CN111783987A - Farmland reference crop evapotranspiration prediction method based on improved BP neural network - Google Patents
Farmland reference crop evapotranspiration prediction method based on improved BP neural network Download PDFInfo
- Publication number
- CN111783987A CN111783987A CN202010675336.3A CN202010675336A CN111783987A CN 111783987 A CN111783987 A CN 111783987A CN 202010675336 A CN202010675336 A CN 202010675336A CN 111783987 A CN111783987 A CN 111783987A
- Authority
- CN
- China
- Prior art keywords
- neural network
- evapotranspiration
- farmland
- data
- improved
- 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 44
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 30
- 238000012549 training Methods 0.000 claims abstract description 28
- 238000012360 testing method Methods 0.000 claims abstract description 25
- 238000003062 neural network model Methods 0.000 claims abstract description 22
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 19
- 238000004364 calculation method Methods 0.000 claims abstract description 14
- 238000005259 measurement Methods 0.000 claims abstract description 14
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 210000002569 neuron Anatomy 0.000 claims description 27
- 230000008569 process Effects 0.000 claims description 17
- 239000011159 matrix material Substances 0.000 claims description 15
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Chemical class O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 14
- 230000005855 radiation Effects 0.000 claims description 7
- 239000002689 soil Substances 0.000 claims description 5
- 230000004907 flux Effects 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 229920006395 saturated elastomer Polymers 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 6
- 238000005457 optimization Methods 0.000 abstract 1
- 230000005068 transpiration Effects 0.000 description 14
- 238000012897 Levenberg–Marquardt algorithm Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 6
- 238000011160 research Methods 0.000 description 6
- 241000209140 Triticum Species 0.000 description 4
- 235000021307 Triticum Nutrition 0.000 description 4
- 240000008042 Zea mays Species 0.000 description 4
- 235000005824 Zea mays ssp. parviglumis Nutrition 0.000 description 4
- 235000002017 Zea mays subsp mays Nutrition 0.000 description 4
- 235000005822 corn Nutrition 0.000 description 4
- 238000002790 cross-validation Methods 0.000 description 3
- 230000002262 irrigation Effects 0.000 description 3
- 238000003973 irrigation Methods 0.000 description 3
- 238000012417 linear regression Methods 0.000 description 3
- 238000012546 transfer Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 208000005156 Dehydration Diseases 0.000 description 1
- 208000003098 Ganglion Cysts Diseases 0.000 description 1
- 244000068988 Glycine max Species 0.000 description 1
- 235000010469 Glycine max Nutrition 0.000 description 1
- 208000005400 Synovial Cyst Diseases 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013529 biological neural network Methods 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000001747 exhibiting effect Effects 0.000 description 1
- 230000003631 expected effect Effects 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 239000003621 irrigation water Substances 0.000 description 1
- 238000012067 mathematical method Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000005477 standard model Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
Images
Classifications
-
- 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
- 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
- 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
- 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—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Molecular Biology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Agronomy & Crop Science (AREA)
- Primary Health Care (AREA)
- Mining & Mineral Resources (AREA)
- Marine Sciences & Fisheries (AREA)
- Animal Husbandry (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a farmland reference crop evapotranspiration prediction method based on an improved BP neural network, which comprises the steps of obtaining meteorological data of a farmland reference crop growth environment according to weather forecast information, and calculating the farmland reference crop evapotranspiration by adopting a PM method; constructing a training set and a testing set, and preprocessing; adopting an LM algorithm to construct a BP neural network model considering rainfall factors, and carrying out training optimization; and predicting the evapotranspiration of the farmland reference crops according to the test set meteorological data by using the optimized BP neural network model. The method utilizes the LM algorithm to construct a BP neural network model considering rainfall factors, takes the evapotranspiration of the farmland reference crops as an actual measurement value to train and optimize the model, improves the convergence degree and the calculation speed of the model, and effectively constructs the nonlinear relation between the evapotranspiration of the farmland reference crops and the driving factors thereof, thereby obviously improving the prediction precision and the prediction effect of the evapotranspiration of the farmland reference crops.
Description
Technical Field
The invention belongs to the technical field of reference crop transpiration prediction, and particularly relates to a farmland reference crop transpiration prediction method based on an improved BP neural network.
