CN108665107A - Crop yield prediction technique and system - Google Patents
Crop yield prediction technique and system Download PDFInfo
- Publication number
- CN108665107A CN108665107A CN201810460935.6A CN201810460935A CN108665107A CN 108665107 A CN108665107 A CN 108665107A CN 201810460935 A CN201810460935 A CN 201810460935A CN 108665107 A CN108665107 A CN 108665107A
- Authority
- CN
- China
- Prior art keywords
- statistics phase
- phase
- statistics
- crop yield
- meteorological data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 64
- 238000003062 neural network model Methods 0.000 claims abstract description 45
- 230000002123 temporal effect Effects 0.000 claims abstract description 17
- 238000012549 training Methods 0.000 claims description 23
- 238000001556 precipitation Methods 0.000 claims description 17
- 239000002689 soil Substances 0.000 claims description 17
- 238000004519 manufacturing process Methods 0.000 claims description 14
- 238000003860 storage Methods 0.000 claims description 9
- 230000004913 activation Effects 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 6
- 230000001052 transient effect Effects 0.000 claims description 6
- 238000005259 measurement Methods 0.000 claims description 4
- 210000002569 neuron Anatomy 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000004891 communication Methods 0.000 claims description 3
- 239000012141 concentrate Substances 0.000 claims 1
- 210000005036 nerve Anatomy 0.000 claims 1
- 210000004218 nerve net Anatomy 0.000 claims 1
- 235000017060 Arachis glabrata Nutrition 0.000 description 6
- 244000105624 Arachis hypogaea Species 0.000 description 6
- 235000010777 Arachis hypogaea Nutrition 0.000 description 6
- 235000018262 Arachis monticola Nutrition 0.000 description 6
- 241000209140 Triticum Species 0.000 description 6
- 235000021307 Triticum Nutrition 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 6
- 235000020232 peanut Nutrition 0.000 description 6
- 238000009313 farming Methods 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000004590 computer program Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000009826 distribution Methods 0.000 description 4
- 241000196324 Embryophyta Species 0.000 description 3
- PEDCQBHIVMGVHV-UHFFFAOYSA-N Glycerine Chemical compound OCC(O)CO PEDCQBHIVMGVHV-UHFFFAOYSA-N 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 241000238631 Hexapoda Species 0.000 description 2
- 238000012897 Levenberg–Marquardt algorithm Methods 0.000 description 2
- 240000007594 Oryza sativa Species 0.000 description 2
- 235000007164 Oryza sativa Nutrition 0.000 description 2
- 241000607479 Yersinia pestis Species 0.000 description 2
- 238000012271 agricultural production Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 230000007812 deficiency Effects 0.000 description 2
- 201000010099 disease Diseases 0.000 description 2
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 2
- 238000011478 gradient descent method Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000013139 quantization Methods 0.000 description 2
- 235000009566 rice Nutrition 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 229920000742 Cotton Polymers 0.000 description 1
- 241001269238 Data Species 0.000 description 1
- 244000068988 Glycine max Species 0.000 description 1
- 235000010469 Glycine max Nutrition 0.000 description 1
- 241000219146 Gossypium Species 0.000 description 1
- 240000007049 Juglans regia Species 0.000 description 1
- 235000009496 Juglans regia Nutrition 0.000 description 1
- 240000008042 Zea mays Species 0.000 description 1
- 235000005824 Zea mays ssp. parviglumis Nutrition 0.000 description 1
- 235000002017 Zea mays subsp mays Nutrition 0.000 description 1
- 235000013339 cereals Nutrition 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 235000005822 corn Nutrition 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000017074 necrotic cell death Effects 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 239000002417 nutraceutical Substances 0.000 description 1
- 235000021436 nutraceutical agent Nutrition 0.000 description 1
- 230000000050 nutritive effect Effects 0.000 description 1
- 235000020265 peanut milk Nutrition 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 102000004169 proteins and genes Human genes 0.000 description 1
- 108090000623 proteins and genes Proteins 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 235000021122 unsaturated fatty acids Nutrition 0.000 description 1
- 150000004670 unsaturated fatty acids Chemical class 0.000 description 1
- 235000020234 walnut Nutrition 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- General Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Business, Economics & Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Biomedical Technology (AREA)
- Mining & Mineral Resources (AREA)
- Animal Husbandry (AREA)
- Agronomy & Crop Science (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Primary Health Care (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Artificial Intelligence (AREA)
- Marine Sciences & Fisheries (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the present invention provides crop yield prediction technique and system.Wherein, method includes:According to the meteorological data in the first statistics phase, the temporal information of the meteorological data in the second statistics phase belonging to the first statistics phase and the first statistics phase, the characteristic of crops is obtained;The characteristic is inputted into trained neural network model, obtains crop yield prediction result.Crop yield prediction technique and system provided in an embodiment of the present invention can quickly and easily obtain crop yield prediction result with high accuracy by Neural Network model predictive crop yield.
Description
Technical field
The present embodiments relate to field of agricultural production technologies more particularly to crop yield prediction techniques and system.
