CN109886477A - Prediction technique, device and the electronic equipment of water pollution - Google Patents
Prediction technique, device and the electronic equipment of water pollution Download PDFInfo
- Publication number
- CN109886477A CN109886477A CN201910074973.2A CN201910074973A CN109886477A CN 109886477 A CN109886477 A CN 109886477A CN 201910074973 A CN201910074973 A CN 201910074973A CN 109886477 A CN109886477 A CN 109886477A
- Authority
- CN
- China
- Prior art keywords
- website
- data
- network model
- neural network
- affecting parameters
- 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.)
- Granted
Links
Classifications
-
- 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
- Y02A20/00—Water conservation; Efficient water supply; Efficient water use
- Y02A20/152—Water filtration
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the present application provides prediction technique, device and the electronic equipment of a kind of water pollution.This method comprises: obtaining the testing data of at least one the second website in the testing data and the corresponding presumptive area of the first website of the first website；The first affecting parameters and the second affecting parameters are obtained using the attention mechanism attention of the first convolution neural network model；The prediction result of the water quality data for the first website is obtained, prediction result is variation tendency of the water quality data of the first website in the following predetermined amount of time；The first affecting parameters, the second affecting parameters and prediction result are handled using default second convolution neural network model, the prediction result after being adjusted.The embodiment of the present application solves in the prior art because of the influence for not considering other flow information of water near measured point, the low problem of caused prediction result accuracy.
Description
Technical field
This application involves technical field of data processing, specifically, this application involves a kind of prediction techniques of water pollution, dress
It sets and electronic equipment.
Background technique
With the continuous development of China's economy, as product dirt caused by water body of some unsuitable mankind's environmental rights
Dye also increasingly increases.In recent years, water contamination accident frequently occurs in China, and fwaater resources protection is very urgent.Therefore, water pollution
Prediction is a groundwork for realizing water environment management planning and water prevention and cure of pollution.
According to the difference of various water pollution prediction technique mechanism, can substantially be divided into statistics prediction method, intelligent predicting method with
And mechanism model predicted method.With the research that deepens continuously of prediction technique, in recent years, every field all increasingly concerns are about general
Artificial intelligence prediction technique is predicted for large-scale data.
In current water pollution prediction technique, mainly using artificial neural network (ANN) and support vector machines (SVM) side
Method.However these methods all can only carry out water pollution prediction to the water flow of water flow to be detected, not consider that other water flows are treated
Detect the influence of water flow water pollution.
Summary of the invention
This application provides a kind of prediction technique of water pollution, device and electronic equipments, can solve above-mentioned technical problem.
The technical solution is as follows:
In a first aspect, this application provides a kind of prediction techniques of water pollution, this method comprises:
Obtain at least one the second website in the testing data and the corresponding presumptive area of the first website of the first website
Testing data, testing data include water quality time series data and hydrology timing information；
The testing data that will acquire the first website is input to preset first convolution neural network model and carries out feature extraction,
And handled using the attention mechanism attention of the first convolution neural network model, the first affecting parameters are obtained,
In, the first affecting parameters are to any hydrology in the water quality time series data of the first website and the hydrology timing information of the first website
Data are to the influence degree of the water quality data of the first website in the following predetermined amount of time, and by the testing data of each second website
It is input to the first convolution neural network model and carries out feature extraction, and utilize the attention mechanism of the first convolution neural network model
Attention combines predetermined each second website to handle the first website influence degree, obtains the second influence ginseng
Number, wherein the second affecting parameters are the water quality time series data of each second website and the hydrology timing letter of each second website
Influence degree of any hydrographic data to the water quality data of the first website in the following predetermined amount of time in breath；
The prediction result of the water quality data for the first website is obtained, prediction result is the water quality data of the first website not
Carry out the variation tendency in predetermined amount of time；
Using default second convolution neural network model to the first affecting parameters, the second affecting parameters and prediction result into
Row processing, the prediction result after being adjusted.
Further, the convolution mode that the first convolution neural network model uses is empty convolution.
Further, using default second convolution neural network model to the first affecting parameters, second affecting parameters
And the prediction result is handled, the prediction result after being adjusted, comprising:
Using the attention mechanism attention of default second convolution neural network model to the first affecting parameters, second
Affecting parameters and prediction result are coupled；
The result of coupling is handled using the CNN module of the second convolution neural network model, it is pre- after being adjusted
Survey result, wherein the second convolution neural network model uses convolution mode for cause and effect convolution.
Further, obtain in the testing data and the corresponding presumptive area of the first website of the first website at least one the
After the testing data of two websites, method further includes；
Detect whether any data is not small in the testing data of the first website of current point in time and/or any second website
In corresponding standard parameter；
If being not less than corresponding standard parameter, generates warning information and determine corresponding website.
Further, the history testing data of the first website is acquired by the detection device for being previously deployed at the first website
It arrives, the history testing data of each second website is collected by the detection device for being previously deployed at each second website, method
Further include:
Each second website is recalculated to the first website influence degree, and with predetermined each second website to
One website influence degree is compared, and judges whether to change；
If each second website changes to the first website influence degree, the detection dress for being deployed in the first website is determined
It sets and is abnormal；
If each second website to the first website influence degree and it is not all change, determine changed second station
The detection device of point is abnormal.
Further, each second website is calculated to the first website influence degree using the first convolution neural network model.
Further, hydrographic information includes at least one of the following: water temperature, water level, flow velocity.
