CN112767544A - Spatial case reasoning method for regional landslide risk evaluation - Google Patents
Spatial case reasoning method for regional landslide risk evaluation Download PDFInfo
- Publication number
- CN112767544A CN112767544A CN202110013321.5A CN202110013321A CN112767544A CN 112767544 A CN112767544 A CN 112767544A CN 202110013321 A CN202110013321 A CN 202110013321A CN 112767544 A CN112767544 A CN 112767544A
- Authority
- CN
- China
- Prior art keywords
- landslide
- case
- unknown
- similarity
- cases
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 47
- 238000011156 evaluation Methods 0.000 title abstract description 49
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 25
- 238000012502 risk assessment Methods 0.000 claims description 6
- 239000000126 substance Substances 0.000 claims description 4
- 238000013527 convolutional neural network Methods 0.000 description 18
- 238000012706 support-vector machine Methods 0.000 description 17
- 238000002474 experimental method Methods 0.000 description 12
- 238000013461 design Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 9
- 231100000807 Critical Body Residue (CBR) model Toxicity 0.000 description 7
- 238000011160 research Methods 0.000 description 6
- 238000000605 extraction Methods 0.000 description 5
- 238000012360 testing method Methods 0.000 description 5
- 238000012549 training Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 238000005191 phase separation Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 2
- 230000001939 inductive effect Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 230000002265 prevention Effects 0.000 description 2
- 208000025174 PANDAS Diseases 0.000 description 1
- 208000021155 Paediatric autoimmune neuropsychiatric disorders associated with streptococcal infection Diseases 0.000 description 1
- 240000000220 Panda oleosa Species 0.000 description 1
- 235000016496 Panda oleosa Nutrition 0.000 description 1
- 235000014676 Phragmites communis Nutrition 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000003139 buffering effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000006698 induction Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000010972 statistical evaluation Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/05—Geographic models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
Abstract
The invention relates to the technical field of geological disaster risk evaluation, and discloses a space case reasoning method for regional landslide risk evaluation, which comprises the steps of extracting attribute characteristics of landslide cases; calculating the spatial proximity of the landslide case and the adjacent grids and the spatial topological relation between the landslide case and the belonged area, thereby extracting the spatial features of the landslide case; calculating the attribute similarity of the unknown landslide case and the known landslide case, respectively determining the spatial proximity and the spatial topological relation similarity of the unknown landslide case and the known landslide case according to the spatial proximity and the spatial topological relation, further determining the spatial comprehensive similarity of the unknown landslide case and the known landslide case, determining the similarity of the unknown landslide case and the known landslide case according to the attribute similarity and the spatial comprehensive similarity, and finally determining the estimated landslide state of the unknown landslide case. The space case reasoning method for regional landslide risk evaluation can accurately evaluate regional landslide risk.
Description
Technical Field
The invention relates to the technical field of geological disaster risk assessment, in particular to a space case reasoning method for regional landslide risk assessment.
Background
Landslide is a common global natural disaster, and because the landslide causes huge economic loss and casualties every year, research on evaluation of the risk of the regional landslide becomes particularly important.
At present, a machine learning method is a method which is commonly used for evaluating the risk of regional landslide, and the risk evaluation of regional landslide is mostly based on the influence of basic factors, such as elevation, rainfall, fault and the like, on landslide. However, there are various factors causing landslide, and the risk assessment of regional landslide based on the basic factors is often unable to be accurately evaluated.
Therefore, how to provide an effective scheme to accurately evaluate the risk of regional landslide is an urgent problem to be solved in the prior art.
Disclosure of Invention
In order to solve the problem of inaccurate risk evaluation on regional landslide in the prior art, the invention aims to provide a spatial case reasoning method for regional landslide risk evaluation so as to accurately evaluate the regional landslide risk.
The invention provides a space case reasoning method for regional landslide risk evaluation, which comprises the following steps:
dividing an area to be evaluated into a plurality of grids, wherein each grid represents a landslide case;
calculating the spatial proximity of the landslide case and the adjacent grids thereof on the fault, the water system and the road network buffer surface;
determining a spatial topological relation between a central point of a landslide case and an equivalent surface of an area to which the landslide case belongs on each influence factor layer, wherein each influence factor layer comprises a slope layer, a lithology layer, a land utilization layer, a rainfall layer, an elevation layer, a population density layer, a vegetation coverage layer, a fault layer, a road network layer, a water system layer and a seismic center layer;
extracting attribute characteristic values of the landslide case in each influence factor layer;
calculating Euclidean distances of the unknown landslide cases and the known landslide cases on the corresponding dimension of each image layer, and obtaining attribute similarity of the unknown landslide cases and each known landslide case based on the Euclidean distances of the unknown landslide cases and each known landslide case on the corresponding dimension of each image layer;
determining the spatial proximity similarity between the unknown landslide cases and each known landslide case based on the spatial proximity of the unknown landslide cases and the adjacent grids thereof in the fault dimension, the water system dimension and the road network dimension, and the spatial proximity of each known landslide case and the adjacent grids thereof in the fault dimension, the water system dimension and the road network dimension;
determining the similarity of the spatial topological relations between the unknown landslide cases and each known landslide case on the basis of the spatial topological relations between the unknown landslide cases and the isosurface of each known landslide case on each layer;
determining the spatial comprehensive similarity of the unknown landslide cases and each known landslide case based on the spatial proximity similarity between the unknown landslide cases and each known landslide case and the spatial topological relation similarity between the unknown landslide cases and each known landslide case;
determining the similarity between the unknown landslide cases and each known landslide case based on the attribute similarity between the unknown landslide cases and each known landslide case and the spatial comprehensive similarity between the unknown landslide cases and each known landslide case;
and determining the estimated landslide state of the unknown landslide case based on the landslide state of the known landslide case with the maximum similarity to the unknown landslide case.
Based on the above content, the method for spatial case reasoning for regional landslide risk evaluation provided by the application calculates spatial proximity of a landslide case and adjacent grids thereof in each dimension, spatial topological relation of a central point of the landslide case and an area thereof in an isosurface of each influence factor layer, Euclidean distance between an unknown landslide case and a known landslide case in a known landslide state in a corresponding dimension of each layer, determines attribute similarity of the unknown landslide case and the known landslide case according to the Euclidean distance, determines spatial proximity similarity between the unknown landslide case and the known landslide case according to the calculated spatial proximity, determines spatial topological relation similarity between the unknown landslide case and the known landslide case according to the calculated spatial topological relation, determines spatial comprehensive similarity between the unknown landslide case and the known landslide case according to the topological relation and the proximity similarity, and finally, determining the estimated landslide state of the unknown landslide case according to the landslide state of the known landslide case with the maximum similarity to the unknown landslide case. According to the regional landslide risk evaluation-oriented space case reasoning method, when regional landslide risk evaluation is conducted, not only are basic factors influencing regional landslides considered, but also space characteristics among geological environment factors are considered, so that regional landslide risk can be accurately evaluated, and landslide disaster prevention and treatment can be guided.
In one possible design, the calculating euclidean distances between the unknown landslide case and a plurality of known landslide cases with known landslide states in the corresponding dimension of each image layer includes:
and calculating Euclidean distances between the unknown landslide case and the plurality of known landslide cases in the known landslide state on the corresponding dimension of each influence factor layer according to the attribute characteristic values of the unknown landslide cases in each influence factor layer and the attribute characteristic values of the plurality of known landslide cases in the known landslide state on each influence factor layer.
In one possible design, the similarity of the attributes of the unknown landslide case to each of the known landslide cases is:
wherein Dist (X, Y) is Euclidean distance between an unknown landslide case and each known landslide case on each influence factor layer, WiAnd the weights are corresponding to the layer i.
In one possible design, the spatial proximity similarity between an unknown landslide case and each of the known landslide cases is:
wherein, if
Then
For the j-th dimension spatial proximity value corresponding to the known landslide case,
a j-th dimension spatial proximity value corresponding to an unknown landslide case of an unknown landslide state,
is the weight corresponding to the proximity value.
In one possible design, the similarity of the spatial topological relationship between the unknown landslide case and each of the known landslide cases is as follows:
wherein the content of the first and second substances,
D(Xi,Yi) The spatial topological relation of the equivalent surfaces of the unknown landslide case and the known landslide case on a layer is represented, 0 represents that the unknown landslide case and the known landslide case do not belong to the same topological relation, 1 represents that the unknown landslide case and the known landslide case belong to the same topological relation, and q is the total number of the layers.
In one possible design, the spatial integrated similarity of an unknown landslide case to each of the known landslide cases is:
wherein, WpWeights corresponding to spatial proximity similarity between an unknown landslide case and a known landslide case, WtAnd weighting corresponding to the spatial topological relation similarity of the unknown landslide case and the known landslide case.
In one possible design, the similarity of an unknown landslide case to each of the known landslide cases is:
wherein, WsA weight W corresponding to the spatial comprehensive similarity of the unknown landslide case and each of the known landslide casesaAnd weighting corresponding to the similarity of the attributes of the unknown landslide case and the known landslide case.
In one possible design, the spatial proximity of a landslide case to its neighboring meshes in the dimensions of the road network, water system, and fault is:
where n denotes the number of adjacent grids, PiThe spatial proximity of the landslide case and an adjacent grid i among the plurality of adjacent grids in the dimensions of a road network, a water system and a fault is shown.
In one possible design, the area where the landslide case is located is an area formed by the grid corresponding to the landslide case and all the adjacent grids thereof.
In one possible design, the divided grid is a square grid of equal length and width.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a flowchart of a spatial case reasoning method for regional landslide risk assessment.
FIG. 2 is a schematic diagram of the intersection of a grid with a correlation factor buffer surface.
Fig. 3 is a schematic diagram of spatial topology extraction.
Fig. 4 is a graph illustrating the weight of the landslide factor determined using the IGR method.
FIG. 5 is a schematic diagram of landslide risk evaluation results under different CBR models.
FIG. 6 is a schematic diagram of landslide hazard area ratios predicted by different CBR models.
FIG. 7 is a schematic diagram of ROC curves for different CBR models.
Fig. 8 is a schematic diagram of a research area landslide risk evaluation result predicted under a CNN model and an SVM model.
FIG. 9 is a schematic diagram of landslide hazard area ratios predicted by CBR-ASSR, CNN and SVM models.
FIG. 10 is a diagram illustrating ROC curves of three models, CBR-ASSR, CNN, and SVM.
Detailed Description
The invention is further described with reference to the following figures and specific embodiments. It should be noted that the description of the embodiments is provided to help understanding of the present invention, but the present invention is not limited thereto. Specific structural and functional details disclosed herein are merely illustrative of example embodiments of the invention. This invention may, however, be embodied in many alternate forms and should not be construed as limited to the embodiments set forth herein.
It will be understood that, although the terms first, second, etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first element could be termed a second element, and, similarly, a second element could be termed a first element, without departing from the scope of example embodiments of the present invention.
It should be understood that, for the term "and/or" as may appear herein, it is merely an associative relationship that describes an associated object, meaning that three relationships may exist, e.g., a and/or B may mean: a exists alone, B exists alone, and A and B exist at the same time; for the term "/and" as may appear herein, which describes another associative object relationship, it means that two relationships may exist, e.g., a/and B, may mean: a exists independently, and A and B exist independently; in addition, for the character "/" that may appear herein, it generally means that the former and latter associated objects are in an "or" relationship.
It will be understood that when an element is referred to herein as being "connected," "connected," or "coupled" to another element, it can be directly connected or coupled to the other element or intervening elements may be present. Conversely, if a unit is referred to herein as being "directly connected" or "directly coupled" to another unit, it is intended that no intervening units are present. In addition, other words used to describe the relationship between elements should be interpreted in a similar manner (e.g., "between … …" versus "directly between … …", "adjacent" versus "directly adjacent", etc.).
It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments of the invention. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises," "comprising," "includes" and/or "including," when used herein, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, numbers, steps, operations, elements, components, and/or groups thereof.
It should also be noted that, in some alternative implementations, the functions/acts noted may occur out of the order noted in the figures. For example, two figures shown in succession may, in fact, be executed substantially concurrently, or the figures may sometimes be executed in the reverse order, depending upon the functionality/acts involved.
It should be understood that specific details are provided in the following description to facilitate a thorough understanding of example embodiments. However, it will be understood by those of ordinary skill in the art that the example embodiments may be practiced without these specific details. For example, systems may be shown in block diagrams in order not to obscure the examples in unnecessary detail. In other instances, well-known processes, structures and techniques may be shown without unnecessary detail in order to avoid obscuring example embodiments.
Examples
Please refer to fig. 1, which is a flowchart of a regional landslide risk evaluation-oriented spatial case-based reasoning method according to an embodiment of the present application, where the regional landslide risk evaluation-oriented spatial case-based reasoning method may include the following steps:
step S101, dividing an area to be evaluated into a plurality of grids, wherein each grid represents a landslide case.
In the embodiment of the application, the area to be evaluated is an area needing landslide risk evaluation, and is different from a traditional case reasoning method only aiming at attribute characteristics. When a landslide case is constructed, data is preprocessed by using Geographic Information System (GIS) software, an area to be evaluated can be divided into a plurality of grids by adopting a regular grid dividing mode, each grid is used as a landslide risk evaluation unit, and each grid represents one landslide case. The size of the grid can be divided according to actual conditions, and the grid divided in the embodiment of the present application is a positive direction grid with equal length and width, for example, the area to be evaluated is divided into grids with the size of 30m × 30 m.
And S102, calculating the spatial proximity of the landslide case and the adjacent grids thereof on the fault dimension, the water system dimension and the road network dimension buffer surface.
Spatial proximity describes the proximity of two geographic cells in space. To extract the spatial proximity of the landslide evaluation unit, 8 meshes around the landslide case can be selected as the extraction range, i.e., 3 × 3 mesh regions.
Because the fault, the water system and the road network have great influence on landslide induction, the three factors are selected to extract the proximity, and the spatial proximity of the landslide case and the adjacent grids thereof on the fault, the water system and the road network buffer surface is calculated.
The traditional spatial proximity formula is
Wherein l represents the common boundary length of the two grids, d is the distance between the centers of the two grids, and since the extraction range is from the center of the grid to the adjacent 8 grids when the spatial proximity of the landslide unit on the fault, the water system and the road network buffer surface is extracted, it is seen that the two extraction targets are not planar factors, so that the embodiment of the application improves the traditional spatial proximity formula and provides a new spatial proximity formula.
The improved spatial proximity formula is:
wherein d is the distance from the center point of the landslide case to the center point of the polygon formed by the intersection of the correlation factor buffer surface (fault, water system or road network) and the adjacent grid, and S is the area of the polygon formed by the intersection of the correlation factor buffer surface and the adjacent grid.
As shown in fig. 2, assuming that the shadow area is an area where a road network is located, the grid where the center point O is located is a landslide case, the polygon ACDFE is an intersecting surface of the road network and an adjacent grid on the left side of the landslide case, B is a geometric center of the intersecting surface, and S is an area of the polygon ACDFE.
As shown in fig. 2, the landslide case has 8 adjacent grids, and the road network intersects with 5 adjacent grids, which correspond to a spatial proximity. The spatial proximity of a landslide case and its neighboring meshes in the road network dimension can be expressed as:
wherein n represents the number of adjacent grids and has a value of 8, PiRepresenting the spatial proximity of the landslide case to the neighboring mesh i of the 8 neighboring meshes in the road network dimension.
And S103, determining the spatial topological relation between the central point of the landslide case and the isosurface of the area to which the landslide case belongs on each influence factor layer.
The spatial topological relation describes the adjacency, association and inclusion relation among spatial target points, lines and surfaces, and refers to the feature which is not changed in topological transformation and reflects the most basic mutual relation among spatial objects. In the topological relation extraction, 2 spatial topological relations including and separated from each other are selected, which are valuable to case reasoning.
In the embodiment of the application, each influence factor layer comprises a slope layer, a lithology layer, a land utilization layer, a rainfall layer, an elevation layer, a population density layer, a vegetation coverage layer, a fault layer, a road network layer, a water system layer and a seismic center layer. The area to which the landslide case belongs may be an area composed of the landslide case and its neighboring cells together, as shown by the dotted area in fig. 2.
As shown in FIG. 3, P1、P2、P3、P4The three-dimensional topological structure of the landslide case represents the isosurface of a certain factor layer, and the spatial topological relations between the central point O of the landslide case and the isosurfaces are phase separation, inclusion and phase separation in sequence. The iso-surface may be at various different influencesAnd the factor layer is divided into planes according to the corresponding attribute characteristic values.
And S104, extracting the attribute characteristic values of the landslide case in each influence factor layer.
The attribute feature value may be a value of a corresponding attribute on the impact factor layer. For example, in the rainfall map layer, the rainfall may be divided into a plurality of sections, the areas in the same rainfall section are divided into the same isosurface, and the attribute feature value in the rainfall map layer is the rainfall. For example, in the elevation map layer, the elevation may be divided into a plurality of sections, areas in the same elevation section may be divided into the same isosurface, and the attribute feature value of the elevation map layer is the elevation.
In the embodiment of the application, the central point of the landslide case represents the landslide case, and the central point is superposed with a slope layer, a lithology layer, a land utilization layer, a rainfall layer, an elevation layer, a population density layer, a vegetation coverage layer, a fault layer, a road network layer, a water system layer and a seismic layer, so that the attribute characteristic value of the central point of the landslide case in each influence factor layer can be determined. And replacing the landslide point/non-landslide point with the central point of the grid where the landslide point/non-landslide point is located, so as to obtain the influence factor value corresponding to each landslide point/non-landslide point, so as to construct a case. Each grid area represents a case, which is expressed as follows:
in the formula, akFor the attribute feature value of each influence factor layer, k is the number of layers (i.e. 12), p1、p2、p3Respectively representing the spatial proximity of the landslide case and the adjacent grids thereof on the fault, the water system and the buffering surface of the road network, Iq1:D,Iq2:W,L,IqmD represents the spatial topological relation between the central point of the landslide case and the isosurface of the region to which the landslide case belongs on each influence factor layer, IqmRepresenting an isosurface of an influence factor layer, D represents the spatial topological relation between the central point of the landslide case and the isosurface of the influence factor layerThe method comprises the steps that W represents that the spatial topological relation between the central point of the landslide case and the isosurface of the influence factor layer is a phase separation, R is a landslide type (landslide or non-landslide), and the R value of the unknown landslide case is null.
And S105, calculating Euclidean distances of the unknown landslide cases and the known landslide cases on the corresponding dimension of each layer, and obtaining the attribute similarity of the unknown landslide cases and each known landslide case based on the Euclidean distances of the unknown landslide cases and each known landslide case on the corresponding dimension of each layer.
The attribute similarity reasoning adopts a nearest neighbor algorithm, and measures the similarity by using the distance of the grids in the feature space. The smaller its distance, the more similar the two grids are. In the embodiment of the application, the euclidean distances of the unknown landslide cases and the known landslide cases in the known landslide state in the corresponding dimensions of each layer can be calculated according to the attribute feature values of the unknown landslide cases in the corresponding dimensions of each layer and the attribute feature values of the known landslide cases in the known landslide states in the corresponding dimensions of each layer.
The formula of Euclidean distances between an unknown landslide case and a plurality of known landslide cases in known landslide states on the corresponding dimension of each layer is as follows:
wherein the content of the first and second substances,
k represents the number of dimensions corresponding to each layer (i.e. 12), D (X)i,Yi) Is the difference value X of the attribute characteristic values of the dimensionalities corresponding to the image layers i of the unknown landslide case and the known landslide caseiAnd YiRespectively are values obtained after normalization processing is carried out on attribute characteristic values of the dimensionality corresponding to the layer i, WiAnd the weights are the attribute characteristic values of the dimensionality corresponding to the layer i.
Based on the euclidean distances of the unknown landslide cases and the known landslide cases in the dimension corresponding to each image layer, the attribute similarity between the unknown landslide cases and each known landslide case can be expressed as follows:
and S106, determining the spatial proximity similarity between the unknown landslide cases and each known landslide case based on the spatial proximity of the unknown landslide cases and the adjacent grids thereof in the fault dimension, the water system dimension and the road network dimension, and the spatial proximity of each known landslide case and the adjacent grids thereof in the fault dimension, the water system dimension and the road network dimension.
Spatial features are extremely important in solving geographic problems, and therefore how to define and calculate spatial proximity similarity becomes critical. In the embodiment of the application, the calculation formula of the spatial proximity similarity between the unknown landslide case and each known landslide case is as follows:
wherein, if
Then
For the j-th dimension spatial proximity value corresponding to the known landslide case,
a j-th dimension spatial proximity value corresponding to an unknown landslide case of an unknown landslide state,
is the weight corresponding to the proximity value.
And S107, determining the similarity of the spatial topological relations of the unknown landslide cases and the known landslide cases on the basis of the spatial topological relations of the equivalent surfaces of the unknown landslide cases and the known landslide cases on the layers.
The spatial topological relation similarity between the unknown landslide case and each known landslide case can be expressed as follows:
wherein the content of the first and second substances,
D(Xi,Yi) The spatial topological relation of the equivalent surfaces of the unknown landslide case and the known landslide case on a layer is represented, 0 represents that the unknown landslide case and the known landslide case do not belong to the same topological relation, 1 represents that the unknown landslide case and the known landslide case belong to the same topological relation, and q is the total number of the layers.
And S108, determining the spatial comprehensive similarity of the unknown landslide case and each known landslide case based on the spatial proximity similarity between the unknown landslide case and each known landslide case and the spatial topological relation similarity between the unknown landslide case and each known landslide case.
The spatial comprehensive similarity is a joint reasoning of spatial proximity similarity and spatial topological relation similarity, and the spatial comprehensive similarity between an unknown landslide case and each known landslide case can be expressed as follows:
wherein, WpWeights corresponding to spatial proximity similarity between an unknown landslide case and a known landslide case, WtAnd weighting corresponding to the spatial topological relation similarity of the unknown landslide case and the known landslide case.
And S109, determining the similarity between the unknown landslide case and each known landslide case based on the attribute similarity between the unknown landslide case and each known landslide case and the spatial comprehensive similarity between the unknown landslide case and each known landslide case.
In the embodiment of the application, when the similarity between an unknown landslide case and each known landslide case is determined, a first weight corresponding to the attribute similarity and a second weight corresponding to the spatial comprehensive similarity can be determined by adopting multiple collinearity analysis. And then carrying out weighting operation based on the attribute similarity of the unknown landslide case and each known landslide case, the spatial comprehensive similarity of the unknown landslide case and each known landslide case, the first weight corresponding to the attribute similarity and the second weight corresponding to the spatial comprehensive similarity to obtain the similarity of the unknown landslide case and each known landslide case.
The similarity of an unknown landslide case to each of the known landslide cases can be expressed as:
wherein, WsA weight W corresponding to the spatial comprehensive similarity of the unknown landslide case and each of the known landslide casesaAnd weighting corresponding to the similarity of the attributes of the unknown landslide case and the known landslide case.
And S110, determining the estimated landslide state of the unknown landslide case based on the landslide state of the known landslide case with the maximum similarity to the unknown landslide case.
In the embodiment of the application, if the landslide of the known landslide case with the largest similarity to the unknown landslide case is in the landslide state, the estimated landslide state of the unknown landslide case is set as the high landslide risk. And if the landslide of the known landslide case with the maximum similarity to the unknown landslide case is in the non-landslide state, setting the estimated landslide state of the unknown landslide case as the low landslide risk. Based on the steps S101 to S110, the estimated landslide state of each grid can be determined, and landslide risk mapping is performed on the area to be evaluated according to the estimated landslide state of each grid.
The spatial case reasoning method for regional landslide risk evaluation provided by the embodiment of the application calculates the spatial proximity of a landslide case and an adjacent grid thereof in each dimension, the spatial topological relation of a central point of the landslide case and an equivalent surface of an area to which the landslide case belongs in each influence factor layer, and the Euclidean distance of an unknown landslide case and a known landslide case in a known landslide state in a corresponding dimension of each layer, determines the attribute similarity of the unknown landslide case and the known landslide case according to the Euclidean distance, determines the spatial proximity similarity between the unknown landslide case and the known landslide case according to the calculated spatial proximity, determines the spatial topological relation similarity between the unknown landslide case and the known landslide case according to the calculated spatial topological relation, determines the spatial comprehensive similarity of the unknown landslide case and the known landslide case according to the topological relation and the proximity similarity, and finally, determining the estimated landslide state of the unknown landslide case according to the landslide state of the known landslide case with the maximum similarity to the unknown landslide case. According to the regional landslide risk evaluation-oriented space case reasoning method, when regional landslide risk evaluation is conducted, not only are basic factors influencing regional landslides considered, but also space characteristics among geological environment factors are considered, space correlation among data can be effectively expressed, interpretability among the data is improved, more valuable space characteristics are mined, landslide data characteristics can be enriched better, integrated reasoning is conducted, the method can be effectively applied to landslide risk evaluation, and accuracy of landslide risk evaluation is improved. Meanwhile, the regional landslide risk evaluation-oriented space case reasoning method has the advantages that the algorithm principle, the parameter setting and the reasoning process are simple, deterministic and effective, the regional landslide risk evaluation-oriented space case reasoning method does not need a complex algorithm principle and a network structure, the model parameter setting and the similarity reasoning principle are simple and clear, the model does not have a training process, and all known landslide cases can be directly used as a case base to carry out landslide risk evaluation. Therefore, the regional landslide risk evaluation-oriented space case reasoning method can be effectively applied to regional landslide risk evaluation through space similarity and attribute similarity integrated reasoning, and further guides prevention and treatment of landslide disasters. And, it can also be applied to the study of the geoscience problem in the similar scene.
In order to verify the performance of the regional landslide risk evaluation scheme provided by the application, the embodiment of the application adopts three evaluation indexes, namely overall precision and statistical index: the Recall rate (Recall), the Precision rate (Precision), the F-Measure (F1) and the ROC curve are used for comprehensively evaluating the superiority and inferiority of the regional landslide risk evaluation scheme provided by the application.
On the basis of the data processing, three dimensions of a fault, a road network and a water system are selected for each grid to calculate the spatial proximity, and the spatial topological relation of the isosurface of 12 layers is extracted. The landslide cases are constructed by basic attribute data, spatial proximity, spatial topological relation and landslide categories, known cases are formed by aiming at 706 known landslide cases with known landslide states, and unknown cases are formed by aiming at 1323696 grids in the whole area of Lushan county and used for risk evaluation of the landslide in the whole area. A layered random sampling method is adopted to enable known case sets to form three groups of experimental data sets according to the modes of 9:1, 8:2 and 7:3, and each data set respectively comprises case libraries and test cases formed by cases in corresponding proportions. Respectively developing attribute similarity reasoning (CBR-ASR), spatial similarity reasoning (CBR-SSR) and attribute and spatial similarity integrated reasoning (CBR-ASSR) experiments aiming at different data sets, wherein 8:1:1 training, verifying and testing data sets for Convolutional Neural Network (CNN) comparison experiments are further constructed aiming at the 9:1 data set, and 9:1 training and testing data sets for Support Vector Machine (SVM) comparison experiments are further constructed.
The output result of each method in the experiment is a probability value, and after the estimated landslide state of each grid is obtained through the regional landslide risk evaluation scheme provided by the application, the grids can be divided into high, medium and low risk levels. The classification principle is as follows: the standard deviation describes the discrete degree of a group of data, namely the difference between a sample data value and a sample average value, and the probability value obtained by subtracting the standard deviation from the landslide probability mean value can be regarded as a lower bound of probability of a unit being landslide of model evaluation, and the lower bound is taken as a boundary line of high risk and medium risk of the landslide; similarly, the probability value obtained by adding the standard deviation to the mean value of the non-landslide probability can be regarded as an upper bound of the non-landslide probability by the model judgment unit, and the bound is used as a bound of low risk and medium risk, so that the evaluation result can be divided into three levels of high risk, medium risk and low risk by the bound.
In order to meet the experiment requirements, an experiment program is developed based on open source libraries such as Numpy, Pandas, Scipy, Sklearn, Keras and the like by adopting Python language under a Windows10 operating system for developing experiments.
The experiments were conducted using a co-linearity test with a sample set of data formed from landslide and non-landslide points, with the results shown in table 1. TOL is the reciprocal of VIF, and it is generally considered that factors are not related when VIF is less than 10. As can be seen from Table 1, the highest VIF is the elevation factor, which is 5.907 and less than 10. Therefore, no correlation exists between the selected factors, and the selected factors can be all used for model training and testing.
TABLE 1
The weight of each factor was calculated using the Information Gain Ratio (IGR), and the result is shown in fig. 4, where the rainfall, earthquake, and gradient weights were the largest, indicating that landslide occurrence in the lushan area is mainly affected by these three factors. Due to the fact that 7.0-level earthquake and large-area rainfall occur in reed mountains in 2013, the two factors are main inducing factors of landslide, and the effectiveness of the IGR method in evaluating the landslide inducing factors is verified.
Before determining each item weight in the CBR by using an Analytic Hierarchy Process (AHP), firstly, carrying out case reasoning experiments by using a single spatial feature or a single reasoning model to obtain the corresponding classification evaluation precision, and carrying out importance ranking on the classification evaluation precision according to the obtained precision and prior knowledge, namely: the spatial topological relation > spatial proximity, and CBR-SSR > CBR-ASR, so that the weights of various items in CBR are specifically determined by an AHP method, and the results are shown in Table 2.
TABLE 2
Three CBR models are adopted to carry out landslide risk evaluation experiments. According to different experimental data sets, the landslide risk probability of 1323696 grids in the research area is calculated, and a landslide risk evaluation graph is drawn in GIS software. Fig. 5 is a graph of the 9:1 experimental data set for landslide risk assessment predicted for different CBR models. It can be seen that: the landslide high-risk area is mainly concentrated in a road network and a water system, and the evaluation results of the CBR-ASSR model and the CBR-SSR model are very similar. The high risk regions in fig. 5(a) are relatively dense, and the high risk regions in fig. 5(b) and (c) are relatively dispersed. FIG. 6 shows the landslide hazard area ratios predicted by different CBR models. The high-risk area occupation ratios of the CBR-ASR model, the CBR-SSR model and the CBR-ASSR model are 20.37%, 17.10% and 15.90% respectively, wherein the high-risk area occupation ratio in the CBR-ASR model is large, and as can be seen by combining with a figure 5(a), most of town areas present high-risk areas and are not in accordance with actual conditions; CBR-SSR; while the area ratio of the high risk region of CBR-ASSR is small, it can be seen from fig. 5(c) that the town region has only a small number of high risk regions and most of them are low risk regions. Compared with the three, the CBR-ASSR model is more suitable for the real situation of the Lushan area. Therefore, the evaluation precision can be effectively improved by integrating the spatial features to carry out spatial similarity reasoning, and the spatial similarity and attribute similarity integrated reasoning has higher application value for landslide risk evaluation.
The index values used to evaluate the model accuracy in table 3 show the variation of model performance under different experimental data sets. As can be seen from Table 3, the Overall Accuracy (OA), accuracy and F1 of the CBR-ASSR model are all higher than those of the CBR-ASR model under the case base with different case numbers. And under the case library of the 9:1 experimental data set, the recall rate of negative cases in the prediction result of the CBR-ASSR model is the same as the evaluation index of the CBR-ASR. Under two experimental data sets of 8:2 and 7:3, OA, recall rate, accuracy rate and F1 of the CBR-ASSR are all higher than those of the CBR-SSR; under a 9:1 experimental data set case library, the accuracy of the CBR-ASSR model is up to 91.66%, but is slightly lower than 93.05% of that of the CBR-SSR model. The result shows that the CBR-ASSR is suitable for landslide risk evaluation and can effectively solve the complex geospatial problem. And no matter what model, the accuracy rate is reduced with the reduction of case base. The accuracy of the CBR model is shown to be in positive correlation with the size and the abundance of the case base.
TABLE 3
The performance of the model can be generally evaluated through an ROC curve, the horizontal axis represents the accumulated percentage of the area of the landslide risk from high to low, the vertical axis represents the accumulated percentage of the quantity of the landslide points corresponding to the risk index, a success rate curve is drawn, the area value (AUC) under the curve is calculated to serve as an accuracy evaluation index, the AUC value is larger when the curve shape is bent to be close to the upper left of an image, and the landslide risk evaluation result is better. FIG. 7 is a ROC curve under the 9:1 experimental data set. As can be seen from FIG. 7, the prediction ability of the CBR-ASSR model performed better than the other models, with an AUC value of 0.910, and CBR-ASR and CBR-SSR values of 0.874 and 0.905, respectively. The method shows that after the spatial features are blended, the integrated inference model of the spatial similarity and the attribute similarity is superior to the traditional attribute similarity inference and the simple spatial similarity inference in the aspect of landslide risk evaluation.
In order to further verify the effectiveness of the model, the CBR-ASSR model with better performance is selected for comparison experiments with the CNN and the SVM. The excellent machine learning method is beneficial to landslide risk evaluation, the CNN has a complex structure but excellent performance, and the model performance exceeds that of ANN through convolution and pooling processing of spatial correlation; a support vector machine model based on a radial basis function is a promising landslide prediction method. The parameters of the CNN and SVM models during the experiment are shown in table 4.
TABLE 4
Comparative experiments used a 9:1 experimental data set. Fig. 8 is a drawing of evaluation of landslide risk in a research area predicted by a CNN and SVM model constructed using 90% of sample data as training data, and fig. 9 is a ratio of landslide risk areas predicted by CBR-ASSR, CNN and SVM model. Fig. 8(a) is a drawing of evaluating the landslide risk of the research area predicted by the SVM model, and fig. 8(b) is a drawing of evaluating the landslide risk of the research area predicted by the CNN model, and it can be seen from fig. 8 that the high-risk regions in the evaluation result of the SVM model are too many and dense, and are similar to CBR-ASR as a whole; the high risk area of landslide is mainly concentrated in the road network and water system. As can be seen from FIG. 9, the high risk area ratio of CBR-ASSR is very similar to the results of CNN, 0.1597 and 0.1587; CNN has the largest area ratio of low risk regions, the second to CBR-ASSR. The experimental results show that the accuracy rates of CBR-ASSR, CNN and SVM are 0.9166, 0.897 and 0.8611 respectively. The CBR-ASSR has better evaluation capability compared with CNN and SVM.
An ROC curve for landslide risk evaluation using the SVM model and the CNN model is shown in fig. 10, where AUC of the CBR-ASSR model is 0.910, and AUC of the SVM model and the CNN model are 0.839 and 0.890, respectively.
In conclusion, the CBR-ASSR model is superior to SVM and CNN models in overall precision, statistical evaluation indexes and ROC curves. The method has the advantages that the introduction of the spatial features and the improvement of the similarity reasoning method have obvious promotion effects on solving the problem of complex geographic environment, and the spatial case reasoning method for regional landslide risk evaluation provided by the embodiment of the application has a good application prospect on the regional landslide risk evaluation.
The embodiments described above are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device to perform the methods described in the embodiments or some portions of the embodiments.
The invention is not limited to the above alternative embodiments, and any other various forms of products can be obtained by anyone in the light of the present invention, but any changes in shape or structure thereof, which fall within the scope of the present invention as defined in the claims, fall within the scope of the present invention.
Claims (10)
1. A space case reasoning method for regional landslide risk assessment is characterized by comprising the following steps:
dividing an area to be evaluated into a plurality of grids, wherein each grid represents a landslide case;
calculating the spatial proximity of the landslide case and the adjacent grids thereof on the fault, the water system and the road network buffer surface;
determining a spatial topological relation between a central point of a landslide case and an equivalent surface of an area to which the landslide case belongs on each influence factor layer, wherein each influence factor layer comprises a slope layer, a lithology layer, a land utilization layer, a rainfall layer, an elevation layer, a population density layer, a vegetation coverage layer, a fault layer, a road network layer, a water system layer and a seismic center layer;
extracting attribute characteristic values of the landslide case in each influence factor layer;
calculating Euclidean distances of the unknown landslide cases and the known landslide cases on the corresponding dimension of each image layer, and obtaining attribute similarity of the unknown landslide cases and each known landslide case based on the Euclidean distances of the unknown landslide cases and each known landslide case on the corresponding dimension of each image layer;
determining the spatial proximity similarity between the unknown landslide cases and each known landslide case based on the spatial proximity of the unknown landslide cases and the adjacent grids thereof in the fault dimension, the water system dimension and the road network dimension, and the spatial proximity of each known landslide case and the adjacent grids thereof in the fault dimension, the water system dimension and the road network dimension;
determining the similarity of the spatial topological relations between the unknown landslide cases and each known landslide case on the basis of the spatial topological relations between the unknown landslide cases and the isosurface of each known landslide case on each layer;
determining the spatial comprehensive similarity of the unknown landslide cases and each known landslide case based on the spatial proximity similarity between the unknown landslide cases and each known landslide case and the spatial topological relation similarity between the unknown landslide cases and each known landslide case;
determining the similarity between the unknown landslide cases and each known landslide case based on the attribute similarity between the unknown landslide cases and each known landslide case and the spatial comprehensive similarity between the unknown landslide cases and each known landslide case;
and determining the estimated landslide state of the unknown landslide case based on the landslide state of the known landslide case with the maximum similarity to the unknown landslide case.
2. The method of claim 1, wherein the calculating Euclidean distances between the unknown landslide case and a plurality of known landslide cases with known landslide states in the corresponding dimension of each image layer comprises:
and calculating Euclidean distances between the unknown landslide case and the plurality of known landslide cases in the known landslide state on the corresponding dimension of each influence factor layer according to the attribute characteristic values of the unknown landslide cases in each influence factor layer and the attribute characteristic values of the plurality of known landslide cases in the known landslide state on each influence factor layer.
3. The method of claim 2, wherein the similarity of the attributes of the unknown landslide case to each of the known landslide cases is:
wherein Dist (X, Y) is Euclidean distance between an unknown landslide case and each known landslide case on each influence factor layer, WiAnd the weights are corresponding to the layer i.
4. The method of claim 3, wherein the spatial proximity similarity between an unknown landslide case and each of the known landslide cases is:
wherein, if
Then
For the j-th dimension spatial proximity value corresponding to the known landslide case,
a j-th dimension spatial proximity value corresponding to an unknown landslide case of an unknown landslide state,
is the weight corresponding to the proximity value.
5. The method of claim 4, wherein the spatial topological relationship similarity of an unknown landslide case to each of the known landslide cases is:
wherein the content of the first and second substances,
D(Xi,Yi) The spatial topological relation of the equivalent surfaces of the unknown landslide case and the known landslide case on a layer is represented, 0 represents that the unknown landslide case and the known landslide case do not belong to the same topological relation, 1 represents that the unknown landslide case and the known landslide case belong to the same topological relation, and q is the total number of the layers.
6. The method of claim 5, wherein the spatially integrated similarity of an unknown landslide case to each of the known landslide cases is:
wherein, WpWeights corresponding to spatial proximity similarity between an unknown landslide case and a known landslide case, WtAnd weighting corresponding to the spatial topological relation similarity of the unknown landslide case and the known landslide case.
7. The method of claim 6, wherein the similarity of an unknown landslide case to each of the known landslide cases is:
wherein, WsA weight W corresponding to the spatial comprehensive similarity of the unknown landslide case and each of the known landslide casesaAnd weighting corresponding to the similarity of the attributes of the unknown landslide case and the known landslide case.
8. The method of claim 1, wherein the spatial proximity of a landslide case to its neighboring meshes in road network, water system, fault dimensions is:
where n denotes the number of adjacent grids, PiThe spatial proximity of the landslide case and an adjacent grid i among the plurality of adjacent grids in the dimensions of a road network, a water system and a fault is shown.
9. The method of claim 1, wherein the area where the landslide case is located is an area formed by the grid corresponding to the landslide case and all the neighboring grids.
10. The method of claim 1, wherein the partitioned grid is a square grid of equal length and width.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110013321.5A CN112767544B (en) | 2021-01-06 | 2021-01-06 | Spatial case reasoning method for regional landslide risk evaluation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110013321.5A CN112767544B (en) | 2021-01-06 | 2021-01-06 | Spatial case reasoning method for regional landslide risk evaluation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112767544A true CN112767544A (en) | 2021-05-07 |
| CN112767544B CN112767544B (en) | 2022-04-01 |
Family
ID=75700108
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110013321.5A Expired - Fee Related CN112767544B (en) | 2021-01-06 | 2021-01-06 | Spatial case reasoning method for regional landslide risk evaluation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112767544B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114170427A (en) * | 2021-11-12 | 2022-03-11 | 河海大学 | Wireless microwave rain attenuation model SSIM image similarity evaluation method based on rain cells |
| CN116244863A (en) * | 2023-03-09 | 2023-06-09 | 北京理工大学 | Reflow soldering spot simulation design decision system based on multi-granularity case flexible reconstruction |
| CN116244863B (en) * | 2023-03-09 | 2024-05-17 | 北京理工大学 | Reflow soldering spot simulation design decision system based on multi-granularity case flexible reconstruction |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1700623A1 (en) * | 2006-06-01 | 2006-09-13 | Pieps GmbH | Avalanche probe and search system for burried persons |
| JP2013238540A (en) * | 2012-05-16 | 2013-11-28 | Meteorological Engineering Center | Observation method and observation system |
| CN104598766A (en) * | 2015-03-05 | 2015-05-06 | 成都理工大学 | Oil and gas reservoir comprehensive assessment oriented spatial case-based reasoning method |
| CN105184050A (en) * | 2015-08-10 | 2015-12-23 | 南京师范大学 | Landslide hazard degree evaluation method under support of GIS (Geographic Information System) and artificial intelligence technology |
| CN105678012A (en) * | 2016-01-27 | 2016-06-15 | 青岛理工大学 | Downslide thrust evaluation method for variable width slicing of side slope |
| CN106327404A (en) * | 2016-08-27 | 2017-01-11 | 重庆九洲星熠导航设备有限公司 | Multi-source data collision detection and resolution method for comprehensive disaster assessment |
| CN106884405A (en) * | 2017-03-08 | 2017-06-23 | 中国水利水电科学研究院 | Inrush type mountain flood assay method for a kind of Cross Some Region Without Data |
| CN109165424A (en) * | 2018-08-03 | 2019-01-08 | 四川理工学院 | A kind of landslide assessment of easy generation method based on domestic GF-1 satellite data |
| CN110489860A (en) * | 2019-08-16 | 2019-11-22 | 兰州交通大学 | A kind of novel landslide risk evaluating method |
| CN111080080A (en) * | 2019-11-25 | 2020-04-28 | 桂林理工大学南宁分校 | Method and system for estimating risk of geological disaster of villages and small towns |
| CN111898315A (en) * | 2020-07-27 | 2020-11-06 | 中国科学院空天信息创新研究院 | Landslide susceptibility assessment method based on fractal-machine learning hybrid model |
-
2021
- 2021-01-06 CN CN202110013321.5A patent/CN112767544B/en not_active Expired - Fee Related
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1700623A1 (en) * | 2006-06-01 | 2006-09-13 | Pieps GmbH | Avalanche probe and search system for burried persons |
| JP2013238540A (en) * | 2012-05-16 | 2013-11-28 | Meteorological Engineering Center | Observation method and observation system |
| CN104598766A (en) * | 2015-03-05 | 2015-05-06 | 成都理工大学 | Oil and gas reservoir comprehensive assessment oriented spatial case-based reasoning method |
| CN105184050A (en) * | 2015-08-10 | 2015-12-23 | 南京师范大学 | Landslide hazard degree evaluation method under support of GIS (Geographic Information System) and artificial intelligence technology |
| CN105678012A (en) * | 2016-01-27 | 2016-06-15 | 青岛理工大学 | Downslide thrust evaluation method for variable width slicing of side slope |
| CN106327404A (en) * | 2016-08-27 | 2017-01-11 | 重庆九洲星熠导航设备有限公司 | Multi-source data collision detection and resolution method for comprehensive disaster assessment |
| CN106884405A (en) * | 2017-03-08 | 2017-06-23 | 中国水利水电科学研究院 | Inrush type mountain flood assay method for a kind of Cross Some Region Without Data |
| CN109165424A (en) * | 2018-08-03 | 2019-01-08 | 四川理工学院 | A kind of landslide assessment of easy generation method based on domestic GF-1 satellite data |
| CN110489860A (en) * | 2019-08-16 | 2019-11-22 | 兰州交通大学 | A kind of novel landslide risk evaluating method |
| CN111080080A (en) * | 2019-11-25 | 2020-04-28 | 桂林理工大学南宁分校 | Method and system for estimating risk of geological disaster of villages and small towns |
| CN111898315A (en) * | 2020-07-27 | 2020-11-06 | 中国科学院空天信息创新研究院 | Landslide susceptibility assessment method based on fractal-machine learning hybrid model |
Non-Patent Citations (2)
| Title |
|---|
| 乔建平 等: "采用本底因子贡献率法的三峡库区滑坡危险度区划", 《山地学报》 * |
| 许英杰: "四川冕宁县滑坡地质灾害发育特征及风险性评价研究", 《中国优秀博硕士学位论文全文数据库(硕士)基础科学辑》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114170427A (en) * | 2021-11-12 | 2022-03-11 | 河海大学 | Wireless microwave rain attenuation model SSIM image similarity evaluation method based on rain cells |
| CN116244863A (en) * | 2023-03-09 | 2023-06-09 | 北京理工大学 | Reflow soldering spot simulation design decision system based on multi-granularity case flexible reconstruction |
| CN116244863B (en) * | 2023-03-09 | 2024-05-17 | 北京理工大学 | Reflow soldering spot simulation design decision system based on multi-granularity case flexible reconstruction |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112767544B (en) | 2022-04-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112733749B (en) | Real-time pedestrian detection method integrating attention mechanism | |
| CN106650809B (en) | A kind of vehicle-mounted laser point cloud objective classification method and system | |
| CN113642849B (en) | Geological disaster risk comprehensive evaluation method and device considering spatial distribution characteristics | |
| CN111626128A (en) | Improved YOLOv 3-based pedestrian detection method in orchard environment | |
| KR101675778B1 (en) | Decision making system corresponding to volcanic disaster | |
| CN106600617A (en) | Method of extracting building contour line from Lidar point cloud data based on curvature | |
| CN103984746B (en) | Based on the SAR image recognition methodss that semisupervised classification and region distance are estimated | |
| CN112767544B (en) | Spatial case reasoning method for regional landslide risk evaluation | |
| Dimitrios | Urban growth prediction modelling using fractals and theory of chaos | |
| CN111598359B (en) | Logistics station site selection method and system | |
| Juyal et al. | A study of landslide susceptibility mapping using machine learning approach | |
| CN104331711B (en) | SAR image recognition methods based on multiple dimensioned fuzzy mearue and semi-supervised learning | |
| CN110705606A (en) | Spatial K-means clustering method based on Spark distributed memory calculation | |
| CN111080080B (en) | Village geological disaster risk prediction method and system | |
| CN110097636B (en) | Site selection planning method based on visual field analysis | |
| Huang et al. | Landslide susceptibility evaluation using different slope units based on BP neural network | |
| CN114511571A (en) | Point cloud data semantic segmentation method and system and related components | |
| LU502699B1 (en) | Method and system for predicting risk of geological disasters in villages and towns | |
| Fang et al. | Zonation and scaling of tropical cyclone hazards based on spatial clustering for coastal China | |
| Zhang et al. | A Simplified Approach to DEM Based on Important Points and the Triangular Irregular Network Angles | |
| Shijin et al. | Comparative study of feature dimension reduction algorithm for high-resolution remote sensing image classification | |
| Zhou et al. | A Distributed Storage Strategy For Trajectory Data Based On Nosql Database | |
| Wang et al. | DeepGCNs-Att for Point Cloud Semantic Segmentation | |
| Kim et al. | An enhanced density-based spatial separation strategy of 3D point clouds | |
| CN103955696A (en) | SAR image classification based on random observation projection and sparse representation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20220401 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |