CN111882138B - Water quality prediction method, device, equipment and storage medium based on space-time fusion - Google Patents
Water quality prediction method, device, equipment and storage medium based on space-time fusion Download PDFInfo
- Publication number
- CN111882138B CN111882138B CN202010792866.6A CN202010792866A CN111882138B CN 111882138 B CN111882138 B CN 111882138B CN 202010792866 A CN202010792866 A CN 202010792866A CN 111882138 B CN111882138 B CN 111882138B
- Authority
- CN
- China
- Prior art keywords
- water quality
- parameter data
- quality parameter
- time
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 416
- 230000004927 fusion Effects 0.000 title claims abstract description 65
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000003860 storage Methods 0.000 title claims abstract description 14
- 239000013535 sea water Substances 0.000 claims abstract description 51
- 230000015654 memory Effects 0.000 claims abstract description 29
- 238000012544 monitoring process Methods 0.000 claims abstract description 28
- 238000012549 training Methods 0.000 claims abstract description 21
- 238000010586 diagram Methods 0.000 claims description 26
- 230000002123 temporal effect Effects 0.000 claims description 26
- 230000007246 mechanism Effects 0.000 claims description 24
- 238000012545 processing Methods 0.000 claims description 21
- 230000001932 seasonal effect Effects 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 10
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 abstract description 9
- 238000004364 calculation method Methods 0.000 description 8
- 230000008901 benefit Effects 0.000 description 6
- 230000008859 change Effects 0.000 description 6
- 230000007774 longterm Effects 0.000 description 6
- 241000196171 Hydrodictyon reticulatum Species 0.000 description 5
- 238000009360 aquaculture Methods 0.000 description 5
- 244000144974 aquaculture Species 0.000 description 5
- 238000004891 communication Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 230000009471 action Effects 0.000 description 3
- 230000001364 causal effect Effects 0.000 description 3
- IOVCWXUNBOPUCH-UHFFFAOYSA-M Nitrite anion Chemical compound [O-]N=O IOVCWXUNBOPUCH-UHFFFAOYSA-M 0.000 description 2
- OAICVXFJPJFONN-UHFFFAOYSA-N Phosphorus Chemical compound [P] OAICVXFJPJFONN-UHFFFAOYSA-N 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 2
- XKMRRTOUMJRJIA-UHFFFAOYSA-N ammonia nh3 Chemical compound N.N XKMRRTOUMJRJIA-UHFFFAOYSA-N 0.000 description 2
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 2
- 229930002868 chlorophyll a Natural products 0.000 description 2
- ATNHDLDRLWWWCB-AENOIHSZSA-M chlorophyll a Chemical compound C1([C@@H](C(=O)OC)C(=O)C2=C3C)=C2N2C3=CC(C(CC)=C3C)=[N+]4C3=CC3=C(C=C)C(C)=C5N3[Mg-2]42[N+]2=C1[C@@H](CCC(=O)OC\C=C(/C)CCC[C@H](C)CCC[C@H](C)CCCC(C)C)[C@H](C)C2=C5 ATNHDLDRLWWWCB-AENOIHSZSA-M 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000004069 differentiation Effects 0.000 description 2
- 238000009313 farming Methods 0.000 description 2
- 229910052760 oxygen Inorganic materials 0.000 description 2
- 239000001301 oxygen Substances 0.000 description 2
- 229910052698 phosphorus Inorganic materials 0.000 description 2
- 239000011574 phosphorus Substances 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 108010003272 Hyaluronate lyase Proteins 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012851 eutrophication Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000009364 mariculture Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
- 230000016776 visual perception Effects 0.000 description 1
- 229940088594 vitamin Drugs 0.000 description 1
- 229930003231 vitamin Natural products 0.000 description 1
- 235000013343 vitamin Nutrition 0.000 description 1
- 239000011782 vitamin Substances 0.000 description 1
- 150000003722 vitamin derivatives Chemical class 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Mining
-
- 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
Abstract
The invention relates to the technical field of seawater cage culture, and discloses a water quality prediction method, device, equipment and storage medium based on space-time fusion, wherein the method comprises the following steps: searching a water quality monitoring record corresponding to the seawater net cage, extracting historical water quality parameter data from the water quality monitoring record, and taking the historical water quality parameter data as time water quality parameter data; acquiring current water quality parameter data acquired by a plurality of preset sensors in a seawater net cage, and taking the current water quality parameter data as space water quality parameter data; determining target data according to the time water quality parameter data and the space water quality parameter data; constructing a network prediction model based on a preset long-short-term memory network and a preset time convolution network; and inputting the target data into a network prediction model for training to obtain a predicted value, and taking the predicted value as a water quality predicted result. Therefore, the water quality is predicted in a space-time fusion mode to obtain a predicted value by acquiring time water quality parameter data and space water quality parameter data, and the accuracy of water quality detection is improved.
Description
Technical Field
The invention relates to the technical field of seawater cage culture, in particular to a water quality prediction method, device and equipment based on space-time fusion and a storage medium.
Background
China is a large country of aquaculture, and along with the continuous promotion of social economy and living standard of people, the demands of people on marine products also show a trend of rising year by year. In recent years, deep sea cage culture has been rapidly developed as a novel fishery culture mode, and the cage culture becomes one of important forms of aquaculture activities, and the deep water storm-resistant cage culture also enables the mariculture industry to be continuously transformed and upgraded. But at the same time, the seawater cage culture has the characteristics of high density and high bait feeding mode, and the phenomena such as eutrophication of water bodies, degradation of substrates, frequent diseases and the like are caused by the overlarge culture density, unreasonable layout and a series of unreasonable culture management measures.
The water quality safety has the greatest influence on the aquaculture, and the stability of the aquaculture water quality environment is ensured to play a vital role in the health of the aquaculture objects. The seawater culture can enlarge the degree and the range of uncontrollable factors in the artificial culture management process due to the factors of the external environment, and the accumulated experience and visual perception of people are not reliable any more along with the deepening of the seawater net cage depth.
Disclosure of Invention
The invention mainly aims to provide a water quality prediction method, a device, equipment and a storage medium based on space-time fusion, which aim at solving the technical problem of how to improve the accuracy of water quality detection.
In order to achieve the above object, the present invention provides a water quality prediction method based on space-time fusion, which comprises the following steps:
searching a water quality monitoring record corresponding to the seawater net cage, and extracting historical water quality parameter data from the water quality monitoring record;
acquiring current water quality parameter data acquired by a plurality of preset sensors in the seawater net cage;
taking the historical water quality parameter data as time water quality parameter data and taking the current water quality parameter data as space water quality parameter data;
determining target data according to the time water quality parameter data and the space water quality parameter data;
constructing a network prediction model based on a preset long-short-term memory network and a preset time convolution network;
and inputting the target data into the network prediction model for training to obtain a predicted value, and taking the predicted value as a water quality predicted result.
Preferably, the determining target data according to the time water quality parameter data and the space water quality parameter data specifically includes:
Taking the time water quality parameter data and the space water quality parameter data as water quality parameter data;
preprocessing the water quality parameter data to generate a time sequence diagram corresponding to the water quality parameter data;
judging whether the water quality parameter data is stable or not according to the time sequence diagram, and determining the period of original sequence data in the time sequence diagram;
when the water quality parameter data is not stable, extracting first water quality parameter data at the current moment and second water quality parameter data at the last period from the water quality parameter data;
and determining target data according to the first water quality parameter data and the second water quality parameter data.
