Landslide risk assessment method, system, device and medium based on multi-level prediction

Abstract

The application discloses a landslide risk assessment method, system, device and medium based on multi-level prediction, relates to the technical field of data processing, and comprises the following steps: acquiring a target area, dividing a plurality of sub-regions, acquiring historical landslide data and historical landslide positions; acquiring a susceptibility index, defining a high-susceptibility region and a low-susceptibility region; counting a terrain erosion sequence, acquiring a to-be-processed erosion index of each high-susceptibility region, acquiring a correction ratio of each high-susceptibility region, and acquiring a target erosion index of each high-susceptibility region; if the target erosion index of the i th high-susceptibility region is lower than a second preset threshold, the i th high-susceptibility region is defined as a medium-susceptibility region, and a landslide risk grade corresponding to each sub-region is acquired according to the high-susceptibility region, the medium-susceptibility region and the low-susceptibility region. The application has the advantages of dynamic refinement, spatial cooperation and hierarchical self-adaptation.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a landslide risk assessment method, system, equipment, and medium based on multi-level prediction. Background Technology

[0002] In the field of geological disaster monitoring, early warning, prevention and mitigation, there is a need for risk assessment of regional and dynamically evolving landslide disasters. Existing assessment methods have some problems in their application.

[0003] Specifically, existing landslide risk assessment methods mainly rely on historical static landslide data (such as location and scale statistics) to macroscopically classify regional susceptibility (high / low risk zones). This static assessment model often treats the designated high-risk areas as a homogeneous whole, ignoring the dynamic differences in geological hazard activity (such as topographic erosion rate and deformation trend) among the sub-regions within a high-risk area. This results in the inability to accurately identify key locations within the high-risk area that are truly in a state of accelerated instability and are likely to experience landslides in the near future. Secondly, existing methods typically view each assessment unit in isolation, failing to effectively quantify the chain effect of disasters between adjacent high-risk units (i.e., the risk amplification phenomenon where the potential instability of one unit may trigger or accelerate the instability of its adjacent units). This leads to an underestimation of the severity of local disaster risk clusters. Thirdly, there is a lack of mechanisms for using dynamic monitoring data (such as continuous topographic displacement monitoring sequences) to provide real-time feedback, verification, and fine-tuning of the preliminary macroscopic susceptibility assessment results. Summary of the Invention

[0004] To address the technical problems in existing technologies where assessment results lag behind the actual dynamic process of disaster development, failing to meet the needs of refined early warning and targeted deployment of monitoring points, resulting in low efficiency in disaster prevention resource allocation, limited accuracy of early warning, and difficulty in supporting timely disaster prevention decisions, this invention provides a landslide risk assessment method, system, equipment, and medium based on multi-level prediction.

[0005] A landslide risk assessment method based on multi-level prediction includes: acquiring a target area and dividing the target area into multiple sub-regions, and acquiring historical landslide data and locations of the target area; acquiring susceptibility indicators for each sub-region based on historical landslide locations and data, and defining each sub-region as a high-susceptibility or low-susceptibility region based on its susceptibility indicators and a first preset threshold; statistically analyzing the topographic erosion sequence of each high-susceptibility region within the current assessment period, acquiring untreated erosion indicators for each high-susceptibility region based on its topographic erosion sequence, acquiring a correction ratio for each high-susceptibility region based on the number of adjacent high-susceptibility regions, and acquiring a target erosion indicator for each high-susceptibility region based on its untreated erosion indicators and correction ratio; if the target erosion indicator of the i-th high-susceptibility region is lower than a second preset threshold, then defining the i-th high-susceptibility region as a medium-susceptibility region, and acquiring the landslide risk level corresponding to each sub-region based on the high-susceptibility, medium-susceptibility, and low-susceptibility regions.

[0006] Optionally, obtaining the susceptibility index of each sub-region based on the historical landslide location and historical landslide data includes: obtaining the historical landslide source coverage of each historical landslide based on the historical landslide location; obtaining the number of times the historical landslide source coverage of each historical landslide includes the j-th sub-region, and using this number as the susceptibility index of the j-th sub-region; and obtaining the susceptibility index of each sub-region by dividing the susceptibility index of each sub-region by the sum of the susceptibility indexes of all sub-regions.

[0007] Optionally, defining each sub-region as a high-risk region or a low-risk region based on the susceptibility index of each sub-region and a first preset threshold includes: if the susceptibility index of the j-th sub-region exceeds the first preset threshold, then the j-th sub-region is designated as a high-risk region; if the susceptibility index of the j-th sub-region does not exceed the first preset threshold, then the j-th sub-region is designated as a low-risk region.

[0008] Optionally, the statistical analysis of the topographic erosion sequence of each high-risk area within the current assessment period includes: statistical analysis of the topographic displacement of each collection segment of the i-th high-risk area within the current assessment period; and arranging the multiple topographic displacements of the i-th high-risk area in the current assessment period in chronological order to form the topographic erosion sequence of the i-th high-risk area.

[0009] Optionally, obtaining the erosion index to be treated for each high-risk area based on the topographic erosion sequence of each high-risk area includes: obtaining the displacement difference between the previous and subsequent topographic displacements in the topographic erosion sequence of each high-risk area; summing all displacement differences in the topographic erosion sequence of each high-risk area to obtain the total difference; dividing the total difference of the topographic erosion sequence of each high-risk area by the number of displacement differences, and defining the calculation result as the erosion index to be treated for the topographic erosion sequence of each high-risk area.

[0010] Optionally, obtaining the correction ratio of each high-risk region based on the number of adjacent high-risk regions of each high-risk region includes: obtaining the number of adjacent sub-regions of each high-risk region and using it as a first number; obtaining the number of adjacent high-risk regions of each high-risk region and using it as a second number; dividing the second number by the first number to obtain the correction ratio of each high-risk region.

[0011] Optionally, obtaining the target erosion index for each high-risk area based on the erosion index to be treated and the correction ratio for each high-risk area includes: adding 1 to the correction ratio for each high-risk area to obtain the weight value for each high-risk area; and multiplying the erosion index to be treated for each high-risk area by the weight value to obtain the target erosion index for each high-risk area.

[0012] A landslide risk assessment system based on multi-level prediction is also provided. The system includes: an acquisition module for acquiring a target area, dividing the target area into multiple sub-regions, and acquiring historical landslide data and locations; a first risk assessment module for acquiring susceptibility indicators for each sub-region based on historical landslide locations and data, and defining each sub-region as a high-susceptibility or low-susceptibility region based on its susceptibility indicators and a first preset threshold; and a second risk assessment module for statistically analyzing the topographic erosion sequence of each high-susceptibility region within the current assessment period, and defining each sub-region as a high-susceptibility or low-susceptibility region based on its susceptibility indicators and a first preset threshold. The topographic erosion sequence of the high-risk area is used to obtain the erosion index to be treated for each high-risk area, and the correction ratio of each high-risk area is obtained according to the number of adjacent high-risk areas. The target erosion index of each high-risk area is obtained according to the erosion index to be treated and the correction ratio. The third risk assessment module is used to define the i-th high-risk area as a medium-risk area when the target erosion index of the i-th high-risk area is lower than the second preset threshold, and to obtain the landslide risk level corresponding to each sub-area according to the high-risk area, medium-risk area and low-risk area.

[0013] An electronic device is also provided, comprising: a memory storing a computer program thereon; and a processor for executing the computer program in the memory to implement a landslide risk assessment method based on multi-level prediction.

[0014] A non-transitory computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, implements a landslide risk assessment method based on multi-level prediction.

[0015] The beneficial effects of this invention are reflected in:

[0016] In the multi-level prediction-based landslide risk assessment method, firstly, based on the macro-level high-risk areas defined by historical landslide source coverage frequency, the dynamic activity of each high-risk sub-region (represented by the erosion index to be processed) is quantified by introducing topographic erosion sequence data within the current monitoring period. This accurately identifies key points within the high-risk areas that are truly on the verge of instability due to recent accelerated erosion. Secondly, by statistically analyzing the proportion of adjacent high-risk areas and calculating the spatial correlation correction coefficient (correction ratio), the chain amplification effect caused by the potential instability of surrounding units is incorporated into the calculation of the target erosion index, solving the problem of underestimating local risk clusters. Finally, the initial high-risk areas are dynamically reclassified (some are downgraded to medium-risk) using the target erosion index combined with a second threshold. Combined with the retained low-risk areas, a comprehensive evaluation system covering high, medium, and low risks is ultimately formed. Attached Figure Description

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0018] Figure 1 This is a schematic diagram illustrating the steps of the landslide risk assessment method based on multi-level prediction of the present invention;

[0019] Figure 2 This is a schematic diagram of part of step S2 in the landslide risk assessment method based on multi-level prediction of the present invention;

[0020] Figure 3 This is a schematic diagram of another part of step S2 in the landslide risk assessment method based on multi-level prediction of the present invention;

[0021] Figure 4 This is a schematic diagram of part S3 in the landslide risk assessment method based on multi-level prediction of the present invention;

[0022] Figure 5This is a schematic diagram of another part of step S3 in the landslide risk assessment method based on multi-level prediction of the present invention;

[0023] Figure 6 This is a schematic diagram of another part of step S3 in the landslide risk assessment method based on multi-level prediction of the present invention;

[0024] Figure 7 This is a schematic diagram of another part of step S3 in the landslide risk assessment method based on multi-level prediction of the present invention;

[0025] Figure 8 This is a block diagram illustrating an electronic device according to an embodiment of the present invention.

[0026] Reference numerals: 700 - Electronic device, 701 - Processor, 702 - Memory, 703 - Multimedia component, 704 - I / O interface, 705 - Communication component. Detailed Implementation

[0027] 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.

[0028] 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.

[0029] 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.

[0030] like Figure 1 As shown, a landslide risk assessment method based on multi-level prediction is provided. In one embodiment, the method includes:

[0031] S1. Obtain the target area, divide the target area into multiple sub-areas, and obtain the historical landslide data and historical landslide locations of the target area;

[0032] S2. Based on historical landslide locations and historical landslide data, obtain the susceptibility index of each sub-region, and define each sub-region as a high-susceptibility region or a low-susceptibility region based on the susceptibility index of each sub-region and the first preset threshold.

[0033] S3. Statistically analyze the topographic erosion sequence of each high-risk area within the current assessment period, obtain the erosion index to be treated for each high-risk area based on the topographic erosion sequence of each high-risk area, obtain the correction ratio of each high-risk area based on the number of adjacent high-risk areas, and obtain the target erosion index of each high-risk area based on the erosion index to be treated and the correction ratio of each high-risk area.

[0034] S4. If the target erosion index of the i-th high-risk area is lower than the second preset threshold, then the i-th high-risk area is defined as a medium-risk area, and the landslide risk level corresponding to each sub-area is obtained according to the high-risk area, medium-risk area and low-risk area.

[0035] In this implementation, it should be noted that in S1, the research object—the geographic space where landslides may occur—is clearly defined, and its internal structure is finely segmented. Simultaneously, key historical information characterizing the disaster background of the area is collected. Specifically, the first step is to "obtain the target area," which means clearly defining the spatial boundaries of the assessment work. This target area is typically an area with a certain degree of topographical complexity, geological conditions, or human activity impact, resulting in a potential landslide risk. Examples include a watershed, a slope area of ​​a mining area, or a mountain road corridor with potential geological hazards. Clearly defined geographic boundaries are crucial, providing a unified spatial framework for all subsequent analyses. After determining the target area, it needs to be "divided into multiple sub-regions." This division process is not arbitrary; it requires reasonable segmentation based on actual topographic features (such as slope and aspect changes), geological structural units, hydrological network distribution, or management needs (such as administrative divisions and monitoring grids). The aim is to discretize the continuous geographic space into a series of basic assessment units (sub-regions) with relatively uniform internal attributes. For example, in steep mountainous areas, division can be based on natural gullies or ridgelines; in broad, gentle slopes, regular grids (such as kilometer or hectare grids) can be used. Each sub-region formed after division will serve as the smallest spatial unit for subsequent calculations of "susceptibility indicators" and statistics on "topographic erosion sequences." The choice of sub-region scale requires a trade-off between accuracy requirements and computational resources; too large a scale will result in the loss of internal variability information (making it impossible to accurately identify high-risk points), while too small a scale will increase the burden of data collection and analysis.

[0036] Furthermore, while completing the spatial structure division, historical landslide data and locations for the target area are acquired. This forms the basis for macro-level susceptibility zoning based on historical static data in risk assessment. Historical landslide data not only refers to the spatial coordinates or record points of landslide events, but more profoundly, it requires "obtaining the historical landslide source coverage of each historical landslide." This includes a precise description of the spatial impact of historical landslide events: the range of the source area (sliding initiation area) of each landslide event, which needs to be obtained through historical records, ultimately forming a polygon of the landslide impact range that can cover one or more sub-regions. For example, a recorded landslide event, after detailed investigation, determined that its source area roughly covered two adjacent 50x50 meter grids (sub-regions) on the hillside, while the deposit extended to cover three grids below where the road passed. In addition, historical landslide data typically includes supplementary information such as the time of occurrence, scale (volume or area), and possible triggering factors (e.g., heavy rainfall, earthquakes). Although not necessarily used directly in the initial S2 zoning, it provides background support for understanding regional disaster patterns and dynamic analysis in subsequent steps (e.g., considering the relationship between erosion rate and precipitation in S3). Historical landslide location information needs to be spatially correlated with the divided sub-regions (e.g., overlay analysis) to determine which sub-regions were specifically affected by each historical landslide event. Ultimately, the structured target area division results (set of sub-regions) and their associated historical landslide information (event list, coverage area) provide input for calculating the "susceptibility index" of each sub-region based on frequency (number of occurrences / total number of occurrences) in S2.

[0037] In S2, the potential tendency (susceptibility) of landslides in the history of all sub-regions divided in S1 is quantified, and a preliminary risk level classification is made based on this, defining the scope for subsequent dynamic analysis and fine-tuning. The core input of this step is the historical landslide data of the target area prepared in S1 and the set of divided sub-regions. The specific process first involves accurately defining the actual impact range of historical landslide events, that is, identifying the "historical landslide source coverage area". This is different from simply recording the location of landslide points, but requires depicting the spatial extent of the source area (landslide source area) where the material began to destabilize and move in each historical landslide event. For example, in the investigation report of a landslide event triggered by heavy rainfall, the landslide source area is identified as an irregular area on a specific slope, whose boundary may cross multiple contour lines and cover several sub-regions in the grid. Subsequently, through spatial analysis techniques (such as overlay analysis in geographic information systems), the landslide source coverage area of ​​each historical landslide event is cross-compared with all sub-region layers to accurately count the number of times each sub-region is covered by different historical landslide source areas.

[0038] For example, if a sub-region located at a specific point on a river valley slope (such as a grid on a hillside) is covered by the source areas of three different landslides out of 10 historical landslide events, then the "proneness count" for that sub-region is 3. After completing this statistical analysis for all sub-regions, the historical "exposure" count for each region is obtained. Finally, for cross-regional comparisons, the proneness count for each sub-region needs to be standardized to calculate its "proneness index." This index represents the relative share of the sub-region in the total "exposure" of all historical landslide events, i.e., the relative probability that the region was historically selected as a landslide source. The calculation logic for this index is to divide the proneness count of a single sub-region by the sum of the proneness counts of all sub-regions in the entire target region (i.e., the cumulative number of times the sub-region is covered by the total area covered by all historical landslide sources). The resulting proneness index is a relative value between 0 and 1; the higher the value, the stronger the tendency of the sub-region to act as a landslide source in historical records.

[0039] Furthermore, after obtaining the susceptibility indicators for all sub-regions, the subsequent operation of S2 aims to classify the sub-regions of the entire target area into a preliminary, binary risk classification based on these indicator values: high-susceptibility areas and low-susceptibility areas. The core decision-making basis is the "first preset threshold." This threshold is typically a critical value set comprehensively based on the overall distribution characteristics of historical data for the study area (such as the average, median, or specific percentile of all susceptibility indicators), the geological background knowledge of the region, and management needs. For example, when determining the first preset threshold, for a mountainous area, the susceptibility indicators for 500 sub-regions show a right-skewed distribution (most areas <0.05), with a few areas >0.3. Through cumulative frequency curve analysis, it was found that areas with indicators ≥0.12 accounted for 10% of the total, and these areas covered 80% of landslide events in the past 5 years. After confirmation by geological experts that this threshold can effectively distinguish between tectonic fracture zones and stable bedrock areas, the first preset threshold was ultimately set at 0.12.

[0040] For each sub-region (denoted as the j-th sub-region), if its calculated susceptibility index exceeds the preset first threshold, then the sub-region is classified as a "high susceptibility region". Conversely, if its index value is equal to or lower than the first threshold, then it is classified as a "low susceptibility region".

[0041] For example, assuming that after expert evaluation and data analysis, a first threshold of 0.1 is set, a sub-region located on a steep, fractured rock slope with a susceptibility index of 0.25 (far above the threshold) is marked as a high-susceptibility area; while another sub-region located on a gentle, stable deposit layer has an index of only 0.03 (below the threshold) and is marked as a low-susceptibility area. The result of this step is the generation of a preliminary risk zoning map of the target area, clearly identifying all clusters of "high-susceptibility areas" that require high attention. This zoning plays a crucial role in connecting the preceding and following steps: it directly relies on historical spatial patterns (static) to identify the source areas most likely to produce landslides (corresponding to the part of the existing methods mentioned in the technical issues section).

[0042] In S3, the processing target is all the high-risk areas initially identified in S2, obtaining quantitative indicators reflecting the geological activity (i.e., recent instability tendency) of these high-risk sub-regions within the current assessment period—target erosion indicators. Its core inputs include the list of high-risk areas output from S2 and the current dynamic monitoring data (topographic displacement) for these areas. Specifically, the process first requires calculating the "topographic displacement" of each high-risk area (denoted as the i-th) within each "collection segment" (e.g., daily, weekly, or per sensor scan cycle) within the current assessment period (e.g., after a rainy season, a heavy rainfall event, or a specific monitoring period). Understanding "topographic displacement" is crucial: it represents the relative distance of movement of surface or shallow material due to rainfall, freeze-thaw cycles, or geological creep within that time period, providing direct evidence of whether slope deformation and erosion are currently occurring and the intensity of such deformation. Considering the potential for locational differences within a sub-region (e.g., significant crack expansion at a particular point), displacement values ​​at multiple representative locations within a single acquisition segment are typically collected. The value with the largest displacement is then selected as the representative value for that sub-region within that acquisition segment (e.g., if point A displaces by 2 cm and point B by 5 cm in a single scan of a sub-region, then 5 cm is taken). These representative values, collected in chronological order, are then compiled and arranged according to acquisition time to form a key data sequence characterizing the dynamic surface / near-surface erosion process of that sub-region during the current assessment period—the topographic erosion sequence. This is essentially a time series, depicting the direct observational trajectory of the region's geological activity over time.

[0043] Furthermore, having sequence data characterizing dynamic erosion, the subsequent S3 operations focus on calculating an intermediate metric that comprehensively reflects the intensity of recent trends in erosion (i.e., the tendency for activity to intensify or weaken), and further integrates the potential amplification effect of cascading risks brought about by its spatial location. First, the changing trends of the topographic erosion sequence in each high-risk area are analyzed: this is done by calculating the change in topographic displacement recorded between two adjacent collection periods in the sequence (e.g., this week and last week), i.e., the "displacement difference." For example, if the displacement in a certain area was 3 cm last week and 8 cm this week, the difference is +5 cm (increasing), while if it was 10 cm the week before last and 3 cm last week, the difference is -7 cm (decreasing). The displacement differences between all such consecutive time points are summed, and then the average change intensity is calculated, resulting in the erosion index to be processed. This index characterizes the average increase or decrease in topographic displacement in the area within the current assessment period: a larger positive value indicates an accelerating average deformation rate (increased risk of erosion intensification), while a negative value may indicate a slowdown in activity. However, considering only its own dynamic trends is insufficient to fully characterize the risk—geological hazards (such as landslides) often exhibit spatial correlations. The impending instability of one region may significantly increase the likelihood of instability in its adjacent high-risk areas through physical mechanisms (such as lateral displacement, loading, and changes in groundwater pathways). To quantify this "neighborhood risk amplification effect," S3 introduces spatial analysis: it counts the total number of all directly adjacent sub-regions around each high-risk area (the first quantity), and the number of these adjacent sub-regions that are themselves identified as high-risk areas by S2 (the second quantity). Dividing the second quantity by the first quantity yields a "correction ratio," which essentially reflects the tightness with which the high-risk area is "surrounded" by other high-risk areas—the higher the ratio, the more centrally located it is within a "high-risk cluster," and the greater the risk of a chain reaction triggered by its instability. Finally, the erosion index to be treated, representing the intensity of its recent dynamic erosion changes, is multiplied by a weighting coefficient of (correction ratio + 1) to calculate the "target erosion index." This indicator not only includes the region's average erosion change in the near term, but also adds the additional risk multiplier effect caused by the instability of surrounding high-risk areas due to its spatial location. It is a more comprehensive and dynamic quantification of the actual danger level of a single sub-region at present.

[0044] In S4, the "high-risk areas" initially identified in S2 are further refined based on the dynamic assessment results of S3, and the classification results of all areas are integrated to generate the final landslide risk level map. This step processes all high-risk areas output from S3 and their corresponding "target erosion indices." For each high-risk area (denoted as the i-th), its target erosion index is compared with the second preset threshold (the initial judgment of S4). If the target erosion index is lower than the second preset threshold, the high-risk area is reclassified as a "medium-risk area." This second preset threshold, similar to the first preset threshold in S2, is usually a critical value set based on a comprehensive analysis of historical data, monitoring experience, and tolerance for manageable risks in the target area. It is mainly used to distinguish areas that, although historically considered high-risk areas (high macro-risk), show relatively gentle dynamic changes in surface erosion during the current monitoring period (indicated by a small average change in the erosion index to be processed) and have a low risk of being affected by potential chain instability in surrounding high-risk areas (indicated by a small correction ratio, resulting in a ultimately lower target erosion index). For example, when setting the second preset threshold, during a certain rainy season monitoring period, the target erosion index of 100 high-risk areas was distributed between [-0.8, 5.2], showing a bimodal distribution. Analysis revealed that the index in the first peak area (active area) was ≥1.2 (accounting for 60%), and the index in the second peak area (decay area) was ≤0.3 (accounting for 30%). Combined with concurrent rainfall data, it was confirmed that the displacement rate in areas with a value <0.5 was below the rainfall-induced critical value.

[0045] Based on this, a second preset threshold of 0.5 was set, and 30 areas with indicators <0.5 were downgraded to medium-risk areas. Subsequent verification revealed that no new landslides occurred in the downgraded areas in the following two months, while three landslide events occurred in the non-downgraded areas.

[0046] Furthermore, S4 downgrades these areas from "high" to "medium," essentially performing a dynamic "screening" or "cooling down" operation on macroscopic high-risk areas. For example, a sub-region located on the edge of a macroscopic high-risk area cluster, whose monitored displacement changes are very small or even slightly stabilized due to recent drought (weak dynamic trend), and whose surrounding neighboring areas are only a few high-risk areas (low spatial correlation risk), might have a target erosion index calculated below the second threshold. Therefore, it would be downgraded to a medium-risk area, indicating that although historically prone to erosion, it is not currently a primary and urgent threat. Areas with target erosion indices higher than or equal to the second threshold, however, retain their "high-risk area" level, indicating that their own activity and the spatial risk environment suggest that their danger remains very prominent in the current period.

[0047] Furthermore, after completing the dynamic risk refinement of the original high-risk area set (partially downgraded to medium-risk), the final goal of S4 is to output the final landslide risk level of all sub-regions in the entire target area. This is the result of integrating risk assessment information from three different sources and of different natures: (1) the initial binary division of all areas (high / low susceptibility) based on historical static data in S2, which characterizes the long-term historical tendency background risk level of the area; (2) the reclassification of the high-risk area subset in S2 based on the dynamic indicators (target erosion indicators) in S3 (adjusted to medium susceptibility), which introduces the activity status and spatial correlation risk of the area in the current period; (3) the low-susceptibility areas originally identified in S2 remain unchanged. Therefore, the final output will contain three distinct risk level categories: high-susceptibility areas (historically high susceptibility and current dynamic indicators reaching the danger threshold), medium-susceptibility areas (historically high susceptibility but current dynamic indicators show that the risk is relatively controllable), and low-susceptibility areas (historical susceptibility has not reached the initial high threshold).

[0048] For example, the entire target area comprises thousands of sub-regions. After initial screening in S2, approximately 15% (e.g., 200) were identified as high-risk areas. In S3, the target erosion index for each of these 200 high-risk areas was calculated, and combined with a second preset threshold, it was found that approximately 50 areas had target erosion indices below the threshold. Therefore, in S4, these areas were redefined as medium-risk areas. The remaining 150 areas with target erosion indices still above the threshold remained at the high-risk level. Furthermore, the 85% of low-risk areas initially classified in S2 remained unchanged. Ultimately, each sub-region was assigned a comprehensive risk level label (high, medium, low) that integrates historical context, current dynamic changes, and spatial correlation effects. This final landslide risk level zoning map, compared to the initial zoning (S2) which relied solely on historical data, has significantly higher timeliness (incorporating the latest monitoring data) and refinement (breaking down macroscopic high-risk areas into high-dynamic-risk and medium-dynamic-risk areas). It more accurately identifies the "hotspots" that urgently require key attention and early warning and are most likely to become unstable in the near future (maintaining high susceptibility), while also identifying areas where the risk level has eased but still needs to be monitored (medium susceptibility), as well as areas with low historical risk (low susceptibility). This provides a scientific, reliable, and realistic basis for decision-making regarding the precise deployment of monitoring equipment, the optimization of emergency resource allocation, and the issuance of tiered early warning information, effectively solving the problems of the lagging and crude nature of existing static methods.

[0049] In summary, the multi-level prediction-based landslide risk assessment method firstly quantifies the dynamic activity (represented by unprocessed erosion indicators) of each high-risk sub-region by introducing topographic erosion sequence data from the current monitoring period, based on the macro-level high-risk zones defined by historical landslide source coverage frequency. This accurately identifies key points within high-risk zones that are truly on the verge of instability due to recent accelerated erosion. Secondly, by statistically analyzing the proportion of adjacent high-risk zones and calculating the spatial correlation correction coefficient (correction ratio), the chain amplification effect caused by potential instability of surrounding units is incorporated into the calculation of the target erosion indicator, solving the problem of underestimating local risk clusters. Finally, the initial high-risk zones are dynamically reclassified (some are downgraded to medium-risk) using the target erosion indicator combined with a second threshold. Combined with the retained low-risk zones, a comprehensive evaluation system covering high, medium, and low risks is ultimately formed. This multi-dimensional integration of historical background (susceptibility indicators), real-time dynamics (topographic erosion sequence), and spatial topology (neighborhood correction ratio) not only maintains the overall control over historical patterns but also captures the non-uniformity and correlation in the disaster evolution process, thereby improving the timeliness, spatial resolution, and rationality of the assessment results.

[0050] like Figure 2 As shown, in one embodiment, S2, obtaining susceptibility indicators for each sub-region based on historical landslide locations and historical landslide data includes:

[0051] S21. Obtain the historical landslide source coverage area of ​​each historical landslide based on the historical landslide location;

[0052] S22. Obtain the number of times the historical landslide source coverage area of ​​each historical landslide includes the j-th sub-region, and use this number as the number of times the j-th sub-region is prone to landslides;

[0053] S23. Divide the frequency of occurrence of each sub-region by the sum of the frequency of occurrence of all sub-regions to obtain the frequency index of each sub-region.

[0054] In this embodiment, it should be noted that in S21, the spatial source area where the unstable material initially began to slide in each historical landslide event is defined, i.e., the landslide source coverage area. For example, based on historical disaster investigation reports or remote sensing image interpretation results, a landslide triggered by heavy rain may be identified as originating from a specific-shaped area of ​​fractured rock mass on a hillside. This landslide source area is geographically depicted as an irregular polygon, and the extent of this polygon may completely or partially cover one or more grid cells in the divided sub-region grid.

[0055] In S22, the polygon layers showing the landslide source coverage of all historical landslide events obtained in S21 are spatially compared with the sub-region grid layers of the entire target area. For each sub-region (the j-th one), the number of different historical landslide source coverage polygons covering that sub-region is counted, and the total number of coverages is taken as the "proneness index" for that sub-region. For example, if a sub-region located on the edge of a canyon is spatially covered by the landslide source area polygons of three independent historical landslide events, then the proneness index for that sub-region is 3. This index reflects how frequently that sub-region has been identified as a landslide origin in history.

[0056] In S23, the "susceptibility frequency" (absolute exposure frequency) of each sub-region obtained in S22 is transformed into a standardized relative risk value that can be used for cross-regional comparison, namely the "susceptibility index". The calculation method is to use the susceptibility frequency of a single sub-region as the numerator and the sum of the susceptibility frequencies of all sub-regions within the entire target region as the denominator, and then perform a division operation. For example, if a sub-region has a susceptibility frequency of 2, and the sum of the susceptibility frequencies of all sub-regions within the entire region is 100, then the susceptibility index of that sub-region is 0.02. This index is a value between 0 and 1, representing the relative share of that sub-region among all regions historically recorded as landslide sources. The higher the value, the stronger its tendency to be selected as a landslide source in historical records, i.e., the higher the historical background risk.

[0057] like Figure 3 As shown, in one embodiment, defining each sub-region as a high-risk region or a low-risk region in S2 based on the susceptibility index of each sub-region and a first preset threshold includes:

[0058] S24. If the susceptibility index of the j-th sub-region exceeds the first preset threshold, then the j-th sub-region is designated as a high-susceptibility region.

[0059] S25. If the susceptibility index of the j-th sub-region does not exceed the first preset threshold, then the j-th sub-region is designated as a low-susceptibility region.

[0060] In this embodiment, it should be noted that in S24, based on the "susceptibility index" calculated in S23, a preliminary binary risk level classification is performed on all sub-regions. A key threshold value—"first preset threshold"—is set. For each sub-region (the j-th one), its susceptibility index is compared with this first preset threshold. If the susceptibility index of the sub-region exceeds the first preset threshold, then the sub-region is classified as a "high-susceptibility region". For example, assuming the first preset threshold is set to 0.08 based on regional risk distribution and management requirements, then a sub-region with a susceptibility index of 0.15 (greater than 0.08) indicates that its historical risk is significantly higher than the regional average, and therefore it is classified as a high-susceptibility region.

[0061] In S25, corresponding to S24, another part of the binary partitioning is completed. For each sub-region (the j-th), if its susceptibility index does not exceed (i.e., is equal to or lower than) a first preset threshold, then the sub-region is classified as a "low-susceptibility region". For example, a sub-region with a susceptibility index of only 0.02 (far below the first preset threshold of 0.08) indicates that it has rarely or never acted as a landslide source in history, reflecting a lower historical background risk, and is therefore classified as a low-susceptibility region.

[0062] like Figure 4 As shown, in one embodiment, the statistical analysis of topographic erosion sequences for each highly susceptible area during the current assessment period in S3 includes:

[0063] S31. Calculate the terrain displacement of the i-th high-risk area in each data collection segment during the current assessment period;

[0064] S32. Arrange the multiple topographic displacements of the i-th high-risk area in the current assessment period in chronological order to form the topographic erosion sequence of the i-th high-risk area.

[0065] In this embodiment, it should be noted that in S31, this step monitors surface deformation for each "collection segment" (e.g., each satellite pass or sensor scan) within the selected current assessment period (e.g., this rainy season) for each high-risk area (i-th one) identified in S2. Since there may be deformation differences at multiple points within a sub-region, the topographic displacement recorded at all valid monitoring locations within a single collection segment is statistically analyzed, and the largest displacement value is selected as the "topographic displacement" representing that collection segment and that sub-region. For example, if point A in a high-risk area displaces by 1 mm, point B by 3 mm, and point C by 5 mm during a certain collection period, then the topographic displacement of that sub-region in that collection segment is recorded as 5 mm.

[0066] In S32, the "terrain displacement" values ​​of S31 (the i-th highly susceptible area) collected in each acquisition segment within the current assessment period are strictly arranged according to their corresponding acquisition time sequence to form an ordered data sequence. For example, if the assessment period includes 4 acquisition segments, and the recorded displacement values ​​are 2 mm, 5 mm, 8 mm, and 3 mm in chronological order, then the constructed terrain erosion sequence for this area is [2, 5, 8, 3].

[0067] like Figure 5 As shown, in one embodiment, obtaining the erosion indicators to be treated for each high-risk area in S3 based on the topographic erosion sequence of each high-risk area includes:

[0068] S33. Obtain the displacement difference between the previous and next topographic displacements in the topographic erosion sequence of each high-risk area.

[0069] S34. Add up all the displacement differences of the topographic erosion sequence in each high-risk area and get the sum of the differences;

[0070] S35. Divide the sum of the differences in the topographic erosion sequences of each high-risk area by the number of displacement differences, and define the calculation result as the erosion index to be processed for the topographic erosion sequences of each high-risk area.

[0071] In this embodiment, it should be noted that in S33, this step analyzes the topographic erosion sequence constructed in S32 and calculates the change (difference) between the topographic displacement recorded in two temporally adjacent acquisition segments within the sequence. For example, for a sequence [a1,a2,a3,a4], it is necessary to calculate a2-a1 (the difference between the second segment and the first segment), a3-a2 (the difference between the third segment and the second segment), and a4-a3 (the difference between the fourth segment and the third segment) sequentially. Positive values ​​represent increased displacement (intensified erosion), and negative values ​​represent decreased displacement (slowed erosion).

[0072] In S34, all the "displacement differences between adjacent sampling segments" calculated in S33 for a high-risk area are summed to obtain a "total difference". This total represents the cumulative intensity of surface displacement changes in the area over a continuous time period within the current assessment period. A larger positive value indicates a more significant and sustained increase in erosion overall, while a larger negative value indicates a general tendency towards stabilization or rebound.

[0073] In S35, the "sum of differences" obtained in S34 is divided by the number of "displacement differences" calculated in S33 of the topographic erosion sequence of the region (i.e., the number of consecutive time intervals). The average value obtained is the "erosion index to be treated". This index quantifies the typical magnitude of topographic displacement change in the region in the current cycle, on average in each consecutive time interval. It is a direct measure of the region's recent average dynamic erosion intensity (rate of intensification or deceleration).

[0074] like Figure 6 As shown, in one embodiment, obtaining the correction ratio of each high-risk region in S3 based on the number of adjacent high-risk regions includes:

[0075] S36. Obtain the number of adjacent sub-regions of each high-incidence region and use it as the first quantity;

[0076] S37. Obtain the number of adjacent high-risk regions for each high-risk region and use it as the second quantity;

[0077] S38. Divide the second quantity by the first quantity to obtain the correction ratio for each high-risk region.

[0078] In this embodiment, it should be noted that in S36, based on the sub-region spatial division relationship of the target region (the adjacency relationship has been determined during the division), for each high-risk region (the i-th one), the number of all sub-regions that are directly adjacent to it in geographic space (shared boundary) is counted. This number is called the "first number" (i.e., the total number of directly adjacent units).

[0079] In S37, among all the adjacent sub-regions counted in S36, those adjacent sub-regions classified as "high-risk areas" in S2 are further filtered out, and their number is counted, called the "second number". This reflects how many neighbors around this area also belong to the historical high-risk background.

[0080] In S38, the ratio of the "second quantity" (number of adjacent high-risk areas) obtained in S37 to the "first quantity" (total number of all adjacent sub-regions) obtained in S36 is calculated to obtain the "corrected ratio". This ratio measures the degree of danger clustering in the local space where the high-risk area is located. The higher the ratio (closer to 1), the more closely it is surrounded by other historically high-risk areas; the lower the ratio (closer to 0), the more likely it is surrounded by low-risk areas or empty boundaries.

[0081] like Figure 7 As shown, in one embodiment, obtaining the target erosion index for each high-risk area in S3 based on the erosion index to be treated and the correction ratio for each high-risk area includes:

[0082] S39. Add 1 to the correction ratio of each high-risk region and obtain the weight value of each high-risk region.

[0083] S310. Multiply the erosion index to be treated in each high-risk area by the weight value to obtain the target erosion index for each high-risk area.

[0084] In this embodiment, it should be noted that in step S39, the "correction ratio" calculated in step S38 is incremented by 1 to obtain the "weight value" of the high-risk area. The "+1" operation ensures that the minimum weight is 1 (when the correction ratio is 0), which means that even without high-risk neighbors, its own dynamic risk still retains the basic weight; the larger the correction ratio, the larger the weight value, indicating that the additional risk it faces due to the high risk of its neighbors is amplified to a greater extent.

[0085] In S310, the erosion index to be treated calculated in S35 is multiplied by the weight value calculated in S39, which reflects the chain risk amplification effect due to its spatial location. The resulting product is the final target erosion index for this highly susceptible area. This index comprehensively quantifies the intensity of the unit's current geological activity and the associated risk amplification that may be brought about by the local environment. It is the core basis for dynamic grading adjustment in S4.

[0086] A landslide risk assessment system based on multi-level prediction is also provided, the system including:

[0087] The acquisition module is used to acquire the target area, divide the target area into multiple sub-regions, and acquire historical landslide data and historical landslide locations of the target area;

[0088] The first risk assessment module is used to obtain the susceptibility index of each sub-region based on the historical landslide location and historical landslide data, and to define each sub-region as a high-susceptibility region or a low-susceptibility region based on the susceptibility index of each sub-region and a first preset threshold.

[0089] The second risk assessment module is used to statistically analyze the topographic erosion sequence of each high-risk area within the current assessment period, obtain the erosion index to be treated for each high-risk area based on the topographic erosion sequence of each high-risk area, obtain the correction ratio of each high-risk area based on the number of adjacent high-risk areas, and obtain the target erosion index of each high-risk area based on the erosion index to be treated and the correction ratio of each high-risk area.

[0090] The third risk assessment module is used to define the i-th high-risk area as a medium-risk area when the target erosion index of the i-th high-risk area is lower than the second preset threshold, and to obtain the landslide risk level corresponding to each sub-area based on the high-risk area, medium-risk area and low-risk area.

[0091] In this embodiment, it should be noted that the specific method of performing the above-mentioned landslide risk assessment system based on multi-level prediction has been described in detail in the embodiments of the landslide risk assessment method based on multi-level prediction, and will not be elaborated here.

[0092] Figure 8 This is a block diagram of an electronic device for a landslide risk assessment method based on multi-level prediction, according to an exemplary embodiment. Figure 8 As shown, the electronic device 700 may include: a processor 701 and a memory 702. The electronic device 700 may also include one or more of a multimedia component 703, an I / O interface 704 (input / output interface), and a communication component 705.

[0093] The processor 701 controls the overall operation of the electronic device 700 to complete all or part of the steps in the multi-level prediction-based landslide risk assessment method described above. The memory 702 stores various types of data to support the operation of the electronic device 700. This data may include, for example, instructions for any application or method operating on the electronic device 700, and application-related data such as contact data, sent and received messages, pictures, audio, video, etc. The memory 702 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. Multimedia component 703 may include a screen and an audio component. The screen may be, for example, a touchscreen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in memory 702 or transmitted via communication component 705. The audio component also includes at least one speaker for outputting audio signals. I / O interface 704 provides an interface between processor 701 and other interface modules, such as a keyboard, mouse, buttons, etc. These buttons may be virtual or physical buttons. Communication component 705 is used for wired or wireless communication between the electronic device 700 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, 4G, NB-IoT, eMTC, or other 5G technologies, or combinations thereof, is not limited here. Therefore, the corresponding communication component 705 may include: a Wi-Fi module, a Bluetooth module, an NFC module, etc.

[0094] In an exemplary embodiment, the electronic device 700 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described landslide risk assessment method based on multi-level prediction.

[0095] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, which, when executed by a processor, implement the steps of the landslide risk assessment method based on multi-level prediction described above. For example, the computer-readable storage medium may be the memory 702 including the program instructions described above, which may be executed by the processor 701 of the electronic device 700 to complete the landslide risk assessment method based on multi-level prediction described above.

[0096] In another exemplary embodiment, a computer program product is also provided, the computer program product comprising a computer program executable by a programmable device, the computer program having a code portion for performing the above-described multi-level prediction-based landslide risk assessment method when executed by the programmable device.

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

[0098] It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, this disclosure will not describe the various possible combinations separately.

[0099] Furthermore, various different embodiments of this disclosure can be combined in any way, as long as they do not violate the spirit of this disclosure, they should also be regarded as the content disclosed in this disclosure.

[0100] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A landslide risk assessment method based on multi-level prediction, characterized in that, include: Obtain the target area, divide the target area into multiple sub-regions, and obtain historical landslide data and historical landslide locations for the target area; Based on historical landslide locations and historical landslide data, susceptibility indicators for each sub-region are obtained, and each sub-region is defined as a high-susceptibility region or a low-susceptibility region based on its susceptibility indicators and a first preset threshold. Calculate the topographic displacement of the i-th high-risk area in each collection segment during the current assessment period; arrange the multiple topographic displacements of the i-th high-risk area in the current assessment period in chronological order to form the topographic erosion sequence of the i-th high-risk area; Obtain the displacement difference between the previous and next topographic displacements in the topographic erosion sequence of each high-risk area; sum all displacement differences in the topographic erosion sequence of each high-risk area and obtain the total difference; divide the total difference of the topographic erosion sequence of each high-risk area by the number of displacement differences, and define the calculation result as the erosion index to be processed for the topographic erosion sequence of each high-risk area. Obtain the number of adjacent sub-regions of each high-risk region and use it as the first quantity; obtain the number of adjacent high-risk regions of each high-risk region and use it as the second quantity; divide the second quantity by the first quantity to obtain the correction ratio of each high-risk region. The target erosion index for each high-risk area is obtained based on the erosion index to be treated and the correction ratio for each high-risk area. If the target erosion index of the i-th high-risk area is lower than the second preset threshold, then the i-th high-risk area is defined as a medium-risk area, and the landslide risk level corresponding to each sub-area is obtained based on the high-risk area, medium-risk area and low-risk area.

2. The landslide risk assessment method based on multi-level prediction according to claim 1, characterized in that, The susceptibility indicators for each sub-region obtained based on historical landslide locations and historical landslide data include: The historical landslide source coverage area of ​​each historical landslide was obtained based on the location of the historical landslide. Obtain the number of times the historical landslide source coverage area of ​​each historical landslide includes the j-th sub-region, and use this number as the likelihood of occurrence in the j-th sub-region; The susceptibility index for each sub-region is obtained by dividing the frequency of occurrence in each sub-region by the sum of the frequency of occurrence in all sub-regions.

3. The landslide risk assessment method based on multi-level prediction according to claim 1, characterized in that, The step of defining each sub-region as a high-risk region or a low-risk region based on the susceptibility index of each sub-region and a first preset threshold includes: If the susceptibility index of the j-th sub-region exceeds the first preset threshold, then the j-th sub-region is designated as a high-susceptibility region. If the susceptibility index of the j-th sub-region does not exceed the first preset threshold, then the j-th sub-region is designated as a low-susceptibility region.

4. The landslide risk assessment method based on multi-level prediction according to claim 1, characterized in that, The process of obtaining the target erosion index for each high-risk area based on the erosion index to be treated and the correction ratio includes: Add 1 to the correction ratio of each high-risk region and obtain the weight value of each high-risk region; The erosion index to be treated in each high-risk area is multiplied by the weight value to obtain the target erosion index for each high-risk area.

5. A landslide risk assessment system based on multi-level prediction, characterized in that, The system implements the landslide risk assessment method based on multi-level prediction as described in any one of claims 1 to 4, the system comprising: The acquisition module is used to acquire the target area, divide the target area into multiple sub-regions, and acquire historical landslide data and historical landslide locations of the target area; The first risk assessment module is used to obtain the susceptibility index of each sub-region based on the historical landslide location and historical landslide data, and to define each sub-region as a high-susceptibility area or a low-susceptibility area based on the susceptibility index of each sub-region and a first preset threshold. The second risk assessment module is used to statistically analyze the topographic erosion sequence of each high-risk area within the current assessment period, obtain the erosion index to be treated for each high-risk area based on the topographic erosion sequence of each high-risk area, obtain the correction ratio of each high-risk area based on the number of adjacent high-risk areas, and obtain the target erosion index of each high-risk area based on the erosion index to be treated and the correction ratio of each high-risk area. The third risk assessment module is used to define the i-th high-risk area as a medium-risk area when the target erosion index of the i-th high-risk area is lower than the second preset threshold, and to obtain the landslide risk level corresponding to each sub-area based on the high-risk area, medium-risk area and low-risk area.

6. An electronic device, characterized in that, include: A memory on which computer programs are stored; A processor is configured to execute the computer program in the memory to implement the landslide risk assessment method based on multi-level prediction as described in any one of claims 1 to 4.

7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the landslide risk assessment method based on multi-level prediction as described in any one of claims 1 to 4.

Images