CN112016831A - AI intelligent forecast-based urban waterlogging landing area identification method - Google Patents
AI intelligent forecast-based urban waterlogging landing area identification method Download PDFInfo
- Publication number
- CN112016831A CN112016831A CN202010878210.6A CN202010878210A CN112016831A CN 112016831 A CN112016831 A CN 112016831A CN 202010878210 A CN202010878210 A CN 202010878210A CN 112016831 A CN112016831 A CN 112016831A
- Authority
- CN
- China
- Prior art keywords
- data
- rainfall
- waterlogging
- model
- minutes
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000004364 calculation method Methods 0.000 claims abstract description 21
- 238000007405 data analysis Methods 0.000 claims abstract description 16
- 238000012544 monitoring process Methods 0.000 claims abstract description 12
- 238000005070 sampling Methods 0.000 claims abstract description 5
- 238000013527 convolutional neural network Methods 0.000 claims description 22
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 21
- 238000012549 training Methods 0.000 claims description 17
- 238000004519 manufacturing process Methods 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 7
- 239000013598 vector Substances 0.000 claims description 6
- 238000009825 accumulation Methods 0.000 claims description 5
- 230000002265 prevention Effects 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 238000001704 evaporation Methods 0.000 claims description 3
- 230000008020 evaporation Effects 0.000 claims description 3
- 230000008595 infiltration Effects 0.000 claims description 3
- 238000001764 infiltration Methods 0.000 claims description 3
- 230000014759 maintenance of location Effects 0.000 claims description 3
- 230000006870 function Effects 0.000 description 16
- 238000012360 testing method Methods 0.000 description 9
- 238000004088 simulation Methods 0.000 description 7
- 238000011161 development Methods 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 5
- 238000011160 research Methods 0.000 description 5
- 238000009826 distribution Methods 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000001788 irregular Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000002123 temporal effect Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013142 basic testing Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000013277 forecasting method Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000001556 precipitation Methods 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- 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
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/05—Geographic models
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- General Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- Entrepreneurship & Innovation (AREA)
- Computational Linguistics (AREA)
- Databases & Information Systems (AREA)
- Development Economics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Geometry (AREA)
- Remote Sensing (AREA)
- Computing Systems (AREA)
- General Business, Economics & Management (AREA)
- Biomedical Technology (AREA)
- Tourism & Hospitality (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Educational Administration (AREA)
- Computer Graphics (AREA)
Abstract
The invention relates to the technical field of natural disaster prediction, in particular to an identification method of urban waterlogging landing areas based on AI intelligent prediction, which comprises the steps of establishing a data acquisition module, a data analysis module, a data modeling module and a data early warning module; the data acquisition module acquires data of a plurality of automatic rainfall stations, 10-year calendar history rainfall data of the monitored area, waterlogging area data of the monitored area and camera monitoring picture data, and the data acquisition sampling interval of the automatic rainfall stations is 5 minutes or 10 minutes or 20 minutes or 30 minutes or 1 hour; the data analysis module calculates and analyzes data acquired from the automatic rainfall station, rainfall prediction is carried out on a monitored area in the past month by combining rainfall data of the past year, and rainfall calculation is carried out by combining the data of the automatic rainfall station with rainfall in 10 minutes, 30 minutes, 1 hour, 3 hours and 6 hours. The method monitors and warns the whole monitored area according to real-time rainfall and displays the rainfall to workers through GIS real-time modeling.
Description
Technical Field
The invention relates to the technical field of natural disaster prediction, in particular to an identification method of urban waterlogging falling areas based on AI intelligent prediction.
Background
Urban rainfall flood becomes a normal state due to the urbanization development and climate change, the characteristics of frequent occurrence, wide occurrence, sudden turning and serious disasters are increasingly remarkable, and the urban rainfall flood problem becomes a difficult problem widely researched at present. From the weather, the rainstorm forecast is still a world problem, the fixed-point quantification cannot be carried out at regular time, the rain island effect mechanism generated along with the urbanization development is not clear, and the urban rainstorm forecast difficulty is increased. From the hydrology, the rainfall runoff is increased through the urbanization development, meanwhile, the earth surface confluence has the confluence characteristic of combining a slope surface and a pipeline, so that the confluence speed is accelerated, the peak flow is increased, the peak time is advanced, and the traditional hydrology model cannot accurately perform simulation calculation. The urban waterlogging water identification system based on video monitoring is developed, urban waterlogging is monitored in real time, data is fed back quantitatively, and research results are significant to development of work of urban flood control related departments and life and production of people.
In recent years, mode recognition and artificial intelligence and rapid development of image processing technology have attracted more and more attention to the way of water level recognition based on image processing, but some problems that the image reading is affected due to stain damage of a water gauge part or insufficient illumination intensity in rainy and foggy weather still remain to be solved. The overcomplete signal sparse representation theory based on image sparse representation is firstly proposed by MallatSG and the like, a Gabor dictionary is used, a matching pursuit algorithm is introduced, and signals are sparsely represented by a successive approximation method; in 2009, Wright j, et al propose a SRBC (Sparse Representation-Based Classification) classifier Based on Sparse Representation, which is successfully applied to face recognition Classification by linearly encoding or representing a small number of low-dimensional images of the same class for a test face image.
In 2006, with the improvement of a deep learning theory, especially the appearance of a layer-by-layer learning and parameter fine-tuning technology, a convolutional neural network is rapidly developed, and the falsification image identification technology is promoted to enter higher-level research. Belhassen Bayer and the like develop a new convolutional layer form for inhibiting the content of medical images and learning operation detection characteristics in a self-adaptive manner, and experiments prove that the method has a good identification effect; the research progress and application of the convolutional neural network in the aspect of image recognition are deeply analyzed by Chang Liang and the like, and the improvement of the application of the convolutional neural network on the image understanding effect is intuitively given; the convolutional neural network image recognition algorithm based on double optimization, which is proposed by Liu Wanjun et al, optimizes the recognition performance of the convolutional neural network, and provides a new idea for the development of the convolutional neural network with various structures. In the aspect of researching image identification by using CNN convolutional neural network, many mature research methods inspire some directions worth of continuous research. The neural network with the multi-channel structure, which is proposed by Zhang Wenda and the like, effectively reduces the influence of downsampling on feature extraction, but is lack of weight sharing and slightly complex in parameter design; yi Chaoren et al propose a multi-channel convolutional neural network identification method, randomly fuse 4 features in different gradient directions, and effectively reduce the identification error rate, but the model network structure is more complex, compared with a single-channel convolutional neural network, the training time is increased by nearly one time, and the error back propagation time is to be further shortened.
Disclosure of Invention
In view of the above, the present invention provides a method for identifying an urban waterlogging landing area based on AI intelligent prediction, so as to solve the problems in the background art.
The invention provides an AI (intelligent asset) forecast-based urban waterlogging landing area identification method, which comprises the steps of firstly establishing a data acquisition module, a data analysis module, a data modeling module and a data early warning module; the data acquisition module acquires data of a plurality of automatic rainfall stations, 10-year calendar historical rainfall data of a monitored area, waterlogging area data of the monitored area and camera monitoring picture data, and the data acquisition sampling interval of the automatic rainfall stations is 5 minutes or 10 minutes or 20 minutes or 30 minutes or 1 hour;
further, the data analysis module calculates and analyzes data acquired from the automatic rainfall station, rainfall prediction is carried out on a monitored area in a past month by combining rainfall data of a past year, different waterlogging risk level thresholds are calculated by combining the data of the automatic rainfall station with rainfall of 10 minutes, 30 minutes, 1 hour, 3 hours and 6 hours, and camera monitoring picture data are identified and calculated through a convolutional neural network.
And further, carrying out early warning and broadcasting of risk levels on the data calculated by the data analysis module through the data early warning module.
Further, the data modeling module carries out GIS modeling and presents the waterlogging landing area through the data calculated by the data analysis module.
Further, the calendar history rainfall data of the monitored area 10 comprises calendar day rainfall information, waterlogging point area information, waterlogging point ponding water depth information, waterlogging point ponding water area information and waterlogging point ponding time information.
Further, pictures covering different time and different water accumulation conditions within two years are acquired through a data acquisition module; firstly, carrying out type division on a picture through a convolutional neural network, representing by adopting a five-dimensional vector, adopting a softmax function as an activation function of the last layer, and adopting a cross-entropy function as a loss function; the softmax function is shown as equation (1), and the loss function is shown as equation (2):
at this time, the process of the present invention,
and is arbitrary
Are all non-negative;
wherein: n is 2 or 5, and m is the number of samples.
Further, only a few pictures are learned through one iteration of the convolutional neural network, the pictures are randomly selected from the database of each waterlogging level in the data acquisition module, and the probability of the selected pictures is consistent, so that the model uniformly learns the characteristics of each waterlogging level when the whole training is finished; the training precision is improved by adopting a dual-model form, firstly, on the basis of 5 classifications according to a data set, 2 classifications are carried out, data are classified into dry and wet categories, and whether accumulated water exists on the ground or not is distinguished. The learning rate of the convolutional neural network is: 0.001; the optimization function adopts SGD, the evaluation function adopts Cross EntropyLoss, and after 8 ten thousand iterations: the recognition accuracy rate on the training set reaches 93.5 percent; the identification accuracy rate on the test set reaches 85.3 percent; the accuracy of the training set is shown in table 1;
TABLE 1 analysis of image recognition accuracy on training data set
Iterating for 8 ten thousand times on a training set, wherein the model precision tends to be stable, and the accuracy rate of 0 predicted to be 0 is 92.2%; 1 the accuracy of 1 is predicted to be 93.3%; 2 the accuracy of 2 is predicted to be 96.3%; 3 the accuracy for a prediction of 3 is: 92.7 percent; 4 the accuracy of 4 is predicted to be 93.1%; the average predicted accuracy of finishing is 93.5%.
Test set accuracy as in Table 2
TABLE 2 analysis of image recognition accuracy on test data set
Further, the data sources for acquiring rainfall by the data acquisition module further comprise early warning information of the weather station, rainfall forecast data based on a radar-based QPF, hourly rainfall forecast of an intelligent fine grid point, a city geographic database and a disaster database.
The rainfall information monitored by the automatic rainfall stations is live rainfall data monitored by the automatic rainfall stations distributed in different regions. And calculating the cumulative rainfall of 10 minutes, 30 minutes, 1 hour, 3 hours, 6 hours and 24 hours according to the rainfall at different sampling intervals, and calculating the surface rainfall through spatial interpolation to be used by the waterlogging model. Rainfall information monitored by the radar based QPF rainfall forecast data Doppler radar is accumulated rainfall obtained by integrating detected radar echo intensity along time according to rainfall intensity inverted by a Z-R relation. The accumulated rainfall data of the future 10 minutes, 20 minutes, 30 minutes, 1 hour and 2 hours are obtained through a calculation method in radar rainfall forecast, the data are converted into rainfall estimation data with longitude and latitude as coordinates, the surface rainfall of each grid is converted, and the rainfall is substituted into an inland inundation model. The intelligent fine grid forecasts rainfall data, the forecasting time efficiency is 0-72 hours, and the rainfall data can be substituted into the waterlogging model.
Further, the data modeling module integrates an urban ground elevation model, a rainfall model, a production convergence model, a drainage model and a mathematical calculation model.
Further, the urban ground elevation model is established by adopting an interpolation method; the rainfall model estimates the area of the point which is represented by each station and is closest to the station in the flow field by assuming that the rainfall of each point on the flow field is represented by the rainfall of the closest rainfall station; the production and confluence model is used for simulating a confluence area and the submerging height of each point in the confluence area, and the production and confluence model is calculated by adopting a formula (3);
wherein Rs (t) is the ground runoff depth at time t, i is the rainfall intensity, in is the vegetation retention rate, e is the evaporation rate, sd is the hole filling rate, and f is the infiltration rate.
Further, the early warning module is connected with a traffic department, a municipal department, a flood prevention department and a meteorological department through a network, and the early warning module directly sends early warning information to the traffic department, the municipal department, the flood prevention department and the meteorological department when the early warning condition occurs.
The method for identifying the urban waterlogging landing area based on AI intelligent forecast has the advantages that: the rainfall station and weather station early warning information, rainfall forecast data based on radar base data, hourly rainfall forecast of Qinzi grid points, urban geographic databases and disaster databases are used for data acquisition, and then data analysis and GIS modeling and early warning are carried out through a data analysis module, so that the whole monitored area is monitored and early warned according to real-time rainfall, and is displayed to workers of all departments through GIS real-time modeling.
Drawings
FIG. 1 is a system diagram of the method of the present invention;
FIG. 2 is an overall architecture diagram of a GIS-based inland water disaster model of the present invention;
FIG. 3 is a water accumulation early warning model based on region meshing of the present invention;
fig. 4 is a schematic view of mesh generation according to the present invention.
Detailed Description
The present invention will be described in detail with reference to the drawings and specific embodiments, and it is to be understood that the described embodiments are only a few embodiments of the present invention, rather than the entire embodiments, and that all other embodiments obtained by those skilled in the art based on the embodiments in the present application without inventive work fall within the scope of the present application.
As shown in fig. 1, in the method for identifying an urban waterlogging landing area based on AI intelligent prediction, a data acquisition module, a data analysis module, a data modeling module and a data early warning module are firstly established; the data acquisition module acquires data of a plurality of automatic rainfall stations, 10-year calendar historical rainfall data of a monitored area, waterlogging area data of the monitored area and camera monitoring picture data, and the data acquisition sampling interval of the automatic rainfall stations is 5 minutes or 10 minutes or 20 minutes or 30 minutes or 1 hour;
in the embodiment, the data analysis module calculates and analyzes data acquired from the automatic rainfall station, rainfall prediction is carried out on a monitored area in a past month by combining rainfall data of a past year, different waterlogging risk level thresholds are calculated by combining the data of the automatic rainfall station with rainfall of 10 minutes, 30 minutes, 1 hour, 3 hours and 6 hours, and camera monitoring picture data are identified and calculated by a convolutional neural network;
in the embodiment, the data calculated by the data analysis module is subjected to early warning and broadcasting of risk levels through the data early warning module;
in this embodiment, the data modeling module performs GIS modeling and presents the data calculated by the data analysis module.
