector can be completed through the generation network.

[0055] Through the above knowledge vector update process, new data can be formed by updating the second knowledge vectors of the two pieces of geological disaster hidden danger point description information, so that the knowledge vectors related to the stage-type information and the knowledge vectors not related to the stage-type information can be successfully extracted. By training the geological disaster warning analysis network with this kind of geological disaster warning analysis network, the recognition accuracy of the geological disaster warning analysis network for the stage-type knowledge vectors can be improved.

[0056] In addition, the set of example geological disaster hidden danger point description information of the embodiments of the present disclosure may further include a second example group, and the description information pair of geological disaster hidden danger points therein is the description information of geological disaster hidden danger points of different warning objects. For the description information pair of geological disaster hidden danger points in the second example, the embodiments of the present disclosure may perform the knowledge vector update process twice on the geological disaster hidden danger point description information in each description information pair of geological disaster hidden danger points.

[0057] For the second set of examples, according to step 303 of the network update method according to an embodiment of the present disclosure, where on the premise that the input geological disaster hidden danger point description information pairs are geological disaster hidden danger point description information for different early warning objects, the performing knowledge vector update processing on each geological disaster hidden danger point description information in the example geological disaster hidden danger point description information set to obtain new geological disaster hidden danger point description information pairs may include: performing knowledge vector update processing twice on the geological disaster hidden danger point description information in the geological disaster hidden danger point description information pair to obtain new geological disaster hidden danger point description information pairs, and this process may include the following content.

[0058] Step 3031: Generate a new first intermediate geological disaster hidden danger point description information by using the first knowledge vector of the first geological disaster hidden danger point description information and the second knowledge vector of the second geological disaster hidden danger point description information in each geological disaster hidden danger point description information pair in the second example, and generate a new second intermediate geological disaster hidden danger point description information by using the second knowledge vector of the first geological disaster hidden danger point description information and the first knowledge vector of the second geological disaster hidden danger point description information.

[0059] Step 3032: Generate the first geological disaster hidden danger point description information of the first by using the first knowledge vector of the first intermediate geological disaster hidden danger point description information and the second knowledge vector of the second intermediate geological disaster hidden danger point description information, and generate new second geological disaster hidden danger point description information by using the second knowledge vector of the first intermediate geological disaster hidden danger point description information and the first knowledge vector of the second intermediate geological disaster hidden danger point description information.

[0060] For example, the first knowledge vector of X obtained through the stage-based feature extraction layer is X1, the first knowledge vector of Y obtained through the stage-based feature extraction layer is Y1, the second knowledge vector of X obtained through the attribute feature extraction layer is X2, and the second knowledge vector of Y obtained through the attribute feature extraction layer is Y2. X and Y are the first geological disaster hidden danger point description information and the second geological disaster hidden danger point description information of different early warning objects respectively. When performing the first knowledge vector update process, the new first intermediate geological disaster hidden danger point description information Xa can be obtained by using the first knowledge vector X1 of X and the second knowledge vector Y2 of Y, and the new second intermediate geological disaster hidden danger point description information Ya can be obtained by using the first knowledge vector Y1 of Y and the second knowledge vector X2 of X. Correspondingly, when performing the second knowledge vector update process, the first knowledge vector Xa1 and the second knowledge vector Xa2 of the first intermediate geological disaster hidden danger point description information Xa, and the first knowledge vector Ya1 and the second knowledge vector Ya2 of the second intermediate geological disaster hidden danger point description information Ya can be obtained again by using the stage-based feature extraction layer and the attribute feature extraction layer respectively, and the update process of the second knowledge vector Xa2 of the first intermediate geological disaster hidden danger point description information Xa and the second knowledge vector Ya2 of the second intermediate geological disaster hidden danger point description information Ya can be further performed by using the generation network, and the new first geological disaster hidden danger point description information X” can be generated by using the first knowledge vector Xa1 of the first intermediate geological disaster hidden danger point description information Xa and the second knowledge vector Ya2 of the second intermediate geological disaster hidden danger point description information Ya, and the new second geological disaster hidden danger point description information Y” can be generated by using the second knowledge vector Xa2 of the first intermediate geological disaster hidden danger point description information Xa and the first knowledge vector Ya1 of the second intermediate geological disaster hidden danger point description information Ya.

[0061] After the knowledge vector update process is completed, the differences between the obtained new geological disaster hidden danger point description information and the corresponding initial geological disaster hidden danger point description information, as well as the differences between the first knowledge vector and the second knowledge vector of the new geological disaster hidden danger point description information and the first knowledge vector and the second knowledge vector of the corresponding initial geological disaster hidden danger point description information can be utilized. At the same time, the first knowledge vector of the new geological disaster hidden danger point description information can also be input into the classification layer to perform classification processing to obtain a second bucketing result. Among them, for the first example, the second bucketing result of the first knowledge vector of the final new geological disaster hidden danger point description information can be directly obtained. For the second example, in addition to obtaining the second bucketing result of the first knowledge vector of the final new geological disaster hidden danger point description information, the second bucketing result of the first knowledge vector of the intermediate geological disaster hidden danger point description information can also be obtained. The embodiments of the present disclosure can update the geological disaster early warning analysis network according to the above second bucketing result, as well as the differences between the new geological disaster hidden danger point description information and the initial geological disaster hidden danger point description information and the differences between the knowledge vectors.

[0062] That is, the embodiments of the present disclosure can perform feedback debugging on the geological disaster early warning analysis network according to the loss of the output results obtained by each layer of the geological disaster early warning analysis network until the loss of the geological disaster early warning analysis network meets the preset requirements. For example, when it is lower than the loss threshold, it can be determined that the preset requirements are met. The loss network of the geological disaster early warning analysis network in the embodiments of the present disclosure can be related to the loss networks of each data analysis layer. For example, it can be the weighted sum of the loss networks of each data analysis layer. Based on this, the loss of the geological disaster early warning analysis network can be obtained by using the loss of each data analysis layer, so as to correct the geological disaster hidden danger analysis parameters of each layer of the geological disaster early warning analysis network until the preset requirement that the loss is lower than the loss threshold is met. Those skilled in the art can set the loss threshold according to needs, and the present disclosure does not specifically limit this.

[0063] Next, the feedback debugging process of the embodiments of the present disclosure will be described in detail.

[0064] Among them, after the first knowledge vector of each geological disaster hidden danger point description information is obtained through the stage-type feature extraction layer, the classification layer can obtain a first bucketing result according to the first knowledge vector. The embodiments of the present disclosure can utilize the first specified geological disaster early warning analysis network to obtain the first loss of the first bucketing result obtained by using the first knowledge vector obtained through the stage-type feature extraction layer. Among them, according to step 400 of the geological disaster hidden danger point description information processing method in the embodiments of the present disclosure, the process of obtaining the first loss includes the following steps.

[0065] Step 401: Obtain the first bucketing result of the first knowledge vector obtained through the stage-type feature extraction layer.

[0066] Step 402: Use the first specified geological disaster warning analysis network to obtain the first loss by combining the first bucketing result and the current bucketing result corresponding to the input geological disaster hidden danger point description information.

[0067] As described in the above embodiment, in step 200, when obtaining the first knowledge vector of the geological disaster hidden danger point description information in the example, the classification and recognition of the first knowledge vector can be performed through the classification layer to obtain the first bucketing result corresponding to the first knowledge vector. The first bucketing result can be represented in the form of a queue, where each attribute represents the possibility corresponding to each stage-type information label. By comparing the first bucketing result with the current bucketing result, a first comparison value can be obtained. The embodiment of the present disclosure can determine the first comparison value as the first loss. Or in other embodiments, the first bucketing result and the current bucketing result can also be input into the first loss network of the classification layer to obtain the first loss. The present disclosure does not make specific limitations on this.

[0068] In the embodiment of the present disclosure, when training the geological disaster warning analysis network through the first example and the second example, the loss networks used can be the same or different. Moreover, the embodiment of the present disclosure can perform fusion processing on the loss of the geological disaster warning analysis network trained through the first example and the loss of the geological disaster warning analysis network trained through the second example to obtain the final loss of the geological disaster warning analysis network, and use this loss to perform feedback debugging processing on the network. During the feedback debugging process, the geological disaster hidden danger analysis parameters of each data analysis layer of the geological disaster warning analysis network can be debugged, or only the geological disaster hidden danger analysis parameters of some of the layers can be debugged. The present disclosure does not make specific limitations on this.

[0069] First, the embodiment of the present disclosure can use the first specified geological disaster warning analysis network to obtain the first loss of the first bucketing result obtained from the first knowledge vector obtained through the stage-type feature extraction layer.

[0070] Through the above geological disaster warning analysis network, the first loss of the first bucketing result obtained by the classification layer can be obtained. Among them, the embodiment of the present disclosure can perform feedback debugging on the geological disaster hidden danger analysis parameters of the stage-type feature extraction layer, the attribute feature extraction layer, and the classification layer according to the first loss, or can also determine the overall loss of the geological disaster warning analysis network according to the first loss and the losses of other data analysis layers, and perform unified feedback debugging on each layer of the geological disaster warning analysis network layer. The present disclosure does not make limitations on this.

[0071] Secondly, the embodiments of the present disclosure can also process the new geological disaster hidden danger point description information pairs generated by the generated data parsing layer to obtain the second loss of the new geological disaster hidden danger point description information pairs and the third loss of the corresponding knowledge vectors. Among them, the geological disaster early warning parsing network for determining the second loss may include: using the second specified geological disaster early warning parsing network to obtain the second loss of the new geological disaster hidden danger point description information pairs obtained through the generated data parsing layer relative to the initial geological disaster hidden danger point description information pairs.

[0072] In the embodiments of the present disclosure, new geological disaster hidden danger point description information pairs can be obtained through the generation network. The embodiments of the present disclosure can determine the second loss according to the difference between the new geological disaster hidden danger point description information pairs and the initial geological disaster hidden danger point description information pairs.

[0073] Through the above geological disaster early warning parsing network, for the first example, the second loss corresponding to the new geological disaster hidden danger point description information pairs generated by the generated data parsing layer can be obtained.

[0074] In addition, the embodiments of the present disclosure can also obtain the third loss corresponding to the knowledge vectors of the new second geological disaster hidden danger point description information pairs, where the third loss can be obtained by using the third specified geological disaster early warning parsing network.

[0075] Through the above geological disaster early warning parsing network, the third loss corresponding to the knowledge vectors of the new geological disaster hidden danger point description information pairs generated by the generated data parsing layer can be obtained by the classification layer.

[0076] Similarly, the embodiments of the present disclosure can respectively perform feedback debugging on the geological disaster hidden danger parsing parameters of the generated data parsing layer according to the second loss and the third loss, or can simultaneously perform feedback debugging on each layer of the geological disaster early warning parsing network in combination with the first loss. For example, in some possible implementation geological disaster early warning parsing networks of the present disclosure, the loss of the geological disaster early warning parsing network can be obtained by using the weighted sum of the above first loss, second loss, and third loss respectively. In other words, the loss network of the geological disaster early warning parsing network is the weighted sum of the above first loss network, second loss network, and third loss network. The weights of each loss network are not specifically limited in the present disclosure, and those skilled in the art can set them according to needs. When the obtained loss is greater than the loss threshold, the geological disaster hidden danger parsing parameters of each data parsing layer are feedback debugged until the loss is less than the loss threshold, that is, the training can be terminated. At this time, the geological disaster early warning parsing network is updated.

[0077] Moreover, in the embodiments of the present disclosure, when training each of the first loss network, the second loss network, and the third loss network based on the first exemplary geological disaster hidden danger point description information pair, and when training based on the second exemplary geological disaster hidden danger point description information pair, each of the first loss network, the second loss network, and the third loss network may be different, but this is not specifically limited in the present disclosure.

[0078] Among them, the embodiments of the present disclosure may input the generated new geological disaster hidden danger point description information pair into the hidden danger information recognition layer of the geological disaster early warning analysis network, and use the fourth specified geological disaster early warning analysis network to obtain the fourth loss of the new geological disaster hidden danger point description information pair.

[0079] In the embodiments of the present disclosure, the training process of the hidden danger information recognition layer can be executed separately, that is, the generated new geological disaster hidden danger point description information and the corresponding current geological disaster hidden danger point description information can be input into the hidden danger information recognition layer, and the hidden danger information recognition layer can be trained based on the above fourth loss network until the loss corresponding to the fourth loss network is lower than the loss threshold required for training.

[0080] Regarding some possible embodiments, the hidden danger information recognition layer can also be trained simultaneously with the foregoing stage-type feature extraction layer, attribute feature extraction layer, and generated data analysis layer. Correspondingly, step 400 of the embodiments of the present disclosure can also obtain the loss of the geological disaster early warning analysis network by using the above first loss, second loss, third loss, and fourth loss. In other words, the loss network of the geological disaster early warning analysis network is the weighted sum of the above first loss network, second loss network, third loss network, and fourth loss network. The weight values of each loss network are not specifically limited in the present disclosure, and those skilled in the art can set them according to requirements. When the obtained loss is greater than the loss threshold, the geological disaster hidden danger analysis parameters of each data analysis layer of the geological disaster early warning analysis network are fed back and debugged until the loss is less than the loss threshold, that is, the training can be terminated, and at this time, the geological disaster early warning analysis network is updated.

[0081] Moreover, in the embodiments of the present disclosure, when training each of the first loss network, the second loss network, and the third loss network based on the first exemplary geological disaster hidden danger point description information pair, and when training based on the second false group of geological disaster hidden danger point description information pair, each of the first loss network, the second loss network, and the third loss network may be different, but this is not specifically limited in the present disclosure.

[0082] During the training process, when the obtained loss is greater than the loss threshold, the geological disaster hidden danger analysis parameters of the geological disaster warning analysis network are fed back for debugging. For example, the geological disaster hidden danger analysis parameters of each data analysis layer (phased feature extraction layer, attribute feature extraction layer, generated data analysis layer, hidden danger information recognition layer, etc.) can be fed back for debugging until the loss of the geological disaster warning analysis network is less than the loss threshold, that is, the training can be terminated. At this time, the geological disaster warning analysis network is updated. Or, in other embodiments, the geological disaster hidden danger analysis parameters of the phased feature extraction layer, attribute feature extraction layer, and classification layer can also be debugged according to the first loss, and the geological disaster hidden danger analysis parameters of the generated data analysis layer can be fed back for debugging according to the second loss and the third loss, and the geological disaster hidden danger analysis parameters of the hidden danger information recognition layer can be debugged according to the fourth loss until the loss is less than the loss threshold of the corresponding loss network, that is, the training is terminated. That is, the embodiments of the present disclosure can perform feedback debugging and training on each layer separately, or can perform unified debugging on each layer of the geological disaster warning analysis network through the loss of the geological disaster warning analysis network. Those skilled in the art can select an appropriate geological disaster warning analysis network to execute this debugging process according to the requirements.

[0083] In addition, in the embodiments of the present disclosure, in order to improve the recognition accuracy of the phased knowledge vector of the geological disaster warning analysis network, noise can also be added to each of the described example geological disaster hidden danger point description information sets before inputting them into the phased feature extraction layer. For example, preprocessing is performed on the geological disaster hidden danger point description information set of the warning object in the M geological disaster hidden danger point description information of the geological disaster hidden danger point description information pair. In the embodiments of the present disclosure, preprocessing is performed on the geological disaster warning analysis network with a covering layer added to a partial set of the geological disaster hidden danger point description information set of the warning object. The size of the covering layer can be set by those skilled in the art according to the requirements, and the present disclosure does not limit this. It should be noted here that the embodiments of the present disclosure only perform preprocessing on the geological disaster hidden danger point description information input into the phased feature extraction layer and do not introduce noise to other data analysis layers. Through this geological disaster warning analysis network, the recognition accuracy of the phased information of the geological disaster warning analysis network can be effectively improved.

[0084] To more clearly illustrate the embodiments of the present disclosure, the training processes of the first example and the second example are described below by way of example.

[0085] Update the second knowledge vector in the description information pair of geological disaster hidden danger points, and use the generator to obtain the updated M intermediate geological disaster hidden danger point description information. Further, use the stage-type feature extraction layer and the attribute feature extraction layer to obtain the first knowledge vector and the second knowledge vector of the M intermediate geological disaster hidden danger point description information, and then update the second knowledge vector of the intermediate geological disaster hidden danger point description information to obtain the new geological disaster hidden danger point description information. At this time, the second loss corresponding to the M new geological disaster hidden danger point description information, and the third loss corresponding to the first knowledge vector and the second knowledge vector of the M new geological disaster hidden danger point description information can be obtained. Input the intermediate geological disaster hidden danger point description information or the new geological disaster hidden danger point description information into the hidden danger information recognition layer D to obtain the fourth loss. At this time, the loss of the geological disaster warning analysis network can be obtained by using the first loss, the second loss, the third loss, and the fourth loss. When the loss is less than the loss threshold, the training is terminated, otherwise, the geological disaster hidden danger analysis parameters of each data analysis layer of the geological disaster warning analysis network are fed back for debugging.

[0086] On this basis, please refer to Figure 2 , a geological disaster warning device 200 is provided, and the device includes: A data acquisition module 210, configured to acquire a set of example geological disaster hidden danger point description information, where the set of example geological disaster hidden danger point description information includes a geological disaster hidden danger point description information pair composed of geological disaster hidden danger point description information of the same warning object, and a geological disaster hidden danger point description information pair composed of geological disaster hidden danger point description information of different warning objects; A result bucketing module 220, configured to obtain the first knowledge vector and the second knowledge vector of each geological disaster hidden danger point description information in the set of example geological disaster hidden danger point description information, and obtain the first bucketing result by using the first knowledge vector of each geological disaster hidden danger point description information, where the first knowledge vector includes a stage-type knowledge vector, and the second knowledge vector includes an attribute knowledge vector; A data update module 230, configured to perform knowledge vector update processing on each geological disaster hidden danger point description information pair in the set of example geological disaster hidden danger point description information to obtain a new geological disaster hidden danger point description information pair, where the knowledge vector update processing is to generate a new first geological disaster hidden danger point description information by using the first knowledge vector of the first geological disaster hidden danger point description information and the second knowledge vector of the second geological disaster hidden danger point description information in the geological disaster hidden danger point description information pair, and generate a new second geological disaster hidden danger point description information by using the second knowledge vector of the first geological disaster hidden danger point description information and the first knowledge vector of the second geological disaster hidden danger point description information; A loss acquisition module 240, configured to utilize a specified geological disaster warning analysis network to obtain a first loss of the first bucketing result, a second loss of the new geological disaster hidden danger point description information pair, and a third loss of the first knowledge vector and the second knowledge vector of the new geological disaster hidden danger point description information pair; A result warning module 250, configured to at least combine the first loss, the second loss, and the third loss to debug the geological disaster hidden danger analysis parameters of the geological disaster warning analysis network until a preset requirement is met, and analyze the hidden danger information corresponding to the preset requirement to obtain a geological disaster hidden danger point warning result.

[0087] On the basis of the above, a geological disaster warning system is shown, including a processor and a memory that communicate with each other, and the processor is configured to read and execute a computer program from the memory to implement the above method.

[0088] On the basis of the above, a computer-readable storage medium is further provided, and the computer program stored thereon implements the above method when running.

[0089] In summary, based on the above solution, obtain the set of description information of exemplary geological disaster hidden danger points; obtain the first knowledge vector and the second knowledge vector of each geological disaster hidden danger point description information in the set of description information of exemplary geological disaster hidden danger points, and use the first knowledge vector of each geological disaster hidden danger point description information to obtain the first bucketing result, where the first knowledge vector includes a stage-based knowledge vector and the second knowledge vector includes an attribute knowledge vector; perform knowledge vector update processing on each pair of geological disaster hidden danger point description information in the set of description information of exemplary geological disaster hidden danger points to obtain a new pair of geological disaster hidden danger point description information. The knowledge vector update processing is to generate a new first geological disaster hidden danger point description information using the first knowledge vector of the first geological disaster hidden danger point description information and the second knowledge vector of the second geological disaster hidden danger point description information within the geological disaster hidden danger point description information pair, and generate a new second geological disaster hidden danger point description information using the second knowledge vector of the first geological disaster hidden danger point description information and the first knowledge vector of the second geological disaster hidden danger point description information; use a specified geological disaster early warning analysis network to obtain the first loss of the first bucketing result, the second loss of the new pair of geological disaster hidden danger point description information, and the third loss of the first knowledge vector and the second knowledge vector of the new pair of geological disaster hidden danger point description information; debug the geological disaster hidden danger analysis parameters of the geological disaster early warning analysis network by combining at least the first loss, the second loss, and the third loss until the preset requirements are met, and analyze the hidden danger information corresponding to the preset requirements to obtain the early warning result of the geological disaster hidden danger point. Each geological disaster hidden danger point description information in the set of description information of exemplary geological disaster hidden danger points in the embodiments of the present disclosure has a corresponding stage-based label, and this stage-based label can correspond to the early warning object in the geological disaster hidden danger point description information, and is used to distinguish the stage-based information of the early warning object in the geological disaster hidden danger point description information. In the embodiments of the present disclosure, each geological disaster hidden danger point description information in the set of description information of exemplary geological disaster hidden danger points may have a current division label corresponding to its corresponding early warning object. According to this current division label, the reliability of the bucketing result of the geological disaster early warning analysis network can be compared, such as determining the corresponding loss. By accurately determining the loss, the accuracy and confidence of the early warning result of the geological disaster hidden danger point can be ensured.

[0090] It should be understood that the systems and their modules shown above can be implemented in various ways. For example, in some embodiments, the systems and their modules can be implemented through hardware, software, or a combination of software and hardware. Among them, the hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated designed hardware. Those skilled in the art can understand that the above methods and systems can be implemented using computer-executable instructions and / or included in processor control code. For example, such code is provided on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The systems and their modules of the present application can be implemented not only by hardware circuits such as very large scale integrated circuits or gate arrays, semiconductors such as logic chips and transistors, or programmable hardware devices such as field programmable gate arrays and programmable logic devices, but also by software executed by various types of processors, or by a combination of the above hardware circuits and software (for example, firmware).

[0091] It should be noted that the beneficial effects that may be produced by different embodiments are different. In different embodiments, the beneficial effects that may be produced can be any one or several combinations of the above, or any other beneficial effects that may be obtained.

[0092] The basic concepts have been described above. Obviously, for those skilled in the art, the above detailed disclosure is only an example and does not constitute a limitation to the present application. Although not explicitly stated here, those skilled in the art may make various modifications, improvements, and corrections to the present application. Such modifications, improvements, and corrections are proposed in the present application, so such modifications, improvements, and corrections still fall within the spirit and scope of the exemplary embodiments of the present application.

[0093] At the same time, the present application uses specific terms to describe the embodiments of the present application. Such as "one embodiment", "an embodiment", and / or "some embodiments" mean a certain feature, structure, or characteristic related to at least one embodiment of the present application. Therefore, it should be emphasized and noted that the "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more at different positions in this specification does not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of the present application can be appropriately combined.

[0094] In addition, those skilled in the art can understand that various aspects of the present application can be illustrated and described by several patentable types or situations, including any new and useful processes, machines, products, or combinations of substances, or any new and useful improvements thereto. Accordingly, each aspect of the present application can be executed entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. The above-mentioned hardware or software can all be referred to as "data blocks", "modules", "engines", "layers", "components", or "systems". In addition, various aspects of the present application may be embodied as a computer product located in one or more computer-readable media, which includes computer-readable program code.

[0095] A computer storage medium may contain a propagated data signal containing computer program code, such as on a baseband or as part of a carrier wave. This propagated signal may have various forms of manifestation, including electromagnetic form, optical form, etc., or a suitable combination of forms. A computer storage medium can be any computer-readable medium other than a computer-readable storage medium, which can be connected to an instruction execution system, apparatus, or device to enable communication, propagation, or transmission for use of a program. The program code located on the computer storage medium can be propagated through any suitable medium, including radio, cable, fiber optic cable, RF, or similar media, or any combination of the above media.

[0096] The computer program code required for the operation of each part of the present application can be written in any one or more programming languages, including object-oriented programming languages such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.YET, Python, etc., conventional procedural programming languages such as C language, Visual Basic, Fortran 2003, Perl, COBOL 2002, PHP, ABAP, dynamic programming languages such as Python, Ruby, and Groovy, or other programming languages. The program code can run entirely on the user's computer, or as an independent software package on the user's computer, or partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer through any network form, such as a local area network (LAY) or a wide area network (WAY), or connected to an external computer (e.g., through the Internet), or in a cloud computing environment, or used as a service such as software as a service (SaaS).

[0097] In addition, unless explicitly stated in the claims, the order of the processing elements and sequences, the use of numerical and alphabetical characters, or the use of other names in this application are not used to limit the order of the processes and methods of this application. Although various examples are discussed in the above disclosure for some currently useful embodiments of the invention, it should be understood that such details are for illustrative purposes only. The appended claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that meet the essence and scope of the embodiments of this application. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only through software solutions, such as installing the described system on existing servers or mobile devices.

[0098] Similarly, it should be noted that, in order to simplify the presentation of the disclosure of this application and thus help the understanding of one or more embodiments of the invention, in the previous description of the embodiments of this application, sometimes multiple features are merged into one embodiment, drawing, or description thereof. However, this method of disclosure does not mean that the features required by the subject matter of this application are more than those mentioned in the claims. In fact, the features of the embodiments are fewer than all the features of the individual embodiments disclosed above.

[0099] In some embodiments, numbers are used to describe components and attribute quantities. It should be understood that such numbers used in the description of the embodiments are modified by the modifiers "about", "approximate", or "substantially" in some examples. Unless otherwise stated, "about", "approximate", or "substantially" indicate that the said numbers allow for adaptive changes. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, and such approximate values may change according to the characteristics required by individual embodiments. In some embodiments, the numerical parameters should consider the specified significant digits and adopt the method of retaining the general number of digits. Although the numerical ranges and parameters used in some embodiments of this application to confirm the breadth of their scope are approximate values, in specific embodiments, such numerical settings are as precise as possible within the feasible range.

[0100] For each patent, patent application, patent application publication, and other materials cited in this application, such as articles, books, specifications, publications, documents, etc., their entire contents are hereby incorporated into this application by reference. Except for the application history documents that are inconsistent with or conflict with the content of this application, and also except for the documents (currently or subsequently appended to this application) that limit the broadest scope of the claims of this application. It should be noted that if there are inconsistencies or conflicts between the descriptions, definitions, and / or uses of terms in the attached materials of this application and the content described in this application, the descriptions, definitions, and / or uses of terms in this application shall prevail.

[0101] Finally, it should be understood that the embodiments described in this application are only used to illustrate the principles of the embodiments of this application. Other variations may also fall within the scope of this application. Therefore, by way of example and not limitation, alternative training of the embodiments of this application can be considered to be consistent with the teachings of this application. Accordingly, the embodiments of this application are not limited to the embodiments explicitly introduced and described in this application.

[0102] The above are only the embodiments of this application and are not intended to limit this application. For those skilled in the art, various changes and modifications can be made to this application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of this application shall be included within the scope of the claims of this application.

Claims

1. A geological disaster warning method, characterized in that, The method includes: Obtaining a set of description information of exemplary geological disaster potential points, where the set of description information of exemplary geological disaster potential points includes pairs of description information of geological disaster potential points composed of description information of geological disaster potential points with the same early warning object, and pairs of description information of geological disaster potential points composed of description information of geological disaster potential points with different early warning objects; Obtaining a first knowledge vector and a second knowledge vector of each piece of description information of geological disaster potential points in the set of description information of exemplary geological disaster potential points, and obtaining a first bucketing result by using the first knowledge vectors of each piece of description information of geological disaster potential points. The first knowledge vector includes a stage-type knowledge vector, and the second knowledge vector includes an attribute knowledge vector; Performing knowledge vector update processing on each pair of description information of geological disaster potential points in the set of description information of exemplary geological disaster potential points to obtain a new pair of description information of geological disaster potential points. The knowledge vector update processing is to generate new first description information of geological disaster potential points by using the first knowledge vector of the first description information of geological disaster potential points and the second knowledge vector of the second description information of geological disaster potential points within the pair of description information of geological disaster potential points, and to generate new second description information of geological disaster potential points by using the second knowledge vector of the first description information of geological disaster potential points and the first knowledge vector of the second description information of geological disaster potential points; Using a specified geological disaster early warning analysis network to obtain a first loss of the first bucketing result, a second loss of the new pair of description information of geological disaster potential points, and a third loss of the first knowledge vector and the second knowledge vector of the new pair of description information of geological disaster potential points; Adjusting the geological disaster potential point analysis parameters of the geological disaster early warning analysis network by at least combining the first loss, the second loss, and the third loss until the preset requirements are met, and analyzing the potential point information corresponding to the preset requirements met to obtain a geological disaster potential point early warning result.

2. The method according to claim 1, characterized in that, The obtaining of the first knowledge vector and the second knowledge vector of each piece of description information of geological disaster potential points in the set of description information of exemplary geological disaster potential points includes: Inputting the M pieces of description information of geological disaster potential points of the pair of description information of geological disaster potential points into the stage-type feature extraction layer and the attribute feature extraction layer of the geological disaster early warning analysis network, where M is equal to 2; Obtaining the first knowledge vector of the M pieces of description information of geological disaster potential points within the pair of description information of geological disaster potential points by using the stage-type feature extraction layer, and obtaining the second knowledge vector of the M pieces of description information of geological disaster potential points within the pair of description information of geological disaster potential points by using the attribute feature extraction layer.

3. The method according to claim 2, wherein The using of a specified geological disaster early warning analysis network to obtain a first loss of the first bucketing result, a second loss of the new pair of description information of geological disaster potential points, and a third loss of the first knowledge vector and the second knowledge vector of the new pair of description information of geological disaster potential points includes: Obtaining the first bucketing result of the first knowledge vector obtained through the stage-type feature extraction layer; Using the first specified geological disaster warning analysis network, combining the first bucketing result and the current bucketing result corresponding to the input geological disaster hidden danger point description information, the first loss is obtained.

4. The method according to claim 2 or 3, characterized in that, Before inputting the M geological disaster hidden danger point description information pairs of the geological disaster hidden danger point description information into the stage-type feature extraction layer, the method further includes: preprocessing the geological disaster hidden danger point description information set of the warning object in the M geological disaster hidden danger point description information pairs of the geological disaster hidden danger point description information.

5. The method according to claim 1, wherein Performing knowledge vector update processing on each geological disaster hidden danger point description information pair in the exemplary geological disaster hidden danger point description information set to obtain a new geological disaster hidden danger point description information pair includes: Inputting the first knowledge vector and the second knowledge vector of each geological disaster hidden danger point description information in the geological disaster hidden danger point description information pair of the exemplary geological disaster hidden danger point description information set into the generated data analysis layer of the geological disaster warning analysis network; Through the generated data analysis layer, performing the knowledge vector update processing on each geological disaster hidden danger point description information pair in the exemplary geological disaster hidden danger point description information set to obtain the new geological disaster hidden danger point description information pair.

6. The method according to claim 1, wherein On the premise that the input geological disaster hidden danger point description information pair is the geological disaster hidden danger point description information of the same warning object, performing knowledge vector update processing on each geological disaster hidden danger point description information pair in the exemplary geological disaster hidden danger point description information set to obtain a new geological disaster hidden danger point description information pair includes: performing knowledge vector update processing once on the geological disaster hidden danger point description information in the geological disaster hidden danger point description information pair to obtain the new geological disaster hidden danger point description information pair, which includes: generating a new first geological disaster hidden danger point description information by using the first knowledge vector of the first geological disaster hidden danger point description information and the second knowledge vector of the second geological disaster hidden danger point description information in the geological disaster hidden danger point description information pair, and generating a new second geological disaster hidden danger point description information by using the second knowledge vector of the first geological disaster hidden danger point description information and the first knowledge vector of the second geological disaster hidden danger point description information.

7. The method according to claim 1, characterized in that On the premise that the input geological disaster hidden danger point description information pairs are the geological disaster hidden danger point description information for different early warning objects, perform knowledge vector update processing on each geological disaster hidden danger point description information pair in the example geological disaster hidden danger point description information set to obtain new geological disaster hidden danger point description information pairs, including: performing two knowledge vector update processes on the geological disaster hidden danger point description information within the geological disaster hidden danger point description information pair to obtain new geological disaster hidden danger point description information pairs, which includes: generating a new first intermediate geological disaster hidden danger point description information using the first knowledge vector of the first geological disaster hidden danger point description information and the second knowledge vector of the second geological disaster hidden danger point description information in the geological disaster hidden danger point description information pair, and generating a new second intermediate geological disaster hidden danger point description information using the second knowledge vector of the first geological disaster hidden danger point description information and the first knowledge vector of the second geological disaster hidden danger point description information; generating the first geological disaster hidden danger point description information of the first using the first knowledge vector of the first intermediate geological disaster hidden danger point description information and the second knowledge vector of the second intermediate geological disaster hidden danger point description information, and generating a new second geological disaster hidden danger point description information using the second knowledge vector of the first intermediate geological disaster hidden danger point description information and the first knowledge vector of the second intermediate geological disaster hidden danger point description information; Among them, the obtaining of the first loss of the first binning result, the second loss of the new geological disaster hidden danger point description information pair, and the third loss of the first knowledge vector and the second knowledge vector of the new geological disaster hidden danger point description information pair by using the specified geological disaster early warning analysis network includes: obtaining, by using a second specified geological disaster early warning analysis network, the second loss of the new geological disaster hidden danger point description information pair relative to the initial geological disaster hidden danger point description information pair obtained through the generated data analysis layer of the geological disaster early warning analysis network.

8. The method according to claim 1, wherein The obtaining of the first loss of the first binning result, the second loss of the new geological disaster hidden danger point description information pair, and the third loss of the first knowledge vector and the second knowledge vector of the new geological disaster hidden danger point description information pair by using the specified geological disaster early warning analysis network includes: according to a third specified geological disaster early warning analysis network, obtaining the third loss of the first knowledge vector and the second knowledge vector of the new geological disaster hidden danger point description information pair based on the first knowledge vector and the second knowledge vector of the new geological disaster hidden danger point description information pair and the first knowledge vector and the second knowledge vector of the corresponding initial geological disaster hidden danger point description information pair.

9. The method according to claim 1, characterized in that, After performing knowledge vector update processing on each geological disaster hidden danger point description information pair in the example geological disaster hidden danger point description information set to obtain new geological disaster hidden danger point description information pairs, the method further includes: Inputting the generated new geological disaster hidden danger point description information pair into the hidden danger information recognition layer of the geological disaster early warning analysis network to obtain a label knowledge vector representing the current degree of the new geological disaster hidden danger point description information pair; Using the fourth specified geological disaster warning analysis network, obtain the fourth loss of the new geological disaster hidden danger point description information pair according to the label knowledge vector; Among them, the geological disaster hidden danger analysis parameters of the geological disaster warning analysis network are adjusted in combination with the first loss, the second loss and the third loss until the preset requirements are met, including: obtaining the loss of the geological disaster warning analysis network by using the first loss, the second loss, the third loss and the fourth loss; adjusting the geological disaster hidden danger analysis parameters of the geological disaster warning analysis network by using the loss of the geological disaster warning analysis network until the preset requirements are met; Among them, obtaining the loss of the geological disaster warning analysis network by using the first loss, the second loss, the third loss and the fourth loss includes: When the set of sample geological disaster hidden danger point description information input to the geological disaster warning analysis network is a pair of geological disaster hidden danger point description information of the same warning object, use the fifth specified geological disaster warning analysis network to obtain the first network loss of the geological disaster warning analysis network according to the first loss, the second loss, the third loss and the fourth loss; When the set of sample geological disaster hidden danger point description information input to the geological disaster warning analysis network is a pair of geological disaster hidden danger point description information of different warning objects, use the sixth specified geological disaster warning analysis network to obtain the second network loss of the geological disaster warning analysis network according to the first loss, the second loss, the third loss and the fourth loss; Obtain the loss of the geological disaster warning analysis network according to the fusion result of the first network loss and the second network loss.

10. A geological disaster early warning system, characterized in that, It includes a processor and a memory that communicate with each other, and the processor is used to read and execute a computer program from the memory to implement the method according to any one of claims 1-9.

Images