Background technology
In agricultural production, the prediction for carrying out crop yield has extremely strong realistic meaning.Crop yield is carried out
Prediction, not only contributes to the cultivated area that peasant timely adjusts crops according to price change, improves the income of peasant, may be used also
To expand or reduce the cultivated area of crops in time according to the export situation of crops.
Currently, the prediction for most crop yields is all mostly by means of plantation experience, deviation is very big, so finding
It is urgent problem that effective method carries out prediction to crop yield.For example, northern China plants a large amount of peanut,
The nutritive value of peanut is higher than grain class, it contains the content of the fat of a large amount of protein, especially unsaturated fatty acid very
Height, the very suitable various nutraceutical of manufacture, such as peanut milk, peanut and walnut dew, but there is presently no accurate prediction flowers
The method of raw yield.
For some crops, prediction model is had been set up.For example, occurring many wheat growth simulation systems in the world
System, such as the DSSAT systems in the U.S., the APSIM systems of Australia, the WheatGrow systems of China.They pass through parsing
The mechanism relationship of " meteorology-soil-technical measures " and wheat physiological and ecological process, the growth and development to wheat and yield composition
Process carries out quantitative prediction.But according to the yield of existing model prediction wheat, the parameter used is more, and data handling procedure is non-
Often complicated, the accuracy and speed of prediction need to be improved.
In addition, when predicting crop yield, time variable was had ignored as unit of year or entire growth period
Crops are led to the precision that crop yield is predicted not by the otherness of meteorology variation in the different growth stage that growth period includes
Foot.
Invention content
For the problem of the precision deficiency of crop yield of the existing technology prediction, the embodiment of the present invention provides farming
Object production prediction method and system.
According to the first aspect of the invention, the embodiment of the present invention provides a kind of crop yield prediction technique, including:
According to the meteorological data in the first statistics phase, the meteorological data in the second statistics phase belonging to the first statistics phase and
The temporal information of one statistics phase, obtains the characteristic of crops;
The characteristic is inputted into trained neural network model, obtains crop yield prediction result.
According to the third aspect of the invention we, the embodiment of the present invention provides a kind of crop yield forecasting system, including:
Characteristic extracting module, for according to the meteorological data in the first statistics phase, the second statistics belonging to the first statistics phase
The temporal information of meteorological data and the first statistics phase in phase, obtains the characteristic of crops;
Production forecast module obtains farming produce for the characteristic to be inputted trained neural network model
Measure prediction result.
According to the third aspect of the invention we, the embodiment of the present invention provides a kind of pre- measurement equipment of crop yield, including:
At least one processor;And
At least one processor being connect with the processor communication, wherein:
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 the analysis method for being able to carry out crop yield of embodiment of the present invention prediction technique and its all alternative embodiments.
According to the fourth aspect of the invention, the embodiment of the present invention provides a kind of non-transient computer readable storage medium, institute
Non-transient computer readable storage medium storage computer instruction is stated, the computer instruction makes the computer execute the present invention
The analysis method of embodiment crop yield prediction technique and its all alternative embodiments.
Crop yield prediction technique provided in an embodiment of the present invention, by Neural Network model predictive crop yield,
Crop yield prediction result with high accuracy can quickly and easily be obtained.Further, by selecting suitable first statistics
Phase can predict the yield or the yield in current growth period of crops current year, when predicting the yield in current growth period, it is contemplated that raw
Crops are influenced by meteorology variation in the different growth stage for including for a long time, and prediction result is more acurrate.
Description of the drawings
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 technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is this hair
Some bright 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 the flow chart of crop yield prediction technique of the embodiment of the present invention;
Fig. 2 is the functional block diagram of the prediction meanss of crop yield of the embodiment of the present invention;
Fig. 3 is the structure diagram of the pre- measurement equipment of crop yield of the embodiment of the present invention.
Specific implementation mode
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
The every other embodiment obtained without creative efforts, shall fall within the protection scope of the present invention.
Fig. 1 is the flow chart of crop yield prediction technique of the embodiment of the present invention.As shown in Figure 1, a kind of crop yield
Prediction technique includes:Step S101, according in the meteorological data in the first statistics phase, the second statistics phase belonging to the first statistics phase
Meteorological data and first statistics the phase temporal information, obtain the characteristic of crops.
It should be noted that crop yield prediction technique provided in an embodiment of the present invention is suitable for appointing in certain area
A kind of yield of crops of anticipating is predicted.Certain area can be the whole nation, several provinces, several cities, several counties,
Several small towns or several villages, but not limited to this.
Crop yield prediction technique provided in an embodiment of the present invention, prediction be crops certain time yield.Example
Such as, the yield in prediction crops this year or the yield of current Growing season.
Preferably, it is influenced caused by the variation of crop acreage to eliminate, crop yield refers to the list of crops
Production.For example, the per unit area yield of crops is per mu yield.
Influence crop yield because being known as edaphic condition, insect pest or plant disease, meteorological condition.But edaphic condition,
Insect pest or plant disease are difficult to obtain the data of quantization, and the quantized data of various meteorological conditions, i.e. meteorological data are easy to obtain.
Meteorological observatory would generally retain the various meteorological datas of certain area day by day.
Specifically, the characteristic of crops is obtained according to meteorological data.
Meteorological data is divided into two parts:The meteorological data in meteorological data and the second statistics phase in first statistics phase.
First statistics phase and the second statistics phase are certain period of time.Second statistics phase counted phase structure by multiple first
At.Each first statistics phase, corresponding duration may be the same or different.
The temporal information of first statistics phase, for the information of the first statistics phase itself.
For example, a length of moon when the first statistics phase is corresponding, the second statistics phase in a length of season when corresponding;First statistics phase was
April, the second statistics phase are the second quarter in season belonging to April;The temporal information of first statistics phase can be 4 months month information,
Or April belongs to 1st month of the second quarter;According to the meteorological data of April and the second quarter and, month information obtain agriculture April
The characteristic of crop, 1st month that the second quarter can also be belonged to according to the meteorological data and April of April and the second quarter obtain agriculture
The characteristic of crop.
For another example, in a length of week when the first statistics phase is corresponding, the second statistics phase, corresponding duration was growthdevelopmental stage;Second statistics
Phase is Seedling Stage, and the first statistics phase was the second week of Seedling Stage;It is seedling that the temporal information of first statistics phase, which was the first statistics phase,
The second week of phase;It is the second week of Seedling Stage according to the meteorological data and the first statistics phase of Seedling Stage and the second week of Seedling Stage
Obtain the characteristic of crops.
First statistics phase was April, and the second statistics phase was the second quarter in season belonging to April;The temporal information of first statistics phase,
Can be 4 months month information, or belong to 1st month of the second quarter April;According to the meteorological data in April and the second quarter,
And information April in month obtains the characteristic of crops, can also according to the meteorological number in April and the second quarter, according to and belong to two April
1st month of season obtains the characteristic of crops.
It is understood that when the ranging from larger region of crop yield prediction, it is such as national, multiple time saving, if agriculture
When the distribution of the major production areas of crop is not concentrated, multiple representational cities can be chosen from the major production areas of crops,
The characteristic of crops is obtained according to the meteorological data in selected each city, if the distribution of the major production areas of crops is concentrated
When or crop yield prediction ranging from smaller region, such as a city when, can be from the major production areas of crops
Multiple cities are chosen, the meteorological data in selected each city is handled, are such as averaged or weighted average, agriculture is obtained
The characteristic of crop, a city can also be selected from the major production areas of crops, and (crop yield is predicted ranging from
When one city, the city is directly selected), the characteristic of crops is obtained according to the meteorological data in the city.
It is understood that further including before step S101:Obtain meteorological data in the first statistics phase, the first statistics phase
The temporal information of meteorological data and the first statistics phase in the second affiliated statistics phase.Meteorological data in first statistics phase,
The meteorological data in the second statistics phase belonging to one statistics phase, can be obtained by daily meteorological observation, can also be from meteorology
Department, Meteorological Services Shang Deng mechanisms obtain.
Step S102, characteristic is inputted into trained neural network model, obtains crop yield prediction result.
After the characteristic for obtaining crops, by characteristic composition characteristic vector, trained neural network mould is inputted
Type obtains crop yield prediction result according to the output of trained neural network model.
Pass through acquisition when for training the quantity of the training sample of neural network model very big according to experimental result
The accuracy rate of training Neural Network model predictive crop yield is higher.For example, by the meteorological data and the moon of every month in 30 years
Part information as training sample, by the accuracy rate of the training Neural Network model predictive crop yield of acquisition up to 90% with
On;Using the meteorological data of every month in 10 years and month information as training sample, pass through the training neural network model of acquisition
Predict the accuracy rate of crop yield up to 75% or more.
The embodiment of the present invention can be quickly and easily by Neural Network model predictive crop yield based on meteorological data
Obtain crop yield prediction result with high accuracy.Further, by selecting the suitable first statistics phase, farming can be predicted
The yield of object current year or the yield in current growth period, when predicting the yield in current growth period, it is contemplated that the difference that growth period includes
Crops are influenced by meteorology variation in growthdevelopmental stage, and prediction result is more acurrate.
Based on above-described embodiment, the specific steps for obtaining trained neural network model include:According to history meteorology number
According to and corresponding crops historical yield build training sample set.
Specifically, trained neural network model is obtained by following steps.
First, training sample set is built according to history meteorological data and corresponding crops historical yield.
Is counted for each of historical data first, according to the meteorological number in each first statistics phase in historical data the phase
According to the meteorological data and the temporal information of the first statistics phase in the second statistics phase belonging to, the first statistics phase, agriculture is obtained
The characteristic of the first statistics phase of crop.
For example, when the first statistics phase was the moon, the historical data of several years before obtaining;For each of historical data
Month, the meteorological data in the season belonging to the meteorological data of every month in historical data, this month and the month of this month believe
Breath, obtains this month corresponding characteristic of crops.
The first statistics phase corresponding crops historical yield is divided according to section, and each section is distributed and is marked
Label.The quantity in the section of division is determined according to the information of the crops such as the type of crops, kind itself.
For example, 600 jin of per mu yield or more be high yield, distribution label be 3;400 to 600 jin are middle production, the label distributed is
2;400 jin or less be low yield, distribution label be 1.
The data input neural network model that training sample is concentrated is trained, trained neural network mould is obtained
Type.
Specifically, the characteristic of each first statistics phase and the first statistics phase in the historical data of crops are obtained
After the label of corresponding crops historical yield, the characteristic of each first statistics phase in the historical data of crops is inputted
Neural network model, and according to the mark of the output of neural network model crops historical yield corresponding with the first statistics phase
Label, the parameter for adjusting neural network model are adjusted, until the output of neural network model meets preset condition, will be exported
Meet the neural network model of preset condition as trained neural network model.
Based on above-described embodiment, the meteorological data in the first statistics phase includes at least:The highest temperature, most of first statistics phase
Low temperature, sunshine-duration and precipitation;The meteorological data in the second statistics phase belonging to first statistics phase, according to belonging to the second system
The temperature on average of each first statistics phase of meter phase obtains.
As a preferred embodiment, the meteorological data in the first statistics phase includes at least the highest gas of the first statistics phase
Temperature, the lowest temperature, sunshine-duration and precipitation, but not limited to this.
The highest temperature, the lowest temperature, sunshine-duration and the precipitation of first statistics phase counts the phase day by day most according to first
High temperature, the lowest temperature, sunshine-duration and precipitation obtain.I.e. to the first statistics phase highest temperature day by day, the lowest temperature, day
It is suitably handled according to time and precipitation, obtains the highest temperature, the lowest temperature, sunshine-duration and the precipitation of the first statistics phase
Amount
For example, when the first statistics phase was October, it can be by October 1 to the daily highest temperature on October 31, minimum gas
Temperature, the average value of sunshine-duration and precipitation, the highest temperature, the lowest temperature, sunshine-duration and precipitation as the first statistics phase
Amount;It can also be using the average values of maximum 5 values in October 1 to the October 31 daily highest temperature as the first statistics phase
The highest temperature, using the average values of 5 values minimum in October 1 to the October 31 daily lowest temperature as the first statistics
The lowest temperature of phase;But not limited to this.
Sunshine-duration can be the accumulative sunlight hourage of the first statistics phase.
Precipitation can be the accumulative precipitation of the first statistics phase, and unit is millimeter.
For different crops, the meteorological data in the first statistics phase can also include number of days, the Continuous Drought of frost
Number of days, continuous rainfall number of days etc..
For the region of diverse geographic location, the meteorological data in the first statistics phase can also include peculiar with geographical location
Meteorological data.For example, for coastal area, the meteorological data in the first statistics phase can also include typhoon number of days, ocean current temperature
Degree etc.;For cold district, such as the Northeast in China, the meteorological data in the first statistics phase can also include snowfall number of days.
The meteorological data in the second statistics phase belonging to first statistics phase includes at least the temperature record of the second statistics phase.
The temperature on average for each first statistics phase for belonging to for the second statistics phase can be handled by suitable method, it will
Handling result conduct, the temperature record as the second statistics phase.
For example, by the average value of 1 to March monthly mean temperature, the temperature record as the first quarter.
The meteorological data in the second statistics phase belonging to first statistics phase can also include precipitation, the day of the second statistics phase
According to data, but not limited to this.
The meteorological conditions such as temperature, sunshine, precipitation can have an impact the growth of crops, and then influence farming produce
Amount, can be preferably using meteorological conditions such as temperature, sunshine, precipitation as characteristic
Based on above-described embodiment, neural network model is VGG models;The activation primitive of each neuron node of VGG models
For MPELU functions.
As a preferred embodiment, neural network model uses VGG models.
VGG models are derived from the paper that Oxford University's Visual Geometry Group (visual geometric group) is write, thus gain the name
VGG models.
VGG models use modular structure, conveniently increase and change.VGG models finally use average pondization
Instead of full articulamentum, accuracy can be improved.Practical VGG models are finally still adding a full articulamentum, are mainly
Convenience is preferably finely adjusted.Although removing full connection, Dropout has still been used in VGG models.In order to avoid
Gradient disappears, and VGG models add additional the softmax of 2 auxiliary for conducting gradient forward.VGG-Net does not use office
Portion responds normalization operation.
Common VGG models include VGG-19 and VGG-16.VGG-19 mono- shares 19 layers, has 16 convolutional layers and 3 complete
Articulamentum.VGG-16 mono- shares 16 layers, there is 13 convolutional layers and 3 full articulamentums.
After characteristic is inputted trained VGG models, convolutional layer first filters characteristic using convolution kernel
Wave;Pond layer is broadly divided into mean value pond and maximizes pond, it is preferred to use average pond (average pooling).
The most common activation primitive of convolutional neural networks be ReLU functions, due to ReLU functions input be less than 0 when gradient
It is 0, negative gradient is caused to be zeroed out in this ReLU, and this neuron is possible to never again by any data activation,
Namely ReLU neuronal necrosis.
Preferably, the activation primitive of each neuron node of VGG models is MPELU functions.
The activation primitive expression formula of MPELU is
Wherein, α, β are parameter, can be set according to canonical, i.e., first initialize a small constant, and the later stage can adjust it
Numerical value;yiIndicate the input of pond layer.
The advantages of advantage of MPELU is to be provided simultaneously with ReLU, PReLU and ELU.First, MPELU has the convergence of ELU
Property, can allow tens layers network convergence without batch normalization;Secondly, as generalized form, MPELU is compared with three
Generalization Ability it is stronger.
Based on above-described embodiment, the specific steps that the data input neural network model that training sample is concentrated is trained
Including:The data that training sample is concentrated input neural network model, according to Levenberg-Marquart method, to neural network
Model is trained.
As a preferred embodiment, when the data input neural network model that training sample is concentrated is trained, root
Neural network model is trained according to Levenberg-Marquart method.
Neural network model is trained, can be realized by Caffe even depth learning frameworks.
Levenberg-Marquart method (Levenberg-Marquardt algorithm) can provide number non-linear minimum
Change the numerical solution of (Local Minimum).Levenberg-Marquart method can reach by modification parameter when executing in conjunction with Gauss-ox
The advantages of algorithm and gradient descent method, and deficiency of the two is made to improve (for example the inverse matrix in Gauss-Newton algorithm is not
In the presence of or initial value it is too far from local minimum).According to Gauss-Newton algorithm and gradient descent method to neural network model
When being trained, when the error of setting is relatively small, it is also possible to lead to over-fitting.Levenberg-Marquart method convergence is more
Soon, mean square error is relatively low, can quickly obtain trained neural network model, and the trained neural network model obtained
Prediction result it is more acurrate.
Using Levenberg-Marquart method (Levenberg-Marquardt algorithm) algorithm to neural network
Model is trained.
For M sample, using m sample come iteration, 1<m<M.The principle of Levenberg-Marquardt methods is as follows
Wherein, (xi,yi) be sample data independent variable and dependent variable, f (x, β) be neural network model curve table
Up to formula, β is the parameter of curve f (x, β), and S (β) is the deviation of curve f (x, β).
Based on above-described embodiment, according to the meteorological data in the first statistics phase, the second statistics phase belonging to the first statistics phase
The temporal information of interior meteorological data and the first statistics phase, the characteristic for obtaining crops further include:According to the first statistics phase
The meteorological data in the second statistics phase, the temporal information of the first statistics phase belonging to interior meteorological data, the first statistics phase and the
The soil data of one statistics phase, obtains the characteristic of crops.
It,, can also basis except according in addition to meteorological data when obtaining the characteristic of crops as an alternative embodiment
The soil data of first statistics phase.
It is understood that above-mentioned soil data is the soil data for being easy quantization.It can be according to being planted in estimation range
The sampled result of multiple sampled points of the soil of crops obtains soil data.
Based on above-described embodiment, the soil data of the first statistics phase includes at least:The soil acidity or alkalinity of first statistics phase.
As a preferred embodiment, the soil data of the first statistics phase includes at least soil acidity or alkalinity.
Soil acidity or alkalinity can be pH value.
Illustrate the prediction technique of crop yield provided by the invention below by three examples.
Example one
Crops are wheat.The characteristic of last month includes:The highest temperature of last month, the lowest temperature of last month, last month
Sunshine hour number, the precipitation of last month, last month month information, the temperature on average in last month in affiliated season.
The characteristic of last month is inputted into trained VGG models, the per unit area yield in prediction wheat this year is high yield or low yield.
Example two
Crops are peanut.The characteristic of last month includes:The highest temperature of last month, the lowest temperature of last month, last month
Sunshine hour number, the precipitation of last month, last month month information, the soil acidity or alkalinity of last month, the average air in last month in affiliated season
Temperature.
The characteristic of last month is inputted into trained VGG models, the per unit area yield in prediction peanut this year is to be produced in high yield or low
Production.
Example three
Crops are rice.The characteristic of Fen Tiller phase first months includes:The highest temperature of this month, the minimum gas of this month
Temperature, the sunshine hour number of this month, the precipitation of this month, the month information (this month Wei Fen Tiller phase first month) of this month, this month
Soil acidity or alkalinity, the soil moisture content of this month, the temperature on average of the paddy field water level, Fen Tiller phases of this month.
The characteristic of this month is inputted into trained VGG models, the per unit area yield in prediction rice current growth period be high yield or
Low yield.
It is understood that crop yield prediction technique provided by the invention, it not only can be to above-mentioned three kinds of crops
Yield is predicted, the crops such as corn and soybean, cotton, rape, type of the embodiment of the present invention to crops are could be applicable to
Without limiting.
It is understood that crop yield prediction technique provided by the invention, it can also be to the difference of identical crops
The yield of kind is predicted.For example, the peanut improved seeds of China's plantation can carry out production forecast respectively up to 30 kinds.
Fig. 2 is the functional block diagram of crop yield forecasting system embodiment of the present invention.Based on above-described embodiment, such as Fig. 2 institutes
Show, a kind of crop yield forecasting system includes:Characteristic extracting module 201, for according in the first statistics phase meteorological data,
The temporal information of the meteorological data and the first statistics phase in the second statistics phase belonging to first statistics phase, obtains the feature of crops
Data;It is pre- to obtain crop yield for characteristic to be inputted trained neural network model for production forecast module 202
Survey result.
Crop yield forecasting system provided by the invention is used to execute crop yield prediction technique provided by the invention,
Each module that crop yield forecasting system includes realizes that the specific method of corresponding function and flow refer to above-mentioned crop yield
The embodiment of prediction technique, details are not described herein again.
The embodiment of the present invention can be quickly and easily by Neural Network model predictive crop yield based on meteorological data
Obtain crop yield prediction result with high accuracy.Further, by selecting the suitable first statistics phase, farming can be predicted
The yield of object current year or the yield in current growth period, when predicting the yield in current growth period, it is contemplated that the difference that growth period includes
Crops are influenced by meteorology variation in growthdevelopmental stage, and prediction result is more acurrate.
Fig. 3 is the structure diagram of the prediction apparatus embodiments of crop yield of the present invention.Based on above-described embodiment, such as Fig. 3
Shown, the pre- measurement equipment of crop yield includes:Processor (processor) 301, memory (memory) 302 and bus
303;Wherein, processor 301 and memory 302 complete mutual communication by bus 303;Processor 301 is deposited for calling
Program instruction in reservoir 302, to execute the method that above-mentioned each method embodiment is provided, such as including:Crop yield is pre-
Survey method;The training method of neural network model;The method for obtaining the characteristic of crops;According to history meteorological data and phase
The method for the crops historical yield structure training sample set answered.
Another embodiment of the present invention discloses a kind of computer program product, and computer program product is non-transient including being stored in
Computer program on computer readable storage medium, computer program include program instruction, when program instruction is held by computer
When row, computer is able to carry out the method that above-mentioned each method embodiment is provided, such as including:Crop yield prediction technique;
The training method of neural network model;The method for obtaining the characteristic of crops;According to history meteorological data and corresponding agriculture
The method that crop historical yield builds training sample set.
Another embodiment of the present invention provides a kind of non-transient computer readable storage medium, non-transient computer readable storage
Medium storing computer instructs, and computer instruction makes computer execute the method that above-mentioned each method embodiment is provided, such as wraps
It includes:Crop yield prediction technique;The training method of neural network model;The method for obtaining the characteristic of crops;According to
The method of history meteorological data and corresponding crops historical yield structure training sample set.
System embodiment described above is only schematical, wherein can be as the unit that separating component illustrates
Or may not be and be physically separated, the component shown as unit may or may not be physical unit, i.e.,
A place can be located at, or may be distributed over multiple network units.It can select according to the actual needs therein
Some or all of module achieves the purpose of the solution of this embodiment.Those of ordinary skill in the art are not paying creative labor
In the case of dynamic, you can to understand and implement.
Through the above description of the embodiments, those skilled in the art can be understood that each embodiment can
It is realized by the mode of software plus required 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 expressed in the form of software products in other words, should
Computer software product can store in a computer-readable storage medium, such as ROM/RAM, magnetic disc, CD, including several fingers
It enables and using so that computer equipment (can be personal computer, server or the network equipment an etc.) execution is above-mentioned each
The method of certain parts of embodiment 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, it will be understood by those of ordinary skill in the art that:It still may be used
With technical scheme described in the above embodiments is modified or equivalent replacement of some of the technical features;
And these modifications or replacements, various embodiments of the present invention technical solution that it does not separate the essence of the corresponding technical solution spirit and
Range.
Claims (10)
1. a kind of crop yield prediction technique, which is characterized in that including:
According to the meteorological data in the first statistics phase, the meteorological data in the second statistics phase belonging to the first statistics phase and the first system
The temporal information of meter phase obtains the characteristic of crops;
The characteristic is inputted into trained neural network model, obtains crop yield prediction result.
2. crop yield prediction technique according to claim 1, which is characterized in that obtain the trained nerve net
The specific steps of network model include:
Training sample set is built according to history meteorological data and corresponding crops historical yield;
The data input neural network model that the training sample is concentrated is trained, the trained neural network is obtained
Model.
3. crop yield prediction technique according to claim 1 or 2, which is characterized in that in the first statistics phase
Meteorological data includes at least:The highest temperature, the lowest temperature, sunshine-duration and the precipitation of first statistics phase;
The meteorological data in the second statistics phase belonging to the first statistics phase, according to belong to it is described second statistics the phase each first
The temperature on average of statistics phase obtains.
4. crop yield prediction technique according to claim 2, which is characterized in that the neural network model is VGG
Model;
The activation primitive of each neuron node of the VGG models is MPELU functions.
5. crop yield prediction technique according to claim 2, which is characterized in that described to concentrate the training sample
The specific steps that are trained of data input neural network model include:
The data that the training sample is concentrated input neural network model, according to Levenberg-Marquart method, to nerve
Network model is trained.
6. crop yield prediction technique according to claim 1, which is characterized in that described according in the first statistics phase
The temporal information of the meteorological data and the first statistics phase in the second statistics phase belonging to meteorological data, the first statistics phase, obtains agriculture
The characteristic of crop further includes:
According to the meteorological data in the first statistics phase, the meteorological data in the second statistics phase belonging to the first statistics phase, the first system
The soil data of the temporal information of meter phase and the first statistics phase, obtains the characteristic of crops.
7. crop yield prediction technique according to claim 6, which is characterized in that the soil number of the first statistics phase
According to including at least:The soil acidity or alkalinity of first statistics phase.
8. a kind of crop yield forecasting system, which is characterized in that including:
Characteristic extracting module, for according in the meteorological data in the first statistics phase, the second statistics phase belonging to the first statistics phase
Meteorological data and first statistics the phase temporal information, obtain the characteristic of crops;
It is pre- to obtain crop yield for the characteristic to be inputted trained neural network model for production forecast module
Survey result.
9. a kind of pre- measurement equipment of crop yield, which is characterized in that including:
At least one processor;And
At least one processor being connect with the processor communication, wherein:
The memory is stored with the program instruction that can be executed by the processor, and the processor calls described program to instruct energy
Enough methods executed as described in claim 1 to 7 is any.
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 |
---|---|---|---|
CN201810460935.6A CN108665107A (en) | 2018-05-15 | 2018-05-15 | Crop yield prediction technique and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810460935.6A CN108665107A (en) | 2018-05-15 | 2018-05-15 | Crop yield prediction technique and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108665107A true CN108665107A (en) | 2018-10-16 |
Family
ID=63778369
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810460935.6A Pending CN108665107A (en) | 2018-05-15 | 2018-05-15 | Crop yield prediction technique and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108665107A (en) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109615150A (en) * | 2018-12-29 | 2019-04-12 | 航天信息股份有限公司 | A kind of method and system of determining rice Meteorological Output |
CN109767038A (en) * | 2019-01-04 | 2019-05-17 | 平安科技(深圳)有限公司 | Crop yield prediction technique, device and computer readable storage medium |
CN109784588A (en) * | 2019-03-16 | 2019-05-21 | 石家庄学院 | It is a kind of to be allied the communists the Output of Sorghum prediction technique of monitoring based on sensor |
CN109934397A (en) * | 2019-03-05 | 2019-06-25 | 中国农业科学院农业信息研究所 | A kind of Forecasting Methods for Agriculture and system |
CN110443420A (en) * | 2019-08-05 | 2019-11-12 | 山东农业大学 | A kind of crop production forecast method based on machine learning |
CN110751094A (en) * | 2019-10-21 | 2020-02-04 | 北京师范大学 | Crop yield estimation technology based on GEE comprehensive remote sensing image and deep learning method |
CN110889547A (en) * | 2019-11-20 | 2020-03-17 | 中国农业大学 | Crop growth period prediction method and device |
CN112183575A (en) * | 2020-08-24 | 2021-01-05 | 盐城工学院 | Method for predicting potato yield based on regional meteorological parameters |
CN113435649A (en) * | 2021-06-29 | 2021-09-24 | 布瑞克农业大数据科技集团有限公司 | Global agricultural data sorting method, system, device and medium |
CN113743762A (en) * | 2021-08-26 | 2021-12-03 | 中国科学院地理科学与资源研究所 | Grassland grass yield inversion method and system based on neural network model |
CN114418235A (en) * | 2022-01-28 | 2022-04-29 | 西安理工大学 | Crop growth condition determining method and system |
CN116595333A (en) * | 2023-05-18 | 2023-08-15 | 中国农业大学 | Soil-climate intelligent rice target yield and nitrogen fertilizer consumption determination method |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107341577A (en) * | 2017-07-25 | 2017-11-10 | 中国农业科学院农业信息研究所 | A kind of crop yield Forecasting Methodology and system |
CN107392376A (en) * | 2017-07-25 | 2017-11-24 | 中国农业科学院农业信息研究所 | A kind of crops Meteorological Output Forecasting Methodology and system |
CN107832892A (en) * | 2017-11-17 | 2018-03-23 | 北京大学 | A kind of lumpily seed selection decision optimization method based on Combinatorial Optimization |
CN108022023A (en) * | 2017-12-20 | 2018-05-11 | 深圳春沐源控股有限公司 | Plantation production prediction method, device and the storage medium of a kind of crops |
-
2018
- 2018-05-15 CN CN201810460935.6A patent/CN108665107A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107341577A (en) * | 2017-07-25 | 2017-11-10 | 中国农业科学院农业信息研究所 | A kind of crop yield Forecasting Methodology and system |
CN107392376A (en) * | 2017-07-25 | 2017-11-24 | 中国农业科学院农业信息研究所 | A kind of crops Meteorological Output Forecasting Methodology and system |
CN107832892A (en) * | 2017-11-17 | 2018-03-23 | 北京大学 | A kind of lumpily seed selection decision optimization method based on Combinatorial Optimization |
CN108022023A (en) * | 2017-12-20 | 2018-05-11 | 深圳春沐源控股有限公司 | Plantation production prediction method, device and the storage medium of a kind of crops |
Non-Patent Citations (1)
Title |
---|
任斌等: ""基于结构化神经网络挖掘的农产品产量预测方法"", 《计算机工程与科学》 * |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109615150B (en) * | 2018-12-29 | 2023-04-18 | 航天信息股份有限公司 | Method and system for determining rice meteorological output |
CN109615150A (en) * | 2018-12-29 | 2019-04-12 | 航天信息股份有限公司 | A kind of method and system of determining rice Meteorological Output |
CN109767038A (en) * | 2019-01-04 | 2019-05-17 | 平安科技(深圳)有限公司 | Crop yield prediction technique, device and computer readable storage medium |
CN109934397A (en) * | 2019-03-05 | 2019-06-25 | 中国农业科学院农业信息研究所 | A kind of Forecasting Methods for Agriculture and system |
CN109784588A (en) * | 2019-03-16 | 2019-05-21 | 石家庄学院 | It is a kind of to be allied the communists the Output of Sorghum prediction technique of monitoring based on sensor |
CN110443420A (en) * | 2019-08-05 | 2019-11-12 | 山东农业大学 | A kind of crop production forecast method based on machine learning |
CN110751094A (en) * | 2019-10-21 | 2020-02-04 | 北京师范大学 | Crop yield estimation technology based on GEE comprehensive remote sensing image and deep learning method |
CN110889547A (en) * | 2019-11-20 | 2020-03-17 | 中国农业大学 | Crop growth period prediction method and device |
CN110889547B (en) * | 2019-11-20 | 2022-04-15 | 中国农业大学 | Crop growth period prediction method and device |
CN112183575A (en) * | 2020-08-24 | 2021-01-05 | 盐城工学院 | Method for predicting potato yield based on regional meteorological parameters |
CN113435649A (en) * | 2021-06-29 | 2021-09-24 | 布瑞克农业大数据科技集团有限公司 | Global agricultural data sorting method, system, device and medium |
CN113743762A (en) * | 2021-08-26 | 2021-12-03 | 中国科学院地理科学与资源研究所 | Grassland grass yield inversion method and system based on neural network model |
CN114418235A (en) * | 2022-01-28 | 2022-04-29 | 西安理工大学 | Crop growth condition determining method and system |
CN116595333A (en) * | 2023-05-18 | 2023-08-15 | 中国农业大学 | Soil-climate intelligent rice target yield and nitrogen fertilizer consumption determination method |
CN116595333B (en) * | 2023-05-18 | 2024-04-09 | 中国农业大学 | Soil-climate intelligent rice target yield and nitrogen fertilizer consumption determination method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108665107A (en) | Crop yield prediction technique and system | |
US20170270446A1 (en) | Agronomic systems, methods and apparatuses for determining yield limits | |
Hundal | Application of the CERES–Wheat model to yield predictions in the irrigated plains of the Indian Punjab | |
CN111898922B (en) | Multi-scale crop yield assessment method and system | |
Tan et al. | Dynamic process-based modelling of crop growth and competitive water extraction in relay strip intercropping: Model development and application to wheat-maize intercropping | |
CN107392376A (en) | A kind of crops Meteorological Output Forecasting Methodology and system | |
CN108617355A (en) | A kind of cluster greenhouse irrigation decision-making technique and system | |
Belouz et al. | Prediction of greenhouse tomato yield using artificial neural networks combined with sensitivity analysis | |
CN113016450A (en) | Greenhouse crop irrigation method and system | |
Aiken | Applying thermal time scales to sunflower development | |
AU2023203307A1 (en) | Controlling Agricultural Production Areas | |
CN113822479A (en) | Multi-objective optimization method for regional agricultural planting structure considering production, environment and economic benefits | |
CN116595333B (en) | Soil-climate intelligent rice target yield and nitrogen fertilizer consumption determination method | |
CN114128608A (en) | Orchard irrigation management method and system | |
Ribas et al. | An update of new flood-irrigated rice cultivars in the SimulArroz model | |
Liao et al. | Novel models for simulating maize growth based on thermal time and photothermal units: Applications under various mulching practices | |
Huang et al. | Simulating the optimal growing season of rice in the Yangtze River Valley and its adjacent area, China | |
Medd et al. | Prediction of the potential distribution of Carduus nutans (nodding thistle) in Australia | |
ABD EL BAKI et al. | Determination of irrigation depths using a numerical model of crop growth and quantitative weather forecast and evaluation of its effect through a field experiment for potato | |
Mendigoria et al. | Optimization of the Hargreaves, HamonV1, and Penman Potential Evapotranspiration Models Using Bio-inspired Algorithms. | |
Inman-Bamber | Some physiological factors affecting the optimum age and season for harvesting sugarcane | |
Wu et al. | Crop yield estimation in the North China plain from 2001 to 2016 using multi-source remote sensing data and process-based Fgm model | |
Cheviron et al. | The Optirrig model for the generation, analysis and optimization of irrigation scenarios: rationale and scopes | |
Cohen-Manrique et al. | Diffuse irrigation systems for the production of watermelon in the Sabanas subregion, Sucre-Colombia | |
Hunink et al. | Past and future trends in crop production and food demand and supply in the Lower Mekong Basin |
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: 20181016 |
|
RJ01 | Rejection of invention patent application after publication |