CN116822333A - Lake area prediction method based on physical coupling deep learning - Google Patents
Lake area prediction method based on physical coupling deep learning Download PDFInfo
- Publication number
- CN116822333A CN116822333A CN202310612241.0A CN202310612241A CN116822333A CN 116822333 A CN116822333 A CN 116822333A CN 202310612241 A CN202310612241 A CN 202310612241A CN 116822333 A CN116822333 A CN 116822333A
- Authority
- CN
- China
- Prior art keywords
- water
- lake
- period
- pixels
- area
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 36
- 230000008878 coupling Effects 0.000 title claims abstract description 19
- 238000010168 coupling process Methods 0.000 title claims abstract description 19
- 238000005859 coupling reaction Methods 0.000 title claims abstract description 19
- 238000013135 deep learning Methods 0.000 title claims abstract description 19
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 137
- 230000000630 rising effect Effects 0.000 claims abstract description 10
- 230000008569 process Effects 0.000 claims abstract description 7
- 230000006870 function Effects 0.000 claims description 14
- 230000001502 supplementing effect Effects 0.000 claims description 3
- 239000011435 rock Substances 0.000 abstract description 7
- 238000004088 simulation Methods 0.000 abstract description 7
- 238000000605 extraction Methods 0.000 description 6
- 230000008859 change Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 238000012937 correction Methods 0.000 description 2
- 238000005520 cutting process Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 206010003694 Atrophy Diseases 0.000 description 1
- 241001092489 Potentilla Species 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000037444 atrophy Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000013277 forecasting method Methods 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 241000894007 species Species 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002194 synthesizing effect Effects 0.000 description 1
- 239000003643 water by type Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/04—Constraint-based CAD
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/08—Fluids
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
- G06F2218/14—Classification; Matching by matching peak patterns
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/30—Assessment of water resources
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Medical Informatics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention relates to the technical field of lake area prediction methods, and discloses a lake area prediction method based on physical coupling deep learning, which comprises the following steps: s1: based on a GEE platform; s2: identifying pixels with water signals stronger than vegetation signals by MNCWI > NDVI or MNCWI > EVI, and ensuring that vegetation pixels or pixels mixed by water and vegetation are removed by utilizing EVI < 0.1; s3: judging pixels meeting MNCWI > NDVI or MNCWI > EVI and EVI <0.1 as water bodies, otherwise, judging the pixels as non-water bodies; s4: multiplying the total number of pixels identified as the water body by the resolution to obtain the lake area value. The simulation result of the AWL-LSTM model on the peak value is better, the influence of the whole, the dead water period, the water rising period, the water enlarging period and the water withdrawing period on the simulation result is fully considered, the physical process reflecting the area of the Dongting lake and the water level of the urban rock is coupled into the model, the model is continuously guided and corrected through the back propagation of the physical constraint, and the simulation precision, the stability and the physical interpretability of the model are improved.
Description
Technical Field
The invention relates to the technical field of lake area prediction methods, in particular to a lake area prediction method based on physical coupling deep learning.
Background
Lakes are important ties for water circulation and play a very important role in regulating runoff, protecting ecological environment and protecting species diversity. But in recent years, most of lake areas in China have a tendency to shrink due to the deterioration of ecological environment and the influence of human activities. Therefore, the lake area is scientifically predicted, and the method has important significance for realizing reasonable allocation of water resources.
The existing lake area extraction prediction method is mainly divided into an extraction method based on remote sensing and an extraction method based on statistics. The extraction method based on remote sensing mainly adopts different water body identification methods (calculating water body indexes or vegetation indexes and the like according to different wave band combinations of remote sensing data) to identify water body pixels according to Landsat and other satellites, and then obtains the water body area according to the number of the water body pixels multiplied by the pixel resolution. If the monitoring classification method is successfully adopted, the water area change of the hole lake in 1996-2014 year is extracted from the land at remote sensing data, and the reason is analyzed, so that the result shows that the hole lake generally has a atrophy trend in year, and the human production activity is the most important indirect reason for the water area change of the hole lake. Ke Wenli and the like extract the water area of the Dongting lake in 2002-2012 by adopting Terra/MODIS L1B remote sensing number, and the result shows that the water area of the Dongting lake is in an overall reduction trend and has higher correlation with the water level of the urban rock. However, the extracted area data is missing and the reliability is not high because of the problems of different resolutions and blurred or missing remote sensing images of the remote sensing data.
The statistical method is mainly to train a model by adopting a machine learning or deep learning method, and fit the area according to the characteristics of flow, water level and the like. Deep learning methods based on deep v3+ such as Jiang Anxin obtain lake distribution data of areas >1km2 in 2003, 2008, 2013 and 2018 in Qinghai-Tibet plateau. However, the deep learning model lacks of physical interpretability due to the fitting characteristic of a black box, so that some simulation results do not conform to the physical rule, and the tracking and correction of errors are not facilitated.
In view of this, it is necessary to propose a lake area prediction method based on physical coupling deep learning, which scientifically predicts the lake area and realizes reasonable utilization of water resources.
Disclosure of Invention
The invention aims to provide a lake area prediction method based on physical coupling deep learning, which solves the problems in the background technology.
In order to achieve the above purpose, the present invention provides the following technical solutions: the lake area prediction method based on the physical coupling deep learning comprises the following steps:
s1: based on a GEE platform;
s2: identifying pixels with water signals stronger than vegetation signals by MNCWI > NDVI or MNCWI > EVI, and ensuring that vegetation pixels or pixels mixed by water and vegetation are removed by utilizing EVI < 0.1;
s3: judging pixels meeting MNCWI > NDVI or MNCWI > EVI and EVI <0.1 as water bodies, otherwise, judging the pixels as non-water bodies;
s4: multiplying the total number of pixels identified as the water body by the resolution to obtain a lake area value;
s5: supplementing the missing data by adopting linear interpolation to obtain a lake area data set;
s6: obtaining lake water level data;
s7: the correlation between the lake and the water level in four periods is established, and the fitting formula is as follows:
wherein Sn is lake area (km) fitted according to a correlation formula 2 ) The method comprises the steps of carrying out a first treatment on the surface of the L is water level data (m); a (1-5), b (1-5), c (1-5), d (1-5) e (1-5) are fitting coefficients of the whole data set, the dead water period, the flood period, the rich water period and the water withdrawal period respectively;
s8: and (3) establishing a lake-water level relation according to the step S7, and improving a loss function of the LSTM model, wherein the formula is as follows:
s9: according to the water level data in the step S6 as characteristic input, the area data in the step S5 as expected output, and an AWL-LSTM model is built;
s10: and judging the fitting degree according to the descending state of the loss function, and finally outputting the fitting area.
Preferably, the specific operation of the step S1 is to adopt improved normalized water body index, normalized vegetation index and enhanced vegetation index for water body identification.
Preferably, the specific formula of S2 is as follows:
wherein ρG is a green light band; ρSMIR is the short wave infrared band; ρNIR is near infrared band; ρR is the red band; ρb is the blue band.
Preferably, the four periods in step S7 are a dead water period, a flood period, a rich water period and a water withdrawal period.
Preferably, lossn in the step S8 is a loss function of the LSTM model of the coupled physical process; n, i are data numbers; si is lake area data set (km) 2 )；Is an area analog value (m); li is water level data; a (1-5), b (1-5), c (1-5), d (1-5) and e (1-5) are fitting coefficients of the whole data set, the dead water period, the water rising period, the water enlarging period and the water withdrawing period respectively.
The invention provides a lake area prediction method based on physical coupling deep learning. The lake area prediction method based on physical coupling deep learning has the following beneficial effects:
1. according to the lake area forecasting method based on physical coupling deep learning, a Google Earth Engine (GEE) platform is adopted, lake areas are extracted by integrating multiple indexes, the correlation between the lake areas and water levels is fitted, the correlation is between the lake areas and the water levels, and the correlation is used as a loss function of an improved LSTM model through physical constraint, an area forecasting model (AWL-LSTM model) of the physical coupling LSTM is built, parameters are fewer, the acquisition difficulty is low, the cost is low, and the fitting precision of the lake areas to peaks is higher due to the coupling physical relationship, the physical interpretability is higher, and the stability is higher.
2. The simulation result of the AWL-LSTM model on the peak value is better, the influence of the whole, the dead water period, the water rising period, the water enlarging period and the water withdrawing period on the simulation result is fully considered, the physical process reflecting the area of the Dongting lake and the water level of the urban rock is coupled into the model, the model is continuously guided and corrected through the back propagation of the physical constraint, and the simulation precision, the stability and the physical interpretability of the model are improved.
Drawings
FIG. 1 is a schematic illustration of the operational flow diagram of the present invention;
FIG. 2 is a schematic illustration of a study area of the present invention;
FIG. 3 is a schematic view of an area image (2016.06, 2016.08, 2019.10, 2019.12) of the present invention;
FIG. 4 is a schematic diagram of the fitting relationship between lake area and water level according to the present invention;
FIG. 5 is a schematic diagram showing the fitting relationship between lake area and water level according to the present invention;
FIG. 6 is a graph of the results of a coupled model fitting of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments.
Examples of the embodiments are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements throughout or elements having like or similar functionality. The embodiments described below by referring to the drawings are illustrative and intended to explain the present invention and should not be construed as limiting the invention.
The preferred embodiment of the lake area prediction method based on physical coupling deep learning provided by the invention is shown in fig. 1-6:
the lake area prediction method based on the physical coupling deep learning comprises the following steps:
s1: based on the GEE platform, adopting an improved normalized water body index, a normalized vegetation index and an enhanced vegetation index to perform water body identification;
s2: the MNDWI > NDVI or MNDWI > EVI is used for identifying pixels with water signals stronger than vegetation signals, and the EVI <0.1 is used for ensuring that vegetation pixels or pixels mixed by water and vegetation are removed, wherein the specific formula is as follows:
wherein ρG is a green light band; ρSMIR is the short wave infrared band; ρNIR is near infrared band; ρR is the red band; ρb is the blue band;
s3: judging pixels meeting MNCWI > NDVI or MNCWI > EVI and EVI <0.1 as water bodies, otherwise, judging the pixels as non-water bodies;
s4: multiplying the total number of pixels identified as the water body by the resolution to obtain a lake area value;
s5: supplementing the missing data by adopting linear interpolation to obtain a lake area data set;
s6: obtaining lake water level data;
s7: the method comprises the steps of establishing a correlation between a lake and a water level in four periods, namely a dead water period, a water rising period, a water enlarging period and a water withdrawing period, wherein a fitting formula is as follows:
wherein Sn is lake area (km) fitted according to a correlation formula 2 ) The method comprises the steps of carrying out a first treatment on the surface of the L is water level data (m); a (1-5), b (1-5), c (1-5), d (1-5) e (1-5) are fitting coefficients of the whole data set, the dead water period, the flood period, the rich water period and the water withdrawal period respectively;
s8: and (3) establishing a lake-water level relation according to the step S7, and improving a loss function of the LSTM model, wherein the formula is as follows:
wherein Lossn is a loss function of the LSTM model of the coupled physical process; n, i are data numbers; si is lake area data set (km) 2 )；Is an area analog value (m); li is water level data; a (1-5), b (1-5), c (1-5), d (1-5) and e (1-5) are fitting coefficients of the whole data set, the dead water period, the water rising period, the water enlarging period and the water withdrawing period respectively.
S9: according to the water level data in the step S6 as characteristic input, the area data in the step S5 as expected output, and an AWL-LSTM model is built;
s10: and judging the fitting degree according to the descending state of the loss function, and finally outputting the fitting area.
Practical examples:
in the invention, an eastern Dongting lake is taken as an example, the Dongting lake is positioned at the middle and downstream of the Yangtze river, and four waters of Xiangjiang river, resource water, yuanjiang river and water enter the lake in the south, and the Dongting lake is collected into the Yangtze river in the North through the urban rock. The main lake region consists of four parts of an east Dongting lake, a south Dongting lake, a west Dongting lake and a Datong lake, wherein the north latitude at the east Dongting lake is 28 degrees 59-29 degrees 38 degrees and the east longitude is 112 degrees 43-113 degrees 15', and the main lake region is positioned at the eastern part of the Dongting lake, is the most perfect lake in the Dongting lake region and is also an important regulation lake in the downstream region in the Yangtze river. The Dongting lake is positioned in a subtropical humid monsoon climate zone, the sunlight is sufficient, the annual change of the water fall is large, and the specific position is shown in figure 2;
s1, based on a GEE platform, carrying out water body identification by adopting an improved normalized water body index (MNDWI), a normalized vegetation index (NDVI) and an Enhanced Vegetation Index (EVI), extracting lake areas of the east Dongting lake 2000-2020, utilizing Landsat series data sets in a platform database as data sources, namely Landsat5 (2000-2012), landsat7 (2012-2014) and Landsat8 (2014-2020), respectively, wherein the spatial resolution is 30 meters, calculating the water body index by selecting a blue light wave band, a green light wave band, a red light wave band, a near infrared wave band and a short wave near infrared wave band, and cutting an extraction range through image splicing, research area cutting and other processes. The data image is subjected to radiometric calibration and atmospheric correction, and cloud removal processing is carried out on the data image by adopting a cloud removal function so as to ensure the accuracy of extracting the water body;
s2, identifying pixels with water signals stronger than vegetation signals by MNCWI > NDVI or MNCWI > EVI, and ensuring that vegetation pixels or pixels mixed by water and vegetation are removed by utilizing EVI <0.1, wherein MNCWI calculates by utilizing green light wave bands and short wave infrared wave bands in images, so that water characteristics can be enhanced, and meanwhile, noises of buildings, lands, vegetation and the like can be effectively restrained; NDVI utilizes near infrared wave band and red light wave band in the image to calculate, and can effectively reflect the growth dynamics of vegetation; the EVI is calculated by utilizing a near infrared wave band, a red light wave band and a blue light wave band in the image, so that the space difference of vegetation in a research area can be effectively reflected;
s3, judging pixels meeting (MNDWI > NDVI or MNDWI > EVI) and (EVI < 0.1) as a water body, otherwise, judging the pixels as a non-water body;
and S4, multiplying the total number of pixels identified as the water body by the resolution to obtain the lake area value. At length, only partial area extraction results are listed, as shown in fig. 3;
s5, linear interpolation is adopted to complement the missing data, and a lake area data set is obtained. Due to the influence of various factors such as equipment, weather and the like, partial remote sensing images are missing, 213 remote sensing images in 2000-2020 are obtained after final processing, and the missing data are complemented by linear interpolation to obtain 252 pieces of data in total;
s6, obtaining lake water level data, and according to investigation and research, the relationship between the area of the Dongting lake and the water level of the urban mountain rock station is the most intimate, so that the relationship between the area of the Dongting lake and the water level is analyzed by adopting the water level data of the 2000-2020 urban mountain rock station provided by the Yangtze river and Potentilla of the Wuhan city, and the consistency and the reliability can be ensured through the whole arrangement of the data by related departments.
S7, establishing correlation between the lake and the whole water level in four periods (the dead water period, the water rising period, the water enlarging period and the water withdrawing period). The overall dataset (1-12 months), dead water period (1-3 months), flood period (4-6 months), flood period (7-9 months) and water withdrawal period (10-12 months) were correlation fitted using the urban rock station water level data as independent variables and the eastern Dongting lake area data as dependent variables using python3.6 programming, as shown in FIG. 4. Synthesizing time complexity, space complexity and computer performance requirements, and respectively taking the following fitting results: y= -1.64x3+131.5x2-3372.97x+28252.77, y=25.9x-306.04, y=6.87x2-239.17x+2102.76, y= -5.19x2+369.13x-5267.96, y=12.92x2-521.62x+5495.38, with R2 of 0.87, 0.12, 0.89, 0.87 and 0.67, respectively.
It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions;
and 8, improving the loss function of the LSTM model according to the lake-water level relationship established in the step 7.
Wherein, loss n A loss function that is an LSTM model of the coupled physical process; n, i are data numbers; s is S i Is a lake area dataset (km 2);is an area analog value (m); l (L) i Is water level data; a, a (1-5) 、b (1-5) Fitting coefficients of the whole data set, the dead water period, the water rising period, the water enlarging period and the water withdrawing period are respectively obtained.
S9, taking the water level data of S6 as characteristic input, and taking the area data of S5 as expected output to establish an AWL-LSTM model. 80% (2000-2017) of the dataset served as the training set and 20% (2017-2020) served as the test set. The number of hidden layers is 3, the number of nodes in each layer is 32, the number of nodes in an output layer is 1, and the maximum training round is 3000;
and S10, judging the fitting degree according to the descending state of the loss function, and finally outputting the fitting area. R2 and the variation trend of the fitting result of the training set and the testing set are shown in figures 5 and 6; from fig. 5 and 6, the AWL-LSTM model can better reflect the area fluctuation change condition of the Dongting lake, and the R2 is 0.97.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (5)
1. The lake area prediction method based on physical coupling deep learning is characterized by comprising the following steps of: the method comprises the following steps:
s1: based on a GEE platform;
s2: identifying pixels with water signals stronger than vegetation signals by MNCWI > NDVI or MNCWI > EVI, and ensuring that vegetation pixels or pixels mixed by water and vegetation are removed by utilizing EVI < 0.1;
s3: judging pixels meeting MNCWI > NDVI or MNCWI > EVI and EVI <0.1 as water bodies, otherwise, judging the pixels as non-water bodies;
s4: multiplying the total number of pixels identified as the water body by the resolution to obtain a lake area value;
s5: supplementing the missing data by adopting linear interpolation to obtain a lake area data set;
s6: obtaining lake water level data;
s7: the correlation between the lake and the water level in four periods is established, and the fitting formula is as follows:
wherein S is n Lake area (km) fitted for correlation formula 2 ) The method comprises the steps of carrying out a first treatment on the surface of the L is water level data (m); a, a (1-5) 、b (1-5) 、c (1-5) 、、d (1-5) e (1-5) Fitting coefficients of the whole data set, the dead water period, the water rising period, the water enlarging period and the water withdrawing period are respectively obtained;
s8: and (3) establishing a lake-water level relation according to the step S7, and improving a loss function of the LSTM model, wherein the formula is as follows:
s9: according to the water level data in the step S6 as characteristic input, the area data in the step S5 as expected output, and an AWL-LSTM model is built;
s10: and judging the fitting degree according to the descending state of the loss function, and finally outputting the fitting area.
2. The lake area prediction method based on physical coupling deep learning of claim 1, wherein: s1, specifically, carrying out water body identification by adopting an improved normalized water body index, a normalized vegetation index and an enhanced vegetation index.
3. The lake area prediction method based on physical coupling deep learning of claim 1, wherein: the specific formula of S2 is as follows:
wherein ρ is G The green light wave band is adopted; ρ SMIR Is in a short wave infrared band; ρ NIR Is in the near infrared band; ρ R Is in a red light wave band; ρ B Is in the blue band.
4. The lake area prediction method based on physical coupling deep learning of claim 1, wherein: the four periods in the step S7 are a dead water period, a water rising period, a water enlarging period and a water withdrawing period.
5. The lake area prediction method based on physical coupling deep learning of claim 1, wherein: loss in S8 step n A loss function that is an LSTM model of the coupled physical process; n, i are data numbers; s is S i For lake area data sets (km) 2 )；Is an area analog value (m); l (L) i Is water level data; a, a (1-5) 、b (1-5) 、c (1-5) 、d (1-5) 、e (1-5) Fitting coefficients of the whole data set, the dead water period, the water rising period, the water enlarging period and the water withdrawing period are respectively obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310612241.0A CN116822333A (en) | 2023-05-29 | 2023-05-29 | Lake area prediction method based on physical coupling deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310612241.0A CN116822333A (en) | 2023-05-29 | 2023-05-29 | Lake area prediction method based on physical coupling deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116822333A true CN116822333A (en) | 2023-09-29 |
Family
ID=88125006
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310612241.0A Pending CN116822333A (en) | 2023-05-29 | 2023-05-29 | Lake area prediction method based on physical coupling deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116822333A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101114023A (en) * | 2007-08-28 | 2008-01-30 | 北京交通大学 | Lake and marshland flooding remote sense monitoring methods based on model |
| CN110685747A (en) * | 2019-08-27 | 2020-01-14 | 中国矿业大学(北京) | Remote sensing extraction method for coal mining subsidence water body of high diving space |
| CN110991705A (en) * | 2019-11-15 | 2020-04-10 | 广州地理研究所 | City expansion prediction method and system based on deep learning |
| CN113140000A (en) * | 2021-03-26 | 2021-07-20 | 中国科学院东北地理与农业生态研究所 | Water body information estimation method based on satellite spectrum |
| KR20220057740A (en) * | 2020-10-30 | 2022-05-09 | 홍익대학교 산학협력단 | Flood forecasting method using recurrent neural network |
-
2023
- 2023-05-29 CN CN202310612241.0A patent/CN116822333A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101114023A (en) * | 2007-08-28 | 2008-01-30 | 北京交通大学 | Lake and marshland flooding remote sense monitoring methods based on model |
| CN110685747A (en) * | 2019-08-27 | 2020-01-14 | 中国矿业大学(北京) | Remote sensing extraction method for coal mining subsidence water body of high diving space |
| CN110991705A (en) * | 2019-11-15 | 2020-04-10 | 广州地理研究所 | City expansion prediction method and system based on deep learning |
| KR20220057740A (en) * | 2020-10-30 | 2022-05-09 | 홍익대학교 산학협력단 | Flood forecasting method using recurrent neural network |
| CN113140000A (en) * | 2021-03-26 | 2021-07-20 | 中国科学院东北地理与农业生态研究所 | Water body information estimation method based on satellite spectrum |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109190538A (en) | A kind of fully-loaded stream river deposition coast band evolution analysis method based on remote sensing technology | |
| CN107247927B (en) | Method and system for extracting coastline information of remote sensing image based on tassel cap transformation | |
| CN106650812A (en) | City water body extraction method for satellite remote sensing image | |
| Zhang et al. | GWL_FCS30: global 30 m wetland map with fine classification system using multi-sourced and time-series remote sensing imagery in 2020 | |
| CN105243367A (en) | Method and device for monitoring scope of water body based on satellite remote sensing data | |
| CN102855494A (en) | Method and device for extracting water body of satellite remote sensing image | |
| CN114387516B (en) | Single-season rice SAR (synthetic aperture radar) identification method for small and medium-sized fields in complex terrain environment | |
| CN112435292A (en) | Rice planting area extraction method based on remote sensing image | |
| CN114581784B (en) | Construction method of long-time-sequence yearly mangrove remote sensing monitoring product | |
| Karna | Mapping above ground carbon using worldview satellite image and lidar data in relationship with tree diversity of forests | |
| CN112580982A (en) | Ecological protection red line implementation assessment based on multi-temporal remote sensing and CASA model | |
| CN112329265A (en) | Satellite remote sensing rainfall refinement space estimation method and system | |
| CN115294460B (en) | Method for determining degradation degree of phyllostachys praecox forest, medium and electronic device | |
| CN110569733B (en) | Lake long time sequence continuous water area change reconstruction method based on remote sensing big data platform | |
| Sui et al. | A high-resolution inland surface water body dataset for the tundra and boreal forests of North America | |
| CN102540169B (en) | Quality evaluation method for water body mapping product based on remote sensing image | |
| CN111879915B (en) | High-resolution monthly soil salinity monitoring method and system for coastal wetland | |
| Bektas Balcik et al. | Determination of magnitude and direction of land use/land cover changes in Terkos Water Basin, Istanbul | |
| CN111881797A (en) | Method, device, equipment and storage medium for finely classifying vegetation on coastal wetland | |
| CN116822333A (en) | Lake area prediction method based on physical coupling deep learning | |
| Liu et al. | Water extraction on the hyperspectral images of gaofen-5 satellite using spectral indices | |
| Xu et al. | Extraction of rivers and lakes on Tibetan Plateau based on Google Earth Engine | |
| Bao et al. | The study of lakes dynamic change based on RS and GIS–Take DaLiNuoEr Lake as an example | |
| CN117409313B (en) | Sentinel-2 optical image-based method for constructing index of spartina alterniflora in weathering period | |
| Thamaga | The impact of land use and land cover changes on wetland productivity and hydrological systems in the Limpopo transboundary river basin, South Africa |
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 |