CN111027686A - Landslide displacement prediction method, device and equipment - Google Patents
Landslide displacement prediction method, device and equipment Download PDFInfo
- Publication number
- CN111027686A CN111027686A CN201911367893.2A CN201911367893A CN111027686A CN 111027686 A CN111027686 A CN 111027686A CN 201911367893 A CN201911367893 A CN 201911367893A CN 111027686 A CN111027686 A CN 111027686A
- Authority
- CN
- China
- Prior art keywords
- displacement
- displacement data
- neural network
- convolution
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- 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
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B21/00—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant
- G01B21/02—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant for measuring length, width, or thickness
-
- 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 invention discloses a method, a device, equipment and a storage medium for predicting landslide displacement, wherein the prediction method comprises the following steps: acquiring displacement data of a monitoring point; and predicting the displacement of the monitoring point based on a pre-established time convolution neural network and the displacement data, wherein the time convolution neural network comprises at least two convolution hiding layers, the output of at least one convolution hiding layer is determined by the latest displacement data with a set number, the output of one convolution hiding layer is determined by all displacement data, and the loss function of the time convolution neural network comprises all value errors and extremum errors. According to the technical scheme of the embodiment of the invention, the landslide displacement is predicted through the time convolution neural network, so that the automatic and accurate prediction of the landslide displacement is realized, the extreme value is considered in the loss function, and the prediction accuracy of the mutation part is greatly improved.
Description
Technical Field
The embodiment of the invention relates to the technical field of data processing, in particular to a landslide displacement prediction method, a landslide displacement prediction device and equipment.
Background
China is a country with multiple geological disasters, and collapse, landslide and debris flow disasters almost spread over mountain and hilly areas of various provinces in China, wherein the landslide disasters account for about 66.8 percent of the total number of geological disasters in China, and have important significance in predicting the landslide disasters. And landslide displacement prediction is one of basic works for landslide disaster prevention and control and is an important part for landslide stability evaluation and deformation damage prediction.
Many scholars have conducted extensive research on landslide displacement prediction at this stage. The displacement prediction model is mainly divided into a statistical prediction model and a nonlinear prediction model. The statistical prediction model comprises a grey prediction model, a regression analysis model, a time series model, an exponential smoothing method and the like. The nonlinear predictive model is mainly a series of methods represented by deep learning. However, the above models have some problems, such as that the statistical prediction model is relatively simple, but it cannot well fit the nonlinear characteristic of the landslide displacement change, so that it is difficult to improve the prediction accuracy. Although the deep learning model has nonlinear processing capability and can obtain better prediction results on the whole, the deep learning model can only perform more accurate prediction on stable changes and has poor prediction effect on mutation conditions.
Disclosure of Invention
The invention provides a landslide displacement prediction method, a landslide displacement prediction device and equipment, which are used for realizing automatic landslide displacement prediction based on a convolutional neural network, wherein a loss function of the convolutional neural network comprises extreme value information, and the prediction precision of a sudden change condition is greatly improved.
In a first aspect, an embodiment of the present invention provides a method for predicting landslide displacement, where the method includes:
acquiring displacement data of a monitoring point;
and predicting the displacement of the monitoring point based on a pre-established time convolution neural network and the displacement data, wherein the time convolution neural network comprises at least two convolution hiding layers, the output of at least one convolution hiding layer is determined by the latest displacement data with a set number, the output of one convolution hiding layer is determined by all displacement data, and the loss function of the time convolution neural network comprises all value errors and extremum errors.
In a second aspect, an embodiment of the present invention further provides a device for predicting landslide displacement, where the device includes:
the displacement data acquisition module is used for acquiring the displacement data of the monitoring point;
and the displacement prediction module is used for predicting the displacement of the monitoring point based on a pre-established time convolution neural network and the displacement data, wherein the time convolution neural network comprises at least two convolution hidden layers, the output of at least one convolution hidden layer is determined by the latest displacement data with a set number, the output of one convolution hidden layer is determined by all displacement data, and the loss function of the time convolution neural network comprises all value errors and extremum errors.
In a third aspect, an embodiment of the present invention further provides a device for predicting landslide displacement, where the device includes:
one or more processors;
a memory for storing one or more programs;
when executed by the one or more processors, cause the one or more processors to implement a method of predicting landslide displacement as provided in any embodiment of the invention.
In a fourth aspect, embodiments of the present invention further provide a storage medium containing computer-executable instructions, which when executed by a computer processor, are configured to perform the method for predicting landslide displacement provided by any of the embodiments of the present invention.
According to the technical scheme of the embodiment of the invention, the landslide displacement is predicted by establishing the time convolution neural network, so that the automatic prediction of the landslide displacement is realized; the nonlinear characteristic of the landslide displacement change is well fitted by adopting the time convolution neural network, the loss function of the time convolution neural network consists of all value errors and extreme value errors, the conditions that the displacement change is stable and sudden change are fully considered, and the precision of displacement prediction is improved.
Drawings
FIG. 1 is a flow chart of a method for predicting landslide displacement according to one embodiment of the present invention;
fig. 2A is a flowchart of a method for predicting landslide displacement according to a second embodiment of the present invention;
FIG. 2B is a schematic diagram of a time convolutional neural network according to a second embodiment of the present invention;
fig. 3 is a schematic diagram of a landslide displacement prediction apparatus according to a third embodiment of the present invention;
fig. 4 is a schematic diagram of a landslide displacement prediction apparatus according to a fourth embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
Example one
Fig. 1 is a flowchart of a method for predicting landslide displacement according to an embodiment of the present invention, where the present embodiment is applicable to a case of predicting landslide displacement, and the method may be executed by a device for predicting landslide displacement, where the device may be implemented by software and hardware, and as shown in fig. 1, the method specifically includes the following steps:
and step 110, obtaining displacement data of the monitoring points.
Wherein, the monitoring point refers to a landslide detection point. Specifically, the displacement data may be obtained by one or more displacement sensors disposed at the monitoring point.
Specifically, the number of monitoring points may be one, five, ten or other values.
And 120, predicting the displacement of the monitoring point based on a pre-established time convolution neural network and the displacement data.
The time convolution neural network comprises at least two convolution hidden layers, the output of at least one convolution hidden layer is determined by the latest displacement data with a set number, the output of one convolution hidden layer is determined by all the displacement data, and the loss function of the time convolution neural network comprises all value errors and extreme value errors.
The number of settings may be 1, 3, 5, 7 or other values. The extreme value error refers to the error of the displacement corresponding to the extreme value position in the original displacement data and the predicted value thereof, and all value errors refer to the errors of all the displacement data and the predicted value thereof.
Specifically, a time Convolutional Neural Network (TCN), also called a time-series Convolutional Neural Network, is a novel algorithm for solving time-series prediction. The method mainly comprises a convolution part and a Recurrent Neural Network (RNN) part. The convolution part is mainly used for performing convolution processing such as causal convolution, expansion convolution and residual convolution on input data, and the recurrent neural network part is mainly used for receiving the output of the convolution part and connecting each node in a recursive mode, wherein the convolution part comprises forward propagation and backward propagation, and therefore a predicted value is output.
Further, the time convolution neural network includes a convolution network and an RNN network, wherein the convolution network includes a plurality of convolution hidden layers, such as 3, and each convolution hidden layer sends its convolution result to a corresponding hidden layer of the RNN network as an input of the hidden layer of the RNN network. The RNN network may be an LSTM network (Long Short Term Memory).
Optionally, the activation function of the time convolution neural network includes a linear rectification function (recti) and a hyperbolic tangent function (tanh).
Specifically, the activation function of the convolution portion of the time-convolved neural network may employ a ReLu function, and the RNN portion may employ a tanh function.
Of course, other functions may be selected as the activation function, such as a Sigmoid function, a tanh function, and the like.
Optionally, the number of the convolution concealment layers of the time convolution neural network is 3, the output of one convolution concealment layer is determined by the latest displacement data of the first set number, the output of one convolution concealment layer is determined by the latest displacement data of the second set number, and the output of one convolution concealment layer is determined by all the displacement data.
The first set number may be 3, 5 or other values, and the second set number may be 7, 9 or other values.
Specifically, let X be [ X ] as the displacement data1x2…xT]That is, the displacement data includes displacement acquired at T times, the time convolution neural network includes 3 convolution hidden layers, where the hole coefficients of each layer are 1, 2, and 4, respectively, and the convolution kernel size is 3, so that the outputs of the three obtained convolution hidden layers are o1、o2And o3Wherein o is1The covered information is xT-2、xT-1And xTI.e. o1Contains only recent displacement data; o2The covered information is xT-6、xT-5、xT-4、xT-3、xT-2、xT-1And xTI.e. o2Contains only near term and medium term displacement data; o3The full displacement data is covered. O to put convolution hidden layer out1、o2And o3And the displacement is input into the cyclic neural network to predict the displacement according to the recent, medium and all displacement data, so that the weights of the recent and medium data are improved, and the prediction precision is further improved.
Specifically, the loss function of the time-convolved nerve consists of two parts, namely, a data error part and an extremum error part. Specifically, the Mean Absolute Error (MAE) and the extreme Mean Error may be used. The extreme value average error refers to an average value of absolute values of errors between the displacement data corresponding to all extreme values in the displacement data and the predicted displacement data.
Further, all value errors of the loss function of the time-convolved nerve are different from the corresponding weights of the extreme value errors.
Optionally, the loss function of the time convolution neural network is:
wherein α is the first weight coefficient, β is the second weight coefficient, yiFor the ith piece of displacement data, the displacement data,is the predicted value corresponding to the ith displacement data, m is the total number of predictions, yext jFor the displacement data corresponding to the jth extreme value,and the predicted value of the displacement data corresponding to the jth extreme value is p, the total number of the extreme values of the displacement data is p, and l (-) is an error loss function.
The loss of all data and the loss of the extreme point are weighted, so that the prediction accuracy of the data is guaranteed, meanwhile, the data corresponding to the extreme point independently form one part of a loss function to be analyzed, and the prediction accuracy of the displacement of a mutation part in the original data is improved.
Specifically, the Error loss function l (-) may be an average absolute Error, a Mean Square Error (MSE), a Log-Cosh loss function, a Huber loss function, or the like.
Illustratively, the error loss function l (-) may be the mean absolute error, and the corresponding loss function of the time-convolved neural network is expressed as:
the error loss function l (-) can be the mean square error, and the corresponding loss function of the time convolutional neural network is expressed as:
according to the technical scheme of the embodiment of the invention, the landslide displacement is predicted by establishing the time convolution neural network, so that the automatic prediction of the landslide displacement is realized; the nonlinear characteristic of the landslide displacement change is well fitted by adopting the time convolution neural network, the loss function of the time convolution neural network consists of all value errors and extreme value errors, the conditions that the displacement change is stable and sudden change are fully considered, and the precision of displacement prediction is improved.
Example two
Fig. 2A is a flowchart of a method for predicting landslide displacement according to a second embodiment of the present invention, where this embodiment is a further refinement and supplement to the previous embodiment, and the method for predicting landslide displacement according to the second embodiment of the present invention further includes: filling missing values in the displacement data based on a Krigin method; and filtering the displacement data and normalizing the displacement data based on a Kalman filtering algorithm.
As shown in fig. 2A, the method includes the steps of:
and step 210, obtaining displacement data of the monitoring points.
And step 220, filling missing values in the displacement data based on a Kriging method.
When the displacement sensor of the monitoring point collects data, displacement data at individual moments are lost due to sensor faults or other reasons, and then the lost value needs to be filled according to the known displacement data.
The Kriging method (Kriging) is also called a space optimal unbiased estimator, and is a regression algorithm based on a covariance function. When missing value supplement or interpolation is carried out by the Krigin method, not only the correlation between the position to be estimated and the known data position is considered, but also the spatial correlation of the data is considered, and the interpolation result is more accurate.
Of course, other algorithms for missing value filling may be used, such as cubic spline algorithm, tree model algorithm (missfiest), etc.
And 230, filtering the displacement data based on a Kalman filtering algorithm.
Because various interferences can be received in the data acquisition and transmission processes, the displacement data needs to be filtered to improve the cleanliness of the displacement data, thereby improving the precision of displacement prediction.
And 240, performing normalization processing on the filtered displacement data.
Specifically, the displacement data sequence of the monitoring point after missing value filling and filtering is Where T is the last moment of monitoring. And carrying out maximum and minimum normalization on the displacement data sequence X, wherein the normalized displacement data sequence is X, and X is [ X ═ X1x2… xT]
Wherein, i is 1, 2, …, T, xiAs displacement data at the ith timeThe normalized displacement data of (a) is,for shifting data sequencesThe minimum value of the displacement in (a),for shifting data sequencesMaximum value of displacement in (1).
