CN112561171B - Landslide prediction method, device, equipment and storage medium - Google Patents
Landslide prediction method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN112561171B CN112561171B CN202011501960.8A CN202011501960A CN112561171B CN 112561171 B CN112561171 B CN 112561171B CN 202011501960 A CN202011501960 A CN 202011501960A CN 112561171 B CN112561171 B CN 112561171B
- Authority
- CN
- China
- Prior art keywords
- landslide
- time sequence
- historical
- training
- factors
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 63
- 238000012549 training Methods 0.000 claims abstract description 151
- 238000012795 verification Methods 0.000 claims abstract description 67
- 238000012360 testing method Methods 0.000 claims description 49
- 238000011156 evaluation Methods 0.000 claims description 11
- 230000003203 everyday effect Effects 0.000 claims description 5
- 239000013049 sediment Substances 0.000 claims description 3
- 239000002689 soil Substances 0.000 claims description 3
- 230000003068 static effect Effects 0.000 claims description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 3
- 238000010200 validation analysis Methods 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 8
- 238000012545 processing Methods 0.000 description 7
- 230000002354 daily effect Effects 0.000 description 6
- 238000012544 monitoring process Methods 0.000 description 5
- 230000003287 optical effect Effects 0.000 description 5
- 238000004590 computer program Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012876 topography Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000013307 optical fiber Substances 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- YBJHBAHKTGYVGT-ZKWXMUAHSA-N (+)-Biotin Chemical compound N1C(=O)N[C@@H]2[C@H](CCCCC(=O)O)SC[C@@H]21 YBJHBAHKTGYVGT-ZKWXMUAHSA-N 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000008707 rearrangement Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000010998 test method Methods 0.000 description 1
- FEPMHVLSLDOMQC-UHFFFAOYSA-N virginiamycin-S1 Natural products CC1OC(=O)C(C=2C=CC=CC=2)NC(=O)C2CC(=O)CCN2C(=O)C(CC=2C=CC=CC=2)N(C)C(=O)C2CCCN2C(=O)C(CC)NC(=O)C1NC(=O)C1=NC=CC=C1O FEPMHVLSLDOMQC-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/047—Probabilistic or stochastic networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
The embodiment of the invention discloses a landslide prediction method, a landslide prediction device, landslide prediction equipment and a storage medium. And acquiring a landslide influence factor of each slope unit of the area to be predicted in a preset first historical time period, and determining the landslide occurrence probability of each slope unit of the current day according to the landslide influence factor of each slope unit of the area to be predicted in the first historical time period based on the landslide prediction model after training. As the history verification factors are added when the landslide prediction model is trained, the trained landslide prediction model can be prevented from being over-fitted. And the historical verification factors are compared with the prediction information corresponding to the historical training factors, so that the time sequence information of training samples is prevented from being disturbed, and the robustness of the landslide prediction model is improved, therefore, the landslide influence factors are input into the landslide prediction model after training, and the landslide occurrence probability of each slope unit on the same day can be accurately determined.
Description
Technical Field
The embodiment of the invention relates to the technical field of landslide monitoring, in particular to a landslide prediction method, a landslide prediction device, landslide prediction equipment and a storage medium.
Background
Landslide is one of the most common disastrous natural disasters, has the characteristics of wide distribution range, high occurrence frequency, multiple occurrence, regional property, severity and the like, and can cause a large amount of casualties and serious environmental and infrastructure losses every year. The method has important significance in evaluating the liability of landslide.
The existing landslide susceptibility prediction can be divided into a deterministic method and a non-deterministic method according to the difference of theoretical basis on which the landslide susceptibility prediction is based. The deterministic method is mainly a directional analysis based on expert experience and knowledge and a landslide process or physical model analysis method, and the prediction accuracy is poor. With the rapid development of computer technology and 3S technology in recent years, non-deterministic methods have been widely used, mainly including fuzzy logic methods, analytic hierarchy processes, decision trees, and the like. However, the landslide factor processed by the method has poor time precision and lacks consideration of time sequence of the rainfall factor, so that the landslide prediction precision is not high and needs to be improved.
Disclosure of Invention
The invention provides a landslide prediction method, a landslide prediction device, landslide prediction equipment and a storage medium, which realize the effect of improving landslide prediction precision.
In a first aspect, an embodiment of the present invention provides a method for predicting a landslide, where the method for predicting a landslide includes:
acquiring landslide influence factors of each slope unit of a region to be predicted in a preset first historical time period every day;
and determining landslide occurrence probability of each slope unit in the current day according to the landslide influence factors of each day in the first historical time period based on the landslide prediction model after training, wherein the landslide prediction model comprises at least two time sequence sub-networks, each time sequence sub-network is obtained by training according to the historical training factors and the historical verification factors of each slope unit in the sample area in the second historical time period, and the time sequence information of the historical verification factors of each time sequence sub-network is larger than the time sequence information of the historical training factors of the time sequence sub-network.
In a second aspect, an embodiment of the present invention further provides a device for predicting a landslide, where the device for predicting a landslide includes:
the landslide influence factor acquisition module is used for acquiring the landslide influence factors of each slope unit of the area to be predicted in a preset first historical time period;
the landslide occurrence probability prediction module is used for determining the landslide occurrence probability of each slope unit in the current day according to the landslide influence factors of each day in the first historical time period based on the landslide prediction model after training, wherein the landslide prediction model comprises at least two time sequence sub-networks, each time sequence sub-network is obtained through training according to the historical training factors and the historical verification factors of each slope unit in the sample area in the second historical time period, and the time sequence information of the historical verification factors of each time sequence sub-network is larger than the time sequence information of the historical training factors of the time sequence sub-network.
