Slope deformation displacement model prediction method and system considering environmental characteristic factors

Abstract

The invention discloses a slope deformation displacement model prediction method and system considering environmental characteristic factors. The method comprises the following steps: S1, constructing a slope deformation historical sample set of a research area; s2, processing the convolutional neural network model to obtain a feature vector; s3, carrying out model learning training on the LSTM model according to a time sequence; s4, collecting future slope environment monitoring data of a research area according to a time sequence, inputting the data into the convolutional neural network model to obtain a feature vector, and outputting a slope displacement preliminary predicted value # imgabs0 # according to the time sequence by the LSTM model based on the feature vector; S5, constructing a linear regression model, fitting to obtain a fitted slope deformation displacement amount yt, and subtracting the fitted slope deformation displacement amount yt from the slope displacement preliminary predicted value # imgabs1 # to obtain a residual sequence; filling the rejected data by adopting an average value of the slope displacement preliminary predicted values before and after the sliding window is taken out; and S6, carrying out optimization fitting on an SG filter according to a time sequence. According to the invention, high-precision slope displacement prediction data can be obtained, and technical support is provided for accurate prediction of slope deformation displacement.

Description

Technical Field

[0001] The present invention relates to the field of slope deformation monitoring, and in particular to a slope deformation displacement model prediction method and system taking environmental characteristic factors into consideration. Background Art

[0002] The prediction of slope SAR displacement can identify potential unstable areas of the slope in advance, help decision makers take measures such as reinforcement and drainage in time, and avoid the occurrence of sudden geological disasters. Especially in high-risk areas such as open-pit mining areas and infrastructure construction areas, the prediction of slope SAR displacement is more critical; however, the previous slope SAR displacement prediction technology has many shortcomings. In the past, the slope SAR displacement prediction method was mainly based on single factor analysis, ignoring the complexity of multi-factor interaction, resulting in low accuracy of the prediction results. In addition, although many traditional machine learning methods have begun to be applied to slope SAR displacement prediction, they usually adopt a single-layer machine learning model and do not consider the characteristic differences between different influencing factors, resulting in the influence of certain key environmental factors being underestimated or ignored, thereby affecting the prediction accuracy. At the same time, the previous slope SAR displacement prediction did not have further optimization prediction, lacked effective prediction refinement measures, and could not effectively detect and remove outliers in the preliminary prediction, which made the prediction results often interfered with by data quality, affecting the stability and reliability of the model. Summary of the invention

[0003] The purpose of the present invention is to solve the technical problems pointed out by the background technology, and to provide a slope deformation displacement model prediction method and system taking into account environmental characteristic factors. The convolutional neural network model is used to extract and identify the features of the slope environmental monitoring data and the long-term dependency feature extraction of the time series data is fused using the long short-term memory network. The data results of the LSTM model and the linear regression model are used to further obtain the degree of deviation, and the outliers are judged by the degree of deviation to achieve the removal and filling of the outliers, which is conducive to eliminating the interference of the outliers on the prediction results. Finally, the initial prediction value of the slope displacement after filling is optimized by the SG filter according to the time series curve fitting, and the slope displacement prediction data with high accuracy can be obtained.

[0004] The purpose of the present invention is achieved through the following technical solutions:

[0005] A slope deformation displacement model prediction method taking into account environmental characteristic factors, the method comprising:

[0006] S1. Construct a slope deformation history sample set in the study area that includes slope environmental monitoring data, slope deformation displacement, and is associated in time series;

[0007] S2, the convolutional neural network model extracts the feature map of environmental characteristic factors from the slope environmental monitoring data of the slope deformation history sample set and flattens it into a feature vector through a fully connected layer;

[0008] S3, LSTM model performs model learning and training on feature vectors and slope deformation displacement according to time series;

[0009] S4. Collect the future slope environmental monitoring data of the study area in time series and input it into the convolutional neural network model to obtain the feature vector. The LSTM model outputs the preliminary prediction value of the slope displacement in time series based on the feature vector.

