System for estimating cluster phenomena

Abstract

The objective of the present invention is to solve the problems of the prior art, namely, to provide a landslide estimation system that can evaluate whether a landslide has occurred by comparing images from two different time periods. [Solution] The ground convergence phenomenon estimation system of the present invention is a system for estimating ground convergence phenomena based on a first terrain model and a second terrain model obtained at different times in the same target area, and comprises a pixel value setting means, a cross-correlation coefficient calculation means, a movement speed calculation means, and a landslide evaluation means. Of these, the landslide evaluation means evaluates that a landslide has occurred in the inspection area when the maximum cross-correlation coefficient exceeds a first correlation threshold and the movement speed falls below a speed threshold, and sets a landslide area based on the inspection area.

Description

Technical Field

[0001] The invention of the present application relates to a technique for estimating mass movement phenomena of the ground such as landslides and collapses. More specifically, it relates to a mass movement estimation system capable of estimating landslide areas and collapse areas by comparing topographic images obtained by imaging a topographic model based on topographic quantities at two times.

Background Art

[0002] It is said that two-thirds of the land in Japan is mountainous. As a result, many people live on land with slopes behind them, and roads and railway lines always have sections passing along the sides of slopes. And slopes have the potential for disasters such as landslides and collapses, and have suffered serious damage many times in the past.

[0003] Both landslides and collapses are part of the mass movement phenomenon of the ground, and both are called "landslide" in English. On the other hand, in Japan, which has been troubled by slope disasters, landslides and collapses are generally classified as follows. That is, "landslide" is a mass movement phenomenon in which the movement is slow, most of the moving soil masses are not crushed and keep their original shapes, and sediment often remains in the slope near the landslide head. In contrast, "collapse" is a mass movement phenomenon in which the movement is fast, almost none of the moving soil masses are loosened enough to keep their original shapes, and it reaches far away from the collapse head.

[0004] While slow-moving landslides are sometimes known and regularly monitored, it's possible that landslides exist whose existence is still unknown, potentially causing latent activity. Although countermeasures can be taken for monitored landslides according to their movement, it's difficult to implement measures for undetected landslides. Therefore, their existence may only become known when significant movement occurs, and in some cases, damage may occur before the landslide is even recognized. Consequently, detecting undetected landslides is extremely beneficial in preventing damage.

[0005] On the other hand, rapid landslides make it difficult to detect precursory phenomena, and therefore, the current situation is that a landslide area is only recognized as such after damage has actually occurred. In other words, it is difficult to detect areas that are likely to collapse (so to speak, potential landslide areas) before a collapse occurs, and because of the rapid movement, it is not practical to take countermeasures once the behavior begins. Moreover, it is expected that there are many cases in which a landslide occurs but no damage occurs, and in these cases, the fact that a landslide has occurred is rarely recognized. On the other hand, the surrounding areas of a landslide should have the same or very similar conditions, such as geology and topography (especially slope). Therefore, detecting areas where a landslide has already occurred but no damage has occurred is thought to suggest the possibility of landslides in the surrounding area, and is extremely useful in particular when there are areas to be protected, as it allows for the taking of preventative measures.

[0006] On slopes showing signs of landslides, measurements are sometimes taken to monitor their movement. For example, measurements using extensometers and extrusion plates, measurements using in-situ inclinometers, and measurements of ground surface displacement have been carried out. Furthermore, for slopes that have been assessed as being at risk of collapse through prior inspections, it is conceivable to estimate the movement speed of surface soil from videos acquired by fixed-point cameras. Patent document 1 proposes a technology for monitoring the movement of a slope using images of the slope. Patent document 2 discloses a technology that visualizes a topographic model based on the topographic quantities of each mesh, and by comparing these images at two different time periods, it is possible to grasp horizontal changes. [Prior art documents] [Patent Documents]

[0007] [Patent Document 1] Japanese Patent Publication No. 2019-201275 [Patent Document 2] Japanese Patent Publication No. 2010-266419 [Overview of the project] [Problems that the invention aims to solve]

[0008] Incidentally, there is currently no known method to quantitatively distinguish between landslides and collapses. As mentioned above, in areas showing signs of landslides, measurements using extensometers and shims, measurements using borehole inclinometers, and measurements of ground surface displacement are sometimes carried out, but these can only be carried out in cases where the extent and depth of the landslide are predetermined. Moreover, while it is possible to monitor the behavior of landslides, it is not possible to detect unknown landslides, much less collapses that have already occurred. Furthermore, even methods such as monitoring using video acquired by fixed-point cameras, as described in Patent Document 1, can only be applied in places where landslides or collapses are predicted in advance, and they cannot quantitatively distinguish between landslides and collapses.

[0009] The object of the present invention is to solve the problems of the prior art, namely, to provide a landslide estimation system that can evaluate whether a landslide has occurred by comparing topographic images based on topographic models from two different time periods. [Means for solving the problem]

[0010] The present invention focuses on distinguishing between "landslides" and "collapses" based on the speed of ground movement that occurred between two periods, and is an invention based on a novel idea.

[0011] The ground convergence phenomenon estimation system of the present invention is a system for estimating ground convergence phenomena based on a first terrain model and a second terrain model obtained at different times in the same target area, and comprises a pixel value setting means, a cross-correlation coefficient calculation means, a movement speed calculation means, and a landslide evaluation means. Of these, the pixel value setting means calculates the topographic quantity of the meshes constituting the first terrain model and the second terrain model, and assigns a pixel value corresponding to the topographic quantity to each mesh. The cross-correlation coefficient calculation means sets an inspection area consisting of multiple meshes for the first terrain model, sets an exploration area larger than the inspection area for the second terrain model, and then scans the exploration area with the inspection area to obtain a cross-correlation coefficient based on the pixel values, and extracts the maximum cross-correlation coefficient which is the largest among the multiple cross-correlation coefficients. The movement speed calculation means calculates the movement speed based on the displacement amount obtained based on the coordinates of the corresponding area showing the maximum cross-correlation coefficient within the exploration area and the inspection area, the acquisition time of the first terrain model, and the acquisition time of the second terrain model. The landslide evaluation means evaluates that a landslide has occurred in the inspection area when the maximum cross-correlation coefficient exceeds a predetermined first correlation threshold and the movement speed falls below a predetermined speed threshold, and sets the landslide area based on the inspection area.

[0012] The collective motion estimation system of the present invention can also be configured to combine two or more adjacent inspection areas that have been evaluated as having experienced a landslide and set them as a landslide area.

[0013] The landslide phenomenon estimation system of the present invention may further include an elevation difference calculation means. This elevation difference calculation means uses the elevations of each mesh located at the same position in the inspection area and the exploration area, and calculates the elevation difference by subtracting the elevation of the mesh from the elevation of the mesh from the elevation of the mesh from the elevation of the mesh from the mesh from the newer period. In this case, the landslide evaluation means sets a landslide area when an elevation difference exceeding a first elevation threshold is detected from a mesh included in the low-elevation end area of ​​the landslide area, and an elevation difference below a second elevation threshold is detected from a mesh included in the high-elevation end area of ​​the landslide area.

[0014] The clustering phenomenon estimation system of the present invention may further include a collapse evaluation means. This collapse evaluation means evaluates that a collapse has occurred in the inspection area when the maximum cross-correlation coefficient falls below a predetermined second correlation threshold and an elevation difference exceeding a predetermined first elevation threshold and an elevation difference below a predetermined second elevation threshold are detected, and sets a collapse area based on the inspection area.

[0015] The landslide phenomenon estimation system of the present invention may also define landslide areas based on the elevation of the mesh. In this case, the landslide evaluation means sets a landslide area based on the inspection area when an elevation difference exceeding a first elevation threshold is detected from a mesh included in the low-elevation end region of the landslide area, and an elevation difference below a second elevation threshold is detected from a mesh included in the high-elevation end region of the landslide area. [Effects of the Invention]

[0016] The clustering phenomenon estimation system of the present invention has the following effects: (1) By using topographic images that cover a wide area, it is possible to detect landslides and collapse areas that have not yet been identified. As a result, damage can be prevented by taking appropriate measures according to the situation. (2) It is possible to objectively estimate the classification of clustering phenomena without relying on the subjective opinions of engineers, etc. (3) Since it does not require expensive measuring instruments and there is no need to arrange monitors over a long period of time, it is possible to estimate the classification of the collective movement phenomenon at a low cost compared to the prior art.

Brief Description of the Drawings

[0017] [Figure 1] A block diagram showing the main configuration of the collective movement phenomenon estimation system of the present invention. [Figure 2] (a) is a model diagram schematically showing an inspection area composed of a plurality of meshes, and (b) is a model diagram schematically showing a search area composed of a plurality of meshes. [Figure 3] A model diagram showing the inspection area superimposed on the corresponding area. [Figure 4] A model diagram schematically showing a landslide area set by aggregating a plurality of partial landslide areas. [Figure 5] (a) is a cross-sectional view schematically showing a slope before sliding occurs, and (b) is a cross-sectional view schematically showing a slope after landslide occurs. [Figure 6] [[ID=JIN24]]A flowchart showing the main processing flow from "calculation of terrain quantity" to "extraction of target inspection area" among the processes related to the collective movement phenomenon estimation system of the present invention. [Figure 7] A flowchart showing the main processing flow up to "setting of landslide area" among the processes related to the collective movement phenomenon estimation system of the present invention. [Figure 8] A flowchart showing the main processing flow up to "setting of collapse area" among the processes related to the collective movement phenomenon estimation system of the present invention. [Figure 9] A discrimination criterion diagram showing an example of conditions for discriminating the classification of the collective movement phenomenon. <着

Mode for Carrying Out the Invention

[0018] [[ID=4r]]An example of an embodiment of the collective movement phenomenon estimation system of the present invention will be described based on the drawings.

[0019] It should be noted that there seems to be an incorrect tag "着0000094" in the original text which is retained as is in the translation. Also, the tag "JIN24" in the translation of line 24 is likely a mislabeling during the translation process and should be corrected to the original tag if possible.Figure 1 is a block diagram showing the main components of the motion clustering phenomenon estimation system 100 of the present invention. As shown in this figure, the motion clustering phenomenon estimation system 100 of the present invention is configured to include a pixel value setting means 101, a cross-correlation coefficient calculation means 102, a movement speed calculation means 103, and a landslide evaluation means 104. It can also be configured to include an elevation difference calculation means 105, a collapse evaluation means 106, a displacement amount calculation means 107, a topography evaluation means 108, a topography model storage means 109, and the like.

[0020] Each component of the motion phenomenon estimation system 100 (particularly the pixel value setting means 101 to the terrain evaluation means 108) can be manufactured as a dedicated unit, or a general-purpose computer device can be used. In other words, the computer device performs calculations according to a predetermined program, thereby performing processing specific to each component. This computer device is equipped with a processor such as a CPU, memory such as ROM or RAM, and may also include input means such as a mouse or keyboard and a display. For example, it can be composed of a personal computer (PC) or a server.

[0021] Furthermore, the terrain model storage means 109 can utilize the storage device of a general-purpose computer (e.g., a personal computer) or it can be built on a database server. When built on a database server, it can be located on a local network (LAN: Local Area Network) or it can be a cloud server that stores data via the internet.

[0022] The following describes in detail each of the main elements constituting the clustering phenomenon estimation system 100 of the present invention.

[0023] (Means for storing topographic models) The terrain model storage means 109 is a means for storing terrain models. When the ground is measured by aerial laser measurement or aerial photogrammetry, a large amount of measurement point data with three-dimensional coordinates (hereinafter referred to as "three-dimensional point cloud") is obtained. Here, "terrain model" refers to a model generated based on this three-dimensional point cloud, and examples include DSM (Digital Surface Model) and DEM (Digital Elevation Model). Typically, terrain models are composed of small regions (hereinafter referred to as "meshes") obtained by dividing the planar range of the measurement target into multiple parts. These meshes are, for example, sections formed by dividing them into orthogonal grids, and each mesh is assigned an elevation. Since the three-dimensional point cloud obtained by aerial laser measurement consists of random data (data with an irregular arrangement in the plane), geometric calculations are often performed to assign an elevation to each mesh. Methods for calculating this height include the Triangulated Irregular Network (TIN) method, which uses an irregular triangular network formed from random data to determine the height; the Nearest Neighbor method, which uses the nearest laser measurement point; Inverse Distance Weighting (IDW); the Kriging method; and the averaging method.

[0024] The terrain model storage means 109 stores two or more terrain models obtained at different times in the same target area. The clustering phenomenon estimation system 100 of the present invention reads out the two terrain models stored in the terrain model storage means 109 and performs the respective processes. For convenience, here we will refer to one of the two terrain models as the "first terrain model" and the other as the "second terrain model". Note that the one measured earlier (i.e., the older one) may be designated as the first terrain model, or the one measured later (i.e., the newer one).

[0025] (Pixel value setting means) The pixel value setting means 101 is a means for assigning a "pixel value" to each mesh that constitutes the terrain model. Here, a pixel value is a value set in a "color model" that can be handled by a computer, or a value set in grayscale. Examples of color models include RGB, which uses three basic colors: red, green, and blue; CMYK, which uses four basic colors: cyan, magenta, yellow, and black (key color); and NCS and the Ostwald color system, which use six basic colors: yellow, red, blue, green, black, and white. The procedure for assigning a pixel value to each mesh by the pixel value setting means 101 will be described below.

[0026] First, the first and second terrain models are read from the terrain model storage means 109. Next, a "terrain quantity" is calculated for each mesh that makes up the first and second terrain models. Here, the terrain quantity is an indicator that represents the characteristics of the terrain, and is calculated based on representative points of the mesh or the original random data. For example, the terrain quantity can be elevation value, Laplacian value, surface openness value, subsurface openness value, slope, curvature, or a combination of these.

[0027] The Laplacian value is used to draw a Laplacian diagram, which represents the rate of change of slope. This Laplacian diagram has the characteristic of being positive for depressed terrain and negative for protruding terrain, with the absolute value being larger where the change in terrain is large. The surface openness value is used to draw an openness diagram. This surface openness diagram represents the area of ​​the sky visible within a certain distance from the point of interest, and has the characteristic of emphasizing protruding mountain peaks and ridges. The subsurface openness value is used to draw an openness diagram. This subsurface openness diagram represents the area of ​​the subsurface within a certain distance when looking underground from the ground surface, and has the characteristic of emphasizing depressions and valleys. The slope amount is used to draw a slope amount diagram. This slope amount diagram shows the degree of slope of the terrain, with a larger slope indicating a greater slope and a smaller slope indicating a gentler slope. Curvature is an index that represents the unevenness of the terrain, with positive curvature indicating a convex shape and negative curvature indicating a concave shape.

[0028] Once the terrain quantities are calculated for each mesh constituting the first and second terrain models, an image is created based on these terrain quantities. For convenience, the image created based on the terrain quantities will be referred to here as a "terrain image," and specifically, the one created from the first terrain model will be called the "first terrain image," and the one created from the second terrain model will be called the "second terrain image." Specifically, the first terrain image is created by assigning pixel values ​​(color model or grayscale) corresponding to the terrain quantities to each mesh constituting the first terrain model, and the second terrain image is created by assigning pixel values ​​corresponding to the terrain quantities to each mesh constituting the second terrain model. Note that in order to create a terrain image, pixels are set and then pixel values ​​are assigned to those pixels, but for convenience, this explanation will use the example where one mesh is one pixel. Of course, it is also possible to use a specification where one pixel is composed of multiple meshes.

[0029] (Means for calculating cross-correlation coefficient) The cross-correlation coefficient calculation means 102 is a means for determining the "cross-correlation coefficient" based on pixel values ​​while scanning the "exploration area" in the "inspection area". The procedure by which the cross-correlation coefficient calculation means 102 calculates the cross-correlation coefficient will be described below.

[0030] First, the cross-correlation coefficient calculation means 102 sets an inspection area for the first terrain model (first terrain image) and sets an exploration area for the second terrain model (second terrain image). As described above, the older of the two terrain models can be designated as the first terrain model, or the newer one can be designated as the first terrain model. In other words, it is possible to set an inspection area for the older terrain model and an exploration area for the newer terrain model, or to set an inspection area for the newer terrain model and an exploration area for the older terrain model.

[0031] The inspection area and the exploration area are each composed of multiple meshes, as shown in Figure 2. However, as shown in this figure, the exploration area is larger than the inspection area. Figure 3(a) is a schematic model diagram showing an inspection area composed of multiple meshes, and Figure 3(b) is a schematic model diagram showing an exploration area composed of multiple meshes. The inspection area is set within the range of the exploration area and is set at multiple locations while moving (arrows in Figure 3(a)) to cover the exploration area. In this case, it is good practice to set multiple inspection areas while moving (shifting) them in units of, for example, one mesh (one pixel). The exploration area, on the other hand, has a fixed range and does not move. However, the exploration area can be set at one location or at two or more locations.

[0032] Once the inspection area and exploration area are defined, a portion of the first terrain image corresponding to the inspection area (hereinafter referred to as the "inspection image") is extracted. Next, the inspection image is used to perform image matching with the image of the second terrain image corresponding to the exploration area (hereinafter referred to as the "exploration image"). Specifically, as shown by the arrows in Figure 3(b), the inspection image is scanned to cover the entire exploration image, and the inspection image and exploration image are matched. At this time, it is preferable to scan while moving (shifting) in mesh units (one or more pixels). The result of this matching is then calculated as a cross-correlation coefficient. Since the inspection image moves within the exploration area, it is matched with the matched area in the exploration image (hereinafter referred to as the "matching area") at each new location, meaning that the cross-correlation coefficient is calculated each time it moves. Furthermore, as mentioned above, the inspection image is set at multiple locations while moving, so the cross-correlation coefficient is calculated while manipulating the exploration image for each set inspection image. For example, if 16 matching areas are set to cover the exploration area, the inspection image will be compared with the exploration image at 16 locations, meaning that in this case, a 16 × 16 cross-correlation coefficient will be calculated.

[0033] The cross-correlation coefficient is a value calculated based on pixel values ​​and is essentially an index value indicating the similarity between the inspection image and the matching region in the exploration image. Various conventional methods can be used to calculate this cross-correlation coefficient, as long as they can represent the similarity between the inspection image and the matching region. For example, it can be calculated using the cross-correlation method in particle image velocity measurement (PIV). The cross-correlation coefficient is determined as the similarity between the inspection image and the matching region, and is calculated for each of multiple inspection images.

[0034] Furthermore, the cross-correlation coefficient calculation means 102 extracts the maximum value of the cross-correlation coefficient (hereinafter referred to as the "maximum cross-correlation coefficient") for each matching area. For example, if one inspection image moves within the exploration area and cross-correlation coefficients are calculated at 16 locations, the maximum of the 16 cross-correlation coefficients is extracted as the maximum cross-correlation coefficient. In addition, the maximum cross-correlation coefficient is extracted for each set inspection image, so if, for example, inspection images are set at 16 locations to cover the exploration area, the maximum cross-correlation coefficient is extracted for each of the 16 inspection images.

[0035] (Means for calculating displacement) The displacement calculation means 107 detects the region showing the maximum cross-correlation coefficient within the matching region (hereinafter referred to as the "corresponding region") and determines the displacement amount based on the coordinates of the corresponding region and the inspection region. The procedure by which the displacement calculation means 107 determines the displacement amount will be described below.

[0036] First, the displacement calculation means 107 detects the range related to the maximum cross-correlation coefficient within the matching area (matching image) for each set inspection area (inspection image) as the "corresponding area". Therefore, the corresponding area will be the same size as the inspection area, as shown in Figure 3, but its position may differ. Note that in Figure 3, the inspection area is displayed superimposed on the second terrain image (exploration image) for convenience.

[0037] When a corresponding area is detected for each inspection area, as shown in Figure 3, a "displacement vector" is calculated for each corresponding area based on the inspection area and the corresponding area, and the displacement amount, which is the magnitude of the displacement vector, is also calculated. However, the movement vector is calculated with the older of the inspection area and the newer of the corresponding area as the starting point and the newer one as the ending point. In other words, Figure 3 shows an example where the inspection area (first terrain image) is older and the corresponding area (second terrain image) is newer. Since the first terrain model and the second terrain model are composed of 3D point clouds, it is also possible to calculate the horizontal displacement vector as a 2D (horizontal plane) vector and then calculate its magnitude (horizontal displacement amount), or to calculate the 3D displacement vector as a 3D (space) vector and then calculate its magnitude (3D displacement amount).

[0038] As shown in Figure 3, one type of displacement vector and displacement amount are determined based on one set of inspection areas and corresponding areas. In subsequent processing, evaluations such as landslides and collapses are performed on the inspection areas. The system can be configured to evaluate all inspection areas, or it can be configured to evaluate only inspection areas that meet predetermined conditions (hereinafter referred to as "target inspection areas"). For example, areas in which the displacement amount exceeds a predetermined threshold (hereinafter referred to as "displacement threshold") can be selected as target inspection areas. For convenience, the subsequent processing will be explained here using an example of selecting target inspection areas.

[0039] (Movement speed calculation means) The movement speed calculation means 103 is a means for calculating the speed at which the target inspection area moves (hereinafter referred to as "movement speed"). Specifically, the movement speed is calculated based on the amount of displacement related to the target inspection area, the time period at which measurements were taken for the first terrain model (hereinafter referred to as "first period"), and the time period at which measurements were taken for the second terrain model (hereinafter referred to as "second period"). In other words, the time difference between the first period and the second period is determined, and the value obtained by dividing the amount of displacement by that time difference is the movement speed.

[0040] (Landslide evaluation method) The landslide evaluation means 104 is a means for evaluating whether or not a landslide has occurred in the target inspection area. The procedure by which the landslide evaluation means 104 evaluates the occurrence of a landslide will be described below.

[0041] As previously described, landslides are a phenomenon of concentrated movement in which the moving soil mass is slow and largely retains its original shape without being crushed. Therefore, the landslide evaluation means 104 compares the maximum cross-correlation coefficient related to the target inspection area with a predetermined threshold (hereinafter referred to as the "first correlation threshold"), and further compares the movement speed related to the target inspection area with a predetermined threshold (hereinafter referred to as the "velocity threshold"). When the maximum cross-correlation coefficient exceeds the first correlation threshold (or becomes equal to or greater than the first correlation threshold), and the movement speed falls below the velocity threshold (or becomes less than or equal to the velocity threshold), the target inspection area is evaluated as having experienced a landslide. In other words, the maximum cross-correlation coefficient exceeding the first correlation threshold is judged to mean that "the moving soil mass is not crushed and retains its original shape," and the movement speed falling below the velocity threshold is judged to mean that "the movement is slow." For convenience, the target inspection area evaluated by the landslide evaluation means 104 as having experienced a landslide is referred to here as a "partial landslide area."

[0042] Furthermore, the landslide evaluation means 104 sets a "landslide area" which represents the entire range of the landslide based on the partial landslide areas. At this time, one partial landslide area can be set as the landslide area as is, or the landslide area can be set by combining multiple adjacent (including partially overlapping) partial landslide areas. For example, in Figure 4, the landslide area (shaded area) is set by combining seven partially overlapping partial landslide areas.

[0043] The landslide evaluation means 104 can also extract partial landslide areas based on the direction of the displacement vector related to the target inspection area, in addition to the conditions based on the first correlation threshold and the velocity threshold. As described above, the older movement vector is considered the starting point and the newer movement vector the ending point, so the target inspection area is evaluated as a partial landslide area when its direction is at least downward. In this case, it is also possible to specify that the target inspection area is evaluated as a partial landslide area when the displacement vector falls within a predetermined range.

[0044] Incidentally, after a landslide or collapse occurs, as shown in Figure 5, the upper part tends to subside, while the lower part tends to retain soil and other debris. Figure 5(a) is a schematic cross-sectional view of the slope before a landslide occurs, and Figure 5(b) is a schematic cross-sectional view of the slope after a landslide occurs. Note that in Figure 5(b), the slope before the landslide occurs is shown with a dashed line. In other words, the elevation of the higher-elevation end region of the landslide area (hereinafter simply referred to as the "high-elevation end region") decreases, while the elevation of the lower-elevation end region (hereinafter simply referred to as the "low-elevation end region") may increase.

[0045] Therefore, the landslide evaluation means 104 can be configured to initially set a landslide area provisionally, but then determine the landslide area according to the change in elevation in the high-elevation end area and the low-elevation end area. In this case, the elevation difference calculation means 105 calculates the "elevation difference". Here, the elevation difference is the value obtained by subtracting the elevation of the old period from the elevation of the new period. More specifically, the elevation difference is obtained by using the elevations of the respective meshes at the same location in the inspection area and the exploration area, and subtracting the elevation of the old period mesh from the elevation of the new period mesh.

[0046] The landslide evaluation means 104 then determines a landslide area when it detects an elevation difference from among the meshes included in the low elevation end area of ​​the provisionally defined landslide area that exceeds a predetermined threshold (hereinafter referred to as the "first elevation threshold") (or is equal to or greater than the first elevation threshold), and when it detects an elevation difference from among the meshes included in the high elevation end area that falls below a predetermined threshold (hereinafter referred to as the "second elevation threshold") (or is equal to or less than the second elevation threshold). However, the first elevation threshold is set as a positive value, and the second elevation threshold is set as a negative value. When defining the low elevation end area and the high elevation end area, a predetermined number (or predetermined proportion) of meshes in descending order of elevation value from among the meshes constituting the provisionally defined landslide area can be designated as the low elevation end area, and a predetermined number (or predetermined proportion) of meshes in descending order of elevation value can be designated as the high elevation end area.

[0047] (Method for evaluating collapse) The collapse evaluation means 106 is a means for evaluating whether or not collapse has occurred in the target inspection area. The procedure by which the collapse evaluation means 106 evaluates the occurrence of collapse will be described below.

[0048] As previously described, collapse is a rapid, concentrated movement phenomenon in which the moving soil mass is sufficiently loosened and hardly any part retains its original shape. Therefore, the collapse evaluation means 106 compares the maximum cross-correlation coefficient for the target inspection area with a predetermined threshold (hereinafter referred to as the "second correlation threshold"), and further focuses on the elevation difference calculated by the elevation difference calculation means 105. When the maximum cross-correlation coefficient falls below the second correlation threshold (or becomes less than or equal to the second correlation threshold), and when elevation differences exceeding the first elevation threshold and elevation differences below the second elevation threshold are detected, the target inspection area is evaluated as having experienced a collapse. In other words, the maximum cross-correlation coefficient exceeds the first correlation threshold, and positive elevation differences (e.g., uplifted areas) and negative elevation differences (e.g., subsided areas) are confirmed, leading to the judgment that "the moving soil mass has been sufficiently loosened and hardly any part retains its original shape." For convenience, the inspection area in which a landslide has occurred, as determined by the displacement calculation means 107, is referred to here as the "partial collapse area."

[0049] Furthermore, the collapse evaluation means 106 sets a "collapse area" which represents the entire range of the collapsed area based on the partial collapse areas. At this time, one partial collapse area can be set as the collapse area as is, or multiple adjacent (including partially overlapping) partial collapse areas can be combined to set the collapse area. In addition, the collapse evaluation means 106, like the landslide evaluation means 104, can be configured to initially set a collapse area provisionally, but then determine the collapse area according to the change in elevation in the high-elevation end area and the low-elevation end area. Specifically, when an elevation difference exceeding the first elevation threshold (or becoming equal to or greater than the first elevation threshold) is detected from among the meshes included in the low-elevation end area of ​​the provisionally set collapse area, and an elevation difference below the second elevation threshold (or becoming equal to or less than the second elevation threshold) is detected from among the meshes included in the high-elevation end area, the area is determined to be a collapse area.

[0050] (Process flow) The main processes of the kinetic phenomenon estimation system 100 of the present invention will be explained in detail below with reference to Figures 6 to 8. Figure 6 is a flowchart showing the flow of the main processes related to the kinetic phenomenon estimation system 100 of the present invention from "calculation of topographic quantities" to "extraction of target inspection areas," Figure 7 is a flowchart showing the flow of the main processes up to "setting of landslide areas," and Figure 8 is a flowchart showing the flow of the main processes up to "setting of collapse areas." In Figures 6 to 8, the central column shows the processes to be performed, the left column shows what is necessary for those processes, and the right column shows what results from those processes.

[0051] In extracting a target inspection area using the clustering phenomenon estimation system 100 of the present invention, as shown in Figure 6, first, the topographic quantity is calculated for each mesh constituting the first topographic model and the second topographic model (Step 201 in Figure 6). Next, a first topographic image and a second topographic image are created by assigning pixel values ​​corresponding to the topographic quantity to each mesh (Step 202 in Figure 6). Once the topographic images are created, the inspection area is set for the first topographic model (first topographic image) (Step 203 in Figure 6), and the exploration area is set for the second topographic model (second topographic image) (Step 204 in Figure 6).

[0052] Once the inspection area and exploration area are set, the system scans the inspection image to cover the exploration image, comparing the inspection area with the matching area (Step 205 in Figure 6), and calculating the cross-correlation coefficient for each (Step 206 in Figure 6). Image matching (Step 205) and calculation of the cross-correlation coefficient (Step 206) are repeated until the exploration image is covered and all set matching areas are covered. Then, the maximum cross-correlation coefficient is extracted for each inspection area (Step 207 in Figure 6), and the corresponding area showing the maximum cross-correlation coefficient among the exploration areas is detected (Step 208 in Figure 6).

[0053] When a corresponding area is detected, the displacement vector and displacement amount are calculated based on the inspection area and the corresponding area (Step 209 in Figure 6). However, the movement vector is calculated using the older of the inspection area and the newer of the corresponding area as the starting point and ending point as the newer one. Then, areas where the displacement amount related to the inspection area exceeds the displacement amount threshold are selected as the target inspection area (Step 210 in Figure 6).

[0054] When setting a landslide area using the collective motion estimation system 100 of the present invention, as shown in Figure 7, the movement speed is first calculated based on the difference between the first period and the second period, and the amount of displacement (Step 211 in Figure 7). Then, when the maximum cross-correlation coefficient exceeds the first correlation threshold (Yes in Step 212 in Figure 7), and the movement speed is below the velocity threshold (Yes in Step 213 in Figure 7), the target inspection area is evaluated as a partial landslide area (Step 214 in Figure 7).

[0055] When a partial landslide area is evaluated, the landslide area is defined based on that partial landslide area (Step 215 in Figure 7). At this time, one partial landslide area can be defined as the landslide area as is, or multiple adjacent partial landslide areas can be combined to define the landslide area. Furthermore, the landslide area defined here is considered provisional, and the landslide area can be determined according to the change in elevation in the high-elevation end area and the low-elevation end area. Specifically, when an elevation difference exceeding the first elevation threshold is detected in the mesh included in the low-elevation end area of ​​the provisionally defined landslide area, and an elevation difference below the second elevation threshold is detected in the mesh included in the high-elevation end area, the area is determined as a landslide area.

[0056] In setting a collapse area using the kinetic phenomenon estimation system 100 of the present invention, the elevation difference is first calculated by the elevation difference calculation means 105, as shown in Figure 8 (Step 221 in Figure 8). Specifically, the elevation of each mesh located at the same position in the inspection area and the exploration area is used, and the elevation difference is calculated by subtracting the elevation of the mesh from

[0057] When a partial collapse area is evaluated, the collapse area is defined based on that partial collapse area (Step 225 in Figure 8). At this time, one partial collapse area can be defined as the collapse area as is, or multiple adjacent partial collapse areas can be combined to define the collapse area. Furthermore, the collapse area defined here is considered provisional, and the collapse area can be determined according to the change in elevation in the high-elevation end area and the low-elevation end area. Specifically, when an elevation difference exceeding the first elevation threshold is detected in a mesh included in the low-elevation end area of ​​the provisionally defined collapse area, and an elevation difference below the second elevation threshold is detected in a mesh included in the high-elevation end area, the area is determined to be a collapse area.

[0058] If an area is not designated as a landslide or collapse area in the processing flow shown in Figures 7 and 8, it can also be identified as another type of seismic activity by the topographic evaluation means 108 (Figure 1). Figure 9 is a discrimination criterion diagram showing an example of the conditions for distinguishing between seismic activity categories. In this diagram, cases are broadly divided into two categories: one where the maximum cross-correlation coefficient exceeds the first correlation threshold (the moving soil mass is not crushed and retains its original shape) and another where the maximum cross-correlation coefficient falls below the second correlation threshold (the moving soil mass is sufficiently loosened and there are almost no parts that retain their original shape). These categories are then further subdivided for discrimination. For example, in the case where the maximum cross-correlation coefficient exceeds the first correlation threshold, when a generally positive elevation difference is detected, it is classified as "local uplift"; when a generally negative elevation difference is detected, it is classified as "local subsidence or sinkhole"; and when a generally minute elevation difference or a generally uniform (positive or negative) elevation difference is detected, it is classified as "regional uplift or subsidence."

[0059] On the other hand, in cases where the maximum cross-correlation coefficient falls below the second correlation threshold, if a generally uniform (positive or negative) elevation difference is detected, it is classified as "artificially constructed land due to land alteration, etc." and if a generally uniform (positive or negative) elevation difference is detected on the coast, it is classified as "land newly formed by coastal uplift, volcanic eruptions, artificial reclamation, etc., or land lost due to coastal subsidence, coastal erosion, etc." [Industrial applicability]

[0060] The landslide phenomenon estimation system of the present invention can be used on all types of terrain, including slopes in mountainous areas, as well as large-scale cut slopes and embankment slopes. According to the present invention, effective measures can be taken to detect potential landslide and collapse areas, and as a result, slope disasters can be prevented. Therefore, the present invention is not only industrially applicable but also has the potential to make a significant contribution to society. [Explanation of symbols]

[0061] 100 The present invention's system for estimating cluster phenomena 101 Pixel value setting means (of the motion accumulation phenomenon estimation system) 102 Means for calculating the cross-correlation coefficient (of the clustering phenomenon estimation system) 103 (Means for calculating the movement speed of the motion concentration phenomenon estimation system) 104 Landslide evaluation means (of the landslide phenomenon estimation system) 105 (Means for calculating elevation difference of the motion concentration phenomenon estimation system) 106 (Means for evaluating the collapse of the clustering phenomenon estimation system) 107 (Means for calculating displacement of the motion accumulation phenomenon estimation system) 108 (Topographic evaluation means of the motion concentration phenomenon estimation system) 109 (Terrain model storage means of the clustering phenomenon estimation system)

Claims

1. A system for estimating ground compaction phenomena based on a first topographic model and a second topographic model obtained at different times in the same target area, A pixel value setting means for calculating the terrain quantity of the meshes constituting the first terrain model and the second terrain model, and for assigning pixel values ​​to each of the meshes according to the terrain quantity, A means for calculating a cross-correlation coefficient that sets an inspection area consisting of multiple meshes for the first terrain model, sets an exploration area larger than the inspection area for the second terrain model, calculates a cross-correlation coefficient based on the pixel values ​​while scanning the exploration area with the inspection area, and extracts the maximum cross-correlation coefficient which is the largest among the multiple cross-correlation coefficients, A means for calculating the moving speed based on the displacement amount determined from the coordinates of the corresponding area showing the maximum cross-correlation coefficient among the exploration areas and the inspection area, the acquisition time of the first terrain model, and the acquisition time of the second terrain model, The system includes a landslide evaluation means that evaluates that a landslide has occurred in the inspection area and sets a landslide area based on the inspection area when the maximum cross-correlation coefficient exceeds a predetermined first correlation threshold and the movement speed falls below a predetermined speed threshold, A system for estimating cluster phenomena, characterized by the following features.

2. The landslide evaluation means sets together two or more adjacent inspection areas that have been evaluated as having experienced a landslide as the landslide area. The clustering phenomenon estimation system according to claim 1, characterized in that it is described above.

3. The system further includes an elevation difference calculation means that uses the elevations of the respective meshes located at the same position in the inspection area and the exploration area to calculate the elevation difference by subtracting the elevation of the mesh in the older period from the elevation of the mesh in the newer period. The landslide evaluation means sets a landslide area when an elevation difference exceeding a first elevation threshold is detected from the mesh included in the low-elevation end region of the landslide area, and an elevation difference below a second elevation threshold is detected from the mesh included in the high-elevation end region of the landslide area. The clustering phenomenon estimation system according to claim 1, characterized in that it is described above.

4. A system for estimating ground compaction phenomena based on a first topographic model and a second topographic model obtained at different times in the same target area, A pixel value setting means for calculating the terrain quantity of the meshes constituting the first terrain model and the second terrain model, and for assigning pixel values ​​to each of the meshes according to the terrain quantity, A means for calculating a cross-correlation coefficient that sets an inspection area consisting of multiple meshes for the first terrain model, sets an exploration area larger than the inspection area for the second terrain model, calculates a cross-correlation coefficient based on the pixel values ​​while scanning the exploration area with the inspection area, and extracts the maximum cross-correlation coefficient which is the largest among the multiple cross-correlation coefficients, An elevation difference calculation means that uses the elevations of the respective meshes located at the same position in the inspection area and the exploration area to calculate the elevation difference by subtracting the elevation of the mesh in the older period from the elevation of the mesh in the newer period, The system includes a collapse evaluation means that, when the maximum cross-correlation coefficient falls below a predetermined second correlation threshold and the elevation difference exceeds a predetermined first elevation threshold, and the elevation difference falls below a predetermined second elevation threshold, it evaluates that collapse has occurred in the inspection area and sets a collapse area based on the inspection area. A system for estimating cluster phenomena, characterized by the following features.

5. The collapse evaluation means sets a collapsed area when an elevation difference exceeding the first elevation threshold is detected from the mesh included in the low-elevation end region of the collapsed area, and an elevation difference below the second elevation threshold is detected from the mesh included in the high-elevation end region of the collapsed area. The motion concentration phenomenon estimation system according to feature 4.

Images