CN109934400B - Rain collecting, regulating and deficiency crop water demand prediction method based on improved neural network - Google Patents
Rain collecting, regulating and deficiency crop water demand prediction method based on improved neural network Download PDFInfo
- Publication number
- CN109934400B CN109934400B CN201910173728.7A CN201910173728A CN109934400B CN 109934400 B CN109934400 B CN 109934400B CN 201910173728 A CN201910173728 A CN 201910173728A CN 109934400 B CN109934400 B CN 109934400B
- Authority
- CN
- China
- Prior art keywords
- factors
- crop
- neural network
- water demand
- stage
- 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
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A20/00—Water conservation; Efficient water supply; Efficient water use
- Y02A20/152—Water filtration
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a rain-collecting and deficiency-regulating crop water demand prediction method based on an improved Elman neural network. The prediction model of the invention can effectively predict the water demand of crops, and can regulate deficiency and drip irrigation while fully utilizing rainfall, thereby realizing multiple water conservation and saving water resources to a greater extent.
Description
Technical Field
The invention relates to a rain collecting, regulating and deficiency crop water demand prediction method based on an improved Elman neural network, and belongs to the technical field of water resource management and communication networks.
Background
China is a large agricultural country with serious shortage of water resources. The lack of water resources has become a bottleneck factor restricting sustainable development in China. In each water sector, agriculture is the largest industry of water resource consumption, water is dominant, irrigation water consumption is about 90% of agricultural water consumption in arid areas, and irrigation water consumption in China is about 80% of total water consumption. According to the prediction of expert, the agricultural water shortage in China reaches 1000X 108m around 2020 3 The water shortage will be more severe. Therefore, the popularization of the water-saving irrigation technology, the implementation of planning water and intelligent irrigation are very urgent to accelerate the modern footsteps of agriculture. The water-saving irrigation system has the advantages of researching the water-saving mode suitable for the crops, improving the water utilization rate of the crops, realizing accurate irrigation, being very important for relieving the water shortage and the water resource crisis of agriculture and having important significance for guaranteeing the grain safety, ecological safety and social sustainable development of China.
Accurate irrigation is the highest goal of water-saving irrigation, which means that the irrigation quantity required by the growth of crops is accurately estimated, and the water quantity is accurately and uniformly irrigated into the soil of the root system layer of the crops by adopting an efficient water-saving irrigation technology. Therefore, the water demand of crops needs to be predicted scientifically and reasonably in order to reasonably utilize water resources and realize accurate irrigation.
At present, in the research of a crop water demand prediction method, most researchers are biased to the optimization innovation of a weather factor-based prediction algorithm in a traditional cultivation mode, soil factors and crop physiological information are not introduced, and the crop water demand is predicted by selecting different factors influencing the crop water demand according to the crop growth stage. However, the crop water demand has close relation with the crop growth stage, crop physiology and other factors. Different influencing factors are selected according to the crop growth stage, and the water demand of crops can be predicted more accurately by combining a high-precision prediction algorithm.
Aiming at a model for predicting the water demand of crops, the prior technical scheme is as follows: zhang Bing et al applied the L-M optimization algorithm BP neural network to a crop water demand prediction model for the disadvantage of low convergence rate of the BP neural network, and performed prediction example analysis on green pepper crop water demand in the university of tennessee, university plateau laboratory in the United states. According to the scheme, although the convergence speed of the network is improved, smaller training errors can be obtained, the input samples are only meteorological data, crop physiological information is not introduced, and different influencing factors are not selected according to the crop growth stage to predict the crop water demand. Shang Zhigen et al propose a model for predicting the water demand of crops by integrating random forests with an MLP neural network, which can obtain better prediction accuracy, but the model input samples are only meteorological data, and the factors influencing the water demand of crops are not analyzed and selected according to the growth stage of the crops. The Dahua Tang et al input data considers two crop physiological indexes of leaf area index and plant height except for the meteorological factors when predicting the crop water demand by using an artificial neural network optimized by a genetic algorithm, but does not consider the crop canopy temperature in the crop physiological information.
For plants, the leaves have multiple functions, the air temperature of the leaves is an important index of plant physiological sensing, the important index can reflect the moisture condition of crops, and a plurality of researches show that the canopy temperature of the crops is a good index reflecting the moisture condition of the crops, and the moisture condition of the crops is closely related to the water demand of the crops. The acquisition of the crop canopy temperature also overcomes the defects of larger sampling error and time consumption existing in the measurement of other parameters. Therefore, it is necessary to consider the crop canopy temperature as an input factor for predicting the crop water demand model. In addition, the planting methods in all the above documents are not the rain-collecting, deficiency-regulating and drip-irrigation methods for crops. In addition, in all the prediction models related to the crop water demand, the factors influencing the crop water demand are not respectively analyzed and selected according to the crop growth stage.
Disclosure of Invention
The prior art, analyzed, found the following disadvantages: 1) Most of the methods for predicting the water demand of crops mainly consider meteorological factors, and lack of consideration of the canopy temperature of the crops as an input factor of physiological information of the crops. 2) The crop water demand model for predicting the rain collecting, regulating and deficient drip irrigation mode is lacking. 3) There is no model for predicting the water demand of crops by respectively analyzing and selecting different input factors according to the growth stage of crops.
Therefore, the invention provides a crop water demand prediction method based on an improved Elman neural network in a rain-collecting, regulating and deficiency planting mode, and the input factors of the method consist of meteorological factors and crop factors. The weather factors mainly consider data such as daily average air temperature, daily sunshine hours, daily average relative humidity, water vapor pressure and the like; crop factors mainly consider leaf area index, plant height and canopy temperature of crops. According to the method, firstly, according to the crop growth stage, the factors influencing the crop water demand are analyzed according to experimental data from years, and then, in different crop growth stages, the factors which have obvious influence on the crop water demand are selected as input factors of a prediction model. And carrying out normalization processing on the data, then adopting an Elman neural network with dynamic modeling capability to predict, and adopting a genetic algorithm (Genetic Algorithm) to select the optimal connection weight and threshold of the Elman neural network to obtain the Elman neural network with the optimal state.
The invention adopts the following technical scheme:
the crop water demand prediction method based on the improved Elman neural network adopts an Elman neural network model with dynamic modeling capability to predict, and is characterized in that: respectively selecting different influencing factors as input factors of a prediction model in different growing stages of crops; and selecting the optimal connection weight and the threshold value of the Elman neural network by adopting a genetic algorithm.
The input factors comprise meteorological factors and crop factors, wherein the meteorological factors comprise daily average air temperature, daily sunshine hours, daily average relative humidity and daily highest air temperature; crop factors include leaf area index, plant height, canopy temperature.
When the crops are green peppers, the growing stage comprises a seedling stage, a flowering and fruit setting stage, a fruiting full stage and a fruiting later stage. According to historical data, carrying out correlation analysis on influence factors of crop water demand in each growth stage of green peppers and the crop water demand, and selecting input factors in different growth stages according to the correlation sorting.
Preferably, 5 input factors are selected for the seedling stage, the flowering and fruit setting stage, the fruiting full stage and the fruiting later, and the daily highest air temperature, the daily average air temperature, the sunshine hours, the canopy temperature and the leaf area index are selected for the seedling stage respectively; selecting average daily air temperature, maximum daily air temperature, leaf area index, plant height and sunshine hours in the flowering and fruit setting period; selecting daily average air temperature, canopy temperature, sunshine hours, leaf area index and plant height in a fruiting period; and selecting the daily average air temperature, the canopy temperature, the daily highest air temperature, the number of sunshine hours and the daily average relative humidity in the later stage of the result.
When prediction is carried out, normalization processing is carried out on all sample data, and the sample data are converted into a 0-1 interval by adopting the following formula:
x' p ＝(x p -x min )/(x max -x min )
wherein x is p (p=1, 2, …, P) is sample data, x max ＝max{x p },x min ＝min{x p }；
After the Elman neural network model, restoring calculation is carried out on network output data, and an actual value is restored:
x yp ＝x' yp (x y max -x y min )+x y min
wherein x is yp The result after the calculation is restored, namely an actual value; x's' yp For the predicted value of the network output, x y max Outputting the maximum value, x of the prediction result for the network y min The minimum value of the predicted outcome is output for the network.
The gradient descent function of momentum and self-adaptive learning rate is adopted as a training algorithm of the Elman neural network, the S function is adopted as a transfer function of neurons in the middle layer, the output vector is the crop water demand, the output vector is a 1-dimensional vector, the output layer is only one neuron, and the linear transfer function is adopted as a transfer function of the output neuron. And optimizing connection weights from an input layer to an implied layer, from the implied layer to an output layer and from the implied layer to a receiving layer of the Elman neural network by adopting a genetic algorithm, and threshold values of the implied layer and the output layer.
Preferably, the water demand prediction is carried out on the condition that the crops in the rain collecting, deficiency adjusting and drip irrigation mode are moderately deficient in water treatment in the later fruiting period and the other growing periods are normally and fully irrigated.
(III) beneficial effects
The effectiveness of the GA-Elman method provided by the invention is verified by using the data of the Hebei engineering university experiment, and the result shows that the prediction model can effectively predict the water demand of crops in a rain-collecting, deficiency-regulating and drip irrigation mode, the performance of the prediction model is obviously superior to that of the prediction model which only inputs meteorological factors in the whole growth period of crops or only selects fixed input factors in the whole growth period of crops, and the prediction model is also superior to that of the prediction model which singly uses an Elman neural network. In addition, the rainfall is fully utilized and simultaneously the deficiency-regulating drip irrigation is carried out, so that multiple water conservation is realized, and the water resource is saved to a greater extent.
Drawings
FIG. 1 is a schematic diagram of a rain-collecting, regulating and deficiency planting mode;
FIG. 2 is a schematic top view of a rain-collecting, regulating and deficiency drip irrigation planting;
FIG. 3 is a flow chart of a green pepper crop water demand prediction method based on an improved Elman neural network; and
fig. 4 is a graph of Elman neural network model.
Detailed Description
The present invention will be further described in detail below with reference to specific embodiments and with reference to the accompanying drawings, in order to make the objects, technical solutions and advantages of the present invention more apparent.
In the invention, green pepper crops with rain collection, deficiency adjustment and drip irrigation are taken as an example, a 'tortoiseshell' type furrow is adopted for collecting rain in a rain collection mode, black mulching films (the film thickness is about 0.012 mm) are covered on the furrow, and the furrow-ridge ratio is set to be 60:40. and a conventional plain action (denoted B) was set for comparison. The flow rate of the drippers is 2.2L/h, and the interval between the drippers is 30cm. The row spacing of the plants is 50cm, and the plant spacing is 30cm. The schematic diagrams of the rain collecting, regulating and deficiency planting modes are shown in fig. 1 and 2. Dividing the whole growth period of green peppers into 4 stages, namely: seedling stage, flowering and fruit setting stage, fruiting stage and fruiting later stage. Two kinds of water shortage treatments are respectively arranged in each growth period, namely medium water shortage (represented by M: 60% -70% of field water holding capacity) and light water shortage (represented by L: 70% -80% of field water holding capacity), and the control group is full water irrigation (represented by N) in the whole growth period. In addition, the rain-collecting drip irrigation planting was compared with the flat planting (R represents rain collection, B represents flat planting), and the specific results are shown in Table 1.
TABLE 1 Water deficit control treatment for different species of plants at different growth periods
The data collected by the field test gives: under the same water regulating and depletion treatment, the growth vigor and yield of green pepper planted by rain-collecting drip irrigation are obviously better than those of the flat planting. And the deficiency-regulating planting mode obviously saves water compared with full irrigation, and obviously affects the water demand of crops. Through analysis of variance, comprehensively considering the aspects of green pepper yield, quality, irrigation water utilization efficiency and the like, the secondary moderate water shortage treatment of the green pepper results in the rain-collecting planting mode is selected, and the planting modes of normal and full irrigation in other growth periods are optimal, so that water is saved by more than 30% compared with the normal full irrigation, the green pepper yield can be remarkably improved, and meanwhile, the Vc content of the green pepper is improved.
As shown in fig. 3, the green pepper crop water demand prediction method based on the improved Elman neural network mainly comprises two parts of data input, preprocessing, prediction algorithm and result output. The weather factors, the crop factors and the soil factors can have certain influence on the water demand of crops, and the data such as the daily average air temperature, the daily sunshine hours, the daily average relative humidity, the daily highest air temperature and the like in the weather factors are selected when the water demand of the green pepper crops which are collected, regulated and drip-irrigated is predicted, and the leaf area index, the plant height and the canopy temperature in the crop factors are used as input data of a prediction model. And respectively analyzing factors influencing the water demand of the crops by using Origin software in each growth period of the green pepper crops.
In the prediction algorithm module, a genetic algorithm is adopted to optimally select the connection weight and each layer of threshold value of the Elman neural network, and then the improved Elman neural network is trained and tested by data obtained by the data input and preprocessing module, so that the effectiveness of the algorithm is proved.
1 data input and preprocessing
1.1 data input
Most of the crop water demand prediction models only consider meteorological factors, and rarely consider the meteorological factors and crop factors simultaneously to predict the water demand of green pepper crops under the condition of rain collecting, regulating and deficiency drip irrigation. Meanwhile, input factors influencing the water demand of crops are not analyzed and selected respectively according to the growth stage of the crops. Therefore, for the input data of predicting the water demand of green pepper crops under the condition of rain collecting, regulating and drip irrigation, the weather factors are considered, the crop factors are considered at the same time, and the input factors are analyzed and selected according to the growing stage, so that more accurate prediction of the water demand of the crops is necessary. The division of green pepper growth stages in the Hebei Handan area is shown in Table 2.
TABLE 2 division of green pepper fertility stage (Hebei province Handan city)
According to the invention, the Origin software is used for carrying out correlation analysis on the influence factors of the crop water demand and the crop water demand in each growth stage of green peppers according to the test data in recent years in the area of the Origin. The factors affecting the water demand of green pepper crops in the region are found to be different in correlation with the water demand of the crops according to different growth stages of the crops. In the rain-collecting, deficiency-regulating and drip-irrigation planting mode, the correlation between factors affecting the water demand of crops and the water demand of the green pepper crops in different growing stages is shown in a table 3.
TABLE 3 analysis of correlation of Green pepper crop influencing factors with crop Water demand
1.2 data preprocessing
In order to ensure the nonlinear effect of neurons, normalization processing is performed on a numerical learning sample, and the normalization processing is performed by using a formula (1) without losing generality, so that all sample data are converted into a range of 0-1. Thus, a smaller number can be used as the connection right W of the network, and the problem of calculation overflow does not occur in the network calculation.
x′ p ＝(x p -x min )/(x max -x min ) (1)
Wherein x is p (p=1, 2, …, P) is sample data, x max ＝max{x p },x min ＝min{x p }。
2 prediction algorithm (Elman neural network improvement module) and result output
2.1Elman neural network
The Elman neural network consists of four layers, namely an input layer, an implicit layer, a receiving layer and an output layer, and is added with the receiving layer as a one-step time delay operator to achieve the purpose of memorizing, so that the network has the capability of adapting to time-varying characteristics, can directly reflect the network characteristics of a dynamic process, achieves the purpose of dynamic modeling, and has a good prediction effect on data with larger volatility. The Elman neural network model used in the present invention is shown in fig. 4.
Although Elman neural networks have improved performance over conventional neural networks, all of the problems encountered with the neural networks remain in the design process: training algorithm, transfer function, network structure and connection weight. The invention adopts a gradient descent function of momentum and self-adaptive learning rate as a training algorithm of an Elman neural network, adopts a common S function as a transfer function of neurons in an intermediate layer, adopts an output vector as crop water demand, is a 1-dimensional vector, adopts an output layer as only one neuron, and adopts a linear transfer function as a transfer function of the output neuron. The selection of the structure (comprising the number of hidden layers, the number of nodes at each layer and the like) and the connection weight of the Elman neural network is very important to the performance of the whole network.
2.2 Elman neural network improved based on genetic algorithm has excellent performances such as nonlinear mapping, self-learning and the like, but the method is necessarily based on the adoption of proper connection weights and thresholds, and the aim of the method is to select the network connection weights and thresholds which optimize the performances of the Elman neural network by using genetic algorithm. The genetic algorithm is a method for searching the optimal solution by simulating the natural evolution process, has high-efficiency global searching capability, and shows very good performance and high efficiency in solving the complex optimization problem.
The genetic algorithm optimizes the connection weights of the Elman neural network input layer to the hidden layer, the hidden layer to the output layer, and the hidden layer to the sink layer, and the thresholds of the hidden layer and the output layer. The parameter settings of the genetic algorithm are shown in table 4. The initial parameter settings for the Elman neural network are shown in table 5.
Table 4 genetic algorithm key parameter settings
Table 5 Elman neural network initial parameter settings
2.3 output of results
And (3) carrying out restoration calculation on the output data of the network, and recovering the actual value. The reduction calculation is the inverse calculation process of formula (1), namely using formula (2):
x yp ＝x' yp (x y max -x y min )+x y min (2)
wherein x is yp For restoring the calculated result, i.e. the actual value. x's' yp For the predicted value of the network output, x y max Outputting the maximum value, x of the prediction result for the network y min The minimum value of the predicted outcome is output for the network. The invention uses the meteorological data collected by the automatic meteorological station of the Hebei engineering university of city in Hebei province and the experimental data collected from years to verify, wherein the meteorological data comprises green pepper crop information. According to the invention, through long-term test, variance analysis is respectively carried out on the yield, the water utilization efficiency and the Vc content of green peppers under different treatments in different years, and the differences of experimental data in different years mainly come from test treatments, but not the annual differences. In addition, the yield, the quality and the water utilization efficiency of the furrow rain-collecting green peppers are found to be remarkably improved. The method comprises the steps of taking the water utilization efficiency, the yield and the quality (Vc content) of green peppers as an evaluation index system for evaluating a high-yield, high-quality and water-saving irrigation system, comprehensively evaluating the production conditions of the green peppers under different water treatment, respectively extracting main components of each factor by adopting a main component analysis method, comprehensively evaluating, and selecting the optimal test treatment technical scheme as follows: the ditch-ridge ratio is 60:40, and a later stage moderately deficient water technical scheme. The water is saved by more than 30% compared with the traditional horizontal-cropping full irrigation.
The temperature of the green pepper canopy is selected from the input crop factors as the input factors, and the correlation between the green pepper canopy temperature and the crop water demand is larger than the leaf area index and the plant height in the seedling stage, the fruiting full stage and the fruiting later stage of the green pepper crop through correlation analysis, so that the green pepper canopy temperature is necessary to be introduced as the input factors of the crop water demand.
The technical scheme is selected by adopting an Elman neural network method improved by a genetic algorithmAnd (5) carrying out simulation analysis on the water demand of the material. Data from 2015 to 2017 are used as training samples of the Elman neural network, and data from 2018 are used as test samples. According to the crop growth stage, the correlation analysis is carried out on the factors influencing the crop water demand respectively, and the correlation between the factors influencing the crop water demand and the crop water demand is found to be changed in different growth stages of green pepper crops, so that the accuracy of the crop water demand prediction model can be improved by selecting different prediction model input factors in different growth stages, and the crop water demand can be predicted more accurately. The invention uses root mean square error (Root Mean Square Error, RMSE), mean absolute error (mean absolute error, MAE) and decision coefficient (coefficient of determination, R) 2 ) Analyzing the prediction results of each model, P i As a predicted value, O i In order to observe the value of the value,
for predicting mean value +.>
For the observation mean, N is the observation number (number of samples in the test set). The results are shown in Table 6 (model R, which only considers meteorological factors due to the whole crop growth period) 2 Lower, about 0.8 or so, and therefore, analysis thereof is omitted from table 6).
It can be seen from table 6 that, in the case of inputting an equal number of input factors (5 input factors are selected according to the correlation from large to small), the Elman neural network taking the same input factors into consideration in the whole growth period has the lowest precision and fitting degree, and the GA-Elman neural network taking only the same input factors into consideration in the whole growth period has higher precision and fitting degree, which is the Elman neural network taking the input factors into consideration according to the growth stage of crops. The GA-Elman neural network with highest precision and fitting degree respectively considers input factors according to the crop growth stage. It can be seen that for the present invention, the selection of different input factors during different incubation periods has a greater impact on model accuracy than merely improving the impact of the Elman neural network on model accuracy. The method used in the present invention has achieved a satisfactory result.
TABLE 6 prediction result analysis
Aiming at the green pepper crop water demand prediction model which lacks consideration of canopy temperature in crop factors in a rain-collecting, deficiency-regulating and drip-irrigation mode, the invention adds the crop canopy temperature except for leaf area index and plant height in the crop factors in input data except for consideration of meteorological factors; on the basis, according to the crop growth stage and the correlation analysis result, different factors influencing the crop water demand are selected respectively, and the improved Elman neural network is adopted to predict the green pepper crop water demand in a rain collecting, regulating and deficiency mode.
While the foregoing is directed to embodiments of the present invention, other and further details of the invention may be embodied and described herein, it will be appreciated that the preceding is merely illustrative of the invention and not limiting, any modifications, equivalents, improvements or improvements made within the spirit and principles of the invention, including but not limited to: changes in the predicted time scale (time, month, year, etc.), increases or decreases in predicted data, changes in predicted locations, changes in predicted crop types, etc., are all intended to be within the scope of the present invention.
Claims (10)
1. The rain collecting and regulating deficiency crop water demand prediction method based on improved Elman neural network adopts an Elman neural network model with dynamic modeling capability to predict, and is characterized in that: respectively selecting different influencing factors as input factors of a prediction model in different growing stages of crops; and selecting the optimal connection weight and the threshold value of the Elman neural network by adopting a genetic algorithm.
2. The method of claim 1, wherein the input factors consist of meteorological factors and crop factors.
3. The method of claim 2, wherein the meteorological factors comprise average daily air temperature, number of hours of insolation, average daily relative humidity, maximum daily air temperature; the crop factors include leaf area index, plant height, canopy temperature.
4. A method according to claim 3, wherein the crop is green pepper and the growth stage comprises a seedling stage, a flowering and fruit setting stage, a fruiting stage and a fruiting post stage.
5. The method according to claim 4, wherein the correlation analysis is carried out on the influence factors of the crop water demand of each growing stage of green peppers and the crop water demand according to the historical data, and the input factors of different growing stages are selected according to the correlation ranking.
6. The method according to claim 5, wherein all sample data are normalized and converted into the interval 0-1:
x′ p ＝(x p -x min )/(x max -x min )
wherein x is p (p=1, 2, …, P) is sample data, x max ＝max{x p },x min ＝min{x p }；
After the Elman neural network model, restoring calculation is carried out on network output data, and an actual value is restored:
x yp ＝x' yp (x ymax -x ymin )+x ymin
wherein x is yp The result after the calculation is restored, namely an actual value; x's' yp For the predicted value of the network output, x ymax Outputting the maximum value, x of the prediction result for the network ymin The minimum value of the predicted outcome is output for the network.
7. The method according to claim 6, wherein a gradient descent function of momentum and self-adaptive learning rate is used as a training algorithm of the Elman neural network, an S function is used as a transfer function of neurons in the middle layer, an output vector is a crop water demand, the output vector is a 1-dimensional vector, the output layer is only one neuron, and a linear transfer function is used as a transfer function of the output neuron.
8. The method of claim 7, wherein the genetic algorithm optimizes the Elman neural network input layer to hidden layer, hidden layer to output layer, and hidden layer to sink layer connection weights, and hidden layer and output layer thresholds.
9. The method according to claim 8, wherein 5 input factors are selected for the seedling stage, the flowering and fruit setting stage, the fruiting full stage and the fruiting later stage, and the daily maximum air temperature, the daily average air temperature, the number of sunshine hours, the canopy temperature and the leaf area index are selected for the seedling stage respectively; selecting average daily air temperature, maximum daily air temperature, leaf area index, plant height and sunshine hours in the flowering and fruit setting period; selecting daily average air temperature, canopy temperature, sunshine hours, leaf area index and plant height in a fruiting period; and selecting the daily average air temperature, the canopy temperature, the daily highest air temperature, the number of sunshine hours and the daily average relative humidity in the later stage of the result.
10. The method according to any one of the preceding claims, wherein the water demand is predicted for the condition that the crops in the rain-collecting, regulating, deficiency and drip irrigation mode are moderately deficient in water in the later fruiting period and are fully irrigated in other growth periods.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910173728.7A CN109934400B (en) | 2019-03-08 | 2019-03-08 | Rain collecting, regulating and deficiency crop water demand prediction method based on improved neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910173728.7A CN109934400B (en) | 2019-03-08 | 2019-03-08 | Rain collecting, regulating and deficiency crop water demand prediction method based on improved neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109934400A CN109934400A (en) | 2019-06-25 |
| CN109934400B true CN109934400B (en) | 2023-05-16 |
Family
ID=66986782
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910173728.7A Active CN109934400B (en) | 2019-03-08 | 2019-03-08 | Rain collecting, regulating and deficiency crop water demand prediction method based on improved neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109934400B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112966221A (en) * | 2021-03-04 | 2021-06-15 | 江西省水利科学院 | Method for converting total water consumption of assessment year in southern rich water region |
| CN114201923A (en) * | 2022-02-16 | 2022-03-18 | 中国地质调查局水文地质环境地质调查中心 | Processing method and device of landslide displacement prediction model |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101169627A (en) * | 2007-11-27 | 2008-04-30 | 中国水利水电科学研究院 | On-line crop water stress irrigation decision monitoring system |
| CN102565150A (en) * | 2012-01-13 | 2012-07-11 | 北京盈胜泰科技术有限公司 | Detection device and monitoring system for soil relative humidity and soil fertility change |
| CN103886392A (en) * | 2014-03-26 | 2014-06-25 | 中国农业科学院农田灌溉研究所 | Winter wheat water consumption predicting method based on weather forecast information |
| CN106212217A (en) * | 2016-08-02 | 2016-12-14 | 浙江恒合景观规划设计有限公司 | Intelligent precision irrigation control system |
| CN106707757A (en) * | 2017-01-23 | 2017-05-24 | 中国农业大学 | Dynamic regulation and control method and system for irrigation time |
| WO2017106962A1 (en) * | 2015-11-17 | 2017-06-29 | Universite Laval | System and method for real-time water management |
| CN107025505A (en) * | 2017-04-25 | 2017-08-08 | 无锡中科智能农业发展有限责任公司 | A kind of paddy water requirement prediction method based on principal component analysis and neutral net |
| CN107302875A (en) * | 2016-04-20 | 2017-10-31 | 青岛农业大学 | A kind of water-saving irrigation method based on the crucial moisture information of soil-plant system |
| CN107358294A (en) * | 2017-07-21 | 2017-11-17 | 河北工程大学 | A kind of water demand prediction method based on Elman neutral nets |
| CN107507096A (en) * | 2017-08-14 | 2017-12-22 | 陈守敏 | A kind of irrigation rainwater harvesting cellar management information system based on big data |
| CN107945042A (en) * | 2017-11-29 | 2018-04-20 | 上海华维节水灌溉股份有限公司 | A kind of plant growth irrigation decision control system |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2417548B1 (en) * | 2009-04-06 | 2021-08-11 | Smartfield, Inc. | Remote analysis and correction of crop condition |
| WO2012065089A1 (en) * | 2010-11-12 | 2012-05-18 | Smartfield, Inc. | Agricultural management using biological signals |
| US9880537B2 (en) * | 2015-08-05 | 2018-01-30 | Clearag, Inc. | Customized land surface modeling for irrigation decision support in a crop and agronomic advisory service in precision agriculture |
-
2019
- 2019-03-08 CN CN201910173728.7A patent/CN109934400B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101169627A (en) * | 2007-11-27 | 2008-04-30 | 中国水利水电科学研究院 | On-line crop water stress irrigation decision monitoring system |
| CN102565150A (en) * | 2012-01-13 | 2012-07-11 | 北京盈胜泰科技术有限公司 | Detection device and monitoring system for soil relative humidity and soil fertility change |
| CN103886392A (en) * | 2014-03-26 | 2014-06-25 | 中国农业科学院农田灌溉研究所 | Winter wheat water consumption predicting method based on weather forecast information |