Preferably, the determining target data according to the first water quality parameter data and the second water quality parameter data specifically includes:
obtaining a differential sequence according to the first water quality parameter data and the second water quality parameter data;
carrying out seasonal difference processing on the difference sequence to obtain a target sequence;
inputting the target sequence into a preset prediction model for training to obtain prediction data;
and carrying out seasonal difference inverse conversion on the predicted data to obtain target data.
Preferably, the inputting the target data into the network prediction model for training to obtain a predicted value, and taking the predicted value as a water quality prediction result specifically includes:
Inputting the target data into the network prediction model, and extracting the spatial characteristics of the target data through a preset spatial attention mechanism and the preset long and short memory network;
extracting the time characteristics of the target data through a preset time attention mechanism and the preset time convolution network;
determining a spatial attention value according to the spatial features, and determining a temporal attention value according to the temporal features;
determining a predicted value from the spatial attention value and the temporal attention value;
and taking the predicted value as a water quality predicted result.
Preferably, the determining a spatial attention value according to the spatial feature and determining a temporal attention value according to the temporal feature specifically includes:
calculating a space weight coefficient of the space feature by using a vector dot product, and calculating a time weight coefficient of the time feature by using a similarity;
respectively carrying out normalization processing on the space weight coefficient and the time weight coefficient through a preset function to obtain a target space weight coefficient and a target time weight coefficient;
carrying out weighted summation processing on the target space weight coefficient to obtain a space attention value;
And carrying out weighted summation processing on the target time weight coefficient to obtain a time attention value.
Preferably, after the target data is input into the network prediction model to perform training to obtain a predicted value and the predicted value is used as a water quality prediction result, the method further includes:
judging whether the predicted value is larger than a maximum value of a preset parameter;
generating first early warning information when the predicted value is larger than the maximum value of the preset parameter;
and carrying out early warning prompt according to the first early warning information.
Preferably, after the target data is input into the network prediction model to perform training to obtain a predicted value and the predicted value is used as a water quality prediction result, the method further includes:
judging whether the predicted value is smaller than a minimum value of a preset parameter;
generating second early warning information when the predicted value is smaller than the minimum value of the preset parameter;
and carrying out early warning prompt according to the second early warning information.
In addition, in order to achieve the above object, the present invention also provides a water quality prediction device based on space-time fusion, the water quality prediction device based on space-time fusion includes:
the historical data module is used for searching a water quality monitoring record corresponding to the seawater net cage and extracting historical water quality parameter data from the water quality monitoring record;
The current data module is used for acquiring current water quality parameter data acquired by a plurality of preset sensors in the seawater net cage;
the space-time data module is used for taking the historical water quality parameter data as time water quality parameter data and taking the current water quality parameter data as space water quality parameter data;
the target data module is used for determining target data according to the time water quality parameter data and the space water quality parameter data;
the model construction module is used for constructing a network prediction model based on a preset long-short-time memory network and a preset time convolution network;
and the prediction result module is used for inputting the target data into the network prediction model for training to obtain a prediction value, and taking the prediction value as a water quality prediction result.
In addition, in order to achieve the above object, the present invention also provides a water quality prediction apparatus based on space-time fusion, the water quality prediction apparatus based on space-time fusion comprising: the system comprises a memory, a processor and a water quality prediction program based on space-time fusion, wherein the water quality prediction program based on space-time fusion is stored on the memory and can run on the processor, and the water quality prediction program based on space-time fusion is provided with the steps for realizing the water quality prediction method based on space-time fusion.
In addition, in order to achieve the above object, the present invention also proposes a storage medium having stored thereon a water quality prediction program based on space-time fusion, which when executed by a processor, implements the steps of the water quality prediction method based on space-time fusion as described above.
According to the water quality prediction method based on space-time fusion, the water quality monitoring records corresponding to the sea water net cage are searched, and historical water quality parameter data are extracted from the water quality monitoring records; acquiring current water quality parameter data acquired by a plurality of preset sensors in the seawater net cage; taking the historical water quality parameter data as time water quality parameter data and taking the current water quality parameter data as space water quality parameter data; determining target data according to the time water quality parameter data and the space water quality parameter data; constructing a network prediction model based on a preset long-short-term memory network and a preset time convolution network; and inputting the target data into the network prediction model for training to obtain a predicted value, and taking the predicted value as a water quality predicted result. Therefore, the water quality is predicted in a space-time fusion mode to obtain a predicted value by acquiring time water quality parameter data and space water quality parameter data, and the accuracy of water quality detection is improved.
Drawings
FIG. 1 is a schematic diagram of a water quality prediction device based on space-time fusion in a hardware operating environment according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart of a first embodiment of a water quality prediction method based on space-time fusion according to the present invention;
FIG. 3 is a schematic diagram of a minimum unit of the spatial distribution of a sensor according to a first embodiment of the water quality prediction method based on space-time fusion;
FIG. 4 is a schematic view showing the lateral and longitudinal expansion of a spatial minimum distribution unit in a first embodiment of a water quality prediction method based on space-time fusion;
FIG. 5 is a schematic flow chart of a second embodiment of a water quality prediction method based on space-time fusion;
FIG. 6 is a schematic flow chart of a third embodiment of a water quality prediction method based on space-time fusion according to the present invention;
FIG. 7 is a logic diagram of an attention mechanism of a third embodiment of a water quality prediction method based on space-time fusion according to the present invention;
FIG. 8 is a schematic diagram of functional modules of a first embodiment of a water quality prediction apparatus based on space-time fusion according to the present invention.
The achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.
Detailed Description
It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Referring to fig. 1, fig. 1 is a schematic structural diagram of a water quality prediction device based on space-time fusion in a hardware operation environment according to an embodiment of the present invention.
As shown in fig. 1, the water quality prediction apparatus based on space-time fusion may include: a processor 1001, such as a central processing unit (Central Processing Unit, CPU), a communication bus 1002, a user interface 1003, a network interface 1004, a memory 1005. Wherein the communication bus 1002 is used to enable connected communication between these components. The user interface 1003 may include a Display, an input unit such as keys, and the optional user interface 1003 may also include a standard wired interface, a wireless interface. The network interface 1004 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM) memory or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also optionally be a storage device separate from the processor 1001 described above.
It will be appreciated by those skilled in the art that the apparatus structure shown in FIG. 1 is not limiting of a space-time fusion-based water quality prediction apparatus, and may include more or fewer components than shown, or certain components may be combined, or a different arrangement of components.
As shown in fig. 1, an operating system, a network communication module, a user interface module, and a water quality prediction program based on space-time fusion may be included in the memory 1005 as one storage medium.
In the water quality prediction device based on space-time fusion shown in fig. 1, the network interface 1004 is mainly used for connecting an external network and performing data communication with other network devices; the user interface 1003 is mainly used for connecting user equipment and communicating data with the user equipment; the apparatus of the present invention invokes the water quality prediction program based on space-time fusion stored in the memory 1005 through the processor 1001, and executes the water quality prediction method based on space-time fusion provided by the embodiment of the present invention.
Based on the hardware structure, the embodiment of the water quality prediction method based on space-time fusion is provided.
