CN109964611A - A kind of field crop Tree Precise Fertilization method and system - Google Patents
A kind of field crop Tree Precise Fertilization method and system Download PDFInfo
- Publication number
- CN109964611A CN109964611A CN201910303769.3A CN201910303769A CN109964611A CN 109964611 A CN109964611 A CN 109964611A CN 201910303769 A CN201910303769 A CN 201910303769A CN 109964611 A CN109964611 A CN 109964611A
- Authority
- CN
- China
- Prior art keywords
- farmland
- sample
- soil
- amount
- component
- 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
Classifications
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01C—PLANTING; SOWING; FERTILISING
- A01C21/00—Methods of fertilising, sowing or planting
- A01C21/005—Following a specific plan, e.g. pattern
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01C—PLANTING; SOWING; FERTILISING
- A01C21/00—Methods of fertilising, sowing or planting
- A01C21/007—Determining fertilization requirements
Abstract
The embodiment of the present invention provides a kind of field crop Tree Precise Fertilization method and system, this method comprises: obtaining the sample data in multiple groups sample farmland；Using wavelet analysis, the crop yield in each sample farmland and the Wavelet Component of each soil nutrient content are obtained；BP neural network according to the sample data and its Wavelet Component in each sample farmland, after obtaining training；According to Non-Linear Programming, obtain the best amount of nitrogen in the target farmland, the optimum phosphorus application amount in the target farmland, the target farmland best amount of potassium applied.The embodiment of the present invention analyzes the crop yield and soil nutrient content that measure the obtained target farmland by wavelet analysis, it can remove because of existing error in all respects during practical measurement crop yield and target Soil Nutrients in Farmland content, so that BP neural network is more accurate to the projected relationship between soil nutrient, dose and crop yield, to realize the accurate prediction of target farmland fertilization amount.
Description
Technical field
The present embodiments relate to agricultural technology field more particularly to a kind of field crop Tree Precise Fertilization method and system.
Background technique
In crop production, fertilising is the important measures of crop high yield, and fertilizer not only directly provides nutrition necessary to crop
Element supplies crop growth demand, and plays the role of improving soil, improves soil supplying nutrient capability.The input amount of chemical fertilizer
Crop output, producer's income and environmental quality are directly affected with utilization rate.
It shows according to the study, production estimation person often blindly increases making for fertilizer to achieve the purpose that high yield and bumper harvest
Dosage.Since specific crops are to the selective absorbing of nutrient, extra soil salt accumulates year by year, leads to soil environment
Deteriorate, crop yield can also reduce year by year, and therefore, scientific Tree Precise Fertilization is composition portion important in production estimation operating system
/ mono-.
Tree Precise Fertilization is the core content in precision agriculture technology, and basic thought is the spatial variability based on soil nutrient
Property carry out variable fertilization.It was verified that Tree Precise Fertilization can save fertilizer, increase grain yield, balanced soil nutrient reduces ring
Border pollution.
The difficult point of Tree Precise Fertilization essentially consists in the formulation of decision-making technique, however existing fertilizing method and current production practices
Incompatibility is many, such as nutrient balance method undetermined coefficient is more, cannot reflect the reciprocation between nutrient, and
Since target output is difficult estimation accurately, cause calculation of fertilization amount error larger；The fertilizer effect method test period is long, weight between the time
Renaturation is poor, is easy to appear " saddle-shape " curve, and test work load is big, predicts dose bigger error, and a large amount of numbers of its accumulation
It is practiced according to Instructing manufacture is difficult to.
Therefore, it is necessary to construct a kind of more effective crop Tree Precise Fertilization method.
Summary of the invention
The embodiment of the present invention provides a kind of field crop Tree Precise Fertilization method and system, predicts in the prior art to solve
Dose bigger error is difficult to use in the problem of Instructing manufacture is practiced.
In a first aspect, the embodiment of the present invention provides a kind of field crop Tree Precise Fertilization method, comprising:
The sample data in multiple groups sample farmland is obtained, the sample data in each sample farmland includes: each sample farmland
Soil nitrogenous amount, the soil phosphorus content in each sample farmland, the soil potassium content in each sample farmland, each sample farmland are applied
The crop yield of nitrogen quantity, the phosphorus application amount in each sample farmland, the amount of potassium applied in each sample farmland and each sample farmland；
Using wavelet analysis, the Wavelet Component of the crop yield in each sample farmland is obtained, each sample agricultural land soil contains
The small echo of the Wavelet Component of nitrogen quantity, the Wavelet Component of each sample agricultural land soil phosphorus content and each sample agricultural land soil potassium content
Component；
According to the small wavelength-division of the Wavelet Component of the crop yield in each sample farmland, each sample agricultural land soil nitrogen content
Amount, the Wavelet Component of the Wavelet Component of each sample agricultural land soil phosphorus content and each sample agricultural land soil potassium content, obtain instruction
BP neural network after white silk；
Obtain the soil nitrogenous amount in target farmland, the soil phosphorus content in the target farmland, the target farmland soil
The history crop yield of potassium content and the target farmland, the soil for being obtained the target farmland respectively using wavelet analysis are contained
The Wavelet Component of nitrogen quantity, the Wavelet Component of the soil phosphorus content in the target farmland, the target farmland soil potassium content
The Wavelet Component of the history crop yield in Wavelet Component and the target farmland；
By the Wavelet Component of the soil nitrogenous amount in the target farmland, the small wavelength-division of the soil phosphorus content in the target farmland
Amount, the target farmland soil potassium content Wavelet Component and the target farmland history crop yield Wavelet Component
In BP neural network after being input to training, and according to Non-Linear Programming, the amount of nitrogen in the target farmland, the target are obtained
The phosphorus application amount in farmland, the target farmland amount of potassium applied.
Second aspect, the embodiment of the present invention provide a kind of field crop Tree Precise Fertilization system, comprising:
Module is obtained, for obtaining the sample data in multiple groups sample farmland, the sample data in each sample farmland includes: every
The soil nitrogenous amount in one sample farmland, the soil phosphorus content in each sample farmland, each sample farmland soil potassium content, each
The amount of nitrogen in sample farmland, the phosphorus application amount in each sample farmland, the amount of potassium applied in each sample farmland and each sample farmland work
Produce amount；
First small echo module obtains the Wavelet Component, every of the crop yield in each sample farmland for utilizing wavelet analysis
The Wavelet Component of one sample agricultural land soil nitrogen content, the Wavelet Component of each sample agricultural land soil phosphorus content and each sample farmland
The Wavelet Component of soil potassium content；
Training module, for being contained according to the Wavelet Component of the crop yield in each sample farmland, each sample agricultural land soil
The small echo of the Wavelet Component of nitrogen quantity, the Wavelet Component of each sample agricultural land soil phosphorus content and each sample agricultural land soil potassium content
Component, the BP neural network after obtaining training；
Second small echo module, for obtaining the soil nitrogenous amount in target farmland, the soil phosphorus content in the target farmland, institute
State the soil potassium content in target farmland and the history crop yield in the target farmland, obtained respectively using wavelet analysis described in
The Wavelet Component of the soil nitrogenous amount in target farmland, the Wavelet Component of the soil phosphorus content in the target farmland, the target agriculture
The Wavelet Component of the history crop yield in the Wavelet Component of the soil potassium content in field and the target farmland；
Planning module, for by the soil of the Wavelet Component of the soil nitrogenous amount in the target farmland, the target farmland
The Wavelet Component of phosphorus content, the target farmland soil potassium content Wavelet Component and the target farmland history crops
The Wavelet Component of yield is input in the BP neural network after training, and according to Non-Linear Programming, obtains the target farmland
Amount of nitrogen, the phosphorus application amount in the target farmland, the target farmland amount of potassium applied.