In a third aspect, an embodiment of the present invention further provides a landslide prediction apparatus, including: one or more processors;
a memory for storing one or more programs;
the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method of predicting a landslide of any one of the first aspects.
In a fourth aspect, embodiments of the present invention also provide a storage medium containing computer-executable instructions for performing the method of predicting a landslide of any one of the first aspects when executed by a computer processor.
According to the technical scheme provided by the embodiment, the daily landslide influence factors of all slope units of the area to be predicted in the preset first historical time period are obtained, the landslide occurrence probability of all slope units of the current day is determined according to the daily landslide influence factors in the first historical time period based on the landslide prediction model after training, wherein the landslide prediction model comprises at least two time sequence sub-networks, each time sequence sub-network is obtained through training according to the historical training factors and the historical verification factors of all slope units of the sample area in the second historical time period, and the time sequence information of the historical verification factors of each time sequence sub-network is larger than the time sequence information of the historical training factors of the time sequence sub-network. As the history verification factors are added when the landslide prediction model is trained, the trained landslide prediction model can be prevented from being over-fitted. And the historical verification factors are compared with the prediction information corresponding to the historical training factors, so that the time sequence information of training samples is prevented from being disturbed, and the robustness of the landslide prediction model is improved, therefore, the landslide influence factors are input into the landslide prediction model after training, and the landslide occurrence probability of each slope unit on the same day can be accurately determined.
Drawings
FIG. 1 is a flow chart of a method for predicting landslide in accordance with a first embodiment of the invention;
FIG. 2 is a flow chart of a method for predicting landslide in accordance with a second embodiment of the present invention;
FIG. 3 is a training schematic of a time-series subnetwork model1 of a landslide prediction model in a second embodiment of the invention;
FIG. 4 is a schematic diagram of a test process of a landslide prediction model in a second embodiment of the invention;
fig. 5 is a schematic structural diagram of a landslide prediction device in accordance with a third embodiment of the present invention;
fig. 6 is a schematic structural diagram of a landslide prediction apparatus in accordance with a fourth embodiment of the present invention.
Detailed Description
The invention is described in further detail below with reference to the drawings and examples. It is to be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention. It should be further noted that, for convenience of description, only some, but not all of the structures related to the present invention are shown in the drawings.
Example 1
Fig. 1 is a flowchart of a landslide prediction method according to an embodiment of the present invention, where the method may be performed by a landslide prediction device, and the method includes the following steps:
s110, acquiring landslide influence factors of each slope unit of the area to be predicted in a preset first historical time period.
Wherein the first historical time period may be a time period prior to the day of landslide occurrence. The first historical time period may be the first 3 days, the first 7 days, the first 15 days, or longer of the day. The area to be predicted is usually an area where landslide occurs, and may be any designated area. The slope units are basic units for the development of geologic hazards such as landslide, collapse and the like, and corresponding attribute values are assigned to the units to represent a data form of the entity. Specifically, a plurality of monitoring points may be set in the target area, and one or more monitoring stations may be set at each monitoring point, so as to read the landslide impact factors of each slope unit of the area to be predicted from the data of the monitoring stations.
Optionally, the landslide impact factor comprises a dynamic factor comprising rainfall and soil moisture and a static factor comprising at least one of elevation, grade, slope, planar curvature, profile curvature, terrain moisture index, water current intensity index, sediment transport index, terrain roughness index, distance to fault, distance to river, distance to road, lithology, land utilization, and vegetation coverage.
S120, determining landslide occurrence probability of each slope unit in the current day according to the landslide influence factors of each day in the first historical time period based on the landslide prediction model after training.
The landslide prediction model comprises at least two time sequence sub-networks, each time sequence sub-network is obtained by training according to the historical training factors and the historical verification factors of the slope units of the sample area in the second historical time period, and the time sequence information of the historical verification factors of each time sequence sub-network is larger than the time sequence information of the historical training factors of the time sequence sub-network.
The second history period may be the same as the first history period or longer than the first history period. The sample area is similar to the features of the terrain, the topography and the like of the area to be predicted, and the sample area and the feature of the area to be predicted can be positioned in the same area or in different areas. It can be understood that the historical training factors and the historical verification factors of each slope unit comprise rainfall, and the rainfall is time sequence data, so that time sequence information of training samples needs to be considered when the landslide prediction model is trained, and the model is prevented from being fitted.
Specifically, when the landslide prediction model is trained, the historical training factors and the historical verification factors are correspondingly divided into at least two time sequence windows, the time sequence information of the historical training factors in each time sequence window is sequentially increased, the time sequence information of the historical verification factors in each time sequence window is sequentially increased, and the time sequence information of the historical verification factors in each time sequence window is larger than the time sequence information of the historical training factors; the method comprises the steps of inputting historical training factors corresponding to each time sequence window into a landslide prediction model to be trained, outputting prediction information of each slope unit corresponding to each time sequence window, comparing the output prediction information of each slope unit corresponding to each time sequence window with historical verification factors, calculating evaluation parameters of the landslide prediction model to be trained, iteratively adjusting model parameters of the landslide prediction model to be trained based on the evaluation parameters, obtaining a landslide prediction model after training is completed, and determining a time sequence sub-network corresponding to each time sequence window in the landslide prediction model after training is completed. As the history verification factors are added when the landslide prediction model is trained, the trained landslide prediction model can be prevented from being over-fitted. And the historical verification factor is compared with the prediction information corresponding to the historical training factor, so that the time sequence information of the training sample is prevented from being disturbed, and the robustness of the landslide prediction model is improved.
Further, when landslide prediction is performed based on the trained landslide prediction model, determining time sequence information of the landslide influence factors of each day before the landslide occurs, inputting the landslide influence factors of each day into the landslide prediction model, determining corresponding time sequence sub-networks based on the time sequence information of the landslide influence factors, and obtaining the landslide occurrence probability of each slope of the same day based on the time sequence sub-networks corresponding to the time sequence information of the landslide influence factors.
According to the technical scheme provided by the embodiment, the daily landslide influence factors of all slope units of the area to be predicted in the preset first historical time period are obtained, the landslide occurrence probability of all slope units of the current day is determined according to the daily landslide influence factors in the first historical time period based on the landslide prediction model after training, wherein the landslide prediction model comprises at least two time sequence sub-networks, each time sequence sub-network is obtained through training according to the historical training factors and the historical verification factors of all slope units of the sample area in the second historical time period, and the time sequence information of the historical verification factors of each time sequence sub-network is larger than the time sequence information of the historical training factors of the time sequence sub-network. As the history verification factors are added when the landslide prediction model is trained, the trained landslide prediction model can be prevented from being over-fitted. And the historical verification factors are compared with the prediction information corresponding to the historical training factors, so that the time sequence information of training samples is prevented from being disturbed, and the robustness of the landslide prediction model is improved, therefore, the landslide influence factors are input into the landslide prediction model after training, and the landslide occurrence probability of each slope unit on the same day can be accurately determined.
Example two
Fig. 2 is a flowchart of a landslide prediction method according to a second embodiment of the present invention, in which a training step of a landslide prediction model is added on the basis of the previous embodiment. Optionally, the training method of the landslide prediction model includes: acquiring a history training factor and a history verification factor of each slope unit of the sample area in a second history time period; inputting the historical training factors into a current model according to the time sequence information to obtain prediction information of each slope unit; and iteratively adjusting model parameters of the current model according to the history verification factors and the prediction information until the current model under the current iteration times reaches stability, and taking the current model under the current iteration times as a landslide prediction model after training. For parts which are not described in detail in this method embodiment, reference is made to the above-described embodiments. Referring specifically to fig. 2, the method may include the steps of:
s210, acquiring a history training factor and a history verification factor of each slope unit of the sample area in the second history period.
Wherein the historical training factors and the historical verification factors include landslide points and non-landslide points.
S220, the historical training factors are input into the current model according to the time sequence information, and the historical landslide prediction information of each slope unit is obtained.
Optionally, the step of inputting the historical training factor into the current model according to the time sequence information to obtain the prediction information of each ramp unit includes: respectively determining a time sequence window corresponding to each time sequence sub-network in the current model; and based on the corresponding relation between the time sequence sub-network and the time sequence window, inputting the historical training factors corresponding to the time sequence window into the time sequence sub-network to obtain the historical landslide prediction information of each slope unit corresponding to the historical training factors.
Specifically, the historical training factors and the historical verification factors are correspondingly divided into at least two time sequence windows, the corresponding relation between each time sequence window and the time sequence sub-network is determined, when the landslide prediction model is trained, the historical training factors in each time sequence window are input to the corresponding time sequence sub-network of the current model based on the corresponding relation between the time sequence window and the time sequence sub-network, and the historical landslide prediction information of each slope unit corresponding to the historical training factors is output based on each time sequence sub-network. Fig. 3 is a training schematic diagram of the time-series subnetwork model1 of the landslide prediction model. The training sample of the time sequence sub-network model1 is S 1 ,S 2 ,…,S T S is selected from the training samples V+1 ,…,S T As a history verification factor, the training process is divided into q+1 sub-processes according to the time sequence information of the training sample, k history training factors are different between adjacent sub-processes, and the k value can be 1, 3, 5, 10, 15, 20 and other data. For each sub-process, inputting V historical training factors to obtain k historical landslide prediction information, wherein each historical landslide prediction information carries a unique label which can be 0 or 1, and after all the sub-processes are executed, the obtained historical landslide prediction information is S' V+1 、S` V+2 、…S` V+k 、…S` V+k+1 、S` V+k+2 、…S` V+2*k …S` V+q*k+1 、S` V+q*k+2 、…S T . The other time-series subnetworks of the landslide prediction model are different from the training samples of the time-series subnetwork model1, and the training process is consistent. S230, iteratively adjusting model parameters of the current model according to the historical verification factors and the historical landslide prediction information until the current model is reachedThe current model under the iteration times is stable, and the current model under the current iteration times is used as a landslide prediction model after training is completed. Optionally, the method for determining the landslide prediction model after training comprises the following steps: calculating an evaluation parameter of the current model based on the historical landslide prediction information and the historical verification factor, and iteratively adjusting the current model based on the evaluation parameter until the current model under the current iteration number reaches stability, and determining the model parameter under the current iteration number, wherein the evaluation parameter comprises at least one of root mean square error, recall ratio, precision ratio, false alarm ratio and landslide density; based on model parameters under the current iteration times, determining the sub-network corresponding to each time sequence window, and taking the current model comprising the time sequence sub-network corresponding to each time sequence window as a landslide prediction model after training is completed. Specifically, S230 determines the historical landslide prediction information S' of the time-series subnetwork model1 V+1 、S` V+2 、…S` V+k 、…S` V+k+1 、S` V+k+2 、…S` V+2*k …S` V+q*k+1 、S` V+q*k+2 、…S T Historical landslide prediction information S' of time sequence sub-network model1 V+1 、S` V+2 、…S` V+k 、…S` V+k+1 、S` V+k+2 、…S` V+2*k …S` V+q*k+1 、S` V+q*k+2 、…S T History verification factor S with timing subnetwork model1 V+1 ,…,S T And (3) comparing the historical landslide prediction information of other time sequence sub-networks with corresponding historical verification factors, iteratively adjusting the model parameters of the current model according to the comparison result until the current model under the current iteration times is stable, taking the model parameters under the current iteration times as optimal parameters, determining corresponding time sequence sub-networks according to the optimal parameters and the historical training factors of any time sequence sub-network, and taking the current model comprising the time sequence sub-network corresponding to each time sequence window as a trained landslide prediction model.