| WO2017106962A1 (en) * | 2015-11-17 | 2017-06-29 | Universite Laval | System and method for real-time water management |
| CN107302875A (en) * | 2016-04-20 | 2017-10-31 | 青岛农业大学 | A kind of water-saving irrigation method based on the crucial moisture information of soil-plant system |
| CN106212217A (en) * | 2016-08-02 | 2016-12-14 | 浙江恒合景观规划设计有限公司 | Intelligent precision irrigation control system |
| CN106707757A (en) * | 2017-01-23 | 2017-05-24 | 中国农业大学 | Dynamic regulation and control method and system for irrigation time |
| CN107025505A (en) * | 2017-04-25 | 2017-08-08 | 无锡中科智能农业发展有限责任公司 | A kind of paddy water requirement prediction method based on principal component analysis and neutral net |
| CN107358294A (en) * | 2017-07-21 | 2017-11-17 | 河北工程大学 | A kind of water demand prediction method based on Elman neutral nets |
| CN107507096A (en) * | 2017-08-14 | 2017-12-22 | 陈守敏 | A kind of irrigation rainwater harvesting cellar management information system based on big data |
| CN107945042A (en) * | 2017-11-29 | 2018-04-20 | 上海华维节水灌溉股份有限公司 | A kind of plant growth irrigation decision control system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109934400A (en) | 2019-06-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112906298B (en) | Blueberry yield prediction method based on machine learning | |
| CN106842923B (en) | Greenhouse environment multi-factor coordination control method based on crop physiology and energy consumption optimization | |
| CN108781926B (en) | Greenhouse irrigation system and method based on neural network prediction | |
| CN110069032B (en) | Eggplant greenhouse environment intelligent detection system based on wavelet neural network | |
| CN116595333B (en) | Soil-climate intelligent rice target yield and nitrogen fertilizer consumption determination method | |
| CN115453868B (en) | Full-growth-period light intensity regulation and control method based on tomato light response difference characteristics | |
| CN111915062A (en) | Greenhouse crop water demand regulation and control method with water utilization rate and photosynthetic rate coordinated | |
| CN109934400B (en) | Rain collecting, regulating and deficiency crop water demand prediction method based on improved neural network | |
| CN116451823A (en) | Apple yield prediction method based on meteorological master control factors | |
| CN113229076A (en) | Wheat breeding method for screening excellent salt-tolerant strains by using salt pond | |
| CN108319134A (en) | A kind of greenhouse environment intelligent control method based on extreme learning machine network | |
| CN112167035B (en) | Hydroponic leaf vegetable production management method and system | |
| Viola et al. | Olive yield and future climate forcings | |
| CN113984964A (en) | Corn variety screening method and system based on multi-source data monitoring drought resistance analysis | |
| CN118278617A (en) | Rain-farming area corn intelligent irrigation method based on remote sensing and Internet of things | |
| CN112083748B (en) | Facility light environment regulation and control method with priority to light quality | |
| Yuan et al. | Photosynthetic rate prediction model based on PSO-LSSVM for optimization and control of greenhouse environment | |
| Zhou et al. | Optimization of cotton irrigation management for different climatic conditions using the CROPGRO-cotton model. | |
| Yuan et al. | Photosynthetic Rate Prediction Model of Greenhouse Crops Based on PLS-ELM | |
| CN117575174B (en) | Intelligent agricultural monitoring and management system | |
| 남두성 | Predicting Transpiration Rates of Hydroponically-Grown Paprika Plants via an Artificial Neural Network Using Environmental and Growth Factors in Greenhouses | |
| CN118627699B (en) | Citrus yield prediction method, device and application thereof | |
| Chetan et al. | Temperature and Precipitation Forecast for Western Transylvanian Plains Using Artificial Intelligence | |
| Homma et al. | Modeling Short-term Yield Changes in Sweet Pepper Based on Dry Matter Production and Fruit Growth | |
| CN112580703B (en) | Method for predicting morbidity of panax notoginseng in high-incidence stage |
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 |