The third aspect, the embodiment of the present invention provide a kind of electronic equipment, comprising:
At least one processor, at least one processor, communication interface and bus；Wherein,
The processor, memory, communication interface complete mutual communication by the bus；
The communication interface is for the information transmission between the test equipment and the communication equipment of display device；
The memory is stored with the program instruction that can be executed by the processor, and the processor calls described program to refer to
Enable a kind of field crop Tree Precise Fertilization method that first aspect offer is provided.
Fourth aspect, the embodiment of the present invention provide a kind of non-transient computer readable storage medium, which is characterized in that described
Non-transient computer readable storage medium stores computer instruction, and the computer instruction makes the computer execute first aspect
A kind of field crop Tree Precise Fertilization method provided.
A kind of field crop Tree Precise Fertilization method and system provided in an embodiment of the present invention, by wavelet analysis to making produce
Amount is analyzed, and can be removed because of existing error in all respects during practical measurement crop yield, be passed through sample farmland
Sample data and carry out wavelet analysis after crop yield BP neural network is trained so that BP neural network is to soil
Projected relationship between nutrient, dose and crop yield is more accurate, to realize the accurate prediction of target farmland fertilization amount.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is this hair
Bright some embodiments for those of ordinary skill in the art without creative efforts, can be with root
Other attached drawings are obtained according to these attached drawings.
Fig. 1 is a kind of flow chart of field crop Tree Precise Fertilization method provided in an embodiment of the present invention；
Fig. 2 indicates the structural schematic diagram of two layers of multiresolution wavelet analysis in the embodiment of the present invention；
Fig. 3 indicates the wavelet analysis result schematic diagram of difference N value in the embodiment of the present invention；
Fig. 4 indicates the schematic diagram of the low frequency general picture in the embodiment of the present invention under different scale；
Fig. 5 indicates crop yield data wavelet analysis numerical value change curve synoptic diagram in the embodiment of the present invention；
Fig. 6 indicates a kind of structural schematic diagram of BP neural network provided in an embodiment of the present invention；
Fig. 7 indicates the structural schematic diagram of another BP neural network provided in an embodiment of the present invention；
Fig. 8 shows the embodiment of the present invention to provide a kind of structural schematic diagram of field crop Tree Precise Fertilization system；
Fig. 9 illustrates the entity structure schematic diagram of a kind of electronic equipment.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
Every other embodiment obtained without creative efforts, shall fall within the protection scope of the present invention.
Currently, the maturation of BP neural network provides new method with development for crop yield and fertilising quantifier elimination.
Certain applies Matlab7.0 artificial neural network tool box in the prior art, devises BP artificial neuron training network,
It realizes with nitrogen fertilizer application amount, phosphate fertilizer amount, potash fertilizer amount and planting density in soil as input parameter, is prediction mesh with the yield of soybean
Target simulative relation predicts that N P and K investment in a certain range matches obtainable maximum production.
It is another that traditional Fertilizer effect function and the fertilizing method based on BP neural network are respectively adopted in the prior art
Comparative study best nitrogen, phosphorus, potassium fertilization ratio relationship of the processing tomato under maximum production, the results showed that, based on BP nerve
The fertilizing method that network method is established, complex nonlinear closes between solution soil nutrient content, dose, ultimate output three
It fastens with advantage, and its resulting fertilizer applications is more accurate, reasonable.
The another prior art proposes a kind of optimization method of BP neural network based on whole learning rate changing, with corn crop
For research object, choosing planting density, amount of nitrogen, phosphorus application amount, amount of potassium applied is experimental factor, and corn yield is influence index progress
Field trial, the results showed that the optimization method fitting function precision that they are taken is high, and optimum results are accurate, shows based on BP mind
Optimization method through network is applied to the reliability that corn planting density and dose optimize.
Although the relationship research between the crop yield and dose established currently based on BP neural network is more, mostly
It is all confined to theoretical research stage, with current production practices there are still some contradictions, existing method is primarily present following several
Point problem:
First, in terms of input and output mode, existing method be all mostly using soil nutrient, target output as input,
Using dose as output.But this kind of mode needs estimation maximum target yield in advance, and usual for the estimation of target output
There is no a quantitative equations, therefore are difficult estimation accurately, inaccurate so as to cause calculation of fertilization amount.
Second, the method established has regional and actual effect, and it is basic to be only applicable to the conditions such as kind, temperature and moisture
Identical ecotope needs improvement method generalization ability, and consider it if promoted in bigger, more complicated region
He influences factor of yield, such as pH value, soil types etc..
Third, existing method are fitted using single network, but the precision of prediction of single Neural is not high, extensive
It is indifferent.
4th, it is poor to the included noise processed ability of sample data.
In recent years, the method that small echo and BP neural network combine is widely used in multiple fields, in application process body
Very powerful Data Analysis Services ability is showed.
Present invention combination soil nutrient information establishes two different fertilising meters based on wavelet analysis and BP neural network
Calculation method, to be precisely controlled to crop fertilization, so that the development for China's Tree Precise Fertilization policy provides new thinking and side
Method.
Fig. 1 is a kind of flow chart of field crop Tree Precise Fertilization method provided in an embodiment of the present invention, as shown in Figure 1, should
Method includes:
S1 obtains the sample data in multiple groups sample farmland, and the sample data in each sample farmland includes: each sample farmland
Soil nitrogenous amount, the soil phosphorus content in each sample farmland, the soil potassium content in each sample farmland, each sample farmland
The crop yield of amount of nitrogen, the phosphorus application amount in each sample farmland, the amount of potassium applied in each sample farmland and each sample farmland；
S2 obtains the Wavelet Component of the crop yield in each sample farmland, each sample agricultural land soil using wavelet analysis
The Wavelet Component of nitrogen content, the Wavelet Component of each sample agricultural land soil phosphorus content and each sample agricultural land soil potassium content it is small
Wave component；
S3, according to the small echo of the Wavelet Component of the crop yield in each sample farmland, each sample agricultural land soil nitrogen content
The Wavelet Component of component, the Wavelet Component of each sample agricultural land soil phosphorus content and each sample agricultural land soil potassium content obtains
BP neural network after training；
S4, obtain the soil nitrogenous amount in target farmland, the soil phosphorus content in the target farmland, the target farmland soil
The history crop yield of earth potassium content and the target farmland obtains the soil in the target farmland using wavelet analysis respectively
The Wavelet Component of nitrogen content, the Wavelet Component of the soil phosphorus content in the target farmland, the target farmland soil potassium content
Wavelet Component and the target farmland history crop yield Wavelet Component；
S5, by the Wavelet Component of the soil nitrogenous amount in the target farmland, the target farmland soil phosphorus content it is small
Wave component, the target farmland soil potassium content Wavelet Component and the target farmland history crop yield small echo
Component is input in the BP neural network after training, and according to Non-Linear Programming, obtains the amount of nitrogen, described in the target farmland
The phosphorus application amount in target farmland, the target farmland amount of potassium applied.
The sample data in multiple groups sample farmland is acquired first, is said by taking 10 groups of sample farmlands as an example in the embodiment of the present invention
It is bright.
Soil nutrient sensor is laid in the soil in sample farmland, to acquire soil nutrient environmental information, and is collected each
The yield data of a sensor acquisition, in the embodiment of the present invention, with certain sample farmland for having disclosed nutrient grade from different places
Obtained " 3414 " test data in 10 corn trials fields for introduce the foundation of specific method, in " 3414 " test,
Nitrogen, phosphorus, potash fertilizer optimum spraying amount be set to 180kg/hm2、75kg/hm2And 75kg/hm2, its ratio be 12:5:5.
Production is sampled, is chemically examined, applied fertilizer and surveyed to sample farmland, by processing, obtains the sample number in 10 sample farmlands
According to.
Table 1 is the soil nutrient content table in 10 groups of sample farmlands in the embodiment of the present invention, and the soil nutrient in sample farmland contains
Amount includes that soil nitrogenous amount N, soil phosphorus content P and soil potassium content K, concrete content are as shown in table 1:
Table 1
Table 2 is the tables of data of the dose in 10 groups of sample farmlands provided in an embodiment of the present invention, and 14 kinds of processing are carried out to it,
The dose in sample farmland includes amount of nitrogen, phosphorus application amount and amount of potassium applied, amount of nitrogen, phosphorus application amount and the amount of potassium applied in 10 groups of sample farmlands
It is as shown in table 2:
Table 2
In sample farm field data other than including the soil nutrient content in sample farmland, the dose in sample farmland, also
Crop yield including sample farmland, be in the embodiment of the present invention using the corn trials field in past somewhere as sample farmland,
Therefore, the crop yield in 10 groups of sample farmlands is all known.
Due to during measuring crop yield, when will receive harvest, soil adhesion degree is different, blade standard water not
Uniformly, the influence of the enchancement factors such as grain osses and operating personnel makes measured value deviate true value.
Therefore wavelet analysis is utilized, the crop yield in this 10 groups of sample farmlands is decomposed, every group of sample farmland is obtained
Wavelet Component.
It should be noted that wavelet analysis is a kind of Time-Frequency Analysis Method of multiresolution based on Fourier transformation, it is more
Resolution analysis is a low frequency general picture part and several high frequency details resolved into former sequence on any scale resolution
A kind of partial data processing method.
Low frequency general picture part embodies the main information of original signal, and high frequency detail part then embodies original signal
Secondary information.
Using wavelet analysis, it is equivalent to crop yield data obtained to measurement, each sample agricultural land soil nitrogen content, every
A sample agricultural land soil phosphorus content and each sample agricultural land soil potassium content are filtered, and remove making an uproar in crop yield data
Part point.
After carrying out wavelet analysis, the Wavelet Component of the crop yield in available each sample farmland, each sample farmland
The Wavelet Component of soil nitrogenous amount, the Wavelet Component of each sample agricultural land soil phosphorus content and each sample agricultural land soil potassium content
Wavelet Component.
According to the sample data in the collected 10 groups of sample farmlands in front, the small wavelength-division of the crop yield in 10 groups of sample farmlands
Amount, the Wavelet Component of the soil nitrogenous amount in 10 groups of sample farmlands, 10 groups of sample farmlands soil phosphorus content Wavelet Component and 10
The Wavelet Component of the soil potassium content in group sample farmland, is trained BP neural network, the BP neural network after being trained.
It should be noted that BP neural network is a kind of feedforward network based on error back propagation, have very
Strong non-linear mapping capability.
Crop yield is that occur under the influencing each other of many factors (including soil nutrient, dose, environment etc.)
Token state, there is being difficult to determining nonlinear optics between these factors and yield, and the powerful mode of BP neural network is known
Not and data fitting can just overcome these problems.
When being predicted using dose of the trained BP neural network to target farmland, it is necessary first to obtain target
The soil nitrogenous amount in farmland, the soil phosphorus content in target farmland, the soil potassium content in target farmland and the history agriculture in target farmland
Crop yield obtains the Wavelet Component of the soil nitrogenous amount in target farmland, the soil phosphorus content in target farmland using wavelet analysis
Wavelet Component, target farmland soil potassium content Wavelet Component and target farmland history crop yield small wavelength-division
Amount.
By the Wavelet Component of the soil nitrogenous amount in target farmland, Wavelet Component, the target of the soil phosphorus content in target farmland
After the Wavelet Component of the history crop yield of the Wavelet Component and target farmland of the soil potassium content in farmland is input to training
In BP neural network, multiple groups target farmland fertilization amount data can be obtained in this way and are applied according to Non-Linear Programming from multiple groups target farmland
Select one group of data as the amount of nitrogen, phosphorus application amount and amount of potassium applied in the target farmland in fertilizer amount.
A kind of field crop Tree Precise Fertilization method provided in an embodiment of the present invention, by wavelet analysis to target agricultural land soil
Nutrient content and crop yield are analyzed, and can be removed because practical measurement soil nutrient content and crop yield exist in the process
Error existing for various aspects by the soil nutrient content after the sample data and progress wavelet analysis in sample farmland and makees produce
Amount is trained BP neural network, so that BP neural network closes the prediction between soil nutrient, dose and crop yield
System is more accurate, to realize the accurate prediction of target farmland fertilization amount.
On the basis of the above embodiments, it is preferable that the wavelet basis of the wavelet analysis is 5, decomposition scale 2.
Fig. 2 indicates the structural schematic diagram of two layers of multiresolution wavelet analysis in the embodiment of the present invention, as shown in Fig. 2, therefore,
Original crop yield can decompose as follows:
S=ca2+cd2+cd1,
Wherein, ca2 is trend component, and cd2 is periodic component, and cd1 is with the component that becomes.
By wavelet transformation theory it is found that small wave converting method can react rule of the crop yield on different scale.