And step 250, predicting the displacement of the monitoring point based on a pre-established time convolution neural network and the displacement data.
The number of the convolution hiding layers of the time convolution neural network is 3, the output of one convolution hiding layer is determined by the latest displacement data of a first set number, the output of one convolution hiding layer is determined by the latest displacement data of a second set number, and the output of one convolution hiding layer is determined by all the displacement data. The loss function of the time convolutional neural network is:
wherein α is the first weight coefficient, β is the second weight coefficient, yiFor the ith piece of displacement data, the displacement data,is the predicted value corresponding to the ith displacement data, m is the total number of predictions, yext jFor the displacement data corresponding to the jth extreme value,and p is the total number of extreme values of the displacement data.
The loss function comprises the errors of the original displacement and the predicted displacement corresponding to the extreme position, and the weight of displacement prediction at the extreme position is enhanced, so that the prediction precision of the mutation condition is improved, and the application range and precision of the prediction method are improved.
Specifically, fig. 2B is a schematic diagram of a time convolution neural network according to a second embodiment of the present invention, and as can be seen from fig. 2B, the time convolution neural network includes 3 convolution hidden layers, a hole coefficient d is 1, 2, 4, a convolution kernel size k is 3, an activation function is a ReLU function, and normalized displacement data Input the time convolutionNeural network, obtaining o by causal convolution1、o2And o3Wherein o is1、o2And o3Respectively, the output of each convolution hidden layer, wherein o1The information covered isAndi.e. o1Contains only recent displacement data; o2The information covered is Andi.e. o2Contains only near term and medium term displacement data; o3The full displacement data is covered. Will o1、o2And o3As input to the RNN network, h0、h1、h2And h3Representing the state of each hidden layer of each recurrent neural network, wherein the state h1Can be according to o1And h0Is obtained in state h2Can be according to o2And h1Is obtained in state h3Can be according to o3And h2Is obtained, and can further be according to h3Displacement prediction is carried out at the time T +1, and displacement data at the time T +1 are obtainedAnd reversely propagating according to the loss function to modify the parameters of the convolution hidden layer and the RNN hidden layer so as to improve the prediction accuracy.
Further, to obtain the predicted displacement at time T +2The prediction of the time T +1 can be shiftedBy analogy, the predicted displacement data for setting the future time can be obtained as input.
Specifically, before training, the time convolution neural network needs to initialize partial parameters, such as a learning rate, the number of hidden neurons, the number of iteration rounds, a weight vector, and the like.
Specifically, in the time convolution neural network training, the historical displacement data may be calculated by a method of 8: scale of 2 is divided into training and validation sets.
Further, after the predicted displacement data is obtained, the predicted displacement data needs to be restored according to normalization processing. The reduction process is normalized inverse operation, and the predicted displacement data output by the time convolution neural networkThe reduction algorithm is specifically as follows:
According to the technical scheme of the embodiment of the invention, the landslide displacement is predicted by establishing the time convolution neural network, so that the automatic prediction of the landslide displacement is realized; the nonlinear characteristic of landslide displacement change is well fitted by adopting a time convolution neural network, a loss function of the time convolution neural network consists of all value errors and extreme value errors, the conditions that the displacement change is stable and sudden change are fully considered, and the weight of displacement prediction at an extreme value position is enhanced, so that the prediction precision of the sudden change condition is improved, and the precision of the displacement prediction is improved.
EXAMPLE III
Fig. 3 is a schematic diagram of a landslide displacement prediction apparatus according to a third embodiment of the present invention, and as shown in fig. 3, the apparatus includes: a displacement data acquisition module 310 and a displacement prediction module 320.
The displacement data acquisition module 310 is configured to acquire displacement data of a monitoring point; and a displacement prediction module 320, configured to predict the displacement of the monitoring point based on a pre-established time convolutional neural network and the displacement data, where the time convolutional neural network includes at least two convolutional concealment layers, an output of at least one convolutional concealment layer is determined by a set number of latest displacement data, an output of one convolutional concealment layer is determined by all displacement data, and a loss function of the time convolutional neural network is formed by an average absolute error and an extremum average error.
