Mountain torrent disaster early warning method based on multi-source data fusion

Abstract

The invention discloses a mountain torrent disaster early warning method based on multi-source data fusion, and the method comprises the steps: 1, building a hydrological-side slope coupling dynamic model based on a TRIGRS landslide dynamic model, and enabling the hydrological-side slope coupling dynamic model to integrate dynamic calculation modules of soil pore water pressure, soil moisture content and slip shear strength; 2, Beidou-R soil humidity inversion data and three-dimensional deformation field monitoring data are accessed in real time, and a fusion data layer is generated; 3, space-time calculation is conducted on the fusion data layer through the hydrology-slope coupling dynamic model, a dynamic forecast value of the landslide safety coefficient Fs is output, Fs = sliding face resistance / sliding force, and when Fslt, Fs = sliding face resistance / sliding force; 1, triggering landslide early warning; according to the method, the fusion data layer is subjected to space-time calculation through the hydrology-slope coupling dynamic model, the dynamic forecast value of the landslide safety coefficient Fs is output, rainfall-induced geological disaster early warning and forecasting are achieved, and the accuracy of geological disaster area risk early warning is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of geological disaster risk early warning, and in particular to a flash flood disaster early warning method based on multi-source data fusion. Background Art

[0002] Geological hazards are secondary hazards that harbor significant potential hazards, and landslides are one of the most significant natural disasters. The timely extraction of landslide information and the assessment of landslide hazard are crucial for landslide early warning. Rainfall-induced landslides are primarily caused by rainfall infiltration into the landslide, affecting pore water pressure, soil moisture content, bulk density, and shear strength within the landslide mass. This impacts the stability of the landslide, ultimately leading to its instability and failure.

[0003] The research and application demonstration project on regional risk warning of geological disasters by integrating precipitation and deformation information mainly targets the problems of low accuracy and spatiotemporal resolution of current regional risk warning products of geological disasters. According to the idea of ​​model research, business construction and application demonstration, the coupling relationship between precipitation and the risk of geological disasters is studied. However, the existing flash flood disaster warning model is not accurate enough and is prone to omissions or false alarms. Therefore, this application proposes a flash flood disaster warning method based on multi-source data fusion. Summary of the Invention

[0004] The purpose of the present invention is to provide a flash flood disaster early warning method based on multi-source data fusion to solve the problems raised in the above background technology.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a flash flood disaster early warning method based on multi-source data fusion, comprising:

[0006] Step 1: construct a hydrological-slope coupling dynamic model based on the TRIGRS landslide dynamic model, wherein the hydrological-slope coupling dynamic model integrates dynamic calculation modules of soil pore water pressure, soil moisture content, and shear strength of the sliding mass;

[0007] Step 2: Real-time access to BeiDou-R soil moisture inversion data and 3D deformation field monitoring data to generate a fused data layer;

[0008] Step 3: Perform spatiotemporal calculations on the fused data layer using the hydrological-slope coupled dynamic model, and output a dynamic forecast value of the landslide safety factor Fs, where Fs = sliding surface resistance / sliding force. When Fs < 1, a landslide warning is triggered.

[0009] Among them, the calculation formula of the landslide safety factor Fs in the hydrological-slope coupled dynamic model is:

[0010] ;

[0011] in, ;

[0012] ;

[0013] ;

[0014] Wherein, the landslide safety factor Fs is the ratio of the sliding surface resistance to the sliding force. When Fs < 1, that is, the resistance is less than the sliding force, the slope will slide.

[0015] is the soil bulk density correction term; θ is the slope angle, φ´ is the soil internal friction angle, m is the hydraulic coupling correction coefficient, n is the porosity, G s is the specific gravity of soil, S r is the soil water saturation, γ w is the specific gravity of water, H is the depth of soil, and C´ is the total cohesion.

[0016] The total cohesion C´ is expressed as:

[0017] ;

[0018] Where c' is the effective internal friction; C φ is the stress correction term; Δs is the three-dimensional displacement; is the weakening effect of humidity on soil structure. The higher the water content m, the greater the loss of cohesion.

[0019] A is the amplification factor that regulates the effect of humidity, and a high A value indicates that the soil is sensitive to humidity;

[0020] is the humidity correction factor. When m=0, it means dry. =A; when m approaches saturation, Approaches A(1−λ).

