A landslide maneuverability confirmation method and system based on an adaptive difference model

Abstract

The application discloses a landslide mobility confirmation method and system based on an adaptive difference model, and relates to the technical field of data processing.The method comprises the following steps: obtaining a sub-slope area, and obtaining a vegetation coverage data sequence; obtaining a difference index, obtaining a first to-be-processed data sequence, obtaining a second to-be-processed data sequence, obtaining a difference sequence of the first to-be-processed data sequence, obtaining a first fluctuation index, and if the first fluctuation index exceeds a third preset threshold value, regarding the sub-slope area corresponding to the first to-be-processed data sequence as a sliding source area boundary area; obtaining a difference sequence of the second to-be-processed data sequence, obtaining a second fluctuation index, and if the second fluctuation index exceeds a fourth preset threshold value, regarding the sub-slope area corresponding to the second to-be-processed data sequence as a landslide area boundary area; and obtaining a target sliding source area, a target landslide area and an area ratio.The application has the advantages of adaptive difference distinguishing, stable and accurate boundary and dynamic quantification.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a landslide mobility confirmation method and system based on an adaptive difference model. BACKGROUND

[0002] The mobility of landslide disaster (i.e. the movement and diffusion ability of landslide body) directly determines the damage range and secondary disaster risk (such as blocking river channels and impacting infrastructure). However, the existing landslide mobility confirmation method and technology cannot quickly and accurately distinguish the sliding source area (initial rock mass fracture area) and the landslide area (overall movement coverage area), often leading to misjudgment of the dynamic characteristics of the landslide due to ambiguous boundary recognition. Secondly, the existing threshold method usually uses a single static threshold to determine the change in vegetation coverage, but the sliding source area causes a sharp drop in vegetation coverage (strong negative change) due to the violent fracture of the rock mass, while the landslide area presents a mild coverage attenuation (weak negative change) due to the overall displacement of the surface soil, and the difference in change characteristics between the two is not effectively quantified, leading to underestimation of the sliding source area or confusion with the landslide area. Furthermore, the landslide boundary is affected by the geological structure and adjacent areas, and the existing algorithm ignores the spatial correlation of the boundary area and the surrounding data (such as the dramatic fluctuations in boundary vegetation coverage), and relying only on point data analysis can easily miss the true boundary, resulting in area calculation bias. SUMMARY

[0003] In view of the technical problems in the prior art that the sliding source area and the landslide area are not accurately divided, reliable landslide diffusion ability evaluation indexes (such as area ratio) cannot be constructed, and the mobility characteristics of the landslide from local fracture to large-scale diffusion cannot be quantitatively evaluated, which seriously restricts the accuracy of disaster warning and the effectiveness of prevention and control decision-making, the present application provides a landslide mobility confirmation method and system based on an adaptive difference model.

[0004] A landslide mobility confirmation method based on an adaptive difference model includes: acquiring remote sensing image information within the acquisition period of the time point when a landslide occurs on a target slope, dividing the target slope into multiple sub-slope regions, and acquiring vegetation cover data sequences for each sub-slope region based on the remote sensing image information; obtaining difference indices based on adjacent vegetation cover data in the vegetation cover data sequences, acquiring vegetation cover data sequences with difference indices less than a first preset threshold as first data sequences to be processed, acquiring vegetation cover data sequences with difference indices less than a second preset threshold as second data sequences to be processed, wherein the first preset threshold is less than the second preset threshold; obtaining a difference sequence between the vegetation cover data sequences of adjacent sub-slope regions corresponding to the i-th first data sequence and the i-th first data sequence to be processed, and then... A first fluctuation index is obtained from the difference sequence of the data sequences to be processed. If the first fluctuation index exceeds a third preset threshold, the sub-slope region corresponding to the i-th first data sequence to be processed is taken as the boundary region of the landslide source area. The difference sequence of the j-th second data sequence to be processed is obtained from the vegetation cover data sequence of the sub-slope region adjacent to the j-th second data sequence to be processed and the j-th second data sequence to be processed. A second fluctuation index is obtained from the difference sequence of the j-th second data sequence to be processed. If the second fluctuation index exceeds a fourth preset threshold, the sub-slope region corresponding to the j-th second data sequence to be processed is taken as the boundary region of the landslide area. The target landslide source area area is obtained from the multiple landslide source area boundary regions, and the target landslide area area is obtained from the multiple landslide area boundary regions. The target landslide source area area is divided by the target landslide area area to obtain the area ratio used to indicate landslide mobility.

[0005] Optionally, obtaining the difference index based on adjacent vegetation cover data in the vegetation cover data sequence includes: subtracting the previous vegetation cover data from the next vegetation cover data in the vegetation cover data sequence to obtain the difference amount; and dividing the smallest difference amount in the vegetation cover data sequence by the average value of all vegetation cover data in the vegetation cover data sequence to obtain the difference index.

[0006] Optionally, the difference sequence of the i th first to-be-processed data sequence is obtained according to the vegetation coverage data sequence of the sub-slope region adjacent to the sub-slope region corresponding to the i th first to-be-processed data sequence, and the vegetation coverage data sequence of the i th first to-be-processed data sequence, and the difference sequence of the i th first to-be-processed data sequence is obtained by: obtaining the vegetation coverage data of each collection point in the vegetation coverage data sequence of the sub-slope region adjacent to the sub-slope region corresponding to the i th first to-be-processed data sequence, obtaining the vegetation coverage data of each collection point in the i th first to-be-processed data sequence, and obtaining the absolute value of the difference between the vegetation coverage data of the same collection point in the two vegetation coverage data sequences, arranging all the absolute values of the differences in time sequence and forming the difference sequence of the i th first to-be-processed data sequence.

[0007] Optionally, the first fluctuation index is obtained according to the difference sequence of the i th first to-be-processed data sequence, and the first fluctuation index is obtained by: obtaining the average value of all data in the difference sequence of the i th first to-be-processed data sequence, subtracting the average value from each data in the difference sequence to obtain a unit fluctuation amount, and adding the absolute values of all unit fluctuation amounts in the difference sequence of the i th first to-be-processed data sequence to obtain the first fluctuation index.

[0008] Optionally, the target sliding source area is obtained according to the plurality of sliding source boundary regions, and the target sliding source area is obtained by: connecting the plurality of adjacent sliding source boundary regions to form a first boundary extension path, connecting the head and tail ends of the first boundary extension path in a straight line to enclose the area surrounded by the first boundary extension path as the target sliding source area, and obtaining the area of all sub-slope regions occupied by the target sliding source area as the target sliding source area.

[0009] Optionally, the target sliding source area is obtained according to the plurality of sliding source boundary regions, and the target sliding source area is obtained by: connecting the plurality of adjacent sliding source boundary regions to form a first boundary extension path, connecting the head and tail ends of the first boundary extension path in a straight line to enclose the area surrounded by the first boundary extension path as the target sliding source area, and obtaining the area of all sub-slope regions occupied by the target sliding source area as the target sliding source area.