Second aspect, this application provides a kind of prediction meanss of water pollution, which includes:
Data acquisition module, in the testing data and the corresponding presumptive area of the first website for obtaining the first website extremely
The testing data of few second website, testing data includes water quality time series data and hydrology timing information；
Determining module is influenced, the testing data for will acquire the first website is input to preset first convolutional neural networks
Model carries out feature extraction, and is handled using the attention mechanism attention of the first convolution neural network model, obtains
First affecting parameters, wherein when the first affecting parameters are the hydrology to the water quality time series data of the first website and the first website
Any hydrographic data is to the influence degree of the water quality data of the first website in the following predetermined amount of time in sequence information, and by each
The testing data of two websites is input to the first convolution neural network model and carries out feature extraction, and utilizes the first convolutional neural networks
The attention mechanism attention of model combine predetermined each second website to the first website influence degree at
Reason, obtains the second affecting parameters, wherein the second affecting parameters are the water quality time series data and each second of each second website
Influence journey of any hydrographic data to the water quality data of the first website in the following predetermined amount of time in the hydrology timing information of website
Degree；
Predicted value obtains module, for obtaining the prediction result of the water quality data for the first website, prediction result the
Variation tendency of the water quality data of one website in the following predetermined amount of time；
Processing module is predicted, for influencing using default second convolution neural network model on the first affecting parameters, second
Parameter and prediction result are handled, the prediction result after being adjusted.
Further, the convolution mode that the first convolution neural network model uses is empty convolution.
Further, processing module is predicted, for the attention mechanism using default second convolution neural network model
Attention couples the first affecting parameters, the second affecting parameters and prediction result；Utilize the second convolutional neural networks
The CNN module of model handles the result of coupling, the prediction result after being adjusted, wherein the second convolutional neural networks
Model uses convolution mode for cause and effect convolution.
Further, data acquisition module obtains the testing data and the corresponding presumptive area of the first website of the first website
It after the testing data of at least one interior the second website, is also used to: detecting the first website of current point in time and/or any second
Whether any data is not less than corresponding standard parameter in the testing data of website；If being not less than corresponding standard parameter, generate
Warning information simultaneously determines corresponding website.
Further, the testing data of the first website is collected by the detection device for being previously deployed at the first website, respectively
The testing data of a second website is collected by the detection device for being previously deployed at each second website, device further include:
Recalculate module, for recalculating each second website to the first website influence degree, and with predefine
Each second website the first website influence degree is compared, judge whether to change；
First processing module determines deployment if changing for each second website to the first website influence degree
It is abnormal in the detection device of the first website；
Second processing module, if for each second website to the first website influence degree and it is not all change, really
The detection device of fixed changed second website is abnormal.
Further, each second website is calculated to the first website influence degree using the first convolution neural network model.
Further, hydrographic information includes at least one of the following: water temperature, water level, flow velocity.
Further, the correlation of the first website and any second website passes through Pearson correlation coefficient and/or linear phase
Relationship number is determined.
The third aspect, this application provides a kind of electronic equipment, which includes:
One or more processors；
Memory；
One or more application program, wherein one or more application programs be stored in memory and be configured as by
One or more processors execute, and one or more programs are configured to: executing the prediction technique of above-mentioned water pollution.
Fourth aspect, this application provides a kind of computer readable storage mediums, are stored thereon with computer program, the journey
The prediction technique of above-mentioned water pollution is realized when sequence is executed by processor.
Technical solution provided by the embodiments of the present application has the benefit that using the first convolution neural network model not
The feature extraction for only realizing the testing data to the first website also achieves the feature extraction of the testing data to the second website,
And shadow can be completed from time dimension, Spatial Dimension by the attention mechanism attention of the first convolution neural network model
The extraction for ringing power factor, obtains the first affecting parameters and the second affecting parameters, not only realize determine the influence of tested website because
Element also achieves the purpose for determining the influence degree of tested other websites of website periphery and corresponding influence factor, as to water
The reference data that pollution prediction result is adjusted, thus improve water pollution prediction precision, solve in the prior art because
The influence of water pollution prediction of the periphery website to tested website, the problem of caused precision of prediction difference can not be examined.
Detailed description of the invention
In order to more clearly explain the technical solutions in the embodiments of the present application, institute in being described below to the embodiment of the present application
Attached drawing to be used is needed to be briefly described.
Fig. 1 is a kind of flow diagram of the prediction technique of water pollution provided by the embodiments of the present application；
Fig. 2 is the process signal that prediction result is adjusted in a kind of prediction technique of water pollution provided by the embodiments of the present application
Figure；
Fig. 3 is a kind of structural schematic diagram of the prediction meanss of water pollution provided by the embodiments of the present application；
Fig. 4 is the structural schematic diagram of the prediction meanss of another water pollution provided by the embodiments of the present application；
Fig. 5 is the structural schematic diagram of a kind of electronic equipment provided by the embodiments of the present application.
Specific embodiment
Embodiments herein is described below in detail, examples of the embodiments are shown in the accompanying drawings, wherein from beginning to end
Same or similar label indicates same or similar element or element with the same or similar functions.Below with reference to attached
The embodiment of figure description is exemplary, and is only used for explaining the application, and cannot be construed to the limitation to the application.
Those skilled in the art of the present technique are appreciated that unless expressly stated, singular " one " used herein, " one
It is a ", " described " and "the" may also comprise plural form.It is to be further understood that being arranged used in the description of the present application
Diction " comprising " refer to that there are the feature, integer, step, operation, element and/or component, but it is not excluded that in the presence of or addition
Other one or more features, integer, step, operation, element, component and/or their group.It should be understood that when we claim member
Part is " connected " or when " coupled " to another element, it can be directly connected or coupled to other elements, or there may also be
Intermediary element.In addition, " connection " used herein or " coupling " may include being wirelessly connected or wirelessly coupling.It is used herein to arrange
Diction "and/or" includes one or more associated wholes for listing item or any cell and all combinations.
To keep the purposes, technical schemes and advantages of the application clearer, below in conjunction with attached drawing to the application embodiment party
Formula is described in further detail.
Existing water pollution prediction is generally all predicted just for the data in current waters, and can not consider surrounding body
Influence.Current waters may cause to pollute because of the inflow in other waters, therefore prediction result is inaccurate.
Prediction technique, device, electronic equipment and the computer readable storage medium of water pollution provided by the present application, it is intended to solve
The certainly technical problem as above of the prior art.
How the technical solution of the application and the technical solution of the application are solved with specifically embodiment below above-mentioned
Technical problem is described in detail.These specific embodiments can be combined with each other below, for the same or similar concept
Or process may repeat no more in certain embodiments.Below in conjunction with attached drawing, embodiments herein is described.
Embodiment one
The embodiment of the present application provides a kind of prediction technique of water pollution, as shown in Figure 1, this method comprises: step S101
To step S105.
Step S101, obtain in the testing data and the corresponding presumptive area of the first website of the first website at least one the
The testing data of two websites, testing data include water quality time series data and hydrology timing information.
In the embodiment of the present application, water quality time series data is used to characterize the Water quality situation in current waters, when practical application,
Different numerical value can be set for water quality data, and according to different ranks is arranged for different numerical intervals.For example, if water quality
When the corresponding numerical value of PH is 10, corresponding rank is VI grades, can not be drunk.
In the embodiment of the present application, hydrographic information is used to characterize the water position status in current waters, such as water temperature, water level, flow, drop
The information such as water.
In the embodiment of the present application, deployment water quality detecting device acquisition water quality data in advance, hydrology detection device can use
Acquire water level information.When practical application, water quality detecting device, hydrology detection device both can use wireless communication technique and be uploaded to
Electronic equipment (can be server)；Or connect water quality detecting device, hydrology detection device with data collection station, so as to
Testing data is obtained by data collection station, specifically, data collection station can have number for mobile phone, Pad, notebook etc.
According to the electronic equipment of store function.
When practical application, testing data can also be carried out by preset water quality information database, hydrographic information database
It obtains；Or the data of third-party platform (database of such as Bureau of Water Resources) are crawled using crawler technology, to obtain the first website
Water quality data and hydrographic information.
In the embodiment of the present application, the second website is the monitoring point near the first website.For example, it is attached that the first website can be set
Monitoring point in nearly 10 kilometer range is the second website.In the embodiment of the present application, the first website is to be predicted website, second station
Point is other websites being predicted near website.
Step S102, the testing data for the first website that will acquire is input to preset first convolution neural network model
Feature extraction is carried out, and is handled using the attention mechanism attention of the first convolution neural network model, obtains first
Affecting parameters, wherein the first affecting parameters are to believe the water quality time series data of the first website and the hydrology timing of the first website
Any hydrographic data is to the influence degree of the water quality data of the first website in the following predetermined amount of time in breath, and by each second station
The testing data of point is input to the first convolution neural network model and carries out feature extraction, and utilizes the first convolution neural network model
Attention mechanism attention combine predetermined each second website the first website influence degree is handled, obtain
To the second affecting parameters, wherein the second affecting parameters are the water quality time series data and each second website of each second website
Hydrology timing information in any hydrographic data to the influence degree of the water quality data of the first website in the following predetermined amount of time.
In the embodiment of the present application, the first affecting parameters are for characterizing on the influential factor of the water quality of the first website tool, such as
It may be water temperature, water temperature is higher, and water quality is poorer.
In the embodiment of the present application, the second affecting parameters are influential on the water quality of the first website in the second website for characterizing
Factor may be water quality, it is also possible to be hydrographic information.For example, if the second website is the upriver of the first website, then if the
Two website water bodys are contaminated, then will affect the water quality data of the first website.
In the embodiment of the present application, the first convolution neural network model can not only realize the testing data to the first website,
The feature extraction of the testing data of second website, and realized using attention mechanism and determine the first affecting parameters, second
The purpose of affecting parameters, so as to comprehensively consider the first website, influence factor and influence degree in the second website.Pass through
Attention mechanism not only can reflect the influence power of different websites, can also be specific to the main influence feature of the website, such as
The variation of B website (i.e. the second website) water temperature influences greatly the first website water quality data, and the flow velocity of C website is to the water of the first website
Prime number is bigger according to power is rung.For example, B website water temperature is 20% to the influence accounting of the first website water quality data, and the water flow of C website
Flow velocity be 30% to the influence accounting of the first website water quality data.
In the embodiment of the present application, each second website between the first website influence degree may with two websites at a distance from have
It closes.For example, if the distance of two websites is closer, then it is assumed that the influence degree of website is higher, therefore, the first website and the second website
Between the correlation of short distance and the two be inversely proportional, i.e., distance is remoter, influences smaller.When practical application, because will only be led with distance
Cause determining each second website to the first website influence degree inaccuracy problem.For example, if the geographical location ten of two websites
Split-phase is close, but is separated by high mountain, then its correlation will be very low.Therefore, when practical application, can consider simultaneously by environment because
Element.
The embodiment of the present application plays the purpose of auxiliary prediction by the acquisition of the testing data of the second website, to improve pre-
Precision is surveyed, the influence because not considering neighbouring other waters in tested waters is avoided, caused precision of prediction is poor, or even asking unsuccessfully
Topic.
In the embodiment of the present application, predetermined each second website can both pass through Pierre to the first website influence degree
Inferior related coefficient is determined, and be can use linearly dependent coefficient and be determined.
Step S103, the prediction result of the water quality data for the first website is obtained, prediction result is the water of the first website
Prime number is according to the variation tendency in the following predetermined amount of time.
When practical application, it can use arbitrary model and the water quality data of the first website predicted, specifically, Ke Yiying
With numerical model, CNN model, RNN model etc., the application is to this without limiting.
Step S104, using default second convolution neural network model to the first affecting parameters, the second affecting parameters and
Prediction result is handled, the prediction result after being adjusted.
The embodiment of the present application not only realizes the spy of the testing data to the first website using the first convolution neural network model
Sign is extracted, and also achieves the feature extraction of the testing data to the second website, and can pass through the first convolution neural network model
Attention mechanism attention from time dimension, Spatial Dimension complete influence power factor extraction, obtain the first affecting parameters
With the second affecting parameters, not only realize the influence factor for determining tested website, also achieve determine tested website periphery other
The purpose of the influence degree of website and corresponding influence factor, as the reference data being adjusted to water pollution prediction result,
To improve the precision of water pollution prediction, solve in the prior art because periphery website can not be examined to the water pollution of tested website
The influence of prediction, the problem of caused precision of prediction difference.
In one implementation, the convolution mode that the first convolution neural network model uses is empty convolution, to increase
It is input to the first convolution neural network model data volume.
The embodiment of the present application has bigger receptive field using the first convolution neural network model of empty convolution, can be with
The historical information for extracting the time earlier, the time series data for having not only acted as control a longer period of time are input to the first convolutional Neural
The effect that network model is handled, and play the role of increasing the data processing amount of the first convolution neural network model.
In the embodiment of the present application, there is bigger receptive field using empty convolution mode the first convolution neural network model,
The time series data for having not only acted as control a longer period of time is input to the effect that the first convolution neural network model is handled, and increases
The big data processing amount of first convolution neural network model；Meanwhile utilizing the CNN module of the first convolution neural network model
Structure feature ensure that the operation that entire prediction process can be parallel on multiple CPU, can effectively control time cost,
Preferably reflect the fluctuation characteristic of curve.
In another implementation, as shown in Fig. 2, utilizing default second convolution neural network model pair in step S104
First affecting parameters, the second affecting parameters and prediction result are handled, comprising: step S1041 (not shown) and step
Rapid S1042 (not shown).
Step S1041, it is influenced using the attention mechanism attention of default second convolution neural network model on first
Parameter, the second affecting parameters and prediction result are coupled；
Step S1042, the result of coupling is handled using the CNN module of the second convolution neural network model, is obtained
Prediction result adjusted, wherein the second convolution neural network model uses convolution mode for cause and effect convolution.
In the embodiment of the present application, the second convolution neural network model uses attention mechanism, the prediction that will tentatively obtain
As a result, the first affecting parameters, the second affecting parameters carry out concate (merging), to be input to convolutional neural networks model
Neural network input layer, to complete feature extraction, the final neural network hidden layer for obtaining the second convolution neural network model
Output as a result, the output result has comprehensively considered the influence power factor taken out from time dimension, Spatial Dimension, solve
Not the technical issues of not considering the influence of ambient stations in the prior art, to improve the purpose of the precision of prediction of subsequent water pollution.
In the embodiment of the present application, by the prediction result of the water quality data for the first website, play to the first website
Water quality preliminary prediction, and the first affecting parameters of the second convolution neural network model then comprehensively considered and second influence
Parameter is realized to the correcting action of the result of tentative prediction, and the work for comprehensively considering the influence factor in periphery website waters is played
With, improve water pollution prediction accuracy.
In another implementation, testing data and the first website that the first website is obtained in step S101 are corresponding
In presumptive area after the testing data of at least one the second website, this method further include:
Detect whether any data is not small in the testing data of the first website of current point in time and/or any second website
In corresponding standard parameter；
If being not less than corresponding standard parameter, generates warning information and determine corresponding website.
The application by different quality data and corresponding predetermined value and/or difference hydrographic data with it is corresponding and predetermined
The comparison of numerical value plays the purpose of preliminary early warning, to carry out emergency processing work.
For example, if the height of water level got is 60cm, easily endangering when the corresponding predetermined value of height of water level is 50cm
Danger is dykes and dams if region locating for water level, it may be necessary to flood discharge processing carried out, maintenance personnel is notified by warning information, so as to
Maintenance personnel is handled.
In another implementation, the history testing data of the first website is filled by the detection for being previously deployed at the first website
It sets and collects, the history testing data of each second website is acquired by the detection device for being previously deployed at each second website
It arrives, using default second convolution neural network model to the first affecting parameters, the second affecting parameters and prediction in step S104
As a result before being handled, this method further include: step S1041 (not shown), step S1042 (not shown) and step
Rapid S1043 (not shown).
Step S1041, each second website is recalculated to the first website influence degree, and with predetermined each
Two websites are compared the first website influence degree, judge whether to change:
If step S1042, each second website changes to the first website influence degree, determination is deployed in first stop
The detection device of point is abnormal；
If step S1043, each second website to the first website influence degree and it is not all change, determination become
The detection device for the second website changed is abnormal.
When practical application, the correlation of the first website and the second website is usually constant, it is assumed that the second website is respectively A
Website, B website and C website can determine if the correlation of the first website and A website, B website and C website changes
One website is abnormal；It is only the first care hair with C website if the first website does not change with the correlation of A website and B website
Changing both can only consider that A website and B website predicted target contaminant, can also be in the exception for excluding C website
Afterwards, the testing data of C website is reacquired, comprehensively considers all second websites realizations to the prediction mesh of target contaminant to realize
's.
Embodiment two
The embodiment of the present application provides a kind of prediction meanss of water pollution, as shown in figure 3, the prediction meanss 30 of the water pollution
It may include: data acquisition module 301, influence determining module 302, predicted value obtains module 303 and predict processing module
304, wherein
Data acquisition module 301, for obtaining the testing data and the corresponding presumptive area of the first website of the first website
The testing data of at least one interior the second website, testing data include water quality time series data and hydrology timing information；
Determining module 302 is influenced, the testing data for will acquire the first website is input to preset first convolutional Neural
Network model carries out feature extraction, and is handled using the attention mechanism attention of the first convolution neural network model,
Obtain the first affecting parameters, wherein first affecting parameters are the water quality time series data and the first website to the first website
Hydrology timing information in any hydrographic data to the influence degree of the water quality data of the first website in the following predetermined amount of time, and
The testing data of each second website is input to the first convolution neural network model and carries out feature extraction, and utilizes the first convolution
The attention mechanism attention of neural network model combines predetermined each second website to the first website influence degree
It is handled, obtains the second affecting parameters, wherein the second affecting parameters are the water quality time series data of each second website and each
Any hydrographic data is to the water quality data of the first website in the following predetermined amount of time in the hydrology timing information of a second website
Influence degree；
Predicted value obtains module 303, and for obtaining the prediction result for being directed to the water quality data of the first website, prediction result is
Variation tendency of the water quality data of first website in the following predetermined amount of time；
Processing module 304 is predicted, for utilizing default second convolution neural network model to the first affecting parameters, the second shadow
It rings parameter and prediction result is handled, the prediction result after being adjusted.
In the embodiment of the present application, not only realized using the first convolution neural network model to the testing data of the first website
Feature extraction also achieves the feature extraction of the testing data to the second website, and can pass through the first convolutional neural networks mould
The attention mechanism attention of type completes the extraction of influence power factor from time dimension, Spatial Dimension, obtains the first influence ginseng
Several and the second affecting parameters not only realize the influence factor for determining tested website, also achieve determine tested website periphery its
The purpose of the influence degree of his website and corresponding influence factor, as the reference number being adjusted to water pollution prediction result
According to, thus improve water pollution prediction precision, solve in the prior art because periphery website can not be examined to the water of tested website
The influence of pollution prediction, the problem of caused precision of prediction difference.
Further, information extraction modules 302 are used for: correlation, first stop by the first website with any second website
The water quality data of point and the water quality data and hydrographic information of hydrographic information and any second website are input to default first mould
Type obtains the first key message of the first website and the second key message of any second website.
Further, the convolution mode that the first convolution neural network model uses is empty convolution.
Further, prediction processing module 304 is used for: utilizing the attention mechanism of default second convolution neural network model
Attention couples the first affecting parameters, the second affecting parameters and prediction result；
The result of coupling is handled using the CNN module of the second convolution neural network model, it is pre- after being adjusted
Survey result, wherein the second convolution neural network model uses convolution mode for cause and effect convolution.
Further, the testing data of first website of the acquisition of data acquisition module 301 and the first website are corresponding predetermined
In region after the testing data of at least one the second website, it is also used to；Detect the first website described in current point in time and/or
Whether any data is not less than corresponding standard parameter in the testing data of any second website；If joining not less than corresponding standard
Number generates warning information and determines corresponding website.
Further, as shown in figure 4, the device 30 further include:
Recalculate module 305, for recalculating each second website to the first website influence degree, and in advance really
Fixed each second website is compared the first website influence degree, judges whether to change；
First processing module 306, if changing for each second website to the first website influence degree, determining section
The detection device affixed one's name in the first website is abnormal；
Second processing module 307, if for each second website to the first website influence degree and it is not all change,
Determine that the detection device of changed second website is abnormal.
Further, hydrographic information includes at least one of the following: water temperature, water level, flow velocity.
Further, the correlation of the first website and any second website passes through Pearson correlation coefficient and/or linear phase
Relationship number is determined.
The prediction technique for the water pollution that the embodiment of the present application one provides can be performed in the prediction meanss of the water pollution of the present embodiment,
Its realization principle is similar, and details are not described herein again.
Embodiment three
The embodiment of the present application provides a kind of electronic equipment, as shown in figure 5, electronic equipment shown in fig. 5 400 includes: place
Manage device 4001 and memory 4003.Wherein, processor 4001 is connected with memory 4003, is such as connected by bus 4002.Into one
Step ground, electronic equipment 400 can also include transceiver 4004.It should be noted that transceiver 4004 is not limited in practical application
One, the structure of the electronic equipment 400 does not constitute the restriction to the embodiment of the present application.
Processor 4001 can be CPU, general processor, DSP, ASIC, FPGA or other programmable logic device, crystalline substance
Body pipe logical device, hardware component or any combination thereof.It, which may be implemented or executes, combines described by present disclosure
Various illustrative logic blocks, module and circuit.Processor 4001 is also possible to realize the combination of computing function, such as wraps
It is combined containing one or more microprocessors, DSP and the combination of microprocessor etc..
Bus 4002 may include an access, and information is transmitted between said modules.Bus 4002 can be pci bus or
Eisa bus etc..Bus 4002 can be divided into address bus, data/address bus, control bus etc..Only to be used in Fig. 5 convenient for indicating
One thick line indicates, it is not intended that an only bus or a type of bus.
Memory 4003 can be ROM or can store the other kinds of static storage device of static information and instruction, RAM
Or the other kinds of dynamic memory of information and instruction can be stored, it is also possible to EEPROM, CD-ROM or other CDs
Storage, optical disc storage (including compression optical disc, laser disc, optical disc, Digital Versatile Disc, Blu-ray Disc etc.), magnetic disk storage medium
Or other magnetic storage apparatus or can be used in carry or store have instruction or data structure form desired program generation
Code and can by any other medium of computer access, but not limited to this.
Memory 4003 is used to store the application code for executing application scheme, and is held by processor 4001 to control
Row.Processor 4001 is for executing the application code stored in memory 4003, to realize that Fig. 3 or 4 illustrated embodiments are mentioned
The movement of the prediction meanss of the water pollution of confession.
The embodiment of the present application not only realizes the spy of the testing data to the first website using the first convolution neural network model
Sign is extracted, and also achieves the feature extraction of the testing data to the second website, and can pass through the first convolution neural network model
Attention mechanism attention from time dimension, Spatial Dimension complete influence power factor extraction, obtain the first affecting parameters
With the second affecting parameters, not only realize the influence factor for determining tested website, also achieve determine tested website periphery other
The purpose of the influence degree of website and corresponding influence factor, as the reference data being adjusted to water pollution prediction result,
To improve the precision of water pollution prediction, solve in the prior art because periphery website can not be examined to the water pollution of tested website
The influence of prediction, the problem of caused precision of prediction difference.
The embodiment of the present application provides a kind of computer readable storage medium, is stored on the computer readable storage medium
Computer program realizes method shown in embodiment one to three any embodiment of embodiment when the program is executed by processor.
The embodiment of the present application provides a kind of computer readable storage medium, compared with prior art, utilizes the first convolution
Neural network model not only realizes the feature extraction of the testing data to the first website, also achieves the number to be measured to the second website
According to feature extraction, and can be by the attention mechanism attention of the first convolution neural network model from time dimension, sky
Between dimension complete the extraction of influence power factor, obtain the first affecting parameters and the second affecting parameters, not only realize determine it is tested
The influence factor of website also achieves the mesh for determining the influence degree of tested other websites of website periphery and corresponding influence factor
, as the reference data being adjusted to water pollution prediction result, to improve the precision of water pollution prediction, solve existing
Have in technology because of the influence that can not examine water pollution prediction of the periphery website to tested website, the problem of caused precision of prediction difference.
The embodiment of the present application provides a kind of computer readable storage medium and is suitable for above method embodiment.Herein no longer
It repeats.
It should be understood that although each step in the flow chart of attached drawing is successively shown according to the instruction of arrow,
These steps are not that the inevitable sequence according to arrow instruction successively executes.Unless expressly stating otherwise herein, these steps
Execution there is no stringent sequences to limit, can execute in the other order.Moreover, at least one in the flow chart of attached drawing
Part steps may include that perhaps these sub-steps of multiple stages or stage are not necessarily in synchronization to multiple sub-steps
Completion is executed, but can be executed at different times, execution sequence, which is also not necessarily, successively to be carried out, but can be with other
At least part of the sub-step or stage of step or other steps executes in turn or alternately.
The above is only some embodiments of the application, it is noted that those skilled in the art are come
It says, under the premise of not departing from the application principle, several improvements and modifications can also be made, these improvements and modifications also should be regarded as
The protection scope of the application.
Claims (10)
1. a kind of prediction technique of water pollution characterized by comprising
Obtain at least one the second website in the testing data and the corresponding presumptive area of first website of the first website
Testing data, the testing data include water quality time series data and hydrology timing information；
The testing data of first website got is input to preset first convolution neural network model and carries out feature
It extracts, and is handled using the attention mechanism attention of the first convolution neural network model, obtain the first influence
Parameter, wherein first affecting parameters are to believe the water quality time series data of the first website and the hydrology timing of the first website
Any hydrographic data is to the influence degree of the water quality data of the first website in the following predetermined amount of time in breath, and by each second station
The testing data of point is input to the first convolution neural network model and carries out feature extraction, and utilizes first convolutional Neural
The attention mechanism attention of network model combines predetermined each second website to carry out the first website influence degree
Processing, obtains the second affecting parameters, wherein second affecting parameters are the water quality time series data of each second website and each
Any hydrographic data is to the water quality data of the first website in the following predetermined amount of time in the hydrology timing information of a second website
Influence degree；
The prediction result of the water quality data for the first website is obtained, the prediction result is the water quality data of first website
Variation tendency in the following predetermined amount of time；
Using default second convolution neural network model to first affecting parameters, second affecting parameters and described pre-
It surveys result to be handled, the prediction result after being adjusted.
2. the method according to claim 1, wherein the convolution side that the first convolution neural network model uses
Formula is empty convolution.
3. the method according to claim 1, wherein described utilize default second convolution neural network model to institute
It states the first affecting parameters, second affecting parameters and the prediction result to be handled, the prediction result after being adjusted,
Include:
Using the attention mechanism attention of default second convolution neural network model to first affecting parameters, described
Second affecting parameters and the prediction result are coupled；
The result of coupling is handled using the CNN module of the second convolution neural network model, it is pre- after being adjusted
Survey result, wherein the second convolution neural network model uses convolution mode for cause and effect convolution.
4. the method according to claim 1, wherein the testing data for obtaining the first website and described the
In the corresponding presumptive area of one website after the testing data of at least one the second website, the method also includes；
Detect whether any data is not small in the testing data of the first website described in current point in time and/or any second website
In corresponding standard parameter；
If being not less than corresponding standard parameter, generates warning information and determine corresponding website.
5. the method according to claim 1, wherein the history testing data of first website by disposing in advance
It is collected in the detection device of the first website, the history testing data of each second website is by being previously deployed at each second station
The detection device of point collects, the method also includes:
Each second website is recalculated to the first website influence degree, and with predetermined each second website to first stop
Point influence degree is compared, and judges whether to change；
If each second website changes to the first website influence degree, the detection device hair for being deployed in the first website is determined
It is raw abnormal；
If each second website to the first website influence degree and it is not all change, determine changed second website
Detection device is abnormal.
6. according to the method described in claim 5, it is characterized in that, being calculated using the first convolution neural network model each
Second website is to the first website influence degree.
7. method according to any one of claim 1 to 6, which is characterized in that the hydrographic information include it is following at least
One: water temperature, water level, flow velocity.
8. a kind of prediction meanss of water pollution characterized by comprising
Data acquisition module, in the testing data and the corresponding presumptive area of first website for obtaining the first website extremely
The testing data of few second website, the testing data includes water quality time series data and hydrology timing information；
Determining module is influenced, for the testing data for obtaining the first website to be input to preset first convolutional neural networks
Model carries out feature extraction, and is handled using the attention mechanism attention of the first convolution neural network model,
Obtain the first affecting parameters, wherein first affecting parameters are the water quality time series data and the first website to the first website
Hydrology timing information in any hydrographic data to the influence degree of the water quality data of the first website in the following predetermined amount of time, and
The testing data of each second website is input to the first convolution neural network model and carries out feature extraction, and described in utilization
The attention mechanism attention of first convolution neural network model combines predetermined each second website to the first website
Influence degree is handled, and the second affecting parameters are obtained, wherein when second affecting parameters are the water quality of each second website
Ordinal number accordingly and in the hydrology timing information of each second website any hydrographic data to the first website in the following predetermined amount of time
Water quality data influence degree；
Predicted value obtains module, and for obtaining the prediction result for being directed to the water quality data of the first website, the prediction result is institute
State variation tendency of the water quality data of the first website in the following predetermined amount of time；
Processing module is predicted, for utilizing default second convolution neural network model to first affecting parameters, described second
Affecting parameters and the prediction result are handled, the prediction result after being adjusted.
9. a kind of electronic equipment, characterized in that it comprises:
One or more processors；
Memory；
One or more application program, wherein one or more of application programs are stored in the memory and are configured
To be executed by one or more of processors, one or more of programs are configured to: executing -7 according to claim 1
The prediction technique of water pollution described in one.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is by processor
The prediction technique of the described in any item water pollutions of claim 1-7 is realized when execution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910074973.2A CN109886477B (en) | 2019-01-25 | 2019-01-25 | Water pollution prediction method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910074973.2A CN109886477B (en) | 2019-01-25 | 2019-01-25 | Water pollution prediction method and device and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109886477A true CN109886477A (en) | 2019-06-14 |
| CN109886477B CN109886477B (en) | 2022-11-04 |
Family
ID=66926982
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910074973.2A Active CN109886477B (en) | 2019-01-25 | 2019-01-25 | Water pollution prediction method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109886477B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112084230A (en) * | 2020-07-30 | 2020-12-15 | 同济大学 | Surface water quality prediction method |
| CN112561876A (en) * | 2020-12-14 | 2021-03-26 | 中南大学 | Image-based pond and reservoir water quality detection method and system |
| CN114414751A (en) * | 2021-12-15 | 2022-04-29 | 华能煤炭技术研究有限公司 | Underground water quality detection method and system |
| CN117371873A (en) * | 2023-12-01 | 2024-01-09 | 四川省生态环境科学研究院 | Sewage assessment method for environmental protection engineering based on big data |
| CN117491585A (en) * | 2024-01-02 | 2024-02-02 | 石家庄首创水汇环境治理有限公司 | Water ecological pollution monitoring method, device and system based on time sequence network |
| CN117787511A (en) * | 2024-02-28 | 2024-03-29 | 福州工小四物联科技有限公司 | Industrial high-density aquaculture monitoring and early warning method and system thereof |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170364803A1 (en) * | 2016-06-17 | 2017-12-21 | International Business Machines Corporation | Time series forecasting to determine relative causal impact |
| US20180053108A1 (en) * | 2016-08-16 | 2018-02-22 | Toyota Jidosha Kabushiki Kaisha | Efficient Driver Action Prediction System Based on Temporal Fusion of Sensor Data Using Deep (Bidirectional) Recurrent Neural Network |
| CN108388900A (en) * | 2018-02-05 | 2018-08-10 | 华南理工大学 | The video presentation method being combined based on multiple features fusion and space-time attention mechanism |
-
2019
- 2019-01-25 CN CN201910074973.2A patent/CN109886477B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170364803A1 (en) * | 2016-06-17 | 2017-12-21 | International Business Machines Corporation | Time series forecasting to determine relative causal impact |
| US20180053108A1 (en) * | 2016-08-16 | 2018-02-22 | Toyota Jidosha Kabushiki Kaisha | Efficient Driver Action Prediction System Based on Temporal Fusion of Sensor Data Using Deep (Bidirectional) Recurrent Neural Network |
| CN108388900A (en) * | 2018-02-05 | 2018-08-10 | 华南理工大学 | The video presentation method being combined based on multiple features fusion and space-time attention mechanism |
Non-Patent Citations (4)
| Title |
|---|
| 刘天亮等: "融合空间-时间双网络流和视觉注意的人体行为识别", 《电子与信息学报》 * |
| 吴双双等: "基于双神经网络通道的时间序列预测框架", 《舰船电子工程》 * |
| 张琦等: "融合注意力机制和CNN-GRNN模型的读者情绪预测", 《计算机工程与应用》 * |
| 董建明: "地下水动态预测模型研究", 《安徽农业科学》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112084230A (en) * | 2020-07-30 | 2020-12-15 | 同济大学 | Surface water quality prediction method |
| CN112561876A (en) * | 2020-12-14 | 2021-03-26 | 中南大学 | Image-based pond and reservoir water quality detection method and system |
| CN112561876B (en) * | 2020-12-14 | 2024-02-23 | 中南大学 | Image-based water quality detection method and system for ponds and reservoirs |
| CN114414751A (en) * | 2021-12-15 | 2022-04-29 | 华能煤炭技术研究有限公司 | Underground water quality detection method and system |
| CN114414751B (en) * | 2021-12-15 | 2024-05-10 | 华能煤炭技术研究有限公司 | Underground water quality detection method and system |
| CN117371873A (en) * | 2023-12-01 | 2024-01-09 | 四川省生态环境科学研究院 | Sewage assessment method for environmental protection engineering based on big data |
| CN117371873B (en) * | 2023-12-01 | 2024-03-26 | 四川省生态环境科学研究院 | Sewage assessment method for environmental protection engineering based on big data |
| CN117491585A (en) * | 2024-01-02 | 2024-02-02 | 石家庄首创水汇环境治理有限公司 | Water ecological pollution monitoring method, device and system based on time sequence network |
| CN117787511A (en) * | 2024-02-28 | 2024-03-29 | 福州工小四物联科技有限公司 | Industrial high-density aquaculture monitoring and early warning method and system thereof |
| CN117787511B (en) * | 2024-02-28 | 2024-05-10 | 福州工小四物联科技有限公司 | Industrial high-density aquaculture monitoring and early warning method and system thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109886477B (en) | 2022-11-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109886477A (en) | Prediction technique, device and the electronic equipment of water pollution | |
| CN110977614B (en) | Health diagnosis method for numerical control machine tool | |
| CN109656793A (en) | A kind of information system performance stereoscopic monitoring method based on multi-source heterogeneous data fusion | |
| CN102148701B (en) | Trend analysis method and device for monitoring indexes in monitoring system | |
| CN110838075A (en) | Training and predicting method and device for prediction model of transient stability of power grid system | |
| Dong et al. | Frequency prediction of power systems in FNET based on state-space approach and uncertain basis functions | |
| CN105608842A (en) | Nuclear reactor fuel failure online monitoring alarm device | |
| CN112433122A (en) | Charging pile available state detection method, device, equipment and storage medium | |
| CN116663747B (en) | Intelligent early warning method and system based on data center infrastructure | |
| CN114356734A (en) | Service abnormity detection method and device, equipment and storage medium | |
| CN104254083A (en) | Method and device for predicting business hot spots | |
| CN107015900A (en) | A kind of service performance Forecasting Methodology of video website | |
| Evans et al. | Evolution and future of urban road incident detection algorithms | |
| CN107277777B (en) | A kind of indoor orientation method and device | |
| CN105489010A (en) | System and method for monitoring and analyzing fast road travel time reliability | |
| CN117314020A (en) | Wetland carbon sink data monitoring system of plankton | |
| Zhang et al. | An approach for real-time urban traffic state estimation by fusing multisource traffic data | |
| CN117331015A (en) | Misalignment early warning method and system for CVT metering error | |
| WO2023159638A1 (en) | Battery fault detection system, method and apparatus | |
| CN115932593A (en) | Power station temperature thermal runaway early warning method, system and terminal | |
| CN116775400A (en) | Log analysis method, device, equipment and storage medium | |
| CN116319255A (en) | Root cause positioning method, device, equipment and storage medium based on KPI | |
| CN108093431A (en) | It is a kind of using do data processing network element TDPNE be handle node data reporting method and device | |
| Zheng | Fuzzy algorithm-based fault analysis for automated production lines | |
| CN112834841A (en) | Method and device for detecting service life of infrared camera and processor |
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 |