According to the technical scheme of the embodiment of the invention, the landslide displacement is predicted by establishing the time convolution neural network, so that the automatic prediction of the landslide displacement is realized; the nonlinear characteristic of the landslide displacement change is well fitted by adopting the time convolution neural network, the loss function of the time convolution neural network consists of all value errors and extreme value errors, the conditions that the displacement change is stable and sudden change are fully considered, and the precision of displacement prediction is improved.
Optionally, the landslide displacement prediction apparatus further includes:
and the missing value filling module is used for filling missing values in the displacement data based on a Kriging method.
Optionally, the landslide displacement prediction apparatus further includes:
and the Kalman filtering module is used for filtering the displacement data based on a Kalman filtering algorithm.
Optionally, the landslide displacement prediction apparatus further includes:
and the normalization processing module is used for performing normalization processing on the displacement data.
Optionally, the activation function of the time convolution neural network is a linear rectification function.
Optionally, the number of the convolution concealment layers of the time convolution neural network is 3, the output of one convolution concealment layer is determined by the latest displacement data of the first set number, the output of one convolution concealment layer is determined by the latest displacement data of the second set number, and the output of one convolution concealment layer is determined by all the displacement data.
Optionally, the loss function of the time convolution neural network is:
wherein α is the first weight coefficient, β is the second weight coefficient, yiFor the ith piece of displacement data, the displacement data,is the predicted value corresponding to the ith displacement data, m is the total number of predictions, yext jFor the displacement data corresponding to the jth extreme value,and the predicted value of the displacement data corresponding to the jth extreme value is p, the total number of the extreme values of the displacement data is p, and l (-) is an error loss function.
The landslide displacement prediction device provided by the embodiment of the invention can execute the landslide displacement prediction method provided by any embodiment of the invention, and has corresponding functional modules and beneficial effects of the execution method.
Example four
Fig. 4 is a schematic structural diagram of a landslide displacement prediction apparatus according to a fourth embodiment of the present invention, as shown in fig. 4, the apparatus includes a processor 410, a memory 420, an input device 430, and an output device 440; the number of the device processors 410 may be one or more, and one processor 410 is taken as an example in fig. 4; the processor 410, the memory 420, the input device 430 and the output device 440 in the apparatus may be connected by a bus or other means, for example, in fig. 4.
The memory 420 serves as a computer-readable storage medium, and may be used to store software programs, computer-executable programs, and modules, such as program instructions/modules corresponding to the prediction method of landslide displacement in the embodiment of the present invention (for example, the displacement data acquisition module 310 and the displacement prediction module 320 in the prediction device of landslide displacement). The processor 410 executes various functional applications of the device and data processing by executing software programs, instructions and modules stored in the memory 420, namely, the above-mentioned prediction method of landslide displacement is realized.
The memory 420 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required for at least one function; the storage data area may store data created according to the use of the terminal, and the like. Further, the memory 420 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some examples, the memory 420 may further include memory located remotely from the processor 410, which may be connected to the device/terminal/server via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The input means 430 may be used to receive input numeric or character information and generate key signal inputs related to user settings and function control of the apparatus. The output device 440 may include a display device such as a display screen.
EXAMPLE five
Fifth embodiment of the present invention also provides a storage medium containing computer-executable instructions which, when executed by a computer processor, perform a method of predicting landslide displacement, the method comprising:
acquiring displacement data of a monitoring point;
and predicting the displacement of the monitoring point based on a pre-established time convolution neural network and the displacement data, wherein the time convolution neural network comprises at least two convolution hiding layers, the output of at least one convolution hiding layer is determined by the latest displacement data with a set number, the output of one convolution hiding layer is determined by all displacement data, and the loss function of the time convolution neural network comprises all value errors and extremum errors.
Of course, the storage medium containing the computer-executable instructions provided by the embodiments of the present invention is not limited to the method operations described above, and may also perform related operations in the method for predicting landslide displacement provided by any embodiment of the present invention.
From the above description of the embodiments, it is obvious for those skilled in the art that the present invention can be implemented by software and necessary general hardware, and certainly, can also be implemented by hardware, but the former is a better embodiment in many cases. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which may be stored in a computer-readable storage medium, such as a floppy disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a FLASH Memory (FLASH), a hard disk or an optical disk of a computer, and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device) to execute the methods according to the embodiments of the present invention.
It should be noted that, in the embodiment of the foregoing landslide displacement prediction apparatus, the included units and modules are merely divided according to functional logic, but are not limited to the above division as long as the corresponding functions can be implemented; in addition, specific names of the functional units are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present invention.