Background
Reference crop transpiration is a process by which soil and crops demonstrate diffusion of moisture into the atmosphere through transpiration and transpiration, and is a key factor in moisture balance and irrigation scheduling. The calculation and prediction method not only becomes an important field for researching water circulation and water balance of a farmland ecosystem, but also has important functions in aspects of specifying a farmland irrigation system, configuring water and soil resources and the like. Therefore, accurate prediction of the reference crop transpiration is urgently needed in order to better manage the crop irrigation water consumption and improve the crop water utilization efficiency.
The reference crop transpiration can be measured directly by high-cost microclimate techniques based on energy balance and water vapor flux transfer methods (e.g., aerodynamic methods, radiation wave ratio), and the like. In a certain research area, the reference crop transpiration only changes along with the change of meteorological data, and is generally estimated by adopting mathematical models with meteorological data as independent variables, the mathematical methods are divided into physical models and empirical models, and a Penman Monteith 56(PM56) model modified in the physical models is widely used, but the models need a large amount of meteorological data, and the data cannot be obtained in some regions.
The reference crop transpiration is affected by various climatic factors, and is in a complex nonlinear relationship with meteorological data. Neural Networks (ANNs) are a mathematical model whose structure is inspired by biological neural networks and are well suited for modeling nonlinear processes. In the past decades, artificial neural networks have been widely used in the fields of system modeling, fault diagnosis and control, pattern recognition, financial forecasting, hydrology, and the like. By comparing the artificial neural network model with other traditional algorithms, most researches consider that the prediction effect of the artificial neural network method is reliable, and the expected effect can be achieved by using less meteorological data; fuzzy reasoning, radial basis networks, wavelet transforms and genetic algorithms are also used in conjunction with artificial neural networks to build better performing reference crop transpiration prediction models.
The FAO-56 recommended PM model is a standard model for predicting the transpiration of the reference crop, however, the application of the PM model is limited due to the lack of necessary meteorological data, and in addition, the complicated nonlinear relationship exists between the transpiration of the reference crop and the driving factors thereof, and the traditional statistical method cannot obtain satisfactory results. In each neural network, a BP model can be used for randomly constructing a nonlinear function, although some researches adopt the BP model to predict the transpiration of the reference crop at present, most of the existing researches train and test conventional input variables, uncertain rainfall conditions are not used as the input variables, rainfall is an important component in water balance, and the rainfall conditions are considered in semi-humid and humid areas to help to improve the prediction accuracy of the transpiration of the reference crop.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a farmland reference crop evapotranspiration prediction method based on an improved BP neural network under the condition of rainfall so as to improve the prediction precision of the farmland reference crop evapotranspiration.
In order to achieve the purpose of the invention, the invention adopts the technical scheme that:
a farmland reference crop evapotranspiration prediction method based on an improved BP neural network comprises the following steps:
s1, acquiring meteorological data of the growth environment of the farmland reference crops according to weather forecast information, and calculating the evapotranspiration of the farmland reference crops by adopting a PM method; the meteorological data comprise the highest air temperature, the lowest air temperature, sunshine hours and rainfall;
s2, respectively constructing a training set and a testing set according to the meteorological data and the farmland reference crop evapotranspiration acquired in the step S1, and preprocessing the data of the training set and the testing set;
s3, constructing a BP neural network model considering rainfall factors by adopting an LM algorithm, and training and optimizing the model by utilizing training set data;
and S4, predicting the evapotranspiration of the farmland reference crops according to the test set data by using the BP neural network model optimized in the step S3.
Further, the preprocessing of the training set and the test set meteorological data in the step S2 specifically includes:
adopting a hyperbolic tangent transformation function, and carrying out standardization processing on the meteorological data measured value according to the weights of the maximum value and the minimum value of the measured value in the training set data and the test set data, wherein the weights are expressed as follows:
wherein X' is the data measured value after the standardization process, X is the data measured value, Xmax、XminThe maximum and minimum values of the data measurements are respectively.
Further, the step S3 is specifically:
constructing a BP neural network model of a three-layer topological structure comprising an input layer, a hidden layer and an output layer, and setting that the input layer comprises 4 neurons, the hidden layer comprises 10 neurons and the output layer comprises 1 neuron in the BP neural network; and training the BP neural network model by adopting an LM algorithm and taking the error minimization of the model output value and the actually measured output value as a target.
Further, the LM algorithm optimizes the neuron weight for each iteration according to the error minimum value of the model output value and the measured output value in the error back propagation process, which is expressed as:
wn+1＝wn+Δw
wherein, wnFor the neuron weight of the nth iteration, Δ w is the adjustment matrix weight.