Referring to fig. 2, fig. 2 is a schematic flow chart of a first embodiment of a water quality prediction method based on space-time fusion according to the present invention.
In a first embodiment, the water quality prediction method based on space-time fusion comprises the following steps:
and step S10, searching a water quality monitoring record corresponding to the seawater net cage, and extracting historical water quality parameter data from the water quality monitoring record.
It should be noted that, the execution body of the present embodiment may be a water quality prediction device based on space-time fusion, or may be other devices that can implement the same or similar functions, which is not limited in this embodiment, and in this embodiment, a water quality prediction device based on space-time fusion is described as an example.
It should be appreciated that with the continued development and upgrading of sea cage farming, both the increase in cage specifications and the expansion of the range are faced with a great impact caused by the variation of the water parameters. The stability of water quality parameters in the environment of the culture water body plays a vital role in the healthy growth of the culture object, and the data of different water quality variables have respective distribution rules at different depth layers of seawater and different positions of the same depth layer. Currently, seawater quality parameters are greatly influenced by external environments, and related researches on regular changes of the seawater quality parameters based on seawater net cage culture environments are less. Therefore, the data acquisition rule analysis is required to be carried out on the seawater cage culture water quality parameters, so that the prediction and the early warning of the water quality parameters in the cage culture can be realized, the stability of the water quality environment and the healthy growth of the culture objects are ensured, the pollution degree of the seawater cage culture to the surrounding environment is reduced, and the maximum economic benefit of the seawater cage culture is obtained. The embodiment explores the change rule of the water quality parameters of the seawater cage culture in time and space, and realizes the prediction and the early warning of the water quality parameters of different depths and different positions of the same depth based on the analysis result.
It is understood that the collection of the seawater cage culture water quality parameter data comprises two aspects of time and space. The sensor is arranged on the time data to acquire data every 10 minutes, and the accumulated data of the net cage culture water quality parameter data for many years is obtained through continuous monitoring; the space data is realized by distributing multi-parameter sensors at a plurality of preset positions, and the water quality parameter information of different depths and different positions of the same depth of the sea water net cage is respectively acquired.
It should be understood that after the sensor collects data, the water quality monitoring record is generated and stored in the database, so that the water quality monitoring record corresponding to the seawater net cage can be searched in the database, and the historical water quality parameter data, which is the water quality parameter data collected and stored before, can be extracted from the water quality monitoring record.
It will be appreciated that the water quality parameter data includes: the water temperature, salinity, PH, dissolved oxygen, COD, chlorophyll a, ammonia nitrogen, active phosphorus, nitrite, etc., may be other data, and the embodiment is not limited thereto.
And S20, acquiring current water quality parameter data acquired by a plurality of preset sensors in the seawater net cage.
It should be noted that, a plurality of preset sensors are preset in the seawater net cage, and the water quality parameter data is collected through the preset sensors, where the preset sensors may be a multi-parameter water quality detection sensor or other sensors, which is not limited in this embodiment.
In a specific implementation, as shown in fig. 3, fig. 3 is a sensor space distribution minimum unit, that is, a sensor distribution minimum acquisition unit of water quality parameter data in space, multiple parameter water quality detection sensors are placed at intervals of 0.5m in depth, and the distance between two adjacent sensors in the same depth is 1m. And constructing minimum acquisition units of space water quality parameter data, wherein each unit comprises 14 position distribution points for acquiring the water quality parameter data. As shown in fig. 4, fig. 4 shows the lateral and longitudinal expansion of the minimum spatial distribution unit, so that water quality information of a plurality of location points can be collected.
Step S30, taking the historical water quality parameter data as time water quality parameter data and taking the current water quality parameter data as space water quality parameter data.
It will be appreciated that after the historical water quality parameter data and the current water quality parameter data are obtained, the historical water quality parameter data is used as time water quality parameter data, and the current water quality parameter data is used as space water quality parameter data for subsequent water quality prediction.
And step S40, determining target data according to the time water quality parameter data and the space water quality parameter data.
It should be understood that the obtained time water quality parameter data and space water quality parameter data are preprocessed, and a difference method is used to eliminate the time dependence of the data due to seasonality, so that a stable time sequence is constructed. According to the stable time sequence, the water quality prediction model is combined to predict future water quality parameter data according to historical data, and the spatial expansion prediction of the water quality parameter data is realized according to the data accumulation of the spatial sensor distribution minimum unit. According to the target to be realized, attention mechanisms with different actions are added into the prediction model so as to grasp the time-space correlation and long-term time sequence dependency relationship between data. And (5) according to the predicted value of the water quality parameter, realizing early warning of the water quality change condition.
And S50, constructing a network prediction model based on a preset long-short-time memory network and a preset time convolution network.
In the embodiment, a deep network water quality parameter prediction model is constructed based on a time convolution network. A deep network architecture, namely a network prediction model, is built by using an algorithm model combining a Long Short-Term Memory (LSTM) and a time convolution network (Temporal Convolutional Network, TCN), and the problem that information can be forgotten for a Long-Term time sequence due to the LSTM is solved well by a causal convolution network layer in the TCN, so that history information of collected data is not missed. For water quality parameter data of a position to be predicted, the historical time sequence monitored by other sensors has indirect influence on the water quality parameter data and changes with time; and the collected water quality parameter data is accumulated information for many years, and time dependence exists between the data. Therefore, attention mechanisms with different functions are added into the deep network structure of the prediction model and are used for giving different weight values to different positions in space and different moments in time, and the long-term time sequence dependency relationship and the dynamic time-space relevance between the water quality parameter information are grasped.
And step S60, inputting the target data into the network prediction model for training to obtain a predicted value, and taking the predicted value as a water quality prediction result.
Further, after the step S60, the method further includes:
judging whether the predicted value is larger than a maximum value of a preset parameter; generating first early warning information when the predicted value is larger than the maximum value of the preset parameter; and carrying out early warning prompt according to the first early warning information.
Further, after the step S60, the method further includes:
judging whether the predicted value is smaller than a minimum value of a preset parameter; generating second early warning information when the predicted value is smaller than the minimum value of the preset parameter; and carrying out early warning prompt according to the second early warning information.
It should be understood that the preset maximum value and the preset minimum value of the parameters can be preset and used as water quality parameter standards, when the predicted value is larger than the preset maximum value of the parameters, the first early warning information is generated, and when the predicted value is smaller than the preset minimum value of the parameters, the second early warning information is generated, so that different early warning prompts can be given when the predicted value exceeds the range or is lower than the range, and a manager can regulate and control in advance to ensure the stability of water quality.
In the embodiment, the water quality monitoring record corresponding to the sea water net cage is searched, and historical water quality parameter data is extracted from the water quality monitoring record; acquiring current water quality parameter data acquired by a plurality of preset sensors in the seawater net cage; taking the historical water quality parameter data as time water quality parameter data and taking the current water quality parameter data as space water quality parameter data; determining target data according to the time water quality parameter data and the space water quality parameter data; constructing a network prediction model based on a preset long-short-term memory network and a preset time convolution network; and inputting the target data into the network prediction model for training to obtain a predicted value, and taking the predicted value as a water quality predicted result. Therefore, the water quality is predicted by acquiring time water quality parameter data and space water quality parameter data in a space-time fusion mode to obtain a predicted value, the accuracy of water quality detection is improved, the prediction of the water quality change trend along with the increase of time or along with the increase of depth in seawater cage culture is realized, the occurrence of disasters is effectively prevented, the healthy growth of culture objects is ensured, and the economic benefit of culture is improved.
In an embodiment, as shown in fig. 5, a second embodiment of the water quality prediction method based on space-time fusion according to the present invention is proposed based on the first embodiment, and the step S40 includes:
and step S401, taking the time water quality parameter data and the space water quality parameter data as water quality parameter data.
Step S402, preprocessing the water quality parameter data to generate a time sequence diagram corresponding to the water quality parameter data.
The water quality parameter data is preprocessed, and a time sequence diagram corresponding to the water quality parameter data is made.
Step S403, judging whether the water quality parameter data is stable according to the time sequence diagram, and determining the period of the original sequence data in the time sequence diagram.
Step S404, when the water quality parameter data is not stable, extracting the first water quality parameter data at the current moment and the second water quality parameter data at the last period from the water quality parameter data.
Step S405, determining target data according to the first water quality parameter data and the second water quality parameter data
Further, the step S405 includes:
obtaining a differential sequence according to the first water quality parameter data and the second water quality parameter data; carrying out seasonal difference processing on the difference sequence to obtain a target sequence; inputting the target sequence into a preset prediction model for training to obtain prediction data; and carrying out seasonal difference inverse conversion on the predicted data to obtain target data.
The period of the original sequence data in the timing chart is determined as S, and a differential sequence is created. The acquired water quality parameter data is provided with a parameter which is observed for each water quality variableTime series of values: n (N) 1 ,N 2 …N t …,N n Seasonal differentiation is performed for time t: subtracting the previous period observed value from the current observed value to obtain a new sequence with the length of n-1, wherein the calculation formula is as follows:
value(t)＝obs(t)-obs(t-S)；
wherein value (t) is a value obtained after season differentiation is performed on the observed value at time t, and obs (t) is an observed value at time t in the collected water quality data.
And inputting the constructed stable time sequence into a constructed water quality parameter prediction model to realize the prediction of the water quality parameters of the cage culture. The resulting predicted values need to be subjected to inverse conversion of the seasonal differences to obtain the same data dimension as the original data. By adding the obtained predicted value yhat (t) to the original observed value at the corresponding moment, inversion is realized, and the calculation formula is as follows:
new(t)＝yhat(t)+obs(t-S)；
wherein new (t) is a new sequence after inversion after water quality parameter prediction, and is the same as the original data in dimension, yhat (t) is a water quality parameter predicted value, and obs (t-S) is an observed value of collected water quality data at the time of t-S.
In the embodiment, the time water quality parameter data and the space water quality parameter data are used as water quality parameter data; preprocessing the water quality parameter data to generate a time sequence diagram corresponding to the water quality parameter data; judging whether the water quality parameter data is stable or not according to the time sequence diagram, and determining the period of original sequence data in the time sequence diagram; when the water quality parameter data is not stable, extracting first water quality parameter data at the current moment and second water quality parameter data at the last period from the water quality parameter data; and determining target data according to the first water quality parameter data and the second water quality parameter data, so that after the water quality parameter data are preprocessed, a time sequence diagram is generated to determine the target data, and further, the accuracy of water quality prediction is improved.
In an embodiment, as shown in fig. 6, a third embodiment of the water quality prediction method based on space-time fusion according to the present invention is provided based on the first embodiment or the second embodiment, in this embodiment, the description is given based on the first embodiment, and the step S60 includes:
step S601, inputting the target data into the network prediction model, and extracting spatial features of the target data through a preset spatial attention mechanism and the preset long and short memory network.
Step S602, extracting the time characteristics of the target data through a preset time attention mechanism and the preset time convolution network.
Step S603, determining a spatial attention value according to the spatial feature, and determining a temporal attention value according to the temporal feature.
Further, the step S603 includes:
calculating a space weight coefficient of the space feature by using a vector dot product, and calculating a time weight coefficient of the time feature by using a similarity; respectively carrying out normalization processing on the space weight coefficient and the time weight coefficient through a preset function to obtain a target space weight coefficient and a target time weight coefficient; carrying out weighted summation processing on the target space weight coefficient to obtain a space attention value; and carrying out weighted summation processing on the target time weight coefficient to obtain a time attention value.
Step S604, determining a predicted value according to the spatial attention value and the temporal attention value.
And step S605, taking the predicted value as a water quality predicted result.
In the space dimension, the water quality parameter conditions at different positions are mutually influenced, so that the dynamic performance is very strong. Dynamic associations between nodes in a spatial dimension are adaptively captured using a spatial attention mechanism, with different weights being assigned to current sensor information and sensor information at other locations. When acquiring water quality information, the acquired time period is T, and N is set s A plurality of water quality parameter sensors, each water quality parameter sensor having N m And (3) a time sequence, wherein m is the number of water quality information parameters to be acquired.
The correlation between the data sequences of the sensors is calculated, taking the attention value of the water quality sensor i to the other sensor l at the moment t as an example, and the calculation formula is as follows:
wherein,attention parameter value X representing sensor l at time t l For the water quality parameter value acquired by the sensor l at the time t, h t-1 ，s t-1 Respectively input characteristic vector and history state, V s ，u s ，n s ∈R T ，W s ∈R T2n ，U s ∈R TT ，Is a parameter that needs to be learned.
The calculation formula of the attention weight is as follows:
wherein,for the weight value obtained by the sensor at time t, P i,l Representing the geographical correlation between sensor i and sensor l, P i,j Representing the geographical correlation between sensor i and sensor j, λ is an adjustable super-parameter.
Finally, an output vector of the spatial attention at the moment t is obtained:
wherein,n representing the ith sensor at the time t m The value of the vitamin feature.
In a specific implementation, as shown in fig. 7, fig. 7 is a logic diagram of an attention mechanism. After being processed by the spatial attention mechanism, the data with different weights for different bit information is input into the LSTM network to update the hidden state of the network. The LSTM network is based on the state h of the last moment t-1 And the current input gamma t To calculate the state h at the current time t The calculation formula of the network is as follows:
h t ＝f e (h t-1 ,γ t ；w lstm )；
wherein h is t-1 Feature vector gamma input for network t-1 moment network t To output vector through spatial attention mechanism, w lstm Is a model parameter of the network LSTM, f e Representing a network element.
After passing through the multilayer LSTM network, the feature vector (h) of each moment output by the last layer network is obtained 0 ,h 1 ,h 2 ,…,h t ) Network input x as time convolutional network TCN 0 ,x 1 ,x 2 ,…,x T By processing the TCN network, we can obtain the predicted output at the corresponding momentIn order to ensure that the predicted value of the time T is related to the time T and the data before the time T and to increase the receptive field of the network, the time convolution network TCN respectively introduces causal convolution and expansion convolution.
The calculation formula of the expansion convolution operation F of the element s in the sequence is as follows:
where d is a spreading factor, k is a filter size, and (s-d.i) represents the past direction.
In the time dimension, correlation exists among water quality parameters in different time intervals, a time attention mechanism is adaptively selected to generate an output sequence by using a relevant hidden layer in a time convolutional network TCN, dynamic time correlation among different time intervals in a predicted sequence is modeled, and effective information of water quality sequence data in time is better captured.
Output hidden feature vector using TCN networkThe water quality data is input into a time attention mechanism, and different weights at different moments of the water quality data are adaptively given by using the attention mechanism.
For a time convolution network, the attention weight at the output time t under each hidden layer is calculated as follows:
1. calculating the weight of the collected water quality data at the time t by using the similarity of Cosine:
wherein,the weight coefficient s of the water quality data acquired at the moment t is t-1 Is the historical state of the network at the time t-1, h p In a hidden state, W d ∈R nn ，U d ∈R TT ，v d ∈R n Is a parameter to be learned.
2. The preset function is a SoftMax () function, the SoftMax () function is used for carrying out normalization processing on the obtained weight, the weight score obtained in the step 1 is changed into probability distribution with the sum of weights of all elements being 1, the weight value with larger action is highlighted, and the calculation formula is as follows:
3. and (3) carrying out weighted summation on the weight coefficients of all the times obtained in the step (2), so as to obtain a time attention value, wherein the calculation formula is as follows:
finally, outputting the time convolution network TCN hidden layer data of the attention mechanismInputting a time attention value r at different moments * ＝c 0 ,c 1 ,…,c T And (5) splicing. The final predicted output variables are:
h * ＝tanh(W p r * +W x y * )；
Wherein W is p And W is x Parameters to be trained in the model; output variable h * The prediction of the water quality variable is realized after passing through a full connection layer and a Softmax classifier.
In this embodiment, the spatial features of the target data are extracted through a preset spatial attention mechanism and the preset long-short memory network by inputting the target data into the network prediction model; extracting the time characteristics of the target data through a preset time attention mechanism and the preset time convolution network; determining a spatial attention value according to the spatial features, and determining a temporal attention value according to the temporal features; determining a predicted value from the spatial attention value and the temporal attention value; and taking the predicted value as a water quality predicted result. The method realizes the prediction of the water quality change trend along with the increase of time or along with the increase of depth in the seawater cage culture, effectively prevents the occurrence of disasters, ensures the healthy growth of culture objects, and improves the economic benefit of culture
In addition, the embodiment of the invention also provides a storage medium, wherein the storage medium is stored with a water quality prediction program based on space-time fusion, and the water quality prediction program based on space-time fusion realizes the steps of the water quality prediction method based on space-time fusion when being executed by a processor.
Because the storage medium adopts all the technical schemes of all the embodiments, the storage medium has at least all the beneficial effects brought by the technical schemes of the embodiments, and the description is omitted here.
In addition, referring to fig. 8, the embodiment of the invention further provides a water quality prediction device based on space-time fusion, where the water quality prediction device based on space-time fusion includes:
the historical data module 10 is used for searching the water quality monitoring record corresponding to the seawater net cage and extracting historical water quality parameter data from the water quality monitoring record.
It should be appreciated that with the continued development and upgrading of sea cage farming, both the increase in cage specifications and the expansion of the range are faced with a great impact caused by the variation of the water parameters. The stability of water quality parameters in the environment of the culture water body plays a vital role in the healthy growth of the culture object, and the data of different water quality variables have respective distribution rules at different depth layers of seawater and different positions of the same depth layer. Currently, seawater quality parameters are greatly influenced by external environments, and related researches on regular changes of the seawater quality parameters based on seawater net cage culture environments are less. Therefore, the data acquisition rule analysis is required to be carried out on the seawater cage culture water quality parameters, so that the prediction and the early warning of the water quality parameters in the cage culture can be realized, the stability of the water quality environment and the healthy growth of the culture objects are ensured, the pollution degree of the seawater cage culture to the surrounding environment is reduced, and the maximum economic benefit of the seawater cage culture is obtained. The embodiment explores the change rule of the water quality parameters of the seawater cage culture in time and space, and realizes the prediction and the early warning of the water quality parameters of different depths and different positions of the same depth based on the analysis result.
It is understood that the collection of the seawater cage culture water quality parameter data comprises two aspects of time and space. The sensor is arranged on the time data to acquire data every 10 minutes, and the accumulated data of the net cage culture water quality parameter data for many years is obtained through continuous monitoring; the space data is realized by distributing multi-parameter sensors at a plurality of preset positions, and the water quality parameter information of different depths and different positions of the same depth of the sea water net cage is respectively acquired.
It should be understood that after the sensor collects data, the water quality monitoring record is generated and stored in the database, so that the water quality monitoring record corresponding to the seawater net cage can be searched in the database, and the historical water quality parameter data, which is the water quality parameter data collected and stored before, can be extracted from the water quality monitoring record.
It will be appreciated that the water quality parameter data includes: the water temperature, salinity, PH, dissolved oxygen, COD, chlorophyll a, ammonia nitrogen, active phosphorus, nitrite, etc., may be other data, and the embodiment is not limited thereto.
The current data module 20 is configured to obtain current water quality parameter data collected by a plurality of preset sensors in the seawater cage.
It should be noted that, a plurality of preset sensors are preset in the seawater net cage, and the water quality parameter data is collected through the preset sensors, where the preset sensors may be a multi-parameter water quality detection sensor or other sensors, which is not limited in this embodiment.
In a specific implementation, as shown in fig. 3, fig. 3 is a sensor space distribution minimum unit, that is, a sensor distribution minimum acquisition unit of water quality parameter data in space, multiple parameter water quality detection sensors are placed at intervals of 0.5m in depth, and the distance between two adjacent sensors in the same depth is 1m. And constructing minimum acquisition units of space water quality parameter data, wherein each unit comprises 14 position distribution points for acquiring the water quality parameter data. As shown in fig. 4, fig. 4 shows the lateral and longitudinal expansion of the minimum spatial distribution unit, so that water quality information of a plurality of location points can be collected.
The space-time data module 30 is configured to take the historical water quality parameter data as time water quality parameter data and take the current water quality parameter data as space water quality parameter data.
It will be appreciated that after the historical water quality parameter data and the current water quality parameter data are obtained, the historical water quality parameter data is used as time water quality parameter data, and the current water quality parameter data is used as space water quality parameter data for subsequent water quality prediction.
And a target data module 40, configured to determine target data according to the time water quality parameter data and the space water quality parameter data.
It should be understood that the obtained time water quality parameter data and space water quality parameter data are preprocessed, and a difference method is used to eliminate the time dependence of the data due to seasonality, so that a stable time sequence is constructed. According to the stable time sequence, the water quality prediction model is combined to predict future water quality parameter data according to historical data, and the spatial expansion prediction of the water quality parameter data is realized according to the data accumulation of the spatial sensor distribution minimum unit. According to the target to be realized, attention mechanisms with different actions are added into the prediction model so as to grasp the time-space correlation and long-term time sequence dependency relationship between data. And (5) according to the predicted value of the water quality parameter, realizing early warning of the water quality change condition.
The model construction module 50 is configured to construct a network prediction model based on a preset long-short-term memory network and a preset time convolution network.
In the embodiment, a deep network water quality parameter prediction model is constructed based on a time convolution network. A deep network architecture, namely a network prediction model, is built by using an algorithm model combining a Long Short-Term Memory (LSTM) and a time convolution network (Temporal Convolutional Network, TCN), and the problem that information can be forgotten for a Long-Term time sequence due to the LSTM is solved well by a causal convolution network layer in the TCN, so that history information of collected data is not missed. For water quality parameter data of a position to be predicted, the historical time sequence monitored by other sensors has indirect influence on the water quality parameter data and changes with time; and the collected water quality parameter data is accumulated information for many years, and time dependence exists between the data. Therefore, attention mechanisms with different functions are added into the deep network structure of the prediction model and are used for giving different weight values to different positions in space and different moments in time, and the long-term time sequence dependency relationship and the dynamic time-space relevance between the water quality parameter information are grasped.
And the prediction result module 60 is configured to input the target data into the network prediction model for training, obtain a prediction value, and use the prediction value as a water quality prediction result.
It should be understood that the preset maximum value and the preset minimum value of the parameters can be preset and used as water quality parameter standards, when the predicted value is larger than the preset maximum value of the parameters, the first early warning information is generated, and when the predicted value is smaller than the preset minimum value of the parameters, the second early warning information is generated, so that different early warning prompts can be given when the predicted value exceeds the range or is lower than the range, and a manager can regulate and control in advance to ensure the stability of water quality.
In the embodiment, the water quality monitoring record corresponding to the sea water net cage is searched, and historical water quality parameter data is extracted from the water quality monitoring record; acquiring current water quality parameter data acquired by a plurality of preset sensors in the seawater net cage; taking the historical water quality parameter data as time water quality parameter data and taking the current water quality parameter data as space water quality parameter data; determining target data according to the time water quality parameter data and the space water quality parameter data; constructing a network prediction model based on a preset long-short-term memory network and a preset time convolution network; and inputting the target data into the network prediction model for training to obtain a predicted value, and taking the predicted value as a water quality predicted result. Therefore, the water quality is predicted in a space-time fusion mode to obtain a predicted value by acquiring time water quality parameter data and space water quality parameter data, and the accuracy of water quality detection is improved.
In one embodiment, the target data module 40 is further configured to use the time water quality parameter data and the space water quality parameter data as water quality parameter data; preprocessing the water quality parameter data to generate a time sequence diagram corresponding to the water quality parameter data; judging whether the water quality parameter data is stable or not according to the time sequence diagram, and determining the period of original sequence data in the time sequence diagram; when the water quality parameter data is not stable, extracting first water quality parameter data at the current moment and second water quality parameter data at the last period from the water quality parameter data; and determining target data according to the first water quality parameter data and the second water quality parameter data.
In an embodiment, the target data module 40 is further configured to obtain a differential sequence according to the first water quality parameter data and the second water quality parameter data; carrying out seasonal difference processing on the difference sequence to obtain a target sequence; inputting the target sequence into a preset prediction model for training to obtain prediction data; and carrying out seasonal difference inverse conversion on the predicted data to obtain target data.
In an embodiment, the prediction result module 60 is further configured to input the target data into the network prediction model, and extract spatial features of the target data through a preset spatial attention mechanism and the preset long-short memory network; extracting the time characteristics of the target data through a preset time attention mechanism and the preset time convolution network; determining a spatial attention value according to the spatial features, and determining a temporal attention value according to the temporal features; determining a predicted value from the spatial attention value and the temporal attention value; and taking the predicted value as a water quality predicted result.
In one embodiment, the prediction result module 60 is further configured to calculate a spatial weight coefficient of the spatial feature using a vector dot product, and calculate a temporal weight coefficient of the temporal feature using a similarity; respectively carrying out normalization processing on the space weight coefficient and the time weight coefficient through a preset function to obtain a target space weight coefficient and a target time weight coefficient; carrying out weighted summation processing on the target space weight coefficient to obtain a space attention value; and carrying out weighted summation processing on the target time weight coefficient to obtain a time attention value.
In an embodiment, the water quality prediction device based on space-time fusion further includes an early warning prompt module, configured to determine whether the predicted value is greater than a preset parameter maximum value; generating first early warning information when the predicted value is larger than the maximum value of the preset parameter; and carrying out early warning prompt according to the first early warning information.
In an embodiment, the early warning prompt module is further configured to determine whether the predicted value is less than a minimum value of a preset parameter; generating second early warning information when the predicted value is smaller than the minimum value of the preset parameter; and carrying out early warning prompt according to the second early warning information.
Other embodiments or specific implementation methods of the water quality prediction device based on space-time fusion according to the present invention may refer to the above method embodiments, and will not be described herein.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The foregoing embodiment numbers of the present invention are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in an estimator-readable storage medium (e.g. ROM/RAM, magnetic disk, optical disk) as described above, comprising instructions for causing an intelligent device (which may be a cell phone, estimator, space-time fusion based water quality prediction device, air conditioner, or network space-time fusion based water quality prediction device, etc.) to perform the method according to the embodiments of the present invention.
The foregoing description is only of the preferred embodiments of the present invention, and is not intended to limit the scope of the invention, but rather is intended to cover any equivalents of the structures or equivalent processes disclosed herein or in the alternative, which may be employed directly or indirectly in other related arts.
Claims (7)
1. The water quality prediction method based on space-time fusion is characterized by comprising the following steps of:
searching a water quality monitoring record corresponding to the seawater net cage, and extracting historical water quality parameter data from the water quality monitoring record;
acquiring current water quality parameter data acquired by a plurality of preset sensors in the seawater net cage;
taking the historical water quality parameter data as time water quality parameter data and taking the current water quality parameter data as space water quality parameter data;
determining target data according to the time water quality parameter data and the space water quality parameter data;
constructing a network prediction model based on a preset long-short-term memory network and a preset time convolution network;
inputting the target data into the network prediction model for training to obtain a predicted value, and taking the predicted value as a water quality predicted result;
The determining target data according to the time water quality parameter data and the space water quality parameter data specifically comprises:
taking the time water quality parameter data and the space water quality parameter data as water quality parameter data;
preprocessing the water quality parameter data to generate a time sequence diagram corresponding to the water quality parameter data;
judging whether the water quality parameter data is stable or not according to the time sequence diagram, and determining the period of original sequence data in the time sequence diagram;
when the water quality parameter data is not stable, extracting first water quality parameter data at the current moment and second water quality parameter data at the last period from the water quality parameter data;
determining target data according to the first water quality parameter data and the second water quality parameter data;
the determining target data according to the first water quality parameter data and the second water quality parameter data specifically includes:
obtaining a differential sequence according to the first water quality parameter data and the second water quality parameter data;
carrying out seasonal difference processing on the difference sequence to obtain a target sequence;
inputting the target sequence into a preset prediction model for training to obtain prediction data;
Carrying out seasonal differential inverse conversion on the predicted data to obtain target data;
inputting the target data into the network prediction model for training to obtain a predicted value, and taking the predicted value as a water quality prediction result, wherein the method specifically comprises the following steps of:
inputting the target data into the network prediction model, and extracting the spatial characteristics of the target data through a preset spatial attention mechanism and the preset long-short-time memory network;
extracting the time characteristics of the target data through a preset time attention mechanism and the preset time convolution network;
determining a spatial attention value according to the spatial features, and determining a temporal attention value according to the temporal features;
determining a predicted value from the spatial attention value and the temporal attention value;
and taking the predicted value as a water quality predicted result.
2. The water quality prediction method based on space-time fusion according to claim 1, wherein the determining a spatial attention value according to the spatial feature and determining a temporal attention value according to the temporal feature specifically comprises:
calculating a space weight coefficient of the space feature by using a vector dot product, and calculating a time weight coefficient of the time feature by using a similarity;
Respectively carrying out normalization processing on the space weight coefficient and the time weight coefficient through a preset function to obtain a target space weight coefficient and a target time weight coefficient;
carrying out weighted summation processing on the target space weight coefficient to obtain a space attention value;
and carrying out weighted summation processing on the target time weight coefficient to obtain a time attention value.
3. The water quality prediction method based on space-time fusion according to claim 1, wherein after the target data is input into the network prediction model for training to obtain a predicted value, and the predicted value is used as a water quality prediction result, the method further comprises:
judging whether the predicted value is larger than a maximum value of a preset parameter;
generating first early warning information when the predicted value is larger than the maximum value of the preset parameter;
and carrying out early warning prompt according to the first early warning information.
4. The water quality prediction method based on space-time fusion according to any one of claims 1 to 3, wherein the training is performed by inputting the target data into the network prediction model to obtain a predicted value, and after taking the predicted value as a water quality prediction result, the method further comprises:
Judging whether the predicted value is smaller than a minimum value of a preset parameter;
generating second early warning information when the predicted value is smaller than the minimum value of the preset parameter;
and carrying out early warning prompt according to the second early warning information.
5. A water quality prediction device based on space-time fusion, which is characterized in that the water quality prediction device based on space-time fusion comprises:
the historical data module is used for searching a water quality monitoring record corresponding to the seawater net cage and extracting historical water quality parameter data from the water quality monitoring record;
the current data module is used for acquiring current water quality parameter data acquired by a plurality of preset sensors in the seawater net cage;
the space-time data module is used for taking the historical water quality parameter data as time water quality parameter data and taking the current water quality parameter data as space water quality parameter data;
the target data module is used for determining target data according to the time water quality parameter data and the space water quality parameter data;
the model construction module is used for constructing a network prediction model based on a preset long-short-time memory network and a preset time convolution network;
the prediction result module is used for inputting the target data into the network prediction model for training to obtain a prediction value, and taking the prediction value as a water quality prediction result;
The target data module is further used for taking the time water quality parameter data and the space water quality parameter data as water quality parameter data;
preprocessing the water quality parameter data to generate a time sequence diagram corresponding to the water quality parameter data;
judging whether the water quality parameter data is stable or not according to the time sequence diagram, and determining the period of original sequence data in the time sequence diagram;
when the water quality parameter data is not stable, extracting first water quality parameter data at the current moment and second water quality parameter data at the last period from the water quality parameter data;
determining target data according to the first water quality parameter data and the second water quality parameter data;
the target data module is further used for obtaining a differential sequence according to the first water quality parameter data and the second water quality parameter data;
carrying out seasonal difference processing on the difference sequence to obtain a target sequence;
inputting the target sequence into a preset prediction model for training to obtain prediction data;
carrying out seasonal differential inverse conversion on the predicted data to obtain target data;
the prediction result module is further used for inputting the target data into the network prediction model, and extracting the spatial characteristics of the target data through a preset spatial attention mechanism and the preset long-short-time memory network;
Extracting the time characteristics of the target data through a preset time attention mechanism and the preset time convolution network;
determining a spatial attention value according to the spatial features, and determining a temporal attention value according to the temporal features;
determining a predicted value from the spatial attention value and the temporal attention value;
and taking the predicted value as a water quality predicted result.
6. A water quality prediction apparatus based on space-time fusion, characterized in that the water quality prediction apparatus based on space-time fusion comprises: a memory, a processor and a water quality prediction program based on space-time fusion stored on the memory and executable on the processor, the water quality prediction program based on space-time fusion being configured with steps to implement the water quality prediction method based on space-time fusion as claimed in any one of claims 1 to 4.
7. A storage medium having stored thereon a water quality prediction program based on space-time fusion, which when executed by a processor, implements the steps of the water quality prediction method based on space-time fusion according to any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010792866.6A CN111882138B (en) | 2020-08-07 | 2020-08-07 | Water quality prediction method, device, equipment and storage medium based on space-time fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010792866.6A CN111882138B (en) | 2020-08-07 | 2020-08-07 | Water quality prediction method, device, equipment and storage medium based on space-time fusion |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111882138A CN111882138A (en) | 2020-11-03 |
| CN111882138B true CN111882138B (en) | 2024-02-23 |
Family
ID=73211375
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010792866.6A Active CN111882138B (en) | 2020-08-07 | 2020-08-07 | Water quality prediction method, device, equipment and storage medium based on space-time fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111882138B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113067891B (en) * | 2021-04-08 | 2022-10-18 | 深圳市中瓴智慧科技有限公司 | Intelligent construction site data acquisition method and device, computer equipment and storage medium thereof |
| CN113834523B (en) * | 2021-09-06 | 2023-07-11 | 哈尔滨工业大学（威海） | Marine pasture intelligent breeding system based on unmanned ship |
| CN113822270B (en) * | 2021-11-24 | 2022-02-11 | 深圳大学 | Method and device for determining abnormal value of high-frequency data of structural health monitoring |
| 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 |
| CN114578011A (en) * | 2022-03-07 | 2022-06-03 | 上海蓝长科技集团有限公司 | Water quality monitoring method based on multi-sensor multi-source data fusion |
| CN114662788B (en) * | 2022-04-19 | 2022-11-29 | 广东海洋大学 | Seawater quality three-dimensional time-space sequence multi-parameter accurate prediction method and system |
| 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 (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103778482A (en) * | 2014-02-12 | 2014-05-07 | 中国农业大学 | Aquaculture dissolved oxygen short-term prediction method based on multi-scale analysis |
| CN105023071A (en) * | 2015-08-14 | 2015-11-04 | 中国科学院重庆绿色智能技术研究院 | Water quality prediction method based on Gaussian cloud transformation and fuzzy time sequence |
| CN107688871A (en) * | 2017-08-18 | 2018-02-13 | 中国农业大学 | A kind of water quality prediction method and device |
| CN108053054A (en) * | 2017-11-22 | 2018-05-18 | 中国农业大学 | A kind of water quality prediction of river method |
| CN108197736A (en) * | 2017-12-29 | 2018-06-22 | 北京工业大学 | A kind of Air Quality Forecast method based on variation self-encoding encoder and extreme learning machine |
| CN108235001A (en) * | 2018-01-29 | 2018-06-29 | 上海海洋大学 | A kind of deep-sea video quality objective assessment model based on space-time characteristic |
| CN108510741A (en) * | 2018-05-24 | 2018-09-07 | 浙江工业大学 | A kind of traffic flow forecasting method based on Conv1D-LSTM neural network structures |
| CN108614422A (en) * | 2018-05-23 | 2018-10-02 | 中国农业大学 | Dissolved oxygen optimal control method, apparatus and system in the cultivation of continental rise plant circulation water |
| CN108665100A (en) * | 2018-05-09 | 2018-10-16 | 中国农业大学 | A kind of water quality prediction technique, system and device |
| CN108717586A (en) * | 2018-05-10 | 2018-10-30 | 江南大学 | A kind of culture environment of aquatic products dissolved oxygen prediction method based on long memory network in short-term |
| CN109492830A (en) * | 2018-12-17 | 2019-03-19 | 杭州电子科技大学 | A kind of mobile pollution source concentration of emission prediction technique based on space-time deep learning |
| CN109522645A (en) * | 2018-11-15 | 2019-03-26 | 北京英视睿达科技有限公司 | Simulation of water quality method and system for river basins |
| CN110414027A (en) * | 2019-01-02 | 2019-11-05 | 徐宇迪 | A kind of data processing method suitable for flood forecasting system |
| CN110852515A (en) * | 2019-11-15 | 2020-02-28 | 北京工业大学 | Water quality index prediction method based on mixed long-time and short-time memory neural network |
| CN110889085A (en) * | 2019-09-30 | 2020-03-17 | 华南师范大学 | Intelligent wastewater monitoring method and system based on complex network multiple online regression |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA3072045A1 (en) * | 2017-08-02 | 2019-02-07 | Strong Force Iot Portfolio 2016, Llc | Methods and systems for detection in an industrial internet of things data collection environment with large data sets |
-
2020
- 2020-08-07 CN CN202010792866.6A patent/CN111882138B/en active Active
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103778482A (en) * | 2014-02-12 | 2014-05-07 | 中国农业大学 | Aquaculture dissolved oxygen short-term prediction method based on multi-scale analysis |
| CN105023071A (en) * | 2015-08-14 | 2015-11-04 | 中国科学院重庆绿色智能技术研究院 | Water quality prediction method based on Gaussian cloud transformation and fuzzy time sequence |
| CN107688871A (en) * | 2017-08-18 | 2018-02-13 | 中国农业大学 | A kind of water quality prediction method and device |
| CN108053054A (en) * | 2017-11-22 | 2018-05-18 | 中国农业大学 | A kind of water quality prediction of river method |
| CN108197736A (en) * | 2017-12-29 | 2018-06-22 | 北京工业大学 | A kind of Air Quality Forecast method based on variation self-encoding encoder and extreme learning machine |
| CN108235001A (en) * | 2018-01-29 | 2018-06-29 | 上海海洋大学 | A kind of deep-sea video quality objective assessment model based on space-time characteristic |
| CN108665100A (en) * | 2018-05-09 | 2018-10-16 | 中国农业大学 | A kind of water quality prediction technique, system and device |
| CN108717586A (en) * | 2018-05-10 | 2018-10-30 | 江南大学 | A kind of culture environment of aquatic products dissolved oxygen prediction method based on long memory network in short-term |
| CN108614422A (en) * | 2018-05-23 | 2018-10-02 | 中国农业大学 | Dissolved oxygen optimal control method, apparatus and system in the cultivation of continental rise plant circulation water |
| CN108510741A (en) * | 2018-05-24 | 2018-09-07 | 浙江工业大学 | A kind of traffic flow forecasting method based on Conv1D-LSTM neural network structures |
| CN109522645A (en) * | 2018-11-15 | 2019-03-26 | 北京英视睿达科技有限公司 | Simulation of water quality method and system for river basins |
| CN109492830A (en) * | 2018-12-17 | 2019-03-19 | 杭州电子科技大学 | A kind of mobile pollution source concentration of emission prediction technique based on space-time deep learning |
| CN110414027A (en) * | 2019-01-02 | 2019-11-05 | 徐宇迪 | A kind of data processing method suitable for flood forecasting system |
| CN110889085A (en) * | 2019-09-30 | 2020-03-17 | 华南师范大学 | Intelligent wastewater monitoring method and system based on complex network multiple online regression |
| CN110852515A (en) * | 2019-11-15 | 2020-02-28 | 北京工业大学 | Water quality index prediction method based on mixed long-time and short-time memory neural network |
Non-Patent Citations (5)
| Title |
|---|
| 基于LSTM模型的海洋水质预测;李彦杰;贺鹏飞;冯巍巍;刘巧利;杨信志;;计算机与数字工程;20200220(第02期);全文 * |
| 基于局部化双向LSTM和状态转移约束的养殖水质分类预测;商艳红;张静;;渔业现代化(第02期);全文 * |
| 基于长短时记忆网络的水质预测模型研究;秦文虎;陈溪莹;;安全与环境学报;20200225(第01期);全文 * |
| 工厂化水产养殖水质监测系统;肖瑞超;位耀光;他旭翔;张龙;;水产学杂志(第05期);全文 * |
| 深度学习方法在农业信息中的研究进展与应用现状;傅隆生;宋珍珍;Zhang Xin;李瑞;王东;崔永杰;;中国农业大学学报(第02期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111882138A (en) | 2020-11-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111882138B (en) | Water quality prediction method, device, equipment and storage medium based on space-time fusion | |
| Cianfrani et al. | Do habitat suitability models reliably predict the recovery areas of threatened species? | |
| Anderson | Real vs. artefactual absences in species distributions: tests for Oryzomys albigularis (Rodentia: Muridae) in Venezuela | |
| Matthiopoulos et al. | Using satellite telemetry and aerial counts to estimate space use by grey seals around the British Isles | |
| Smart et al. | Empirical realised niche models for British higher and lower plants–development and preliminary testing | |
| Punt et al. | How well can FMSY and BMSY be estimated using empirical measures of surplus production? | |
| Doerr et al. | Fractal analysis can explain individual variation in dispersal search paths | |
| Gusset et al. | Wild dog reintroductions in South Africa: a systematic review and cross-validation of an endangered species recovery programme | |
| Row et al. | Landscape characteristics influencing the genetic structure of greater sage‐grouse within the stronghold of their range: A holistic modeling approach | |
| Lahoz‐Monfort et al. | Exploring the consequences of reducing survey effort for detecting individual and temporal variability in survival | |
| Balestrieri et al. | Far from the madding crowd: Tolerance toward human disturbance shapes distribution and connectivity patterns of closely related Martes spp. | |
| Ranjani et al. | Crop yield prediction using machine learning algorithm | |
| Mentis | Monitoring in South African grasslands | |
| Henry et al. | Foraging behaviour of a frugivorous bat helps bridge landscape connectivity and ecological processes in a fragmented rainforest | |
| CN114242156A (en) | Real-time prediction method and system for relative abundance of pathogenic vibrios on marine micro-plastic | |
| Radomski et al. | An assessment model for a standard gill net incorporating direct and indirect selectivity applied to walleye | |
| Stephens et al. | Sustainable exploitation of social species: a test and comparison of models | |
| Illian et al. | Fitting a log Gaussian Cox process with temporally varying effects—a case study | |
| Vanausdall et al. | Detection and density of breeding marsh birds in Iowa wetlands | |
| Thompson et al. | Detecting seasonal episodic‐like spatio‐temporal memory patterns using animal movement modelling | |
| CN116307879A (en) | Efficient cultivation method, system and medium for penaeus monodon larvae | |
| Korňan et al. | Pairwise null model analyses of temporal patterns of bird assemblages contradict the assumptions of competition theory | |
| Brooks et al. | Quantifying conservation concern–Bayesian statistics, birds and the red lists | |
| Hu et al. | Moving‐resting process with measurement error in animal movement modeling | |
| Forshee et al. | Predator avoidance influences selection of neonatal lambing habitat by Sierra Nevada bighorn sheep |
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 |