TWI806321B - Inundation assessment method and computing device - Google Patents

Abstract

A flood assessment method implemented by a computing device, when the computing device receives multiple pieces of rainfall data related to a region, according to a plurality of different observation coordinates of each rainfall data and its corresponding rainfall, use An interpolation method obtains the rainfall corresponding to each location coordinate as the processed rainfall data, and normalizes multiple location coordinates of the coordinate data to obtain a normalized coordinate data, and each normalizing the rainfall of the processed rainfall data to obtain a normalized rainfall data, and using a flood assessment model based on the normalized coordinate data and the normalized rainfall data, Obtain a flood assessment result including multiple flood depths of the locations in the area at an assessment time point.

Description

Inundation assessment method and computing device

The present invention relates to a flooding assessment method, in particular to a flooding assessment method for assessing a flooding depth related to an area and a computing device thereof.

In the past, when the climate changed, countries around the world were prone to serious flooding problems. Taking Taiwan as an example, there was a severe drought in early 2021, but in early August 2021, the southwesterly airflow brought heavy rain, causing considerable agricultural and economic losses in the south. Aiming at the problem that climate change may rapidly cause flooding in any area, its existing flooding assessment method used in flood prevention has one-dimensional machine learning and two-dimensional flooding physical model, of which one-dimensional machine learning can only be based on real-time observations The data assesses the degree of flooding in the next 1 to 3 hours from the current time point. Since the evaluation data based on it must be real-time observation data, the use is relatively limited, and the time point of flooding that can be estimated is also limited by the availability of For the next 1 to 3 hours of real-time observation data, the evaluation results of the two-dimensional flooding physical model need to consume relatively high computing time.

Therefore, if a method can be proposed to reduce the calculation time and evaluate the future time point that is longer than the current time point, the flood risk can be predicted quickly and early, so that disaster prevention plans can be formulated and pre-disaster preparations can be made. stage can be properly executed.

Therefore, the purpose of the present invention is to provide a flood assessment method that can instantly and automatically assess the flood depth of a region at a future time point that is longer than the current time point.

Therefore, the flood assessment method of the present invention is implemented by a computing device, which stores a flood assessment model for assessing a flood depth related to a region, and a plurality of different models related to the region Coordinate data of a location, the coordinate data includes a plurality of location coordinates corresponding to the locations, and the flood assessment method includes a step (A), a step (B), a step (C), and a step (D).

The step (A) is when the computing device receives multiple pieces of rainfall data related to the region in different time periods, each rainfall data includes the observation coordinates of multiple different observation points, and multiple corresponding observation points The rainfall amount, these time periods are respectively defined by a plurality of different previous time points that are earlier than an evaluation time point and the evaluation time point, for each rainfall data, the calculation device is based on the rainfall data of the Observation coordinates and their corresponding rainfall are obtained by using an interpolation method to obtain the rainfall corresponding to each location coordinate as the processed rainfall data.

In the step (B), the computing device normalizes the location coordinates of the coordinate data to obtain a normalized coordinate data.

In the step (C), for each processed rainfall data, the calculation device normalizes the rainfall amount of the processed rainfall data to obtain a normalized rainfall data.

The step (D) is that the calculation device obtains an estimation time point of the locations including the region by using the flood assessment model based on the normalized coordinate data and the normalized rainfall data. Flood assessment results for multiple flood depths.

Another object of the present invention is to provide a computing device capable of instantly and automatically evaluating the flooding depth of an area at a future time point that is longer than the current time point.

Therefore, the computing device of the present invention includes a storage module and a processing module.

The storage module is used to store a flood assessment model for assessing a flood depth related to a region, and a coordinate data related to a plurality of different locations in the region, the coordinate data includes a plurality of corresponding The location coordinates of the location.

The processing module is electrically connected to the storage module.

Among them, when the processing module receives multiple pieces of rainfall data related to the region in different time periods, each rainfall data includes the observation coordinates of multiple different observation points, and multiple rainfall corresponding to the same observation point , the time intervals are respectively defined by a plurality of different previous time points earlier than an evaluation time point and the evaluation time point, for each rainfall data, the processing module is based on the observations of the rainfall data Coordinates and their corresponding rainfall, use an interpolation method to obtain the rainfall corresponding to each location coordinate, as the processed rainfall data, the processing module will save the coordinate data stored in the storage module The location coordinates are normalized to obtain a normalized coordinate data, and for each processed rainfall data, the processing module normalizes the rainfall of the processed rainfall data to obtain a normalized According to the normalized coordinate data and the normalized rainfall data, the processing module uses the flood assessment model stored in the storage module to obtain a list of the locations including the region Flood assessment results for multiple flood depths at that assessment time point.

The effect of the present invention is that: the processing module obtains the processed rainfall data by using the interpolation method according to the rainfall data, so that the locations covered by the processed rainfall data can be consistent with the locations in the area. In addition, Obtain the normalized coordinate data and the normalized rainfall data through the processing module, and use the flood assessment model to obtain the flood assessment result, because the rainfall data does not need to be limited to real-time The observation data can be estimated rainfall data, so as long as the estimated rainfall data corresponding to different time periods before the future time point can be obtained, the flooding status of the area at the future time point can be evaluated in real time.

Before the present invention is described in detail, it should be noted that in the following description, similar elements are denoted by the same numerals.

Referring to FIG. 1 , an embodiment of the flood assessment method of the present invention is implemented by a computing device, which includes a storage module 1 and a processing module 2 electrically connected to the storage module 1 .

The storage module 1 stores a flood assessment model for assessing a flood depth related to the region, a coordinate data of a plurality of different points related to the region, and the points that are not likely to be flooded in the region Non-flooded location coordinates of at least one non-flooded location of water, the coordinate data includes a plurality of location coordinates corresponding to the location, each location coordinate is a three-dimensional coordinate, and includes an X coordinate, a Y coordinate, and a Z-coordinate.

The storage module 1 also stores a plurality of flooding event data related to the flooding event in the area, and each flooding event data includes multiple flooding rainfall data corresponding to different rainfall time periods, and a pair of The inundation data should wait for the inundation rainfall data. Each flooding data includes the flooding depths of multiple corresponding locations measured at a corresponding flooding measurement time point when the corresponding flooding event occurs, and each flooding rainfall data includes multiple The measurement coordinates of different measurement points, and the multiple rainfall corresponding to these measurement points when the corresponding flooding event occurs, the rainfall time intervals corresponding to each flooding event data are respectively from earlier than A plurality of different rainfall time points corresponding to the flooding measurement time point when the corresponding flooding event occurs are respectively defined by the corresponding flooding measurement time point when the corresponding flooding event occurs.

For example, if the flood measurement time point is 10 am on November 10, 2018, then the rainfall time points are respectively 9 am, 8 am, 7 am, 6 am on November 10, 2018. at 12:00 on the night of November 9, 2018, and at 10:00 on the night of November 9, 2018, that is, the rainfall data are earlier than the flooding measurement time point The cumulative rainfall from 1 hour before the flooding measurement time point to the flooding measurement time point, the accumulated rainfall from the 2 hours before the flooding measurement time point to the flooding measurement time point, and the flooding measurement time point earlier than the flooding The accumulated rainfall from 3 hours before the measurement time point to the inundation measurement time point, the accumulated rainfall from the 4 hours before the inundation measurement time point to the inundation measurement time point, the early The cumulative rainfall from 6 hours before the flooding measurement time point to the flooding measurement time point, the 8 hours earlier than the flooding measurement time point, and the rainfall earlier than the flooding measurement time point The cumulative rainfall from the previous 10 hours to the flooding measurement time point, and the accumulated rainfall from the previous 12 hours to the flooding measurement time point earlier than the flooding measurement time point.

Referring to FIG. 1, the computing device can be a tablet computer, a notebook computer, a server or a personal computer, but not limited thereto.

The details of the operation of each element in the computing device will be described below in conjunction with the embodiment of the flood assessment method of the present invention. The embodiment of the flood assessment method includes a flood assessment model for establishing the flood assessment model Establish procedures, and a flood assessment procedure for assessing the flood depth of the area.

The flood assessment model building procedure includes a step 51 , a step 52 , a step 53 , a step 54 , a step 55 , and a step 56 .

The flood assessment procedure includes a step 61 , a step 62 , a step 63 , and a step 64 .

Referring to Fig. 1 and Fig. 2, the establishment procedure of the flood assessment model includes the following steps.

In step 51, the processing module 2 obtains a plurality of target flooding data through a data screening method through the flooding event data stored in the storage module 1, wherein the data screening method is a percentile method (Percentile), that is, the percentile of the amount of data that samples the flooding event data.

In step 52, for each flooding rainfall data of the target flooding data, the processing module 2 uses an interpolation method to obtain The rainfall corresponding to each location coordinate is used as the flooded rainfall data after processing. Generally speaking, the quantity of these measurement points will be less than the quantity of these locations in this area, and these measurement coordinates and these location coordinates are not completely consistent, by the execution of step 52, each The processed inundation rainfall data all contain the rainfall at all location coordinates, and are consistent with the locations covered by the corresponding inundation data.

In step 53 , the processing module 2 normalizes the location coordinates of the coordinate data stored in the storage module 1 to obtain a normalized coordinate data. It is worth mentioning that since flooding has a certain relationship with the location, the location coordinates of these locations in the area are also indispensable items when establishing the flood assessment model.

In step 54, for each processed flood rainfall data, the processing module 2 normalizes the rainfall of the processed flood rainfall data to obtain a normalized flood rainfall data

In step 55, for each target flood data, the processing module 2 corresponds to the normalized flood rainfall data corresponding to the target flood data, the normalized coordinate data, and the corresponding Depth of flooding, as a set of flooding training data.

In step 56, the processing module 2 uses a machine learning algorithm to establish the flood assessment model according to the flood training data, wherein the machine learning algorithm can be a convolutional neural network (CNN, Convolutional Neural Network) calculation model.

Referring to Figure 1 and Figure 3, the flood assessment procedure includes the following steps.

In step 61, when the processing module 2 receives multiple pieces of rainfall data related to the region in different time periods, each rainfall data includes observation coordinates of multiple different observation points, and multiple corresponding observations The rainfall at the point, these time intervals are respectively defined by a plurality of different previous time points earlier than an evaluation time point and the evaluation time point, for each rainfall data, the processing module 2 according to the rainfall For the observation coordinates of the data and the corresponding rainfall, the interpolation method is used to obtain the rainfall corresponding to each location coordinate as the processed rainfall data.

For example, if the evaluation time point is 10:00 am tomorrow, the rainfall data of the time intervals defined by the previous time points and the evaluation time point can be obtained through the forecast of the Meteorological Bureau, where These previous time points are 9 o'clock, 8 o'clock, 7 o'clock, 6 o'clock, 4 o'clock, 2 o'clock tomorrow morning, 12 o'clock this evening, and 10 o'clock this evening. Compared with conventional methods that can only assess the flooding degree 1 to 3 hours from the current time point based on real-time observation data, if the weather forecast can provide rainfall data at a future time point that is longer than the current time point, it can be used The inundation assessment model estimates the inundation depth at a further future point in time.

In step 62, the processing module 2 normalizes the location coordinates of the coordinate data stored in the storage module 1 to obtain the normalized coordinate data.

Referring to FIG. 1 and FIG. 4 , it is worth noting that step 62 includes the following sub-steps.

In step 621 , the processing module 2 normalizes the X coordinates of the location coordinates stored in the storage module 1 to obtain a plurality of normalized X coordinates.

In step 622 , the processing module 2 normalizes the Y coordinates of the location coordinates stored in the storage module 1 to obtain a plurality of normalized Y coordinates.

In step 623 , the processing module 2 normalizes the Z coordinates of the location coordinates stored in the storage module 1 to obtain a plurality of normalized Z coordinates.

Wherein, the normalized X coordinates, the normalized Y coordinates, and the normalized Z coordinates together constitute the normalized coordinate data.

Referring to FIG. 1 and FIG. 5 , it is worth noting that step 623 includes the following sub-steps.

In step 623a, the processing module 2 sets the Z coordinate of the at least one non-flooded location coordinate stored in the storage module 1 as the maximum value among the Z coordinates of the location coordinates. Since the larger the coordinate value of the Z coordinate, it means that the terrain of the site is higher, and the higher the terrain, the less likely it is to be flooded. Therefore, by setting the Z coordinate of the at least one non-flooded location coordinate as one of the location coordinates The maximum value in the Z coordinate is such that the flooding depth corresponding to at least one non-flooded location substantially indicates no flooding when the flooding assessment is performed.

In step 623b, the processing module 2 normalizes the Z coordinates of the location coordinates set in step 623a to obtain the normalized Z coordinates.

In step 63, for each processed rainfall data, the processing module 2 processes each rainfall data in the processed rainfall data according to the maximum rainfall in the processed rainfall data corresponding to the longest time segment Normalize the amount to obtain a normalized rainfall data.

In step 64, the processing module 2 uses the flood assessment model stored in the storage module 1 according to the normalized coordinate data and the normalized rainfall data to obtain a map containing the area. The results of the inundation assessment for the sites at various inundation depths at that assessment time point.

To sum up, in the flood assessment method of the present invention, the processing module 2 receives the rainfall data related to the region in different time periods, and uses the interpolation method to obtain the processed rainfall data according to each rainfall data. The rainfall data can make the locations covered by the processed rainfall data consistent with those locations in the area. In addition, the normalized coordinate data and the normalized rainfall data can be obtained through the processing module 2 data, and use the flood assessment model to obtain the flood assessment results. Since such rainfall data need not be limited to real-time observation data, it can be estimated rainfall data, so as long as the estimated rainfall data corresponding to different time periods before the future time point can be obtained, the The flooding situation of the region at the future time point, so the purpose of the present invention can be really achieved.

But the above-mentioned ones are only embodiments of the present invention, and should not limit the scope of the present invention. All simple equivalent changes and modifications made according to the patent scope of the present invention and the content of the patent specification are still within the scope of the present invention. Within the scope covered by the patent of the present invention.

1: Storage module

2: Processing module

51~56: Steps

61~64: Steps

621~623: Steps

623a~623b: steps

Other features and effects of the present invention will be clearly presented in the implementation manner with reference to the drawings, wherein: FIG. 1 is a block diagram illustrating a computing device for performing one embodiment of the flooding assessment method of the present invention; Fig. 2 is a flowchart illustrating a flood assessment model building procedure of this embodiment of the flood assessment method of the present invention; FIG. 3 is a flow chart illustrating a flood assessment procedure of the embodiment of the flood assessment method of the present invention; Fig. 4 is a flow chart illustrating the detailed flow of how a processing module normalizes coordinate data; and FIG. 5 is a flow chart illustrating the detailed process of how the processing module normalizes the Z coordinates of multiple location coordinates.

61~64: Steps

Claims (6)

A flood assessment method implemented by a computing device storing a flood assessment model for assessing a flood depth related to a region, coordinate data related to a plurality of different locations in the region , and the non-flooded location coordinates of at least one non-flooded location among the locations in the region that will not be flooded, the coordinate data includes a plurality of location coordinates corresponding to such locations, and each location coordinate is a three-dimensional coordinate, And including an X coordinate, a Y coordinate, and a Z coordinate, the flood assessment method includes the following steps: (A) when the computing device receives multiple pieces of rainfall data related to the region in different time periods, each A rainfall data includes the observation coordinates of a plurality of different observation points, and a plurality of rainfall corresponding to the observation points, and these time segments are respectively connected with the evaluation time points by a plurality of different previous time points earlier than an evaluation time point As defined by the time point, for each rainfall data, the calculation device uses an interpolation method to obtain the rainfall corresponding to each location coordinate according to the observation coordinates of the rainfall data and the corresponding rainfall, as a processing (B) the computing device normalizes the location coordinates of the coordinate data to obtain a normalized coordinate data, wherein step (B) includes the following sub-steps, (B-1) The computing device normalizes the X coordinates of the location coordinates to obtain a plurality of normalized X coordinates, and (B-2) the computing device normalizes the Y coordinates of the location coordinates to obtain a plurality of The normalized Y coordinate, (B-3) The computing device normalizes the Z coordinates of the location coordinates to obtain a plurality of normalized Z coordinates, wherein the normalized X coordinates, the normalized Y coordinates , and the normalized Z coordinates together constitute the normalized coordinate data, wherein step (B-3) includes the following sub-steps, (B-3-1) the computing device takes at least one non-flooded The Z coordinate of the location coordinates is set to the maximum value of the Z coordinates of the location coordinates, and (B-3-2) the computing device will set the Z coordinates of the location coordinates after step (B-3-1) Coordinates are normalized to obtain the normalized Z coordinates; (C) For each processed rainfall data, the computing device normalizes the rainfall of the processed rainfall data to obtain a normalized and (D) the computing device, based on the normalized coordinate data and the normalized rainfall data, uses the flooding assessment model to obtain a set of locations including the region in the assessment Inundation assessment results for multiple inundation depths at time points. The flood assessment method as described in Claim 1, wherein, in step (C), for each processed rainfall data, the computing device calculates the maximum rainfall in the processed rainfall data corresponding to the longest time segment Each rainfall amount in the processed rainfall data is normalized to obtain the normalized rainfall data. According to the flood assessment method described in claim 1, the computing device also stores multiple flood event data related to the flood events that occurred in the area, and each flood The water event data includes multiple flooding rainfall data corresponding to different rainfall time periods, and a pair of flooding data corresponding to the flooding rainfall data. Each flooding data contains a corresponding flooding event when the corresponding flooding event occurs The flooding depths measured at multiple corresponding locations at the flooding measurement time point, each flooding rainfall data includes the measurement coordinates of multiple different measurement points, and the occurrence of the corresponding flooding event The multiple rainfall amounts corresponding to the measurement points at that time, and the rainfall time segments corresponding to each flood event data are respectively earlier than the corresponding flood measurement time points when the corresponding flood event occurs A plurality of different rainfall time points are defined by the corresponding flood measurement time points when the corresponding flood event occurs, wherein, before step (A), the following steps are also included: (E) the computing device Through the flooding event data, a data screening method is used to obtain multiple pieces of target flooding data; (F) for each flooding rainfall data of the target flooding data, the computing device is based on the flooding rainfall data Equal measurement coordinates and their corresponding rainfall, use the interpolation method to obtain the rainfall corresponding to the coordinates of each location, as the processed flood rainfall data; (G) the computing device will use the coordinate data The location coordinates are normalized to obtain the normalized coordinate data; (H) for each processed flood rainfall data, the computing device normalizes the rainfall of the processed flood rainfall data to obtain Obtain a normalized flooding rainfall data; (1) for each target flooding data, the computing device corresponding to the normalized flooding rainfall data of the target flooding data, the normalized constellation The target data and the corresponding flooding depths are used as a set of flooding training data; and (J) the computing device uses a machine learning algorithm to establish the flooding assessment model based on the flooding training data. A computing device for estimating flooding depth, comprising: a storage module for storing a flooding assessment model for estimating a flooding depth related to an area, a plurality of different locations related to the area , and the non-flooded location coordinates of at least one non-flooded location among the locations in the region that will not be flooded, the coordinates data includes multiple location coordinates corresponding to such locations, and each location coordinate is a Three-dimensional coordinates, including an X coordinate, a Y coordinate, and a Z coordinate; and a processing module electrically connected to the storage module; wherein, when the processing module receives multiple transactions related to the region in different time zones When the rainfall data for a period of time is used, each rainfall data contains the observation coordinates of multiple different observation points, and multiple rainfall amounts corresponding to the observation points. For each rainfall data defined by the previous time points and the evaluation time point, the processing module uses an interpolation method to obtain the coordinates of each location according to the observation coordinates of the rainfall data and the corresponding rainfall. The corresponding rainfall is used as the processed rainfall data. The processing module normalizes the location coordinates of the coordinate data stored in the storage module to obtain a normalized coordinate data, wherein the processing The module normalizes the X coordinates of the location coordinates to obtain multiple normalized X coordinates, and the processing module normalizes the Y coordinates of the location coordinates to obtain multiple normalized Y coordinates, the processing module normalizes the Z coordinates of the location coordinates to obtain a plurality of normalized Z coordinates, and the normalized X coordinates, the normalized The Y coordinate and the normalized Z coordinates together constitute the normalized coordinate data, wherein the processing module sets the Z coordinate of the at least one non-flooded location coordinate stored in the storage module as the The maximum value of the Z coordinates of the location coordinates, and normalize the Z coordinates of the location coordinates after setting to obtain the normalized Z coordinates, and for each processed rainfall data, the processing The module normalizes the rainfall of the processed rainfall data to obtain a normalized rainfall data. The processing module uses the normalized coordinate data and the normalized rainfall data to use The flooding assessment model stored in the storage module obtains a flooding assessment result including multiple flooding depths of the locations in the area at the assessment time point. The computing device according to claim 4, wherein, for each processed rainfall data, the processing module calculates the processed rainfall data according to the maximum rainfall in the processed rainfall data corresponding to the longest time segment Each rainfall amount in the data is normalized to obtain the normalized rainfall data. The computing device as described in claim 4, wherein the storage module also stores a plurality of flooding event data related to flooding events in the region, and each flooding event data includes multiple records corresponding to different rainfall times The flooding rainfall data of the section, and the flooding data corresponding to the flooding rainfall data, each flooding data includes the measurement at a corresponding flooding measurement time point when the corresponding flooding event occurs There are multiple flooding depths corresponding to the corresponding locations, and each flooding rainfall data contains the measurement coordinates of multiple different measurement points, And when the corresponding flooding event occurs, the rainfall corresponding to the multiple measurement points, and the rainfall time period corresponding to each flooding event data is from earlier than the corresponding flooding event. A plurality of different rainfall time points corresponding to the flooding measurement time point are defined by the corresponding flooding measurement time point when the corresponding flooding event occurs, and the processing module is stored by the storage module For the flooding event data, a data screening method is used to obtain multiple pieces of target flooding data, and for each flooding rainfall data of the target flooding data, the processing module is based on the measurement of the flooding rainfall data Coordinates and their corresponding rainfall, use the interpolation method to obtain the rainfall corresponding to the coordinates of each location, as the flooded rainfall data after processing, the processing module will store the coordinate data stored in the module The coordinates of the locations are normalized to obtain the normalized coordinate data, and for each processed flood rainfall data, the processing module normalizes the rainfall of the processed flood rainfall data to obtain Obtaining a normalized flooding rainfall data, for each target flooding data, the processing module corresponds to the normalized flooding rainfall data corresponding to the target flooding data, the normalized coordinate data, and The corresponding flooding depths are used as a set of flooding training data, and according to the flooding training data, a machine learning algorithm is used to establish the flooding assessment model.

Images