JP2021174013A - Disaster prediction system and disaster prediction method - Google Patents

Abstract

To provide a system that can predict and detect an occurrence of landslide disaster more stably than ever, the system being hardly affected by an environmental state.SOLUTION: A disaster prediction system includes: first detecting means for detecting a change of the state of land features in a target area; second detecting means for detecting a change of the state of land features based on a detection principle different from that of the first detecting means; predicting means for predicting the degree of the risk of occurrence of landslide disaster in the target area based on the change of the state of land features obtained by the detection results of the first detecting means and the second detecting means; and display means for displaying the target area, the change of the state of land features, and the degree of the risk of the occurrence of landslide disaster.SELECTED DRAWING: Figure 1

Description

The present invention relates to a disaster prediction system for sediment-related disasters and a disaster prediction method.

In light of recent climate change and the frequent occurrence of natural disasters, it is becoming more important to predict disasters in each region according to weather information and the shape of the region. By predicting a disaster, it is possible to guide evacuation before the occurrence of a disaster and to grasp the degree of danger in the area, and it is possible to control the damage caused by the disaster.

For example, Patent Document 1 proposes a technique of measuring the distance to the ground using a laser scanner mounted on an airplane and identifying the altitude and the shape of the terrain on the ground from the position information of the airplane. Further, in Patent Document 2, a monitoring target slope that may cause a sediment disaster is photographed by a plurality of cameras from different viewpoints to acquire three-dimensional coordinate data, and a minute change in the monitoring target slope is detected from the three-dimensional coordinate data. A method has been proposed. Further, Patent Document 3 proposes a method of recognizing a collapse of a slope from a camera image and outputting an alarm from an alarm device.

Japanese Unexamined Patent Publication No. 2014-352232 Japanese Unexamined Patent Publication No. 2016-180681 Japanese Unexamined Patent Publication No. 2003-219398

According to the method of Patent Document 1, there is a problem that various information cannot be obtained in a situation where an airplane cannot fly, and for example, damage caused by a typhoon cannot be detected in real time. Further, according to the methods of Patent Document 2 and Patent Document 3, although the image captured by the camera is used, the image of the slope to be monitored cannot be obtained in a situation where the field of view is not effective due to heavy rain or the like, so that the field of view is, for example. There was a problem that landslides could not be detected in real time in a situation where it did not work.

In view of the above problems, an object of the present invention is to provide a disaster prediction system that is less affected by environmental conditions and can predict and detect the occurrence of sediment-related disasters more stably than before.

In order to solve the above problems, the present invention has the following configuration. That is,
It ’s a disaster prediction system.
The first detection means for detecting changes in the state of features in the target area,
A second detection means for detecting a change in the state of the feature by a detection method having a different detection principle from the first detection means,
Based on the change in the state of the feature obtained by the detection results of the first detection means and the second detection means, the prediction means for predicting the risk of sediment-related disaster in the target area, and the prediction means.
It has a display means for displaying the change in the state of the target area and the feature and the risk of occurrence of the sediment-related disaster.

According to the present disclosure, it is possible to predict and detect the occurrence of sediment-related disasters more stably than before, which is not easily affected by environmental conditions.

The schematic diagram which shows the example of the whole structure of the disaster prediction system which concerns on one Embodiment of this invention. The figure which shows the example of the functional structure of each apparatus which concerns on one Embodiment of this invention. The figure which shows the example of the functional structure of each apparatus which concerns on one Embodiment of this invention. The figure for demonstrating the feature of each detection means which concerns on one Embodiment of this invention. The flowchart of the process which concerns on 1st Embodiment. The flowchart of the process which concerns on 2nd Embodiment. The figure which shows the example of the functional structure of the disaster prediction server which concerns on 3rd Embodiment. The figure for demonstrating learning which concerns on 3rd Embodiment. The flowchart of the process which concerns on 3rd Embodiment. The figure which shows the display example which concerns on one Embodiment of this invention.

Hereinafter, embodiments for carrying out the present invention will be described with reference to drawings and the like. It should be noted that the embodiments described below are embodiments for explaining the present invention, and are not intended to be interpreted in a limited manner, and are described in each embodiment. Not all configurations are essential configurations for solving the problems of the present invention. Further, in each drawing, the same component is given the same reference number to indicate the correspondence.

<First Embodiment>
Hereinafter, the first embodiment of the present invention will be described. In the following description, the terrain and target objects to be monitored are collectively referred to as "features". This is not intended to exclude all moving objects such as vehicles, and may be treated as a feature including these as necessary.

[System configuration]
FIG. 1 is a schematic diagram showing an example of the overall configuration of the disaster prediction system according to the present embodiment. The disaster prediction system according to the present embodiment includes a disaster prediction server 1, a weather information providing server 2, a camera 3, a laser scanner 4, a radar distance measuring device 5, a ground surface position measuring device 6, and a display device 7. NS. Each device is communicably connected via the network 8. The communication standard of the network 8 and the wired / wireless are not particularly limited, and a plurality of communication standards may be combined and realized. Further, in FIG. 1, one device is shown for each device, but a plurality of devices may be included. In particular, with respect to the camera 3, the laser scanner 4, the radar distance measuring device 5, and the ground surface position measuring device 6 as the detecting means, one or a plurality of detecting means are installed according to the shape of the feature for disaster monitoring. It shall be. Further, a subscript n (≧ 2) indicating that a plurality of detection means exists is added to each of the plurality of types of detection means, but this does not mean that the number of each detection means is the same. Each may be a different number. Further, when the same type of detection means is comprehensively explained, the subscripts are omitted, and when individual explanations are required, the subscripts are added.

The disaster prediction server 1 makes a disaster prediction in the area to be monitored based on the result detected by the detection means according to the present embodiment. A plurality of disaster prediction servers 1 may be provided according to the purpose of load balancing, the difference in the area to be monitored, and the like. The weather information providing server 2 provides weather information in the area to be monitored. The display device 7 is a device for displaying the result of the disaster prediction by the disaster prediction server 1. The display device 7 may be, for example, an information processing device such as a PC (Personal Computer), a mobile terminal such as a smartphone or a tablet terminal, or an HMD (Head Mounted Display).

[Functional configuration]
FIG. 2 is a diagram showing an example of the functional configuration of the device as each detection means according to the present embodiment. FIG. 3 is a diagram showing an example of a functional configuration of a device other than the detection means according to the present embodiment.

(camera)
As shown in FIG. 2A, the camera 3 according to the present embodiment is configured as a network camera, and includes an imaging unit 301, a storage unit 302, a control unit 303, and a communication unit 304. The image pickup unit 301 is an image pickup section for acquiring an image of the periphery where the camera 3 is installed, and includes a lens (not shown), an image pickup element, and the like. The storage unit 302 is a storage unit that stores an image captured via the image pickup unit 301, and is composed of, for example, a non-volatile storage area such as a flash memory or an HDD (Hard Disk Drive). The storage unit 302 stores the captured image and information at the time of shooting (shooting date / time, shooting setting, etc.) in association with each other. The control unit 303 is a part that controls the entire camera 3, and is composed of, for example, a CPU (Central Processing Unit) as a processing unit, a dedicated circuit, and the like. The communication unit 304 is a part for communicating with an external device, and the wired / wireless communication standard and the like are not particularly limited.

The camera 3 may have a configuration in which a shooting range is predetermined, or may have a configuration capable of shooting in all directions. Further, the camera 3 may have a configuration in which the shooting settings (for example, brightness, contrast, resolution, etc.) can be switched according to the situation of the surrounding environment and the like. If the shooting range and shooting settings can be switched, the control unit 303 controls. The shooting timing by the camera 3 may be performed at predetermined intervals, or may be performed due to a request from the outside (for example, the disaster prediction server 1). Alternatively, shooting may be performed at the timing when it is detected that a predetermined event has occurred. The predetermined event may be, for example, the case where shaking or light emission exceeding a predetermined value is detected, and these may be detected by using a seismic intensity sensor or a light receiving element (not shown).

The control unit 303 may be configured to perform predetermined image processing on the captured image captured by the imaging unit 301. For example, the control unit 303 may be configured to perform image processing such as filtering on the captured image. Further, when the imaging unit 301 has a configuration capable of shooting a moving image, the control unit 303 performs a process of extracting frames from the captured moving image at predetermined intervals to obtain a captured image for transmission. May be good. Alternatively, the control unit 303 may perform a process of extracting a frame in which the content of the image has changed and using it as a captured image for transmission.

The communication unit 304 timely transmits the captured image stored in the storage unit 302 to the outside (for example, the disaster prediction server 1). The transmission timing here may be performed at predetermined intervals, or may be performed due to a request from the outside (for example, the disaster prediction server 1). Alternatively, it may be transmitted when the capacity of the captured image stored in the storage unit 302 exceeds a predetermined threshold value. The captured image stored in the storage unit 302 may be deleted when time elapses after being transmitted to the outside. When a plurality of cameras 3-1 to 3-n are installed, the configurations of the plurality of cameras 3-1 to 3-n do not have to be the same. A camera 3 having different functions may be provided depending on the installation position, the characteristics of the area to be photographed, and the like.

(Laser scanner)
As shown in FIG. 2B, the laser scanner 4 according to the present embodiment includes a distance measuring unit 401, a storage unit 402, a control unit 403, and a communication unit 404. The distance measuring unit 401 irradiates the measurement target with a laser beam and calculates the measurement result from the reflected light. As the measurement result here, in addition to the three-dimensional coordinates (X, Y, Z) in the predetermined coordinate system, the color information (R, G, B), the reflection intensity, the reflectance, the angle information, etc. are the measurement targets. It may be acquired for each constituent point. As the measurement method using the laser scanner 4, a known method such as a time-of-flight method or a phase shift method may be used, and may be used according to the measurement target. In the following description, the data of a plurality of points constituting the measurement target is also referred to as point cloud data (two-dimensional point cloud data or three-dimensional point cloud data).

The storage unit 402 is a storage unit that stores point group data acquired via the distance measurement unit 401, and is composed of, for example, a non-volatile storage area such as a flash memory. The storage unit 402 stores the point cloud data and the information at the time of measurement (measurement date and time, measurement setting, etc.) in association with each other. The control unit 403 is a portion that controls the entire laser scanner 4, and is composed of, for example, a CPU as a processing unit, a dedicated circuit, or the like. The communication unit 404 is a part for communicating with an external device, and the wired / wireless communication standard and the like are not particularly limited.

The laser scanner 4 may have a configuration in which a measurement range is predetermined, or may have a configuration in which the measurement range and measurement settings can be switched within a predetermined range. If the measurement range and measurement settings can be switched, the control unit 403 controls. The measurement timing by the laser scanner 4 may be performed at predetermined intervals, or may be performed due to a request from the outside (for example, the disaster prediction server 1). Alternatively, the measurement may be performed at the timing when it is detected that a predetermined event has occurred. The predetermined event may be, for example, the case where shaking of a predetermined value or more is detected, and these may be detected by using a seismic intensity sensor (not shown).

The control unit 403 may be configured to perform predetermined processing on the point cloud data measured by the distance measurement unit 401. For example, the control unit 403 may be configured to perform a predetermined analysis process on the point cloud data.

The communication unit 404 transmits the point cloud data stored in the storage unit 402 to the outside (for example, the disaster prediction server 1) in a timely manner. The transmission timing here may be performed at predetermined intervals, or may be performed due to a request from the outside (for example, the disaster prediction server 1). Alternatively, it may be transmitted when the capacity of the point cloud data stored in the storage unit 402 exceeds a predetermined threshold value. The point cloud data stored in the storage unit 402 may be deleted when a predetermined time elapses after being transmitted to the outside. When a plurality of laser scanners 4-1 to 4-n are installed, the configurations of the plurality of laser scanners 4-1 to 4-n do not have to be the same. A laser scanner 4 having different functions may be provided depending on the installation position, the characteristics of the measurement target, and the like.

(Radar distance measuring device)
As shown in FIG. 2C, the radar distance measuring device 5 according to the present embodiment includes a distance measuring unit 501, a storage unit 502, a control unit 503, and a communication unit 504. The distance measuring unit 501 transmits radio waves of a predetermined wavelength from a transmitter (not shown), and receives the reflected wave by a receiver (not shown) to provide information on the distance, speed, and angle to the object. Etc. are calculated. Here, millimeter waves, which are radio waves having a frequency in the band of 30 to 300 GHz, are used. The bandwidth of millimeter waves that can be used varies depending on the region, etc., and is not particularly limited here. In the following description, for convenience, the result measured by the radar distance measuring device 5 is also referred to as radar measurement data.

The storage unit 502 is a storage unit that stores radar measurement data acquired via the distance measurement unit 501, and is composed of, for example, a non-volatile storage area such as a flash memory. The storage unit 502 stores the radar measurement data in association with the information at the time of measurement (measurement date and time, measurement setting, etc.). The control unit 503 is a part that controls the entire radar distance measuring device 5, and is composed of, for example, a CPU as a processing unit, a dedicated circuit, or the like. The communication unit 504 is a part for communicating with an external device, and the wired / wireless communication standard and the like are not particularly limited.

The measurement timing by the radar distance measuring device 5 may be performed at predetermined intervals, or may be performed due to a request from the outside (for example, the disaster prediction server 1). Alternatively, the measurement may be performed at the timing when it is detected that a predetermined event has occurred. The predetermined event may be, for example, the case where shaking or light emission exceeding a predetermined value is detected, and these may be detected by using a seismic intensity sensor or a light receiving element (not shown).

The control unit 503 may be configured to perform predetermined processing on the radar measurement data measured by the distance measurement unit 501. For example, the control unit 503 may be configured to perform a predetermined analysis process, filtering process, or the like on the radar measurement data.

The communication unit 504 timely transmits the radar measurement data stored in the storage unit 502 to the outside (for example, the disaster prediction server 1). The transmission timing here may be performed at predetermined intervals, or may be performed due to a request from the outside (for example, the disaster prediction server 1). Alternatively, it may be transmitted when the capacity of the radar measurement data stored in the storage unit 502 exceeds a predetermined threshold value. The radar measurement data stored in the storage unit 502 may be deleted when a predetermined time elapses after being transmitted to the outside. When a plurality of radar distance measuring devices 5-1 to 5n are installed, the configurations of the plurality of radar distance measuring devices 5-1 to 5n do not have to be the same. A radar distance measuring device 5 having different functions may be provided depending on the characteristics of the installation position and the like.

(Ground surface displacement measuring device)
As shown in FIG. 2D, the ground surface position measuring device 6 according to the present embodiment includes a satellite signal receiving unit 601, a storage unit 602, a control unit 603, and a communication unit 604. The satellite signal receiving unit 601 receives the signal transmitted from the positioning satellite and detects the position of its own device. Here, a global positioning system (GNSS: Global Navigation Satellite System) that uses a positioning satellite such as GPS (Global Positioning System) is used. A plurality of positioning satellites may be used. The received signal (radio wave) includes the transmission time from the satellite, and the satellite signal receiving unit 601 uses the transmission time, the propagation time of the radio wave obtained from the reception time, and the speed of the radio wave to itself. Detect the position of the device. Latitude, longitude, and altitude correspond to the position of the own device. In the following description, for convenience, the result measured by the ground surface position measuring device 6 is also referred to as position data.

The storage unit 602 is a storage unit that stores position data acquired via the satellite signal reception unit 601 and is composed of, for example, a non-volatile storage area such as a flash memory. The storage unit 402 stores the position data and the information at the time of measurement (measurement date and time, etc.) in association with each other. The control unit 603 is a portion that controls the entire ground surface position measuring device 6, and is composed of, for example, a CPU as a processing unit, a dedicated circuit, or the like. The communication unit 604 is a part for communicating with an external device, and the wired / wireless communication standard and the like are not particularly limited.

The measurement timing by the ground surface position measuring device 6 may be performed at predetermined intervals, or may be performed due to a request from the outside (for example, the disaster prediction server 1). Alternatively, the measurement may be performed at the timing when it is detected that a predetermined event has occurred. The predetermined event may be, for example, the case where shaking or light emission exceeding a predetermined value is detected, and these may be detected by using a seismic intensity sensor or a light receiving element (not shown).

The control unit 603 may be configured to perform predetermined processing on the position data measured by the satellite signal reception unit 601. For example, the control unit 603 may be configured to detect a change in position data and instruct the communication unit 604 to transmit to the outside when the change occurs.

The communication unit 604 transmits the position data stored in the storage unit 602 to the outside (for example, the disaster prediction server 1) in a timely manner. The transmission timing here may be performed at predetermined intervals, or may be performed due to a request from the outside (for example, the disaster prediction server 1). Alternatively, it may be transmitted when the capacity of the position data stored in the storage unit 602 exceeds a predetermined threshold value. The position data stored in the storage unit 602 may be deleted when a predetermined time elapses after being transmitted to the outside.

As an example of the detection means according to the present embodiment, four types of detection means are shown. These need not be configured separately, and one device may be configured to include a plurality of detection units. For example, one device may be configured to include an imaging unit and a distance measuring unit using a laser. Further, each detection means is not limited to a configuration in which it is installed in a fixed state. For example, the camera 3 may be realized in a configuration mounted on a moving body (for example, an unmanned aerial vehicle such as a drone). As described above, in the configuration in which the detecting means is mounted on the moving body, it is desirable that the position information of the own device is also acquired at the time of measurement by a GNSS device or the like.

(Disaster prediction server)
As shown in FIG. 3A, the disaster prediction server 1 according to the present embodiment includes a data collection unit 101, a three-dimensional model generation unit 102, a weather information analysis unit 103, a communication control unit 104, and a shape change detection unit 105. It includes a history information management unit 106, a danger range identification unit 107, and a display control unit 108.

The data collection unit 101 collects measurement data from various detection means according to the present embodiment. The timing of collection may be collected by receiving the data at the timing transmitted from the various detection means, or may be collected by the data collection unit 101 making a request to the various detection means. Further, the collection timing may be different for each detection means. In addition, the data collection unit 101 acquires weather information from the weather information providing server 2.

The three-dimensional model generation unit 102 generates a three-dimensional model based on the measurement data collected from various detection means. For example, a three-dimensional model is generated using the captured image acquired from the camera 3 and the point cloud data acquired from the laser scanner 4. A known method may be used as the method for generating the three-dimensional model, and the method is not particularly limited. Here, an example of a method for generating a three-dimensional model will be described.

When creating a three-dimensional model, a plurality of captured images that may contain the same object are collected, and candidate feature points of each captured image are detected using a feature amount calculation function. As this function, Harris, AKAZE, ORB (Oriented FAST and Rotated BRIEF), or the like can be used. Next, when the value of the feature amount of the detected feature point candidate exceeds a preset threshold value, the candidate is used as the feature point (feature point extraction).

When the feature points detected in each image are caused by the same object, those feature points correspond between the point cloud data. Therefore, feature points are associated between images (feature point matching). Since the feature points for which feature point matching is possible indicate a certain part of the same object, the three-dimensional coordinates of the object are calculated based on various information included in the measurement result. Examples of the information here include the orientation and position information of the detection means. The plurality of three-dimensional coordinate points calculated in this way form a three-dimensional point cloud.

Since the three-dimensional point cloud is simply a collection of three-dimensional coordinate points, it is converted into a three-dimensional model so that humans can easily see it. In that case, the area surrounded by the three feature points in the vicinity of the image is attached to the three points of the three-dimensional coordinates formed by the three feature points (texture mapping). As a result, a three-dimensional model composed of triangular polygon planes can be formed. It is also possible to generate a three-dimensional model in the same manner by using the point cloud data acquired from the laser scanner 4.

The meteorological information analysis unit 103 uses the meteorological information collected from the meteorological information providing server 2 to analyze changes in the meteorological conditions in the measurement target area. The content of the analysis here is not particularly limited, and may be, for example, a change in precipitation for each predetermined region in the measurement range. The communication control unit 104 is a part for performing communication with an external device, and the wired / wireless and communication standards are not particularly limited. The shape change detection unit 105 detects the change in the surface shape based on the measurement data collected from various detection means. The change here may be detected based on the accumulated past measurement data. For example, the shape change detection unit 105 may detect the amount of change by comparing the generated three-dimensional models, or may detect the amount of change according to the change in the position information.

The history information management unit 106 manages past weather information, ground surface shape, measurement data, and the like as history information. The history information management unit 106 may manage changes in each measurement range based on the history information. The danger range specifying unit 107 identifies the danger range in the measurement range by using each measurement data and the analysis result. The danger range here may be specified for each range of a predetermined size, or may be determined according to the shape of the terrain. In addition, the danger range specifying unit 107 may specify the danger range for each type of sediment-related disaster (slope collapse, landslide, debris flow, flood damage, avalanche, etc.) that is predicted to occur. In the present embodiment, an example of four levels (levels 1 to 4) as the degree of risk will be described. Level 1 indicates that there is no possibility of a disaster or that the situation is extremely low. Indicates that the higher the level, the higher the likelihood of a disaster. The particle size of the degree of risk is an example and is not limited to this. Also, different levels may be used depending on the target disaster.

The display control unit 108 controls the display of the three-dimensional model generated by the three-dimensional model generation unit 102 and the danger range specified by the danger range identification unit 107. The display control here includes superimposing the danger range specified by the danger range specifying unit 107 on the three-dimensional model and displaying the display contents, and switching the display contents according to the degree of danger. An example of the display will be described later with reference to FIG. The display control unit 108 provides the display device 7 with display data including a three-dimensional model and danger range data in response to a request from the display device 7.

(Weather information providing server)
As shown in FIG. 3B, the weather information providing server 2 according to the present embodiment includes a weather information providing unit 201, a weather information management unit 202, and a communication control unit 203. The weather information providing unit 201 provides weather information for each region. The provision method here may be provided in response to a request from the outside (for example, the disaster prediction server 1), or may be provided at predetermined intervals. Alternatively, when urgent weather information occurs, it may be provided at that timing.

The weather information management unit 202 manages the weather information. The meteorological information may include future forecast information in addition to past information. In addition, information that affects the occurrence of sediment-related disasters may be included in the meteorological information. For example, earthquake occurrence information, volcanic activity information, and the like may be included in the meteorological information. Further, the particle size (time interval and range) indicated by the meteorological information is not particularly limited, and may differ depending on the region and the like. The communication control unit 203 is a part for performing communication with an external device, and the wired / wireless and communication standards are not particularly limited.

(Display device)
As shown in FIG. 3C, the display device 7 according to the present embodiment includes a three-dimensional model display unit 701, an operation reception unit 702, and a communication control unit 703. The three-dimensional model display unit 701 displays the display data provided by the disaster prediction server 1 on the display (not shown) of the display device 7. The display of the three-dimensional model may be displayed by projection of a projector (not shown) or the like, or may be displayed as virtual reality (VR). The operation reception unit 702 receives an operation from the user via the operation unit (not shown) of the display device 7. Examples of the operation here include operations such as switching the display of the three-dimensional model displayed on the display. The communication control unit 703 is a part for performing communication with an external device, and the wired / wireless and communication standards are not particularly limited. Each part of the display device 7 may be realized by an application installed in the display device 7.

[Characteristics of detection means]
As described above, in the disaster prediction system according to the present embodiment, a plurality of types of detection means are used in combination. Since each detection means has a different detection principle, the characteristics related to detection are different. Hereinafter, the advantages and disadvantages based on the characteristics of each of the plurality of types of detection means will be described.

FIG. 4 shows an outline of the characteristics of each of the plurality of types of detection means. The camera 3 acquires a peripheral image by the imaging unit 301. At this time, high-resolution point cloud data can be acquired from the measurement data (captured image) acquired by the camera 3. On the other hand, it is assumed that the camera 3 cannot measure the image depending on the weather conditions (rain, fog, etc.) and the change in visibility (illuminance change, etc.) at the time of shooting, or obtains a low-quality photographed image.

The laser scanner 4 can acquire point cloud data with medium resolution. Here, the medium shows a relative relationship, and is between the camera 3 and the radar distance measuring device 5. On the other hand, it is assumed that the laser scanner 4 cannot perform measurement depending on the weather conditions (rain, fog, etc.) at the time of measurement, or acquires point cloud data with low accuracy.

The radar distance measuring device 5 can measure without being affected by the weather conditions. On the other hand, the radar distance measuring device 5 acquires low-resolution radar measurement data as compared with the measurement data of the camera 3 and the laser scanner 4.

The ground surface position measuring device 6 can measure without being affected by the weather conditions. In addition, the ground surface position measuring device 6 can measure the position of its own device with high accuracy. On the other hand, since the ground surface position measuring device 6 detects the position of its own device, it cannot detect the position of another device.

In the present embodiment, by using a combination of a plurality of types of detection means having different characteristics as described above, the detection accuracy of the entire system is improved and the oversight of detection is reduced. Further, by combining these plurality of types of detection means, it is possible to respond to changes in the environment during measurement and to supplement information when generating a three-dimensional model. Then, by improving the detection accuracy of the entire system and reducing the oversight of detection, the accuracy of generating a three-dimensional model generated from the data obtained as the detection result and the accuracy of disaster prediction are improved.

In this embodiment, four types of detection means are shown. There are 11 combinations of patterns in which at least two types of detection means are used in combination, and a combination suitable for the target area may be used in consideration of the characteristics of each detection means. Further, the number of each detection means used may vary depending on the target area.

[Processing flow]
Hereinafter, the processing flow of the disaster prediction server 1 according to the present embodiment will be described. The processing shown below is realized, for example, by the CPU (not shown) included in the disaster prediction server 1 reading and executing the program stored in the storage device. The following processing may be started at predetermined intervals. Further, each detection means is installed around the area to be measured so that the measurement data can be provided to the disaster prediction server 1.

In S501, the disaster prediction server 1 sets the degree of risk for the amount of change in the measurement target area. In addition, a plurality of risk levels may be set according to the type of disaster. The method of setting the degree of risk is not particularly limited, and may be specified according to the shape in the area to be measured. For example, the degree of risk for the amount of change within a predetermined time from a certain point in time may be set as a reference. Alternatively, the degree of risk for the amount of change of the terrain from the reference position at a certain point in time may be set. In the present embodiment, the degree of risk for the amount of change obtained based on the detection results detected by each of the plurality of types of detection means is set.

In S502, the disaster prediction server 1 acquires measurement data from each of the installed detection means. At this time, the measurement data to be used for the subsequent processing may be extracted according to the content of the acquired measurement data. For example, when the image quality of the captured image acquired from the camera 3 is low, the configuration may be such that it is excluded from the data used for the subsequent processing. Alternatively, if the measurement data cannot be acquired from the detection means at a certain timing, the disaster prediction server 1 may request the measurement data from the detection means. Alternatively, if the measurement data cannot be acquired, the information may be treated as the measurement data, assuming that some kind of disaster may have occurred in the measurement range.

In S503, the disaster prediction server 1 calculates the amount of change in the ground surface shape of the target area based on the measurement data acquired from each detection means. For example, the disaster prediction server 1 newly generates a three-dimensional model from the measurement data acquired in S502, and from the difference obtained by comparing this with the already generated three-dimensional model, the land of the target area. The amount of change in surface shape may be calculated. It is assumed that the generated three-dimensional model is held in the storage device (not shown) each time. Alternatively, the disaster prediction server 1 may calculate the amount of change (variation) in the position of the measurement target based on the measurement data.

In S504, the disaster prediction server 1 predicts the degree of danger in the target area based on the amount of change in the ground surface shape calculated in S503 and the degree of danger set in S501. In the present embodiment, among the detection results by the plurality of types of detection means, the value with the highest risk is predicted as the risk in the area. The prediction method is not limited to this, and even if the risk level is predicted based on the combination of risk levels set corresponding to the detection results by a plurality of types of detection means and the comprehensive evaluation of the detection results. good. The comprehensive evaluation here may be determined by weighting each detection result, or may be evaluated based on a predetermined criterion.

In S505, the disaster prediction server 1 controls the display of the three-dimensional model of the target area based on the prediction result in S504. The display here is performed by superimposing a display according to the degree of risk on the three-dimensional model and providing it to the display device 7 as display data.

In S506, the disaster prediction server 1 issues an alarm based on the prediction result in S504. The alarm here may be configured to blink or give an evacuation warning at the time of display on the display device 7. Alternatively, the configuration may be such that an evacuation warning or information on an area predicted to be dangerous is notified via an alarm device (not shown) such as a speaker installed around the display device 7. Then, this processing flow is terminated.

As described above, according to the present embodiment, it is possible to predict and detect the occurrence of sediment-related disasters more stably than before, which is less affected by the environmental conditions, in consideration of the advantages and disadvantages of the detection means provided for the target area.

<Second embodiment>
Hereinafter, a second embodiment of the present invention will be described. As a second embodiment, regarding an embodiment for determining whether or not there is a risk of a sediment-related disaster using correlation data showing the relationship between the ground surface shape and precipitation in the area where a past sediment-related disaster has occurred. explain. It should be noted that the description of the configuration overlapping with the first embodiment will be omitted, and the description will be given focusing on the difference.

The correlation data according to the present embodiment is data showing the correlation between the occurrence of disasters based on the past precipitation and the predicted precipitation amount and the ground surface shape. As the correlation data according to the present embodiment, for example, those provided by a public institution or a predetermined organization can be used. An example of correlation data is a hazard map provided by an administrative agency having jurisdiction over the target area.

[Processing flow]
Hereinafter, the processing flow of the disaster prediction server 1 according to the present embodiment will be described. The processing shown below is realized, for example, by the CPU (not shown) included in the disaster prediction server 1 reading and executing the program stored in the storage device. The following processing may be started at predetermined intervals. Further, each detection means is installed around the area to be measured so that the measurement data can be provided to the disaster prediction server 1.

In S601, the disaster prediction server 1 acquires the correlation data from a predetermined provider. It is assumed that the provider of the correlation data is specified in advance and information about the provider is retained. It should be noted that the correlation data does not need to be acquired every time this process is executed, and may be configured to be acquired at the timing when the correlation data is updated.

In S602, the disaster prediction server 1 acquires measurement data from each of the installed detection means. At this time, the measurement data to be used for the subsequent processing may be extracted according to the content of the acquired measurement data. For example, when the image quality of the captured image acquired from the camera 3 is low, the configuration may be such that it is excluded from the data used for the subsequent processing. Further, when the measurement data cannot be acquired from the detection means at a certain timing, the disaster prediction server 1 may request the measurement data from the detection means. Alternatively, if the measurement data cannot be acquired, the information may be treated as the measurement data, assuming that some kind of disaster may have occurred in the measurement range.

In S603, the disaster prediction server 1 identifies the ground surface shape of the target area by generating a three-dimensional model based on the measurement data acquired from each detection means. The disaster prediction server 1 newly generates a three-dimensional model from the measurement data acquired in S602. It is assumed that the generated three-dimensional model is held in the storage device each time.

In S604, the disaster prediction server 1 sets the degree of risk based on the correlation data acquired in S601 and the ground surface shape generated in S603. In addition, a plurality of risk levels may be set according to the type of disaster. In the present embodiment, the degree of risk for the amount of change obtained based on the detection results detected by each of the plurality of types of detection means is set.

In S605, the disaster prediction server 1 acquires the weather information in the target area from the weather information providing server 2. The meteorological information acquired here includes past precipitation and future precipitation forecast in the target area. The period included in the meteorological information (past period, predicted period) is not particularly limited. For example, the target period may vary depending on the surface shape, the total amount of precipitation in the past, and the like.

In S606, the disaster prediction server 1 predicts the degree of danger in the target area based on the weather information acquired in S605 and the degree of danger set in S604. In the present embodiment, among the detection results by the plurality of types of detection means, the value with the highest risk is predicted as the risk in the area. The prediction method is not limited to this, and even if the risk level is predicted based on the combination of risk levels set corresponding to the detection results by a plurality of types of detection means and the comprehensive evaluation of the detection results. good. The comprehensive evaluation here may be determined by weighting each detection result, or may be evaluated based on a predetermined criterion.

In S607, the disaster prediction server 1 controls the display of the three-dimensional model of the target area based on the prediction result in S606. The display here is performed by superimposing a display according to the degree of risk on the three-dimensional model and providing it to the display device 7 as display data.

In S608, the disaster prediction server 1 issues an alarm based on the prediction result in S607. The alarm here may be configured to blink or give an evacuation warning at the time of display on the display device 7. Alternatively, the configuration may be such that an evacuation warning or information on an area predicted to be dangerous is notified via an alarm device (not shown) such as a speaker installed around the display device 7. Then, this processing flow is terminated.

As described above, in the present embodiment, a three-dimensional model is generated by combining a plurality of types of detection means in consideration of the advantages and disadvantages of the detection means provided for the target area, and the three-dimensional model can be obtained. Disaster prediction is performed based on the ground surface shape and the degree of risk obtained from the correlation data. This makes it possible to predict and detect the occurrence of sediment-related disasters, which is less affected by environmental conditions and is more stable than before.

<Third embodiment>
Hereinafter, a third embodiment of the present invention will be described. In the third embodiment, an embodiment in which the trained model is used when setting the degree of risk will be described. It should be noted that the description of the configuration overlapping with the second embodiment will be omitted, and the description will be given focusing on the difference.

[Functional configuration]
FIG. 7 is a diagram showing an example of the functional configuration of the disaster prediction server 1 according to the present embodiment. In the first embodiment, it is used in place of the configuration described with reference to FIG. 3 (a). The disaster prediction server 1 according to the present embodiment replaces the weather information analysis unit 103, the shape change detection unit 105, and the danger range identification unit 107 with the learning data generation unit 111, the learning processing unit 112, and the disaster prediction unit 113. Consists of including.

The learning data generation unit 111 generates learning data, which will be described later, using the measurement information, weather information, and the like collected by the data collection unit 101. The learning processing unit 112 performs learning processing using the learning data generated by the learning data generation unit 111, and generates a trained model. The disaster prediction unit 113 makes a disaster prediction using the learned model generated by the learning processing unit 112.

[Learning process]
In the present embodiment, the historical information of the ground surface shape and precipitation in the area where the past sediment-related disaster has occurred is used as input data, and a trained model for outputting the degree of risk is generated. The learning method according to the present embodiment will be described as using supervised learning by a neural network, but other methods may be used.

FIG. 8 is a diagram for explaining the concept of the learning process according to the present embodiment. The learning data used in this embodiment is composed of a pair of input data and teacher data. The input data includes, for example, the inclination angle of the land in the target area, the type of land surface, the funnel-shaped land area, the precipitation amount, the cumulative precipitation amount, and the precipitation prediction amount. The type of land surface indicates the constituent materials of the land surface (trees, stones, sand, etc.). Precipitation indicates past precipitation during a predetermined period (time, day, etc.). Cumulative precipitation indicates the total amount of precipitation in a predetermined period. Precipitation forecast indicates future precipitation forecast at a predetermined time point. The input data is not limited to the above, and may include other information. In addition, the teacher data is a level indicating the degree of risk of disaster occurrence. This level of risk may be defined by the magnitude of past disasters.

When the input data is input to the learning model, the output data indicating the probability of being classified into the risk level 1 to level 4 is output. Then, the output data and the teacher data are compared using the loss function, and the weights in the learning model are adjusted to update the parameters of the learning model. A trained model is generated by repeating this process. That is, in the present embodiment, the trained model operates as a classifier. The learning process may be repeated every time training data is added, and the trained model may be updated according to the learning result.

[Processing flow]
Hereinafter, the processing flow of the disaster prediction server 1 according to the present embodiment will be described. The processing shown below is realized, for example, by the CPU (not shown) included in the disaster prediction server 1 reading and executing the program stored in the storage device. The following processing may be started at predetermined intervals. Further, each detection means is installed around the area to be measured so that the measurement data can be provided to the disaster prediction server 1. Further, it is assumed that the above-mentioned learning process is performed and the trained model has already been generated.

In S901, the disaster prediction server 1 acquires measurement data from each of the installed detection means.

In S902, the disaster prediction server 1 acquires the weather information in the target area from the weather information providing server 2. The meteorological information acquired here includes past precipitation and future precipitation forecast in the target area.

In S903, the disaster prediction server 1 generates input data to be input to the trained model by using the measurement data acquired in S901 and the weather information acquired in S902. It is assumed that the information included in the input data includes, for example, the items shown in FIG.

In S904, the disaster prediction server 1 predicts the risk of disaster in the target area by inputting the input data generated in S903 into the trained model. More specifically, by applying the input data to the trained model, it is output as output data which of the plurality of levels that the target area is set as the risk level corresponds to.

In S905, the disaster prediction server 1 controls the display of the three-dimensional model of the target area based on the prediction result in S904. The display here is performed by superimposing a display according to the degree of risk on the three-dimensional model and providing it to the display device 7 as display data.

In S906, the disaster prediction server 1 issues an alarm based on the prediction result in S907. The alarm here may be configured to blink or give an evacuation warning at the time of display on the display device 7. Alternatively, the configuration may be such that an evacuation warning or information on an area predicted to be dangerous is notified via an alarm device (not shown) such as a speaker installed around the display device 7. Then, this processing flow is terminated.

[Display example]
FIG. 10 is a diagram showing a configuration example of a screen displayed by the display device 7 according to the present embodiment. In FIG. 10, region 1001 indicates the range in which the degree of risk is predicted as level 3. Region 1002 indicates the range in which the risk is predicted as level 2. In FIG. 10, the regions other than the region 1001 and the region 1002 indicate the range predicted as level 1. As shown in FIG. 10, the display content is switched according to the level of risk. The display method may be a method of displaying in different colors or blinking.

Further, the display device 7 may display the evacuation route according to the level of danger. As the route at this time, the route may be calculated so as not to approach the high-risk area, and a place where damage is unlikely to occur may be set as the destination. When calculating the route, for example, it may be linked with a navigation system (not shown) using map information. Further, when the display device 7 is a mobile terminal or the like, the evacuation guidance may be performed in conjunction with the navigation application or the like mounted on the display device 7. The display method and the notification method may be configured to be arbitrarily set by the user of the system according to the present embodiment.

As described above, in the present embodiment, in addition to the effect of the first embodiment, by using the trained model obtained by the learning process, it is possible to speed up the process at the time of prediction.

Further, in the present invention, one or more programs or applications for realizing the functions of one or more embodiments described above are supplied to a system or device using a network or a storage medium, and the system or device is used in a computer. It can also be realized by the process of reading and executing the program by the processor of.

Further, it may be realized by a circuit that realizes one or more functions (for example, an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array)). Further, in machine learning, a GPU (Graphical Processing Unit) or the like may be used in addition to the CPU (Central Processing Unit).

1 Disaster prediction server 2 Meteorological information providing server 3 Camera 4 Laser scanner 5 Radar distance measuring device 6 Ground surface position measuring device 7 Display device 8 Network 101 Data collection unit 102 3D model generation unit 103 Meteorological information analysis unit 104 Communication control unit 105 Shape change detection unit 106 History information management unit 107 Danger range identification unit 108 Display control unit

In order to solve the above problems, the present invention has the following configuration. That is,
It ’s a disaster prediction system.
The first detection means for detecting changes in the state of features in the target area,
A second detection means for detecting a change in the state of the feature by a detection method having a different detection principle from the first detection means,
The first detection means in the target area and the said Second detection means A prediction means for predicting the risk of sediment-related disasters in the target area based on the detection results of each, and a prediction means.
It has a display means for displaying the change in the state of the target area and the feature and the risk of occurrence of the sediment-related disaster.

Claims (11)

The first detection means for detecting changes in the state of features in the target area,
A second detection means that detects a change in the state of the feature by a detection principle different from that of the first detection means, and
Based on the change in the state of the feature obtained by the detection results of the first detection means and the second detection means, the prediction means for predicting the risk of sediment-related disaster in the target area, and the prediction means.
A disaster prediction system characterized by having a display means for displaying a change in the state of the target area and the feature and the risk of occurrence of the sediment-related disaster. It also has a means to acquire historical information including information on the shape and precipitation of features where sediment-related disasters have occurred.
The first aspect of the present invention is characterized in that the prediction means sets the risk of sediment-related disaster occurrence for the detection results of each of the first detection means and the second detection means based on the history information. Described disaster prediction system. The disaster prediction system according to claim 2, wherein the historical information is information showing a correlation between precipitation and ground surface shape obtained by a statistical method using past statistical data. The prediction means comprehensively evaluates changes in the state of the feature obtained from the detection results of the first detection means and the second detection means, and based on the comprehensive evaluation, in the target area. The disaster prediction system according to any one of claims 1 to 3, wherein the risk of occurrence of a sediment-related disaster is predicted. The first detection means in the target area and the said The disaster prediction system according to any one of claims 1 to 4, further comprising a prediction means for predicting the degree of danger of the target feature based on the detection result of each of the second detection means. The combination of the first detecting means and the second detecting means is
Consists of at least two combinations of a camera that acquires image data, a laser scanner that derives point cloud data, a GNSS (Global Navigation Satellite System) device that detects position data, and a radar distance measurement device that measures distance data. The disaster prediction system according to any one of claims 1 to 5, wherein the disaster prediction system is characterized. Further having a modeling means for modeling a three-dimensional ground surface shape using the information obtained by the first detecting means and the second detecting means.
The display means is characterized in that the risk of occurrence of a sediment-related disaster predicted by the prediction means is visualized and displayed on a three-dimensional ground surface shape modeled by the modeling means. The disaster prediction system according to any one of 6. The display means visualizes and displays the risk of sediment-related disasters predicted by the prediction means on a three-dimensional ground surface shape modeled by the modeling means so as to change in chronological order. The disaster prediction system according to claim 7, wherein the disaster prediction system is performed. The display means is characterized in that the three-dimensional ground surface shape modeled by the modeling means is constructed in a virtual space and output, or projected onto a real space and displayed. The disaster prediction system described in. The disaster prediction system according to any one of claims 1 to 9, further comprising a notification means for notifying the evacuation behavior in the target area according to the degree of danger predicted by the prediction means. The process of acquiring the change in the state of the feature in the target area detected by the first detection means, and
A step of acquiring a change in the state of the feature detected by a second detection means having a detection principle different from that of the first detection means, and
A step of predicting a change in the state of the feature obtained by the detection results of each of the first detection means and the second detection means, and a step of predicting the risk of occurrence of a sediment-related disaster in the target area.
A disaster prediction method comprising a display step for displaying a change in the state of the target area, the feature, and the risk of occurrence of the sediment-related disaster.

Images