Further, the calculation method of the adjustment matrix weight is as follows:
Δw＝-[JT(w)J(w)+μI]-1J(w)e(w)
e(w)＝[e1(w),e2(w),……ei(w)]T
wherein, JT(w) is Jacobian transpose matrix, J (w) is Jacobian matrix, μ is Magazine parameter, I is identity matrix, e (w) is error vector.
Further, the convergence condition for optimizing the neuron weights by the LM algorithm is represented as:
wherein, tiFor actually measured output value, aiThe model output value.
Further, the calculation formula of the measured output value is as follows:
wherein, ET0For reference to crop evapotranspiration, delta is the slope of the saturated water-steam pressure curve, Rn is the net surface radiation, G is the soil heat flux, gamma is the dry-wet constant, TmeanIs the average daily temperature u2To set the altitude position wind speed, esSaturated water vapor pressure, eaThe actual water vapor pressure.
The invention has the following beneficial effects:
the method utilizes the LM algorithm to construct a BP neural network model considering rainfall factors, utilizes the farmland reference crop evapotranspiration calculated by PM as an actual measurement value to train and optimize the model, improves the convergence degree and the calculation speed of the model, and effectively constructs the nonlinear relation between the farmland reference crop evapotranspiration and the driving factors thereof, thereby obviously improving the prediction precision and the prediction effect of the farmland reference crop evapotranspiration.
Drawings
FIG. 1 is a flow chart of a farmland reference crop evapotranspiration prediction method based on an improved BP neural network;
FIG. 2 is a graph comparing the predicted results of the multiple linear regression model of the present invention with those of the present invention;
FIG. 3 is a schematic diagram illustrating a cross-validation result of a predicted value and an actual measurement value of a multiple linear regression model according to an embodiment of the present invention;
fig. 4 is a schematic diagram illustrating a cross-validation result of a predicted value and an actual value according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate the understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the invention as defined and defined in the appended claims, and all matters produced by the invention using the inventive concept are protected.
As shown in fig. 1, an embodiment of the present invention provides a method for predicting an evapotranspiration of a farmland reference crop based on an improved BP neural network, including the following steps S1 to S4:
s1, acquiring meteorological data of a farmland reference crop growth environment according to weather forecast information, wherein the meteorological data comprise the highest air temperature, the lowest air temperature, sunshine hours and rainfall;
in this embodiment, the present invention uses the great happy area of Beijing as the research area to explain the related data. The great-rise district (39 degrees 26 '-39 degrees 51' N, 116 degrees 13 '-116 degrees 43' E) in Beijing City is positioned in the eternal river strike plain in the north of the North China plain, the total area is 1031km2, the district belongs to temperate zone semi-humid monsoon climate, and the average temperature of many years is 12.1 ℃. The average rainfall in many years is 540mm, the rainfall is more in 7 months and 9 months, and the rainfall accounts for more than 80% of the total rainfall in all years. The bottom surface of a great-rise test station in Beijing city is mainly a farmland, and comprises corn/wheat and soybean, wherein the corn and wheat are mainly used, the whole growth period of winter wheat is about 260 days (10 months 1 day-6 months 30 days in the next year), and the winter wheat needs to be supplemented with irrigation in normal years so as to ensure the requirement of crops on water. The growth period of summer corn is about 90 days (7 months 1 day-9 months 30 days), and water is not added into the summer corn in the whole growth period. No water stress exists in the growth stage of the crops in the research area. Compared with the great-rise area, the climate and the underlying surface condition of the test station are analyzed, and the test station has better typicality.
According to the invention, historical meteorological data of 2015-2017 years are collected in a China meteorological science sharing service network, and 1 d-7 d forecast period day-by-day weather forecast data of 2018-2019 years are collected in a weather network. The historical meteorological data includes: air pressure PaWind speed U and maximum air temperature TmaxMinimum air temperature TminAverage relative humidity RH, sunshine hours n, rainfall P, and the like. Weather forecast data and information include: maximum air temperature TmaxMinimum air temperature TminWeather conditions, etc.
The rainfall total grade is estimated to be 11 grades according to the rainfall grade table according to various phenomena caused by the influence of rainfall on the ground. And then determining the rainfall value according to the rainfall forecast information of the weather forecast. The rainfall rating table is shown in table 1.
TABLE 1 rainfall rating table
The number of hours of sunshine in the present invention means that the intensity of the radiation of the sun on a plane perpendicular to its rays per day exceeds or equals 120w/m2The length of time of (c). Radiation R per day of the year according to regional geographical location parametersaCan be calculated by the solar constant, the solar inclination angle and the like, and the calculation formula is as follows:
wherein R isaFor radiation, c is the speed of light, GscIs the sun constant, drThe relative distance between the sun and the earth, W is the sun inclination angle, h is the local latitude, and radian units are adopted; k is a radical ofsThe sunset hour angle.
S2, respectively constructing a training set and a testing set according to the meteorological data and the reference crop evapotranspiration acquired in the step S1, and preprocessing the data of the training set and the testing set;
in this embodiment, since a part of the meteorological data in 2015 to 2019 acquired in step S1 may be absent or abnormal due to environmental interference or human operation, 1800 sets of data sets are finally obtained after the acquired meteorological data is screened, and 400 sets of data are respectively randomly selected as a training set and a test set.
Because the dimension of the data in the data set is different and the magnitude is larger, the invention needs to preprocess the data in the training set and the test set; specifically, the magnitude of the data of the training set and the test set is standardized, so that the over-training is avoided, the convergence degree and the calculation speed of the ganglion points of certain layers are improved, and the calculation precision is improved.
The invention adopts a hyperbolic tangent transformation function, and carries out standardization processing on data measurement values according to the weights of the maximum value and the minimum value of the measurement values in the data of a training set and a test set, wherein the weights are expressed as follows:
wherein X' is the data measured value after the standardization process, X is the data measured value, Xmax、XminThe maximum and minimum values of the data measurements are respectively.
By normalizing the meteorological data measurements, the meteorological data measurements can be normalized to be in the range of [ -1,1], thereby exhibiting the most nonlinear characteristics.
S3, constructing a BP neural network model considering rainfall factors by adopting an LM algorithm, and training and optimizing the model by utilizing training gas-collecting image data;
in the embodiment, a BP neural network model with a three-layer topological structure comprising an input layer, a hidden layer and an output layer is constructed to predict the evapotranspiration of the farmland reference crops; each of which is composed of several interconnected neurons.
All neurons in the upper layer of the BP neural network realize full connection to all neurons in the lower layer through a transfer function, and no association exists between neurons in the same layer. When the learning samples are provided to the neural network, the neural network first performs a forward propagation process. If the error between the output and the target output exceeds the expectation, the forward propagation process is switched to a backward propagation process, the error signal is returned along the original connecting path, and the error signal is reduced by modifying the weight of each layer of neuron. With the continuous correction of the error back propagation, the accuracy of the network to the input mode response is continuously improved, and finally the applicable precision is achieved.
Since the number of neurons in input and output is determined by the target and the connection is fixed, the structure depends mainly on the number of hidden layer neurons. Too few hidden nodes may affect the functionality of the network, while too many hidden nodes may result in an over-adaptation of the network to the data. Therefore, the optimal BP neural network architecture 4-10-1 is set in the invention, namely an input layer in the BP neural network comprises 4 neurons, a hidden layer comprises 10 neurons and an output layer comprises 1 neuron; and the learning rate and iteration of training are set to 0.1 and 5000, respectively.
The learning of the neural network is to adjust the weight of the neuron according to the error between the actual output value and the target output value. The change in the weight is proportional to the negative value of the error derivative. During the network training process, error minimization is achieved through repeated iteration. In the back propagation training algorithm, the neuron processing of the hidden layer and the output layer and the input thereof need to multiply each input by the weight value thereof, sum the products, and then process the sum by using a nonlinear transfer function.
Although the traditional BP algorithm adopts a gradient descent algorithm to determine the weights in the network, the calculation speed of the method is slow due to linear convergence. Therefore, the invention adopts a Levenberg-Marquardt algorithm (LM) to train the BP neural network model.
The LM algorithm consists of an information feedforward process and an error back propagation process. In the error back propagation process, optimizing the neuron weight of each iteration according to the error minimum value of the model output value and the measured output value, wherein the error minimum value is expressed as:
wn+1＝wn+Δw
wherein, wnFor the neuron weight of the nth iteration, Δ w is the adjustment matrix weight.
The calculation method of the adjustment matrix weight is as follows:
Δw＝-[JT(w)J(w)+μI]-1J(w)e(w)
e(w)＝[e1(w),e2(w),……ei(w)]T
wherein, JT(w) is Jacobian transpose matrix, J (w) is Jacobian matrix, μ is Magazine parameter, I is identity matrix, e (w) is error vector.
The error minimization process is repeated until an acceptable convergence criterion is reached. The convergence condition for optimizing the weight of the neuron by setting the LM algorithm is expressed as follows:
wherein, tiFor actually measured output value, aiThe model output value.
The invention takes the farmland reference crop evapotranspiration calculated by a Penman-Monteith (PM) model recommended by FAO-56 as an actual measurement output value, and the calculation formula is as follows:
wherein, ET0For reference to crop evapotranspiration, delta is the slope of the saturated water-steam pressure curve, Rn is the net surface radiation, G is the soil heat flux, gamma is the dry-wet constant, TmeanIs the average daily temperature u2To set the altitude position wind speed, esSaturated water vapor pressure, eaThe actual water vapor pressure.
And S4, predicting the evapotranspiration of the farmland reference crops according to the test set meteorological data by using the BP neural network model optimized in the step S3.
In order to verify the prediction effect of the rainfall improvement BP neural network model, the rainfall improvement BP neural network model and the multiple linear regression model are consideredComparing the patterns (MLR) and using the coefficient of determination (R)2) The fit accuracy (Acc), the Root Mean Square Error (RMSE) and the Relative Error (RE) to evaluate the performance of the model.
The calculation formula of the evaluation parameters is as follows:
wherein x isiFor a prediction value of evapotranspiration of a farmland reference crop, yiIs an actual measurement output value of the evapotranspiration of the farmland reference crop, i is a prediction sample sequence,
is the average of the predicted value and the measured output value sequence, and n is the sample number of the predicted value.
As shown in fig. 2, the predicted value and the measured value of the improved BP neural network model considering rainfall according to the present invention have a high fitting effect.
To determine the fit performance of the test data set, cross-validation analysis of calculated and simulated values was performed on the test data set. As shown in FIGS. 3 and 4, the decision coefficient (R) of the improved BP neural network model considering rainfall according to the present invention2) And the fitting precision (Acc) are respectively higher than 0.89 and 90 percent, compared with an MLR model, the rainfall-considered improved BP neural network model has the prediction effect obviously better than that of the MLR model, and the ET can be more effectively constructed by the method0And its complex non-linear relationship between the driving factors.
In conclusion, the improved BP neural network model considering rainfall is higher in prediction accuracy than a traditional multiple regression model, can capture any complex and nonlinear input-output relationship from data without any restrictive assumption on the functional form of a bottom layer process, and considers the uncertain influence of rainfall on the evapotranspiration of farmland reference crops, so that the consistency between the predicted value and the measured value of the evapotranspiration of the farmland reference crops is improved, the prediction accuracy of the evapotranspiration of the farmland reference crops is greatly improved, and the development of ecological hydrology science is further promoted.
It will be appreciated by those of ordinary skill in the art that the embodiments described herein are intended to assist the reader in understanding the principles of the invention and are to be construed as being without limitation to such specifically recited embodiments and examples. Those skilled in the art can make various other specific changes and combinations based on the teachings of the present invention without departing from the spirit of the invention, and these changes and combinations are within the scope of the invention.
Claims (7)
1. A farmland reference crop evapotranspiration prediction method based on an improved BP neural network is characterized by comprising the following steps:
s1, acquiring meteorological data of the growth environment of the farmland reference crops according to weather forecast information, and calculating the evapotranspiration of the farmland reference crops by adopting a PM method; the meteorological data comprise the highest air temperature, the lowest air temperature, sunshine hours and rainfall;
s2, respectively constructing a training set and a testing set according to the meteorological data and the farmland reference crop evapotranspiration acquired in the step S1, and preprocessing the data of the training set and the testing set;
s3, constructing a BP neural network model considering rainfall factors by adopting an LM algorithm, and training and optimizing the model by utilizing training set data;
and S4, predicting the evapotranspiration of the farmland reference crops according to the test set meteorological data by using the BP neural network model optimized in the step S3.
2. The method for predicting the evapotranspiration of farmland reference crops based on the improved BP neural network as claimed in claim 1, wherein the preprocessing of the training set and the test set data in the step S2 is specifically as follows:
adopting a hyperbolic tangent transformation function, and carrying out standardization processing on the data measurement values according to the weights of the maximum value and the minimum value of the measurement values in the data of the training set and the test set, wherein the weights are expressed as follows:
wherein X' is the data measured value after the standardization process, X is the data measured value, Xmax、XminThe maximum and minimum values of the data measurements are respectively.
3. The method for predicting the evapotranspiration of the farmland reference crops based on the improved BP neural network as claimed in claim 2, wherein the step S3 is specifically as follows:
constructing a BP neural network model of a three-layer topological structure comprising an input layer, a hidden layer and an output layer, and setting that the input layer comprises 4 neurons, the hidden layer comprises 10 neurons and the output layer comprises 1 neuron in the BP neural network; and training the BP neural network model by adopting an LM algorithm and taking the error minimization of the model output value and the actually measured output value as a target.
4. The improved BP neural network-based farmland reference crop evapotranspiration prediction method of claim 3, wherein the LM algorithm optimizes the neuron weights for each iteration according to the error minimum value of the model output value and the expected output value in the error back propagation process, and is expressed as:
wn+1＝wn+Δw
wherein, wnFor the neuron weight of the nth iteration, Δ w is the adjustment matrix weight.
5. The farmland reference crop evapotranspiration prediction method based on the improved BP neural network as claimed in claim 4, wherein the calculation mode of the adjustment matrix weight is as follows:
Δw＝-[JT(w)J(w)+μI]-1J(w)e(w)
e(w)＝[e1(w)，e2(w)，……ei(w)]T
wherein, JT(w) is Jacobian transpose matrix, J (w) is Jacobian matrix, μ is Magazine parameter, I is identity matrix, e (w) is error vector.
6. The improved BP neural network-based farmland reference crop evapotranspiration prediction method of claim 5, wherein the convergence condition for optimizing the neuron weight by the LM algorithm is represented as:
wherein, tiFor actually measured output value, aiThe model output value.
7. The improved BP neural network-based farmland reference crop evapotranspiration prediction method of claim 6, wherein the calculation formula of the measured output value is as follows:
wherein, ET0For reference to crop evapotranspiration, delta is the slope of the saturated water-steam pressure curve, Rn is the net surface radiation, G is the soil heat flux, gamma is the dry-wet constant, TmeanIs the average daily temperature u2To set the altitude position wind speed, esSaturated water vapor pressure, eaThe actual water vapor pressure.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010675336.3A CN111783987A (en) | 2020-07-14 | 2020-07-14 | Farmland reference crop evapotranspiration prediction method based on improved BP neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010675336.3A CN111783987A (en) | 2020-07-14 | 2020-07-14 | Farmland reference crop evapotranspiration prediction method based on improved BP neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111783987A true CN111783987A (en) | 2020-10-16 |
Family
ID=72767708
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010675336.3A Pending CN111783987A (en) | 2020-07-14 | 2020-07-14 | Farmland reference crop evapotranspiration prediction method based on improved BP neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111783987A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112785096A (en) * | 2021-03-17 | 2021-05-11 | 郑州大学 | PCA and L-M algorithm-based reference crop evapotranspiration amount calculation method |
| CN112801364A (en) * | 2021-01-26 | 2021-05-14 | 昆明理工大学 | Reference crop evapotranspiration amount prediction method based on MIV and improved bat algorithm |
| CN112966399A (en) * | 2021-04-15 | 2021-06-15 | 苏州大学张家港工业技术研究院 | Pulse tube refrigerator working condition prediction method and system based on machine learning |
| CN113016574A (en) * | 2021-03-05 | 2021-06-25 | 武汉理工大学 | Street lamp with water conservation afforestation irrigation function |
| CN113159439A (en) * | 2021-04-30 | 2021-07-23 | 兰州里丰正维智能科技有限公司 | Crop yield prediction method and system, storage medium and electronic equipment |
| CN113962369A (en) * | 2021-11-29 | 2022-01-21 | 北京工业大学 | Radial basis function neural network optimization method based on improved Levenberg-Marquardt |
| CN114021486A (en) * | 2022-01-05 | 2022-02-08 | 中化现代农业有限公司 | Crop evapotranspiration prediction method and device |
| CN114169503A (en) * | 2021-12-09 | 2022-03-11 | 中国民航大学 | Method for accurately predicting landing distance of airplane |
| CN114186478A (en) * | 2021-11-15 | 2022-03-15 | 北京奥信化工科技发展有限责任公司 | Blasting block degree prediction method based on RBF neural network |
| CN114943361A (en) * | 2022-03-15 | 2022-08-26 | 水利部交通运输部国家能源局南京水利科学研究院 | Method for estimating evapotranspiration of reference crops in data-lacking areas |
| CN115310370A (en) * | 2022-10-09 | 2022-11-08 | 南开大学 | Regional vegetation transpiration prediction method coupling deep learning and physical mechanism |
| CN117636056A (en) * | 2023-12-13 | 2024-03-01 | 中南大学 | Agricultural information monitoring method and system based on big data |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5870302A (en) * | 1994-02-17 | 1999-02-09 | Waterlink Systems, Inc. | Evapotranspiration remote irrigation control system |
| CN101419219A (en) * | 2008-12-09 | 2009-04-29 | 中国农业科学院农业资源与农业区划研究所 | Method for determining evapotranspiration rate of referential crops |
| CN103645295A (en) * | 2013-12-03 | 2014-03-19 | 中国科学院遥感与数字地球研究所 | Multilayer soil moisture simulation method and multilayer soil moisture simulation system |
| CN108077042A (en) * | 2017-12-04 | 2018-05-29 | 北京农业智能装备技术研究中心 | A kind of winter wheat time of infertility irrigates early warning decision method |
| US20190050948A1 (en) * | 2017-08-08 | 2019-02-14 | Indigo Ag, Inc. | Machine learning in agricultural planting, growing, and harvesting contexts |
| CN109460925A (en) * | 2018-11-14 | 2019-03-12 | 江苏电力信息技术有限公司 | A kind of platform area group of assets Performance Evaluation Methods based on BP neural network |
| CN110728401A (en) * | 2019-10-10 | 2020-01-24 | 郑州轻工业学院 | Short-term power load prediction method of neural network based on squirrel and weed hybrid algorithm |
-
2020
- 2020-07-14 CN CN202010675336.3A patent/CN111783987A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5870302A (en) * | 1994-02-17 | 1999-02-09 | Waterlink Systems, Inc. | Evapotranspiration remote irrigation control system |
| CN101419219A (en) * | 2008-12-09 | 2009-04-29 | 中国农业科学院农业资源与农业区划研究所 | Method for determining evapotranspiration rate of referential crops |
| CN103645295A (en) * | 2013-12-03 | 2014-03-19 | 中国科学院遥感与数字地球研究所 | Multilayer soil moisture simulation method and multilayer soil moisture simulation system |
| US20190050948A1 (en) * | 2017-08-08 | 2019-02-14 | Indigo Ag, Inc. | Machine learning in agricultural planting, growing, and harvesting contexts |
| CN108077042A (en) * | 2017-12-04 | 2018-05-29 | 北京农业智能装备技术研究中心 | A kind of winter wheat time of infertility irrigates early warning decision method |
| CN109460925A (en) * | 2018-11-14 | 2019-03-12 | 江苏电力信息技术有限公司 | A kind of platform area group of assets Performance Evaluation Methods based on BP neural network |
| CN110728401A (en) * | 2019-10-10 | 2020-01-24 | 郑州轻工业学院 | Short-term power load prediction method of neural network based on squirrel and weed hybrid algorithm |
Non-Patent Citations (3)
| Title |
|---|
| SAMAN MAROUFPOOR等: "Reference evapotranspiration estimating based on optimal input combination and hybrid artificial intelligent model: Hybridization of artificial neural network with grey wolf optimizer algorithm", 《JOURNAL OF HYDROLOGY》, vol. 588, pages 1 - 2 * |
| 郑重等: "基于BP神经网络的农田蒸散量预报模型", 《水利学报》, vol. 39, no. 2, pages 231 - 232 * |
| 霍再林等: "参考作物潜在蒸散量的人工神经网络模型研究", 《农业工程学报》, vol. 20, no. 1, pages 40 - 42 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112801364A (en) * | 2021-01-26 | 2021-05-14 | 昆明理工大学 | Reference crop evapotranspiration amount prediction method based on MIV and improved bat algorithm |
| CN113016574A (en) * | 2021-03-05 | 2021-06-25 | 武汉理工大学 | Street lamp with water conservation afforestation irrigation function |
| CN113016574B (en) * | 2021-03-05 | 2022-07-08 | 武汉理工大学 | Street lamp with water conservation afforestation irrigation function |
| CN112785096A (en) * | 2021-03-17 | 2021-05-11 | 郑州大学 | PCA and L-M algorithm-based reference crop evapotranspiration amount calculation method |
| CN112966399A (en) * | 2021-04-15 | 2021-06-15 | 苏州大学张家港工业技术研究院 | Pulse tube refrigerator working condition prediction method and system based on machine learning |
| CN112966399B (en) * | 2021-04-15 | 2023-08-22 | 苏州大学张家港工业技术研究院 | Pulse tube refrigerator working condition prediction method and system based on machine learning |
| CN113159439A (en) * | 2021-04-30 | 2021-07-23 | 兰州里丰正维智能科技有限公司 | Crop yield prediction method and system, storage medium and electronic equipment |
| CN114186478A (en) * | 2021-11-15 | 2022-03-15 | 北京奥信化工科技发展有限责任公司 | Blasting block degree prediction method based on RBF neural network |
| CN114186478B (en) * | 2021-11-15 | 2023-01-17 | 北京奥信化工科技发展有限责任公司 | Blasting block degree prediction method based on RBF neural network |
| CN113962369A (en) * | 2021-11-29 | 2022-01-21 | 北京工业大学 | Radial basis function neural network optimization method based on improved Levenberg-Marquardt |
| CN113962369B (en) * | 2021-11-29 | 2024-05-31 | 北京工业大学 | Radial basis function neural network optimization method based on improved Levenberg-Marquardt |
| CN114169503A (en) * | 2021-12-09 | 2022-03-11 | 中国民航大学 | Method for accurately predicting landing distance of airplane |
| CN114021486A (en) * | 2022-01-05 | 2022-02-08 | 中化现代农业有限公司 | Crop evapotranspiration prediction method and device |
| CN114943361A (en) * | 2022-03-15 | 2022-08-26 | 水利部交通运输部国家能源局南京水利科学研究院 | Method for estimating evapotranspiration of reference crops in data-lacking areas |
| CN115310370A (en) * | 2022-10-09 | 2022-11-08 | 南开大学 | Regional vegetation transpiration prediction method coupling deep learning and physical mechanism |
| CN115310370B (en) * | 2022-10-09 | 2023-08-29 | 南开大学 | Regional vegetation transpiration prediction method coupled with deep learning and physical mechanism |
| CN117636056A (en) * | 2023-12-13 | 2024-03-01 | 中南大学 | Agricultural information monitoring method and system based on big data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111783987A (en) | Farmland reference crop evapotranspiration prediction method based on improved BP neural network | |
| CN111833202B (en) | Farmland evapotranspiration short-term prediction method considering crop coefficient dynamic change and rainfall | |
| CN111310889B (en) | Evaporation waveguide profile estimation method based on deep neural network | |
| CN112288164B (en) | Wind power combined prediction method considering spatial correlation and correcting numerical weather forecast | |
| CN111080173B (en) | Estimation method of carbon flux of forest system | |
| CN112287294B (en) | Space-time bidirectional soil water content interpolation method based on deep learning | |
| CN113255972B (en) | Short-term rainfall prediction method based on Attention mechanism | |
| CN114638146A (en) | Crop irrigation water demand prediction method based on AquaCrop model and SVR | |
| Çakır et al. | Yield prediction of wheat in south-east region of Turkey by using artificial neural networks | |
| CN114154401B (en) | Soil erosion modulus calculation method and system based on machine learning and observation data | |
| Bülbül et al. | Optimization, modeling and implementation of plant water consumption control using genetic algorithm and artificial neural network in a hybrid structure | |
| CN108399470A (en) | A kind of indoor PM2.5 prediction techniques based on more example genetic neural networks | |
| CN114372631A (en) | Data-lacking area runoff prediction method based on small sample learning and LSTM | |
| Taki et al. | Application of Neural Networks and multiple regression models in greenhouse climate estimation | |
| CN114529097A (en) | Multi-scale crop phenological period remote sensing dimensionality reduction prediction method | |
| CN115796040A (en) | Facility greenhouse temperature prediction method based on small samples | |
| CN109543911B (en) | Sunlight radiation prediction method and system | |
| CN104361399A (en) | Solar irradiation intensity minute-scale predication method | |
| CN117494419A (en) | Multi-model coupling drainage basin soil erosion remote sensing monitoring method | |
| CN115907204A (en) | Forest transpiration water consumption prediction method for optimizing BP neural network by sparrow search algorithm | |
| Adeyemo | Soft Computing techniques for weather and Climate change studies | |
| Morid et al. | Solar radiation estimation using temperature-based, stochastic and artificial neural networks approaches | |
| Üçler et al. | Estimating daily pan evaporation data using adaptive neuro fuzzy inference system: Case study within Van local station-Turkey | |
| CN112784477A (en) | WRF-LES and BP-PSO-Bagging combined wind power prediction method | |
| Hu et al. | Research on Reference Crop Evapotranspiration Forecast Based on FOA-GRNN |
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 |