For improvement method precision, the present invention carries out multiresolution wavelet point using yield data of the matlab to acquisition
Analysis, main contents are as follows:
(1) selection of wavelet basis
Using crop yield as the former sequence s of wavelet analysis, using dbN mode, when the N in dbN takes 1,3,5,7 and 9,
Fig. 3 indicates the wavelet analysis result schematic diagram of difference N value in the embodiment of the present invention, and the low frequency of former sequence s and its two layer analysis are general
The comparison of looks part dbN is as shown in Figure 3:
As seen from Figure 3, as N=1 or 3, Decomposition Sequence is serrated, with the increase of N, curve increasingly light
It is sliding, and former sequence is approached, and as N=7 or 9, curve is excessively smooth, gradually weakens the embodiment effect to former sequence peaks；Work as N
When=5, the substantially general picture of former sequence can have both been showed, and has also had for peak value and embodies effect well.Therefore selection db5 makees
For the wavelet basis of yield data wavelet analysis.
(2) selection of decomposition scale
Using db5 as the wavelet basis of crop yield data wavelet analysis, when decomposition scale chooses 1,2,3 or 4 respectively, figure
4 indicate the schematic diagram of the low frequency general picture in the embodiment of the present invention under different scale, and former sequence s is compared with low frequency general picture part caN
As shown in Figure 4.
As seen from Figure 4, with the increase of N, Decomposition Sequence is more and more smooth, but as N=3, and curve is existing gradually
Smooth trend, as N=4, curve is excessively smooth, cannot fully show the general picture of former sequence s.
By carrying out comprehensive analysis comparison to figure, point when choosing N=2 as crop yield data wavelet analysis is determined
Solve scale.
(3) wavelet analysis of crop yield
According to the above analysis, wavelet basis db3 is selected to carry out the wavelet multi_resolution analysis of two layers of scale, Fig. 5 to former sequence s
It indicates crop yield data wavelet analysis numerical value change curve synoptic diagram in the embodiment of the present invention, analyzes shown in result figure 5,
In, s is crop yield data, and ca2 is its low frequency general picture part, and cd1, cd2 are its high frequency detail part, s=ca2+cd2+
cd1。
The embodiment of the present invention filters out the noise data in crop yield data using wavelet analysis, improves farmland fertilization amount
Accuracy in computation.
Fig. 6 indicates a kind of structural schematic diagram of BP neural network provided in an embodiment of the present invention, as shown in fig. 6, above-mentioned
On the basis of embodiment, it is preferable that Wavelet Component, each sample farmland soil of the crop yield according to each sample farmland
The Wavelet Component of earth nitrogen content, the Wavelet Component of each sample agricultural land soil phosphorus content and each sample agricultural land soil potassium content
Wavelet Component, the BP neural network after obtaining training, specifically includes:
By the trend of the trend component of the soil nitrogenous amount in each sample farmland, the soil phosphorus content in each sample farmland point
Phosphorus is applied in amount, the trend component of soil potassium content in each sample farmland, the amount of nitrogen in each sample farmland, each sample farmland
The trend component of amount, the amount of potassium applied in each sample farmland and the crop yield in each sample farmland is as the defeated of BP neural network
Enter, output of the crop yield in each sample farmland as the BP neural network is trained the BP neural network, institute
Stating Wavelet Component includes trend component, periodic component and with the component that becomes；
BP neural network after obtaining training.
Specifically, BP neural network is trained, is included the steps that are as follows:
The trend component of crop yield, the trend component of soil nitrogenous amount, the soil in above 10 groups of sample farmlands is phosphorous
The input of the trend component of amount, the trend component of soil potassium content, amount of nitrogen, phosphorus application amount and amount of potassium applied as BP neural network,
Output of the trend component in each sample farmland as the BP neural network according to the method instructs BP neural network
Practice, the BP neural network after being trained.
It should be noted that trend component be after being decomposed using wavelet analysis to crop yield obtained component it
One.
The embodiment of the present invention (is namely become by the yield low frequency general picture part ca2 after wavelet analysis removal detailed information
Gesture component) as one of neural network input, to improve prediction effect.
Fig. 7 indicates the structural schematic diagram of another BP neural network provided in an embodiment of the present invention, as shown in fig. 7, to BP
Another method that neural network is trained is, on the basis of the above embodiments, it is preferable that described according to each sample agriculture
The Wavelet Component of the crop yield in field, the Wavelet Component of each sample agricultural land soil nitrogen content, each sample agricultural land soil are phosphorous
The Wavelet Component of the Wavelet Component of amount and each sample agricultural land soil potassium content, specifically includes:
By the trend of the trend component of the soil nitrogenous amount in each sample farmland, the soil phosphorus content in each sample farmland point
Phosphorus is applied in amount, the trend component of soil potassium content in each sample farmland, the amount of nitrogen in each sample farmland, each sample farmland
The input of amount, the amount of potassium applied in each sample farmland and the crop yield in each sample farmland as the first BP nerve subnetwork, often
Output of the trend component of the crop yield yield in one sample farmland as the first BP nerve subnetwork, to the first BP
Neural subnetwork is trained, and the Wavelet Component includes trend component, periodic component and with the component that becomes；
Using the crop yield in each sample farmland as the input of the 2nd BP nerve subnetwork, the crop in each sample farmland
Output of the periodic component of yield as the 2nd BP nerve subnetwork, is trained the 2nd BP nerve subnetwork；
Using the crop yield in each sample farmland as the input of the 3rd BP nerve subnetwork, each sample farmland with becoming
Output of the component as the 3rd BP nerve subnetwork, is trained the 3rd BP nerve subnetwork；
BP neural network after obtaining training, the BP neural network after the training is by the first BP nerve subnetting after training
The 3rd BP nerve subnetwork composition after the 2nd BP nerve subnetwork and training after network, training.
In the method, BP neural network is by the first BP nerve subnetwork, the 2nd BP nerve subnetwork and the 3rd BP mind
It is formed through subnetwork, in training, needs respectively to be trained these three networks.
The trend component of soil nitrogenous amount, the trend component of soil phosphorus content, the soil in above-mentioned 10 groups of sample farmlands are contained
Input of trend component, amount of nitrogen, phosphorus application amount, amount of potassium applied and the crop yield of potassium amount as the first BP nerve subnetwork, each
Output of the trend component of the crop yield yield in sample farmland as the first BP neural network, with this to the first BP nerve point
Network is trained.
Using the crop yield in above-mentioned 10 groups of sample farmlands as the input of the 2nd BP nerve subnetwork, each sample farmland
Output of the crop yield periodic component as the 2nd BP nerve subnetwork, the 2nd BP nerve subnetwork is trained with this.
Using the crop yield in above-mentioned 10 groups of sample farmlands as the input of the 3rd BP nerve subnetwork, each sample farmland
With output of the component as the 3rd BP nerve subnetwork that becomes, the 3rd BP nerve subnetwork is trained with this.
BP neural network after training has the first BP nerve subnetwork, the 2nd BP nerve subnetwork and the 3rd BP nerve subnetting
Network composition.
Each component that the embodiment of the present invention is obtained for wavelet analysis establishes different BP nerve subnetworks to it respectively
It is predicted, finally the result that each subnetwork exports is overlapped and is summarized, obtain the prediction of final result and crop yield
As a result.
Low frequency general picture component ca2 is the trend term of crop yield original sequence, should be with the correlation analysis result of crop yield
It is similar, related coefficient it is bigger have nitrogen content, amount of nitrogen and crop yield.Therefore selection nitrogen content, amount of nitrogen and work
Input variable of the produce amount as the first BP nerve subnetwork, using low frequency general picture component as output variable.
Periodic component cd2 and with the component cd1 that becomes be all crop yield original sequence details coefficients, it is related to each nutrient parameter
Property is little.
Cd2 is the periodic term of former sequence, is substantially only influenced by crop yield, therefore selects crop yield for the 2nd BP mind
Input variable through subnetwork.
Cd1 is former sequence with the ingredient that becomes, and there is dependences in short-term with yield, therefore also select yield for the 3rd BP
The input variable of neural subnetwork.
On the basis of the above embodiments, it is preferable that the Non-Linear Programming includes maximum production dose and maximum effect
Beneficial dose.
Specifically, using Non-Linear Programming, amount of nitrogen, phosphorus application amount and the amount of potassium applied in target farmland are obtained.
Due to that can obtain the amount of nitrogen in multiple groups target farmland, phosphorus application amount and apply potassium using the BP neural network after training
Amount is as a result, so need to select from multiple groups result a kind of as most using Non-Linear Programming according to the target of Non-Linear Programming
Good result.
Specifically, crop optimum fertilizing amount is calculated by Non-Linear Programming:
It, can be according to its soil nitrogenous amount and pre- for each sample farmland after BP neural network network struction success
It surveys yield and obtains its optimum fertilizing amount, thus conduct science Tree Precise Fertilization.
Relationship between yield and soil nitrogenous amount and amount of nitrogen can indicate are as follows:
Y=ANN (SN, SP, SK, FN, FP, FK),
Under normal circumstances, there are two types of the manifestation modes of optimum fertilizing amount: maximum production dose and greatest benefit dose.
If Non-Linear Programming is that maximum production dose calculates every group of fertilising result from the multiple groups result in target farmland
Corresponding crop yield is chosen so that the maximum amount of nitrogen of the crop yield in target farmland, phosphorus application amount, amount of potassium applied, as mesh
Mark the dose in farmland.Maximum production calculation of fertilization amount formula is as follows:
Wherein, fnmaxIndicate the Maximal amount of nitrogenous fertilizer, fpmaxIndicate the Maximal amount of phosphate fertilizer, fkmaxIndicate potash fertilizer
Maximal amount, sn indicate the soil nitrogenous amount in sample farmland, and sp indicates that the soil phosphorus content in sample farmland, sk indicate sample agriculture
The soil potassium content in field.
If Non-Linear Programming is greatest benefit dose, target agriculture can be made by selecting one group in the result that needs to apply fertilizer from multiple groups
Field benefit is maximum as a result, final amount of nitrogen, phosphorus application amount and amount of potassium applied as target farmland.
Greatest benefit calculation of fertilization amount formula is as follows:
Wherein, fnmaxIndicate the Maximal amount of nitrogenous fertilizer, fpmaxIndicate the Maximal amount of phosphate fertilizer, fkmaxIndicate potash fertilizer
Maximal amount, sn indicate the soil nitrogenous amount in sample farmland, and sp indicates that the soil phosphorus content in sample farmland, sk indicate sample agriculture
The soil potassium content in field, pyIndicate the market price of crop, pnIndicate the market price of nitrogenous fertilizer, ppIndicate the market price of phosphate fertilizer,
pkIndicate the market price of potash fertilizer.
The embodiment of the present invention is based on small echo-BP neural network crop Tree Precise Fertilization method, it is necessary first to using being laid in
Soil nutrient sensor in soil acquires soil nutrient environmental information, is then based on small echo-BP neural network and establishes soil and supports
Divide the relationship between information, crop forecast production and crop fertilization amount, so that it is determined that the dose of crop.
Using communication network, obtained crop fertilization amount information is transferred to fertilizer apparatus, to complete Tree Precise Fertilization.
The embodiment of the present invention provides a kind of field crop Tree Precise Fertilization method, identification and table based on soil nutrient information
It reaches, and forms its communication network, improve the monitoring dynamics to Soil Factors.Wavelet analysis is built in conjunction with BP neural network
Found two kinds of Different Crop Tree Precise Fertilization methods.
Fig. 8 shows the embodiment of the present invention to provide a kind of structural schematic diagram of field crop Tree Precise Fertilization system, such as Fig. 8 institute
Show, which includes: to obtain module 801, the first small echo module 802, training module 803, the second small echo module 804 and planning mould
Block 805, in which:
The sample data that module 801 is used to obtain multiple groups sample farmland is obtained, the sample data in each sample farmland includes:
The soil nitrogenous amount in each sample farmland, the soil phosphorus content in each sample farmland, each sample farmland soil potassium content, every
The amount of nitrogen in one sample farmland, the phosphorus application amount in each sample farmland, the amount of potassium applied in each sample farmland and each sample farmland
Crop yield；
First small echo module 802 be used for utilize wavelet analysis, obtain the crop yield in each sample farmland Wavelet Component,
The Wavelet Component of each sample agricultural land soil nitrogen content, the Wavelet Component of each sample agricultural land soil phosphorus content and each sample agriculture
The Wavelet Component of field soil potassium content；
Training module 803 is used for the Wavelet Component of the crop yield according to each sample farmland, each sample agricultural land soil
The Wavelet Component of nitrogen content, the Wavelet Component of each sample agricultural land soil phosphorus content and each sample agricultural land soil potassium content it is small
Wave component, the BP neural network after obtaining training；
Second small echo module 804 be used to obtain the soil nitrogenous amount in target farmland, the target farmland soil phosphorus content,
The history crop yield of the soil potassium content in the target farmland and the target farmland, institute is obtained using wavelet analysis respectively
State Wavelet Component, the target of the Wavelet Component of the soil nitrogenous amount in target farmland, the soil phosphorus content in the target farmland
The Wavelet Component of the history crop yield in the Wavelet Component of the soil potassium content in farmland and the target farmland；
Planning module 805 is used for the soil of the Wavelet Component of the soil nitrogenous amount in the target farmland, the target farmland
The Wavelet Component of earth phosphorus content, the target farmland soil potassium content Wavelet Component and the target farmland history farming
The Wavelet Component of produce amount is input in the BP neural network after training, and according to Non-Linear Programming, obtains the target farmland
Amount of nitrogen, the phosphorus application amount in the target farmland, the target farmland amount of potassium applied.
The specific implementation procedure of this system and the specific implementation procedure of the above method are identical, and details please refer to above method reality
Example is applied, details are not described herein for this system embodiment.
A kind of field crop Tree Precise Fertilization system provided in an embodiment of the present invention carries out crop yield by wavelet analysis
Analysis can remove because of existing error in all respects during practical measurement crop yield, pass through the sample in sample farmland
Data and carry out wavelet analysis after crop yield BP neural network is trained so that BP neural network to soil nutrient,
Projected relationship between dose and crop yield is more accurate, to realize the accurate prediction of target farmland fertilization amount.
Fig. 9 illustrates the entity structure schematic diagram of a kind of electronic equipment, as shown in figure 9, the server may include: processing
Device (processor) 910, communication interface (CommunicationsInterface) 920, memory (memory) 930 and bus
940, wherein processor 910, communication interface 920, memory 930 complete mutual communication by bus 940.Processor 910
The logical order in memory 930 can be called, to execute following method:
The sample data in multiple groups sample farmland is obtained, the sample data in each sample farmland includes: each sample farmland
Soil nitrogenous amount, the soil phosphorus content in each sample farmland, the soil potassium content in each sample farmland, each sample farmland are applied
The crop yield of nitrogen quantity, the phosphorus application amount in each sample farmland, the amount of potassium applied in each sample farmland and each sample farmland；
Using wavelet analysis, the Wavelet Component of the crop yield in each sample farmland is obtained；
BP nerve according to the Wavelet Component of the sample data in each sample farmland and each sample farmland, after obtaining training
Network；
Obtain the soil nitrogenous amount in target farmland, the soil phosphorus content in the target farmland, the target farmland soil
The prediction crop yield of potassium content and the target farmland is input in the BP neural network after training, and according to non-linear
Planning, obtain the amount of nitrogen in the target farmland, the phosphorus application amount in the target farmland, the target farmland amount of potassium applied.
In addition, the logical order in above-mentioned memory 930 can be realized by way of SFU software functional unit and conduct
Independent product when selling or using, can store in a computer readable storage medium.Based on this understanding, originally
Substantially the part of the part that contributes to existing technology or the technical solution can be in other words for the technical solution of invention
The form of software product embodies, which is stored in a storage medium, including some instructions to
So that a computer equipment (can be personal computer, server or the network equipment etc.) executes each implementation of the present invention
The all or part of the steps of example the method.And storage medium above-mentioned include: USB flash disk, mobile hard disk, read-only memory (ROM,
Read-OnlyMemory), random access memory (RAM, RandomAccessMemory), magnetic or disk etc. are various can
To store the medium of program code.
The present embodiment provides a kind of non-transient computer readable storage medium, the non-transient computer readable storage medium
Computer instruction is stored, the computer instruction makes the computer execute method provided by above-mentioned each method embodiment, example
Such as include:
The sample data in multiple groups sample farmland is obtained, the sample data in each sample farmland includes: each sample farmland
Soil nitrogenous amount, the soil phosphorus content in each sample farmland, the soil potassium content in each sample farmland, each sample farmland are applied
The crop yield of nitrogen quantity, the phosphorus application amount in each sample farmland, the amount of potassium applied in each sample farmland and each sample farmland；
Using wavelet analysis, the Wavelet Component of the crop yield in each sample farmland is obtained；
BP nerve according to the Wavelet Component of the sample data in each sample farmland and each sample farmland, after obtaining training
Network；
Obtain the soil nitrogenous amount in target farmland, the soil phosphorus content in the target farmland, the target farmland soil
The prediction crop yield of potassium content and the target farmland is input in the BP neural network after training, and according to non-linear
Planning, obtain the amount of nitrogen in the target farmland, the phosphorus application amount in the target farmland, the target farmland amount of potassium applied.
Those of ordinary skill in the art will appreciate that: realize that all or part of the steps of above method embodiment can pass through
The relevant hardware of program instruction is completed, and program above-mentioned can be stored in a computer readable storage medium, the program
When being executed, step including the steps of the foregoing method embodiments is executed；And storage medium above-mentioned includes: ROM, RAM, magnetic disk or light
The various media that can store program code such as disk.
The apparatus embodiments described above are merely exemplary, wherein described, unit can as illustrated by the separation member
It is physically separated with being or may not be, component shown as a unit may or may not be physics list
Member, it can it is in one place, or may be distributed over multiple network units.It can be selected according to the actual needs
In some or all of the modules achieve the purpose of the solution of this embodiment.Those of ordinary skill in the art are not paying creativeness
Labour in the case where, it can understand and implement.
Through the above description of the embodiments, those skilled in the art can be understood that each embodiment can
It realizes by means of software and necessary general hardware platform, naturally it is also possible to pass through hardware.Based on this understanding, on
Stating technical solution, substantially the part that contributes to existing technology can be embodied in the form of software products in other words, should
Computer software product may be stored in a computer readable storage medium, such as ROM/RAM, magnetic disk, CD, including several fingers
It enables and using so that a computer equipment (can be personal computer, server or the network equipment etc.) executes each implementation
Method described in certain parts of example or embodiment.
Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although
Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: it still may be used
To modify the technical solutions described in the foregoing embodiments or equivalent replacement of some of the technical features；
And these are modified or replaceed, technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution spirit and
Range.
Claims (10)
1. a kind of field crop Tree Precise Fertilization method characterized by comprising
The sample data in multiple groups sample farmland is obtained, the sample data in each sample farmland includes: the soil in each sample farmland
Nitrogen is applied in nitrogen content, the soil phosphorus content in each sample farmland, the soil potassium content in each sample farmland, each sample farmland
The crop yield of amount, the phosphorus application amount in each sample farmland, the amount of potassium applied in each sample farmland and each sample farmland；
Using wavelet analysis, Wavelet Component, each sample agricultural land soil nitrogen content of the crop yield in each sample farmland are obtained
Wavelet Component, the Wavelet Component of each sample agricultural land soil phosphorus content and each sample agricultural land soil potassium content small wavelength-division
Amount；
According to the Wavelet Component of the crop yield in each sample farmland, Wavelet Component of each sample agricultural land soil nitrogen content, every
The Wavelet Component of one sample agricultural land soil phosphorus content and the Wavelet Component of each sample agricultural land soil potassium content, after obtaining training
BP neural network；
Obtain the soil nitrogenous amount in target farmland, the soil phosphorus content in the target farmland, the target farmland soil contain potassium
The history crop yield of amount and the target farmland, the soil nitrogenous amount in the target farmland is obtained using wavelet analysis respectively
Wavelet Component, the Wavelet Component of soil phosphorus content in the target farmland, the target farmland soil potassium content small echo
The Wavelet Component of component and the history crop yield in the target farmland；
By the Wavelet Component of the soil nitrogenous amount in the target farmland, the Wavelet Component of the soil phosphorus content in the target farmland,
The Wavelet Component of the history crop yield in the Wavelet Component of the soil potassium content in the target farmland and the target farmland is defeated
Enter in the BP neural network to after training, and according to Non-Linear Programming, obtains the amount of nitrogen in the target farmland, the target agriculture
The phosphorus application amount in field, the target farmland amount of potassium applied.
2. method according to claim 1, which is characterized in that the wavelet basis of the wavelet analysis is 5, decomposition scale 2.
3. method according to claim 1, which is characterized in that the small wavelength-division of the crop yield according to each sample farmland
Amount, the Wavelet Component of each sample agricultural land soil nitrogen content, the Wavelet Component of each sample agricultural land soil phosphorus content are as every
The Wavelet Component of this agricultural land soil potassium content, the BP neural network after obtaining training, specifically includes:
By the trend component of the soil nitrogenous amount in each sample farmland, the soil phosphorus content in each sample farmland trend component,
The trend component of the soil potassium content in each sample farmland, the amount of nitrogen in each sample farmland, each sample farmland phosphorus application amount,
Input of the trend component of the crop yield in the amount of potassium applied in each sample farmland and each sample farmland as BP neural network,
Output of the crop yield in each sample farmland as the BP neural network, is trained the BP neural network, described
Wavelet Component includes trend component, periodic component and with the component that becomes；
BP neural network after obtaining training.
4. method according to claim 1, which is characterized in that the small wavelength-division of the crop yield according to each sample farmland
Amount, the Wavelet Component of each sample agricultural land soil nitrogen content, the Wavelet Component of each sample agricultural land soil phosphorus content are as every
The Wavelet Component of this agricultural land soil potassium content, specifically includes:
By the trend component of the soil nitrogenous amount in each sample farmland, the soil phosphorus content in each sample farmland trend component,
The trend component of the soil potassium content in each sample farmland, the amount of nitrogen in each sample farmland, each sample farmland phosphorus application amount,
Input of the crop yield in the amount of potassium applied in each sample farmland and each sample farmland as the first BP nerve subnetwork, per the same
Output of the trend component of the crop yield in this farmland as the first BP nerve subnetwork, to the first BP nerve subnetting
Network is trained, and the Wavelet Component includes trend component, periodic component and with the component that becomes；
Using the crop yield in each sample farmland as the input of the 2nd BP nerve subnetwork, the crop yield in each sample farmland
Output of the periodic component as the 2nd BP nerve subnetwork, the 2nd BP nerve subnetwork is trained；
Using the crop yield in each sample farmland as the input of the 3rd BP nerve subnetwork, the crop yield in each sample farmland
With output of the component as the 3rd BP nerve subnetwork that becomes, the 3rd BP nerve subnetwork is trained；
BP neural network after obtaining training, BP neural network after the training by after training the first BP nerve subnetwork,
The 3rd BP nerve subnetwork composition after the 2nd BP nerve subnetwork and training after training.
5. method according to claim 1, which is characterized in that the Non-Linear Programming includes maximum production dose and maximum
Benefit dose.
6. method according to claim 5, which is characterized in that the calculation formula of the maximum production dose is as follows:
Wherein, fnmaxIndicate the Maximal amount of nitrogenous fertilizer, fpmaxIndicate the Maximal amount of phosphate fertilizer, fkmaxIndicate the maximum of potash fertilizer
Amount of application, cn indicate the soil nitrogenous amount in sample farmland, and cp indicates that the soil phosphorus content in sample farmland, ck indicate sample farmland
Soil potassium content.
7. method according to claim 5, which is characterized in that the calculation formula of the greatest benefit dose is as follows:
Wherein, fnmaxIndicate the Maximal amount of nitrogenous fertilizer, fpmaxIndicate the Maximal amount of phosphate fertilizer, fkmaxIndicate the maximum of potash fertilizer
Amount of application, cn indicate the soil nitrogenous amount in sample farmland, and cp indicates that the soil phosphorus content in sample farmland, ck indicate sample farmland
Soil potassium content, pyIndicate the market price of crop, pnIndicate the market price of nitrogenous fertilizer, ppIndicate the market price of phosphate fertilizer, pkTable
Show the market price of potash fertilizer.
8. a kind of field crop Tree Precise Fertilization system characterized by comprising
Module is obtained, for obtaining the sample data in multiple groups sample farmland, the sample data in each sample farmland includes: per the same
The soil nitrogenous amount in this farmland, the soil phosphorus content in each sample farmland, the soil potassium content in each sample farmland, each sample
Produce are made in the amount of nitrogen in farmland, the phosphorus application amount in each sample farmland, the amount of potassium applied in each sample farmland and each sample farmland
Amount；
First small echo module obtains the Wavelet Component of the crop yield in each sample farmland, per the same for utilizing wavelet analysis
The Wavelet Component of this agricultural land soil nitrogen content, the Wavelet Component of each sample agricultural land soil phosphorus content and each sample agricultural land soil
The Wavelet Component of potassium content；
Training module, for Wavelet Component, each sample agricultural land soil nitrogen content according to the crop yield in each sample farmland
Wavelet Component, the Wavelet Component of each sample agricultural land soil phosphorus content and each sample agricultural land soil potassium content small wavelength-division
Amount, the BP neural network after obtaining training；
Second small echo module, for obtaining the soil nitrogenous amount in target farmland, the soil phosphorus content in the target farmland, the mesh
The soil potassium content in farmland and the history crop yield in the target farmland are marked, obtains the target respectively using wavelet analysis
The Wavelet Component of the soil nitrogenous amount in farmland, the Wavelet Component of the soil phosphorus content in the target farmland, the target farmland
The Wavelet Component of the history crop yield in the Wavelet Component of soil potassium content and the target farmland；
Planning module, for the soil of the Wavelet Component of the soil nitrogenous amount in the target farmland, the target farmland is phosphorous
The Wavelet Component of amount, the target farmland soil potassium content Wavelet Component and the target farmland history crop yield
Wavelet Component be input in the BP neural network after training, and according to Non-Linear Programming, obtain the target farmland applies nitrogen
Amount, the phosphorus application amount in the target farmland, the target farmland amount of potassium applied.
9. a kind of electronic equipment characterized by comprising
At least one processor, at least one processor, communication interface and bus；Wherein,
The processor, memory, communication interface complete mutual communication by the bus；
The communication interface is for the information transmission between the test equipment and the communication equipment of display device；
The memory is stored with the program instruction that can be executed by the processor, and the processor calls described program to instruct energy
It is enough to execute such as method of any of claims 1-7.
10. a kind of non-transient computer readable storage medium, which is characterized in that the non-transient computer readable storage medium is deposited
Computer instruction is stored up, the computer instruction makes the computer execute the method as described in claim 1 to 7 is any.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910303769.3A CN109964611A (en) | 2019-04-16 | 2019-04-16 | A kind of field crop Tree Precise Fertilization method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910303769.3A CN109964611A (en) | 2019-04-16 | 2019-04-16 | A kind of field crop Tree Precise Fertilization method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109964611A true CN109964611A (en) | 2019-07-05 |
Family
ID=67084767
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910303769.3A Pending CN109964611A (en) | 2019-04-16 | 2019-04-16 | A kind of field crop Tree Precise Fertilization method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109964611A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110741795A (en) * | 2019-11-13 | 2020-02-04 | 北京农业智能装备技术研究中心 | Facility crop fertilization method and system based on in-situ soil nutrient concentration |
| CN111754060A (en) * | 2019-10-18 | 2020-10-09 | 广州极飞科技有限公司 | Variable rate fertilization method and device, electronic equipment and storage medium |
| CN113221446A (en) * | 2021-04-21 | 2021-08-06 | 兰州里丰正维智能科技有限公司 | Method and system for acquiring water and fertilizer strategies of saline soil, storage medium and electronic equipment |
| CN115316097A (en) * | 2022-08-02 | 2022-11-11 | 中联智慧农业股份有限公司 | Method and processor for determining nitrogen fertilizer application amount aiming at soil of planting area |
| CN117322214A (en) * | 2023-11-30 | 2024-01-02 | 余姚市农业技术推广服务总站 | Crop fertilizer accurate application method and system based on neural network |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103646299A (en) * | 2013-12-19 | 2014-03-19 | 浙江省公众信息产业有限公司 | Neural network based crop prediction method and device |
| CN103823371A (en) * | 2014-02-12 | 2014-05-28 | 无锡中科智能农业发展有限责任公司 | Neural network model-based agricultural precise fertilization system and fertilization method thereof |
| CN208298002U (en) * | 2018-07-06 | 2018-12-28 | 山东省农业科学院科技信息研究所 | A kind of water-fertilizer integral control system Internet-based |
-
2019
- 2019-04-16 CN CN201910303769.3A patent/CN109964611A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103646299A (en) * | 2013-12-19 | 2014-03-19 | 浙江省公众信息产业有限公司 | Neural network based crop prediction method and device |
| CN103823371A (en) * | 2014-02-12 | 2014-05-28 | 无锡中科智能农业发展有限责任公司 | Neural network model-based agricultural precise fertilization system and fertilization method thereof |
| CN208298002U (en) * | 2018-07-06 | 2018-12-28 | 山东省农业科学院科技信息研究所 | A kind of water-fertilizer integral control system Internet-based |
Non-Patent Citations (5)
| Title |
|---|
| 于合龙等: "《基于改进的BP 神经网络集成的作物精准施肥模型》", 《农 业 工 程 学 报》 * |
| 于平福等: "《基于小波广义回归神经网络的粮食产量预测模型》", 《湖北农业科学》 * |
| 刘东升等: "《融合物联感知与移动监控的智慧农业公共服务技术研究》", 30 November 2015, 浙江工商大学出版社 * |
| 苑超等: "《基于MapReduce的BP神经网络在精准施肥中的应用＊》", 《中国农机化学报》 * |
| 马成林等: "《基于数据包络分析和人工神经网络的变量施肥决策方法研究》", 《农业工程学报》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111754060A (en) * | 2019-10-18 | 2020-10-09 | 广州极飞科技有限公司 | Variable rate fertilization method and device, electronic equipment and storage medium |
| CN110741795A (en) * | 2019-11-13 | 2020-02-04 | 北京农业智能装备技术研究中心 | Facility crop fertilization method and system based on in-situ soil nutrient concentration |
| CN110741795B (en) * | 2019-11-13 | 2021-04-06 | 北京农业智能装备技术研究中心 | Facility crop fertilization method and system based on in-situ soil nutrient concentration |
| CN113221446A (en) * | 2021-04-21 | 2021-08-06 | 兰州里丰正维智能科技有限公司 | Method and system for acquiring water and fertilizer strategies of saline soil, storage medium and electronic equipment |
| CN115316097A (en) * | 2022-08-02 | 2022-11-11 | 中联智慧农业股份有限公司 | Method and processor for determining nitrogen fertilizer application amount aiming at soil of planting area |
| CN117322214A (en) * | 2023-11-30 | 2024-01-02 | 余姚市农业技术推广服务总站 | Crop fertilizer accurate application method and system based on neural network |
| CN117322214B (en) * | 2023-11-30 | 2024-02-09 | 余姚市农业技术推广服务总站 | Crop fertilizer accurate application method and system based on neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109964611A (en) | A kind of field crop Tree Precise Fertilization method and system | |
| Liu et al. | A neural network for setting target corn yields | |
| CN110909679B (en) | Remote sensing identification method and system for fallow crop rotation information of winter wheat historical planting area | |
| CN1324507C (en) | Device and method for forming prescription for monitoring crop growing state and nutrition spreading fertilizer | |
| WO2019081567A1 (en) | Yield estimation in the cultivation of crop plants | |
| CN107392376A (en) | A kind of crops Meteorological Output Forecasting Methodology and system | |
| CN105181607B (en) | Readily available nutrient of soil inversion method and device | |
| CN110119767B (en) | Intelligent cucumber greenhouse temperature detection device based on LVQ neural network | |
| CN106779452A (en) | The method for setting up crops data growth model | |
| CN117322214B (en) | Crop fertilizer accurate application method and system based on neural network | |
| CN117036088A (en) | Data acquisition and analysis method for identifying growth situation of greening plants by AI | |
| Avigal et al. | Learning seed placements and automation policies for polyculture farming with companion plants | |
| Abdelhafeez et al. | Identify the most Productive Crop to Encourage Sustainable Farming Methods in Smart Farming using Neutrosophic Environment | |
| Gautam et al. | Optimized high performance deep neural network for crop recommendation | |
| CN113052433B (en) | Crop unit yield estimation method based on key time phase and farmland landscape characteristic parameters | |
| Timsina et al. | Addressing sustainability of rice-wheat systems: analysis of long-term experimentation and simulation | |
| Hazman | Crop irrigation schedule expert system | |
| CN115017798A (en) | Soil plough layer moisture temperature prediction method based on deep learning | |
| Park et al. | Determination on environmental factors and growth factors affecting tomato yield using pattern recognition techniques | |
| Chaudhary et al. | A critical review on hybrid framework for precise farming with application of Machine Learning (ML) and Internet of Things (IoT) | |
| Wang | The models of niche and their application | |
| Jasti et al. | Crop Intelligent: Weather based Crop Selection using Machine Learning | |
| CN109191320A (en) | A kind of agricultural production management guidance method and system | |
| Zeng et al. | Calibration of the Crop model in the Community Land Model | |
| Gunasekara et al. | Expert prediction system for spice plants grown in Sri Lanka: An incentive for planters |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190705 |
|
| RJ01 | Rejection of invention patent application after publication |