[0010] Also provided is a landslide mobility confirmation system based on an adaptive difference model, the system comprising: an acquisition module configured to acquire remote sensing image information in a collection period in which a target slope occurs at a time point of a landslide, divide the target slope into a plurality of sub-slope regions, and acquire a vegetation coverage data sequence of each sub-slope region according to the remote sensing image information; a first data processing module configured to acquire a difference index according to adjacent vegetation coverage data in the vegetation coverage data sequence, acquire vegetation coverage data sequences with a difference index less than a first preset threshold as first to-be-processed data sequences, and acquire vegetation coverage data sequences with a difference index less than a second preset threshold as second to-be-processed data sequences, wherein the first preset threshold is less than the second preset threshold; a second data processing module configured to acquire a difference sequence of the i th first to-be-processed data sequence according to vegetation coverage data sequences of adjacent sub-slope regions of a sub-slope region corresponding to the i th first to-be-processed data sequence and the i th first to-be-processed data sequence, and acquire a first fluctuation index according to the difference sequence of the i th first to-be-processed data sequence, wherein if the first fluctuation index exceeds a third preset threshold, the sub-slope region corresponding to the i th first to-be-processed data sequence is taken as a sliding source area boundary region; a third data processing module configured to acquire a difference sequence of the j th second to-be-processed data sequence according to vegetation coverage data sequences of adjacent sub-slope regions of a sub-slope region corresponding to the j th second to-be-processed data sequence and the j th second to-be-processed data sequence, and acquire a second fluctuation index according to the difference sequence of the j th second to-be-processed data sequence, wherein if the second fluctuation index exceeds a fourth preset threshold, the sub-slope region corresponding to the j th second to-be-processed data sequence is taken as a landslide area boundary region; and a mobility confirmation module configured to acquire a target sliding source area according to a plurality of sliding source area boundary regions, acquire a target landslide area according to a plurality of landslide area boundary regions, and divide the target sliding source area by the target landslide area to obtain an area ratio for indicating landslide mobility.

[0011] Optionally, the first data processing module is further configured to: subtract a previous vegetation coverage data from a next vegetation coverage data of the adjacent vegetation coverage data in the vegetation coverage data sequence to obtain a difference amount; and divide a minimum difference amount in the vegetation coverage data sequence by an average value of all vegetation coverage data in the vegetation coverage data sequence to obtain the difference index.

[0012] Optionally, the second data processing module is further configured to: acquire vegetation coverage data of each collection point in the vegetation coverage data sequence of the adjacent sub-slope region of the sub-slope region corresponding to the i th first to-be-processed data sequence, acquire vegetation coverage data of each collection point in the i th first to-be-processed data sequence, and acquire an absolute value of a difference between the vegetation coverage data of the same collection point in the two vegetation coverage data sequences; and arrange all absolute values of the differences in a time sequence to form the difference sequence of the i th first to-be-processed data sequence.

[0013] Optionally, the second data processing module is further configured to: obtain the average value of all data in the difference sequence of the i-th first to-be-processed data sequence, subtract the average value from each data in the difference sequence to obtain unit fluctuation quantity; and add the absolute values of all unit fluctuation quantities in the difference sequence of the i-th first to-be-processed data sequence to obtain the first fluctuation index.

[0014] The beneficial effects of the present application are embodied in:

[0015] In the whole landslide mobility confirmation method based on the adaptive difference model, firstly, a double-threshold dynamic judgment mechanism is adopted, a normalized difference index is calculated according to the adjacent change characteristics of the vegetation coverage data sequence, and a strict first preset threshold and a relatively loose second preset threshold are set respectively, so as to accurately separate two types of regions with different damage characteristics in space, and reduce the region misjudgment caused by the existing single-threshold method; secondly, a spatial neighborhood fluctuation analysis model is innovatively introduced, a difference sequence is formed by calculating the absolute difference values of the vegetation coverage data of the candidate sub-region and the adjacent region at all time points, and the first and second fluctuation indexes are constructed based on the average value of the sequence and the absolute sum of the deviation of each point, and the instability characteristics in the time dimension are used to identify the real boundary region; finally, a differentiated area extraction mechanism based on boundary characteristics is established: for the sliding source area, the discrete boundary points are connected and then enclosed by a straight line to form a non-significant boundary (because it is wrapped in the landslide body), and for the landslide area, the discrete boundary points are directly connected to form a natural closed path, and then the total area of the sub-region in the boundary is counted to obtain the accurate division range, and finally the area ratio of the sliding source area to the landslide area is taken as the mobility parameter (a small ratio means that the rupture core area is small but the diffusion range is large, which represents strong mobility), realizing the whole-link quantitative evaluation from dynamic change capture, spatial boundary identification to disaster movement diffusion ability, and greatly improving the disaster warning accuracy and the scientificity of the prevention and control strategy. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the drawings needed in the specific embodiments or the prior art description will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, each element or part is not necessarily drawn according to the actual proportion.

[0017] Figure 1 A step schematic diagram of the landslide mobility confirmation method based on the adaptive difference model of the present application;

[0018] Figure 2 A part of step schematic diagram of S2 in the landslide mobility confirmation method based on the adaptive difference model of the present application;

[0019] Figure 3This is a schematic diagram of part of step S3 in the landslide mobility verification method based on the adaptive difference model of the present invention;

[0020] Figure 4 This is a schematic diagram of part of step S5 in the landslide mobility verification method based on the adaptive difference model of the present invention;

[0021] Figure 5 This is a schematic diagram of another step in S5 of the landslide mobility verification method based on the adaptive difference model of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0023] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0024] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0025] like Figure 1 As shown, a landslide mobility verification method based on an adaptive difference model is provided. In one embodiment, the method includes:

[0026] S1. Obtain remote sensing image information within the acquisition period of the time point when the landslide occurred on the target slope, divide the target slope into multiple sub-slope areas, and obtain the vegetation cover data sequence of each sub-slope area based on the remote sensing image information.

[0027] S2. Obtain a difference index based on adjacent vegetation cover data in the vegetation cover data sequence, obtain a vegetation cover data sequence with a difference index less than a first preset threshold and use it as the first data sequence to be processed, obtain a vegetation cover data sequence with a difference index less than a second preset threshold and use it as the second data sequence to be processed, wherein the first preset threshold is less than the second preset threshold.

[0028] S3, obtaining a difference sequence of the i-th first to-be-processed data sequence according to the vegetation coverage data sequence of the sub-slope region adjacent to the sub-slope region corresponding to the i-th first to-be-processed data sequence and the i-th first to-be-processed data sequence, and obtaining a first fluctuation index according to the difference sequence of the i-th first to-be-processed data sequence, if the first fluctuation index exceeds a third preset threshold, regarding the sub-slope region corresponding to the i-th first to-be-processed data sequence as a sliding source area boundary region;

[0029] S4, obtaining a difference sequence of the j-th second to-be-processed data sequence according to the vegetation coverage data sequence of the sub-slope region adjacent to the sub-slope region corresponding to the j-th second to-be-processed data sequence and the j-th second to-be-processed data sequence, and obtaining a second fluctuation index according to the difference sequence of the j-th second to-be-processed data sequence, if the second fluctuation index exceeds a fourth preset threshold, regarding the sub-slope region corresponding to the j-th second to-be-processed data sequence as a sliding source area boundary region;

[0030] S5, obtaining a target sliding source area according to a plurality of sliding source area boundary regions, and obtaining a target sliding area according to a plurality of sliding area boundary regions, dividing the target sliding source area by the target sliding area to obtain an area ratio for indicating sliding mobility.

[0031] In the embodiment, it should be noted that in S1, high-quality and structured input information is provided for subsequent accurate identification of sliding source area and sliding area boundary and calculation of area ratio. First, a plurality of time-series remote sensing image data covering the target slope region and within a complete acquisition period when the landslide event occurs needs to be obtained. This acquisition period is crucial, and it must include relatively stable state images before the landslide occurs and result images that can clearly reflect changes after the landslide occurs, ensuring that the dynamic change process of vegetation coverage caused by the landslide can be captured (which is a prerequisite for the difference index calculation and to-be-processed sequence screening in S2). The selection of remote sensing image information is critical, and it should be able to effectively retrieve vegetation coverage conditions (for example, through normalized vegetation index, etc.), and have sufficient spatial resolution and time resolution to clearly distinguish subtle changes on the slope and track their time trajectory.

[0032] After the image acquisition, the next step is to divide the whole target slope region into several continuous and regular-shaped (e.g. grid) sub-slope regions. The accuracy of this division directly affects the accuracy of the final boundary identification and area calculation: too coarse division may blur the subtle difference between the source area and the landslide area, making it difficult to filter out specific sub-regions with significant negative changes (i.e. the first and second data sequences to be processed) in S2; too fine division may introduce noise. The purpose of reasonable division is to assign each spatial unit its own vegetation change feature sequence, laying the foundation for subsequent detailed analysis and spatial neighborhood comparison (difference sequence calculation in S3 and S4).

[0033] Further, after successfully structuring the target slope into multiple sub-slope regions and obtaining high-quality multi-temporal remote sensing images, the next key operation in S1 is to extract and construct the vegetation cover data sequence for each sub-slope region. This process requires using remote sensing image information to extract the quantitative value of vegetation cover for each independent sub-slope region within the entire acquisition period (including multiple equally spaced or unequally spaced acquisition time points). This value is a comprehensive representation of the surface vegetation condition (e.g. vegetation index), reflecting the density or health status changes over time. Specifically, each vegetation cover data in the sequence is constructed from the visible red band and near-infrared band reflectance values, with the core expression being vegetation cover data = (near-infrared band reflectance - red band reflectance) / (near-infrared band reflectance + red band reflectance).

[0034] The essence of constructing this data sequence is to depict a "trajectory line" reflecting the vegetation cover fluctuation or trend for each spatial unit (sub-region) in the time dimension. The integrity and quality of this sequence are crucial: it needs to be able to sensitively capture the sudden drop in vegetation cover (possibly corresponding to the severe destruction of the source area) or gradual drop (possibly corresponding to the overall movement of the landslide area) caused by landslide events. After completing this step, each sub-slope region has a dedicated data sequence representing the time evolution of its vegetation cover during the key acquisition period. This sequence is directly used as input for S2 to calculate the changes between adjacent time points (difference indicators) and filter out potential candidate areas (first and second data sequences to be processed).

[0035] In S2, change analysis is performed on each sub-slope region vegetation coverage data sequence generated in S1, aiming to identify sub-regions where vegetation coverage has significantly decreased due to landslide events, and preliminarily classify them according to the intensity of the decrease, laying the foundation for subsequent accurate distinction between the source area and the landslide area. The entire S2 calculates the vegetation coverage data sequence for each sub-region, and the calculation process mainly focuses on the difference between consecutive time points. Specifically, it compares the vegetation coverage value at the later time point with the value at the earlier time point in the sequence, obtaining a set of single time interval changes (referred to as difference values). If a landslide occurs (especially when the landslide mass moves and destroys the surface vegetation), these difference values often show negative values (indicating a decrease in vegetation coverage). In order to quantitatively characterize the maximum degree of vegetation coverage decay in the entire key collection period (which is usually most relevant to the intensity of damage caused by landslides), S2 selects the difference value with the smallest value in the set of difference values (i.e., the most negative change).

[0036] However, in order to avoid the influence of the original vegetation coverage level on the absolute change value (for example, a high coverage area may have a large absolute value even if the change value is large), and to ensure the comparability of the index, the minimum difference value needs to be normalized. This is achieved by dividing the minimum difference value by the average coverage level of all vegetation coverage data in the entire collection period for the sub-region. The final result obtained in this way is the difference index, which is a relative change intensity measure that combines the maximum degree of decline and the average coverage background of the region. The negative value of the index directly represents the severity of the vegetation damage caused by landslides: the more negative (i.e., the larger the negative value), the more severe the sudden drop in vegetation coverage relative to its own average level at a certain time point.

[0037] Further, after obtaining the difference index for each sub-slope region, S2 uses the first and second preset thresholds to discriminate and screen the index, thereby preliminarily identifying two sets of candidate sub-regions with different landslide impact characteristics. The first and second preset thresholds are used to distinguish between source area (strong damage) and landslide area (weak damage) candidate regions; the data is based on limited historical landslide event data, and the difference index distribution characteristics of the two types of regions are analyzed. In the source area, the vegetation coverage decreases suddenly due to rock mass fracture, and the difference index shows very low negative values (such as -0.8 to -0.4); in the landslide area, the difference index shows relatively moderate negative values (such as -0.3 to -0.1) due to the overall displacement of the soil. The first preset threshold (source area threshold) is set to the upper boundary of the difference index distribution of the source area, for example, if 95% of the difference indexes of the source area in the historical data are ≤-0.5, then the first preset threshold is set to -0.5. The second preset threshold (landslide area threshold) is set to the significant quantile of the difference index of the landslide area, for example, if the lowest negative value covering 90% of the landslide area data is -0.2, then the first preset threshold is set to -0.2.

[0038] Firstly, it needs to find out those sub-regions whose difference indicators are less than a first preset threshold. Since the difference indicator is a negative value representing the intensity of the destruction, a smaller (i.e. more negative) value means a stronger and more violent vegetation destruction process. The first preset threshold is set as a relatively strict (numerically more negative) threshold, aiming to filter out sub-regions with very small difference indicators (i.e. the areas where the vegetation coverage has decreased most drastically). As known from the background, the source area as the core area of the initial rupture of the rock mass, its vegetation is often destroyed or buried instantly due to the violent fracture of the rock mass, resulting in a very significant cliff-like drop in the vegetation coverage data sequence, so its difference indicator will present a very small negative value (such as -0.5). Through the filtering of this first preset threshold, a set of candidate sub-regions with extremely violent vegetation destruction can be obtained, and these regions are considered as potential source areas, and their corresponding vegetation coverage data sequences are classified as the first to-be-processed data sequences.

[0039] Secondly, S2 step finds out those sub-regions whose difference indicators are less than a second preset threshold. The second preset threshold value is larger than the first preset threshold value (i.e. the negative degree is slightly weaker), so it is a relatively loose threshold. The threshold aims to capture those areas where the vegetation coverage has also decreased, but the decrease is not so extreme and is relatively moderate (such as an indicator of -0.2). In the landslide event, in addition to the core rupture source area, the overall coverage or disturbance area of the landslide body during the movement (i.e. the landslide area) will also cause the vegetation coverage to decrease, but the surface soil of the area may not be completely crushed, and the vegetation may be pushed, buried or partially damaged, and its coverage decay is manifested as a slow decline rather than an instantaneous collapse. Therefore, the sub-regions that meet the second threshold (but usually do not meet the first threshold) are considered as potential landslide areas (including the sliding coverage area other than the source area), and their sequences are classified as the second to-be-processed data sequences. Through this double-threshold strategy, S2 step effectively distinguishes two sets of candidate regions with different vegetation destruction patterns (the first set represents extremely strong destruction, and the second set represents relatively strong destruction), significantly improving the pertinence, and providing an explicit and clearly classified input data basis for the subsequent S3 and S4 steps to focus on more accurately identifying the source area boundary and the landslide area boundary, respectively.

[0040] In S3, the boundary area of the source area is accurately located in the first to-be-processed data sequence (representing the set of sub-regions of the potential source area with the most violent vegetation destruction) preliminarily screened out in S2. Since the sub-regions on the true boundary of the source area, the time variation pattern of the vegetation coverage will have a significant spatial discontinuity (or spatial difference) with the sub-regions adjacent to it which have not been destroyed violently. This difference is reflected in the comparison of the vegetation coverage data sequences of the two regions in the entire acquisition period.

[0041] Therefore, this step first needs to build a key difference sequence for each set of first to-be-processed data sequences (denoted as the ith, corresponding to a specific potential landslide source area sub-region). The construction process focuses on the calculation of spatial differences in the time dimension. Specifically, it is necessary to obtain the vegetation cover data sequence of all sub-slope regions spatially adjacent (i.e. directly bordering in geographical location) to the ith target sub-region. Then, at each same time point (collection point), the absolute difference between the vegetation cover data of these adjacent sub-regions and the vegetation cover data of the ith sub-region itself is calculated respectively. The reason for using absolute value is to avoid the mutual offset of positive and negative differences, and to simply quantify the difference degree of the two at any time.

[0042] Finally, arranging these absolute value differences calculated in time point order in turn forms a difference sequence specifically for describing the inconsistency degree of the ith target sub-region and its surrounding regions in vegetation cover change. The length of this sequence is the same as the original vegetation cover sequence, and each element of it represents how much deviation occurs between the target region and its neighbors in vegetation state at the corresponding time point. The difference sequence essentially depicts the deviation degree of the target sub-region from its neighborhood environment in the time trajectory in the dynamic evolution process of vegetation cover.

[0043] Further, after building the difference sequence reflecting the difference characteristics of the ith potential landslide source area sub-region and its surrounding regions, step S3 further quantitatively evaluates the fluctuation characteristics of this difference in time sequence to determine whether the sub-region is indeed at the boundary of the landslide source area. The key logic here is that a sub-region that is truly at the boundary position is simultaneously affected by the intense destruction (vegetation sudden drop) within the landslide source area and the relative stability (or non-destructive change) of the external region, and its difference value with the neighborhood vegetation cover will show strong instability or oscillation throughout the collection period - it may be very different at some time points (especially during the dramatic change period before and after the landslide), and it may be smaller at other time points. In order to quantify the intensity of this oscillation, step S3 designs a calculation method for the first fluctuation index.

[0044] Further, first, the average value of all elements of the sequence is obtained to determine the "baseline" level of the difference value over time; then, each element in the sequence is subtracted from this average value to obtain the unit fluctuation amount at each time point relative to the average level. The absolute value of the unit fluctuation amount reflects the magnitude of the difference from the average state at the time point. Finally, the absolute values of all unit fluctuation amounts are summed to obtain the final first fluctuation index. The higher the index value, the greater and more intense the fluctuation of the entire difference sequence around its mean, and the poorer the temporal stability. Such high volatility is the most likely characteristic of the boundary region of the sliding source area: before and after the landslide, the vegetation cover originally similar to the neighborhood will quickly differentiate and repeatedly fluctuate due to the destruction or not.

[0045] Further, the third preset threshold is used to determine the sliding source area boundary (high volatility region), and when the third preset threshold is taken, first, the first fluctuation index of the sliding source area boundary and the non-boundary region in the limited historical information is counted. The boundary area is affected by the difference between the inside and outside of the sliding source area, and the fluctuation index is significantly higher (the time series difference is intense). The fluctuation index of the non-boundary area is relatively flat. Then, the optimal segmentation point is determined through the ROC curve to maximize the boundary recognition accuracy. The value corresponding to this optimal segmentation point is the third preset threshold. Therefore, the first fluctuation index is compared with a preset third preset threshold: when the index exceeds this threshold, that is, when the fluctuation is intense enough, the ith sub-region is formally identified as the sliding source area boundary region. Through this mechanism, the S3 step realizes the accurate identification of the sub-regions located at the junction between the sliding source area and the non-destroyed region in the set of intense destruction regions preliminarily screened out in S2, providing a boundary basis for the definition of the core region (sliding source area) and subsequent area calculation.

[0046] In S4, the boundary region of the landslide area (i.e., the sliding body coverage or disturbance range) is accurately identified in the second set of sub-regions preliminarily screened out in S2 (representing a potential landslide area with a significant but relatively moderate decrease in vegetation coverage). The input data of this step comes from the second set of sub-regions output by S2. The basic principle is similar to that of S3: the sub-regions on the true boundary of the landslide area will also exhibit a specific spatial discontinuity (difference) in the time series evolution pattern of vegetation coverage compared to their adjacent surrounding areas that are not affected by the landslide. This difference also manifests as a dynamic comparison of the vegetation coverage data sequences of the two regions over the entire collection period.

[0047] Therefore, S4 first needs to construct a difference sequence for each second to-be-processed data sequence (denoted as the jth, corresponding to a specific potential landslide sub-region). The construction process follows similar but different logic than S3: obtain the vegetation coverage data sequence of all sub-slope regions adjacent to the jth target sub-region space. At each same time point (collection point), calculate the absolute difference between the vegetation coverage values of these adjacent sub-regions and the corresponding vegetation coverage value of the jth sub-region itself. Again, the absolute value is used to ensure that the deviation of the target region from the neighborhood vegetation state at each time is purely quantified. Finally, arrange all the absolute differences calculated at all time points in chronological order to form a difference sequence that describes the inconsistency between the jth target sub-region and its surrounding neighbors in vegetation coverage dynamics over the entire study period. This sequence fully describes the dynamic difference trajectory between the candidate unit of the potential landslide and its geographical neighborhood environment in vegetation state over time.

[0048] Further, after successfully constructing the difference sequence reflecting the dynamic difference between the jth candidate landslide sub-region and its surrounding regions, the S4 step further needs to quantitatively evaluate the time fluctuation characteristics exhibited by the sequence to determine whether the sub-region is indeed located on the boundary of the landslide area. The core discrimination logic here also inherits from S3 but takes into account the characteristics of the landslide area: a sub-region that is truly located on the boundary of the landslide coverage area will be subject to the "pulling" effect between the internal disturbance-induced vegetation coverage degradation (weak negative change) and the external stable (or normal change) region. This location characteristic will typically manifest as a certain degree of instability or oscillation in its difference sequence with the neighborhood. Before and after the landslide occurs, especially during the peak of the movement or the stabilization stage, this difference may be relatively prominent at some time points (e.g., when the landslide advances to the edge of the region), while at other time points the difference may be smaller due to natural similarity in vegetation changes or the landslide impact has not yet arrived or has left.

[0049] To objectively quantify the oscillation amplitude of the entire difference sequence, the S4 step uses the method of calculating the second fluctuation index. The calculation process is consistent with S3: first, calculate the average value of all data in the difference sequence of the jth sub-region as the time reference for the difference level; then, subtract this average value from the data at each time point in the difference sequence to obtain the unit fluctuation amount at each time point relative to the average difference level; finally, sum all the absolute values of this unit fluctuation amount to obtain the final second fluctuation index. The size of this index value reflects the total amplitude of the difference sequence fluctuating above and below its average value (sum of absolute values). The higher the value, the greater the fluctuation and the more unstable the difference sequence.

[0050] Further, the fourth preset threshold value is used to determine the landslide area boundary. The value of the fourth preset threshold value is determined based on the fact that the fluctuation of the landslide area boundary is lower than that of the landslide source area but higher than that of the stable area. The typical fluctuation index of the landslide area boundary is about 60% to 80% of that of the landslide source area. Therefore, the threshold value can be directly linked and designed based on the third preset threshold value. For example, the fourth threshold value = the third threshold value * 0.7. When the value of the second fluctuation index exceeds the fourth preset threshold value, it is considered that the fluctuation is significant enough, indicating that the jth sub-region is located at the junction of the landslide affected area and the non-affected area. Therefore, the jth sub-region is determined as the landslide area boundary region.

[0051] Through this mechanism, the S4 step realizes accurate identification of the sub-region located at the edge of the actual influence range of the landslide body from the set of regions reflecting relatively mild vegetation degradation that are preliminarily screened out in the S2 step, thereby providing key spatial boundary information for subsequent accurate delineation of the landslide area contour and calculation of the area of the landslide area (S5 step).

[0052] In the S5 step, the total area of the target landslide source area and the total area of the target landslide area are accurately calculated based on the spatial information of the landslide source area boundary region and the landslide area boundary region identified in the S3 and S4 steps, respectively. Finally, the mobility of the landslide is quantitatively evaluated based on the proportional relationship between the two.

[0053] First, the landslide source area calculation is processed: Since the boundary regions identified in the S3 step constitute the clearest dividing line between the landslide source area and the surrounding areas that are not severely damaged (usually non-landslide areas), these boundary regions are usually discontinuous and distributed in discrete points in space. Therefore, the first step of the S5 step is to connect all the landslide source area boundary regions that are adjacent to each other in space to form an irregular, possibly branched but continuous first boundary extension path. This path outlines the main part of the landslide source area contour. However, the landslide source area, as a broken core region wrapped inside a larger landslide body, has a relatively ambiguous boundary with the non-landslide area (i.e., other parts of the landslide body) that is not severely damaged within the landslide body, which is not easy to accurately define directly through remote sensing images or change characteristics. In view of this, in order to completely delineate the landslide source area and make it form a closed region for area calculation, the S5 step adopts a reasonable estimation method: the starting point and the ending point of the first boundary extension path are directly connected by a straight line. This straight line connection is based on the consideration of the continuity of the geological structure, providing a determined and closed boundary of the landslide source area contour for subsequent calculation. At this time, the complete and closed polygonal region formed by the connection path of the actually identified boundary regions and the connecting line of the starting point and the ending point is determined as the target landslide source area.

[0054] Finally, the total area of the target landslide source region can be obtained by summing up the areas of all the sub-slope regions covered by the target landslide source region. The calculation logic of the total area of the target landslide region is similar but simpler: since the boundary regions of the target landslide region identified by S4 constitute a clear spatial dividing line between the overall movement coverage of the landslide and the stable region outside which is completely unaffected by the landslide, the boundary features are very clear. Therefore, S5 directly connects the spatially adjacent boundary regions of the target landslide region in sequence, without the need for additional first and last connection operations, to naturally form a completely closed second boundary extension path. The space enclosed by the closed path is identified as the target landslide region. Similarly, the total area of the target landslide region can be obtained by summing up the areas of all the sub-slope regions contained in the target landslide region.

[0055] Further, after obtaining the total area of the target landslide source region and the total area of the target landslide region based on the spatial boundary identification results, S5 performs the final quantitative evaluation of the landslide mobility. The core basis of the evaluation is to calculate the ratio of the total area of the target landslide source region to the total area of the target landslide region (i.e., the area ratio).

[0056] The physical meaning of this ratio value is very clear: it directly reflects the size of the rupture core region (landslide source region) relative to the total area of the landslide body affected by the expansion (landslide region). The larger the area ratio (i.e., the closer the value to 1), the higher the proportion of the landslide source region in the entire landslide region, and the smaller the landslide coverage expansion range beyond the landslide source region. This situation usually means that the landslide body has weak movement and diffusion ability, and the rupture does not trigger large-scale soil and rock migration, i.e., the landslide mobility is poor. The potential threat of geological disasters is more concentrated near the rupture source region.

[0057] On the contrary, the smaller the area ratio (i.e., the closer the value to 0), the smaller the core rupture region that triggers the landslide, but it eventually leads to a very broad range of surface material sliding and coverage. The landslide body shows strong movement and diffusion ability, effectively converting the energy released at the rupture point into large-area long-distance migration of soil and rock materials, i.e., the landslide mobility is good or high. Such a landslide often has stronger destructive power and can threaten infrastructure and people at a greater distance.

[0058] By this area ratio index, the complex boundary recognition results based on the spatio-temporal analysis of vegetation changes in S1 to S4 are finally converted into a simple, intuitive and clear quantitative parameter with geological significance and disaster evaluation value in S5 step. This index overcomes the problem of inaccurate division of the landslide source area and landslide area in the traditional method, which leads to evaluation distortion, and provides a more reliable quantitative basis for the movement and diffusion capacity of the landslide body for disaster warning and risk prevention and control decision-making. After calculation, the area ratio is the final output value describing the mobility characteristics of the landslide event, serving subsequent analysis and application.

[0059] In summary, in the entire landslide mobility confirmation method based on the adaptive difference model, first, a double-threshold dynamic judgment mechanism is adopted to calculate the normalized difference index according to the adjacent change characteristics of the vegetation cover data sequence, and a strict first preset threshold and a relatively loose second preset threshold are set to accurately separate the two types of regions with different damage characteristics in space, reducing the area misjudgment caused by the existing single-threshold method. Second, the spatial neighborhood fluctuation analysis model is innovatively introduced. The absolute difference between the candidate sub-region and the adjacent region in the vegetation cover data at all time points is calculated to form a difference sequence, and the first and second fluctuation indexes are constructed based on the average value of the sequence and the absolute sum of the deviations of each point. The instability characteristics in the time dimension are used to identify the true boundary region. Finally, a differentiated area extraction mechanism based on boundary characteristics is established. For the landslide source area, the discrete boundary points are connected to form a straight line to close the non-significant boundary (which is wrapped in the landslide body), and for the landslide area, the discrete boundary points are directly connected to form a natural closed path. Then, the total area of the sub-region within the boundary is calculated to obtain the accurate division range. Finally, the area ratio of the landslide source area to the landslide area is taken as the mobility parameter (a small ratio indicates a small rupture core area but a large diffusion range, representing strong mobility). This realizes the full-link quantitative evaluation from dynamic change capture, spatial boundary recognition to disaster movement and diffusion capacity, greatly improving the scientificity of disaster warning accuracy and prevention and control strategy formulation.

[0060] As shown in FIG. 1, in one embodiment, the difference index is obtained according to the adjacent vegetation cover data in the vegetation cover data sequence in S2. Figure 2

[0061] S21, subtracting the previous vegetation cover data from the next vegetation cover data of the adjacent vegetation cover data in the vegetation cover data sequence to obtain a difference amount.

[0062] S22, dividing the smallest difference amount in the vegetation cover data sequence by the average value of all vegetation cover data in the vegetation cover data sequence to obtain a difference index.

[0063] ​In the embodiment, it is to be noted that in S21, the vegetation coverage data sequence is traversed, and the difference between the vegetation coverage value of the next time point and the vegetation coverage value of the previous time point in the sequence is calculated one by one. The difference is essentially an absolute numerical expression of the change in vegetation state between adjacent time points. In the case of landslide damage, the landslide impact will cause the surface vegetation to be destroyed, buried or significantly damaged, so that the vegetation coverage value at the subsequent time point is generally lower than the previous value, and thus the calculated difference is usually negative. The size of the negative value intuitively reflects the degree of vegetation coverage attenuation between specific consecutive time points. The larger the negative value (the larger the absolute value), the more severe the vegetation damage in the time period. S21 generates a series of discrete differences representing the time gradient of vegetation coverage for each sub-region.

[0064] In S22, a comprehensive index representing the strongest vegetation damage event in the entire acquisition period of the sub-region is extracted from the numerous single differences calculated in S21. The method is to select the one with the smallest value (i.e., the one with the most negative change, representing the maximum magnitude of vegetation coverage reduction) among all the differences. However, to avoid distortion in the direct comparison of change magnitudes between regions with different basic levels of vegetation coverage (e.g., dense forest areas and sparse shrub areas), normalization processing is required. Therefore, the selected minimum difference is divided by the average value of all vegetation coverage data in the entire acquisition period of the sub-region. The resulting ratio is the difference index. The index is a relative quantity that not only reflects the absolute peak intensity of vegetation damage, but also eliminates the interference of the average coverage level of the region itself, making different basic vegetation coverage sub-regions comparable in terms of vegetation attenuation degree.

[0065] In one embodiment, the vegetation coverage data sequence of the adjacent sub-slope region of the sub-slope region corresponding to the i-th first to-be-processed data sequence and the difference sequence of the i-th first to-be-processed data sequence in S3 according to the i-th first to-be-processed data sequence include:

[0066] The vegetation coverage data of each acquisition point in the vegetation coverage data sequence of the adjacent sub-slope region of the sub-slope region corresponding to the i-th first to-be-processed data sequence is obtained, the vegetation coverage data of each acquisition point in the i-th first to-be-processed data sequence is obtained, and the absolute value of the difference between the vegetation coverage data of the same acquisition point in the above two vegetation coverage data sequences is obtained. All absolute values are arranged in time sequence and form the difference sequence of the i-th first to-be-processed data sequence.

[0067] In this embodiment, it should be noted that for the i-th potential landslide source area sub-region selected through S2 (whose vegetation cover data sequence belongs to the first data sequence to be processed), a key data point—the difference sequence—is constructed. The purpose of this sequence is to quantify the dynamic inconsistency of the vegetation cover status between the candidate region and all its geographically directly adjacent sub-regions over the entire time dimension. The construction process is as follows: First, the vegetation cover data sequence of the i-th sub-region and the vegetation cover data sequences corresponding to each of its spatially neighboring sub-regions are acquired simultaneously.

[0068] Then, for each corresponding sampling point on the time axis (i.e., the same observation time), the absolute value of the difference between the vegetation cover value of the i-th sub-region and the vegetation cover value of each neighboring sub-region at that time is calculated (i.e., ignoring the positive and negative directions, only measuring the magnitude of the difference). Finally, these difference magnitude values ​​calculated at the same time (i.e., the absolute difference between the target area and the adjacent areas at the same point) are arranged in chronological order to form a time series. This newly constructed series is the difference series unique to the i-th sub-region, where each element reflects the instantaneous spatial difference in vegetation state between the target area and its immediate surrounding environment (neighborhood) at a specific point in time, and the entire series depicts the trajectory of this spatial difference changing over time.

[0069] like Figure 3 As shown, in one embodiment, obtaining the first fluctuation index based on the difference sequence of the i-th first data sequence to be processed in step S3 includes:

[0070] S31. Obtain the average value of all data in the difference sequence of the i-th first data sequence to be processed, subtract the average value from each data in the difference sequence to obtain the unit fluctuation amount;

[0071] S32. Add up the absolute values ​​of all unit fluctuations in the difference sequence of the i-th first data sequence to be processed, and obtain the first fluctuation index.

[0072] In this embodiment, it should be noted that S31 is the first step in calculating the first fluctuation index. The goal is to establish a time baseline level for the vegetation cover difference between the i-th potential landslide source area boundary candidate sub-region and its surrounding neighborhood. Specifically, for the already constructed sequence describing the absolute difference in vegetation cover between the target sub-region and all spatially adjacent sub-regions at all collection time points (i.e., data at each point in the difference sequence), the arithmetic mean of all elements in the sequence is calculated. This mean represents an overall central value or stable state level of the degree of difference in vegetation cover status between the candidate sub-region and its immediate neighbors on the time axis throughout the entire collection period. It abstractly reflects the average intensity of the difference and provides a calculation benchmark for subsequent evaluation of the fluctuation range of the difference sequence around the average level.

[0073] In S32, after calculating the average value of the difference sequence in S31, S32 is responsible for quantifying the degree of dispersion (fluctuation) of the entire sequence data around the average value, i.e., obtaining the first fluctuation index. The specific method is: for each time point data in the difference sequence, subtract the average value of the sequence to obtain a new sequence called "unit fluctuation amount". The unit fluctuation amount represents the deviation of the difference value at that time from the average difference level. Then, take the absolute value of the unit fluctuation amount calculated at all time points (ignore the positive and negative directions, only care about the size of the deviation), and finally sum all the absolute values of the unit fluctuation amount. This sum is the first fluctuation index. The first fluctuation index is essentially a measure of the total amplitude of the deviation of the target candidate sub-region from the average level of the difference in vegetation coverage in the neighborhood over the entire time period. The larger the value of the first fluctuation index, the more dramatic the fluctuation of the difference sequence over time, and the more unstable the difference level.

[0074] As shown in FIG. 5, in one embodiment, obtaining the target sliding source area according to the plurality of sliding source boundary regions in S5 includes: Figure 4

[0075] S51, connecting a plurality of adjacent sliding source boundary regions and forming a first boundary extension path;

[0076] S52, linearly connecting the head and tail ends of the first boundary extension path, so as to take the area surrounded by the first boundary extension path as the target sliding source area;

[0077] S53, obtaining the area of all sub-slope regions occupied by the target sliding source area as the target sliding source area.

[0078] In this embodiment, it should be noted that in S51, the spatial aggregation of the sliding source boundary regions identified by S3 is used to prepare for the sliding source range delineation. The object it processes is a plurality of sub-slope regions (polygons) that are spatially discrete and determined as sliding source boundaries. The specific operation is: find the adjacency relationship of these boundary regions in space, and connect the boundary regions that are connected in position one by one to form an irregular polyline path that may contain multiple line segments (because the sub-regions that are continuously adjacent in space are connected), called the first boundary extension path. This path outlines the main part of the sliding source area outside contour, reflecting the preliminary boundary line between the sliding source area and the surrounding area that has not been severely damaged (usually non-sliding area).

[0079] ​In S52, the first boundary extension path formed by the connection in S51 is usually unconnected (not closed), and its interior is the undetermined landslide source area, while its exterior is mainly a non-landslide area. However, since the landslide source area, as a rupture core, is enclosed within the entire landslide area (determined by the subsequently identified landslide area boundary), its boundary with the non-landslide source area (i.e., the inner boundary of the landslide source area) often lacks obvious variation characteristics and is difficult to accurately detect using the aforementioned methods. Therefore, step S52 adopts an approximation strategy: on the spatial map, a straight line segment is used to directly connect the start and end points of the first boundary extension path. In this way, the first boundary extension path plus this connecting line together constitutes a closed polygonal boundary. This treatment is based on the geological assumption that boundary position changes usually have local continuity, providing a calculable outline of the landslide source area.

[0080] In S53, the basis is the spatial region defined by the closed polygon boundary determined in step S52, which is defined as the target landslide source area. The specific calculation method is as follows: First, determine which original sub-slope regions, divided in step S1, are completely covered or occupied by the target landslide source area; then, sum the area values ​​of all these delineated sub-slope regions (the area of ​​each sub-region was determined during the division in S1). This sum is the total area of ​​the target landslide source area, reflecting the spatial scale of the core rupture region.

[0081] like Figure 5 As shown, in one embodiment, obtaining the target landslide area based on multiple landslide area boundary regions in step S5 includes:

[0082] S54. Connect multiple adjacent landslide area boundaries to form a closed second boundary extension path, and take the area enclosed by the second boundary extension path as the target landslide area.

[0083] S55. Obtain the area of ​​all sub-slope areas occupied by the target landslide area and use it as the area of ​​the target landslide area.

[0084] In this embodiment, it should be noted that S54 is similar to S51, but it processes the landslide area boundary region identified in step S4. Its boundary characteristics are very clear, namely, a clear dividing line between the final impact area of ​​the landslide movement and the completely unaffected stable area outside. Therefore, by simply connecting all spatially connected landslide area boundary regions directly according to their natural geographical location, a closed zigzag path, i.e., the second boundary extension path, can be naturally formed. The interior of this closed path represents the entire extent of the landslide area, eliminating the need for additional manual connection operations, as the detected boundary already completely and encloses the range of landslide movement diffusion.

[0085] In S55, the total area of the target landslide area is calculated, which has exactly the same logic as S53, except that the target object is changed to the landslide area. The basis is the space area enclosed by the second boundary extension path formed in step S54, which is defined as the target landslide area. The specific operation is: determine all S1 divided original sub-slope areas completely contained in the enclosed boundary path; then, the area values of all the contained sub-slope areas are added up. This total is the total area of the target landslide area, reflecting the total range of the surface sliding and covering impact caused by the landslide disaster. The area value is one of the key input parameters for calculating the final landslide mobility evaluation index (area ratio).

[0086] Also provided is a landslide mobility confirmation system based on an adaptive difference model, the system comprising:

[0087] An acquisition module is configured to acquire remote sensing image information in a collection period in which a target slope occurs a landslide, divide the target slope into a plurality of sub-slope areas, and acquire a vegetation coverage data sequence of each sub-slope area according to the remote sensing image information;

[0088] A first data processing module is configured to acquire a difference index according to adjacent vegetation coverage data in the vegetation coverage data sequence, acquire vegetation coverage data sequence with a difference index less than a first preset threshold as a first to-be-processed data sequence, and acquire vegetation coverage data sequence with a difference index less than a second preset threshold as a second to-be-processed data sequence, wherein the first preset threshold is less than the second preset threshold;

[0089] A second data processing module is configured to acquire a difference sequence of the i th first to-be-processed data sequence according to vegetation coverage data sequences of adjacent sub-slope areas of a sub-slope area corresponding to the i th first to-be-processed data sequence and the i th first to-be-processed data sequence, and acquire a first fluctuation index according to the difference sequence of the i th first to-be-processed data sequence, and if the first fluctuation index exceeds a third preset threshold, the sub-slope area corresponding to the i th first to-be-processed data sequence is taken as a landslide source area boundary area;

[0090] A third data processing module is configured to acquire a difference sequence of the j th second to-be-processed data sequence according to vegetation coverage data sequences of adjacent sub-slope areas of a sub-slope area corresponding to the j th second to-be-processed data sequence and the j th second to-be-processed data sequence, and acquire a second fluctuation index according to the difference sequence of the j th second to-be-processed data sequence, and if the second fluctuation index exceeds a fourth preset threshold, the sub-slope area corresponding to the j th second to-be-processed data sequence is taken as a landslide area boundary area;

[0091] The mobility confirmation module is configured to obtain a target sliding source area according to the plurality of sliding source boundary areas, obtain a target sliding area according to the plurality of sliding area boundary areas, divide the target sliding source area by the target sliding area, and obtain an area ratio indicating the sliding mobility.

[0092] In one embodiment, the first data processing module is further configured to: subtract a previous vegetation coverage data from a next vegetation coverage data of adjacent vegetation coverage data in the vegetation coverage data sequence to obtain a difference amount; and divide a minimum difference amount in the vegetation coverage data sequence by an average value of all vegetation coverage data in the vegetation coverage data sequence to obtain a difference index.

[0093] In one embodiment, the second data processing module is further configured to: obtain vegetation coverage data of each collection point in a vegetation coverage data sequence of an adjacent sub-slope area of a sub-slope area corresponding to the i th first to-be-processed data sequence, obtain vegetation coverage data of each collection point in the i th first to-be-processed data sequence, and obtain an absolute value of a difference between the vegetation coverage data of the same collection point in the two vegetation coverage data sequences; and arrange all the absolute values of the differences in a time sequence to form a difference sequence of the i th first to-be-processed data sequence.

[0094] In one embodiment, the second data processing module is further configured to: obtain an average value of all data in the difference sequence of the i th first to-be-processed data sequence, and subtract the average value from each data in the difference sequence to obtain a unit fluctuation amount.

[0095] In one embodiment, the second data processing module is further configured to: obtain an average value of all data in the difference sequence of the i th first to-be-processed data sequence, and subtract the average value from each data in the difference sequence to obtain a unit fluctuation amount.

[0096] In the embodiment, it should be noted that the specific manner of performing operations of the landslide mobility confirmation system based on the adaptive difference model has been described in detail in the embodiments of the landslide mobility confirmation method based on the adaptive difference model, and will not be described in detail here.

[0097] The preferred embodiments of the present disclosure are described in detail above with reference to the accompanying drawings, but the present disclosure is not limited to the specific details in the above embodiments. Within the technical concept of the present disclosure, various simple modifications can be made to the technical solutions of the present disclosure, and these simple modifications all belong to the protection scope of the present disclosure.

[0098] In addition, it should be noted that each specific technical feature described in the above specific embodiments can be combined in any appropriate manner without contradiction. In order to avoid unnecessary repetition, the present disclosure will not further describe various possible combinations.

[0099] In addition, any combination of the various embodiments of the present disclosure can also be made, as long as it does not deviate from the idea of the present disclosure, it should also be considered as disclosed by the present disclosure.

[0100] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or equivalently replace some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application, and they should all be covered in the scope of the claims and description of the present application.

Claims

1. A method for confirming landslide mobility based on an adaptive difference model, characterized by, The method comprises the following steps: acquiring remote sensing image information in a collection cycle in which a time point at which a target slope slides occurs, dividing the target slope into a plurality of sub-slope regions, and acquiring a vegetation coverage data sequence of each sub-slope region according to the remote sensing image information; acquiring a difference index according to adjacent vegetation coverage data in the vegetation coverage data sequence, acquiring vegetation coverage data sequence with a difference index less than a first preset threshold as a first to-be-processed data sequence, and acquiring vegetation coverage data sequence with a difference index less than a second preset threshold as a second to-be-processed data sequence, wherein the first preset threshold is less than the second preset threshold; acquiring vegetation coverage data of each collection point in a vegetation coverage data sequence of a sub-slope region adjacent to a sub-slope region corresponding to an i-th first to-be-processed data sequence, acquiring vegetation coverage data of each collection point in the i-th first to-be-processed data sequence, acquiring an absolute value of a difference between the vegetation coverage data of the same collection point in the two vegetation coverage data sequences, arranging all the absolute values of the differences in time sequence in turn to form a difference sequence of the i-th first to-be-processed data sequence; acquiring a first fluctuation index according to the difference sequence of the i-th first to-be-processed data sequence, and regarding the sub-slope region corresponding to the i-th first to-be-processed data sequence as a sliding source area boundary region if the first fluctuation index exceeds a third preset threshold; acquiring a difference sequence of the j-th second to-be-processed data sequence according to a vegetation coverage data sequence of a sub-slope region adjacent to a sub-slope region corresponding to the j-th second to-be-processed data sequence and the j-th second to-be-processed data sequence, acquiring a second fluctuation index according to the difference sequence of the j-th second to-be-processed data sequence, and regarding the sub-slope region corresponding to the j-th second to-be-processed data sequence as a sliding area boundary region if the second fluctuation index exceeds a fourth preset threshold; acquiring a target sliding source area according to a plurality of sliding source area boundary regions, acquiring a target sliding area according to a plurality of sliding area boundary regions, and obtaining an area ratio for indicating sliding mobility by dividing the target sliding source area by the target sliding area.

2. The landslide mobility confirmation method based on an adaptive difference model according to claim 1, characterized by, The method further comprises the following steps: subtracting a previous vegetation coverage data from a next vegetation coverage data of adjacent vegetation coverage data in the vegetation coverage data sequence to obtain a difference amount; dividing the minimum difference amount in the vegetation coverage data sequence by an average value of all vegetation coverage data in the vegetation coverage data sequence to obtain the difference index. 3.The landslide mobility confirmation method based on an adaptive difference model according to claim 1, wherein, The method further comprises the following steps: obtaining an average value of all data in the difference sequence of the i-th first to-be-processed data sequence, subtracting the average value from each data in the difference sequence to obtain a unit fluctuation amount, and adding absolute values of all unit fluctuation amounts in the difference sequence of the i-th first to-be-processed data sequence to obtain the first fluctuation index. The method further comprises the following steps:

4. The landslide mobility confirmation method based on an adaptive difference model according to claim 1, characterized by, connecting a plurality of adjacent sliding source area boundary regions to form a first boundary extension path. ​ linearly connecting the first and the last ends of the first boundary extension path, so as to take the area surrounded by the first boundary extension path as a target sliding source area; acquire the area of all sub-slope areas occupied by the target sliding source area as the target sliding source area.

5. The landslide mobility confirmation method based on an adaptive difference model according to claim 1, characterized by, The method further includes: connecting the plurality of adjacent sliding source area boundary regions to form a closed second boundary extension path, and taking the area surrounded by the second boundary extension path as a target sliding source area; acquire the area of all sub-slope areas occupied by the target sliding source area as the target sliding source area. 6.A landslide mobility confirmation system based on an adaptive difference model, characterized by The system is used to implement the landslide mobility confirmation method based on the adaptive difference model, and the system comprises: an acquisition module configured to acquire remote sensing image information in a collection period in which a time point at which a target slope occurs a landslide is located, divide the target slope into a plurality of sub-slope areas, and acquire vegetation coverage data sequences of the sub-slope areas according to the remote sensing image information; a first data processing module configured to acquire a difference index according to adjacent vegetation coverage data in the vegetation coverage data sequences, acquire vegetation coverage data sequences with a difference index less than a first preset threshold as first to-be-processed data sequences, and acquire vegetation coverage data sequences with a difference index less than a second preset threshold as second to-be-processed data sequences, wherein the first preset threshold is less than the second preset threshold; a second data processing module configured to acquire a difference sequence of an i th first to-be-processed data sequence according to vegetation coverage data sequences of adjacent sub-slope areas of a sub-slope area corresponding to the i th first to-be-processed data sequence, and acquire a first fluctuation index according to the difference sequence of the i th first to-be-processed data sequence, and if the first fluctuation index exceeds a third preset threshold, the sub-slope area corresponding to the i th first to-be-processed data sequence is taken as a sliding source area boundary region; a third data processing module configured to acquire a difference sequence of a j th second to-be-processed data sequence according to vegetation coverage data sequences of adjacent sub-slope areas of a sub-slope area corresponding to the j th second to-be-processed data sequence, and acquire a second fluctuation index according to the difference sequence of the j th second to-be-processed data sequence, and if the second fluctuation index exceeds a fourth preset threshold, the sub-slope area corresponding to the j th second to-be-processed data sequence is taken as a sliding source area boundary region; a mobility confirmation module configured to acquire a target sliding source area according to a plurality of sliding source area boundary regions, and acquire a target sliding source area according to a plurality of sliding source area boundary regions, divide the target sliding source area by the target sliding source area to obtain an area ratio for indicating landslide mobility.

7. The adaptive difference model based landslide mobility confirmation system of claim 6, wherein, The first data processing module is further configured to: subtract a previous vegetation coverage data from a next vegetation coverage data of adjacent vegetation coverage data in the vegetation coverage data sequences to obtain a difference amount; divide a minimum difference amount in the vegetation coverage data sequences by an average value of all vegetation coverage data in the vegetation coverage data sequences to obtain a difference index.

8. The adaptive difference model based landslide mobility confirmation system of claim 6, wherein, The second data processing module is further configured to: The average value of all data in the difference sequence of the i-th first to-be-processed data sequence is obtained, each data in the difference sequence is subtracted by the average value, and a unit fluctuation amount is obtained; the absolute values of all unit fluctuation amounts in the difference sequence of the i-th first to-be-processed data sequence are added, and a first fluctuation index is obtained.

Images