[0010] S5. Construct a linear regression model with time as the independent variable and slope deformation displacement as the dependent variable to perform regression fitting to obtain the fitted slope deformation displacement y t , the fitting slope deformation displacement y t Preliminary prediction of slope displacement Make a difference to get the residual sequence and calculate the deviation degree Z of the residual sequence t , the deviation degree Z t The preliminary predicted values ​​of slope displacement with absolute values ​​greater than 2 are eliminated, and the eliminated data are filled with the average value of the preliminary predicted values ​​of slope displacement before and after the sliding window is taken out;

[0011] S6. The initial predicted value of the slope displacement after filling is subjected to curve optimization fitting in time series using the SG filter to obtain the slope displacement prediction curve and slope displacement prediction data after optimization fitting.

[0012] In order to better implement the present invention, the linear regression fitting expression of the linear regression model is as follows:

[0013] y t =β 0 +β 1 t+ε t , where y t is the fitted slope deformation displacement of linear regression, t is the continuous time of the study area including history and future, β 0 , β 1 is the regression coefficient, ε t is the residual term;

[0014] The degree of deviation Z t The expression is as follows:

[0015] Where Z t represents the deviation at time t, μ represents the mean of the residual sequence, σ represents the standard deviation of the residual sequence, and r t represents the residual at time t in the residual sequence.

[0016] Preferably, the LSTM model includes a forget gate, an input gate, a unit state and an output gate. The forget gate outputs a real number vector between 0 and 1 through a sigmoid function, and the expression is as follows:

[0017] f=σ(W*[h t-1 , x t ]+b), where σ is the activation function, W and b represent the weight and bias terms respectively, [h t-1 , x t ] represents the concatenation of the unit state of the t-1th layer and the input data of the tth layer, h t-1 represents the unit state of the t-1th layer, that is, the hidden state of the t-1th layer;

[0018] The input gate expression is as follows:

[0019] i t =σ(W xi x t +W hi h t-1 +b i

[0020]

[0021] x t is the input data, W hi , W xi , W xg , W hg Respectively represent weights, b i 、b g Respectively represent the bias value, tanh is the activation function, i t is the information of the incoming unit status, To update the content; the output gate of the LSTM model outputs the preliminary predicted value of the slope displacement The expression is as follows:

[0022] Where W OUT is the weight of the output layer, b OUT is the bias term, h t represents the hidden state of the tth layer.

[0023] Preferably, the slope environment monitoring data includes surface environment temperature data, rainfall data, atmospheric refractive index data, atmospheric pressure data, and atmospheric relative humidity data; the environmental characteristic factors include surface environment temperature factor, rainfall factor, atmospheric refractive index factor, atmospheric pressure factor, and atmospheric relative humidity factor; the slope environment monitoring data of the slope deformation history sample set and the future slope environment monitoring data of the study area are both normalized as follows:

[0024] P is the characteristic data of slope environmental monitoring data, P min is the minimum value of the characteristic data of the slope environmental monitoring data, P max is the maximum value of the characteristic data of the slope environment monitoring data, and P is the normalized value of the characteristic data of the slope environment monitoring data.

[0025] Preferably, the method for filling the preliminary predicted value of slope displacement after elimination is as follows: using a sliding window to take out the preliminary predicted values ​​of slope displacement before and after the position of the preliminary predicted value of slope displacement eliminated and calculate the average value, and filling the average value to the position of the preliminary predicted value of slope displacement eliminated.

[0026] Preferably, in step S2, the convolutional neural network model outputs a feature E graph Z through a convolution layer, where Z=W*A+B, wherein W represents a convolution kernel, which is used to slide the input data A to extract the features of all environmental characteristic factors; in the convolution operation, the convolution kernel W performs element-wise multiplication and summation operation on the input data A, and B is a bias term; the input data A is slope environmental monitoring data, and the input data A sequence matrix expression is as follows: Among them A n，m represents the mth feature of the nth sample, N is the total number of samples, and M is the total number of features; the feature map Z is pooled to obtain the feature map Z′, which is the feature map of the environmental characteristic factor.