Illustratively, in connection with FIG. 3, when the current model at the current iteration number reaches steady, the optimal parameters are determined, and based on training samples (e.g., S T-V+1 ,S T-V+2 ,…,S T-V+k ,S T-V+k+1 ,…,S T-1 ，S T ) And optimal parameters, determining the time sequence sub-network model1. And similarly, determining other time sequence sub-networks according to training samples and optimal parameters of the other time sequence sub-networks, and further taking a current model of the time sequence sub-network corresponding to each time sequence window as a landslide prediction model after training.
Because the landslide prediction model comprises a plurality of time sequence sub-networks, training is carried out according to the training samples corresponding to each time sequence sub-network, and the training processes of the time sequence sub-networks are carried out simultaneously and are not interfered with each other, the time sequence information of the training samples can be prevented from being disturbed, and the robustness of the landslide prediction model is improved.
In order to improve the generalization capability of the landslide prediction model after training, the landslide prediction model after training is obtained can also be tested based on the historical test factors. Optionally, the method for testing the landslide prediction model includes: acquiring a history test factor corresponding to each time sequence sub-network of the landslide prediction model after training; and testing the landslide prediction model based on the historical test factors, the historical training factors and the historical verification factors corresponding to each time sequence sub-network.
Wherein the testing the landslide prediction model based on the history test factor, the history training factor, and the history verification factor corresponding to each of the time-series sub-networks includes: inputting the corresponding history training factors and the history verification factors into each time sequence sub-network to obtain the corresponding prediction probability of each time sequence sub-network; and determining a test result of the landslide prediction model according to the prediction probability corresponding to each time sequence sub-network and the history test factor corresponding to each time sequence sub-network.
Specifically, a time sequence sub-network corresponding to each historical test factor is determined based on time sequence information of the historical test factors, a time sequence sub-network corresponding to each historical test factor and a time sequence sub-network corresponding to each historical test factor are used for obtaining prediction probability corresponding to each time sequence sub-network, the prediction probability corresponding to each time sequence sub-network is compared with the historical test factor corresponding to the time sequence sub-network, test results of the trained landslide prediction model are determined according to comparison results of all the time sequence sub-networks, and the trained landslide prediction model is evaluated according to the test results.
Fig. 4 is a schematic diagram showing a test procedure of a landslide prediction model, where the landslide prediction model includes a time-series sub-network model1 and a time-series sub-network model 2..time-series sub-network model l+1, and a history test factor of the landslide prediction model is S T+1 ，S T+2 ，S T+k+1 ，S T+k+2 ，...，S m-1 ，S m Wherein the history test factor of the time sequence sub-network model1 is S T+1 ，S T+2 ，...，S T+k The history test factor of the time sequence sub-network model2 is S T+1+k ，S T+2+k ，...，S T+2*k The historical test factor of the time sequence sub-network model L+1 is S T+1+L*k ，S T+2+L*k ，...，S m . In testing the sequential subnetwork model1, training samples S of the sequential subnetwork model1 are tested 1 ，S 2 ，...，S T The history training factor and the history verification factor corresponding to the time sequence sub-network model1 are input into the time sequence sub-network model1 to obtain the prediction probability corresponding to the time sequence sub-network model1, namely S' T+1 、S` T+2 、…S` T+k A label of (a); similarly, training samples of other time sequence sub-networks are respectively input into corresponding time sequence sub-networks, namely, history training factors and history verification factors of other time sequence sub-networks are input into corresponding time sequence sub-networks, and prediction probabilities corresponding to the time sequence sub-networks are obtained; comparing the prediction probability corresponding to each time sequence sub-network with the history test factor corresponding to the time sequence sub-network, and determining the test result of the landslide prediction model after training according to the comparison result of all the time sequence sub-networks.
When the landslide prediction model is tested, the landslide prediction model is divided into different time sequence sub-networks, corresponding tests are conducted according to the historical test factors corresponding to the time sequence sub-networks, the test processes of the time sequence sub-networks are conducted simultaneously and are not interfered with each other, time sequence information of test samples can be prevented from being disturbed, and the test efficiency and reliability of the landslide prediction model are improved.
S240, acquiring landslide influence factors of each slope unit of the area to be predicted in a preset first historical time period every day.
S250, determining landslide occurrence probability of each slope unit in the current day according to the landslide influence factors of each day in the first historical time period based on the landslide prediction model after training.
The landslide prediction model comprises at least two time sequence sub-networks, each time sequence sub-network is obtained by training according to the historical training factors and the historical verification factors of the slope units of the sample area in the second historical time period, and the time sequence information of the historical verification factors of each time sequence sub-network is larger than the time sequence information of the historical training factors of the time sequence sub-network.
According to the technical scheme, the landslide prediction model comprises a plurality of time sequence sub-networks, when the landslide prediction model is trained, training is carried out according to training samples corresponding to each time sequence sub-network, the training processes of the time sequence sub-networks are carried out simultaneously and are not interfered with each other, time sequence information of the training samples can be prevented from being disturbed, and the robustness of the landslide prediction model is improved. When the landslide prediction model is tested, the landslide prediction model is divided into different time sequence sub-networks, corresponding tests are carried out according to the historical test factors corresponding to the time sequence sub-networks, the test processes of the time sequence sub-networks are carried out simultaneously and are not interfered with each other, and time sequence information of a test sample can be prevented from being disturbed, so that the test efficiency and the reliability of the landslide prediction model are improved.
Example III
Fig. 5 is a schematic diagram of the result of a landslide prediction device according to a third embodiment of the present invention, where, as shown in fig. 5, the landslide prediction device includes: a landslide impact factor acquisition module 310 and a landslide occurrence probability prediction module 320.
The landslide influence factor obtaining module 310 is configured to obtain a landslide influence factor of each slope unit of the area to be predicted in a preset first historical time period;
the landslide occurrence probability prediction module 320 is configured to determine the landslide occurrence probability of each ramp unit in the current day according to the landslide influence factor of each day in the first historical period based on the trained landslide prediction model, where the landslide prediction model includes at least two time sequence sub-networks, each time sequence sub-network is obtained by training according to the historical training factor and the historical verification factor of each ramp unit in the sample area in the second historical period, and the time sequence information of the historical verification factor of each time sequence sub-network is greater than the time sequence information of the historical training factor of the time sequence sub-network.
According to the technical scheme, the daily landslide influence factors of all slope units of the area to be predicted in the preset first historical time period are obtained, the landslide occurrence probability of all slope units of the current day is determined according to the daily landslide influence factors in the first historical time period based on a trained landslide prediction model, wherein the landslide prediction model comprises at least two time sequence sub-networks, each time sequence sub-network is obtained through training according to the historical training factors and the historical verification factors of all slope units of the sample area in the second historical time period, and the time sequence information of the historical verification factors of each time sequence sub-network is larger than the time sequence information of the historical training factors of the time sequence sub-network. As the history verification factors are added when the landslide prediction model is trained, the trained landslide prediction model can be prevented from being over-fitted. And the historical verification factors are compared with the prediction information corresponding to the historical training factors, so that the time sequence information of training samples is prevented from being disturbed, and the robustness of the landslide prediction model is improved, therefore, the landslide influence factors are input into the landslide prediction model after training, and the landslide occurrence probability of each slope unit on the same day can be accurately determined.
Optionally, the apparatus further comprises: a training module of the landslide prediction model; the training module of the landslide prediction model is used for acquiring historical training factors and historical verification factors of slope units of the sample area in a second historical time period;
inputting the historical training factors into a current model according to the time sequence information to obtain historical landslide prediction information of each slope unit;
and iteratively adjusting model parameters of the current model according to the historical verification factors and the historical landslide prediction information until the current model under the current iteration times reaches stability, and taking the current model under the current iteration times as a landslide prediction model after training.
Optionally, the training module of the landslide prediction model is further configured to determine a timing window corresponding to each of the timing sub-networks in the current model;
and based on the corresponding relation between the time sequence sub-network and the time sequence window, inputting the historical training factors corresponding to the time sequence window into the time sequence sub-network to obtain the historical landslide prediction information of each slope unit corresponding to the historical training factors.
Optionally, the training module of the landslide prediction model is further configured to calculate an evaluation parameter of the current model based on the historical landslide prediction information and the historical verification factor, and iteratively adjust the current model based on the evaluation parameter until the current model under the current iteration number reaches stability, and determine a model parameter under the current iteration number, where the evaluation parameter includes at least one of a root mean square error, a recall ratio, an accuracy ratio, a false alarm ratio, and a landslide density;
determining the sub-network corresponding to each time sequence window based on the model parameters under the current iteration times, and taking the current model comprising the time sequence sub-network corresponding to each time sequence window as the landslide prediction model after training is completed.
Optionally, the apparatus further comprises: a test module of a landslide prediction model; the test module of the landslide prediction model is used for acquiring a history test factor corresponding to each time sequence sub-network of the landslide prediction model after training;
and testing the landslide prediction model based on the historical test factors, the historical training factors and the historical verification factors corresponding to each time sequence sub-network.
Optionally, the test module of the landslide prediction model is further configured to input the corresponding historical training factors and the historical verification factors into each time sequence sub-network to obtain a prediction probability corresponding to each time sequence sub-network;
and determining a test result of the landslide prediction model according to the prediction probability corresponding to each time sequence sub-network and the history test factor corresponding to each time sequence sub-network.
