Deduction method and system for debris flow occurrence

Abstract

The invention discloses a debris flow occurrence deduction method and system, and the method comprises the steps: determining the data scale of to-be-configured configuration data as the input data of the deduction degree contained in a deduction data set in response to an execution task for the deduction data set; splicing all decision units in the target debris flow deduction thread to obtain a splicing result for the target debris flow deduction thread; based on the input data, the thread calculation performance corresponding to the last decision unit of the target debris flow deduction thread and the splicing result for the target debris flow deduction thread, determining a debris flow occurrence result output by the target debris flow deduction thread; based on the debris flow occurrence result output by each decision-making unit of the debris flow deduction thread, an allocated long and short-term memory module is obtained; and executing a data operation relationship by using the distributed long and short-term memory module to obtain a debris flow deduction result for configuring characters for the target debris flow deduction thread. Therefore, the deduction precision is improved.

Description

Technical Field

[0001] This disclosure relates to the technical field of debris flow occurrence prediction, and in particular to a debris flow occurrence prediction method and system. Background Technology

[0002] Debris flows, as a frequent and sudden geological disaster in mountainous areas, are characterized by rapid onset, strong destructive force, and wide-ranging impact. Their occurrence and evolution are influenced by a complex interplay of multiple factors, including topography, geological lithology, precipitation conditions, and sediment supply, posing a serious threat to the lives and property of mountain residents, infrastructure development, and the ecological environment. Accurately predicting the probability of debris flows, their onset time, movement paths, and impact range is a core prerequisite for disaster risk early warning, emergency response planning, and the implementation of scientific prevention and control measures. It is of paramount importance for enhancing the initiative and effectiveness of debris flow disaster prevention and control.

[0003] With the rapid development and deep integration of technologies such as remote sensing, UAV aerial surveying, ground sensor networks, and meteorological monitoring, the ability to collect debris flow-related data has been significantly improved, resulting in a large-scale dataset covering multiple dimensions, including topographic elevation data, geological structure data, precipitation time-series data, vegetation cover data, and sediment reserves data. This provides rich data support for debris flow occurrence prediction, driving the transformation of prediction technology from traditional empirical models and statistical analysis to data-driven intelligent methods.

[0004] Therefore, how to overcome the above-mentioned technical challenges and construct a debris flow occurrence prediction method that can dynamically adapt to complex environments, efficiently integrate multi-source data, accurately depict temporal correlations, and balance computational efficiency and prediction accuracy, so as to achieve accurate simulation and prediction of the debris flow occurrence process, has become the core problem that urgently needs to be solved in the field of debris flow disaster prevention and control. It is also the key technical support for promoting the transformation of disaster prediction technology from "static simulation" to "dynamic intelligent prediction". Summary of the Invention

[0005] To address the technical problems existing in related technologies, this disclosure provides a method and system for predicting debris flow occurrence.

[0006] A method for predicting debris flow occurrence, the method includes: Obtain the set of inference data obtained by character mining based on the debris flow inference thread configuration; In response to the execution task for the simulation dataset, the data size of the configuration data to be configured is determined as the input data for the simulation level covered in the simulation dataset; The decision units in the debris flow simulation thread are spliced ​​together to obtain the splicing result for the debris flow simulation thread. Based on the input data, the thread computing performance corresponding to the last decision unit of the debris flow simulation thread, and the splicing results of the debris flow simulation thread, the debris flow occurrence result output by the debris flow simulation thread is determined. Based on the debris flow occurrence results output by each decision unit of the debris flow inference thread, the allocated long short-term memory modules are obtained. By utilizing the allocated long short-term memory modules to perform data operations, the debris flow prediction results are obtained for the characters configured in the debris flow prediction thread.

[0007] In this application, based on the input data, the thread computing performance corresponding to each decision unit in the debris flow simulation thread, and the splicing results corresponding to each decision unit in the debris flow simulation thread, the debris flow occurrence results output by each decision unit in the debris flow simulation thread are determined, including: Based on the input data, the thread computing performance corresponding to the first decision unit of the debris flow simulation thread, and the splicing results, the debris flow occurrence result output by the first decision unit of the debris flow simulation thread is determined. For each decision unit other than the first decision unit, the debris flow occurrence result output by that decision unit is determined according to the following steps; Based on the debris flow occurrence result output by the previous decision unit of the current debris flow simulation thread, the thread computing performance corresponding to the previous decision unit of the current debris flow simulation thread, and the splicing result, the debris flow occurrence result output by the current decision unit of the current debris flow simulation thread is determined. This process continues until the debris flow occurrence result is obtained from the output of the last decision unit of the debris flow prediction thread.

[0008] In this application, data operations are performed using allocated long short-term memory modules to obtain debris flow projection results, including: The data values ​​of the configuration data to be configured, as well as the thread coefficient values ​​generated by each decision unit of the debris flow simulation thread during the excavation process, are determined as the input data. Using the allocated long short-term memory modules, the input data is processed according to the data calculation rules to obtain the debris flow prediction results.

[0009] In this application, the operation of determining the debris flow occurrence results output by each decision unit of the debris flow prediction thread runs in the first calculation module, and the operation of executing data calculation relationships using the allocated long short-term memory module runs in the second calculation module; Among them, the first operation module and the second operation module are heterogeneous operation modules.

[0010] In this application, the inference process and data operation relationships are performed in one of the following ways: After the first calculation module completes the calculation of the degree of each decision unit of the debris flow simulation thread, the second calculation module then executes the data calculation relationship of each decision unit of the debris flow simulation thread. After the first calculation module completes the calculation of the current decision unit of the debris flow calculation thread, and the second calculation module executes the data calculation relationship of the current decision unit of the debris flow calculation thread, the first calculation module executes the calculation of the next decision unit of the debris flow calculation thread in parallel.

[0011] In this application, after determining the debris flow occurrence results output by each decision unit of the debris flow simulation thread, the method further includes: Verify whether the debris flow occurrence results output by each decision unit in the debris flow simulation thread conform to the preset debris flow occurrence results; Under the premise of verifying that the debris flow occurrence results are consistent with the preset debris flow occurrence results, the steps of obtaining the allocated long short-term memory modules are executed based on the debris flow occurrence results output by each decision unit of the debris flow inference thread.

[0012] The technical solutions provided by the embodiments of this disclosure may include the following beneficial effects.

[0013] A method and system for debris flow occurrence prediction involves obtaining a prediction data set based on the character mining of a specific debris flow prediction thread; responding to an execution task on the prediction data set, determining the data size of the configuration data to be configured as the input data for the prediction level covered in the prediction data set; concatenating the decision units of the specific debris flow prediction thread to obtain the concatenation result for the specific debris flow prediction thread; determining the debris flow occurrence result output by the specific debris flow prediction thread based on the input data, the thread computing performance corresponding to the last decision unit of the specific debris flow prediction thread, and the concatenation result for the specific debris flow prediction thread; obtaining allocated long short-term memory modules based on the debris flow occurrence results output by each decision unit of the specific debris flow prediction thread; and using the allocated long short-term memory modules to execute data operation relationships to obtain the debris flow prediction result for the configuration characters of the specific debris flow prediction thread. This improves the accuracy of the prediction.

[0014] It should be understood that the above general description and the following detailed description are merely exemplary and do not limit this disclosure. Attached Figure Description

[0015] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the specification, serve to explain the principles of this application.

[0016] Figure 1 A flowchart illustrating a debris flow occurrence estimation method provided in this application embodiment; Detailed Implementation

[0017] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0018] Based on the above, please refer to the following: Figure 1 This is a flowchart illustrating a debris flow occurrence prediction method provided in this application embodiment. Further, a debris flow occurrence prediction method may specifically include the content described in the following steps S101-S102.

[0019] S101: Obtain the set of inference data obtained by character mining based on the debris flow inference thread configuration; S102: In response to the execution task for the simulation data set, based on the configuration data to be configured and the simulation degree and data operation relationship covered in the simulation data set, execute to obtain the debris flow simulation result for the configuration character of the target debris flow simulation thread.

[0020] One aspect is the degree of projection, which is the splicing result corresponding to each decision unit in the debris flow projection thread, represented in data form. This splicing result can include the order of shape calculations and the specific calculation rules used. In this way, even if the execution phase using configuration data of a different size than the excavation phase uses the projection dataset, it can still be accommodated. Furthermore, once the shape of the input configuration data is determined, the debris flow occurrence results output by each subsequent decision unit can also be determined based on the aforementioned splicing result.

[0021] Furthermore, in this embodiment, each execution of the simulation data set corresponds to a debris flow simulation result. This debris flow simulation result can be determined based on the corresponding splicing result and data calculation rules, provided that the configuration data is input into the simulation level and data operation relationship. That is, the execution of the thread characters can be understood as a thread configuration process. In each configuration process, the output result can be determined, and then the relevant thread coefficients determined in the excavation stage can be adjusted. By adjusting the thread coefficients, the thread coefficient value of the configured debris flow simulation thread can be obtained.

[0022] The data processing method provided in this embodiment can be executed according to the following steps: [Debris flow prediction thread configuration character] Step 1: Based on the configuration data to be configured and the degree of simulation covered in the simulation data set, determine the debris flow occurrence results output by each decision unit of the debris flow simulation thread; Step 2: Based on the debris flow occurrence results output by each decision unit of the debris flow simulation thread, obtain the allocated long short-term memory modules; Step 3: Utilize the allocated long short-term memory modules to perform data calculations and obtain the debris flow projection results.

[0023] Based on the above description, the following embodiments of this disclosure will explain the execution process of the deduction degree and data operation relationship from two aspects.

[0024] Firstly, the embodiments disclosed herein can be performed according to the following steps: Step 1: Determine the data size of the configuration data to be configured as the input data covering the degree of inference in the inference data set; Step 2: Based on the input data, the thread computing performance of each decision unit in the debris flow simulation thread, and the splicing results of each decision unit in the debris flow simulation thread, determine the debris flow occurrence results output by each decision unit in the debris flow simulation thread.

[0025] Here, the data scale of the configuration data can be determined as the input data for the simulation level. In this way, based on the input data, the thread computing performance of each decision unit in the debris flow simulation thread, and the splicing results of each decision unit in the debris flow simulation thread, the debris flow occurrence results output by each decision unit in the debris flow simulation thread can be determined.

[0026] In this embodiment of the present disclosure, the debris flow occurrence results output by each decision unit of the debris flow prediction thread can be determined unit by unit. That is, the debris flow occurrence result output by the first decision unit can be determined first, and then the debris flow occurrence results output by the second decision unit, the third decision unit and so on until the last decision unit are determined based on this output result.

[0027] In addition, regarding other decision-making units, the debris flow occurrence result output by the previous decision-making unit of the current debris flow simulation thread and the thread computing performance corresponding to the previous decision-making unit of the current debris flow simulation thread can be substituted into the splicing result to determine the debris flow occurrence result output by the current decision-making unit of the current debris flow simulation thread.

[0028] The specific splicing results are similar to the description of the first decision unit mentioned above, and will not be repeated here.

[0029] Given the debris flow occurrence result output by the last decision unit, the loss function value of the target debris flow simulation thread to be configured can be determined based on this debris flow occurrence result. Based on this loss function value, the target debris flow simulation thread to be configured can be adjusted. By re-executing the simulation data set, the adjusted target debris flow simulation thread can be reconfigured until the thread reaches convergence, thus obtaining the configured target debris flow simulation thread.

[0030] Secondly, the data operation relationship can be performed according to the following steps in the embodiments of this disclosure: Step 1: Determine the data values ​​of the configuration data to be configured, as well as the thread coefficient values ​​generated by each decision unit of the debris flow simulation thread during the excavation process, as the input data; Step 2: Using the allocated long short-term memory modules, perform calculations on the input data based on the data calculation rules to obtain the debris flow prediction results.

[0031] In this embodiment of the disclosure, the second operation module may execute the data operation relationship corresponding to each decision unit of the debris flow simulation thread after the first operation module has completed the simulation of the simulation degree corresponding to each decision unit of the debris flow simulation thread. Alternatively, the first operation module may complete the simulation of the current decision unit of the debris flow simulation thread, and the second operation module may execute the data operation relationship corresponding to the current decision unit of the debris flow simulation thread, while the first operation module executes the simulation degree corresponding to the next decision unit of the debris flow simulation thread in parallel.

[0032] Here, to ensure the accuracy of the debris flow prediction results, it can be verified whether the debris flow occurrence results output by each decision unit of the current debris flow prediction thread conform to the preset debris flow occurrence results. Only after verification that they conform to the preset debris flow occurrence results will the step of obtaining the allocated long short-term memory modules based on the debris flow occurrence results output by each decision unit of the current debris flow prediction thread be executed, thereby ensuring the accuracy of subsequent debris flow prediction results. It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for predicting the occurrence of debris flows, characterized in that, The methods include: Obtain the set of inference data obtained by character mining based on the debris flow inference thread configuration; In response to the execution task for the simulation dataset, the data size of the configuration data to be configured is determined as the input data for the simulation level covered in the simulation dataset; The decision units in the debris flow simulation thread are spliced ​​together to obtain the splicing result for the debris flow simulation thread. Based on the input data, the thread computing performance corresponding to the last decision unit of the debris flow simulation thread, and the splicing results of the debris flow simulation thread, the debris flow occurrence result output by the debris flow simulation thread is determined. Based on the debris flow occurrence results output by each decision unit of the debris flow inference thread, the allocated long short-term memory modules are obtained. By utilizing the allocated long short-term memory modules to perform data operations, the debris flow prediction results are obtained for the characters configured in the debris flow prediction thread.

2. The method according to claim 1, characterized in that, Based on the input data, the thread computation performance of each decision unit in the debris flow simulation thread, and the splicing results of each decision unit in the debris flow simulation thread, the debris flow occurrence results output by each decision unit in the debris flow simulation thread are determined, including: Based on the input data, the thread computing performance corresponding to the first decision unit of the debris flow simulation thread, and the splicing results, the debris flow occurrence result output by the first decision unit of the debris flow simulation thread is determined. For each decision unit other than the first decision unit, the debris flow occurrence result output by that decision unit is determined according to the following steps; Based on the debris flow occurrence result output by the previous decision unit of the current debris flow simulation thread, the thread computing performance corresponding to the previous decision unit of the current debris flow simulation thread, and the splicing result, the debris flow occurrence result output by the current decision unit of the current debris flow simulation thread is determined. This process continues until the debris flow occurrence result is obtained from the output of the last decision unit of the debris flow prediction thread.

3. The method according to claim 1 or 2, characterized in that, By utilizing the allocated long short-term memory modules to perform data operations, the debris flow projection results are obtained, including: The data values ​​of the configuration data to be configured, as well as the thread coefficient values ​​generated by each decision unit of the debris flow simulation thread during the excavation process, are determined as the input data. Using the allocated long short-term memory modules, the input data is processed according to the data calculation rules to obtain the debris flow prediction results.

4. The method according to any one of claims 1 to 3, characterized in that, The operation of determining the debris flow occurrence results output by each decision unit in the debris flow prediction thread runs in the first calculation module, and the operation of executing data calculation relationships using the allocated long short-term memory module runs in the second calculation module; Among them, the first operation module and the second operation module are heterogeneous operation modules.

5. The method according to claim 4, characterized in that, The inference process and data operation relationships shall be performed in one of the following ways: After the first calculation module completes the calculation of the degree of each decision unit of the debris flow simulation thread, the second calculation module then executes the data calculation relationship of each decision unit of the debris flow simulation thread. After the first calculation module completes the calculation of the current decision unit of the debris flow calculation thread, and the second calculation module executes the data calculation relationship of the current decision unit of the debris flow calculation thread, the first calculation module executes the calculation of the next decision unit of the debris flow calculation thread in parallel.

6. The method according to any one of claims 1 to 5, characterized in that, After determining the debris flow occurrence results output by each decision unit in the debris flow simulation thread, the method also includes: Verify whether the debris flow occurrence results output by each decision unit in the debris flow simulation thread conform to the preset debris flow occurrence results; Under the premise of verifying that the debris flow occurrence results are consistent with the preset debris flow occurrence results, the steps of obtaining the allocated long short-term memory modules are executed based on the debris flow occurrence results output by each decision unit of the debris flow inference thread.

7. A debris flow occurrence prediction system, characterized in that, It includes a processor and a memory that communicate with each other, the processor being used to read a computer program from the memory and execute it to implement the method of any one of claims 1-6.

Images