[0027] Preferably, the SG filter uses a sliding window to perform optimization fitting using a polynomial least squares method, and the expression is as follows:

[0028] where y 最终 represents the slope displacement prediction data after SG filtering, represents the initial predicted value of slope displacement after filling, c λ represents the coefficients obtained by least squares fitting of the polynomial, λ 1 represents the data point on the side of the sliding window boundary centered on the sliding window, λ 2 Represents the data points on the other side of the sliding window boundary centered on the sliding window.

[0029] Preferably, the residual expression at time t in the residual sequence is as follows:

[0030] A slope deformation displacement model prediction system taking into account environmental characteristic factors includes a slope deformation history sample set, a data acquisition module, a convolutional neural network model, an LSTM model and a linear regression model. The slope deformation history sample set includes slope environmental monitoring data and slope deformation displacement associated in time series. The convolutional neural network model extracts a feature map of environmental characteristic factors from the slope environmental monitoring data of the slope deformation history sample set and flattens it into a feature vector through a fully connected layer. The LSTM model performs model learning and training on the feature vector and the slope deformation displacement in time series. The data acquisition module is used to collect the future slope environmental monitoring data in time series of the study area and input it into the convolutional neural network model to obtain a feature vector. The LSTM model outputs a preliminary prediction value of the slope displacement in time series based on the feature vector. The linear regression model is constructed with time as the independent variable and slope deformation displacement as the dependent variable, and regression fitting is performed to obtain the fitted slope deformation displacement y t , the fitting slope deformation displacement y t Preliminary prediction of slope displacement Make a difference to get the residual sequence and calculate the deviation degree Z of the residual sequence t , the deviation degree Z t The preliminary predicted values ​​of slope displacement with absolute values ​​greater than 2 are eliminated, and the eliminated data are filled with the average value of the preliminary predicted values ​​of slope displacement before and after the sliding window is taken out; the linear regression model includes an SG filter, and the linear regression model uses the SG filter to perform curve optimization fitting on the preliminary predicted values ​​of slope displacement after filling in time series to obtain the slope displacement prediction curve and slope displacement prediction data after optimized fitting.

[0031] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0032] (1) The present invention uses a convolutional neural network model to extract and identify features of slope environmental monitoring data and uses a long short-term memory network to fuse the feature extraction of long-term dependence of time series data. The data results of the LSTM model and the linear regression model are subtracted to further obtain the degree of deviation, and the outliers are judged by the degree of deviation to achieve the removal and filling of outliers, which is conducive to eliminating the interference of outliers on the prediction results. Finally, the initial predicted value of the slope displacement after filling is optimized and fitted by the SG filter according to the time series curve, which can obtain high-precision slope displacement prediction data, which not only effectively enhances the stability and accuracy of the model, but also provides a solid technical guarantee for the accurate identification of the trend of slope deformation displacement. Its application in the field of slope monitoring has important practical value and forward-looking significance.

[0033] (2) The present invention fully considers multiple factors such as environmental characteristic factors and time series characteristics, extracts comprehensive features through a convolutional neural network model, and can accurately predict the preliminary prediction data of slope displacement through the learning and training of the LSTM model. It also adopts rigorous outlier processing and missing value filling processing, and then introduces SG filtering to optimize the data. The present invention adopts a refined and rigorous slope deformation displacement model prediction method, which significantly improves the data quality of the slope deformation displacement time series, thereby greatly improving the accuracy and stability of the prediction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a structural schematic diagram of the present invention;

[0035] Figure 2 It is a schematic diagram of the process principle of the embodiment;

[0036] Figure 3 It is a schematic diagram of a time series of the slope deformation history sample set in the embodiment according to the time sequence correlation between the environmental characteristic factors and the slope deformation displacement;

[0037] Figure 4 The preliminary prediction data of slope displacement in the study area including future time series and in time series in the embodiment;

[0038] Figure 5 The slope displacement optimization prediction data after outlier removal and sliding filling in the embodiment;

[0039] Figure 6 It is the slope displacement prediction data after filtering, fusion and optimization fitting in the embodiment. DETAILED DESCRIPTION

[0040] The present invention is further described in detail below in conjunction with embodiments:

[0041] Example

[0042] like Figure 1 As shown, a slope deformation displacement model prediction method taking into account environmental characteristic factors includes:

[0043] S1. Construct a slope deformation history sample set in the study area that includes slope environmental monitoring data, slope deformation displacement, and is associated in time series. Figure 2 As shown, the slope environmental monitoring data in the slope deformation history sample set of this embodiment includes surface environmental temperature data, rainfall data, atmospheric refractive index data, atmospheric pressure data, and atmospheric relative humidity data. The slope environmental monitoring data in the slope deformation history sample set is normalized to eliminate the scale differences between data of different dimensions, which is convenient for the subsequent extraction and identification of environmental characteristic factors.

[0044] For example, the study area takes an open-pit coal mine in Xinjiang as an example. The monitoring radar is IBIS360ArcSAR (used to detect the deformation and displacement of the slope at the measuring point). The slope deformation history sample set obtains 300 hours of historical data hourly. In order to better illustrate the principle, the environmental characteristic factors in the model include surface environmental temperature factor, rainfall factor, atmospheric refractive index factor, atmospheric pressure factor, and atmospheric relative humidity factor. The slope environmental monitoring data can be multi-source data. The convolutional neural network model is based on the slope environmental monitoring data and obtains the time series data according to the environmental characteristic factors and the slope deformation and displacement according to the time series (the total time series is 300 hours), such as Figure 3 shown.

[0045] S2, the convolutional neural network model extracts the feature map of environmental characteristic factors from the slope environmental monitoring data of the slope deformation history sample set and flattens it into a feature vector through the fully connected layer. Figure 2 As shown in the figure, the features that are associated with the slope deformation and displacement and are of high importance (such as environmental characteristic factors such as surface environmental temperature factor, rainfall factor, atmospheric refractive index factor, atmospheric pressure factor, atmospheric relative humidity factor, etc.) are extracted from the slope environmental monitoring data. These features are flattened into feature vectors in time series according to their association with the slope deformation and displacement, and model training is performed to facilitate the subsequent LSTM model to output the preliminary predicted value of the slope displacement in time series based on the feature vector.

[0046] In some embodiments, the convolutional neural network model outputs a feature map Z through a convolution layer, where Z=W*A+B, where W represents a convolution kernel, which is used to slide the input data A to extract the features of all environmental characteristic factors. In the convolution operation (the convolution can be set to 3x3), the convolution kernel W performs element-wise multiplication and summation operations on the input data A, and B is a bias term. The input data A is slope environmental monitoring data, and the input data A sequence matrix expression is as follows: Among them A n，m It represents the mth feature of the nth sample, N is the total number of samples (for example, N slope environmental monitoring data and N slope deformation displacements are collected hourly in the study area in time series, N is the total amount of sample data collected hourly, in the case of a Xinjiang open-pit coal mine selected in the study area, N is 300), M is the total number of features (including the total number of features such as surface environmental temperature factor, rainfall factor, atmospheric refractive index factor, atmospheric pressure factor, atmospheric relative humidity factor, etc.). Feature map Z (after the convolution operation, the generated feature map Z contains a large amount of feature information, which significantly increases the data dimension and is conducive to the significant improvement of feature extraction in time series) is obtained after pooling processing (reducing the size of the feature map and retaining the most critical information related to the prediction task, which can effectively reduce the amount of calculation and improve the processing efficiency). Feature map Z′ is the feature map of the environmental characteristic factor.

[0047] S3. LSTM model performs model learning and training on feature vectors and slope deformation displacement in time series.

[0048] S4. Collect the slope environmental monitoring data of the study area in the future in time series (in the case of an open-pit coal mine in Xinjiang, the previous sample size is 300, and the future time series is 200, so the total number of historical samples and future time series is 500). Input the convolutional neural network model to obtain the feature vector. The LSTM model outputs the preliminary prediction value of the slope displacement in time series based on the feature vector.

[0049] In some embodiments, the slope environmental monitoring data includes surface environmental temperature data, rainfall data, atmospheric refractive index data, atmospheric pressure data, and atmospheric relative humidity data, and the environmental characteristic factors include surface environmental temperature factor, rainfall factor, atmospheric refractive index factor, atmospheric pressure factor, and atmospheric relative humidity factor. The slope environmental monitoring data of the slope deformation history sample set and the future slope environmental monitoring data of the study area are both normalized as follows:

[0050] P is the characteristic data of slope environmental monitoring data, P min is the minimum value of the characteristic data of the slope environmental monitoring data, P max is the maximum value of the characteristic data of the slope environment monitoring data, and P′ is the normalized value of the characteristic data of the slope environment monitoring data. Taking the surface environment temperature data as an example, if the surface environment temperature is 27°C, the maximum value of the surface temperature characteristic is 33°C, and the minimum value of the surface temperature characteristic is 17°C, then the surface temperature 27°C is normalized to:

[0051] In this embodiment, the LSTM model includes a forget gate, an input gate, a unit state (or an update memory unit), and an output gate. The forget gate outputs a real number vector between 0 and 1 through a sigmoid function, and the expression is as follows:

[0052] f=σ(W*[h t-1 , x t ]+b), where σ is the activation function, W and b represent the weight and bias terms respectively, [h t-1 , x t ] represents the concatenation of the unit state of the t-1th layer and the input data of the tth layer, h t-1 It represents the state of the t-1th layer unit, that is, the hidden state of the t-1th layer.

[0053] The input gate expression is as follows:

[0054] i t =σ(W xi x t +Whi h t-1 +b i

[0055]

[0056] x t is the input data, W hi , W xi , W xg , W hg Respectively represent weights, b i 、b g Respectively represent the bias value, tanh is the activation function, i t is the information of the incoming unit status, The output gate of the LSTM model outputs the preliminary predicted value of the slope displacement. The expression is as follows:

[0057] Where W OUT is the weight of the output layer, b OUT is the bias term, h t Represents the hidden state of the tth layer. Update the memory unit, which is combined with the forget gate f t and input gate i t The new memory cell state combines the historical information controlled by the forget gate and the current information controlled by the input gate, which ultimately affects the prediction of the slope radar displacement time series. t =f t *C t-1 +i t *C t , the output gate determines the final information extracted from the memory cell as the current hidden state h t Output

[0058] o t =σ(W o *[h t-1 , F t ]+b o )

[0059] Among them, t is the activation of the input gate, W o is the weight matrix of the input gate, b o is the bias term.

[0060] Calculate the hidden state, the final hidden state h t It is determined by the state of the memory unit and the control of the output gate. t =o t *tanh(C t ).

[0061] The LSTM model is first trained using the slope deformation history sample set to obtain the trained LSTM model. The trained LSTM model can obtain the feature vector prediction output of the preliminary predicted value of the slope displacement (in time series) based on the future environmental characteristic factors of the study area in time series. In the case of an open-pit coal mine in Xinjiang selected in the study area, the slope deformation history sample set (containing 300 hourly data, of which 250 are used as training sets and 50 are used as prediction verification sets) is used to predict the slope displacement prediction data for the next 200 hours, and the preliminary predicted data of the slope displacement containing the future time series and in time series is obtained, such as Figure 4 shown.

[0062] S5. Construct a linear regression model with time as the independent variable and slope deformation displacement as the dependent variable to perform regression fitting to obtain the fitted slope deformation displacement y t In some embodiments, the linear regression fitting expression of the linear regression model is as follows:

[0063] y t =β 0 +β 1 t+ε t , where y t is the fitted slope deformation displacement of linear regression, t is the continuous time of the study area including history and future, β 0 , β 1 is the regression coefficient, ε t is the residual term.

[0064] The fitted slope deformation displacement y t Preliminary prediction of slope displacement The residual sequence is obtained by difference. The residual expression at time t in the residual sequence is as follows: Calculate the degree of deviation Z of the residual sequence t In some embodiments, the deviation degree Z t The expression is as follows:

[0065] Where Z t represents the deviation at time t, μ represents the mean of the residual sequence, σ represents the standard deviation of the residual sequence, and r t represents the residual at time t in the residual sequence.

[0066] The deviation degree Z t Preliminary predicted slope displacement values ​​with absolute values ​​greater than 2 are eliminated (deviation degree Z t The absolute value is greater than 2, which is the deviation degree Z tThe preliminary predicted values ​​of slope displacement that are greater than 2 or less than -2, and whose absolute value of the degree of deviation is greater than 2, are determined to be outliers, and the outliers are first eliminated, and the eliminated data bits are missing values), and the average value of the preliminary predicted values ​​of slope displacement before and after the sliding window is taken out to fill the eliminated data. In some embodiments, the method of filling the preliminary predicted values ​​of slope displacement after elimination is as follows: the preliminary predicted values ​​of slope displacement before and after the position of the eliminated preliminary predicted values ​​of slope displacement are taken out using a sliding window and the average value is calculated, and the average value is filled to the position of the eliminated preliminary predicted values ​​of slope displacement (that is, the missing values ​​are filled). In the case of an open-pit coal mine in Xinjiang selected in the study area, the slope deformation history sample set (containing 300 hourly data, of which 250 are used as training sets and 50 are used as prediction verification sets) is used to predict the slope displacement prediction data for the next 200 hours, and the optimized prediction data of slope displacement after the elimination of outliers and sliding filling of future time series is obtained, such as Figure 5 shown.

[0067] S6. The initial predicted value of the slope displacement after filling is optimized and fitted by the SG filter in time series to obtain the slope displacement prediction curve and slope displacement prediction data after optimization fitting. In the case of an open-pit coal mine in Xinjiang selected in the study area, the slope deformation history sample set (containing 300 hourly data, of which 250 are used as training sets and 50 are used as prediction verification sets) is used to predict the slope displacement prediction data for the next 200 hours, and the final slope displacement prediction data containing the future time series and in time series is obtained, such as Figure 6 shown.

[0068] In some embodiments, the SG filter of the present invention uses a sliding window to optimize the fitting using a polynomial least squares method, and the expression is as follows:

[0069] where y 最终 represents the slope displacement prediction data after SG filtering (weighted sum of weighted coefficients of data points in the sliding window), represents the initial predicted value of slope displacement after filling, c λ represents the coefficient obtained by least squares fitting of the polynomial (i.e., the weight of the target point smoothing), λ 1 represents the data point on the side of the sliding window boundary centered on the sliding window, λ 2 represents the data point on the other side of the sliding window boundary centered on the sliding window, λ 1 , 2It is the position of the left and right endpoints of the data in the sliding window. The present invention uses a convolutional neural network model to perform feature extraction and identification on the slope environmental monitoring data (extract and identify environmental influencing factors) and uses a long short-term memory network to fuse the feature extraction of the long-term dependence of the time series data. The data results of the LSTM model and the linear regression model are subtracted to further obtain the degree of deviation, and the outliers are judged by the degree of deviation to achieve the removal and filling of the outliers, which is conducive to eliminating the interference of outliers on the prediction results. Finally, the initial predicted value of the slope displacement after filling is optimized and fitted by the SG filter according to the time series, and the slope displacement prediction data with high accuracy can be obtained, which not only effectively enhances the stability and accuracy of the model, but also provides a solid technical guarantee for the accurate identification of the trend of slope deformation displacement. Its application in the field of slope monitoring has important practical value and forward-looking significance.

[0070] A slope deformation displacement model prediction system that takes into account environmental characteristic factors includes a slope deformation history sample set, a data acquisition module, a convolutional neural network model, an LSTM model and a linear regression model. The slope deformation history sample set includes slope environmental monitoring data and slope deformation displacement associated in time series. The convolutional neural network model extracts the characteristic graph of environmental characteristic factors from the slope environmental monitoring data of the slope deformation history sample set and flattens it into a characteristic vector through a fully connected layer. The LSTM model performs model learning and training on the characteristic vector and slope deformation displacement in time series. The data acquisition module is used to collect the future slope environmental monitoring data in the study area in time series and input it into the convolutional neural network model to obtain a characteristic vector. The LSTM model outputs a preliminary prediction value of the slope displacement in time series based on the characteristic vector. The linear regression model is constructed with time as the independent variable and slope deformation displacement as the dependent variable, and regression fitting is performed to obtain the fitted slope deformation displacement y t , the fitting slope deformation displacement y t Preliminary prediction of slope displacement Make a difference to get the residual sequence and calculate the deviation degree Z of the residual sequence t , the deviation degree Z t The initial predicted values ​​of slope displacement with an absolute value greater than 2 are eliminated, and the average value of the initial predicted values ​​of slope displacement before and after the sliding window is taken out is used to fill the eliminated data. The linear regression model includes an SG filter. The linear regression model uses the SG filter to perform curve optimization fitting on the initial predicted values ​​of slope displacement after filling in time series to obtain the slope displacement prediction curve and slope displacement prediction data after optimization fitting.

[0071] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A slope deformation displacement model prediction method taking into account environmental characteristic factors, characterized by: The methods include: S1. Construct a slope deformation history sample set in the study area that includes slope environmental monitoring data, slope deformation displacement, and is associated in time series; S2, the convolutional neural network model extracts the feature map of environmental characteristic factors from the slope environmental monitoring data of the slope deformation history sample set and flattens it into a feature vector through a fully connected layer; S3, LSTM model performs model learning and training on feature vectors and slope deformation displacement according to time series; S4. Collect the future slope environmental monitoring data of the study area in time series and input it into the convolutional neural network model to obtain the feature vector. The LSTM model outputs the preliminary prediction value of the slope displacement in time series based on the feature vector. S5. Construct a linear regression model with time as the independent variable and slope deformation displacement as the dependent variable to perform regression fitting to obtain the fitted slope deformation displacement y t , the fitting slope deformation displacement y t Preliminary prediction of slope displacement Make a difference to get the residual sequence and calculate the deviation degree Z of the residual sequence t , the deviation degree Z t The preliminary predicted values ​​of slope displacement with absolute values ​​greater than 2 are eliminated, and the eliminated data are filled with the average value of the preliminary predicted values ​​of slope displacement before and after the sliding window is taken out; S6. The initial predicted value of the slope displacement after filling is subjected to curve optimization fitting in time series using the SG filter to obtain the slope displacement prediction curve and slope displacement prediction data after optimization fitting.

2. The slope deformation displacement model prediction method taking into account environmental characteristic factors according to claim 1 is characterized by: The linear regression fitting expression of the linear regression model is as follows: y t =β0+β1t+ε t , where y t is the fitted slope deformation displacement of linear regression, t is the continuous time of the study area including history and future, β0 and β1 are regression coefficients, ε t is the residual term; The degree of deviation Z t The expression is as follows: Where Z t represents the deviation at time t, μ represents the mean of the residual sequence, σ represents the standard deviation of the residual sequence, and r t represents the residual at time t in the residual sequence.

3. The slope deformation displacement model prediction method taking into account environmental characteristic factors according to claim 1 is characterized by: The LSTM model includes a forget gate, an input gate, a cell state, and an output gate. The forget gate outputs a real number vector between 0 and 1 through a sigmoid function. The expression is as follows: f=σ(W*[h t-1 , x t ]+b), where σ is the activation function, W and b represent the weight and bias terms respectively. [h t-1 , x t ] represents the concatenation of the unit state of the t-1th layer and the input data of the tth layer, h t-1 represents the unit state of the t-1th layer, that is, the hidden state of the t-1th layer; The input gate expression is as follows: i t =σ(W xi x t +W hi h t-1 +b i x t is the input data, W hi , W xi , W xg , W hg Respectively represent weights, b i , b g Respectively represent the bias value, tanh is the activation function, i t is the information of the incoming unit status, To update the content; the output gate of the LSTM model outputs the preliminary predicted value of the slope displacement The expression is as follows: Where W OUT is the weight of the output layer, b OUT is the bias term, h t represents the hidden state of the tth layer.

4. The slope deformation displacement model prediction method taking into account environmental characteristic factors according to claim 1 is characterized by: The slope environmental monitoring data include surface environmental temperature data, rainfall data, atmospheric refractive index data, atmospheric pressure data, and atmospheric relative humidity data; the environmental characteristic factors include surface environmental temperature factor, rainfall factor, atmospheric refractive index factor, atmospheric pressure factor, and atmospheric relative humidity factor; the slope environmental monitoring data of the slope deformation history sample set and the future slope environmental monitoring data of the study area are both normalized as follows: P is the characteristic data of slope environmental monitoring data, P min is the minimum value P of the characteristic data of the slope environment monitoring data max The maximum value P′ of the characteristic data of the slope environment monitoring data is the normalized value of the characteristic data of the slope environment monitoring data.

5. The slope deformation displacement model prediction method taking into account environmental characteristic factors according to claim 1 is characterized by: The method for filling the preliminary predicted value of slope displacement after elimination is as follows: use a sliding window to take out the preliminary predicted values ​​of slope displacement before and after the position of the preliminary predicted value of slope displacement eliminated and calculate the average value, and fill the average value to the position of the preliminary predicted value of slope displacement eliminated.

6. The slope deformation displacement model prediction method taking into account environmental characteristic factors according to claim 1 is characterized by: In step S2, the convolutional neural network model outputs a feature map Z through a convolutional layer, where Z=W*A+B, where W represents a convolution kernel, which is used to slide the input data A to extract the features of all environmental feature factors; In the convolution operation, the convolution kernel W performs element-wise multiplication and summation operation with the input data A, and B is the bias term; the input data A is the slope environment monitoring data, and the input data A sequence matrix expression is as follows: Among them A n，m represents the mth feature of the nth sample, N is the total number of samples, and M is the total number of features; the feature map Z is pooled to obtain the feature map Z′, which is the feature map of the environmental characteristic factor.

7. The slope deformation displacement model prediction method taking into account environmental characteristic factors according to claim 1 is characterized by: The SG filter uses a sliding window to optimize the fitting using the polynomial least squares method, and the expression is as follows: where y 最终 represents the slope displacement prediction data after SG filtering, represents the initial predicted value of slope displacement after filling, c λ represents the coefficient obtained by least squares fitting the polynomial, λ1 represents the data point on one side of the sliding window boundary centered on the sliding window, and λ2 represents the data point on the other side of the sliding window boundary centered on the sliding window.

8. The slope deformation displacement model prediction method taking into account environmental characteristic factors according to claim 2 is characterized by: The residual expression at time t in the residual sequence is as follows:

9. A slope deformation displacement model prediction system taking into account environmental characteristic factors, characterized by: The invention comprises a slope deformation history sample set, a data acquisition module, a convolutional neural network model, an LSTM model and a linear regression model. The slope deformation history sample set comprises slope environmental monitoring data and slope deformation displacement associated in time series. The convolutional neural network model extracts a characteristic graph of environmental characteristic factors from the slope environmental monitoring data of the slope deformation history sample set and flattens it into a characteristic vector through a fully connected layer. The LSTM model performs model learning and training on the characteristic vector and the slope deformation displacement in time series. The data acquisition module is used to collect the slope environmental monitoring data of the study area in the future in time series and input it into the convolutional neural network model to obtain a characteristic vector. The LSTM model outputs a preliminary prediction value of the slope displacement in time series based on the characteristic vector. The linear regression model is constructed with time as the independent variable and slope deformation displacement as the dependent variable, and regression fitting is performed to obtain the fitted slope deformation displacement y t , the fitting slope deformation displacement y t Preliminary prediction of slope displacement Make a difference to get the residual sequence and calculate the deviation degree Z of the residual sequence t , the deviation degree Z t The preliminary predicted values ​​of slope displacement with absolute values ​​greater than 2 are eliminated, and the eliminated data are filled with the average value of the preliminary predicted values ​​of slope displacement before and after the sliding window is taken out; the linear regression model includes an SG filter, and the linear regression model uses the SG filter to perform curve optimization fitting on the preliminary predicted values ​​of slope displacement after filling in time series to obtain the slope displacement prediction curve and slope displacement prediction data after optimized fitting.

Images