Optionally, the landslide impact factor comprises a dynamic factor comprising rainfall and soil humidity and a static factor comprising elevation, slope, plane curvature, profile curvature, topography humidity index, water current intensity index, sediment transport index, topography roughness index, distance to fault
The landslide prediction device provided by the embodiment of the invention can execute the landslide prediction method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
Example IV
Fig. 6 is a schematic structural diagram of a landslide prediction apparatus according to a fourth embodiment of the present invention. Fig. 6 shows a block diagram of an exemplary landslide prediction apparatus 12 suitable for use in implementing embodiments of the invention. The landslide prediction apparatus 12 shown in fig. 6 is merely an example, and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.
As shown in fig. 6, the prediction device 12 of the landslide is in the form of a general purpose computing device. Components of the landslide prediction apparatus 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, a bus 18 that connects the various system components, including the system memory 28 and the processing units 16.
Bus 18 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, and a local bus using any of a variety of bus architectures. By way of example, and not limitation, such architectures include Industry Standard Architecture (ISA) bus, micro channel architecture (MAC) bus, enhanced ISA bus, video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus.
Landslide prediction device 12 typically comprises a variety of computer system readable media. Such media can be any available media that can be accessed by the landslide prediction device 12 and includes both volatile and nonvolatile media, removable and non-removable media.
The system memory 28 may include computer system readable media in the form of volatile memory, such as Random Access Memory (RAM) 30 and/or cache memory 32. Landslide prediction device 12 may further comprise other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 34 may be used to read from or write to non-removable, nonvolatile magnetic media (not shown in FIG. 6, commonly referred to as a "hard disk drive"). Although not shown in fig. 6, a magnetic disk drive for reading from and writing to a removable non-volatile magnetic disk (e.g., a "floppy disk"), and an optical disk drive for reading from or writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In such cases, each drive may be coupled to bus 18 through one or more data medium interfaces. The system memory 28 may include at least one program product having a set of program modules (e.g., a landslide impact factor acquisition module 310 and a landslide occurrence probability prediction module 320 of a landslide prediction device) configured to perform the functions of the various embodiments of the invention.
The program/utility 44 having a set of program modules 46 (e.g., landslide impact factor acquisition module 310 and landslide occurrence probability prediction module 320 of a landslide prediction device) may be stored in, for example, system memory 28, such program modules 46 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each or some combination of which may include an implementation of a network environment. Program modules 46 generally perform the functions and/or methods of the embodiments described herein.
The landslide prediction device 12 may also communicate with one or more external devices 14 (e.g., keyboard, pointing device, display 24, etc.), with one or more devices that enable a user to interact with the landslide prediction device 12, and/or with any device (e.g., network card, modem, etc.) that enables the landslide prediction device 12 to communicate with one or more other computing devices. Such communication may occur through an input/output (I/O) interface 22. Also, the landslide prediction device 12 may also communicate with one or more networks such as a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network such as the internet via the network adapter 20. As shown, the network adapter 20 communicates with other modules of the landslide prediction device 12 via the bus 18. It should be appreciated that although not shown in the figures, other hardware and/or software modules may be used in connection with the landslide prediction apparatus 12, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, data backup storage systems, and the like.
The processing unit 16 executes various functional applications and data processing by running programs stored in the system memory 28, for example, to implement a landslide prediction method provided by an embodiment of the present invention, the method including:
acquiring landslide influence factors of each slope unit of a region to be predicted in a preset first historical time period every day;
and determining landslide occurrence probability of each slope unit in the current day according to the landslide influence factors of each day in the first historical time period based on the landslide prediction model after training, wherein the landslide prediction model comprises at least two time sequence sub-networks, each time sequence sub-network is obtained by training according to the historical training factors and the historical verification factors of each slope unit in the sample area in the second historical time period, and the time sequence information of the historical verification factors of each time sequence sub-network is larger than the time sequence information of the historical training factors of the time sequence sub-network.
The processing unit 16 executes various functional applications and data processing by running programs stored in the system memory 28, for example, to implement a landslide prediction method provided by an embodiment of the present invention.
Of course, those skilled in the art will understand that the processor may also implement the technical solution of the landslide prediction method provided in any embodiment of the present invention.
Example five
The fifth embodiment of the present invention further provides a computer readable storage medium having a computer program stored thereon, the program when executed by a processor implementing a landslide prediction method as provided by the embodiment of the present invention, the method comprising:
acquiring landslide influence factors of each slope unit of a region to be predicted in a preset first historical time period every day;
and determining landslide occurrence probability of each slope unit in the current day according to the landslide influence factors of each day in the first historical time period based on the landslide prediction model after training, wherein the landslide prediction model comprises at least two time sequence sub-networks, each time sequence sub-network is obtained by training according to the historical training factors and the historical verification factors of each slope unit in the sample area in the second historical time period, and the time sequence information of the historical verification factors of each time sequence sub-network is larger than the time sequence information of the historical training factors of the time sequence sub-network.
Of course, the computer-readable storage medium provided by the embodiments of the present invention, on which the computer program stored, is not limited to the above-described method operations, but may also perform the related operations in a rainfall peak type division method provided by any of the embodiments of the present invention.
The computer storage media of embodiments of the invention may take the form of any combination of one or more computer-readable media. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. The computer readable storage medium can be, for example but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, or device, or any combination of the above. More specific examples (a non-exhaustive list) of the computer-readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, or device.
The computer readable signal medium may include a landslide impact factor, a landslide occurrence probability, etc., having computer readable program code embodied therein. Such forms of the transmitted landslide influence factor, landslide occurrence probability, and the like. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, or device.
Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
Computer program code for carrying out operations of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, smalltalk, C ++ and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider).
It should be noted that, in the embodiment of the landslide prediction apparatus described above, each included module is only divided according to the functional logic, but not limited to the above-mentioned division, as long as the corresponding function can be implemented; in addition, the specific names of the functional units are also only for distinguishing from each other, and are not used to limit the protection scope of the present invention.
Note that the above is only a preferred embodiment of the present invention and the technical principle applied. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, while the invention has been described in connection with the above embodiments, the invention is not limited to the embodiments, but may be embodied in many other equivalent forms without departing from the spirit or scope of the invention, which is set forth in the following claims.
Claims (8)
1. A method of predicting landslide, comprising:
acquiring landslide influence factors of each slope unit of a region to be predicted in a preset first historical time period every day;
determining landslide occurrence probability of each slope unit of the current day according to the landslide influence factors of each day in the first historical time period based on the trained landslide prediction model, wherein the landslide prediction model comprises at least two time sequence sub-networks, each time sequence sub-network is trained according to the historical training factors and the historical verification factors of each slope unit of a sample area in the second historical time period, and the time sequence information of the historical verification factors of each time sequence sub-network is larger than the time sequence information of the historical training factors of the time sequence sub-network;
the training method of the landslide prediction model comprises the following steps:
acquiring a history training factor and a history verification factor of each slope unit of the sample area in a second history time period;
inputting the historical training factors into a current model according to the time sequence information to obtain historical landslide prediction information of each slope unit;
iteratively adjusting model parameters of the current model according to the historical verification factors and the historical landslide prediction information until the current model under the current iteration times reaches stability, and taking the current model under the current iteration times as a landslide prediction model after training;
the step of inputting the historical training factors into the current model according to the time sequence information to obtain the historical landslide prediction information of each slope unit comprises the following steps:
respectively determining a time sequence window corresponding to each time sequence sub-network in the current model;
based on the corresponding relation between the time sequence sub-network and the time sequence window, inputting a history training factor corresponding to the time sequence window into the time sequence sub-network to obtain history landslide prediction information of each slope unit corresponding to the history training factor;
the training processes of the at least two time sequence sub-networks are performed simultaneously and do not interfere with each other.
2. The method according to claim 1, wherein iteratively adjusting the model parameters of the current model according to the history verification factor and the prediction information until the current model at the current iteration number reaches a stability, and using the current model at the current iteration number as the trained landslide prediction model comprises:
calculating an evaluation parameter of the current model based on the historical landslide prediction information and the historical verification factor, and iteratively adjusting the current model based on the evaluation parameter until the current model under the current iteration number reaches stability, and determining a model parameter under the current iteration number, wherein the evaluation parameter comprises at least one of root mean square error, recall ratio, precision ratio, false alarm ratio and landslide density;
determining the sub-network corresponding to each time sequence window based on the model parameters under the current iteration times, and taking the current model comprising the time sequence sub-network corresponding to each time sequence window as the landslide prediction model after training is completed.
3. The method of claim 1, wherein after obtaining the trained landslide prediction model, the method further comprises:
acquiring a history test factor corresponding to each time sequence sub-network of the landslide prediction model after training;
and testing the landslide prediction model based on the historical test factors, the historical training factors and the historical verification factors corresponding to each time sequence sub-network.
4. A method according to claim 3, wherein said testing said landslide prediction model based on said historical test factors, said historical training factors and historical validation factors corresponding to each of said time series sub-networks comprises:
inputting the corresponding history training factors and the history verification factors into each time sequence sub-network to obtain the corresponding prediction probability of each time sequence sub-network;
and determining a test result of the landslide prediction model according to the prediction probability corresponding to each time sequence sub-network and the history test factor corresponding to each time sequence sub-network.
5. The method of claim 1, wherein the landslide impact factor comprises a dynamic factor comprising rainfall and soil moisture and a static factor comprising at least one of elevation, grade, slope, plane curvature, profile curvature, terrain moisture index, water current intensity index, sediment transport index, terrain roughness index, fault distance, river distance, road distance, lithology, land utilization, and vegetation coverage.
6. A landslide prediction apparatus comprising:
the landslide influence factor acquisition module is used for acquiring the landslide influence factors of each slope unit of the area to be predicted in a preset first historical time period;
the landslide occurrence probability prediction module is used for determining the landslide occurrence probability of each slope unit in the current day according to the landslide influence factor of each day in the first historical time period based on the landslide prediction model after training, wherein the landslide prediction model comprises at least two time sequence sub-networks, each time sequence sub-network is obtained by training according to the historical training factors and the historical verification factors of each slope unit in a sample area in the second historical time period, and the time sequence information of the historical verification factors of each time sequence sub-network is larger than the time sequence information of the historical training factors of the time sequence sub-network;
the training module of the landslide prediction model is used for acquiring historical training factors and historical verification factors of each slope unit of the sample area in the second historical time period;
inputting the historical training factors into a current model according to the time sequence information to obtain historical landslide prediction information of each slope unit;
iteratively adjusting model parameters of the current model according to the historical verification factors and the historical landslide prediction information until the current model under the current iteration times reaches stability, and taking the current model under the current iteration times as a landslide prediction model after training;
the training module of the landslide prediction model is further used for respectively determining a time sequence window corresponding to each time sequence sub-network in the current model;
based on the corresponding relation between the time sequence sub-network and the time sequence window, inputting a history training factor corresponding to the time sequence window into the time sequence sub-network to obtain history landslide prediction information of each slope unit corresponding to the history training factor;
the training processes of the at least two time sequence sub-networks are performed simultaneously and do not interfere with each other.