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. 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, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims (10)
1. A method of predicting landslide displacement, comprising:
acquiring displacement data of a monitoring point;
and predicting the displacement of the monitoring point based on a pre-established time convolution neural network and the displacement data, wherein the time convolution neural network comprises at least two convolution hiding layers, the output of at least one convolution hiding layer is determined by the latest displacement data with a set number, the output of one convolution hiding layer is determined by all displacement data, and the loss function of the time convolution neural network comprises all value errors and extremum errors.
2. The method of claim 1, after obtaining displacement data for the monitoring point, further comprising:
and filling missing values in the displacement data based on a Krigin method.
3. The method of claim 1, after obtaining displacement data for the monitoring point, further comprising:
and filtering the displacement data based on a Kalman filtering algorithm.
4. The method of claim 1, after obtaining displacement data for the monitoring point, further comprising:
and carrying out normalization processing on the displacement data.
5. The method of claim 1, wherein the activation function of the time-convolved neural network comprises a linear rectification function and a hyperbolic tangent function.
6. The method of claim 1, wherein the number of convolutional concealment layers in the time convolutional neural network is 3, the output of one convolutional concealment layer is determined by a first set number of the latest displacement data, the output of one convolutional concealment layer is determined by a second set number of the latest displacement data, and the output of one convolutional concealment layer is determined by all the displacement data.
7. The method of claim 1, wherein the loss function of the time-convolutional neural network is:
wherein α is the first weight coefficient, β is the second weight coefficient, yiFor the ith piece of displacement data, the displacement data,is the predicted value corresponding to the ith displacement data, m is the total number of predictions, yext jFor the displacement data corresponding to the jth extreme value,and the predicted value of the displacement data corresponding to the jth extreme value is p, the total number of the extreme values of the displacement data is p, and l (-) is an error loss function.
8. A landslide displacement prediction apparatus comprising:
the displacement data acquisition module is used for acquiring the displacement data of the monitoring point;
and the displacement prediction module is used for predicting the displacement of the monitoring point based on a pre-established time convolution neural network and the displacement data, wherein the time convolution neural network comprises at least two convolution hidden layers, the output of at least one convolution hidden layer is determined by the latest displacement data with a set number, the output of one convolution hidden layer is determined by all displacement data, and the loss function of the time convolution neural network comprises all value errors and extremum errors.
9. An apparatus for predicting landslide displacement, said apparatus comprising:
one or more processors;
a memory for storing one or more programs;
when executed by the one or more processors, cause the one or more processors to implement a method of predicting landslide displacement as recited in any one of claims 1-7.
10. A storage medium containing computer-executable instructions for performing the method of predicting landslide displacement according to any one of claims 1-7 when executed by a computer processor.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911367893.2A CN111027686B (en) | 2019-12-26 | 2019-12-26 | Landslide displacement prediction method, device and equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911367893.2A CN111027686B (en) | 2019-12-26 | 2019-12-26 | Landslide displacement prediction method, device and equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111027686A true CN111027686A (en) | 2020-04-17 |
CN111027686B CN111027686B (en) | 2023-06-16 |
Family
ID=70214741
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201911367893.2A Active CN111027686B (en) | 2019-12-26 | 2019-12-26 | Landslide displacement prediction method, device and equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111027686B (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112242060A (en) * | 2020-10-16 | 2021-01-19 | 深圳先进技术研究院 | Traffic flow prediction method and apparatus, computer device, and readable storage medium |
CN112507780A (en) * | 2020-10-21 | 2021-03-16 | 西安工程大学 | Landslide sensitivity evaluation method based on TCN deep learning |
CN112561171A (en) * | 2020-12-17 | 2021-03-26 | 杭州鲁尔物联科技有限公司 | Landslide prediction method, device, equipment and storage medium |
CN113065455A (en) * | 2021-03-30 | 2021-07-02 | 中国水利水电科学研究院 | Landslide risk inspection method and system based on deep learning |
CN113069117A (en) * | 2021-04-02 | 2021-07-06 | 中山大学 | Electroencephalogram emotion recognition method and system based on time convolution neural network |
CN113076754A (en) * | 2021-03-23 | 2021-07-06 | 山东师范大学 | False comment detection method and system based on knowledge integration |
CN116739184A (en) * | 2023-08-08 | 2023-09-12 | 四川川核地质工程有限公司 | Landslide prediction method and system |
CN116993803A (en) * | 2023-09-26 | 2023-11-03 | 南方科技大学 | Landslide deformation monitoring method and device and electronic equipment |
CN117196353A (en) * | 2023-11-07 | 2023-12-08 | 山东省济宁生态环境监测中心(山东省南四湖东平湖流域生态环境监测中心) | Environmental pollution assessment and monitoring method and system based on big data |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140136140A1 (en) * | 2012-11-13 | 2014-05-15 | Elwha Llc | Systems and methods for detecting overhead line motion |
CN107578093A (en) * | 2017-09-14 | 2018-01-12 | 长安大学 | The Elman neural network dynamic Forecasting Methodologies of Landslide Deformation |
CN109492793A (en) * | 2018-09-29 | 2019-03-19 | 桂林电子科技大学 | A kind of dynamic grey Fil Haast neural network landslide deformation prediction method |
CN110213287A (en) * | 2019-06-12 | 2019-09-06 | 北京理工大学 | A kind of double mode invasion detecting device based on ensemble machine learning algorithm |
CN110427654A (en) * | 2019-07-09 | 2019-11-08 | 中国地质大学(武汉) | A kind of predictive model of landslide construction method and system based on sensitiveness |
CN110470259A (en) * | 2019-07-09 | 2019-11-19 | 西安工程大学 | Landslide displacement dynamic prediction method based on LSTM |
-
2019
- 2019-12-26 CN CN201911367893.2A patent/CN111027686B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140136140A1 (en) * | 2012-11-13 | 2014-05-15 | Elwha Llc | Systems and methods for detecting overhead line motion |
CN107578093A (en) * | 2017-09-14 | 2018-01-12 | 长安大学 | The Elman neural network dynamic Forecasting Methodologies of Landslide Deformation |
CN109492793A (en) * | 2018-09-29 | 2019-03-19 | 桂林电子科技大学 | A kind of dynamic grey Fil Haast neural network landslide deformation prediction method |
CN110213287A (en) * | 2019-06-12 | 2019-09-06 | 北京理工大学 | A kind of double mode invasion detecting device based on ensemble machine learning algorithm |
CN110427654A (en) * | 2019-07-09 | 2019-11-08 | 中国地质大学(武汉) | A kind of predictive model of landslide construction method and system based on sensitiveness |
CN110470259A (en) * | 2019-07-09 | 2019-11-19 | 西安工程大学 | Landslide displacement dynamic prediction method based on LSTM |
Non-Patent Citations (3)
Title |
---|
PEIHONG XIE ET AL.: "The Application of Long Short-Term Memory(LSTM) Method on Displacement Prediction of Multifactor-Induced Landslides" * |
宣杰: "基于数据挖掘技术的信任模型研究" * |
曹国清 等: "基于PCA-LSTM的多变量矿山排土场滑坡预警研究" * |
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2022077767A1 (en) * | 2020-10-16 | 2022-04-21 | 深圳先进技术研究院 | Traffic flow prediction method and apparatus, computer device, and readable storage medium |
CN112242060A (en) * | 2020-10-16 | 2021-01-19 | 深圳先进技术研究院 | Traffic flow prediction method and apparatus, computer device, and readable storage medium |
CN112242060B (en) * | 2020-10-16 | 2022-06-21 | 深圳先进技术研究院 | Traffic flow prediction method and apparatus, computer device, and readable storage medium |
CN112507780A (en) * | 2020-10-21 | 2021-03-16 | 西安工程大学 | Landslide sensitivity evaluation method based on TCN deep learning |
CN112561171B (en) * | 2020-12-17 | 2023-12-05 | 杭州鲁尔物联科技有限公司 | Landslide prediction method, device, equipment and storage medium |
CN112561171A (en) * | 2020-12-17 | 2021-03-26 | 杭州鲁尔物联科技有限公司 | Landslide prediction method, device, equipment and storage medium |
CN113076754A (en) * | 2021-03-23 | 2021-07-06 | 山东师范大学 | False comment detection method and system based on knowledge integration |
CN113065455B (en) * | 2021-03-30 | 2021-09-17 | 中国水利水电科学研究院 | Landslide risk inspection method and system based on deep learning |
CN113065455A (en) * | 2021-03-30 | 2021-07-02 | 中国水利水电科学研究院 | Landslide risk inspection method and system based on deep learning |
CN113069117A (en) * | 2021-04-02 | 2021-07-06 | 中山大学 | Electroencephalogram emotion recognition method and system based on time convolution neural network |
CN116739184A (en) * | 2023-08-08 | 2023-09-12 | 四川川核地质工程有限公司 | Landslide prediction method and system |
CN116739184B (en) * | 2023-08-08 | 2023-11-07 | 四川川核地质工程有限公司 | Landslide prediction method and system |
CN116993803A (en) * | 2023-09-26 | 2023-11-03 | 南方科技大学 | Landslide deformation monitoring method and device and electronic equipment |
CN116993803B (en) * | 2023-09-26 | 2024-01-19 | 南方科技大学 | Landslide deformation monitoring method and device and electronic equipment |
CN117196353A (en) * | 2023-11-07 | 2023-12-08 | 山东省济宁生态环境监测中心(山东省南四湖东平湖流域生态环境监测中心) | Environmental pollution assessment and monitoring method and system based on big data |
CN117196353B (en) * | 2023-11-07 | 2024-02-27 | 山东省济宁生态环境监测中心(山东省南四湖东平湖流域生态环境监测中心) | Environmental pollution assessment and monitoring method and system based on big data |
Also Published As
Publication number | Publication date |
---|---|
CN111027686B (en) | 2023-06-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111027686B (en) | Landslide displacement prediction method, device and equipment | |
CN107808122B (en) | Target tracking method and device | |
CN108549841A (en) | A kind of recognition methods of the Falls Among Old People behavior based on deep learning | |
CN110686633B (en) | Landslide displacement prediction method and device and electronic equipment | |
CN110309537B (en) | Intelligent health prediction method and system for aircraft | |
Ayodeji et al. | Causal augmented ConvNet: A temporal memory dilated convolution model for long-sequence time series prediction | |
CN110647974A (en) | Network layer operation method and device in deep neural network | |
CN111832228B (en) | Vibration transmission system based on CNN-LSTM | |
CN115587666A (en) | Load prediction method and system based on seasonal trend decomposition and hybrid neural network | |
CN114004383A (en) | Training method of time series prediction model, time series prediction method and device | |
Sadr et al. | An anomaly detection method for satellites using Monte Carlo dropout | |
CN113253336B (en) | Earthquake prediction method and system based on deep learning | |
Pandhiani et al. | Time series forecasting by using hybrid models for monthly streamflow data | |
CN111158918B (en) | Supporting point parallel enumeration load balancing method, device, equipment and medium | |
KR20220145007A (en) | Data Processing Method of Detecting and Recovering Missing Values, Outliers and Patterns in Tensor Stream Data | |
CN115346125B (en) | Target detection method based on deep learning | |
CN110222840B (en) | Cluster resource prediction method and device based on attention mechanism | |
CN113688770B (en) | Method and device for supplementing long-term wind pressure missing data of high-rise building | |
CN115830865A (en) | Vehicle flow prediction method and device based on adaptive hypergraph convolution neural network | |
CN115510748A (en) | Landslide displacement prediction method based on variational modal decomposition and CNN-GRU | |
CN112242959B (en) | Micro-service current-limiting control method, device, equipment and computer storage medium | |
CN117493837B (en) | Machine learning-based shield tunneling machine attitude item prediction method | |
CN117556261B (en) | MCNN-based diaphragm pump check valve service life prediction method and system | |
Vishnevskaya et al. | Comparison of the applicability of synergistic models with dense neural networks on the example of mobile device security | |
CN117131465B (en) | Single-pipe tower damage identification method and device, electronic equipment and 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 | ||
CB02 | Change of applicant information |
Address after: Room 206-002, 2 / F, building 8, Xixi bafangcheng, Wuchang Street, Yuhang District, Hangzhou City, Zhejiang Province Applicant after: HANGZHOU RUHR TECHNOLOGY Co.,Ltd. Address before: A4-4-201, No. 643, Shuangliu, Zhuantang street, Xihu District, Hangzhou City, Zhejiang Province Applicant before: HANGZHOU RUHR TECHNOLOGY Co.,Ltd. |
|
CB02 | Change of applicant information | ||
GR01 | Patent grant | ||
GR01 | Patent grant |