[0021] Among them, in step 2, the Beidou-R soil moisture inversion data adopts a microwave remote sensing inversion algorithm, specifically including: constructing a soil volume moisture inversion model through the ground reflection intensity and phase delay of the Beidou navigation satellite signal, combined with the surface cover type classification data, with an inversion resolution of 100m×100m, and a data update frequency of not less than 1 hour.

[0022] Among them, the inversion formula of the soil volumetric water content inversion model is:

[0023] ;

[0024] Where m is the volumetric water content of the soil; is the L-band microwave backscattering coefficient; is the dry soil background scattering value; is the soil moisture sensitivity coefficient; is the noise correction term.

[0025] Among them, in step 2, the three-dimensional deformation field monitoring data is obtained through the Beidou ground augmentation station network, specifically including: deploying no less than 3 Beidou monitoring points on the surface of the landslide body, using carrier phase difference technology to solve the three-dimensional displacement, the deformation monitoring accuracy reaches the millimeter level, and the data sampling interval is ≤5 minutes.

[0026] The calculation formula for the three-dimensional displacement is:

[0027] ;

[0028] Where, is the three-dimensional displacement; is the radar wavelength; is the InSAR phase difference; is the angle of incidence; for Data weight coefficient.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] The present invention constructs a hydrological-slope coupling dynamic model based on the TRIGRS landslide dynamic model, then generates a fused data layer from BeiDou-R soil moisture inversion data and three-dimensional deformation field monitoring data, and performs spatiotemporal calculations on the fused data layer through the hydrological-slope coupling dynamic model to output a dynamic forecast value of the landslide safety factor Fs, where Fs = anti-slip force / slip force. When Fs < 1, a landslide warning is triggered, realizing early warning and forecasting of precipitation-induced geological disasters, and improving the accuracy of regional risk warnings for geological disasters. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a schematic diagram of the process of the present invention;

[0032] Figure 2 It is a schematic diagram of the cross-sectional structure of the landslide of the present invention. DETAILED DESCRIPTION

[0033] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0034] See also Figure 1-2 , Figure 1It is a schematic diagram of the process of the present invention; Figure 2 This is a schematic diagram of the cross-sectional structure of a landslide according to the present invention. The upper half of the figure shows the spatial distribution of monitoring points of the three-dimensional deformation field on the surface; the lower half of the cross-sectional diagram shows the sliding surface, pore water pressure distribution, and key parameter measurement points. The combination of the three-dimensional and cross-sectional diagrams realizes the "surface deformation-underground structure: coupling analysis". In the figure, I is the length of the sliding surface; h is the total water head (position head + pressure head); is the pore water pressure head (water head height converted from pore water pressure); is the soil slope (the angle between the sliding surface and the horizontal plane); σ is the normal stress (the compressive stress perpendicular to the sliding surface); s is the effective stress (the stress actually borne by the soil skeleton, s=σ−u); u is the pore water pressure (the pressure generated by water in the soil pores); w is the angle between the shear stress direction and the surface of maximum principal stress; is the shear stress (shear force along the sliding surface); c is the cohesion of the soil (the bonding force that resists shear failure); is the internal friction angle of soil (friction characteristics between soil particles).

[0035] The present invention provides a technical solution: a flash flood disaster early warning method based on multi-source data fusion, comprising:

[0036] Step 1: construct a hydrological-slope coupling dynamic model based on the TRIGRS landslide dynamic model, wherein the hydrological-slope coupling dynamic model integrates dynamic calculation modules of soil pore water pressure, soil moisture content, and shear strength of the sliding mass;

[0037] Step 2: Real-time access to BeiDou-R soil moisture inversion data and 3D deformation field monitoring data to generate a fused data layer;

[0038] Step 3: Perform spatiotemporal calculations on the fused data layer using the hydrological-slope coupled dynamic model, and output a dynamic forecast value of the landslide safety factor Fs, where Fs = sliding surface resistance / sliding force. When Fs < 1, a landslide warning is triggered.

[0039] Furthermore, a hydrological-slope coupling dynamic model is constructed based on the TRIGRS landslide dynamic model, and then a fused data layer is generated by the Beidou-R soil moisture inversion data and the three-dimensional deformation field monitoring data. The fused data layer is temporally and spatially solved by the hydrological-slope coupling dynamic model, and a dynamic forecast value of the landslide safety factor Fs is output, where Fs = anti-sliding force / sliding force. When Fs < 1, a landslide warning is triggered. The Beidou-R soil moisture inversion data and the three-dimensional deformation field monitoring data are comprehensively integrated to establish a hydrological-slope coupling dynamic model based on the TRIGRS landslide dynamic model, realize the early warning and forecast of precipitation-induced geological disasters, and improve the accuracy of regional risk warnings of geological disasters.

[0040] The dynamic calculation method of soil pore water pressure is: based on the saturated-unsaturated seepage theory of the TRIGRS model, using the Richard equation:

[0041] ;

[0042] Where, is the volumetric moisture content, the volume ratio of water in unit volume of soil (water volume / total soil volume), 0≤ ≤ ( is the saturated moisture content); is the pressure head, the height of the water column corresponding to the soil pore water pressure, h>0: pore water pressure is positive (saturated zone), h<0: pore water pressure is negative (unsaturated zone); The permeability coefficient is the ability of soil to allow water to flow through, which varies with the moisture content. change; is the height of the position, the vertical coordinate corresponding to the gravitational potential (usually the surface is taken as the reference plane z = 0, and downward is the negative direction); is the source-sink term, which is the net inflow / outflow of water per unit volume of soil per unit time; t is the time dimension of dynamic calculation, which reflects the evolution of pore water pressure over time; is the gradient / divergence operator, is the head gradient (pressure gradient + gravity gradient); is the spatial variation of water flux; It represents the dynamic changes of soil moisture content and pressure head, that is, soil pore water pressure.

[0043] in, ;

[0044] Where, is the saturated permeability coefficient, the soil is completely saturated = The maximum permeability coefficient at , which is directly related to the soil pore structure and connectivity; is the residual moisture content, the amount of immobile water that the soil can hold under very high suction (it cannot be taken up by plants or drained by gravity); is the saturated moisture content, the maximum moisture content when the soil pores are completely filled with water, = porosity; is the volumetric moisture content, which is the volume percentage of water in unit volume of soil; n is an empirical index, which characterizes the nonlinear parameter of soil pore size distribution. The larger the value, the steeper the change of permeability coefficient with moisture content. Determination method: obtained through experimental data fitting (such as pressure plate test), the typical range is 1.2-2.5.

[0045] The dynamic calculation method of the shear strength of the sliding body is as follows: combining the Mohr-Coulomb criterion and introducing the humidity correction term:

[0046] ;

[0047] Where, The shear strength of the sliding mass, the maximum ability of the soil to resist shear failure, is the core parameter for determining slope stability; is the effective cohesion, the inherent bonding force between soil particles (the strength component that is independent of pore water pressure); is the effective normal stress, the compressive stress perpendicular to the sliding surface (net stress after deducting the pore water pressure); is the effective internal friction angle, the friction characteristics between soil particles, reflecting the ability of shear strength to increase with normal stress; is the moisture sensitivity coefficient, which quantifies the influence of moisture content on the attenuation of soil shear strength ( The larger the value, the more significant the humidity effect); is a nonlinear index that controls the nonlinear degree of the effect of moisture content on shear strength ( >1, the strength of the high moisture content area decays faster); is the soil moisture content, which is the volume percentage of water in unit volume of soil.

[0048] Among them, the calculation formula of the landslide safety factor Fs in the hydrological-slope coupled dynamic model is:

[0049] ;

[0050] in, ;

[0051] ;

[0052] ;

[0053] Where, the landslide safety factor Fs is the slope safety factor, which measures the ratio of the slope's anti-sliding force to its sliding force and is used to assess landslide risk: F S >1 indicates stability, F S <1 indicates instability. When Fs<1, that is, the resistance is less than the sliding force, the slope will slide;

[0054] is the soil bulk density correction term; θ is the slope angle, the angle between the sliding surface and the horizontal plane; φ´ is the soil internal friction angle, which reflects the strength parameter of the friction characteristics between soil particles; m is the hydraulic coupling correction coefficient, which represents the nonlinear effect of precipitation infiltration on pore water pressure; n is the porosity, the ratio of the pore volume to the total volume of the soil; G sis the specific gravity of soil, the ratio of the density of soil particles to the density of pure water at 4°C, S r is the soil water saturation, the ratio of the volume of water in the pores to the total volume of the pores, γ w is the specific gravity of water, reflecting the density characteristics of water; H is the soil depth, the vertical height of the potential sliding body, and the vertical depth from the trailing edge of the landslide to the sliding surface; C' is the total cohesion, the strength parameter of the inherent bonding force between soil particles.

[0055] Reflects the effect of slope gradient on anti-sliding resistance. The larger the slope angle (θ↑), The smaller it is, the lower the safety factor;

[0056] Characterizes the degree to which the friction strength in the soil is effectively exerted;

[0057] represents the effective mass proportion of soil skeleton;

[0058] represents the pore water filling effect;

[0059] The overall reflection of the composite bulk density characteristics of the soil in a partially saturated state.

[0060] middle, It represents the amplified nonlinear effect of porosity and reflects the pore reconstruction effect caused by continuous seepage. When n>0.3, this term significantly reduces the overall strength (especially in sandy soil layers).

[0061] is the cohesion influence coefficient, represents the integration of sliding surface inclination and geometric scale effect, and the denominator Characterizes the reducing effect of hydrostatic pressure on shear strength.

[0062] The total cohesion C´ is expressed as:

[0063] ;

[0064] Where, the total cohesion C' is the combined effective cohesion and stress / humidity-corrected cohesion, which is used to describe the strength characteristics of unsaturated soil; c' is the effective internal friction, the inherent bonding force between soil particles, without considering the influence of pore water pressure. The inherent cohesive strength of soil is related to mineral composition and compaction degree and is relatively stable in dry or saturated state; C φ is the stress correction term, which is the additional cohesive force component caused by external stress or humidity change; Δs is the three-dimensional displacement, which is the progressive change of cohesion with shear displacement;

[0065] is the weakening effect of humidity on soil structure; λ is the humidity influence coefficient, which characterizes the regulatory effect of water content changes on cohesion; m represents the water content, which is the ratio of the volume of water in the soil to the total volume; α is the nonlinear correction index, which describes the nonlinear relationship between the influence of humidity on cohesion; the higher the water content m, the greater the cohesion loss;

[0066] A is the amplification factor that regulates the effect of humidity. A high A value indicates that the soil is sensitive to humidity (e.g., expansive soil).

[0067] is the humidity correction factor. When m=0, it means dry. =A; when m approaches saturation, Approaches A(1−λ).

[0068] Among them, in step 2, the Beidou-R soil moisture inversion data adopts a microwave remote sensing inversion algorithm, specifically including: constructing a soil volume moisture inversion model through the ground reflection intensity and phase delay of the Beidou navigation satellite signal, combined with the surface cover type classification data, with an inversion resolution of 100m×100m, and a data update frequency of not less than 1 hour.

[0069] Among them, the soil volumetric water content is also the soil moisture content. The inversion formula of the soil volumetric water content (soil moisture content) inversion model is:

[0070] ;

[0071] Where m is the volumetric water content of the soil; is the L-band microwave backscattering coefficient; is the dry soil background scattering value; is the soil moisture sensitivity coefficient; is the noise correction term.

[0072] Among them, the noise correction term The calculation method is to make multiple measurements at the same location, record the fluctuation range of the measured values, use standard soil samples with known soil volume moisture content for measurement, and compare the difference between the measured values ​​and the true values, which is the noise correction term. .

[0073] Among them, in step 2, the three-dimensional deformation field monitoring data is obtained through the Beidou ground augmentation station network, specifically including: deploying no less than 3 Beidou monitoring points on the surface of the landslide body, using carrier phase difference technology to solve the three-dimensional displacement, the deformation monitoring accuracy reaches the millimeter level, and the data sampling interval is ≤5 minutes.

[0074] The calculation formula for the three-dimensional displacement is:

[0075] ;

[0076] Where, is the three-dimensional displacement; is the radar wavelength; is the InSAR phase difference; is the angle of incidence; for Data weight coefficient.

[0077] Among them, InSAR phase difference The acquisition of is achieved through interferometric processing of satellite radar images. The specific calculation method is:

[0078] Step 1: Use a satellite equipped with synthetic aperture radar (SAR) (such as Sentinel-1 and ALOS) to observe the target area twice or more, obtaining two radar images (primary and secondary) of the same area. SAR generates high-resolution images of surface reflection intensity by transmitting microwave signals and receiving echoes. Because radar signals travel in a straight line, slight changes in their round-trip path will result in phase differences.

[0079] Step 2: Perform sub-pixel registration of the auxiliary image with the primary image to ensure that the pixels of the two images are completely aligned. The registration error must be controlled within 1 / 10 of the wavelength (e.g., approximately 5.6 cm in the C band), correcting for the time difference between the two observations and eliminating geometric distortion caused by Earth rotation, orbital deviation, etc.

[0080] Step 3: Multiply the complex data of the primary image and the secondary image to generate an interference pattern:

[0081] ;

[0082] Where, is the phase of the main image; The phase of the auxiliary image; is the terrain phase (caused by surface elevation); It is the atmospheric delay phase (caused by water vapor and temperature changes).

[0083] Step 4: The interferometric phase value is limited to [−π,π). The actual deformation phase needs to be restored by "using multi-temporal data to constrain the phase evolution law (such as PS-InSAR technology)" to obtain the true cumulative value, which is the phase change after unwrapping, thereby obtaining the deformation variable ;

[0084] Step 5: Convert the phase difference into deformation data in the geographic coordinate system, i.e. InSAR phase difference for:

[0085] ;

[0086] Where, is the radar wavelength; is the phase change after unwrapping; is the radar incident angle.

[0087] in, It is the abbreviation of Global Navigation Satellite System. It is a general term that refers to any system that provides global navigation satellite services. represents the measurement value provided by the i-th global navigation satellite system, Expressing different The system's measurements are weighted summed.

[0088] Among them, the Beidou-R soil moisture inversion data and the three-dimensional deformation field monitoring data generate a fused data layer. By means of spatiotemporal alignment, complementary enhancement and dynamic weight optimization, the two heterogeneous data of soil moisture (reflecting the physical state of the soil) and deformation field (reflecting structural stability) are integrated to generate comprehensive risk indicators and improve the sensitivity and accuracy of landslide warning.

[0089] The specific framework of the fusion method model is:

[0090] 1. Data preprocessing:

[0091] BeiDou-R soil moisture retrieval data:

[0092] Spatiotemporal interpolation: Based on the inversion resolution (100m×100m) and update frequency (≥1h), Kriging interpolation or inverse distance weighted method is used to fill in the missing areas.

[0093] Noise filtering: Savitzky-Golay filter is used to eliminate random noise of microwave remote sensing signals. The formula is: ;

[0094] in, is the spatiotemporal humidity value (or displacement), the humidity observation value (or deformation value) at time t and spatial position (x, y), indexed as i; i is a spatial or temporal index, which may represent the row number in the grid division (such as the i-th column) or the time step number; K is the sliding window size, which controls the time or spatial window length of the summation range (for example, K=3 means taking three adjacent points); is the weight coefficient, dynamic weight (such as distance decay weight, time decay weight); 、 is the spatial sampling interval; j is the column index, the grid column number corresponding to i (such as the position (i, j) in the two-dimensional array).

[0095] 3D deformation field data:

[0096] Coordinate registration: Align the displacement data (millimeter-level accuracy) of the Beidou monitoring points with the landslide DEM model in the geographic information system (GIS).

[0097] Deformation rate calculation: Perform sliding window difference on the displacement time series to obtain the deformation rate field. The formula is: ;

[0098] in, is the displacement change at time t, defined as the displacement difference from the current time t to t+Δt (i.e., Δs(t)=s(t+Δt)-s(t)); is the displacement change at time t+Δt, the displacement difference from time t+Δt to time t+2Δt in the future (i.e., Δs(t+Δt)=s(t+2Δt)-s(t+Δt)); is the time interval, the time difference between two adjacent observations, ≤5min; s(t) is the displacement at time t, which is the instantaneous value of a physical quantity (such as landslide displacement) at time t.

[0099] 2. Feature-level fusion:

[0100] Moisture feature extraction: calculate soil moisture gradient, , characterizing the uneven spatial distribution of moisture; is the total variation of m in space (or gradient amplitude), which indicates the comprehensive variation intensity of m in the x and y directions and is used to measure the spatial heterogeneity of m; is the partial derivative of m with respect to x, describing the rate of change of m in the x direction (such as the gradient of soil moisture in the east-west direction); is the partial derivative of m with respect to y, describing the rate of change of m in the y direction (such as the gradient of soil moisture in the north-south direction); x and y are spatial coordinates.

[0101] Assume that the data of a landslide monitoring point is:

[0102] The rate of change of soil moisture m in the x direction: = 0.1% / m;

[0103] The rate of change of soil moisture m in the y direction: =0.2% / m;

[0104] but: =0.22% / m, indicating that the humidity in this area presents a significant spatial gradient change and may be a high-risk area.

[0105] The effect of humidity on soil strength is corrected by combining the surface cover type (vegetation, bare soil), and the formula is: ; is the cover type correction factor (e.g. 0.8 for vegetation area, 1.2 for bare soil area); is the initial cohesion (intrinsic cohesion of soil), the basic value of cohesion of soil in dry or water-free state, unit: kPa or MPa; is an empirical correction coefficient, a dimensionless coefficient related to soil type and structure; m is the soil volume water content, the volume percentage of water in the soil, dimensionless (0-1), and λ is an empirical correction coefficient used to quantify the effect of soil volume water content (m) on the modified cohesion ( ) by adjusting the humidity term The contribution ratio of humidity to cohesion is used to control the weakening effect of humidity changes on soil mechanical properties (such as cohesion); the larger λ is, the more significant the negative impact of humidity (m) on cohesion is (i.e., cohesion decreases faster under high humidity); the smaller λ is, the weaker the effect of humidity changes on cohesion is (i.e., the soil is insensitive to humidity).

[0106] in the formula The exponent α introduces a nonlinear response (α ≥ 1), and λ further adjusts the strength of this nonlinear effect. For example, when α = 2 (the squared moisture term), λ may amplify or suppress the effect of the squared moisture term on cohesion. The value of λ should be determined based on the specific soil type (e.g., clay, sand) and pore structure. For example, clay typically has a higher λ value than sand (because clay is more sensitive to moisture changes).

[0107] Deformation feature extraction:

[0108] Extract deformation rate threshold (such as =5mm / day), divide the stable area and the deformation area; calculate the deformation acceleration ), identifying the accelerated deformation stage; a(t) is the acceleration (a function of time t), which describes the rate of change of velocity with time, that is, the acceleration characteristics of deformation or motion; v(t) is the velocity (a function of time t), which describes the rate of change of displacement with time, that is, the deformation rate of the landslide or soil; t is time, the time variable for monitoring or calculation; dt is the small change in time, the differential of time, indicating that the time interval in the acceleration calculation approaches infinitesimal.

[0109] 3. Decision-level fusion:

[0110] Spatiotemporal correlation model: Constructing the humidity-deformation coupling matrix:

[0111] ;

[0112] in, is the spatiotemporal coupling risk score, which is the comprehensive risk value (such as landslide warning score) at time t and spatial position (i, j); α, β, γ are weight coefficients, which are calibrated by historical data (such as least squares method) to control humidity respectively. ,deformation and the contribution ratio of its interaction term; is the humidity parameter (humidity value of the i-th grid), soil volume water content or pore water pressure, reflecting the degree of soil moisture; is the deformation parameter (displacement of the jth monitoring point), the millimeter-level deformation variable monitored by Beidou, reflecting the slope deformation rate; t is the time variable, monitoring time; through linear weighting and interaction terms, the coupling relationship between humidity and deformation is quantified. For example: when humidity Elevated and deformed When accelerating, the interaction term Significantly increased, triggering a higher risk warning.

[0113] Dynamic weight allocation:

[0114] Dynamically adjust weights based on data reliability:

[0115] When humidity dominates the scene (high humidity + low deformation rate): =0.7, =0.3

[0116] When deformation dominates the scene (low humidity + high deformation rate): =0.3, =0.7.

[0117] Comprehensive risk score:

[0118] ;

[0119] is the fused comprehensive risk score, the final fusion result at time t and spatial position (x, y), which is used for landslide warning decision-making; 、 is the weight coefficient, which controls the humidity data and deformation data The contribution ratio must meet + =1; w is the total weight normalization factor, ensuring that the sum of the weights is 1; Moisture-related characteristic values, such as soil volumetric water content and moisture gradient, reflect the impact of soil moisture on risk; It is the deformation-related characteristic value, such as the displacement and deformation rate monitored by Beidou, which reflects the impact of slope structure stability on risk.

[0120] Input the fusion results into the hydrological-slope coupled dynamic model: ;

[0121] like > (The threshold is determined by historical disaster cases and is the critical value that triggers warnings or system status changes. When the threshold is exceeded, the system determines it is an abnormal state (such as landslide risk). The threshold is a critical value set by historical data or engineering experience), and F S <1, a landslide warning is triggered. This fusion model realizes the complementary advantages of multi-source data through spatiotemporal synchronization, feature coupling, dynamic weight optimization and physical constraints, solves the limitations of a single data source in spatiotemporal resolution and sensitivity, and provides high-precision and high-reliability technical support for urban flood control and flash flood disaster warning.

[0122] In summary, this application constructs a hydrological-slope coupling dynamic model based on the TRIGRS landslide dynamic model, and then generates a fused data layer from the Beidou-R soil moisture inversion data and the three-dimensional deformation field monitoring data. The fused data layer is temporally and spatially solved by the hydrological-slope coupling dynamic model, and the dynamic forecast value of the landslide safety factor Fs is output, where Fs = anti-sliding force / sliding force. When Fs < 1, a landslide warning is triggered. The Beidou-R soil moisture inversion data and the three-dimensional deformation field monitoring data are comprehensively integrated to establish a hydrological-slope coupling dynamic model based on the TRIGRS landslide dynamic model, realize the early warning and forecast of precipitation-induced geological disasters, and improve the accuracy of regional risk warnings of geological disasters.

[0123] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A flash flood disaster early warning method based on multi-source data fusion, characterized in that: include: Step 1: construct a hydrological-slope coupling dynamic model based on the TRIGRS landslide dynamic model, wherein the hydrological-slope coupling dynamic model integrates dynamic calculation modules of soil pore water pressure, soil moisture content, and shear strength of the sliding mass; Step 2: Real-time access to BeiDou-R soil moisture inversion data and 3D deformation field monitoring data to generate a fused data layer; Step 3: Perform spatiotemporal calculations on the fused data layer using the hydrological-slope coupled dynamic model, and output a dynamic forecast value of the landslide safety factor Fs, where Fs = sliding surface resistance / sliding force. When Fs < 1, a landslide warning is triggered.

2. The flash flood disaster early warning method based on multi-source data fusion according to claim 1, characterized in that: The calculation formula of landslide safety factor Fs in the hydrological-slope coupled dynamic model is: ； in, ; ； ； Wherein, the landslide safety factor Fs is the ratio of the sliding surface resistance to the sliding force. When Fs < 1, that is, the resistance is less than the sliding force, the slope will slide. is the soil bulk density correction term; θ is the slope angle, φ´ is the soil internal friction angle, m is the hydraulic coupling correction coefficient, n is the porosity, G s is the specific gravity of soil, S r is the soil water saturation, γ w is the specific gravity of water, H is the depth of soil, and C´ is the total cohesion.

3. The flash flood disaster early warning method based on multi-source data fusion according to claim 2 is characterized by: The total cohesion C´ is expressed as: ； Where c' is the effective internal friction; C φ is the stress correction term; Δs is the three-dimensional displacement; is the weakening effect of humidity on soil structure. The higher the water content m, the greater the loss of cohesion. A is the amplification factor that regulates the effect of humidity, and a high A value indicates that the soil is sensitive to humidity; is the humidity correction factor. When m=0, it means dry. =A; when m approaches saturation, Approaches A(1−λ).

4. The flash flood disaster early warning method based on multi-source data fusion according to claim 1, characterized in that: In step 2, the Beidou-R soil moisture inversion data adopts a microwave remote sensing inversion algorithm, specifically including: constructing a soil volume moisture inversion model through the ground reflection intensity and phase delay of the Beidou navigation satellite signal, combined with the surface cover type classification data, with an inversion resolution of 100m×100m, and a data update frequency of not less than 1 hour.

5. The flash flood disaster early warning method based on multi-source data fusion according to claim 4 is characterized by: The inversion formula of the soil volumetric water content inversion model is: ； Where m is the volumetric water content of the soil; is the L-band microwave backscattering coefficient; is the dry soil background scattering value; is the soil moisture sensitivity coefficient; is the noise correction term.

6. The flash flood disaster early warning method based on multi-source data fusion according to claim 1, characterized in that: In step 2, the three-dimensional deformation field monitoring data is obtained through the Beidou ground augmentation station network, which specifically includes: deploying no less than 3 Beidou monitoring points on the surface of the landslide body, using carrier phase difference technology to solve the three-dimensional displacement, the deformation monitoring accuracy reaches the millimeter level, and the data sampling interval is ≤5 minutes.

7. The flash flood disaster early warning method based on multi-source data fusion according to claim 6, characterized in that: The calculation formula for the three-dimensional displacement is: ； Where, is the three-dimensional displacement; is the radar wavelength; is the InSAR phase difference; is the angle of incidence; for Data weight coefficient.

Images