7. A landslide prediction apparatus characterized by comprising:
one or more processors;
a memory for storing one or more programs;
when executed by the one or more processors, causes the one or more processors to implement the landslide prediction method of any one of claims 1-5.
8. A storage medium containing computer executable instructions which, when executed by a computer processor, are for performing a method of landslide prediction as claimed in any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011501960.8A CN112561171B (en) | 2020-12-17 | 2020-12-17 | Landslide prediction method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011501960.8A CN112561171B (en) | 2020-12-17 | 2020-12-17 | Landslide prediction method, device, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112561171A CN112561171A (en) | 2021-03-26 |
| CN112561171B true CN112561171B (en) | 2023-12-05 |
Family
ID=75063404
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011501960.8A Active CN112561171B (en) | 2020-12-17 | 2020-12-17 | Landslide prediction method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112561171B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114329663B (en) * | 2021-12-27 | 2022-08-05 | 中国自然资源航空物探遥感中心 | Slope unit dividing method based on scale of historical geological disasters |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2017101992A (en) * | 2015-12-01 | 2017-06-08 | エーエルティー株式会社 | Disaster situation monitoring/warning/evacuation guidance system |
| CN111027686A (en) * | 2019-12-26 | 2020-04-17 | 杭州鲁尔物联科技有限公司 | Landslide displacement prediction method, device and equipment |
| CN111079084A (en) * | 2019-12-04 | 2020-04-28 | 清华大学 | Information forwarding probability prediction method and system based on long-time and short-time memory network |
| CN111784044A (en) * | 2020-06-29 | 2020-10-16 | 杭州鲁尔物联科技有限公司 | Landslide prediction method, device, equipment and storage medium |
| US10824940B1 (en) * | 2016-11-30 | 2020-11-03 | Amazon Technologies, Inc. | Temporal ensemble of machine learning models trained during different time intervals |
-
2020
- 2020-12-17 CN CN202011501960.8A patent/CN112561171B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2017101992A (en) * | 2015-12-01 | 2017-06-08 | エーエルティー株式会社 | Disaster situation monitoring/warning/evacuation guidance system |
| US10824940B1 (en) * | 2016-11-30 | 2020-11-03 | Amazon Technologies, Inc. | Temporal ensemble of machine learning models trained during different time intervals |
| CN111079084A (en) * | 2019-12-04 | 2020-04-28 | 清华大学 | Information forwarding probability prediction method and system based on long-time and short-time memory network |
| CN111027686A (en) * | 2019-12-26 | 2020-04-17 | 杭州鲁尔物联科技有限公司 | Landslide displacement prediction method, device and equipment |
| CN111784044A (en) * | 2020-06-29 | 2020-10-16 | 杭州鲁尔物联科技有限公司 | Landslide prediction method, device, equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 降雨型滑坡灾害的特征聚合决策树预测模型;赵晓萌;灾害学;第35卷(第1期);27-31 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112561171A (en) | 2021-03-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108563548B (en) | Abnormality detection method and apparatus | |
| Calafat et al. | Trends in Europe storm surge extremes match the rate of sea-level rise | |
| CN109543906B (en) | Atmospheric visibility prediction method and equipment | |
| US11206555B2 (en) | Method for implementing antenna azimuth correction based on user data | |
| CN112200358A (en) | Landslide grade prediction method, device, equipment and storage medium | |
| US20110113009A1 (en) | Outlier data point detection | |
| CN110705115A (en) | Meteorological forecasting method and system based on deep belief network | |
| CN114169604A (en) | Performance index abnormality detection method, abnormality detection device, electronic apparatus, and storage medium | |
| CN113361690A (en) | Water quality prediction model training method, water quality prediction device, water quality prediction equipment and medium | |
| CN111369056B (en) | Geological disaster prediction method and electronic equipment | |
| CN112561171B (en) | Landslide prediction method, device, equipment and storage medium | |
| CN110134754B (en) | Method, device, server and medium for predicting operation duration of region interest point | |
| CN115659335A (en) | Block chain intelligent contract vulnerability detection method and device based on mixed fuzzy test | |
| CN113253336B (en) | Earthquake prediction method and system based on deep learning | |
| CN112651172B (en) | Rainfall peak type dividing method, device, equipment and storage medium | |
| Voutchkova et al. | Parameter sensitivity of automated baseflow separation for snowmelt‐dominated watersheds and new filtering procedure for determining end of snowmelt period | |
| CN113159934A (en) | Method and system for predicting passenger flow of network, electronic equipment and storage medium | |
| CN110555352A (en) | interest point identification method, device, server and storage medium | |
| US10878329B1 (en) | Probabilistic accumulation approach to assess primary uncertainty in catastrophe models | |
| CN111027680B (en) | Monitoring quantity uncertainty prediction method and system based on variational self-encoder | |
| RU2591239C1 (en) | Systems and methods for assessment of capability of collector system | |
| CN115755131A (en) | Satellite positioning method, device and medium | |
| CN115391746A (en) | Interpolation method, device, electronic device and medium for meteorological element data | |
| CN112529315A (en) | Landslide prediction method, device, equipment and storage medium | |
| CN114842303A (en) | Self-training optimization method and device, electronic equipment and computer-readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |