Disaster prevention map preparation method

Abstract

To provide a disaster prevention map preparation method capable of generating a disaster prevention map with which dangerous places in disaster prevention are specified with high accuracy without manual data processing.SOLUTION: A disaster prevention map preparation method for preparing a disaster prevention map with which dangerous places are grasped includes: an original data acquisition step of acquiring original data composed of three-dimensional point group data obtained by measuring the ground from the sky; a ground surface vicinity data generation step of automatically generating ground surface vicinity data for automatically extracting only points in the ground surface vicinity in the original data using a computer; and a grid data generation step of automatically generating grid data by interpolating the ground surface vicinity data using the computer.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a method for creating a disaster prevention map that allows users to grasp dangerous locations. [Background technology]

[0002] A topographical map is created by shining a laser from the sky onto the ground to obtain three-dimensional point cloud data, which is then processed by computer. Using the topographical map obtained in this way, it is possible to analyze the terrain in three dimensions and identify dangerous areas where rockfalls and landslides are likely to occur. For this reason, it is expected that these topographical maps will be used in the field of disaster prevention, and research into this field is becoming increasingly active.

[0003] For example, Non-Patent Document 1 studies a method for detecting locations where rockfalls are likely to occur based on a topographical map generated from three-dimensional point cloud data. Also, Non-Patent Document 2 introduces a method for investigating the amount of sediment moved by landslides using laser surveying. [Prior art documents] [Patent documents]

[0004] [Non-Patent Document 1] Akimoto Sakida et al., "Detection of Rockfall Sources Using Support Vector Machines with Topographic Maps," 2022, Geotechnical Journal, Vol. 17, No. 2, pp. 147-157 [Non-patent document 2] Akihiko Ikeda, “Field survey on estimation of sediment flow scale and facility effects”, 2021, Journal of the Sabo Society, Vol. 74, No. 3, pp. 91-95 Summary of the Invention [Problem to be solved by the invention]

[0005] When creating a topographical map from 3D point cloud data, the 3D point cloud data is often converted into grid data, which makes it possible to analyze the undulations of the earth's surface.

[0006] However, conventionally, instead of directly converting the original data (raw data) of 3D point cloud data into grid data, ground data is created by extracting a point cloud representing the earth's surface from the original data, and then the ground data is converted into grid data. This is because the original data also includes a point cloud measuring ground surface obstructions such as trees and buildings, and the point cloud appears to be distributed vertically and has a width, so it is not possible to generate grid data by using this data as is.

[0007] However, when creating ground data, extraction of the point cloud representing the earth's surface is usually done manually. As already mentioned, in the original data, the point cloud appears as a vertically distributed area with a certain width, and the worker extracts the lowest point of each point from the point cloud. However, in areas where the ground is completely covered by an obstruction, the point cloud corresponding to the earth's surface is missing from the original data, so the worker must interpolate the points in those areas. The worker cannot confirm the on-site situation using photographs or the like, and must perform the above work relying only on the original data. This requires the worker to have advanced skills, and creating ground data requires a huge amount of effort.

[0008] Furthermore, the work performed by the workers described above may be detrimental to identifying dangerous areas from a disaster prevention perspective, because in dangerous areas, the ground surface rarely changes smoothly, but rather has sharp and distinctive changes, and the work performed by the workers described above may overlook these distinctive changes in the ground surface as being caused by obstructions.

[0009] The present invention has been made to solve the above-mentioned problems, and provides a method for creating a disaster prevention map that can generate a disaster prevention map that can identify dangerous areas with high accuracy without manual data processing. [Means for solving the problem]

[0010] The above issues are: A method for creating a disaster prevention map that allows users to grasp dangerous areas, comprising: an original data acquisition step of acquiring original data consisting of three-dimensional point cloud data obtained by measuring the ground from the sky; a near-surface data generating step of automatically generating near-surface data by automatically extracting only points near the surface of the ground from the original data using a computer; a grid data generation step of automatically generating grid data by interpolating the near-surface data using a computer; A method for creating a disaster prevention map, characterized by This is solved by providing

[0011] In the disaster prevention map creation method of the present invention, the creation of ground data, which has conventionally been done manually, is replaced by an automated process of generating near-surface data. Therefore, manual work can be omitted when creating a disaster prevention map. Therefore, a disaster prevention map can be created without the need for highly skilled workers. Furthermore, the effort required for creating a disaster prevention map can be significantly reduced. Furthermore, as will be described later, using near-surface data automatically generated by a computer enables more accurate identification of dangerous areas than using manually created ground data.

[0012] In the disaster prevention map creation method of the present invention, the near-ground data generation step is preferably performed by automatically removing, by a computer, point clouds that are distributed vertically and have a certain width in the original data, and that are located above a predetermined height H from the lowest point of each point. The predetermined height H is not particularly limited. However, if the predetermined height H is set too small, the point clouds remaining in the near-ground data will be too sparse, making it difficult to perform interpolation in the grid data generation step. Furthermore, if the predetermined height H is set too large, many point clouds measuring ground obstructions will remain in the near-ground data, and the influence of ground obstructions will be more likely to appear in the grid data generated in the grid data generation step. For this reason, the predetermined height H is preferably set in the range of 10 to 100 cm.

[0013] In the disaster prevention map creation method of the present invention, the interpolation algorithm in the grid data generation step is not particularly limited, and various algorithms can be used. Among them, the Kringing method is suitable for interpolation in the disaster prevention map creation method of the present invention. This makes it possible to represent in the grid data an undulation close to the actual earth surface, making it easier to identify dangerous areas.

[0014] In the disaster prevention map creation method of the present invention, the density of the point cloud acquired in the original data acquisition step is not particularly limited. However, if the density of the point cloud in the original data is too low, the number of points in the data near the ground surface will be small, and even if interpolation is performed in the grid data generation step, the actual undulations of the ground surface may not be accurately represented in the resulting grid data. For this reason, the density of the point cloud acquired in the original data acquisition step is set to 10 points / m 2 There is no particular upper limit to the density of the point cloud in the original data. In order to create a highly accurate disaster prevention map, it is advantageous to make the density of the point cloud in the original data as high as possible. However, taking into consideration the load on the computer in the near-surface data generation process and the grid data generation process, currently, it is recommended to make the density of the point cloud in the original data as high as possible. 5 points / m 2 This is thought to be the limit. [Effects of the Invention]

[0015] As described above, the present invention makes it possible to provide a method for creating disaster prevention maps that can identify disaster-prevention hazardous areas with high accuracy, without manual data processing. [Brief explanation of the drawings]

[0016] [Figure 1] 1 is a diagram showing an example of a disaster prevention map created by a disaster prevention map creating method according to the present invention; [Figure 2] FIG. 1 is a flowchart showing an example of the procedure of a disaster prevention map creation method according to the present invention. [Figure 3] FIG. 10 is a diagram showing an example of a vertical cross section of original data. [Figure 4] FIG. 10 is a diagram showing an example of a vertical cross section of near-surface data. [Figure 5] FIG. 2 is a diagram showing an example of grid data. [Figure 6] 1A and 1B are diagrams showing disaster prevention maps generated by a conventional method and a method according to the present invention, respectively. [Figure 7] FIG. 7 is a diagram illustrating a coloring method for the disaster prevention map of FIG. 6. [Figure 8] FIG. 7 is an enlarged view of a part of the disaster prevention map in FIG. 6(b). [Figure 9] This is a photograph taken at the site near the specific location (α part) on the disaster prevention map in Figure 8. [Figure 10] FIG. 1 is a diagram showing a vertical cross section of near-ground data in the vicinity of a specific location (α part) on a disaster prevention map. DETAILED DESCRIPTION OF THE INVENTION

[0017] An embodiment of the method for creating a disaster prevention map of the present invention will be specifically described with reference to the drawings. The configuration described below is merely a preferred embodiment, and the technical scope of the method for creating a disaster prevention map of the present invention is not limited to the embodiment described below. The method for creating a disaster prevention map of the present invention can be modified as appropriate within the scope that does not impair the spirit of the invention.

[0018] The present invention relates to a method for creating a disaster prevention map (a disaster prevention map creating method). FIG. 1 shows an example of a disaster prevention map created by the disaster prevention map creating method according to the present invention. The disaster prevention map shows the slope and undulations of the ground. Therefore, by looking at the disaster prevention map, it is possible to grasp in advance dangerous areas where rockfalls, landslides, etc. are likely to occur.

[0019] Figure 2 shows an example of the procedure for the method for creating a disaster prevention map according to the present invention. In this embodiment, as shown in Figure 1, a disaster prevention map is created through an original data acquisition process, a near-surface data generation process, a grid data generation process, and a disaster prevention map generation process. Each process will be explained below in order.

[0020] 1. Original data acquisition process The original data acquisition process is a process of acquiring original data consisting of three-dimensional point cloud data measured on the ground from the sky. Figure 3 shows an example of original data. The original data in Figure 3 is shown in a vertical cross section (a cross section cut by a plane parallel to the xz plane), and the point cloud on the original data is distributed three-dimensionally (it also exists in the y-axis direction in Figure 3). Looking at Figure 3, it can be seen that the original data contains a point cloud (ground line) that reflects the ground (surface) and a point cloud that reflects ground surface obstructions such as trees. In this way, the original data is acquired with a width in the direction perpendicular to the ground (surface).

[0021] The original data acquisition process involves flying an aircraft equipped with a laser scanner into the sky and illuminating the ground with a laser beam from the scanner. The distance from the aircraft to the measurement point on the ground can be determined from the time it takes for the laser beam (illuminated light) to be emitted from the laser scanner, reflected by the ground, and returned to the laser scanner. In addition to knowing the direction of the laser beam, the laser scanner is also connected to a GNSS receiver such as a GPS and an inertial measurement unit (IMU) that measures tilt, allowing it to determine the aircraft's current position and orientation. This information can then be used to identify the position (GNSS coordinates) of the measurement point. The laser scanner scans the ground with a laser beam. The original data is acquired by plotting the positions of the numerous measurement points measured in this way in three-dimensional space.

[0022] As already mentioned, the density (number of measurement points per unit area) of the point cloud (the numerous measurement points mentioned above) acquired in the original data acquisition process is 10 points / m 2 In order to create a more accurate disaster prevention map, the density of the point cloud acquired in the original data acquisition process should be 20 points / m 2 More preferably, 30 points / m or more. 2 It is more preferable to set the density of the point cloud in the original data to 100 points / m or more. The higher the density of the point cloud in the original data, the better. If the processing speed of the computer used in the near-surface generation process and grid data generation process described later allows, the density of the point cloud in the original data should be 100 points / m or more. 2 It can be set to 1000 points / m 2 It can also be done as follows.

[0023] The aircraft may be either an unmanned aircraft or a manned aircraft. However, in order to measure more detailed topography on the ground, it is preferable to use an aircraft capable of hovering or flying at low speed. Examples of such aircraft include propeller-type drones and helicopters. Among these, propeller-type drones are preferred because they can fly in the sky at low cost.

[0024] 2. Near-surface data generation process The near-ground data generation process is a process in which a computer is used to automatically generate near-ground data that automatically extracts only points near the ground surface (ground line) in the original data (Fig. 3). Fig. 4 shows an example of near-ground data. The near-ground data in Fig. 4 is shown in a vertical cross section (a cross section cut by a plane parallel to the xz plane), and like the original data described above, the point cloud in the near-ground data is also distributed three-dimensionally (it also exists in the y-axis direction in Fig. 4). Fig. 4 shows that the near-ground data does not include point clouds that reflect ground obstructions such as trees, and only includes point clouds near the ground surface (ground lines).

[0025] The algorithm for the near-surface data generation process is not particularly limited. For example, a method can be adopted in which the original data is divided horizontally (x and y directions) into multiple fine regions, and the point in each fine region with the smallest vertical coordinate (z coordinate) is extracted. However, if this algorithm is used to extract a point cloud near the ground surface (ground line), the point cloud reflecting the ground surface (ground line) will also be thinned out, which reduces the significance of obtaining a high-density point cloud of the original data.

[0026] For this reason, in this embodiment, a method is adopted in which, from among the point clouds that appear in the original data as being distributed with a width in the vertical direction (z-axis direction), point clouds that exceed a predetermined height H from the lowest point of each point are automatically removed by a computer. This prevents point clouds near the ground surface (ground line) from being thinned out, making it possible to reflect the actual ground surface (ground line) in high resolution in the data near the ground surface. Furthermore, because this is a relatively simple process, it is also possible to reduce the load on the computer.

[0027] In addition, as already mentioned, in the conventional method, ground data corresponding to the near-surface data is created manually, which requires highly skilled workers. Even if the worker has the skills, there is a risk that point clouds that should not be deleted from a disaster prevention perspective will be deleted, and dangerous areas will not be reflected in the resulting disaster prevention map. Furthermore, creating ground data requires a huge amount of time and effort. In this regard, in the present embodiment, the near-surface data is created automatically by computer rather than manually, thereby solving the above problems that occurred in the conventional method at once.

[0028] As described above, in this embodiment, the near-ground surface data generation step is performed by automatically removing point clouds that are located at a height greater than a predetermined height H from the lowest point of each point on the original data using a computer. As already mentioned, the height H at which the point clouds are filtered is preferably set to a range of 10 to 100 cm. However, from the viewpoint of facilitating interpolation in the grid data generation step described later, the height H at which the point clouds are filtered is more preferably 20 cm or greater, and even more preferably 30 cm or greater. On the other hand, from the viewpoint of minimizing the influence of ground obstructions on the grid data generated in the grid data generation step described later, the height H at which the point clouds are filtered is more preferably 70 cm or less, and even more preferably 50 cm or less. In this embodiment, the height H at which the point clouds are filtered is set to 40 cm.

[0029] 3. Grid data generation process The grid data generation process is a process in which grid data is automatically generated from the near-surface data (Fig. 4) using a computer. An example of grid data is shown in Fig. 5. Looking at Fig. 5, it can be seen that the point cloud is arranged in a grid pattern. When the grid data (Fig. 5) is viewed from above in the vertical direction (positive side in the z-axis direction), each point overlaps with a lattice point position on the horizontal plane (xy plane). Although the point cloud in the near-surface data is not arranged in a grid pattern, such grid data can be generated by performing interpolation on the point cloud in the near-surface data.

[0030] The interpolation algorithm used in the grid data generation process is not particularly limited as long as it can estimate the coordinates of points (unknown points) located between known points based on the coordinates of the known points. Examples of such interpolation methods include the inverse distance weighting (IDW) method, the spline method, the natural neighbor method, and the Kriging method.

[0031] The inverse distance weighting (IDW) method is a method for determining the coordinates of an unknown point by averaging the values ​​of multiple known points near the unknown point. When averaging, the values ​​of known points closer to the unknown point are evaluated higher than the values ​​of known points farther from the unknown point. The spline method is a method for determining the coordinates of an unknown point from a spline surface (a surface with minimum curvature) that passes through all known points. The natural neighbor method is a method for determining the coordinates of an unknown point by taking into account natural neighbors near the known point. The kriging method is a method for determining the coordinates of an unknown point by using the values ​​of known points to model the spatial correlation between known points and the known points, and then using this modeled spatial correlation.

[0032] In this embodiment, grid data is generated by interpolating the coordinates of the estimated points using the Kriging method, which allows the grid data to represent undulations that are close to the actual earth surface.

[0033] 4. Disaster prevention map generation process The disaster prevention map generation process uses a computer to generate a disaster prevention map, such as that shown in Figure 2, from the grid data (Figure 5) generated in the grid data generation process. Disaster prevention maps are generated by generating surfaces that pass through the gaps between the point clouds (lattice points) on the grid data. Contour lines can also be displayed on disaster prevention maps. Furthermore, it is possible to display convex areas brightly and concave areas darkly. Figure 2 shows that the disaster prevention map clearly depicts the topography, including the undulations and slopes of the terrain it contains. This makes it easy to identify dangerous areas where there is a risk of rockfalls, etc. For example, a point that is both sloped and locally convex can be determined to be at risk of rockfalls, etc.

[0034] The created disaster prevention map can be displayed on the screen of a display device (such as a liquid crystal display) connected to a computer such as a personal computer. By operating the cursor, the displayed disaster prevention map can be rotated three-dimensionally on the screen, and specific points on the disaster prevention map can be enlarged or reduced. This makes it easier to grasp dangerous areas.

[0035] 5.Other For other configurations not specifically mentioned above, those that comply with "Part 4: Topographic Surveying and Photogrammetry (Three-dimensional Point Cloud Surveying)" in the "Standards for Work Regulations" of the Ministry of Land, Infrastructure, Transport and Tourism Notification No. 250 of March 31, 2023 may be adopted.

[0036] 6. Verification In order to verify the extent to which the disaster prevention map created by the method of this embodiment is more advantageous in identifying dangerous areas than the disaster prevention map created by the conventional method (a method in which ground data equivalent to data near the ground surface is manually created), disaster prevention maps were created using each method from the same original data. Figure 6(a) shows the disaster prevention map created by the conventional method, and Figure 6(b) shows the disaster prevention map created by the method of this embodiment.

[0037] While Fig. 6 is displayed in black and white, the original map of Fig. 6 is displayed in color. In the original map of Fig. 6, points on the disaster prevention map created using the method of this embodiment that appear at higher elevations than those on the disaster prevention map created using the conventional method are colored blue, and the opposite points (points on the disaster prevention map created using the conventional method that appear at higher elevations than those on the disaster prevention map created using the method of this embodiment) are colored red. Fig. 7 shows the coloring method for the disaster prevention map (color drawing) of Fig. 6. The greater the difference in elevation, the darker the blue or red.

[0038] When comparing Figures 6(a) and 6(b) with the actual topography, as shown in Figure 7, it was found that the disaster prevention map created using the method of this embodiment tends to have higher elevations above the cliff than the disaster prevention map created using the conventional method, and that the disaster prevention map created using the method of this embodiment tends to have lower elevations below the cliff than the disaster prevention map created using the conventional method. In other words, it was found that the disaster prevention map created using the conventional method smooths out the actual topographical undulations, while the disaster prevention map created using the method of this embodiment tends to faithfully represent the actual topographical undulations. This shows that the disaster prevention map created using the method of this embodiment can more accurately identify dangerous areas.

[0039] In this regard, FIG. 6(b) shows a circle marked with a thick solid line (blue) and a circle marked with a thick dashed line (red). The circle marked with a thick solid line (blue) shows an area that appears as unevenness in FIG. 6(a), but where the unevenness is more emphasized in FIG. 6(b). The circle marked with a thick dashed line (red) shows an area that cannot be recognized as unevenness in FIG. 6(a), but can be recognized as unevenness in FIG. 6(b). Looking at these circle marks, it is clear that the disaster prevention map created using the method of this embodiment is able to identify more dangerous areas.

[0040] FIG. 8 is an enlarged view of a portion of the disaster prevention map of FIG. 6(b). FIG. 9 is a photograph of an actual scene near a specific location (part α) on the disaster prevention map of FIG. 8. FIG. 10 is a view showing a vertical cross section of the near-ground surface data near the specific location (part α) on the disaster prevention map. Although FIGS. 8 to 10 are displayed in black and white, the original maps of FIGS. 8 to 10 are also displayed in color, as in FIG. 6. In FIG. 8, the data is colored blue and red in the same manner as in FIG. 6. Furthermore, in FIG. 10, the point cloud in the near-ground surface data generated by the method of this embodiment is shown by light blue dots, and the corresponding points in FIG. 9 are also shown by light blue dots. In addition, in FIG. 10, the point cloud in the ground data generated by the conventional method is shown by yellow dots, and the corresponding points in FIG. 9 are also shown by yellow dots.

[0041] The rock in part α in Fig. 9 is a dangerous area where there is a risk of rockfall, and Fig. 10 shows that the near-surface data generated by the method of this embodiment shows the outline of the rock outcrop more clearly than the ground data created by the conventional method. Therefore, even in the disaster prevention map of Fig. 8, the elevation of the bedrock area in part α is higher in the disaster prevention map generated by the method of this embodiment, making it easier to recognize the dangerous area.

[0042] As described above, the disaster prevention map generated by the method of this embodiment can detect more minute dangerous areas than the disaster prevention map generated by the conventional method. 2 Below, 3m 2 Below, 1m 2 A few meters or less 2 It can detect dangerous areas in the following areas:

Claims

1. A method for creating a disaster prevention map that allows users to grasp dangerous areas, comprising: an original data acquisition step of acquiring original data consisting of three-dimensional point cloud data obtained by measuring the ground from the sky; a near-surface data generating step of automatically generating near-surface data by automatically extracting only points near the surface of the ground from the original data using a computer; a grid data generation step of automatically generating grid data by interpolating the near-surface data using a computer; A method for creating a disaster prevention map, comprising the steps of:

2. The near-surface data generation step includes: Among the point clouds that appear in the original data as being distributed vertically and having a certain width, the point clouds that exceed a certain height H from the lowest point of each point are automatically removed by computer. This is done by The method for creating a disaster prevention map according to claim 1.

3. 3. The method for creating a disaster prevention map according to claim 2, wherein the predetermined height H is set to 10 to 100 cm.

4. 4. The method for creating a disaster prevention map according to claim 3, wherein the interpolation in the grid data generating step is performed by the Kringing method.

5. In the original data acquisition process, 10 points / m ２ 5. The method for creating a disaster prevention map according to claim 4, wherein the point cloud is measured at a density equal to or greater than the above.

Images