CN114881381B - Urban ponding water level prediction method and system based on improved convolutional neural network - Google Patents

Urban ponding water level prediction method and system based on improved convolutional neural network Download PDF

Info

Publication number
CN114881381B
CN114881381B CN202210815291.4A CN202210815291A CN114881381B CN 114881381 B CN114881381 B CN 114881381B CN 202210815291 A CN202210815291 A CN 202210815291A CN 114881381 B CN114881381 B CN 114881381B
Authority
CN
China
Prior art keywords
water level
data
preset time
urban
neural network
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202210815291.4A
Other languages
Chinese (zh)
Other versions
CN114881381A (en
Inventor
智协飞
吕阳
季焱
朱寿鹏
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing University of Information Science and Technology
Original Assignee
Nanjing University of Information Science and Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing University of Information Science and Technology filed Critical Nanjing University of Information Science and Technology
Priority to CN202210815291.4A priority Critical patent/CN114881381B/en
Publication of CN114881381A publication Critical patent/CN114881381A/en
Application granted granted Critical
Publication of CN114881381B publication Critical patent/CN114881381B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/24—Querying
    • G06F16/245—Query processing
    • G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
    • G06F16/2462—Approximate or statistical queries
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/24—Querying
    • G06F16/245—Query processing
    • G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
    • G06F16/2474—Sequence data queries, e.g. querying versioned data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/29—Geographical information databases
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/10—Services
    • G06Q50/26—Government or public services
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
    • Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Abstract

The invention discloses an urban waterlogging water level prediction method and system based on an improved convolutional neural network, belonging to the technical field of urban waterlogging water level prediction, wherein the method comprises the following steps: acquiring current ponding water level, urban ground elevation data and rainfall forecast data within a future preset time; identifying a rainribbon around a target station based on precipitation forecast data in a future preset time, and extracting object attributes of the rainribbon; preprocessing rainfall forecast data, urban ground elevation data and object attributes of a rain belt to form input variables; inputting the input variable into a pre-trained deep learning model based on an improved convolutional neural network to obtain the ponding variability of the target station within the future preset time, and obtaining the ponding water level within the future preset time by combining the current ponding water level; the model has high nonlinearity and strong robustness after being trained, and compared with the prior art, the technical scheme of the invention has longer prediction timeliness and extremely high application value.

Description

Urban ponding water level prediction method and system based on improved convolutional neural network
Technical Field
The invention relates to an urban ponding water level prediction method and system based on an improved convolutional neural network, and belongs to the technical field of urban waterlogging water level prediction.
Background
Under the background of global warming, disastrous weather such as rainstorm and the like frequently occurs, and urban waterlogging is serious day by day, so that great threat is brought to the life and property safety of people; under the background, high-quality monitoring and forecasting of the water level of the accumulated water have important significance; at present, urban accumulated water level monitoring stations are deployed in part of cities, so that real-time monitoring of accumulated water levels in important road sections can be realized, and the urban monitoring system plays an important role in disaster prevention and reduction of cities, but high-quality prediction of the accumulated water levels still remains an important scientific and technical problem to be solved urgently.
The traditional hydrological model-based water level prediction method has the advantages that the long-short term memory neural network-based extrapolation of the ponding water level is researched, the ponding water level prediction in the future 3 hours can be provided, however, the scheme is difficult to output the ponding water level prediction data with longer prediction time effect, and the application value is limited; in addition, the traditional scheme mostly aims at minimizing the mean square error in model training, so that the smoothing effect is brought, and the prediction of the ponding water level under the large-magnitude rainfall scene is not facilitated.
Disclosure of Invention
The invention aims to provide a city ponding water level prediction method and system based on an improved convolutional neural network, and solves the problems of short prediction time, low application value and the like in the prior art.
In order to realize the purpose, the invention is realized by adopting the following technical scheme:
in a first aspect, the invention provides a method for predicting the water level of urban ponding based on an improved convolutional neural network, which comprises the following steps:
acquiring the current ponding water level, urban ground elevation data and rainfall forecast data in the future preset time;
identifying a rainribbon around a target station based on precipitation forecast data in a future preset time, and extracting object attributes of the rainribbon;
preprocessing rainfall forecast data, urban ground elevation data and object attributes of a rain belt to form input variables;
inputting the input variable into a pre-trained deep learning model based on an improved convolutional neural network to obtain the ponding variability of the target station within the future preset time, and obtaining the ponding water level within the future preset time by combining the current ponding water level;
the object attributes of the rain belt comprise a centroid position, an area and precipitation intensity, and are extracted by the following method:
selecting the geometric center of a rain belt as a centroid position, representing the area through the number of grids in the rain belt range, and taking the 10 th percentile of rainfall in the rain belt range according to the rainfall intensity;
the deep learning model based on the improved convolutional neural network comprises a convolutional layer, a pooling layer, an expansion layer and a full-link layer, and the last layer isAdding object attribute of rain belt into full connection layer, and adding object attribute of rain belt into full connection layerTSScoring as a loss function of the deep learning model;
the training method of the deep learning model based on the improved convolutional neural network comprises the following steps:
acquiring historical ponding water level, urban ground elevation data and rainfall forecast data, standardizing the ponding water level, the urban ground elevation data and the rainfall forecast data, and putting the data into a constructed hydrological feature data set;
identifying a rainzone based on rainfall forecast data, extracting object attributes of the rainzone, standardizing the rainzone, and forming a training set together with a hydrological feature data set;
and training the deep learning model by using the training set to obtain the trained deep learning model.
With reference to the first aspect, further, the current ponding water level, the urban ground elevation data, and the precipitation forecast data in the future preset time are obtained by the following method:
the method comprises the steps of collecting the current ponding water level in real time through an urban ponding monitoring station, obtaining urban ground elevation data from a geographic space data cloud open platform, and obtaining rainfall forecast data within the future preset time through a data interface of a meteorological department.
With reference to the first aspect, further, the rain belt is identified by:
performing feature extraction on precipitation forecast data in a future preset time through a MODE spatial inspection technology so as to identify a rain belt;
firstly, spatial smoothing is carried out, and the calculation formula is as follows:
Figure 100002_DEST_PATH_IMAGE001
wherein, the first and the second end of the pipe are connected with each other,
Figure 100002_DEST_PATH_IMAGE002
is a function of the filtering of the received signal,
Figure 100002_DEST_PATH_IMAGE003
is the original forecastThe field(s) is (are) such that,
Figure 100002_DEST_PATH_IMAGE004
is a water-falling field after filtration,
Figure 100002_DEST_PATH_IMAGE005
and
Figure 100002_DEST_PATH_IMAGE006
is the coordinates of the grid point or points,Ris the convolution radius;
then, controlling the precipitation threshold value, wherein the calculation formula is as follows:
Figure 100002_DEST_PATH_IMAGE007
wherein the content of the first and second substances,
Figure 100002_DEST_PATH_IMAGE008
is the field of the mask or masks to be used,Tis the precipitation threshold;
finally, identifying the rain belt, wherein the calculation formula is as follows:
Figure 100002_DEST_PATH_IMAGE009
wherein the content of the first and second substances,
Figure 100002_DEST_PATH_IMAGE010
is the identified rain strip.
With reference to the first aspect, further, the preprocessing includes removing outliers and normalizing the data, and the formula of the normalization process is as follows:
Figure 100002_DEST_PATH_IMAGE011
wherein, the first and the second end of the pipe are connected with each other,
Figure 100002_DEST_PATH_IMAGE012
is the result of the normalization of the results,
Figure 100002_DEST_PATH_IMAGE013
is that
Figure 100002_DEST_PATH_IMAGE014
Is determined by the average value of (a),
Figure 100002_DEST_PATH_IMAGE015
is that
Figure 598705DEST_PATH_IMAGE014
The standard deviation of (a) is determined,
Figure 250267DEST_PATH_IMAGE014
is the value of the data.
In combination with the first aspect, further, theTSThe score is calculated as follows:
Figure 100002_DEST_PATH_IMAGE016
wherein, the first and the second end of the pipe are connected with each other,
Figure 100002_DEST_PATH_IMAGE017
is the number of times that the forecast is correct,
Figure DEST_PATH_IMAGE018
the number of times of the empty report is,
Figure 100002_DEST_PATH_IMAGE019
is the number of false negative.
In a second aspect, the present invention also provides a system for predicting the water level of urban ponding based on an improved convolutional neural network, including:
a real-time data acquisition module: the system is used for acquiring the current ponding water level, urban ground elevation data and rainfall forecast data within the future preset time;
the rain belt object attribute extraction module: the system comprises a target station, a rainfall forecast database, a target station and a target server, wherein the target station is used for identifying a rainzone around the target station based on the rainfall forecast data in a future preset time and extracting object attributes of the rainzone;
an input variable construction module: the system is used for preprocessing rainfall forecast data, urban ground elevation data and object attributes of a rain belt to form input variables;
the accumulated water level prediction module: and the method is used for inputting input variables into a pre-trained deep learning model based on the improved convolutional neural network to obtain the ponding variability of the target station within the future preset time, and obtaining the ponding water level within the future preset time by combining the current ponding water level.
Compared with the prior art, the invention has the following beneficial effects:
(1) According to the urban ponding water level prediction method and system based on the improved convolutional neural network, rainfall forecast data is utilized, urban ground elevation data are fused, the obtained real-time data are processed to obtain input variables, the input variables are input into a pre-trained deep learning model based on the improved convolutional neural network, ponding water level prediction is completed by combining the current ponding water level, and the model has high nonlinearity and strong robustness after being trained, has longer prediction timeliness compared with the prior art, and has extremely high application value;
(2) The rainfall forecast data in the future preset time is subjected to feature extraction through the MODE spatial inspection technology so as to identify the raindrops, extract object attributes of the raindrops, and add the object attributes of the raindrops into the deep learning model for training, so that the feature enhancement effect is achieved, the model can better capture peripheral rainfall information, and the ponding water level prediction capability of the model is improved;
will be provided withTSThe score is used as a loss function of the deep learning model, so that the smoothing effect caused by taking the mean square error as the loss function is avoided, and the prediction capability of the ponding water level under the large-scale rainfall scene is effectively improved.
Drawings
FIG. 1 is a flow chart of a city ponding water level prediction method based on an improved convolutional neural network provided by an embodiment of the invention;
fig. 2 is a schematic structural diagram of a deep learning model based on an improved convolutional neural network provided by an embodiment of the present invention.
Detailed Description
The present invention is further described with reference to the accompanying drawings, and the following examples are only for clearly illustrating the technical solutions of the present invention, and should not be taken as limiting the scope of the present invention.
Example 1
As shown in fig. 1, the method for predicting the urban ponding water level based on the improved convolutional neural network provided by the embodiment of the present invention includes the following steps:
s1, acquiring current ponding water level, urban ground elevation data and rainfall forecast data in a future preset time.
The method comprises the steps that a ponding monitoring station acquires the current ponding water level of the ponding monitoring station in real time, urban ground elevation data are acquired from a geospatial data cloud open platform, precipitation forecast data within the future preset time are acquired through a data interface of a meteorological department, and the precipitation forecast data are numerical mode precipitation forecast data with high space-time resolution.
When the gridded ponding water level, the precipitation forecast data and the urban ground elevation data around each ponding monitoring station are extracted, the resolution is 1km, 20 multiplied by 20 fine grids around each ponding monitoring station are extracted, the range can be adjusted according to actual conditions, and the ponding water level data is processed into a ponding variability (ponding water level change within 3 hours) within 3 hours.
S2, identifying a rain strip around the target station based on the rainfall forecast data in the future preset time, and extracting the object attribute of the rain strip.
Performing feature extraction on precipitation forecast data in a future preset time through a MODE spatial inspection technology so as to identify a rain belt;
firstly, spatial smoothing is carried out, and the calculation formula is as follows:
Figure 390523DEST_PATH_IMAGE001
wherein the content of the first and second substances,
Figure 595852DEST_PATH_IMAGE002
is a function of the filtering of the received signal,
Figure 269279DEST_PATH_IMAGE003
is the original forecast field, and the forecast field,
Figure 170370DEST_PATH_IMAGE004
is a filtered water-falling field and is provided with a filter,
Figure 47190DEST_PATH_IMAGE005
and
Figure 714932DEST_PATH_IMAGE006
is the coordinates of the grid point or points,Ris the convolution radius;
then, carrying out precipitation threshold control, wherein the calculation formula is as follows:
Figure 977286DEST_PATH_IMAGE007
wherein the content of the first and second substances,
Figure 518120DEST_PATH_IMAGE008
is the field of the mask or masks to be used,Tis the precipitation threshold;
finally, identifying the rain belt, wherein the calculation formula is as follows:
Figure 6870DEST_PATH_IMAGE009
wherein the content of the first and second substances,
Figure 337357DEST_PATH_IMAGE010
is the identified rain strip.
Extracting an object attribute (rain belt characteristic) of the rain belt after the rain belt is identified; the method comprises the steps of selecting a geometric center of a rain belt as a centroid position, representing the area size by using the number of grids in a rain belt range, and taking the 10 th percentile of rainfall in the rain belt range as rainfall intensity.
And S3, preprocessing the precipitation forecast data, the urban ground elevation data and the object attributes of the rain belt to form input variables.
The preprocessing includes removing outliers and normalizing by the formula:
Figure 399424DEST_PATH_IMAGE011
wherein the content of the first and second substances,
Figure 317351DEST_PATH_IMAGE012
is the result of the normalization of the results,
Figure 965501DEST_PATH_IMAGE013
is that
Figure 37362DEST_PATH_IMAGE014
Is determined by the average value of (a),
Figure 415254DEST_PATH_IMAGE015
is that
Figure 91698DEST_PATH_IMAGE014
The standard deviation of (a) is determined,
Figure 289461DEST_PATH_IMAGE014
is the value of the data.
And S4, inputting the input variable into a pre-trained deep learning model based on the improved convolutional neural network to obtain the ponding variability of the target station within the future preset time, and obtaining the ponding water level within the future preset time by combining the current ponding water level.
The deep learning model based on the improved convolutional neural network is constructed in advance, the structure of the deep learning model is shown in figure 2, the deep learning model comprises a convolutional layer, a pooling layer, an expansion layer and a full connection layer, the object attribute (rain zone characteristic) of a rain zone is added into the last full connection layer, the operation plays a role in characteristic enhancement, the model is favorable for capturing peripheral rainfall information better, and therefore the ponding prediction capability of the model is improved.
The convolutional layer is a 5 × 5 convolution with a step size of 1, containing 6 convolution kernels; the pooling layer is 2 × 2 pooling and comprises 6 pooling cores; the expanded layer comprises 512 cells.
In the deep learning model also writeTSThe score is used as a function of the loss,TSthe calculation formula of the score is as follows:
Figure 961751DEST_PATH_IMAGE016
wherein the content of the first and second substances,arepresenting the number of correct forecasts (both forecasts and observations are of a specified magnitude),bindicating the number of empty reports (forecasts reach a specified magnitude without observations),crepresenting the number of false positives (observations of a specified magnitude but not forecasts).
Will be provided withTSThe score is used as a loss function, so that the smoothing effect caused by taking the mean square error as the loss function can be effectively avoided, and the forecasting capacity of the water level of the ponding under the large-scale rainfall scene is improved.
Acquiring historical ponding water level, urban ground elevation data and rainfall forecast data, standardizing the ponding water level, the urban ground elevation data and the rainfall forecast data, and putting the data into a constructed hydrological feature data set;
identifying a raindrop based on rainfall forecast data, extracting the object attribute of the raindrop, standardizing the raindrop, and forming a training set together with the hydrological feature data set;
and training the deep learning model by using the training set to obtain the trained deep learning model.
And training convergence conditions are that the loss function drop does not exceed a preset loss function threshold value, and the model weight change between two iterations does not exceed a preset weight threshold value, so that a deep learning model which is most suitable for the target station is obtained after training convergence.
The predetermined time period in this example is 1 to 7 days.
Inputting input vectors containing rainfall forecast data, urban ground elevation data and object attributes of a rain belt into a pre-trained deep learning model based on an improved convolutional neural network to obtain a ponding variability of a target site every 3 hours in 1-7 days in the future, and obtaining a ponding water level every 3 hours in 1-7 days in the future by combining with the current ponding water level.
Example 2
Based on the urban ponding water level prediction method based on the improved convolutional neural network described in embodiment 1, this embodiment provides an urban ponding water level prediction system based on the improved convolutional neural network, which includes:
a real-time data acquisition module: the system is used for acquiring the current ponding water level, urban ground elevation data and rainfall forecast data in the future preset time;
the rain belt object attribute extraction module: the system comprises a target station, a rainfall forecast database, a target station and a target server, wherein the target station is used for identifying a rainzone around the target station based on the rainfall forecast data in a future preset time and extracting object attributes of the rainzone;
an input variable construction module: the system is used for preprocessing rainfall forecast data, urban ground elevation data and object attributes of a rain belt to form input variables;
a ponding water level prediction module: and the method is used for inputting input variables into a pre-trained deep learning model based on the improved convolutional neural network to obtain the ponding variability of the target station within the future preset time, and obtaining the ponding water level within the future preset time by combining the current ponding water level.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.

Claims (6)

1. The urban ponding water level prediction method based on the improved convolutional neural network is characterized by comprising the following steps of:
acquiring the current ponding water level, urban ground elevation data and rainfall forecast data in the future preset time;
identifying a rainbelt around a target station based on rainfall forecast data in a future preset time, and extracting object attributes of the rainbelt;
preprocessing rainfall forecast data, urban ground elevation data and object attributes of a rain belt to form input variables;
inputting the input variable into a pre-trained deep learning model based on an improved convolutional neural network to obtain the ponding variability of the target station within the future preset time, and obtaining the ponding water level within the future preset time by combining the current ponding water level;
the object attributes of the rain belt comprise a centroid position, an area and precipitation intensity, and are extracted by the following method:
selecting the geometric center of a rain belt as a centroid position, representing the area through the number of grids in the rain belt range, and taking the 10 th percentile of rainfall in the rain belt range according to the rainfall intensity;
the deep learning model based on the improved convolutional neural network comprises a convolutional layer, a pooling layer, an expansion layer and a full-connection layer, the object attribute of the rain belt is added into the last full-connection layer, and the object attribute of the rain belt is added into the last full-connection layerTSScoring as a loss function of the deep learning model;
the training method of the deep learning model based on the improved convolutional neural network comprises the following steps:
acquiring historical ponding water level, urban ground elevation data and rainfall forecast data, standardizing the ponding water level, the urban ground elevation data and the rainfall forecast data, and putting the data into a constructed hydrological feature data set;
identifying a raindrop based on rainfall forecast data, extracting the object attribute of the raindrop, standardizing the raindrop, and forming a training set together with the hydrological feature data set;
and training the deep learning model by using the training set to obtain the trained deep learning model.
2. The improved convolutional neural network-based urban ponding water level prediction method of claim 1, wherein the current ponding water level, urban ground elevation data, and precipitation forecast data in a preset time in the future are obtained by the following methods:
the method comprises the steps of collecting the current ponding water level in real time through an urban ponding monitoring station, obtaining urban ground elevation data from a geographic space data cloud open platform, and obtaining rainfall forecast data in the future preset time through a data interface of a meteorological department.
3. The improved convolutional neural network-based urban ponding water level prediction method of claim 1, wherein the rain belt is identified by:
performing feature extraction on precipitation forecast data in a future preset time through a MODE spatial inspection technology so as to identify a rain belt;
firstly, spatial smoothing is carried out, and the calculation formula is as follows:
Figure DEST_PATH_IMAGE001
wherein the content of the first and second substances,
Figure DEST_PATH_IMAGE002
is a function of the filtering of the received signal,
Figure DEST_PATH_IMAGE003
is the original forecast field, and the forecast field,
Figure DEST_PATH_IMAGE004
is a water-falling field after filtration,
Figure DEST_PATH_IMAGE005
and
Figure DEST_PATH_IMAGE006
is the coordinate of a grid point and is,Ris the convolution radius;
then, carrying out precipitation threshold control, wherein the calculation formula is as follows:
Figure DEST_PATH_IMAGE007
wherein, the first and the second end of the pipe are connected with each other,
Figure DEST_PATH_IMAGE008
is the field of the mask or masks to be used,Tis the precipitation threshold;
finally, identifying the rain belt, wherein the calculation formula is as follows:
Figure DEST_PATH_IMAGE009
wherein the content of the first and second substances,
Figure DEST_PATH_IMAGE010
is the identified rain strip.
4. The method for predicting the urban ponding water level based on the improved convolutional neural network as claimed in claim 1, wherein the preprocessing comprises removing outliers and normalizing the data, and the formula of the normalization processing is as follows:
Figure DEST_PATH_IMAGE011
wherein, the first and the second end of the pipe are connected with each other,
Figure DEST_PATH_IMAGE012
is the result of the normalization and is,
Figure DEST_PATH_IMAGE013
is that
Figure DEST_PATH_IMAGE014
Is determined by the average value of (a),
Figure DEST_PATH_IMAGE015
is that
Figure 904324DEST_PATH_IMAGE014
The standard deviation of (a) is determined,
Figure 290306DEST_PATH_IMAGE014
is the value of the data.
5. The improved convolutional neural network-based urban ponding water level prediction method of claim 1, wherein the method is characterized in thatTSThe score is calculated as follows:
Figure DEST_PATH_IMAGE016
wherein the content of the first and second substances,
Figure DEST_PATH_IMAGE017
is the number of times that the forecast is correct,
Figure DEST_PATH_IMAGE019
the number of times of the empty report is,
Figure DEST_PATH_IMAGE020
is the number of false positives.
6. Urban ponding water level prediction system based on improve convolutional neural network, its characterized in that includes:
a real-time data acquisition module: the system is used for acquiring the current ponding water level, urban ground elevation data and rainfall forecast data within the future preset time;
the rain belt object attribute extraction module: the system comprises a target station, a rainfall forecast database, a target station and a target server, wherein the target station is used for identifying a rainzone around the target station based on the rainfall forecast data in a future preset time and extracting object attributes of the rainzone;
an input variable construction module: the system is used for preprocessing rainfall forecast data, urban ground elevation data and object attributes of a rain belt to form input variables;
the accumulated water level prediction module: and the method is used for inputting input variables into a pre-trained deep learning model based on the improved convolutional neural network to obtain the ponding variability of the target station within the future preset time, and obtaining the ponding water level within the future preset time by combining the current ponding water level.
CN202210815291.4A 2022-07-11 2022-07-11 Urban ponding water level prediction method and system based on improved convolutional neural network Active CN114881381B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210815291.4A CN114881381B (en) 2022-07-11 2022-07-11 Urban ponding water level prediction method and system based on improved convolutional neural network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210815291.4A CN114881381B (en) 2022-07-11 2022-07-11 Urban ponding water level prediction method and system based on improved convolutional neural network

Publications (2)

Publication Number Publication Date
CN114881381A CN114881381A (en) 2022-08-09
CN114881381B true CN114881381B (en) 2022-10-21

Family

ID=82683192

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210815291.4A Active CN114881381B (en) 2022-07-11 2022-07-11 Urban ponding water level prediction method and system based on improved convolutional neural network

Country Status (1)

Country Link
CN (1) CN114881381B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115860272B (en) * 2023-02-22 2023-06-30 山东捷讯通信技术有限公司 Reservoir multi-time-point intelligent water level prediction method and system based on deep learning
CN116702627B (en) * 2023-07-31 2024-01-30 华南理工大学 Urban storm waterlogging rapid simulation method based on deep convolutional neural network
CN117237677B (en) * 2023-11-15 2024-03-26 南京信息工程大学 Precipitation prediction correction method for overall similarity of strong precipitation space based on deep learning

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10025983B2 (en) * 2015-09-21 2018-07-17 The Climate Corporation Ponding water detection on satellite imagery
CN111650671B (en) * 2020-05-13 2021-10-22 苏翔 Rainfall forecast space inspection method based on multi-object optimal matching technology
CN111624680B (en) * 2020-05-13 2021-10-19 苏翔 Precipitation forecast displacement correction algorithm based on object diagnosis
CN113780668A (en) * 2021-09-15 2021-12-10 泰华智慧产业集团股份有限公司 Urban ponding waterlogging prediction method and system based on historical data

Also Published As

Publication number Publication date
CN114881381A (en) 2022-08-09

Similar Documents

Publication Publication Date Title
CN114881381B (en) Urban ponding water level prediction method and system based on improved convolutional neural network
CN109446992B (en) Remote sensing image building extraction method and system based on deep learning, storage medium and electronic equipment
CN107610469B (en) Day-dimension area traffic index prediction method considering multi-factor influence
JP2020526826A (en) Vessel automatic tracking methods and systems based on deep learning networks and average shifts
CN108564588B (en) Built-up area automatic extraction method based on depth features and graph segmentation method
CN111538798B (en) Urban catchment area refined extraction method considering DSM and DLG
CN111861027A (en) Urban traffic flow prediction method based on deep learning fusion model
CN109407177B (en) Machine learning and conventional meteorological observation-based fog identification system and application method
CN116645616A (en) Method for evaluating damage degree of building in natural disasters by using remote sensing images
CN116227362A (en) Municipal drainage pipe network state prediction method based on graph and deep learning
CN115859577A (en) Water bloom early warning method, device, medium and equipment based on weather forecast
CN115497015A (en) River floating pollutant identification method based on convolutional neural network
CN115272656A (en) Environment detection alarm method and device, computer equipment and storage medium
CN114067535A (en) Geological disaster early warning method and device based on artificial intelligence and storage medium
CN116703004B (en) Water system river basin intelligent patrol method and device based on pre-training model
CN110765900A (en) DSSD-based automatic illegal building detection method and system
CN115760866A (en) Crop drought detection method based on remote sensing image
CN110414027A (en) A kind of data processing method suitable for flood forecasting system
CN115330088A (en) Flood control fine prediction method for small reservoir
CN115482489A (en) Improved YOLOv 3-based power distribution room pedestrian detection and trajectory tracking method and system
CN115100592A (en) Method and device for identifying hidden danger of external damage of power transmission channel and storage medium
CN114067103A (en) Intelligent pipeline third party damage identification method based on YOLOv3
CN112711793A (en) Mountain cutting construction identification method and device
CN117078045B (en) Holographic inspection and monitoring method and system for traffic infrastructure
Zhang et al. Dense and low-rise residential areas detection by shadow data Mining in urban high-resolution images

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant