Landslide susceptibility assessment method and device considering factor variation trends

Abstract

The present invention discloses a method and device for assessing landslide susceptibility taking into account the trend of factor changes, wherein the method comprises: based on the scope of the target area to be assessed, clipping the time-series landslide impact factor data in the target area to be assessed, obtaining a time-series landslide impact factor set with consistent scope size and corresponding spatial position; quantifying each factor in the factor set into a trend landslide impact factor; de-dimensionalizing each trend landslide impact factor in the trend landslide impact factor set; performing a multicollinearity test on the trend landslide impact factor set after the de-dimensionalization; inputting the trend landslide impact factor set after the multicollinearity test into a pre-established landslide susceptibility assessment model, and obtaining a landslide susceptibility assessment result of the target area to be assessed. The present invention can fully reflect the temporal and spatial change trend and cumulative effect of the landslide impact factor through the quantified trend landslide impact factor, thereby improving the accuracy of the landslide susceptibility assessment.

Description

Technical Field

[0001] The present invention relates to the technical field of geological disaster risk assessment, and in particular to a landslide susceptibility assessment method and device taking into account factor change trends. Background Art

[0002] This section is intended to provide a background or context to the embodiments of the invention recited in the claims. No admission is made that the description herein is prior art by inclusion in this section.

[0003] Landslide susceptibility assessment refers to a method of quantitatively assessing the area and probability of landslides by analyzing the geological, geomorphological, hydrological, and meteorological factors that affect the occurrence of landslides. Accurate landslide susceptibility assessment can identify potential landslide risk areas and timely warn of landslide disasters. Therefore, the study of landslide susceptibility assessment is of great significance for the scientific formulation of disaster prevention and mitigation strategies and the protection of people's lives and property.

[0004] Landslide impact factors are the basis of landslide susceptibility assessment and determine the results of the assessment. Existing landslide susceptibility assessment methods usually use static landslide impact factors of a specific year or use the difference of factors at different time points as dynamic factors for landslide susceptibility assessment. However, landslide impact factors have a certain trend of change. When this trend reaches a certain level, it leads to the occurrence of landslides. Using static factors cannot reflect the changes in impact factors, and using dynamic factors can only reflect the changes between two time points, and cannot capture the changing trend of landslide impact factors.

[0005] Therefore, the existing landslide susceptibility assessment methods still have shortcomings, which will affect the accuracy of the final landslide susceptibility assessment. Summary of the invention

[0006] The embodiment of the present invention provides a landslide susceptibility assessment method taking into account the change trend of factors, which is used to fully reflect the temporal and spatial change trend and cumulative effect of landslide influencing factors through quantified trend landslide influencing factors, and improve the accuracy of landslide susceptibility assessment. The method includes:

[0007] Obtain the time series landslide impact factor data in the target area to be evaluated;

[0008] Based on the scope of the target area to be evaluated, the time series landslide impact factor data are clipped to obtain a set of time series landslide impact factors with consistent scope and corresponding spatial location;

[0009] Quantify each factor in the time series landslide impact factor set with consistent range and corresponding spatial position as a trend landslide impact factor to obtain the trend landslide impact factor set;

[0010] De-dimensionalize each trend landslide impact factor in the trend landslide impact factor set to obtain the de-dimensionalized trend landslide impact factor set;

[0011] Conduct multicollinearity test on the de-dimensionalized trend landslide influencing factor set to obtain the trend landslide influencing factor set after multicollinearity test;

[0012] The landslide influencing factor set after the multicollinearity test is input into a pre-established landslide susceptibility assessment model to obtain the landslide susceptibility assessment result of the target area to be assessed; the landslide susceptibility assessment model is pre-established based on the relationship sample data between the landslide influencing factor set after the historical multicollinearity test and the landslide susceptibility assessment result.

[0013] The embodiment of the present invention further provides a landslide susceptibility assessment device taking into account the factor change trend, which is used to fully reflect the temporal and spatial change trend and cumulative effect of the landslide influencing factor through quantified trend landslide influencing factor, and improve the accuracy of landslide susceptibility assessment. The device includes:

[0014] An acquisition unit, used for acquiring time series landslide impact factor data in the target area to be evaluated;

[0015] A clipping processing unit is used to clip the time series landslide impact factor data based on the scope of the target area to be evaluated, so as to obtain a set of time series landslide impact factors with consistent scope size and corresponding spatial position;

[0016] A trend quantification processing unit is used to quantify each factor in the time series landslide impact factor set with the same range size and corresponding spatial position into a trend landslide impact factor, so as to obtain a trend landslide impact factor set;

[0017] A dimension removal processing unit is used to perform dimension removal processing on each trend landslide impact factor in the trend landslide impact factor set to obtain the trend landslide impact factor set after dimension removal processing;

[0018] A multicollinearity test unit is used to perform a multicollinearity test on the set of trend landslide influencing factors after dimension removal, and obtain a set of trend landslide influencing factors after the multicollinearity test;

[0019] The evaluation processing unit is used to input the trend landslide influencing factor set after the multicollinearity test into a pre-established landslide susceptibility evaluation model to obtain the landslide susceptibility evaluation result of the target area to be evaluated; the landslide susceptibility evaluation model is pre-established based on the relationship sample data between the trend landslide influencing factor set after the historical multicollinearity test and the landslide susceptibility evaluation result.

[0020] An embodiment of the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned landslide susceptibility assessment method taking into account the factor change trend when executing the computer program.

[0021] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned landslide susceptibility assessment method taking into account the factor change trend.

[0022] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the landslide susceptibility assessment method taking into account the factor change trend is implemented.

[0023] Compared with the prior art landslide susceptibility assessment method which only uses static or dynamic factors to perform landslide susceptibility assessment, which has shortcomings and leads to low accuracy of landslide susceptibility assessment, the landslide susceptibility assessment scheme taking into account the factor change trend provided in the embodiment of the present invention has the beneficial technical effect as follows: the embodiment of the present invention takes into account that the occurrence of landslide is a dynamically changing process, and fully reflects the temporal and spatial change trend and cumulative effect of landslide influencing factors through quantified trend landslide influencing factors, thereby improving the accuracy of landslide susceptibility assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0025] Figure 1 A schematic flow chart of a landslide susceptibility assessment method taking into account factor variation trends in an embodiment of the present invention;

[0026] Figure 2 It is a flow chart of a landslide susceptibility assessment method taking into account the factor variation trend in another embodiment of the present invention;

[0027] Figure 3 It is a flow chart of a method for preprocessing time series landslide impact factor data in an embodiment of the present invention;

[0028] Figure 4 A set of time series landslide impact factor diagrams in an embodiment of the present invention;

[0029] Figure 5A trend landslide impact factor diagram in an embodiment of the present invention;

[0030] Figure 6 A set of trend landslide impact factor diagrams in an embodiment of the present invention;

[0031] Figure 7 It is a flow chart of a method for quantifying the influence factor of a trend landslide in an embodiment of the present invention;

[0032] Figure 8 is a static landslide impact factor diagram in an embodiment of the present invention;

[0033] Fig. 9 Schematic diagram of the process of dimension removal processing in an embodiment of the present invention;

[0034] Fig.10 A discrete landslide impact factor diagram in an embodiment of the present invention;

[0035] Fig.11 It is an interval result diagram after dimension removal of continuous landslide impact in an embodiment of the present invention;

[0036] Fig.12 Schematic diagram of the process of multicollinearity testing method in an embodiment of the present invention;

[0037] Fig.13 This is a graph showing the multicollinearity test results of landslide influencing factors in an embodiment of the present invention;

[0038] Fig.14 This is a graph showing the collinearity test results of rainfall and slope in an embodiment of the present invention;

[0039] Fig.15 This is a landslide susceptibility assessment result diagram in an embodiment of the present invention;

[0040] Fig.16 This is a landslide susceptibility result accuracy assessment diagram in an embodiment of the present invention;

[0041] Fig.17 The result diagram of the proportion of landslide points at different levels and the proportion of the area at the corresponding level in the embodiment of the present invention;

[0042] Fig.18 A confusion matrix diagram in an embodiment of the present invention;

[0043] Fig.19 Schematic diagram of the structure of a landslide susceptibility assessment device taking into account factor variation trends in an embodiment of the present invention. DETAILED DESCRIPTION

[0044] To make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0045] The acquisition, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of laws and regulations.

[0046] The existing landslide susceptibility assessment is usually a static landslide impact factor of a specific year or uses the difference of factors at different time points as a dynamic factor to conduct landslide susceptibility assessment. However, the inventor has discovered a technical problem. The occurrence of landslides is a dynamic process and is a geological disaster caused by the cumulative effect of impact factors over time. That is, the existing landslide susceptibility assessment ignores that the natural environment and human activities in the area where the landslide is located are constantly changing. The landslide impact factor has a certain trend of change. When this change accumulates to a certain extent, it leads to the occurrence of landslides. There are deficiencies in using only static or dynamic factors to conduct landslide susceptibility assessment. When evaluating landslide susceptibility, the use of static factors and dynamic factors cannot reflect the temporal and spatial change trend and cumulative effect of the landslide impact factor, which will affect the accuracy and credibility of the susceptibility assessment results. Therefore, the inventor proposes a landslide susceptibility assessment scheme that takes into account the trend of factor changes. The scheme solves the above problem by quantifying the temporal and spatial change trend and cumulative effect of the landslide impact factor. The following is a detailed introduction to the landslide susceptibility assessment scheme that takes into account the trend of factor changes.

[0047] Figure 1 FIG. 4 is a flow chart of a landslide susceptibility assessment method taking into account the change trend of factors in an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0048] Step 101: Acquire time series landslide impact factor data in the target area to be evaluated;

[0049] Step 102: based on the scope of the target area to be evaluated, the time series landslide impact factor data is clipped to obtain a time series landslide impact factor set with consistent scope size and corresponding spatial position;

[0050] Step 103: quantify each factor in the time series landslide impact factor set with the same range size and corresponding spatial position into a trend landslide impact factor to obtain a trend landslide impact factor set;

[0051] Step 104: performing dimension-removing processing on each trend landslide impact factor in the trend landslide impact factor set to obtain a trend landslide impact factor set after dimension-removing processing;

[0052] Step 105: performing a multicollinearity test on the de-dimensionalized trend landslide influencing factor set to obtain a trend landslide influencing factor set after the multicollinearity test;

[0053] Step 106: Input the set of trend landslide influencing factors after the multicollinearity test into a pre-established landslide susceptibility assessment model to obtain a landslide susceptibility assessment result of the target area to be assessed; the landslide susceptibility assessment model is pre-established based on sample data of the relationship between the set of trend landslide influencing factors after the historical multicollinearity test and the landslide susceptibility assessment result.

[0054] Compared with the prior art landslide susceptibility assessment method that only uses static or dynamic factors to perform landslide susceptibility assessment, which has shortcomings and leads to low accuracy of landslide susceptibility assessment, the landslide susceptibility assessment scheme that takes into account the factor change trend provided in the embodiment of the present invention has the following beneficial technical effects: the embodiment of the present invention takes into account that the occurrence of landslides is a dynamic process, and fully reflects the temporal and spatial change trend and cumulative effect of landslide influence factors through quantified trend landslide influence factors, thereby improving the accuracy of landslide susceptibility assessment. The landslide susceptibility assessment method that takes into account the factor change trend is introduced in detail below.

[0055] The existing landslide susceptibility assessment ignores the fact that the natural environment and human activities in the area where the landslide occurs are constantly changing. The landslide influencing factors have a certain trend of change. When this change accumulates to a certain extent, it leads to the occurrence of landslides. The use of static factors and dynamic factors in landslide susceptibility assessment cannot reflect the temporal and spatial change trend and cumulative effect of landslide influencing factors, which will affect the accuracy and credibility of the susceptibility assessment results. Based on this, the embodiment of the present invention provides a landslide susceptibility assessment method that takes into account the trend of factor changes. Figures 2 to 14 Detailed description.

[0056] The problem to be solved by the embodiments of the present invention is that in the geological disaster risk assessment, the static landslide impact factor of a specific year or the difference of the factor at different time points is used as a dynamic factor to evaluate the landslide susceptibility. However, the occurrence of landslides is a dynamic process and is a geological disaster caused by the cumulative effect of the impact factor over time. Only using static or dynamic factors cannot reflect the temporal and spatial variation trend and cumulative effect of the landslide impact factor, which affects the accuracy and credibility of the landslide susceptibility assessment. Therefore, a landslide susceptibility assessment method that takes into account the factor change trend is provided. Figure 2 This is a flow chart of a landslide susceptibility assessment method taking into account factor variation trends in another embodiment of the present invention, which specifically includes the following steps S1 to S9:

[0057] Step S1: Obtain the time series landslide impact factor data of the study area (target area to be evaluated), and preprocess it to obtain the grid time series landslide impact factor.

[0058] In specific implementation, when establishing a landslide susceptibility assessment model, what is obtained is the time series landslide impact factor data of the historical study area. When using the landslide susceptibility assessment model to conduct actual landslide susceptibility assessment, what is obtained is the time series landslide impact factor data of a new target area to be assessed.

[0059] In specific implementation, the time series landslide impact factor data obtained in step 101 are multiple types of factor data within a preset time period. The multiple types of factors can be the discrete evaluation factors and continuous landslide impact factors mentioned below, that is, the preset time period includes multiple groups of time series factor data, such as a group of geological lithology factor data with time marks, a group of soil category factor data with time series, or a group of land use factor data with time series, etc.

[0060] The landslide impact factor data of different sources and different resolutions are resampled, projected, and grid-converted to obtain the grid-time series landslide impact factor with consistent spatial resolution and coordinate system. The specific steps include:

[0061] (1) Unify the coordinate system of all landslide influencing factors from different sources.

[0062] (2) Resample the data after the coordinate system is unified to a uniform resolution of 30 meters. Finally, the raster data is converted to obtain the raster time series landslide impact factor data with consistent spatial resolution and coordinate system. The following example is set to 5 periods of data, that is, Figure 4 The five-period landslide impact factor data shown in: Figure 4 a is the normalized difference vegetation index in 2013, Figure 4 b is the normalized difference vegetation index in 2014, Figure 4 c is the normalized difference vegetation index in 2015, Figure 4 d in the middle is the normalized difference vegetation index in 2016, Figure 4 e in the figure is the Normalized Difference Vegetation Index in 2017.

[0063] From the above, it can be seen that in one embodiment, Figure 3 FIG. 1 is a flow chart of a method for preprocessing time series landslide impact factor data in an embodiment of the present invention. Figure 3 As shown, the above-mentioned landslide susceptibility assessment method taking into account the factor change trend may also include: preprocessing the time series landslide influencing factor data according to the following steps:

[0064] Step 201: unifying the coordinate systems of the time series landslide impact factor data from different sources to obtain the time series landslide impact factor data of coordinate system 1;

[0065] Step 202: resampling the time series landslide impact factor data of coordinate system one to obtain the time series landslide impact factor data of coordinate system one with consistent spatial resolution;

[0066] Step 203: Performing raster data conversion processing on the time series landslide impact factor data with consistent spatial resolution and coordinate system 1 to obtain raster time series landslide impact factor data with consistent spatial resolution and coordinate system.

[0067] In specific implementation, the above-mentioned preprocessing implementation method can further improve the accuracy of landslide susceptibility assessment.

[0068] Step S2: Clip the above time series landslide impact factors with the scope of the study area to obtain the grid time series landslide impact factors with the same scope size and corresponding spatial position, that is, the grid time series landslide impact factors with the same scope size and corresponding spatial position are obtained by: based on the polygon elements of the study area, retain the pixels in the target data set that fall within the scope or intersect with the polygon elements, and clip other pixels outside the scope, thereby generating a new raster data set. All grid time series landslide impact factors use the same study area polygon elements, and finally obtain a set of grid time series landslide impact factors with the same scope size and corresponding spatial position. Figure 4 1 is a set of time series landslide impact factor diagrams in an embodiment of the present invention.

[0069] As can be seen from the above, in one embodiment, based on the scope of the target area to be evaluated, the time series landslide impact factor data is clipped to obtain a time series landslide impact factor set with the same scope size and corresponding spatial position, which may include:

[0070] Based on the polygonal elements of the target area to be evaluated, the pixels in the time series landslide impact factor dataset that fall within the range of the polygonal elements or intersect with the polygonal elements are retained, and the pixels outside the range of the polygonal elements are clipped to generate a new raster time series landslide impact factor dataset, thus obtaining a raster time series landslide impact factor set with the same range size and corresponding spatial position.

[0071] Step S3: quantify the above-mentioned time series landslide impact factors into trend landslide impact factors one by one; the specific method of quantifying the trend landslide impact factors is:

[0072] (1) According to the formula Calculate the point pair value S of the same pixel at different time points of a single time series landslide influencing factor ij, a point pair is a pair of data points at the same spatial position in different years such as 2013 and 2014, 2013 and 2015, 2013 and 2016. The point pair value S ij is the quotient of the data value difference and time difference of the same spatial location point pair at different time points (for example, different years), where i and j are the time of the previous year and the next year in the time series, respectively. <i<j ;y j and i are the data values ​​of the corresponding years, t j and t i are the time of the corresponding year. Generally, i is the initial year of the time series landslide impact factor, and j is any year thereafter. ij After sorting the values ​​in ascending order, take the median, denoted as G, that is, ij The median is recorded as G, which is the estimated value of the change trend of the pixel. The G value of all pixels is calculated according to the above formula, thereby quantifying the time series landslide impact factor into the trend landslide impact factor. Figure 5 This is a trend landslide impact factor diagram in an embodiment of the present invention.

[0073] (2) According to the above step (1), the remaining time-series landslide impact factors are quantified into trend landslide impact factors, thereby obtaining a set of trend landslide impact factors.

[0074] According to the above two steps, the quotient of the data value difference and the time difference of all the point pairs of landslide impact factors is obtained, and the median is taken as the value of the point after sorting them in ascending order, thereby obtaining the trend landslide impact factor of a single grid. Finally, the remaining time series landslide impact factors are quantified as trend landslide impact factors according to the above. Figure 6 1 is a set of trend landslide impact factor diagrams in an embodiment of the present invention.

[0075] From the above, it can be seen that in one embodiment, Figure 7 FIG. 4 is a flow chart of a method for quantifying a trend landslide impact factor in an embodiment of the present invention; Figure 7 As shown, quantifying each factor in the time series landslide impact factor set with the same range size and corresponding spatial position into a trend landslide impact factor to obtain the trend landslide impact factor set may include: for each factor in the time series landslide impact factor set with the same range size and corresponding spatial position, performing the following quantization into the trend landslide impact factor operation steps 301 to 302, and finally obtaining the trend landslide impact factor set:

[0076] Step 301: determining the point pair value of the same pixel at different time points for each landslide influencing factor, where the point pair value is the quotient of the data value difference of the point pair at the same spatial position and the time difference;

[0077] Step 302: sort the point pair values ​​of all the point pairs in ascending order and take the median as the estimated value of the change trend of the pixel. After the estimated values ​​of the change trend of all the pixels are calculated, the landslide impact factor is quantified as the change trend landslide impact factor.

[0078] In specific implementation, the above implementation method of quantifying the landslide trend influencing factor can further improve the accuracy of landslide susceptibility assessment.

[0079] Step S4: Obtain the influencing factors that are not easy to change or change slowly in the short term in the study area, cut the above landslide influencing factors with the scope of the study area, and process them into static landslide influencing factors.

[0080] The static landslide impact factors from different sources and different resolutions are projected, resampled, and processed into raster data. The pixels within the study area or intersecting with it are retained, and other pixels outside the range are cropped. After this processing, the static landslide impact factors of the grid with consistent spatial resolution, coordinate system, and range are obtained. Figure 8 It is a static landslide impact factor diagram in an embodiment of the present invention.

[0081] It can be seen from the above that, in one embodiment, the landslide susceptibility assessment method taking into account the factor variation trend may also include:

[0082] The static landslide impact factors of different sources and different resolutions are projected, resampled, and processed by raster data conversion, and the pixels within or intersecting the target area to be evaluated are retained, and the pixels outside the range are cropped to obtain the static landslide impact factors of the grid with consistent spatial resolution, consistent coordinate system, and consistent range; the static landslide impact factors are used to input into a pre-established landslide susceptibility assessment model together with the trend landslide impact factor set after the multicollinearity test to obtain the landslide susceptibility assessment results of the target area to be evaluated.

[0083] When implementing it, Figure 2 As shown, the static landslide influencing factors may also be subjected to the dimension removal step of step S5 and the multicollinearity test step of step S6 to further improve the accuracy of landslide susceptibility assessment.

[0084] In specific implementation, considering static landslide influencing factors as model input can further improve the accuracy of landslide susceptibility assessment.

[0085] Step S5: De-dimensionalizing the landslide impact factors, that is, comparing and processing data of different magnitudes to eliminate the unit influence of the landslide impact factors extracted above.

[0086] (1) Discrete evaluation factors (geological lithology factors, soil category factors, and land use factors) are discretized by reclassifying their categories. Fig.10 It is a discrete landslide impact factor diagram in an embodiment of the present invention.

[0087] (2) For continuous landslide influencing factors such as rainfall, slope, digital elevation model, normalized difference vegetation index and other landslide influencing factors. Fig.11 As shown in the figure, the continuous landslide impact factor is first divided into several preliminary intervals according to different value ranges, and each interval contains only one data point, which is recorded as C i . Calculate the proportion of each interval in the entire data set, denoted as P i (ratio value), and then use the formula The value of this interval is calculated and recorded as L i , calculate all intervals The sum (sum value) is minus sign and recorded as E. After this step, it is also necessary to calculate each interval C i The merged value, such as C i and C i+1 Merging is computing and The sum is denoted as E i Then according to the formula Calculate the weighted value of each interval, recorded as Eq. Finally, calculate the difference between E and each interval Eq (merging basis difference), recorded as LG. When used, those intervals that can bring a larger LG value (greater than the preset LG value threshold) after merging will be selected for merging, and the merged new interval will become the new candidate interval. Repeat the above steps until the merged LG value is less than the preset threshold, and finally get a set of discretized intervals, and divide each interval into a category. In this way, the discretization of the continuous landslide influencing factors is completed, that is, the second discrete trend landslide influencing factor set is finally obtained. Fig.11 This is an interval result diagram after dimension removal of the impact of continuous landslide in an embodiment of the present invention.

[0088] From the above, it can be seen that in one embodiment, Fig. 9 Schematic diagram of the process of de-dimensionalization processing in an embodiment of the present invention. Fig. 9 As shown, de-dimensionalizing each trend landslide impact factor in the trend landslide impact factor set to obtain the de-dimensionalized trend landslide impact factor set may include the following steps:

[0089] Step 401: for each discrete trend landslide impact factor, discretize it according to the discrete trend landslide impact factor category reclassification method to obtain a first discrete trend landslide impact factor set;

[0090] Step 402: for each continuous trend landslide impact factor, the following discretization processing operation of the continuous landslide impact factor is cyclically performed, and the following operations are performed in each cycle:

[0091] Step 4021: Divide the continuous trend landslide impact factor into a number of preliminary intervals according to different value ranges, each preliminary interval contains only one data point, and determine the proportion value of each preliminary interval in the entire data set;

[0092] Step 4022: Determine the value of each preliminary interval according to the ratio value of each preliminary interval;

[0093] Step 4023: Determine the sum of all preliminary intervals according to the ratio value of each preliminary interval and the value of each preliminary interval;

[0094] Step 4024: Determine the combined value of the one data point of all preliminary intervals according to the sum of all preliminary intervals;

[0095] Step 4025: Determine a weighted value for each preliminary interval according to the negative value of the sum of each preliminary interval and the combined value of the one data point of all preliminary intervals;

[0096] Step 4026: merging the preliminary intervals to obtain candidate intervals according to the minus sign value of the sum of each preliminary interval and the merging basis difference of the weighted value of each preliminary interval; until the merging basis difference after merging is less than a preset threshold to obtain a group of discretized intervals, and finally obtain a second discrete type of trend landslide influencing factor set;

[0097] Step 403: taking the first discrete type of trend landslide impact factor set and the second discrete type of trend landslide impact factor set as the trend landslide impact factor set after dimension removal processing.

[0098] In specific implementation, the above-mentioned dimension removal process can further improve the accuracy of landslide susceptibility assessment.

[0099] Step S6: Perform multicollinearity test on the landslide influencing factors after dimension removal to ensure the independence of the influencing factors. Fig.13 This is a graph showing the multicollinearity test results of landslide influencing factors in an embodiment of the present invention.

[0100] (1) Using the multivariate linear regression model, we took a single landslide influencing factor as the independent variable and the remaining landslide influencing factors as the dependent variables. First, we used the formula Calculate the sum of squares of the differences between the means of the independent variable and the dependent variable, denoted as Ys, and then use the formula Calculate the sum of squares of the differences between the predicted values ​​of the multivariate linear model and the true values, recorded as Yr, and finally calculate the difference between 1 and Ys divided by Yr, recorded as X. .

[0101] (2) Calculate the quotient of 1 and 1-X, denoted as Z, where Z is the coefficient of determination of the multivariate linear regression model. All landslide influencing factors are calculated according to the above steps. Finally, the landslide influencing factors with Z values ​​greater than 10 are eliminated. Fig.14 This is a graph showing the results of the rainfall and slope collinearity test in an embodiment of the present invention.

[0102] From the above, it can be seen that in one embodiment, Fig.12 Schematic diagram of the multicollinearity test method in the embodiment of the present invention. Fig.12 As shown in the figure, the multicollinearity test is performed on the de-dimensionalized trend landslide influencing factor set, and the trend landslide influencing factor set after the multicollinearity test is obtained, which may include:

[0103] Step 501: Perform the following operation to determine the value of the determination coefficient for each trend landslide impact factor in the trend landslide impact factor set:

[0104] Step 5011: using a multiple linear regression model with a single landslide influencing factor as an independent variable and the remaining landslide influencing factors as dependent variables, determine the sum of squares of the differences between the means of the independent variable and the dependent variable;

[0105] Step 5012: Determine the sum of squares of the differences between the predicted values ​​and the true values ​​of the multiple linear regression model;

[0106] Step 5013: Determine the difference between 1 minus the sum of squares of the differences between the means of the independent variable and the dependent variable divided by the sum of squares of the differences between the predicted value and the true value;

[0107] Step 5014: Determine the coefficient of determination of the multiple linear regression model according to the difference;

[0108] Step 502: After determining the determination coefficient values ​​corresponding to all trend landslide influencing factors, the landslide influencing factors with determination coefficient values ​​greater than 10 are eliminated to obtain a set of trend landslide influencing factors after multicollinearity test.

[0109] In specific implementation, the above-mentioned implementation method of multicollinearity test can further improve the accuracy of landslide susceptibility assessment.

[0110] The above steps S1 to S6 can be general steps for implementing the establishment of a landslide susceptibility assessment model and the actual landslide susceptibility assessment. When implementing the establishment of a landslide susceptibility assessment model, the following steps S7 to S9 are performed. When implementing the actual landslide susceptibility assessment, step 106 is implemented with reference to the landslide susceptibility assessment step when constructing the model in step S8.

[0111] Step S7: When implementing the landslide susceptibility assessment model, a training data set and an evaluation data set for landslide susceptibility assessment are prepared. The same number of non-landslide points as landslide points are selected in places far away from landslides, faults, roads, water systems, with high vegetation coverage and gentle slopes. The landslide points and non-landslide points are merged and randomly divided into 7 / 3, with 70% of the data used for training and 30% of the data used for evaluating the accuracy of the results.

[0112] Step S8: Landslide susceptibility assessment. Based on the machine learning model, the trend landslide influence factor is combined with the static landslide influence factor to assess the landslide susceptibility, that is, the trend landslide influence factor and the static landslide influence factor are assigned to 70% of the training data as the input variables of the model to train the model. After completion, the landslide probability corresponding to each pixel in the study area is evaluated, and the trend landslide influence factor and the static landslide influence factor are assigned to 70% of the data as the input variables of the model, and the output variable is the landslide probability corresponding to each pixel. Therefore, the input of the landslide susceptibility assessment model constructed in this way can be the trend landslide influence factor (or combined with the static landslide influence factor), and the output can be the landslide susceptibility assessment result (landslide probability and landslide area). Fig.15 This is a graph showing the landslide susceptibility assessment results in an embodiment of the present invention.

[0113] Step S9: Assess the accuracy and reliability of the landslide susceptibility assessment results. Assign the assessment results to the remaining 30% of the assessment data, and use the confusion matrix to build a 2x2 table for the accuracy of the assessment results. Fig.18 First, the number of landslides that were originally landslides and evaluated as landslides by the model is recorded as A, the number of non-landslides that were originally non-landslides and evaluated as non-landslides by the model is recorded as B, the number of non-landslides that were originally non-landslides and evaluated as landslides by the model is recorded as C, and the number of landslides that were originally landslides and evaluated as non-landslides by the model is recorded as D. The calculated result is used as the accuracy index of the result. The landslide susceptibility assessment results are divided into five levels, and the quotient of the proportion of landslide points in different levels and the proportion of the area in the same level is calculated to measure the reliability of the susceptibility assessment results. Fig.17 The zoning results in the landslide susceptibility assessment results shown should have a high degree of consistency with the actual landslide distribution. With the increase of the susceptibility level, the quotient value gradually increases, the value of the extremely low susceptibility area is the smallest, and the value of the extremely high susceptibility area is the largest. This zoning is more in line with the actual situation of landslide occurrence. Fig.16 This is a landslide susceptibility result accuracy assessment diagram in an embodiment of the present invention. Fig.17 This is a result diagram of the proportion of landslide points at different levels and the proportion of the area at the same level in an embodiment of the present invention.

[0114] It can be seen from the above that, in one embodiment, the above landslide susceptibility assessment method taking into account the factor variation trend may further include: evaluating the accuracy and reliability of the landslide susceptibility assessment result.

[0115] As can be seen from the above, in one embodiment, the accuracy and reliability of the landslide susceptibility assessment result are evaluated, including:

[0116] Based on the previously retained 30% evaluation data, a 2x2 table is constructed using the confusion matrix. The statistical results are divided into multiple categories (for example, A, B, C, and D). According to the formula Calculation is performed and the calculation result is used as the accuracy evaluation index. The closer the value is to 1, the higher the accuracy.

[0117] The landslide susceptibility assessment results are divided into five levels, and the quotient of the proportion of landslide points in different levels and the proportion of the area of ​​the same level is calculated to measure the reliability of the susceptibility assessment results.

[0118] In specific implementation, the verification step of evaluating the accuracy and reliability of the landslide susceptibility assessment results in the embodiment of the present invention further improves the assessment accuracy of the landslide susceptibility assessment model, thereby further improving the accuracy of subsequent landslide susceptibility assessment of the area to be assessed using the landslide susceptibility assessment model.

[0119] The beneficial effect of the landslide susceptibility assessment method that takes into account the changing trend of factors provided by the embodiment of the present invention is: the present invention uses drone data and remote sensing data as data sources to construct a landslide susceptibility assessment method that can take into account the changing trend and cumulative effect of landslide influencing factors. The trend landslide influencing factors quantified by this method can fully reflect the temporal and spatial changing trends and cumulative effects of landslide influencing factors, and provide a data basis for the accurate assessment of landslide susceptibility. The trend landslide influencing factors and static landslide influencing factors are combined to assess landslide susceptibility, and the results have high accuracy, and can accurately and clearly assess the risk level and spatial location of landslide disasters in the study area, which has important practical significance for geological disaster early warning and disaster prevention and mitigation.

[0120] The present invention also provides a device for assessing the susceptibility of landslides taking into account the trend of factor changes, as described in the following embodiments. Since the principle of solving the problem by the device is similar to that of the method for assessing the susceptibility of landslides taking into account the trend of factor changes, the implementation of the device can refer to the implementation of the method for assessing the susceptibility of landslides taking into account the trend of factor changes, and the repeated parts will not be repeated.

[0121] Fig.19 FIG. 4 is a schematic diagram of the structure of a landslide susceptibility assessment device taking into account the factor variation trend in an embodiment of the present invention. Fig.19 As shown, the device comprises:

[0122] An acquisition unit 01 is used to acquire the time series landslide impact factor data in the target area to be evaluated;

[0123] A clipping processing unit 02 is used to clip the time series landslide impact factor data based on the scope of the target area to be evaluated, so as to obtain a time series landslide impact factor set with a consistent scope size and corresponding spatial position;

[0124] The trend quantification processing unit 03 is used to quantify each factor in the time series landslide impact factor set with the same range size and corresponding spatial position into a trend landslide impact factor, so as to obtain a trend landslide impact factor set;

[0125] A dimension removal processing unit 04 is used to perform dimension removal processing on each trend landslide impact factor in the trend landslide impact factor set to obtain a trend landslide impact factor set after dimension removal processing;

[0126] The multicollinearity test unit 05 is used to perform a multicollinearity test on the trend landslide influencing factor set after dimension removal, and obtain the trend landslide influencing factor set after the multicollinearity test;

[0127] The evaluation processing unit 06 is used to input the trend landslide impact factor set after the multicollinearity test into a pre-established landslide susceptibility evaluation model to obtain the landslide susceptibility evaluation result of the target area to be evaluated; the landslide susceptibility evaluation model is pre-established based on the relationship sample data between the trend landslide impact factor set after the historical multicollinearity test and the landslide susceptibility evaluation result.

[0128] In one embodiment, the above-mentioned landslide susceptibility assessment device taking into account the factor variation trend may further include a preprocessing unit for preprocessing the time series landslide influencing factor data according to the following method:

[0129] Unify the coordinate systems of the time series landslide impact factor data from different sources to obtain the time series landslide impact factor data of coordinate system 1;

[0130] Resample the time series landslide impact factor data of coordinate system one to obtain the time series landslide impact factor data of coordinate system one with consistent spatial resolution;

[0131] The time series landslide impact factor data with consistent spatial resolution and coordinate system are converted into raster data to obtain the raster time series landslide impact factor data with consistent spatial resolution and coordinate system.

[0132] In one embodiment, the clipping processing unit is specifically used for:

[0133] Based on the polygonal elements of the target area to be evaluated, the pixels in the time series landslide impact factor dataset that fall within the range of the polygonal elements or intersect with the polygonal elements are retained, and the pixels outside the range of the polygonal elements are clipped to generate a new raster time series landslide impact factor dataset, thus obtaining a raster time series landslide impact factor set with the same range size and corresponding spatial position.

[0134] In one embodiment, the trend quantification processing unit is specifically used to: for each factor in the time series landslide impact factor set with the same range size and corresponding spatial position, perform the following quantization operation into a trend landslide impact factor, and finally obtain the trend landslide impact factor set:

[0135] Determine the point pair value of the same pixel at different time points for each landslide influencing factor. The point pair value is the quotient of the data value difference of the point pair at the same spatial position and the time difference.

[0136] After sorting the point pair values ​​of all the point pairs in ascending order, the median is taken as the estimated value of the change trend of the pixel. The estimated value of the change trend is the quantified change trend of the landslide impact factor.

[0137] In one embodiment, the dimension removal processing unit is specifically used for:

[0138] For each discrete trend landslide impact factor, discretize it according to the discrete trend landslide impact factor category reclassification method to obtain the first discrete trend landslide impact factor set;

[0139] For each continuous trend landslide impact factor, the following discretization processing of the continuous landslide impact factor is performed cyclically, and the following operations are performed in each cycle:

[0140] The continuous trend landslide impact factor is divided into several preliminary intervals according to different value ranges, each preliminary interval contains only one data point, and the proportion value of each preliminary interval in the entire data set is determined;

[0141] Determine the value of each preliminary interval according to the proportion value of each preliminary interval;

[0142] Determine the sum of all preliminary intervals according to the proportion value of each preliminary interval and the value of each preliminary interval;

[0143] Determine, according to the sum of all preliminary intervals, a combined value of the one data point of all preliminary intervals;

[0144] Determine a weighted value for each preliminary interval according to the negative value of the sum of each preliminary interval and the combined value of the one data point of all preliminary intervals;

[0145] According to the minus sign value of the sum of each preliminary interval and the merging basis difference of the weighted value of each preliminary interval, preliminary intervals are merged to obtain candidate intervals; until the merging basis difference after merging is less than a preset threshold, a group of discretized intervals is obtained, and finally a second discrete type of trend landslide influencing factor set is obtained;

[0146] The first discrete type of trend landslide impact factor set and the second discrete type of trend landslide impact factor set are taken as the trend landslide impact factor set after dimension removal.

[0147] In one embodiment, the multicollinearity test unit is specifically used for:

[0148] For each trend landslide impact factor in the trend landslide impact factor set, the following operation is performed to determine the determination coefficient value:

[0149] A multivariate linear regression model was used with a single landslide influencing factor as an independent variable and the remaining landslide influencing factors as dependent variables to determine the sum of squares of the differences between the means of the independent and dependent variables;

[0150] Determine the sum of squares of the differences between the predicted values ​​and the true values ​​of the multiple linear regression model;

[0151] Determine the difference between 1 minus the sum of the squares of the differences between the means of the independent and dependent variables divided by the sum of the squares of the differences between the predicted and true values;

[0152] According to the difference, determining the coefficient of determination value of the multiple linear regression model;

[0153] After determining the determination coefficient values ​​corresponding to all trend landslide influencing factors, the landslide influencing factors with determination coefficient values ​​greater than 10 were eliminated to obtain the trend landslide influencing factor set after multicollinearity test.

[0154] In one embodiment, the above-mentioned landslide susceptibility assessment device taking into account the factor variation trend may further include a verification unit for evaluating the accuracy and reliability of the landslide susceptibility assessment result.

[0155] In one embodiment, the verification unit is specifically used to:

[0156] Based on the previously retained 30% evaluation data, a 2x2 table is constructed using the confusion matrix. The statistical results are divided into multiple categories (for example, A, B, C, and D). According to the formula Calculation is performed and the calculation result is used as the accuracy evaluation index. The closer the value is to 1, the higher the accuracy.

[0157] The landslide susceptibility assessment results are divided into five levels, and the quotient of the proportion of landslide points in different levels and the proportion of the area in the same level is calculated to measure the reliability of the susceptibility assessment results.

[0158] In one embodiment, the above-mentioned landslide susceptibility assessment device taking into account the trend of factor changes may also include a static landslide influence factor processing unit, which is used to: perform projection conversion, resampling, and raster data conversion processing on static landslide influence factors of different sources and different resolutions, retain pixels within or intersecting the target area to be evaluated, and cut off pixels outside the range to obtain static landslide influence factors of the grid with consistent spatial resolution, consistent coordinate system, and consistent range; the static landslide influence factors are used to input into a pre-established landslide susceptibility assessment model together with a set of trend landslide influence factors after multicollinearity test to obtain a landslide susceptibility assessment result for the target area to be evaluated.

[0159] An embodiment of the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned landslide susceptibility assessment method taking into account the factor change trend when executing the computer program.

[0160] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned landslide susceptibility assessment method taking into account the factor change trend.

[0161] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the landslide susceptibility assessment method taking into account the factor change trend is implemented.

[0162] Compared with the prior art landslide susceptibility assessment method which only uses static or dynamic factors to perform landslide susceptibility assessment, which has shortcomings and leads to low accuracy of landslide susceptibility assessment, the landslide susceptibility assessment scheme taking into account the factor change trend provided in the embodiment of the present invention has the beneficial technical effect as follows: the embodiment of the present invention takes into account that the occurrence of landslide is a dynamically changing process, and fully reflects the temporal and spatial change trend and cumulative effect of landslide influencing factors through quantified trend landslide influencing factors, thereby improving the accuracy of landslide susceptibility assessment.

[0163] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0164] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0165] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0166] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0167] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A landslide susceptibility assessment method taking into account the changing trend of factors, characterized in that: include: Obtain the time series landslide impact factor data in the target area to be evaluated; Based on the scope of the target area to be evaluated, the time series landslide impact factor data are clipped to obtain a set of time series landslide impact factors with consistent scope and corresponding spatial location; Each factor in the time series landslide impact factor set with the same range and corresponding spatial position is quantified as a trend landslide impact factor, and a trend landslide impact factor set is obtained, which includes: for each factor in the time series landslide impact factor set with the same range and corresponding spatial position, the following quantization operation is performed to quantify it into a trend landslide impact factor, and finally the trend landslide impact factor set is obtained: the point pair value of the same pixel at different time points of each landslide impact factor is determined, and the point pair value is the quotient of the data value difference of the point pair at the same spatial position and the time difference; wherein, according to the formula Calculate the point pair value S of the same pixel at different time points of a single time series landslide influencing factor ij ; The point value S ij is the quotient of the data value difference and time difference of the same spatial position point pair at different time points, where i and j are the time of the previous year and the next year in the time series, respectively. <i<j ;y j and i are the data values ​​of the corresponding years, t j and t i are the time of the corresponding year respectively; after sorting the point pair values ​​of all the point pairs in ascending order, the median is taken as the estimated value of the change trend of the pixel, and the estimated value of the change trend is the quantitative change trend of the landslide impact factor; De-dimensionalize each trend landslide impact factor in the trend landslide impact factor set to obtain the de-dimensionalized trend landslide impact factor set; Conduct multicollinearity test on the de-dimensionalized trend landslide influencing factor set to obtain the trend landslide influencing factor set after multicollinearity test; The landslide influencing factor set after the multicollinearity test is input into a pre-established landslide susceptibility assessment model to obtain the landslide susceptibility assessment result of the target area to be assessed; the landslide susceptibility assessment model is pre-established based on the relationship sample data between the landslide influencing factor set after the historical multicollinearity test and the landslide susceptibility assessment result.

2. The method according to claim 1, characterized in that Also includes: The time series landslide influencing factor data were preprocessed as follows: Unify the coordinate systems of the time series landslide impact factor data from different sources to obtain the time series landslide impact factor data of coordinate system 1; Resample the time series landslide impact factor data of coordinate system one to obtain the time series landslide impact factor data of coordinate system one with consistent spatial resolution; The time series landslide impact factor data with consistent spatial resolution and coordinate system are converted into raster data to obtain the raster time series landslide impact factor data with consistent spatial resolution and coordinate system.

3. The method according to claim 1, characterized in that Based on the scope of the target area to be evaluated, the time series landslide impact factor data are clipped to obtain a set of time series landslide impact factors with consistent scope and corresponding spatial location, including: Based on the polygonal elements of the target area to be evaluated, the pixels in the time series landslide impact factor dataset that fall within the range of the polygonal elements or intersect with the polygonal elements are retained, and the pixels outside the range of the polygonal elements are clipped to generate a new raster time series landslide impact factor dataset, thus obtaining a raster time series landslide impact factor set with the same range size and corresponding spatial position.

4. The method according to claim 1, characterized in that Each trend landslide impact factor in the trend landslide impact factor set is de-dimensionalized to obtain the trend landslide impact factor set after de-dimensionalization, including: For each discrete trend landslide impact factor, discretize it according to the discrete trend landslide impact factor category reclassification method to obtain the first discrete trend landslide impact factor set; For each continuous trend landslide impact factor, the following discretization processing of the continuous landslide impact factor is performed cyclically, and the following operations are performed in each cycle: The continuous trend landslide impact factor is divided into several preliminary intervals according to different value ranges, and each preliminary interval contains only one data point C i , determine the proportion value P of each preliminary interval in the entire data set i ; According to the proportion value of each preliminary interval, the value of each preliminary interval is determined as follows: , denoted as L i ; According to the proportion value of each preliminary interval and the value of each preliminary interval, the sum of all preliminary intervals is determined as follows: , take the negative sign and record it as E; According to the sum of all preliminary intervals, the value of the one data point after merging all preliminary intervals is determined, and the formula is as follows: C i and C i+1 ,calculate and The sum is denoted as E i ; According to the negative value of the sum of each preliminary interval and the combined value of the one data point of all preliminary intervals, the weighted value of each preliminary interval is determined, and the formula is as follows: Calculate the weighted value of each interval, denoted as Eq; According to the minus sign value of the sum of each preliminary interval and the merging basis difference of the weighted value of each preliminary interval, the preliminary intervals are merged to obtain candidate intervals; until the merging basis difference after merging is less than the preset threshold, a group of discretized intervals are obtained, and finally a second discrete type of trend landslide influencing factor set is obtained, and the merging basis difference is: the difference between E and each interval Eq is calculated; The first discrete type of trend landslide impact factor set and the second discrete type of trend landslide impact factor set are taken as the trend landslide impact factor set after dimension removal.

5. The method according to claim 1, characterized in that The multicollinearity test is performed on the de-dimensionalized trend landslide influencing factor set, and the trend landslide influencing factor set after the multicollinearity test is obtained, including: For each trend landslide impact factor in the trend landslide impact factor set, the following operation is performed to determine the determination coefficient value: A multivariate linear regression model was used to take a single landslide influencing factor as an independent variable and the remaining landslide influencing factors as dependent variables. The sum of squares of the differences between the means of the independent and dependent variables was determined using the formula: , denoted as Ys; Determine the sum of squares of the differences between the predicted values ​​and the true values ​​of the multiple linear regression model, using the formula: , denoted as Yr; Determine the difference between 1 minus the sum of the squares of the differences between the means of the independent and dependent variables divided by the sum of the squares of the differences between the predicted values ​​and the true values, using the formula: , denoted as X; According to the difference, determine the coefficient of determination of the multiple linear regression model, wherein the quotient of 1 and 1-X is calculated, denoted as Z, and Z is the coefficient of determination of the multiple linear regression model; After determining the determination coefficient values ​​corresponding to all trend landslide influencing factors, the landslide influencing factors with determination coefficient values ​​greater than 10 were eliminated to obtain the trend landslide influencing factor set after multicollinearity test.

6. A landslide susceptibility assessment device taking into account the trend of factor changes, characterized in that: include: An acquisition unit, used for acquiring time series landslide impact factor data in the target area to be evaluated; A clipping processing unit is used to clip the time series landslide impact factor data based on the scope of the target area to be evaluated, so as to obtain a set of time series landslide impact factors with consistent scope size and corresponding spatial position; The trend quantification processing unit is used to quantify each factor in the time series landslide impact factor set with the same range size and corresponding spatial position into a trend landslide impact factor, and obtain the trend landslide impact factor set, which includes: for each factor in the time series landslide impact factor set with the same range size and corresponding spatial position, the following quantization operation into a trend landslide impact factor is performed to finally obtain the trend landslide impact factor set: determine the point pair value of the same pixel at different time points of each landslide impact factor, and the point pair value is the quotient of the data value difference of the point pair at the same spatial position and the time difference; wherein, according to the formula Calculate the point pair value S of the same pixel at different time points of a single time series landslide influencing factor ij ; The point value S ij is the quotient of the data value difference and time difference of the same spatial position point pair at different time points, where i and j are the time of the previous year and the next year in the time series, respectively. <i<j ;y j and i are the data values ​​of the corresponding years, t j and t i are the time of the corresponding year respectively; after sorting the point pair values ​​of all the point pairs in ascending order, the median is taken as the estimated value of the change trend of the pixel, and the estimated value of the change trend is the quantitative change trend of the landslide impact factor; A dimension removal processing unit is used to perform dimension removal processing on each trend landslide impact factor in the trend landslide impact factor set to obtain the trend landslide impact factor set after dimension removal processing; A multicollinearity test unit is used to perform a multicollinearity test on the set of trend landslide influencing factors after dimension removal, and obtain a set of trend landslide influencing factors after the multicollinearity test; The evaluation processing unit is used to input the trend landslide influencing factor set after the multicollinearity test into a pre-established landslide susceptibility evaluation model to obtain the landslide susceptibility evaluation result of the target area to be evaluated; the landslide susceptibility evaluation model is pre-established based on the relationship sample data between the trend landslide influencing factor set after the historical multicollinearity test and the landslide susceptibility evaluation result.

7. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

9. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images