A method and system for monitoring the space-time deformation of a slope based on multi-source remote sensing data

Abstract

The application discloses a kind of based on multi-source remote sensing data's slope deformation space-time monitoring method and system, specifically related to the technical field of surface deformation detection and measurement using radio waves, for solving the problem that existing periodic monitoring mode cannot capture non-steady-state accelerated deformation triggered by rainfall in time due to fixed time interval.The present application constructs a rainfall-deformation response sensitivity baseline, converts the monitoring of the micro-deformation section from periodic re-measurement to dynamic response discrimination driven by the current rainfall event, identifies sensitive enhancement sections with enhanced deformation response sensitivity compared to the same period in history, confirms effective sensitive sections through the trend of backscattering characteristic space distribution, extracts two-dimensional deformation field and spatiotemporal distribution of cracks from high-resolution images of effective sensitive sections, and assesses the risk level of accelerated slip.

Description

Technical Field

[0001] This invention relates to the technical field of detecting and measuring surface deformation using radio waves, and more specifically, to a method and system for spatiotemporal monitoring of slope deformation based on multi-source remote sensing data. Background Technology

[0002] Numerous excavated slopes along highways and railways are composed of special rock and soil materials such as red bed soft rock and expansive soil. Regional deformation monitoring of these slopes is fundamental to ensuring operational safety. Current technologies commonly employ spaceborne synthetic aperture radar interferometry to conduct area surveys of the slopes along the route, obtaining large-scale surface deformation rates. Based on this, potential hazard sections showing signs of deformation are identified. Furthermore, for the selected key hazard slopes, precise and periodic verification is conducted using methods such as global navigation satellite systems and UAV aerial surveys. This forms a monitoring model of general survey delineation and detailed verification. This model concentrates monitoring frequency and resources on key areas where deformation has been confirmed. For a large number of slopes identified as having minimal deformation and low risk levels, periodic re-measurements are conducted at longer intervals, or manual inspections are relied upon.

[0003] The aforementioned periodic monitoring model is essentially a discrete spatial sampling based on fixed time intervals. This results in discontinuous gaps in the monitoring of a large number of non-key slopes in the time domain. For slopes where deformation does not develop at a uniform rate but is triggered by short-term strong disaster-causing factors such as extreme rainfall, resulting in an unsteady accelerated response, when the timing of this acceleration process is misaligned with the monitoring window, key deformation information will be missed, making the early warning of slope instability lack the necessary timely response capability in actual implementation. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method and system for spatiotemporal monitoring of slope deformation based on multi-source remote sensing data to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for spatiotemporal monitoring of slope deformation based on multi-source remote sensing data includes the following steps:

[0007] S1: Acquire synthetic aperture radar image sequences and rainfall data for the slope area to delineate areas with minor deformation;

[0008] S2: Extract the deformation response sequence of each small deformation segment under historical rainfall events, and construct a rainfall-deformation response sensitivity baseline based on the deformation rate before rainfall;

[0009] S3: Input the current rainfall event into the rainfall-deformation response sensitivity baseline, determine whether there are small deformation segments with enhanced deformation response sensitivity compared to the same period in history, and identify them as sensitive enhanced segments;

[0010] S4: Extract the spatial distribution of synthetic aperture radar backscattering features of sensitive enhancement sections before and after the current rainfall event, analyze the changing trend of the spatial distribution of backscattering features, and determine the effective sensitive sections by matching the changing trend with the evolution law of existing landslide backscattering features.

[0011] S5: Acquire high-resolution synthetic aperture radar and optical images of the effective sensitive section, and extract the two-dimensional deformation field and spatiotemporal distribution of cracks on the slope of the effective sensitive section.

[0012] S6: Based on the two-dimensional deformation field and spatiotemporal distribution of cracks in the effective sensitive section, assess the risk level of accelerated slip in the effective sensitive section.

[0013] Furthermore, synthetic aperture radar (SAR) image sequences and rainfall data for the slope area are acquired to delineate areas of minor deformation. This includes: acquiring SAR image sequences covering the slope area from spaceborne SAR and acquiring rainfall data for the slope area from a ground-based meteorological network; performing differential interferometry on the SAR image sequences to generate differential interferogram sequences and deformation rate maps for the slope area, and extracting the changes in deformation rate and cumulative deformation of each pixel in the slope area over time; performing time-series analysis on the rainfall data to identify the start and end time windows of historical rainfall events and the rainfall intensity distribution within each time window; and delineating the set of pixels with deformation rates below a preset rate threshold and cumulative deformation exhibiting slight fluctuations over time as areas of minor deformation within the slope area.

[0014] Furthermore, the deformation response sequences of each minor deformation segment under historical rainfall events are extracted, and a rainfall-deformation response sensitivity baseline is constructed based on the deformation rate before rainfall. This includes: for each minor deformation segment, extracting the change process of deformation rate over time within the start and end time window of the historical rainfall event, and using the difference sequence between the deformation rate at each moment within the start and end time window and the deformation rate before rainfall as the deformation response sequence; arranging the deformation response sequences according to the chronological order of the historical rainfall events, generating historical response curves for each minor deformation segment under different historical rainfall events, and using the mean envelope of the historical response curves as the rainfall-deformation response sensitivity baseline.

[0015] Furthermore, the mean envelope of the historical response curve is used as the baseline for rainfall-deformation response sensitivity. This includes: calculating the mean value of the deformation rate response value of each small deformation segment at the same relative time under different historical rainfall events, and connecting the mean values ​​at each relative time to form a mean response curve; taking the upper and lower envelope values ​​of the deformation rate response value at each relative time point on the mean response curve, connecting the upper envelope values ​​to form an upper envelope line, and connecting the lower envelope values ​​to form a lower envelope line, and using the upper and lower envelope lines as the envelope range of the rainfall-deformation response sensitivity baseline.

[0016] Furthermore, the current rainfall event is input into the rainfall-deformation response sensitivity baseline to determine whether there are any small deformation segments whose deformation response sensitivity is enhanced compared to the same period in history. These segments are then identified as sensitive enhancement segments. This process includes: obtaining the start and end time windows and rainfall intensity distribution of the current rainfall event; extracting the deformation rate change process of each small deformation segment over time within the start and end time windows of the current rainfall event; generating the current response curve; comparing the current response curve with the rainfall-deformation response sensitivity baseline; calculating the amplitude and duration of the current response curve exceeding the envelope of the rainfall-deformation response sensitivity baseline; and identifying small deformation segments whose amplitude and duration exceed the preset discrimination conditions as sensitive enhancement segments.

[0017] Furthermore, the spatial distribution of synthetic aperture radar backscattering features of sensitive enhancement sections before and after the current rainfall event is extracted, and the changing trend of the spatial distribution of backscattering features is analyzed. Sections whose changing trends match the evolution patterns of existing landslide backscattering features are identified as effective sensitive sections. This includes: extracting the backscattering coefficients of sensitive enhancement sections before and after the current rainfall event from the synthetic aperture radar image sequence, generating spatial distribution maps of backscattering features before and after rainfall; performing differential processing on the spatial distribution maps of backscattering features before and after rainfall, extracting areas where the difference values ​​are concentrated and continuously distributed in a planar pattern in the spatial distribution difference map; matching the spatial morphology of these areas with concentrated difference values ​​and continuously distributed in a planar pattern with the evolution patterns of existing landslide backscattering features, and identifying sensitive enhancement sections with a matching degree exceeding a preset condition as effective sensitive sections.

[0018] Furthermore, high-resolution synthetic aperture radar (SAR) and optical images of the effective sensitive section are acquired, and the two-dimensional deformation field and spatiotemporal distribution of the slope in the effective sensitive section are extracted. This includes: acquiring spaceborne high-resolution SAR and spaceborne high-resolution optical images covering the effective sensitive section; performing differential interferometry processing on the spaceborne high-resolution SAR images to generate the two-dimensional deformation field of the slope in the effective sensitive section; and performing continuous temporal image recognition on the spaceborne high-resolution optical images to extract the location of new fractures, the direction of expansion, and the width change process of the slope fractures in the effective sensitive section as the spatiotemporal distribution of the fractures.

[0019] Furthermore, based on the two-dimensional deformation field and spatiotemporal distribution of fractures in the effective sensitive section, the risk level of accelerated slip in the effective sensitive section is assessed, including: extracting the location of the deformation concentration area and the deformation rate gradient direction from the two-dimensional deformation field, and extracting the location of fracture formation and the direction of fracture propagation from the spatiotemporal distribution of fractures; analyzing the spatial angle relationship between the deformation rate gradient direction of the deformation concentration area and the fracture propagation direction, and simultaneously analyzing the time overlap relationship between the deformation rate growth period of the deformation concentration area and the fracture width change process; when the spatial angle is less than a preset angle threshold and the time overlap ratio exceeds a preset overlap threshold, the effective sensitive section is assessed as having a high accelerated slip risk level.

[0020] Furthermore, the determination of the preset angle threshold and the preset overlap threshold includes: obtaining the two-dimensional deformation field and spatiotemporal distribution of cracks in known landslide cases within the slope area; extracting the spatial angle between the deformation rate gradient direction and the crack propagation direction in the deformation concentration area of ​​the known landslide cases, as well as the time overlap ratio between the deformation rate growth period and the crack width change process in the deformation concentration area; and using the maximum value of the spatial angle and the minimum value of the time overlap ratio in the known landslide cases as the preset angle threshold and the preset overlap threshold, respectively.

[0021] On the other hand, the present invention provides a spatiotemporal monitoring system for slope deformation based on multi-source remote sensing data, comprising the following modules:

[0022] The data acquisition module is used to acquire synthetic aperture radar image sequences and rainfall data of the slope area and delineate areas with minor deformations.

[0023] The baseline construction module is used to extract the deformation response sequence of each small deformation segment under historical rainfall events and construct a rainfall-deformation response sensitivity baseline based on the deformation rate before rainfall.

[0024] The segment identification module is used to input the current rainfall event into the rainfall-deformation response sensitivity baseline, determine whether there are small deformation segments whose deformation response sensitivity is enhanced compared to the same period in history, and identify them as sensitive enhanced segments.

[0025] The section determination module is used to extract the spatial distribution of synthetic aperture radar backscattering features of sensitive enhanced sections before and after the current rainfall event, analyze the changing trend of the spatial distribution of backscattering features, and determine the effective sensitive sections that match the changing trend with the evolution law of existing landslide backscattering features.

[0026] The deformation extraction module is used to acquire high-resolution synthetic aperture radar and optical images of the effective sensitive section, and extract the two-dimensional deformation field and spatiotemporal distribution of cracks on the slope of the effective sensitive section.

[0027] The risk assessment module is used to assess the risk level of accelerated slip in the effective sensitive section based on the two-dimensional deformation field and spatiotemporal distribution of the fracture.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] 1. By constructing a rainfall-deformation response sensitivity baseline and using continuous rainfall data as a triggering condition, the monitoring of small deformation sections is transformed from periodic retesting at fixed time intervals to dynamic response discrimination driven by rainfall events. When rainfall occurs, the deformation response sequence is automatically compared with the historical baseline, and sections exceeding the envelope are immediately identified as sensitive enhanced sections. This allows rainfall-triggered accelerated deformation occurring between two periodic retests to be captured in a timely manner, filling the monitoring blind spot in the time domain of conventional survey models and improving the timeliness of early warning for slopes.

[0030] 2. The variation trend of the spatial distribution of backscattering characteristics is used to reconfirm sensitive enhancement sections, eliminating spurious responses caused by factors such as changes in surface humidity, and narrowing the scope of high-resolution fine verification. Subsequently, high-resolution images are acquired for the confirmed effective sensitive sections, and the two-dimensional deformation field and fracture spatiotemporal distribution are extracted. The spatial angle relationship between the deformation gradient direction and the fracture propagation direction, as well as the temporal overlap relationship between the deformation growth period and the fracture change period, are analyzed. This enables the assessment of accelerated slip risk level from the perspective of deformation-fracture feedback mechanism, improving the reliability of early identification of slopes that generate unsteady accelerated responses triggered by rainfall. Attached Figure Description

[0031] Figure 1 This is a flowchart of a method for spatiotemporal monitoring of slope deformation based on multi-source remote sensing data according to the present invention;

[0032] Figure 2 This is a schematic diagram of the structure of a slope deformation spatiotemporal monitoring system based on multi-source remote sensing data according to the present invention. Detailed Implementation

[0033] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0034] Example 1: Figure 1 This invention presents a spatiotemporal monitoring method for slope deformation based on multi-source remote sensing data, comprising the following steps:

[0035] S1: Acquire synthetic aperture radar image sequences and rainfall data for the slope area to delineate areas with minor deformation;

[0036] S2: Extract the deformation response sequence of each small deformation segment under historical rainfall events, and construct a rainfall-deformation response sensitivity baseline based on the deformation rate before rainfall;

[0037] S3: Input the current rainfall event into the rainfall-deformation response sensitivity baseline, determine whether there are small deformation segments with enhanced deformation response sensitivity compared to the same period in history, and identify them as sensitive enhanced segments;

[0038] S4: Extract the spatial distribution of synthetic aperture radar backscattering features of sensitive enhancement sections before and after the current rainfall event, analyze the changing trend of the spatial distribution of backscattering features, and determine the effective sensitive sections by matching the changing trend with the evolution law of existing landslide backscattering features.

[0039] S5: Acquire high-resolution synthetic aperture radar and optical images of the effective sensitive section, and extract the two-dimensional deformation field and spatiotemporal distribution of cracks on the slope of the effective sensitive section.

[0040] S6: Based on the two-dimensional deformation field and spatiotemporal distribution of cracks in the effective sensitive section, assess the risk level of accelerated slip in the effective sensitive section.

[0041] During S1, a sequence of synthetic aperture radar (SAR) images covering the slope area is acquired from a spaceborne SAR satellite. The spaceborne SAR is a C-band SAR satellite operating in a sun-synchronous orbit. It images the slope area with the same orbit number at revisit intervals, acquiring single-view complex images containing phase and amplitude information. The single-view complex images acquired from each imaging session are arranged chronologically to form a SAR image sequence covering the slope area.

[0042] Rainfall data for the slope area was obtained from a network of surface meteorological stations. Surface meteorological stations were deployed around the slope area, and the rain gauges at each station recorded real-time rainfall at a sampling frequency of minutes. The rainfall data from each surface meteorological station corresponding to the slope area were aligned by time and interpolated to obtain continuous rainfall data covering the slope area. The continuous rainfall data is in millimeters and includes rainfall intensity values ​​per minute.

[0043] Differential interferometry (DI) is performed on synthetic aperture radar (SAR) image sequences to generate a sequence of differential interferograms and a deformation rate map for the slope region. Image pairs within the SAR sequence whose imaging time and spatial baseline are both within a defined range are selected. Using an external digital elevation model (DEM) as a terrain reference, the main image and auxiliary image are multiplied by complex conjugate to generate an interferogram. Then, differential interferograms are obtained through de-flattening and terrain-removing phase processing. DI is performed on all available image pairs sequentially according to their imaging time to generate a sequence of differential interferograms for the slope region. For each differential interferogram in the sequence, phase unwrapping and orbit refinement are performed. The deformation rate of each pixel along the radar line-of-sight is obtained through least-squares inversion. The deformation rates of all pixels together constitute the deformation rate map of the slope region.

[0044] The deformation rate and cumulative deformation of each pixel in the slope area are extracted over time. On the deformation rate map, taking each pixel as a unit, the deformation rate at each imaging moment is extracted sequentially within the time span covered by the differential interferogram sequence, forming the deformation rate variation over time. Using the initial imaging moment of the differential interferogram sequence as the zero point of time, the deformation rates at each moment are accumulated over time to obtain the cumulative deformation variation over time.

[0045] Time-series analysis was performed on rainfall data to identify the start and end time windows of historical rainfall events and the rainfall intensity distribution within each time window. Continuous rainfall data was traversed, with a period of sustained rainfall intensity greater than zero considered as a rainfall event. The start time window was defined as the moment the rainfall intensity jumped from zero, and the end time window was defined as the moment the rainfall intensity returned to zero and remained stable. Within the start and end time windows of each rainfall event, rainfall intensity values ​​were extracted at the minute-by-minute sampling level, forming the rainfall intensity distribution within each time window.

[0046] Within the slope area, the set of pixels with deformation rates below a preset rate threshold and whose cumulative deformation exhibits slight fluctuations over time is designated as a minor deformation zone. The preset rate threshold is set based on the statistical quantile of the deformation rates of all pixels on the deformation rate map; for example, the statistical median of the deformation rates of all pixels is used as the preset rate threshold. The deformation rate value of each pixel is compared with the preset rate threshold, and all pixels with deformation rates below the preset rate threshold are selected. Among the selected pixels, the change process of cumulative deformation over time is analyzed, and pixels whose cumulative deformation fluctuates around zero over time with a fluctuation amplitude not exceeding a preset fluctuation threshold are extracted. The preset fluctuation threshold is set based on the deformation measurement uncertainty caused by radar system thermal noise; for example, a multiple of the deformation measurement uncertainty corresponding to radar system thermal noise is used as the preset fluctuation threshold. The set of pixels that simultaneously meet the criteria of a deformation rate below the preset rate threshold and a cumulative deformation exhibiting slight fluctuations over time is designated as a minor deformation zone.

[0047] In implementing S2, for each minor deformation segment, the deformation rate over time is extracted within the start and end time windows of historical rainfall events. Historical rainfall events are those identified in S1 through time-series analysis of continuous rainfall data. The start and end time windows are the start and end time windows of each historical rainfall event. For each minor deformation segment delineated in S1, the statistical average of the deformation rates of all pixels within its pixel set is taken as the representative deformation rate of the minor deformation segment. Within the start and end time windows of each historical rainfall event, representative deformation rate values ​​for each minor deformation segment are extracted one by one according to the time sampling points corresponding to the deformation rate's change over time.

[0048] The deformation response sequence is formed by the difference between the deformation rate at each moment within the start and end time window and the deformation rate before rainfall. The deformation rate before rainfall is the statistical average of the representative deformation rates of the small deformation segment within a preset stable period prior to the start time window of the historical rainfall event. The duration of the preset stable period is set according to the imaging time interval of the differential interferogram sequence, for example, taking an integer multiple of the imaging time interval of the differential interferogram sequence, and the duration of the preset stable period is not less than the empirical duration that ensures no trend change in the representative deformation rate within the preset stable period. For each time sampling point within the start and end time window, the difference between the representative deformation rate value at each time sampling point and the deformation rate before rainfall is calculated. The differences of all time sampling points are arranged chronologically to form the deformation response sequence of the small deformation segment under a historical rainfall event. Positive values ​​in the deformation response sequence represent deformation rates greater than the deformation rate before rainfall, and negative values ​​represent deformation rates less than the deformation rate before rainfall.

[0049] Deformation response sequences are arranged chronologically according to historical rainfall events to generate historical response curves for each minor deformation segment under different historical rainfall events. For each minor deformation segment, all historical rainfall events are arranged chronologically according to their starting time windows. For each historical rainfall event, the deformation response sequence is plotted with the starting time window as the zero point, the relative time of subsequent time sampling points to the zero point as the x-axis, and the difference in deformation rate at each time sampling point in the deformation response sequence as the y-axis, to create a response curve. The response curves generated for the same minor deformation segment under all historical rainfall events are then arranged chronologically to form the historical response curve for that minor deformation segment.

[0050] Calculate the mean deformation rate response value of each minor deformation segment at the same relative moment under different historical rainfall events. For each minor deformation segment, take the historical response curves of all historical rainfall events, and extract each deformation rate response value corresponding to the same relative moment on all historical response curves, using relative time as the alignment reference. Statistically average each extracted deformation rate response value to calculate the mean deformation rate response value for each relative moment. Iterate through all relative moment points within the start and end time window to obtain the mean deformation rate response value corresponding to each relative moment point.

[0051] Connect the mean values ​​at each relative time point to form a mean response curve. Plotting relative time on the x-axis and the mean deformation rate response value at each relative time point on the y-axis, connect adjacent relative time points to form a continuous curve, which serves as the mean response curve. The mean response curve represents the average deformation response level of a small deformation segment under historical rainfall events.

[0052] The upper and lower envelope values ​​of the deformation rate response values ​​at each relative time point on the mean response curve are taken. The upper envelope value is the mean of the deformation rate response values ​​at the same relative time point in all historical response curves of all historical rainfall events, plus the deviation between the mean and the maximum deformation rate response value. The lower envelope value is the mean of the deformation rate response values ​​at the same relative time point minus the deviation between the mean and the minimum deformation rate response value. The degree of deviation between the mean and the maximum and minimum values ​​is characterized by the statistical standard deviation of the deformation rate response values. For example, the upper envelope value is taken as the mean + 2 times the statistical standard deviation, and the lower envelope value is taken as the mean - 2 times the statistical standard deviation. The statistical standard deviation is calculated based on the dispersion of the deformation rate response values ​​of the same small deformation segment at the same relative time point under all historical rainfall events.

[0053] Connecting the upper envelope values ​​forms the upper envelope line, and connecting the lower envelope values ​​forms the lower envelope line. These upper and lower envelope lines constitute the envelope range of the rainfall-deformation response sensitivity baseline. Using relative time as the abscissa, connecting the upper envelope values ​​at each relative time point forms the upper envelope line, and connecting the lower envelope values ​​at each relative time point forms the lower envelope line. The region between the upper and lower envelope lines constitutes the envelope range of the rainfall-deformation response sensitivity baseline. The rainfall-deformation response sensitivity baseline represents the normal fluctuation range of the deformation response of a small deformation segment to historical rainfall events. When the deformation response caused by subsequent rainfall events exceeds the envelope range between the upper and lower envelope lines, it indicates that the deformation response sensitivity of the small deformation segment has increased.

[0054] When implementing S3, the start and end time windows and rainfall intensity distribution of the current rainfall event are obtained. The current rainfall event is the rainfall process occurring in the slope area. In S1, from the continuous rainfall data obtained from the surface meteorological station network, the period in which the most recent rainfall intensity jumps from zero and remains above zero is identified as the current rainfall event. The start time window of the current rainfall event is the moment when the rainfall intensity jumps from zero, and the end time window is the moment when the rainfall intensity falls back to zero and remains stable. The rainfall intensity distribution of the current rainfall event is a sequence of rainfall intensity values ​​extracted at the minute-level sampling time within the start and end time windows of the current rainfall event.

[0055] The deformation rate of each minor deformation segment changes over time within the start and end time window of the current rainfall event, generating the current response curve. For each minor deformation segment delineated in S1, the statistical average of the deformation rates of all pixels in its pixel set is taken as the representative deformation rate of the minor deformation segment. Within the start and end time window of the current rainfall event, representative deformation rate values ​​of minor deformation segments are extracted one by one according to the time sampling points corresponding to the deformation rate change over time. Taking the start time window of the current rainfall event as the zero point, the relative time of the time sampling points after the zero point relative to the zero point is used as the abscissa, and the difference between the representative deformation rate value of each time sampling point within the start and end time window and the deformation rate before rainfall is used as the ordinate, the current response curve is plotted. The deformation rate before rainfall is the statistical average of the representative deformation rate of the small deformation segment within a preset stable period before the start time window of the current rainfall event. The preset stable period is set in the same way as the preset stable period of historical rainfall events in S2, for example, taking an integer multiple of the imaging time interval of the differential interferogram sequence, and the duration of the preset stable period is not less than the empirical duration that makes the representative deformation rate have no trend change within the preset stable period.

[0056] The current response curve is compared with the rainfall-deformation response sensitivity baseline to calculate the extent to which the current response curve exceeds the envelope of the rainfall-deformation response sensitivity baseline. The rainfall-deformation response sensitivity baseline is the envelope between the upper and lower envelopes constructed in S2 for the same small deformation segment. At the same relative time point between the current response curve and the rainfall-deformation response sensitivity baseline, the difference in deformation rate of the current response curve is compared with the values ​​of the upper and lower envelopes. When the difference in deformation rate of the current response curve at a certain relative time point is greater than the value of the upper envelope, the difference between the difference in deformation rate of the current response curve and the value of the upper envelope is calculated as the excess magnitude at that relative time point. When the difference in deformation rate of the current response curve at a certain relative time point is less than the value of the lower envelope, the difference between the value of the lower envelope and the difference in deformation rate of the current response curve is calculated as the excess magnitude at that relative time point. This process is repeated for all relative time points of the current response curve to obtain the excess magnitude corresponding to each relative time point where an excess occurs.

[0057] Calculate the duration for which the current response curve exceeds the rainfall-deformation response sensitivity baseline envelope. Starting from the relative point in time when the current response curve first shows an excess amplitude, and ending at the end of the interval where the time interval between adjacent relative points of excess amplitude does not exceed a preset interval threshold, the duration of a single excess is defined as the time length from the start to the end of the interval. The preset interval threshold is set by taking the time sampling interval of the deformation rate change over time, multiplied by an empirical coefficient. This empirical coefficient is determined based on statistical analysis of the time span from anomaly to recovery in the deformation response during historical rainfall events in the slope area; for example, it is an integer value within the range of 2 to 5 times the time sampling interval. If multiple discontinuous excess intervals exist within the entire time period of the current response curve, the duration of each excess interval is calculated separately.

[0058] Minor deformation sections exceeding both the amplitude and duration of preset discrimination criteria are identified as sensitive enhancement sections. The preset discrimination criteria include a preset over-amplitude threshold and a preset timeout threshold. The preset over-amplitude threshold is obtained by acquiring the rainfall-deformation response sensitivity baseline generated in S2, extracting the maximum deviation of the deformation rate response value relative to the mean response curve at each relative time point in the historical response curves of the minor deformation section under all historical rainfall events, and using this maximum deviation as the preset over-amplitude threshold. The preset timeout threshold is obtained by collecting known landslide cases that have experienced accelerated sliding within the slope area, extracting the time interval from the first abnormal deformation response to the occurrence of accelerated sliding in the known landslide cases, taking the statistical minimum of the time intervals of all known landslide cases as the shortest warning time, and multiplying the shortest warning time by a preset proportional coefficient to obtain the preset timeout threshold. The preset proportional coefficient is set based on a trade-off between warning timeliness and false alarm rate, for example, a percentage value within the range of 30% to 70% of the shortest warning time. For each minor deformation segment, the maximum exceedance value of the current response curve of the minor deformation segment is compared with a preset exceedance threshold, and the maximum duration value of the current response curve of the minor deformation segment is compared with a preset timeout threshold. When the maximum exceedance value of the minor deformation segment is greater than or equal to the preset exceedance threshold, and the maximum duration value is greater than or equal to the preset timeout threshold, the minor deformation segment is identified as a sensitive enhancement segment.

[0059] In S4, the backscattering coefficients of sensitive enhancement segments before and after the current rainfall event are extracted from the synthetic aperture radar (SAR) image sequence. The SAR image sequence is a sequence of single-view complex images covering the slope area acquired from spaceborne SAR in S1, arranged chronologically. The current rainfall event is the rainfall process occurring in the slope area acquired in S3. Before the current rainfall event refers to the time before the start time window of the current rainfall event, and after the current rainfall event refers to the time after the end time window of the current rainfall event. The backscattering coefficient is the quantified value of the back reflection intensity of the ground target to the radar incident wave during SAR imaging. It is converted from the image pixel values ​​in the single-view complex images of the SAR image sequence into backscattering coefficients through radiometric calibration. For each sensitive enhancement segment identified in S3, the backscattering coefficient value of each pixel in the pixel set of a single-view complex image before the current rainfall event is extracted as the backscattering coefficient before and after rainfall, generating a spatial distribution map of the backscattering characteristics before and after rainfall. The backscattering coefficient value of each pixel within a single-view complex image of the sensitive enhancement section after the current rainfall event is extracted as the backscattering coefficient after rainfall, generating a spatial distribution map of the backscattering characteristics after rainfall. Both the spatial distribution map of backscattering characteristics before and after rainfall and the spatial distribution map of backscattering characteristics after rainfall use pixel positions as planar coordinates and backscattering coefficient values ​​as pixel values, expressed in decibels.

[0060] Differential processing is performed on the spatial distribution maps of backscattering characteristics before and after rainfall, and after rainfall, to extract regions in the spatial distribution difference map where the difference values ​​are concentrated and continuously distributed in a planar pattern. The differential processing involves subtracting the backscattering coefficient value of the corresponding pixel location in the spatial distribution maps before and after rainfall from the backscattering coefficient value of each pixel in the spatial distribution map after rainfall. This yields the change in backscattering coefficient for each pixel, and the spatial distribution difference map is constructed from the changes in backscattering coefficients of all pixels. In the spatial distribution difference map, a positive change in backscattering coefficient indicates increased backscattering after rainfall, while a negative change indicates decreased backscattering. Regions with concentrated difference values ​​refer to areas in the spatial distribution difference map where pixels in the backscattering feature spatial distribution difference map have an absolute value of backscattering coefficient change exceeding a preset difference threshold and are clustered together. The preset difference threshold is set by selecting a number of pixels in a stable region outside the sensitive enhancement zone, calculating the statistical standard deviation of the backscattering coefficient changes for these pixels, and taking a multiple of the statistical standard deviation as the preset difference threshold, for example, twice the statistical standard deviation. A region exhibiting a continuous planar distribution refers to a region where the signs of the backscattering coefficient changes are the same and the regions are spatially connected, with the connected area exceeding the preset area threshold. The preset area threshold is set by statistical analysis of the minimum area of ​​the backscattering characteristic change region caused by a single rainfall event in known landslide cases in the slope area; for example, the statistical median of the smallest area in known landslide cases is taken as the preset area threshold.

[0061] The spatial morphology of areas with concentrated, continuous, planar differences is matched with the evolution patterns of backscattering characteristics of existing landslides. Sensitive enhancement sections with a matching degree exceeding a preset matching degree threshold are identified as effective sensitive sections. The evolution patterns of backscattering characteristics of existing landslides are the evolution patterns of the spatial distribution of backscattering characteristics before and after a landslide, extracted from known landslide cases in the slope area. The extraction method for existing backscattering characteristic evolution patterns involves collecting known landslide cases that have experienced accelerated sliding in the slope area, acquiring synthetic aperture radar images of the known landslide cases before and after the landslide event, extracting the spatial distribution map of backscattering characteristics before and after the landslide event, and performing differential processing. The spatial morphology of areas with concentrated, continuous, planar differences in the spatial distribution map of backscattering characteristics is taken as the evolution patterns of existing landslide backscattering characteristics. The spatial morphology includes the outline shape of the areas with concentrated, continuous, planar differences, the distribution gradient direction of the backscattering coefficient changes within the area, and the gradient change rate. The matching process involves calculating the similarity between the spatial morphology of areas with concentrated, continuous, planar differences corresponding to sensitive enhancement sections and the spatial morphology of existing landslide backscattering feature evolution patterns. The similarity calculation method involves extracting the contour shape features and internal gradient distribution features of these areas from the sensitive enhancement sections, comparing them one by one with the contour shape features and internal gradient distribution features of existing landslide backscattering feature evolution patterns, calculating the feature deviation for each feature, and using the weighted sum of these deviations as the matching degree. The weight of each feature is set based on its discriminative power among known landslide cases; features with stronger discriminative power are assigned higher weight values. The preset matching degree threshold is set based on the statistical lower limit of spatial morphological similarity among known landslide cases; for example, the statistical minimum similarity among known landslide cases is used as the preset matching degree threshold. When the matching degree of a sensitive enhancement section is greater than or equal to the preset matching degree threshold, the sensitive enhancement section is determined to be a valid sensitive section.

[0062] During S5 implementation, spaceborne high-resolution synthetic aperture radar (SAR) images and spaceborne high-resolution optical images covering the effective sensitive area are acquired. The effective sensitive area is defined as the area whose spatial morphology, characterized by concentrated and continuously distributed area-like differences in S4, is matched with the evolution pattern of existing landslide backscattering characteristics. Sensitive enhancement areas with a matching degree exceeding a preset matching degree threshold are identified. The acquisition method for spaceborne SAR images involves submitting an imaging request to a spaceborne SAR satellite operating in a sun-synchronous orbit. The imaging request includes the geographic coordinates of the effective sensitive area and imaging mode parameters. The spaceborne SAR satellite images the effective sensitive area in a focused mode, acquiring single-view complex images with a spatial resolution better than the preset high-resolution threshold. The preset high-resolution threshold is set based on the identifiable scale of surface cracks and subtle deformations on the slope, for example, 1 meter is used as the preset high-resolution threshold. The acquisition of high-resolution optical imagery by spaceborne satellites involves submitting an imaging request to a high-resolution optical satellite operating in a sun-synchronous orbit. This request includes the geographic coordinates of the effective sensitive area and the imaging time period. The satellite then performs continuous temporal imaging of the effective sensitive area in stereo pair mode, acquiring multispectral images with a spatial resolution better than a preset optical resolution threshold. The preset optical resolution threshold is set based on the identifiable scale of the cracks on the slope surface, for example, 0.5 meters. Continuous temporal imaging refers to the satellite imaging the same effective sensitive area multiple times within multiple revisit periods, with the interval between each imaging session being the satellite's orbital revisit period.

[0063] Differential interferometry (DI) is applied to spaceborne high-resolution synthetic aperture radar (SAR) imagery to generate a two-dimensional slope deformation field for the effective sensitive section. DI involves interferometry processing of at least two spaceborne SAR images taken at different times, covering the same effective sensitive section. One image is taken after the current rainfall event, and the other is taken before the current rainfall event. The DI process involves using an external digital elevation model (DEM) as a terrain reference, multiplying the main image and the auxiliary image using complex conjugates to generate an interferogram. Then, a differential interferogram is obtained through de-flattening and terrain-removing phase processing. The differential interferogram is then unwrapped to extract the deformation along the radar line-of-sight direction. This deformation along the radar line-of-sight direction is decomposed using the projection relationship between the line-of-sight direction and the horizontal and vertical directions to obtain the horizontal and vertical deformations of each pixel on the slope surface of the effective sensitive section. These horizontal and vertical deformations constitute a two-dimensional slope deformation field in the slope coordinate system. The slope coordinate system uses the slope dip as the vertical axis and the slope strike as the horizontal axis. The two-dimensional deformation field of the slope includes the horizontal and vertical deformation rates of each pixel.

[0064] Image recognition is performed on high-resolution spaceborne optical images across consecutive time phases to extract the location, direction of expansion, and width variation of slope cracks in the effective sensitive section, representing the spatiotemporal distribution of cracks. Continuous time-phase image recognition involves identifying crack features in all time-phase high-resolution spaceborne optical images covering the same effective sensitive section. Crack features include linear features and texture difference features. The identification method involves performing image edge detection and image segmentation on the high-resolution spaceborne optical images of each time phase, extracting linearly continuous image regions with widths exceeding a preset crack width threshold as crack identification results. The preset crack width threshold is set based on the spatial resolution of the high-resolution spaceborne optical images and the identifiability of slope cracks; for example, a multiple of the spatial resolution of the high-resolution spaceborne optical images is used as the preset crack width threshold. The crack identification results of the first time phase are used as the initial crack distribution. The crack identification results of the second and subsequent time phases are spatially compared with the initial crack distribution. Crack identification results that appear in subsequent time phases but are not present in the initial crack distribution are marked as the location of crack formation. Spatially compare the fracture identification results of the same fracture in subsequent time phases with the fracture identification results of the previous adjacent time phase to extract the direction of change of the fracture tip position, and take the direction of change of the tip position as the fracture propagation direction. Arrange the fracture width of the same fracture in each time phase of consecutive time phases in chronological order to extract the process of fracture width change. Combine the fracture formation location, propagation direction, and width change process as the spatiotemporal distribution of the fracture.

[0065] When implementing S6, the location of the deformation concentration region and the deformation rate gradient direction are extracted from the two-dimensional deformation field. The two-dimensional deformation field is the slope two-dimensional deformation field of the effective sensitive section generated in S5 by differential interferometry processing of spaceborne high-resolution synthetic aperture radar imagery, containing the horizontal and vertical deformation rates of each pixel. The deformation concentration region refers to the area in the slope two-dimensional deformation field where pixels with a modulus exceeding a preset deformation rate are clustered together. The preset deformation rate is set by calculating the statistical median of the moduli of the horizontal and vertical deformation rates of all pixels in the slope two-dimensional deformation field, and taking a multiple of the statistical median as the preset deformation rate, for example, 1.5 times or 2 times the statistical median. The location of the deformation concentration region is the set of coordinates of the centroid pixel and the boundary pixel of the deformation concentration region. The deformation rate gradient direction is the direction in which the modulus of the deformation rate changes most steeply within the deformation concentration region. The method for extracting the deformation rate gradient direction is to substitute the horizontal and vertical deformation rates of each pixel in the deformation concentration area into the gradient operator for calculation. The gradient operator takes the difference in deformation rates of adjacent pixels in the horizontal and vertical directions respectively, synthesizes the difference in the horizontal direction and the difference in the vertical direction into a gradient vector, and takes the direction with the largest magnitude of the gradient vector as the deformation rate gradient direction.

[0066] The spatiotemporal distribution of fractures extracts the location of fracture formation and the direction of fracture propagation. The spatiotemporal distribution of fractures is the process of changes in the location of fracture formation, the direction of propagation, and the width of slope fractures in the effective sensitive section extracted through continuous temporal image recognition of high-resolution satellite optical images in S5. The location of fracture formation is the set of pixel coordinates marking the first appearance of a fracture in the spatiotemporal distribution. The direction of fracture propagation is the direction vector of the line connecting the tip positions of the same fracture in different temporal phases within the spatiotemporal distribution. The method for extracting the direction of fracture propagation is to subtract the tip position coordinates of the same fracture in the previous temporal phase from the tip position coordinates in the later temporal phase, obtaining the tip displacement vector, and using the direction of the tip displacement vector as the direction of fracture propagation.

[0067] This study analyzes the spatial angular relationship between the deformation rate gradient direction and the crack propagation direction in deformation concentration zones. The deformation rate gradient direction in the deformation concentration zones is extracted from the two-dimensional deformation field, and the crack propagation direction is extracted from the spatiotemporal distribution of cracks. The spatial angular relationship is analyzed by performing a dot product operation on the deformation rate gradient direction vector and the crack propagation direction vector, and calculating the angle between the two direction vectors, with the angle ranging from 0 degrees to 180 degrees. When multiple deformation concentration zones exist, the angle between the deformation rate gradient direction of each zone and the propagation direction of each crack is calculated separately, and the minimum value among all angles is taken as the spatial angle. When multiple cracks exist, the spatial angle corresponding to each crack is calculated separately, and the minimum value among all spatial angles corresponding to all cracks is taken as the spatial angle of the effective sensitive section.

[0068] This study analyzes the overlap between the deformation rate growth period and the crack width change process in the deformation concentration zone. The deformation rate growth period in the deformation concentration zone refers to the time interval within the two-dimensional deformation field of the slope where the modulus of the deformation rate continuously increases over time. The method for extracting the deformation rate growth period is as follows: using continuous-phase spaceborne high-resolution synthetic aperture radar (SAR) images as the data source, the modulus of the deformation rate within the deformation concentration zone is extracted from the two-dimensional deformation field of the slope in each phase. The modulus values ​​are arranged chronologically, and a continuous time period in which the modulus monotonically increases and the increase exceeds a preset growth threshold is extracted as the deformation rate growth period. The preset growth threshold is set by calculating the standard deviation of the modulus of the deformation rate in the stable region of the two-dimensional deformation field of the slope over time, and taking a multiple of the standard deviation as the preset growth threshold, for example, twice the standard deviation. The crack width change process is the crack width value sequence of the same crack in each phase of the crack spatiotemporal distribution extracted in S5. The time interval during which the crack width increases with time during the crack width change process is called the crack width growth period. The method for analyzing the overlap relationship of these periods is to compare the start and end times of the deformation rate growth period with the start and end times of the crack width growth period, and calculate the proportion of the time overlap portion of the two periods to the total duration of the deformation rate growth period, which is taken as the time overlap ratio. When there are multiple deformation concentration zones and multiple cracks, the maximum value among all time overlap ratios is taken as the time overlap ratio of the effective sensitive section.

[0069] When the spatial angle is less than a preset angle threshold and the time period overlap ratio exceeds a preset overlap threshold, the effective sensitive section is assessed as having a high accelerated slip risk level. The preset angle threshold and preset overlap threshold are determined by obtaining the two-dimensional deformation field and spatiotemporal distribution of fractures from known landslide cases within the slope area. Known landslide cases are landslide events within the slope area that have been confirmed to have experienced accelerated slip. The deformation rate gradient direction of the deformation concentration area of ​​the known landslide cases is extracted from the two-dimensional deformation field of the known landslide cases, and the fracture propagation direction of the known landslide cases is extracted from the spatiotemporal distribution of fractures in the known landslide cases. The spatial angle between the deformation rate gradient direction of the deformation concentration area of ​​the known landslide cases and the fracture propagation direction of the known landslide cases is calculated. Multiple known landslide cases are collected within the slope area, the spatial angle of each known landslide case is calculated, and the maximum value among all spatial angles is taken as the preset angle threshold. This study extracts the deformation rate growth periods of the deformation concentration areas from the two-dimensional deformation fields of known landslide cases, and extracts the crack width variation processes from the spatiotemporal distribution of cracks in known landslide cases. The study then calculates the time overlap ratio between the deformation rate growth periods of the deformation concentration areas and the crack width variation processes of the known landslide cases. Multiple known landslide cases are collected within the slope area, and the time overlap ratio for each case is calculated. The minimum value among all time overlap ratios is taken as the preset overlap threshold.

[0070] Example 2: Figure 2 A schematic diagram of a spatiotemporal monitoring system for slope deformation based on multi-source remote sensing data is provided according to the present invention. The system includes the following modules:

[0071] The data acquisition module is used to acquire synthetic aperture radar image sequences and rainfall data of the slope area to delineate areas with minor deformation.

[0072] The baseline construction module is used to extract the deformation response sequence of each small deformation segment under historical rainfall events and construct a rainfall-deformation response sensitivity baseline based on the deformation rate before rainfall.

[0073] The segment identification module is used to input the current rainfall event into the rainfall-deformation response sensitivity baseline, determine whether there are small deformation segments whose deformation response sensitivity is enhanced compared to the same period in history, and identify them as sensitive enhanced segments.

[0074] The section determination module is used to extract the spatial distribution of synthetic aperture radar backscattering features of sensitive enhanced sections before and after the current rainfall event, analyze the changing trend of the spatial distribution of backscattering features, and determine the effective sensitive sections that match the changing trend with the evolution law of existing landslide backscattering features.

[0075] The deformation extraction module is used to acquire high-resolution synthetic aperture radar and optical images of the effective sensitive section, and to extract the two-dimensional deformation field and spatiotemporal distribution of cracks on the slope of the effective sensitive section.

[0076] The risk assessment module is used to assess the risk level of accelerated slip in the effective sensitive section based on the two-dimensional deformation field and spatiotemporal distribution of the fracture.

[0077] All calculations involved in the embodiments are dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0078] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0079] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and inventive constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0080] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0081] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0082] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0083] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for spatiotemporal monitoring of slope deformation based on multi-source remote sensing data, characterized in that, Includes the following steps: S1: Acquire synthetic aperture radar image sequences and rainfall data for the slope area to delineate areas with minor deformation; S2: Extract the deformation response sequence of each minor deformation segment under historical rainfall events, and construct a rainfall-deformation response sensitivity baseline based on the deformation rate before rainfall. This includes: for each minor deformation segment, extracting the change process of deformation rate over time within the start and end time window of the historical rainfall event, and using the difference sequence between the deformation rate at each moment within the start and end time window and the deformation rate before rainfall as the deformation response sequence; arranging the deformation response sequence according to the time sequence of the historical rainfall events, generating historical response curves for each minor deformation segment under different historical rainfall events, and using the mean envelope of the historical response curves as the rainfall-deformation response sensitivity baseline. S3: Input the current rainfall event into the rainfall-deformation response sensitivity baseline, determine whether there are small deformation segments with enhanced deformation response sensitivity compared to the same period in history, and identify them as sensitive enhanced segments; S4: Extract the spatial distribution of synthetic aperture radar backscattering features of sensitive enhancement sections before and after the current rainfall event, analyze the changing trend of the spatial distribution of backscattering features, and determine the effective sensitive sections by matching the changing trend with the evolution law of existing landslide backscattering features. S5: Acquire high-resolution synthetic aperture radar and optical images of the effective sensitive section, and extract the two-dimensional deformation field and spatiotemporal distribution of cracks on the slope of the effective sensitive section. S6: Based on the two-dimensional deformation field and spatiotemporal distribution of cracks in the effective sensitive section, assess the risk level of accelerated slip in the effective sensitive section.

2. The method for spatiotemporal monitoring of slope deformation based on multi-source remote sensing data according to claim 1, characterized in that, Acquiring synthetic aperture radar (SAR) image sequences and rainfall data for the slope area, and delineating areas of minor deformation, includes: acquiring SAR image sequences covering the slope area from spaceborne SAR and acquiring rainfall data for the slope area from a ground-based meteorological network; performing differential interferometry (DIAC) on the SAR image sequences to generate a differential interferogram sequence and deformation rate map for the slope area, and extracting the changes in deformation rate and cumulative deformation of each pixel in the slope area over time; performing time-series analysis on the rainfall data to identify the start and end time windows of historical rainfall events and the rainfall intensity distribution within each time window; and delineating the set of pixels within the slope area whose deformation rate is below a preset rate threshold and whose cumulative deformation shows slight fluctuations over time as areas of minor deformation.

3. The method for spatiotemporal monitoring of slope deformation based on multi-source remote sensing data according to claim 1, characterized in that, The mean envelope of the historical response curve is used as the baseline for rainfall-deformation response sensitivity. This includes: calculating the mean value of the deformation rate response value of each small deformation segment at the same relative time under different historical rainfall events, and connecting the mean values ​​at each relative time to form a mean response curve; taking the upper and lower envelope values ​​of the deformation rate response value at each relative time point on the mean response curve, connecting the upper envelope values ​​to form an upper envelope line, and connecting the lower envelope values ​​to form a lower envelope line, and using the upper and lower envelope lines as the envelope range of the rainfall-deformation response sensitivity baseline.

4. The method for spatiotemporal monitoring of slope deformation based on multi-source remote sensing data according to claim 1, characterized in that, The current rainfall event is input into the rainfall-deformation response sensitivity baseline. The system identifies any small deformation segments with enhanced deformation response sensitivity compared to historical data for the same period, classifying them as enhanced sensitive segments. This process includes: acquiring the start and end time windows and rainfall intensity distribution of the current rainfall event; extracting the deformation rate change over time for each small deformation segment within the start and end time windows of the current rainfall event; generating the current response curve; comparing the current response curve with the rainfall-deformation response sensitivity baseline; calculating the amplitude and duration by which the current response curve exceeds the envelope of the rainfall-deformation response sensitivity baseline; and identifying small deformation segments whose amplitude and duration exceed preset discrimination conditions as enhanced sensitive segments.

5. The method for spatiotemporal monitoring of slope deformation based on multi-source remote sensing data according to claim 1, characterized in that, The process involves extracting the spatial distribution of synthetic aperture radar (SAR) backscattering features of sensitive enhancement sections before and after the current rainfall event, analyzing the changing trends of the spatial distribution of backscattering features, and identifying effective sensitive sections as those whose changing trends match the evolution patterns of existing landslide backscattering features. This includes: extracting the backscattering coefficients of sensitive enhancement sections before and after the current rainfall event from SAR image sequences to generate spatial distribution maps of backscattering features before and after rainfall; performing differential processing on the spatial distribution maps of backscattering features before and after rainfall to extract areas where the difference values ​​are concentrated and continuously distributed in a planar pattern; matching the spatial morphology of these areas with concentrated difference values ​​and continuous planar distribution with the evolution patterns of existing landslide backscattering features, and identifying sensitive enhancement sections with a matching degree exceeding a preset condition as effective sensitive sections.

6. The method for spatiotemporal monitoring of slope deformation based on multi-source remote sensing data according to claim 1, characterized in that, Acquiring high-resolution synthetic aperture radar (SAR) and optical images of the effective sensitive section, and extracting the two-dimensional deformation field and spatiotemporal distribution of the slope surface and cracks in the effective sensitive section, includes: acquiring spaceborne high-resolution SAR and optical images covering the effective sensitive section; performing differential interferometry processing on the spaceborne high-resolution SAR images to generate the two-dimensional deformation field of the slope surface in the effective sensitive section; and performing continuous temporal image recognition on the spaceborne high-resolution optical images to extract the location of crack formation, propagation direction, and width variation process of the slope surface cracks in the effective sensitive section as the spatiotemporal distribution of cracks.

7. The method for spatiotemporal monitoring of slope deformation based on multi-source remote sensing data according to claim 1, characterized in that, Based on the two-dimensional deformation field and spatiotemporal distribution of fractures in the effective sensitive section, the risk level of accelerated slip in the effective sensitive section is assessed, including: extracting the location of the deformation concentration area and the deformation rate gradient direction from the two-dimensional deformation field, and extracting the location of fracture formation and the direction of fracture propagation from the spatiotemporal distribution of fractures; analyzing the spatial angle relationship between the deformation rate gradient direction of the deformation concentration area and the fracture propagation direction, and simultaneously analyzing the time overlap relationship between the deformation rate growth period of the deformation concentration area and the fracture width change process; when the spatial angle is less than a preset angle threshold and the time overlap ratio exceeds a preset overlap threshold, the effective sensitive section is assessed as having a high accelerated slip risk level.

8. The method for spatiotemporal monitoring of slope deformation based on multi-source remote sensing data according to claim 7, characterized in that, The determination of the preset angle threshold and the preset overlap threshold includes: obtaining the two-dimensional deformation field and spatiotemporal distribution of cracks in known landslide cases within the slope area; extracting the spatial angle between the deformation rate gradient direction and the crack propagation direction in the deformation concentration area of ​​the known landslide cases; and the time overlap ratio between the deformation rate growth period and the crack width change process in the deformation concentration area; and using the maximum value of the spatial angle and the minimum value of the time overlap ratio in the known landslide cases as the preset angle threshold and the preset overlap threshold, respectively.

9. A spatiotemporal monitoring system for slope deformation based on multi-source remote sensing data, used to implement the spatiotemporal monitoring method for slope deformation based on multi-source remote sensing data as described in any one of claims 1-8, characterized in that, Includes the following modules: The data acquisition module is used to acquire synthetic aperture radar image sequences and rainfall data of the slope area and delineate areas with minor deformations. The baseline construction module is used to extract the deformation response sequence of each small deformation segment under historical rainfall events and construct a rainfall-deformation response sensitivity baseline based on the deformation rate before rainfall. The segment identification module is used to input the current rainfall event into the rainfall-deformation response sensitivity baseline, determine whether there are small deformation segments whose deformation response sensitivity is enhanced compared to the same period in history, and identify them as sensitive enhanced segments. The section determination module is used to extract the spatial distribution of synthetic aperture radar backscattering features of sensitive enhanced sections before and after the current rainfall event, analyze the changing trend of the spatial distribution of backscattering features, and determine the effective sensitive sections that match the changing trend with the evolution law of existing landslide backscattering features. The deformation extraction module is used to acquire high-resolution synthetic aperture radar and optical images of the effective sensitive section, and extract the two-dimensional deformation field and spatiotemporal distribution of cracks on the slope of the effective sensitive section. The risk assessment module is used to assess the risk level of accelerated slip in the effective sensitive section based on the two-dimensional deformation field and spatiotemporal distribution of the fracture.

Images