In this embodiment, the calendar history rainfall data of the monitored area 10 includes the calendar daily rainfall information, the waterlogging point area information, the waterlogging point ponding water depth information, the waterlogging point ponding area information, and the waterlogging point ponding time information.
In the embodiment, pictures covering different times and different water accumulation conditions within two years are acquired through the data acquisition module; firstly, the picture is classified by a convolutional neural network, five-dimensional vectors are used for representing, different grades are represented by classification variables and 5-dimensional vectors are used for representing, and the neural network constructed at the moment outputs one 5-dimensional vector. The first element of the 5-dimensional vector represents that the picture is of the second class waterlogging level, if the output result of a certain picture is (0.8, 0.1, 0.05, 0.05, 0), it is considered that 80% of the probability is 0, 10% of the probability is 1, 5% of the probability is 2, 5% of the probability is 3, and 0% of the probability is 4.
The activation function of the last layer adopts a softmax function, and the loss function adopts a cross-entropy function; the softmax function is shown as equation (1), and the loss function is shown as equation (2):
at this time, the process of the present invention,
and is arbitrary
Are all non-negative;
wherein: n is 2 or 5, and m is the number of samples.
In the embodiment, only a few pictures are learned through one iteration of the convolutional neural network, the pictures are randomly selected from the database of each waterlogging level in the data acquisition module, and the probability of the selected pictures is consistent, so that the model uniformly learns the characteristics of each waterlogging level when the whole training is finished; the training precision is improved by adopting a dual-model form, firstly, on the basis of 5 classifications according to a data set, 2 classifications are carried out, data are classified into dry and wet categories, and whether accumulated water exists on the ground or not is distinguished.
Taking the city of western ann as an example, the video information of a camera at 31 positions in the city is researched, the collected data covers pictures at different time and under different water accumulation conditions within two years, and the total number of the pictures is 15573. The classification of the collected picture information is shown in table 3:
TABLE 3 training data and test data set quantity statistics
| Statistics of quantity | Training set | Test set |
| 0 | 4410 | 1824 |
| 1 | 2717 | 1160 |
| 2 | 1814 | 786 |
| 3 | 1216 | 513 |
| 4 | 793 | 340 |
| Total of | 10950 | 4623 |
The learning rate of the convolutional neural network is: 0.001; the optimization function adopts SGD, the evaluation function adopts Cross EntropyLoss, and after 8 ten thousand iterations: the recognition accuracy rate on the training set reaches 93.5 percent; the identification accuracy rate on the test set reaches 85.3 percent; the accuracy of the training set is shown in table 1;
TABLE 1 analysis of image recognition accuracy on training data set
Iterating for 8 ten thousand times on a training set, wherein the model precision tends to be stable, and the accuracy rate of 0 predicted to be 0 is 92.2%; 1 the accuracy of 1 is predicted to be 93.3%; 2 the accuracy of 2 is predicted to be 96.3%; 3 the accuracy for a prediction of 3 is: 92.7 percent; 4 the accuracy of 4 is predicted to be 93.1%; the average predicted accuracy of finishing is 93.5%. Test set accuracy as in Table 2
TABLE 2 analysis of image recognition accuracy on test data set
In this embodiment, the data sources for the data acquisition module to acquire rainfall further include early warning information of the weather station, rainfall forecast data based on radar-based data, an hourly rainfall forecast of qinzhi grid points, a city geographic database, and a disaster database.
Carrying out waterlogging simulation calculation by using the data, wherein the waterlogging simulation calculation comprises surface rainfall calculation, grid setting, grid internal hydraulic flow dynamic model construction, generalization of urban terrain and ground objects and generalization of urban drainage system facilities;
the surface rainfall is calculated for various numerical forecasting products and rainfall output by different forecasting methods, and the monitoring rainfall obtained by automatic rainfall stations and radar monitoring is converted into a surface rainfall forecast value and a surface rainfall actual monitoring value of a non-structural irregular grid by adopting an effective interpolation method, so that the surface rainfall forecast value and the surface rainfall actual monitoring value are used as rainfall boundary conditions of an inland inundation system; as the surface rainfall can not be measured to be a true value, the method takes urban ponding process data of 10 times in Shaanxi province and typical rainstorm waterlogging instances as basic test data, surface rainfall values calculated by several surface rainfall calculation methods such as an arithmetic mean method, a Thiessen polygon method, an isochronal line method and the like are respectively used as rainfall boundary conditions of an waterlogging system, the rainfall boundary conditions are substituted into an waterlogging simulation model for simulation calculation, the absolute error between the ponding depth calculated by the model and the actual ponding depth is used as a judgment standard for judging the quality of the surface rainfall calculation method, and finally, the surface rainfall calculation method with relatively small absolute error is selected and applied to the waterlogging simulation subsystem.
The grid setting adopts unstructured irregular grid design calculation areas. The grid can be designed as a triangle, a quadrangle or a pentagon. Each side of the grid is defined as a channel, the normal direction of the grid can be any direction, each side of the grid is defined as a channel, and the joint between the channels is defined as a node. In the calculation process, the runoff quantity, the drainage quantity and the ponding depth in each grid are calculated by taking each grid as a unit.
The method is characterized in that a hydraulic flow mechanics model in a grid is constructed, a two-dimensional plane unsteady flow equation is adopted for calculating an area in the grid, the water flow motion of the urban earth surface is simulated and calculated through the two-dimensional plane unsteady flow equation, and a one-dimensional unsteady flow equation is adopted for each channel, so that the water flow motion in drainage channels and rivers in the city is simulated and calculated. Various dams in cities or places with higher terrain can be simulated and calculated by adopting a wide-top weir flow formula, and the model structure is shown as the lower graph of the figure and is shown as the figure 3.
The generalized urban drainage system facilities are provided with a drainage pump station, a gate, a temporary pump station and a grid (connected with a first-level river channel or a second-level river channel or a regional boundary) for submerging an outflow pipeline as a drainage outlet of a pipe network, and the generalized drainage capacity is given. The grid design of the waterlogging model considers different terrain and landform characteristics, gives consideration to channel attributes, and is processed into various grid types, dense grids in densely populated areas, areas where water is easy to accumulate in cities, dense grids in main functional areas and the like, and sparse grids in less populated areas, as shown in fig. 4.
As shown in fig. 2: in this embodiment, the data modeling module integrates an urban ground elevation model, a rainfall model, a production convergence model, a drainage model, and a mathematical computation model.
In the embodiment, the urban ground elevation model is established by adopting an interpolation method; the rainfall model estimates the area of the point which is represented by each station and is closest to the station in the flow field by assuming that the rainfall of each point on the flow field is represented by the rainfall of the closest rainfall station; the production and confluence model is used for simulating a confluence area and the submerging height of each point in the confluence area, and the production and confluence model is calculated by adopting a formula (3);
wherein Rs (t) is the ground runoff depth at time t, i is the rainfall intensity, in is the vegetation retention rate, e is the evaporation rate, sd is the hole filling rate, and f is the infiltration rate. First, all the depressions in the DEM data are found, all boundary points of all the depressions are recorded, the minimum value of the boundary points of all the depressions is found to serve as a notch point, and the inclusion relation among all the depressions is determined.
Constructing a waterlogging model structure, and using a ground elevation model (DEM) to simulate the height and the relief of the urban ground, wherein the model is the basis for calculating and simulating the waterlogging water of the urban; the rainfall model is used for simulating and calculating the spatial and temporal distribution of urban rainfall; the production convergence model is used for simulating the generation process and the convergence process of the ground runoff; the drainage model is used for calculating the urban drainage according to different urban drainage modes. The comprehensive calculation of the models provides a basis for calculating urban waterlogging. The mathematical computation model and the GIS space analysis application module are the 'operation center' of the whole urban waterlogging model. The mathematical calculation model provides key theories and formulas including precipitation calculation, comprehensive displacement calculation, residual rainfall calculation, ponding height calculation and the like.
Rainfall is the main driving variable in the model and is the most important people transportation data in the computation of the waterlogging simulation model. The rainfall of a city varies with the duration of the rainfall, and has non-uniformity in time. Meanwhile, the rainfall is different in size at each place in the range of the cage cover, and the nonuniformity of rainfall amount spatial distribution is shown. Therefore, in constructing a rainfall model of a city, the spatial and temporal distribution of rainfall must be considered. The urban inland inundation is mostly formed by rainstorm, so in order to reflect the change of rainfall time, the model utilizes hydrology knowledge to establish an empirical formula of rainstorm intensity according to rainfall observation data of the city for many years to calculate the average rainfall intensity in different time periods. And calculating the rainstorm intensity of 1, 2, 3, 5, 10 and other years in the recurrence period according to different durations to prepare a statistical table of the recurrence period T, the rainstorm intensity i and the rainfall duration T.
In this embodiment, in order to reflect the change in the rainfall space, the model records the actually measured rainfall at each rainfall station of a certain rain field on the topographic map of the whole city range, and draws the rainfall space distribution map by using a suitable algorithm. The surface rain approximation model parameter table is shown in table 4.
TABLE 4 generalized model parameter Table
Although the present invention has been described in detail with reference to the preferred embodiments, it will be understood by those skilled in the art that various changes may be made and equivalents may be substituted without departing from the spirit and scope of the invention as defined in the appended claims. The techniques, shapes, and configurations not described in detail in the present invention are all known techniques.
Claims (8)
1. An identification method of urban waterlogging landing areas based on AI intelligent forecast is specifically implemented according to the following steps:
s1: firstly, establishing a data acquisition module, a data analysis module, a data modeling module and a data early warning module;
s1.1: the data acquisition module acquires data of a plurality of automatic rainfall stations, 10-year calendar historical rainfall data of a monitored area, waterlogging area data of the monitored area and camera monitoring picture data, and the data acquisition sampling interval of the automatic rainfall stations is 5 minutes or 10 minutes or 20 minutes or 30 minutes or 1 hour;
s1.2: the data analysis module calculates and analyzes data acquired from the automatic rainfall station, rainfall prediction is carried out on a monitored area in a past month by combining rainfall data of a past year, different waterlogging risk level thresholds are calculated by combining the data of the automatic rainfall station with rainfall of 10 minutes, 30 minutes, 1 hour, 3 hours and 6 hours, and identification calculation is carried out on camera monitoring picture data through a convolutional neural network;
s2: the data calculated by the data analysis module is subjected to early warning and broadcasting of risk levels through the data early warning module;
s3: and the data modeling module carries out GIS modeling and presents the waterlogging falling area through the data calculated by the data analysis module.
2. The method for identifying urban waterlogging landing areas based on AI intelligent prediction as claimed in claim 1, wherein: the calendar history rainfall data of the monitored area 10 comprises calendar day rainfall information, waterlogging point area information, waterlogging point ponding water depth information, waterlogging point ponding water area information and waterlogging point ponding time information.
3. The method for identifying urban waterlogging landing areas based on AI intelligent prediction as claimed in claim 1, wherein: in step S1.2: acquiring pictures covering different time and different water accumulation conditions within two years through a data acquisition module;
s3.1: firstly, carrying out type division on a picture through a convolutional neural network, then representing the picture by adopting a five-dimensional vector, wherein the convolutional neural network learns that an activation function of the last layer adopts a softmax function, and a loss function adopts a cross-entry function; the softmax function is shown as equation (1), and the loss function is shown as equation (2):
at this time, the process of the present invention,
and is arbitrary
Are all non-negative;
wherein: n is 2 or 5, and m is the number of samples.
4. The method for identifying urban waterlogging landing areas based on AI intelligent prediction as claimed in claim 3, wherein: only a few pictures are learned through one iteration of the convolutional neural network, the pictures are randomly selected from the database of each waterlogging level in the data acquisition module, and the probability of each selected picture is consistent, so that the model uniformly learns the characteristics of each waterlogging level when the whole training is finished; the training precision is improved by adopting a dual-model form, firstly, on the basis of 5 classifications according to a data set, 2 classifications are carried out, data are classified into dry and wet categories, and whether accumulated water exists on the ground or not is distinguished.
5. The method for identifying urban waterlogging landing areas based on AI intelligent prediction as claimed in claim 1, wherein: the data acquisition module acquires rainfall data from early warning information of weather stations, rainfall forecast data based on a radar-based QPF (quench-Polish-Filter) module, hourly rainfall forecast of intelligent fine grid points, a city geographic database and a disaster database.
6. The method for identifying urban waterlogging landing areas based on AI intelligent prediction as claimed in claim 1, wherein: the data modeling module integrates an urban ground elevation model, a rainfall model, a production convergence model, a drainage model and a mathematical calculation model.
7. The method for identifying urban waterlogging landing areas based on AI intelligent prediction as claimed in claim 1, wherein: the urban ground elevation model is established by adopting an interpolation method; the rainfall model estimates the area of the point which is represented by each station and is closest to the station in the flow field by assuming that the rainfall of each point on the flow field is represented by the rainfall of the closest rainfall station; the production and confluence model is used for simulating a confluence area and the submerging height of each point in the confluence area, and the production and confluence model is calculated by adopting a formula (3);
wherein Rs (t) is the ground runoff depth at time t, i is the rainfall intensity, in is the vegetation retention rate, e is the evaporation rate, sd is the hole filling rate, and f is the infiltration rate.
8. The method for identifying urban waterlogging landing areas based on AI intelligent prediction as claimed in claim 1, wherein: the early warning module is connected with the traffic department, the municipal department, the flood prevention department and the meteorological department through a network, and the early warning information is directly sent to the traffic department, the municipal department, the flood prevention department and the meteorological department when the early warning condition occurs.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010878210.6A CN112016831A (en) | 2020-08-27 | 2020-08-27 | AI intelligent forecast-based urban waterlogging landing area identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010878210.6A CN112016831A (en) | 2020-08-27 | 2020-08-27 | AI intelligent forecast-based urban waterlogging landing area identification method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112016831A true CN112016831A (en) | 2020-12-01 |
Family
ID=73502300
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010878210.6A Pending CN112016831A (en) | 2020-08-27 | 2020-08-27 | AI intelligent forecast-based urban waterlogging landing area identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112016831A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112558190A (en) * | 2020-12-17 | 2021-03-26 | 水利部南京水利水文自动化研究所 | Water level and rainfall integrated monitoring system for urban rainstorm flood |
| CN112578479A (en) * | 2020-12-29 | 2021-03-30 | 北京维艾思气象信息科技有限公司 | Urban waterlogging warning device and method based on rainfall forecast |
| CN112785053A (en) * | 2021-01-15 | 2021-05-11 | 北京市水科学技术研究院 | Method and system for forecasting urban drainage basin flood |
| CN113344291A (en) * | 2021-06-29 | 2021-09-03 | 珠江水利委员会珠江水利科学研究院 | Urban inland inundation range forecasting method, device, medium and equipment |
| CN113343923A (en) * | 2021-07-01 | 2021-09-03 | 江苏舆图信息科技有限公司 | Real-time river drainage port drainage state identification method based on video images |
| CN113378668A (en) * | 2021-05-28 | 2021-09-10 | 上海商汤智能科技有限公司 | Method, device and equipment for determining accumulated water category and storage medium |
| CN113435049A (en) * | 2021-06-30 | 2021-09-24 | 国能大渡河大数据服务有限公司 | Rainfall equipment fault analysis system based on logistic regression |
| CN114858987A (en) * | 2022-03-30 | 2022-08-05 | 河海大学 | River and lake water quantity and quality monitoring and management system based on Internet of things |
| CN115661438A (en) * | 2022-10-21 | 2023-01-31 | 广州地铁集团有限公司 | Outdoor water level monitoring and flood prevention early warning method based on machine vision |
| CN117473791A (en) * | 2023-12-22 | 2024-01-30 | 水发科技信息(山东)有限公司 | Public data storage management system based on artificial intelligence |
| CN117746342A (en) * | 2024-02-19 | 2024-03-22 | 广州市突发事件预警信息发布中心(广州市气象探测数据中心) | Method for identifying road ponding by utilizing public video |
| CN117911750A (en) * | 2023-12-24 | 2024-04-19 | 振宁(无锡)智能科技有限公司 | Edge intelligent data processing method and edge server |
| CN118735312A (en) * | 2024-09-04 | 2024-10-01 | 无锡学院 | Urban flood early warning system and method based on artificial intelligence |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107832931A (en) * | 2017-10-31 | 2018-03-23 | 上海市政工程设计研究总院（集团）有限公司 | A kind of Modularity analysis method of plain river network region waterlogging risk |
| CN110956783A (en) * | 2019-12-04 | 2020-04-03 | 江河瑞通（北京）技术有限公司 | Urban waterlogging monitoring method and electronic equipment |
-
2020
- 2020-08-27 CN CN202010878210.6A patent/CN112016831A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107832931A (en) * | 2017-10-31 | 2018-03-23 | 上海市政工程设计研究总院（集团）有限公司 | A kind of Modularity analysis method of plain river network region waterlogging risk |
| CN110956783A (en) * | 2019-12-04 | 2020-04-03 | 江河瑞通（北京）技术有限公司 | Urban waterlogging monitoring method and electronic equipment |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112558190B (en) * | 2020-12-17 | 2021-10-22 | 水利部南京水利水文自动化研究所 | Water level and rainfall integrated monitoring system for urban rainstorm flood |
| CN112558190A (en) * | 2020-12-17 | 2021-03-26 | 水利部南京水利水文自动化研究所 | Water level and rainfall integrated monitoring system for urban rainstorm flood |
| CN112578479A (en) * | 2020-12-29 | 2021-03-30 | 北京维艾思气象信息科技有限公司 | Urban waterlogging warning device and method based on rainfall forecast |
| CN112785053B (en) * | 2021-01-15 | 2023-11-03 | 北京市水科学技术研究院 | Method and system for forecasting urban drainage basin flood |
| CN112785053A (en) * | 2021-01-15 | 2021-05-11 | 北京市水科学技术研究院 | Method and system for forecasting urban drainage basin flood |
| CN113378668A (en) * | 2021-05-28 | 2021-09-10 | 上海商汤智能科技有限公司 | Method, device and equipment for determining accumulated water category and storage medium |
| CN113344291A (en) * | 2021-06-29 | 2021-09-03 | 珠江水利委员会珠江水利科学研究院 | Urban inland inundation range forecasting method, device, medium and equipment |
| CN113344291B (en) * | 2021-06-29 | 2022-04-15 | 珠江水利委员会珠江水利科学研究院 | Urban inland inundation range forecasting method, device, medium and equipment |
| CN113435049A (en) * | 2021-06-30 | 2021-09-24 | 国能大渡河大数据服务有限公司 | Rainfall equipment fault analysis system based on logistic regression |
| CN113435049B (en) * | 2021-06-30 | 2022-05-03 | 国能大渡河大数据服务有限公司 | Rainfall equipment fault analysis system based on logistic regression |
| CN113343923A (en) * | 2021-07-01 | 2021-09-03 | 江苏舆图信息科技有限公司 | Real-time river drainage port drainage state identification method based on video images |
| CN114858987A (en) * | 2022-03-30 | 2022-08-05 | 河海大学 | River and lake water quantity and quality monitoring and management system based on Internet of things |
| CN115661438A (en) * | 2022-10-21 | 2023-01-31 | 广州地铁集团有限公司 | Outdoor water level monitoring and flood prevention early warning method based on machine vision |
| CN115661438B (en) * | 2022-10-21 | 2024-03-08 | 广州地铁集团有限公司 | Off-site water level monitoring and flood prevention early warning method based on machine vision |
| CN117473791A (en) * | 2023-12-22 | 2024-01-30 | 水发科技信息(山东)有限公司 | Public data storage management system based on artificial intelligence |
| CN117473791B (en) * | 2023-12-22 | 2024-03-29 | 水发科技信息(山东)有限公司 | Public data storage management system based on artificial intelligence |
| CN117911750A (en) * | 2023-12-24 | 2024-04-19 | 振宁(无锡)智能科技有限公司 | Edge intelligent data processing method and edge server |
| CN117746342A (en) * | 2024-02-19 | 2024-03-22 | 广州市突发事件预警信息发布中心(广州市气象探测数据中心) | Method for identifying road ponding by utilizing public video |
| CN117746342B (en) * | 2024-02-19 | 2024-05-17 | 广州市突发事件预警信息发布中心(广州市气象探测数据中心) | Method for identifying road ponding by utilizing public video |
| CN118735312A (en) * | 2024-09-04 | 2024-10-01 | 无锡学院 | Urban flood early warning system and method based on artificial intelligence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112016831A (en) | AI intelligent forecast-based urban waterlogging landing area identification method | |
| CN111047099B (en) | Regional torrential flood risk prediction method and system | |
| CN111507375B (en) | Urban waterlogging risk rapid assessment method and system | |
| CN111625993B (en) | Small watershed surface rainfall interpolation method based on mountainous terrain and rainfall characteristic prediction | |
| CN109448361B (en) | Resident traffic travel flow prediction system and prediction method thereof | |
| CN110646867A (en) | Urban drainage monitoring and early warning method and system | |
| Gourley et al. | Remote collection and analysis of witness reports on flash floods | |
| CN111144656A (en) | Disaster evaluation analysis method based on GIS | |
| CN118070619B (en) | Urban waterlogging model sensitive parameter identification optimization method by utilizing machine learning | |
| Zhai et al. | Simulating flash flood hydrographs and behavior metrics across China: Implications for flash flood management | |
| CN117494586B (en) | Mountain torrent space-time prediction method based on deep learning | |
| Karyotis et al. | Deep learning for flood forecasting and monitoring in urban environments | |
| CN113987810A (en) | Numerical forecasting method for rainstorm flood under drainage basin scale | |
| CN115270632A (en) | System and method for predicting water bloom risk of drinking water source | |
| CN116205136A (en) | Large-scale river basin deep learning flood forecasting method based on runoff lag information | |
| CN117875564A (en) | Surface radial flow source sink analysis method for urban three-dimensional landscape | |
| CN116110210B (en) | Data-driven landslide hazard auxiliary decision-making method in complex environment | |
| CN116758707A (en) | Geological disaster monitoring system and method based on big data | |
| Zanella et al. | Sensor networks, data processing, and inference: the hydrology challenge | |
| CN118090526B (en) | Soil micro-plastic migration intelligent monitoring method based on water and soil loss | |
| CN117852449B (en) | Fine ponding treatment method based on large-scale city rainstorm waterlogging model | |
| CN117094448B (en) | Big data analysis method and system | |
| CN114612801B (en) | Flood early warning method based on high-resolution remote sensing satellite geometric positioning model | |
| CN117196305A (en) | Meteorological hydrologic early warning method suitable for plain water network area | |
| CN117540878A (en) | Sea ice condition risk level prediction method and system based on artificial intelligence |
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 |