CN110428107A - A kind of corn yield remote sensing prediction method and system - Google Patents
A kind of corn yield remote sensing prediction method and system Download PDFInfo
- Publication number
- CN110428107A CN110428107A CN201910719979.0A CN201910719979A CN110428107A CN 110428107 A CN110428107 A CN 110428107A CN 201910719979 A CN201910719979 A CN 201910719979A CN 110428107 A CN110428107 A CN 110428107A
- Authority
- CN
- China
- Prior art keywords
- remote sensing
- vegetation index
- sensing image
- corn
- property parameters
- 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
- 240000008042 Zea mays Species 0.000 title claims abstract description 70
- 235000002017 Zea mays subsp mays Nutrition 0.000 title claims abstract description 70
- 235000005824 Zea mays ssp. parviglumis Nutrition 0.000 title claims abstract description 68
- 235000005822 corn Nutrition 0.000 title claims abstract description 68
- 238000000034 method Methods 0.000 title claims abstract description 28
- 238000003062 neural network model Methods 0.000 claims abstract description 38
- 230000002708 enhancing effect Effects 0.000 claims abstract description 25
- 238000004519 manufacturing process Methods 0.000 claims abstract description 23
- 239000008267 milk Substances 0.000 claims abstract description 18
- 210000004080 milk Anatomy 0.000 claims abstract description 18
- 235000013336 milk Nutrition 0.000 claims abstract description 18
- 238000012549 training Methods 0.000 claims description 34
- 238000013528 artificial neural network Methods 0.000 claims description 27
- 238000012937 correction Methods 0.000 claims description 13
- 230000003595 spectral effect Effects 0.000 claims description 13
- 230000005855 radiation Effects 0.000 claims description 12
- 238000002310 reflectometry Methods 0.000 claims description 12
- 238000007781 pre-processing Methods 0.000 claims description 10
- 210000005036 nerve Anatomy 0.000 claims description 3
- 230000007774 longterm Effects 0.000 claims description 2
- 240000007594 Oryza sativa Species 0.000 claims 1
- 235000007164 Oryza sativa Nutrition 0.000 claims 1
- 230000035800 maturation Effects 0.000 claims 1
- 235000009566 rice Nutrition 0.000 claims 1
- 238000004422 calculation algorithm Methods 0.000 description 8
- 230000006870 function Effects 0.000 description 7
- 230000002068 genetic effect Effects 0.000 description 7
- 238000005259 measurement Methods 0.000 description 4
- 210000004218 nerve net Anatomy 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 241000208340 Araliaceae Species 0.000 description 2
- 241000196324 Embryophyta Species 0.000 description 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 2
- 235000003140 Panax quinquefolius Nutrition 0.000 description 2
- 235000016383 Zea mays subsp huehuetenangensis Nutrition 0.000 description 2
- 235000008434 ginseng Nutrition 0.000 description 2
- 235000009973 maize Nutrition 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000003306 harvesting Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 239000000843 powder Substances 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000001507 sample dispersion Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/086—Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/80—Geometric correction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N2021/1793—Remote sensing
- G01N2021/1797—Remote sensing in landscape, e.g. crops
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
- G06T2207/30188—Vegetation; Agriculture
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Business, Economics & Management (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Human Resources & Organizations (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Strategic Management (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- Economics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Marketing (AREA)
- Quality & Reliability (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Physiology (AREA)
- Agronomy & Crop Science (AREA)
- Primary Health Care (AREA)
- Operations Research (AREA)
- Genetics & Genomics (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Animal Husbandry (AREA)
- Marine Sciences & Fisheries (AREA)
- Mining & Mineral Resources (AREA)
Abstract
The invention discloses a kind of corn yield remote sensing prediction method and system.This method comprises: obtaining the remote sensing image of corn setting growth period in region to be measured, the setting growth period includes milk stage and maturity period；Determine that the property parameters of each remote sensing image, the property parameters include difference vegetation index, ratio vegetation index, enhancing vegetation index and green degree vegetation index；Property parameters are inputted to the yield of corn in GA-BP Neural Network model predictive region to be measured, GA-BP neural network model is that the history remote sensing image that growth period is set according to corn in region to be measured and corresponding historical production data train obtained model.Corn yield remote sensing prediction method and system provided by the invention have the characteristics that precision of prediction is high.
Description
Technical field
The present invention relates to crop yield electric powder prediction, more particularly to a kind of corn yield remote sensing prediction method and
System.
Background technique
Currently with remotely-sensed data carry out corn yield estimation be commonly based on regression model, that is, establish yield with it is a certain
Linear relationship between variable.But regression model needs a large amount of local actual measurement yield datas and supports, and calculating process is opposite
Time-consuming, moreover, describing linear relationship merely, structure is too simple, and precision of prediction is not high.
Summary of the invention
The object of the present invention is to provide a kind of corn yield remote sensing prediction method and system, have easy and precision of prediction high
The characteristics of.
To achieve the above object, the present invention provides following schemes:
A kind of corn yield remote sensing prediction method, comprising:
Obtain the remote sensing image of the setting growth period of corn in region to be measured, the setting growth period include milk stage with
Maturity period；
Determine the property parameters of each remote sensing image, the property parameters include that difference vegetation index, ratio vegetation refer to
Number, enhancing vegetation index and green degree vegetation index；
It is described by the yield of corn in region to be measured described in property parameters input GA-BP Neural Network model predictive
GA-BP neural network model is that the history remote sensing image of growth period and corresponding is set according to corn in the region to be measured
The model that historical production data training obtains.
Optionally, it obtains described in region to be measured after the remote sensing image of corn setting growth period, in the determination
Before the property parameters of each remote sensing image, further includes:
The remote sensing image is pre-processed, the pretreatment includes: radiation calibration, atmospheric correction and ortho-rectification；
According to the reflectivity for each pixel of the remote sensing image that pretreatment obtains, wave band calculating is carried out to each pixel, it is described
Wave band includes near infrared band, red spectral band, blue wave band and green light band, and the wave band is used for difference vegetation index, ratio
The calculating of vegetation index, enhancing vegetation index and green degree vegetation index.
Optionally, the training of the GA-BP neural network model includes:
Obtain the history remote sensing image and corresponding historical yield number of corn setting growth period in the region to be measured
According to；
Determine that the property parameters of each history remote sensing image, the property parameters include difference vegetation index, ratio plant
By index, enhancing vegetation index and green degree vegetation index；
It is input with the property parameters of the history remote sensing image, is output, training GA-BP with the historical production data
Neural network obtains GA-BP neural network model.
Optionally, in the history remote sensing image and correspondence for obtaining corn setting growth period in the region to be measured
Historical production data after, before the property parameters of each history remote sensing image of the determination, further includes:
The history remote sensing image is pre-processed, the pretreatment includes: radiation calibration, atmospheric correction and just penetrates school
Just；
According to the reflectivity for each pixel of history remote sensing image that pretreatment obtains, wave band calculating is carried out to each pixel,
The wave band includes near infrared band, red spectral band, blue wave band and green light band, the wave band for difference vegetation index,
The calculating of ratio vegetation index, enhancing vegetation index and green degree vegetation index.
Optionally, after the property parameters of each history remote sensing image of the determination, in the trained GA-BP nerve
Before network, further includes:
Determine sampled point；
Each property parameters that the sampled point respectively sets growth period are obtained, for as the training GA-BP nerve net
Input when network model.
The present invention also provides a kind of corn yield remote sensing prediction systems, comprising:
Remote sensing image obtains module, described to set for obtaining the remote sensing image of corn setting growth period in region to be measured
Determining growth period includes milk stage and maturity period；
Property parameters determining module, for determining that the property parameters of each remote sensing image, the property parameters include poor
It is worth vegetation index, ratio vegetation index, enhancing vegetation index and green degree vegetation index；
Production forecast module, for the property parameters to be inputted region to be measured described in GA-BP Neural Network model predictive
The yield of interior corn, the GA-BP neural network model are the history that growth period is set according to corn in the region to be measured
The model that remote sensing image and the training of corresponding historical production data obtain.
Optionally, the system also includes:
Remote sensing image preprocessing module, for pre-processing to the remote sensing image, the pretreatment includes: radiation mark
Fixed, atmospheric correction and ortho-rectification；
Remote sensing image wave band computing module, the reflectivity of each pixel of the remote sensing image for being obtained according to pretreatment,
Wave band calculating is carried out to each pixel, the wave band includes near infrared band, red spectral band, blue wave band and green light band, described
Calculating of the wave band for difference vegetation index, ratio vegetation index, enhancing vegetation index and green degree vegetation index.
Optionally, the system also includes:
Historical data obtains module, for obtaining the history remote sensing image of corn setting growth period in the region to be measured
And corresponding historical production data；
History remote sensing image property parameters determining module, for determining the property parameters of each history remote sensing image, institute
Stating property parameters includes difference vegetation index, ratio vegetation index, enhancing vegetation index and green degree vegetation index；
Neural network model training module is gone through for being input with the property parameters of the history remote sensing image with described
History yield data is output, and training GA-BP neural network obtains GA-BP neural network model.
Optionally, the system also includes:
History remote sensing image preprocessing module, for being pre-processed to the history remote sensing image, the pretreatment packet
It includes: radiation calibration, atmospheric correction and ortho-rectification；
History remote sensing image wave band computing module, each pixel of history remote sensing image for being obtained according to pretreatment
Reflectivity carries out wave band calculating to each pixel, and the wave band includes near infrared band, red spectral band, blue wave band and green light wave
Section, calculating of the wave band for difference vegetation index, ratio vegetation index, enhancing vegetation index and green degree vegetation index.
Optionally, the system also includes:
Sampled point determining module, for determining sampled point；
Sampled point property parameters obtain module, and each property parameters of growth period are respectively set for obtaining the sampled point,
Input when for the GA-BP neural network model described as training.
The specific embodiment provided according to the present invention, the invention discloses following technical effects: corn provided by the invention
Yield remote sensing prediction method and system obtain the remote sensing image of corn setting growth period in region to be measured first, and the setting is raw
Long-term includes milk stage and maturity period；Then, it is determined that the property parameters of each remote sensing image, which includes that difference is planted
By index, ratio vegetation index, enhancing vegetation index and green degree vegetation index；Finally, property parameters are inputted GA-BP nerve net
The yield of corn in network model prediction region to be measured, wherein GA-BP neural network model is to set growth period according to corn
The model that history remote sensing image and the training of corresponding historical production data obtain.As it can be seen that the present invention according to corn milk stage and
Difference vegetation index, ratio vegetation index, enhancing vegetation index and the green degree vegetation index in maturity period, using GA-BP nerve net
Network model predicts the yield of corn, wherein falls into conjunction with what genetic algorithm can be avoided that BP neural network is easy to appear
The problem of locally optimal solution, accelerates the speed that yield estimation is carried out using neural network, meanwhile, GA-BP neural network model
Use, keep the prediction of yield more easy, and improve the accuracy of prediction.
Detailed description of the invention
It in order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, below will be to institute in embodiment
Attached drawing to be used is needed to be briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations of the invention
Example, for those of ordinary skill in the art, without creative efforts, can also obtain according to these attached drawings
Obtain other attached drawings.
Fig. 1 is corn yield of embodiment of the present invention remote sensing prediction method flow diagram；
Fig. 2 is the training flow chart of GA-BP of embodiment of the present invention neural network model；
Fig. 3 is GA-BP of the embodiment of the present invention and BP neural network comparative result figure；
Fig. 4 is corn yield of embodiment of the present invention remote sensing prediction system construction drawing.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
The object of the present invention is to provide a kind of corn yield remote sensing prediction method and system, have easy and precision of prediction high
The characteristics of.
In order to make the foregoing objectives, features and advantages of the present invention clearer and more comprehensible, with reference to the accompanying drawing and specific real
Applying mode, the present invention is described in further detail.
The first aspect of the present invention provides a kind of corn yield remote sensing prediction method, as shown in Figure 1, this method include with
Lower step:
Step 101: obtaining the remote sensing image of corn setting growth period in region to be measured, setting growth period may include
Milk stage and maturity period；Wherein it is possible to using GF1 WFV satellite image data as data source, obtain respectively milk stage (August), at
The remote sensing image of ripe phase (September), the data spatial resolution are 16m, including 4 wave bands, respectively blue wave band (0.45-
0.52um), green wave band (0.52-0.59um), red wave band (0.63-0.69um), nearly red wave band (0.77-0.89um)；
Step 102: determining that the property parameters of each remote sensing image, property parameters may include difference vegetation index, ratio plant
By index, enhancing vegetation index and green degree vegetation index；
Step 103: property parameters being inputted to the yield of corn in GA-BP Neural Network model predictive region to be measured, and raw
At the corn yield figure in region to be measured, wherein GA (genetic algorithm)-BP neural network model is to set growth period according to corn
History remote sensing image and the obtained neural network model of corresponding historical production data training；
Using the Decision Classfication tool of ENVI software, the corn yield of the region to be measured of prediction everywhere is classified.According to production
Magnitude sorts from low to high, in conjunction with local normal output numerical value, is divided into the underproduction 8 into above (almost having no harvest), underproduction 7-8
Amount to 8 grades at following at, the underproduction 3 at, underproduction 3-4 at, underproduction 4-5 at, underproduction 5-6 at, underproduction 6-7.
Due to traditional BP neural network, initial weight and threshold value are the random numbers between several groups 0 to 1, and network reaches
More the number of iterations is needed to convergence, and genetic algorithm can obtain optimum solution by global resolving, thus using heredity
Algorithm optimizes the initial weight of BP network and threshold value, improves the arithmetic speed of BP neural network and reduces trained network
Required time.
In one embodiment, between step 101 and step 102, can also include:
Remote sensing image is pre-processed, pretreatment includes: radiation calibration, atmospheric correction and ortho-rectification, wherein utilize
ENVI software carries out radiation calibration, atmospheric correction, the pretreatment of ortho-rectification.By the initial data downloaded to by original pixel
DN value is processed into reflectivity；
According to the reflectivity for each pixel of remote sensing image that pretreatment obtains, wave band calculating is carried out to each pixel, wave band includes
Near infrared band, red spectral band, blue wave band and green light band, wave band for carry out difference vegetation index, ratio vegetation index,
Enhance vegetation index and green degree vegetation index calculates.Wherein, using the wave band calculating instrument of ENVI software to the reflectivity of acquisition
Data carry out wave band calculating.Obtain normalized site attenuation (NDVI), ratio vegetation index (RVI) enhances vegetation index
(EVI) green degree vegetation index (G) image data.Formula is as follows:
NDVI=(ρNIR-ρR)/(ρNIR+ρR)
RVI=ρNIR/ρR
EVI=2.5 × (ρNIR-ρR)/(ρNIR+6.0×ρR-7.5×ρB+1)
G=ρNIR/ρG
Wherein, ρNIRFor near infrared band, ρRFor red spectral band, ρBFor blue wave band, ρGFor green light band.
The training process of GA-BP neural network model in the above-described embodiments is as follows:
Obtain the history remote sensing image and corresponding historical production data of corn setting growth period in region to be measured；
Determine that the property parameters of each history remote sensing image, property parameters include difference vegetation index, ratio vegetation index, increasing
Strong vegetation index and green degree vegetation index；Using ArcMap software, the latitude and longitude information according to yield monitoring point establishes dotted want
Element, and " multivalue is extracted a little " tool is utilized, by NDVI, RVI, EVI, G, value of totally 4 kinds of images at sampled point is extracted,
Generate the storage of Excel table.Table totally 9 column, first row is yield (kg/ha), it is rear 8 column be milk stage NDVI, milk stage RVI,
Milk stage EVI, milk stage G, maturity period NDVI, maturity period RVI, maturity period EVI, maturity period G.By the number in Excel table
Value is used as yield data parameter set；
It is input with the property parameters of history remote sensing image, is output, training GA-BP nerve net with historical production data
Network obtains GA-BP neural network model.Yield data parameter set is read by MATLAB software code realization, realization will be newborn
Ripe phase NDVI, milk stage RVI, milk stage EVI, milk stage G, maturity period NDVI, maturity period RVI, maturity period EVI, maturity period G number
It is worth the input as GA-BP neural network, using yield (units/kg/ha) numerical value as the desired output of GA-BP neural network.It produces
Amount parameter data set is divided into training set and verifying collection, and training set accounts for the 2/3 of sum, and verifying collection accounts for the 1/3 of sum.
Trained flow chart is as shown in Fig. 2, the initialization network of left part is that neural network reads yield ginseng in figure
The process of number data set.Genetic algorithm first encodes original input data, using BP neural network training error as
The expression formula of fitness function, fitness function F is shown below.
In formula, yjIt is the desired output of j-th of node of network, ojIt is the prediction output of j-th of node of network, n is node
Number, k are adjustment factors.The individual for completing aforesaid operations is subjected to fitness value judgement, meets the individual of condition as BP mind
Optimization weight and optimization threshold value through network.The error amount calculation formula of BP neural network is as follows:
From the above equation, we can see that yjIt is the output valve of node j, ojIt is the desired output of node j.The one of the two squared difference sum
Half is the error E (ω, b) of whole network.When network reaches the cycle-index of setting or reaches precision, deconditioning is simultaneously generated
Export result.
Wherein, in one embodiment, in acquired region to be measured corn setting growth period history remote sensing image
Later, further includes:
History remote sensing image is pre-processed, pretreatment includes: radiation calibration, atmospheric correction and ortho-rectification；
According to the reflectivity for each pixel of history remote sensing image that pretreatment obtains, wave band calculating, wave band are carried out to each pixel
Including near infrared band, red spectral band, blue wave band and green light band, wave band is for carrying out difference vegetation index, ratio vegetation
Index, enhancing vegetation index and green degree vegetation index calculate.
In one embodiment, before training GA-BP neural network, can also include:
Determine sampled point；
Each property parameters that sampled point respectively sets growth period are obtained, when for as training GA-BP neural network model
Input.
Corn yield remote sensing prediction method provided by the present application is verified below:
The present invention is constructed in the verification process with BP neural network (BP) and with genetic algorithm optimization BP neural network
(GA-BP) two kinds of Yield Estimation Models.It is respectively 8,9,1 that input layer, hidden layer, output layer number of nodes, which is arranged, in BP neural network
It is a.Select tansig function as hidden layer training function, purelin function is as output layer training function, training precision
0.0001, e-learning rate is set as 0.1, at the same be arranged trained maximum cycle be 2000, using sim function to result into
Row emulation.Four kinds of vegetation indexs are as input, and actual measurement yield is as desired output, BP neural network result figure and GA-BP nerve
Web results figure is as shown in Figure 3.
From the figure 3, it may be seen that GA-BP model is for opposite low yield (800kg/ha or so) and with respect to high yield (1200kg/ha)
It is all better than BP modelling effect to simulate effect.Compare check post relative error between BP model and GA-BP model, the maximum of BP model
Relative error is -59.16%, and minimum relative error is 12.20%, average relative error 29.23%.GA-BP model is most
Big relative error is 11.59%, and minimum relative error is -0.86%, average relative error 5.27%.BP model is for low yield
The analog result of amount and the difference of actual value are larger, and GA-BP model is smaller for high yield and low-producing analog case error.
Precision test
Related coefficient (R2) and root-mean-square error (RMSE) be the parameter for judgment models result, R2Calculation formula such as
Under:
In formula, OjIndicate forecast production,Indicate the mean value of forecast production；yjIndicate actual measurement yield, yjIndicate actual measurement yield
Mean value, n indicate number of samples.R2Range be [0,1].R2Closer to 1, indicate that the degree of fitting is higher.
Root-mean-square error RMSE is that error sum of squares is averaged, then extracts square root, and is the numerical value for indicating sample dispersion degree,
Calculation formula is as follows:
The R of BP neural network model and GA-BP neural network model2It is worth as shown in table 1 with RMSE value:
The R of table 1 BP model and GA-BP model2With RMSE
As shown in Table 1, the Maize yeild estimation model R constructed using BP neural network2Reach 0.8452, there is preferable yield
Computational estimation competence, RMSE (%) are 28.37, show to have differences between the predicted value of model and true value.It is excellent using genetic algorithm
Change the Maize yeild estimation model R of BP neural network building2Reaching 0.9850, RMSE (%) is 6.70, shows the predictive ability of model
It is relatively strong, and the difference very little between predicted value and true value, play the role of good yield estimation, and for larger yield
The simulation of smaller yield all has stronger learning ability.
The second aspect of the present invention provides a kind of corn yield remote sensing prediction system, as shown in figure 4, the system includes:
Remote sensing image obtains module 401, for obtaining the remote sensing image of corn setting growth period in region to be measured, setting
Growth period includes milk stage and maturity period；
Property parameters determining module 402, for determining that the property parameters of each remote sensing image, property parameters include difference vegetation
Index, ratio vegetation index, enhancing vegetation index and green degree vegetation index；
Production forecast module 403, for property parameters to be inputted corn in GA-BP Neural Network model predictive region to be measured
Yield, GA-BP neural network model be according to corn set growth period history remote sensing image and corresponding history produce
The neural network model that amount data training obtains.
In one embodiment, corn yield remote sensing prediction system provided by the invention can also include:
Remote sensing image preprocessing module, for pre-processing to remote sensing image, pretreatment includes: radiation calibration, atmosphere
Correction and ortho-rectification；
Remote sensing image wave band computing module, the reflectivity of each pixel of remote sensing image for being obtained according to pretreatment, to each
Pixel carries out wave band calculating, and wave band includes near infrared band, red spectral band, blue wave band and green light band, and wave band is for carrying out
Difference vegetation index, ratio vegetation index, enhancing vegetation index and green degree vegetation index calculate.
In the above-described embodiments, further include following for training the module of GA-BP neural network:
Historical data obtains module, for obtain the setting growth period of corn in region to be measured history remote sensing image and
Corresponding historical production data；
History remote sensing image preprocessing module, for pre-processing to history remote sensing image, pretreatment includes: radiation mark
Fixed, atmospheric correction and ortho-rectification；
History remote sensing image wave band computing module, the reflection of each pixel of history remote sensing image for being obtained according to pretreatment
Rate carries out wave band calculating to each pixel, and wave band includes near infrared band, red spectral band, blue wave band and green light band, and wave band is used
It is calculated in carrying out difference vegetation index, ratio vegetation index, enhancing vegetation index and green degree vegetation index；
History remote sensing image property parameters determining module, for determining the property parameters of each history remote sensing image, attribute ginseng
Number includes difference vegetation index, ratio vegetation index, enhancing vegetation index and green degree vegetation index；
Sampled point determining module, for determining sampled point；
Sampled point property parameters obtain module, and each property parameters of growth period are respectively set for obtaining sampled point, are used for
Input when as training GA-BP neural network model.
Neural network model training module, for being input with the property parameters of history remote sensing image, with historical yield number
According to export, training GA-BP neural network obtains GA-BP neural network model.
Corn yield remote sensing prediction method and system provided by the invention are planted according to the difference in corn milk stage and maturity period
By index, ratio vegetation index, enhancing vegetation index and green degree vegetation index, using GA-BP neural network model to corn
Yield is predicted, wherein falls into asking for locally optimal solution in conjunction with what genetic algorithm can be avoided that BP neural network is easy to appear
Topic accelerates the speed that yield estimation is carried out using neural network, meanwhile, the use of GA-BP neural network model makes yield
Prediction is more easy, improves the accuracy of prediction.
Each embodiment in this specification is described in a progressive manner, the highlights of each of the examples are with other
The difference of embodiment, the same or similar parts in each embodiment may refer to each other.For system disclosed in embodiment
For, since it is corresponded to the methods disclosed in the examples, so being described relatively simple, related place is said referring to method part
It is bright.
Used herein a specific example illustrates the principle and implementation of the invention, and above embodiments are said
It is bright to be merely used to help understand method and its core concept of the invention；At the same time, for those skilled in the art, foundation
Thought of the invention, there will be changes in the specific implementation manner and application range.In conclusion the content of the present specification is not
It is interpreted as limitation of the present invention.
Claims (10)
1. a kind of corn yield remote sensing prediction method characterized by comprising
The remote sensing image of corn setting growth period in region to be measured is obtained, the setting growth period includes milk stage and maturation
Phase；
Determine that the property parameters of each remote sensing image, the property parameters include difference vegetation index, ratio vegetation index, increasing
Strong vegetation index and green degree vegetation index；
By the yield of corn in region to be measured described in property parameters input GA-BP Neural Network model predictive, the GA-BP
Neural network model is the history remote sensing image that growth period is set according to corn in the region to be measured and corresponding history
The model that yield data training obtains.
2. corn yield remote sensing prediction method according to claim 1, which is characterized in that obtained in region to be measured described
After corn sets the remote sensing image of growth period, before the property parameters of each remote sensing image of the determination, further includes:
The remote sensing image is pre-processed, the pretreatment includes: radiation calibration, atmospheric correction and ortho-rectification；
According to the reflectivity for each pixel of the remote sensing image that pretreatment obtains, wave band calculating, the wave band are carried out to each pixel
Including near infrared band, red spectral band, blue wave band and green light band, the wave band is used for difference vegetation index, ratio vegetation
The calculating of index, enhancing vegetation index and green degree vegetation index.
3. corn yield remote sensing prediction method according to claim 1, which is characterized in that the GA-BP neural network mould
The training of type includes:
Obtain the history remote sensing image and corresponding historical production data of corn setting growth period in the region to be measured；
Determine the property parameters of each history remote sensing image, the property parameters include that difference vegetation index, ratio vegetation refer to
Number, enhancing vegetation index and green degree vegetation index；
It is input with the property parameters of the history remote sensing image, is output, training GA-BP nerve with the historical production data
Network obtains GA-BP neural network model.
4. corn yield remote sensing prediction method according to claim 3, which is characterized in that obtain the area to be measured described
It is each described in the determination in domain after the history remote sensing image and corresponding historical production data of corn setting growth period
Before the property parameters of history remote sensing image, further includes:
The history remote sensing image is pre-processed, the pretreatment includes: radiation calibration, atmospheric correction and ortho-rectification；
According to the reflectivity for each pixel of history remote sensing image that pretreatment obtains, wave band calculating is carried out to each pixel, it is described
Wave band includes near infrared band, red spectral band, blue wave band and green light band, and the wave band is used for difference vegetation index, ratio
The calculating of vegetation index, enhancing vegetation index and green degree vegetation index.
5. corn yield remote sensing prediction method according to claim 3 or 4, which is characterized in that each described in the determination
After the property parameters of history remote sensing image, before the trained GA-BP neural network, further includes:
Determine sampled point；
Each property parameters that the sampled point respectively sets growth period are obtained, for as the training GA-BP neural network mould
Input when type.
6. a kind of corn yield remote sensing prediction system characterized by comprising
Remote sensing image obtains module, for obtaining the remote sensing image of corn setting growth period in region to be measured, the setting life
Long-term includes milk stage and maturity period；
Property parameters determining module, for determining that the property parameters of each remote sensing image, the property parameters include that difference is planted
By index, ratio vegetation index, enhancing vegetation index and green degree vegetation index；
Production forecast module, for will the property parameters input it is beautiful in region to be measured described in GA-BP Neural Network model predictive
The yield of rice, the GA-BP neural network model are the history remote sensing that growth period is set according to corn in the region to be measured
The model that image and the training of corresponding historical production data obtain.
7. corn yield remote sensing prediction system according to claim 6, which is characterized in that the system also includes:
Remote sensing image preprocessing module, for being pre-processed to the remote sensing image, it is described pretreatment include: radiation calibration,
Atmospheric correction and ortho-rectification；
Remote sensing image wave band computing module, the reflectivity of each pixel of the remote sensing image for being obtained according to pretreatment, to each
Pixel carries out wave band calculating, and the wave band includes near infrared band, red spectral band, blue wave band and green light band, the wave band
Calculating for difference vegetation index, ratio vegetation index, enhancing vegetation index and green degree vegetation index.
8. corn yield remote sensing prediction system according to claim 6, which is characterized in that the system also includes:
Historical data obtains module, for obtain the setting growth period of corn in the region to be measured history remote sensing image and
Corresponding historical production data；
History remote sensing image property parameters determining module, for determining the property parameters of each history remote sensing image, the category
Property parameter include difference vegetation index, ratio vegetation index, enhancing vegetation index and green degree vegetation index；
Neural network model training module, for being input with the property parameters of the history remote sensing image, with history production
Measuring data is output, and training GA-BP neural network obtains GA-BP neural network model.
9. corn yield remote sensing prediction system according to claim 8, which is characterized in that the system also includes:
History remote sensing image preprocessing module, for pre-processing to the history remote sensing image, the pretreatment includes: spoke
Penetrate calibration, atmospheric correction and ortho-rectification；
History remote sensing image wave band computing module, the reflection of each pixel of history remote sensing image for being obtained according to pretreatment
Rate carries out wave band calculating to each pixel, and the wave band includes near infrared band, red spectral band, blue wave band and green light band, institute
State calculating of the wave band for difference vegetation index, ratio vegetation index, enhancing vegetation index and green degree vegetation index.
10. corn yield remote sensing prediction system according to claim 8 or claim 9, which is characterized in that the system also includes:
Sampled point determining module, for determining sampled point；
Sampled point property parameters obtain module, and each property parameters of growth period are respectively set for obtaining the sampled point, are used for
Input when the GA-BP neural network model described as training.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910719979.0A CN110428107A (en) | 2019-08-06 | 2019-08-06 | A kind of corn yield remote sensing prediction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910719979.0A CN110428107A (en) | 2019-08-06 | 2019-08-06 | A kind of corn yield remote sensing prediction method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110428107A true CN110428107A (en) | 2019-11-08 |
Family
ID=68414271
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910719979.0A Pending CN110428107A (en) | 2019-08-06 | 2019-08-06 | A kind of corn yield remote sensing prediction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110428107A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111445023A (en) * | 2020-02-18 | 2020-07-24 | 中南林业科技大学 | Genetic algorithm optimized BP neural network GF-2 image forest classification method |
| CN111982822A (en) * | 2020-09-28 | 2020-11-24 | 武汉工程大学 | Long-time sequence high-precision vegetation index improvement algorithm |
| CN114612896A (en) * | 2022-03-07 | 2022-06-10 | 广东省科学院广州地理研究所 | Rice yield prediction method, device and equipment based on remote sensing image |
| CN114782829A (en) * | 2022-06-22 | 2022-07-22 | 浙江甲骨文超级码科技股份有限公司 | Method and system for constructing yield estimation model and yield estimation method and system |
| CN115471760A (en) * | 2022-10-31 | 2022-12-13 | 吉林高分遥感应用研究院有限公司 | Corn growth monitoring method and system, electronic equipment and computer storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104123409A (en) * | 2014-07-09 | 2014-10-29 | 江苏省农业科学院 | Field winter wheat florescence remote sensing and yield estimating method |
| CN107609695A (en) * | 2017-09-04 | 2018-01-19 | 北京信德智图科技有限公司 | Crop yield remote sensing estimation method based on adjustable vegetation index |
-
2019
- 2019-08-06 CN CN201910719979.0A patent/CN110428107A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104123409A (en) * | 2014-07-09 | 2014-10-29 | 江苏省农业科学院 | Field winter wheat florescence remote sensing and yield estimating method |
| CN107609695A (en) * | 2017-09-04 | 2018-01-19 | 北京信德智图科技有限公司 | Crop yield remote sensing estimation method based on adjustable vegetation index |
Non-Patent Citations (3)
| Title |
|---|
| 吴朝阳: ""应用遗传算法和神经网络预测粮食总产量"", 《科学技术与工程》 * |
| 墨蒙: ""基于遗传算法优化的BP 神经网络研究应用"", 《现代电子技术》 * |
| 安秦: ""长春地区玉米产量遥感估算方法研究"", 《中国优秀硕士学位论文全文数据库 农业科技辑》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111445023A (en) * | 2020-02-18 | 2020-07-24 | 中南林业科技大学 | Genetic algorithm optimized BP neural network GF-2 image forest classification method |
| CN111445023B (en) * | 2020-02-18 | 2022-04-19 | 中南林业科技大学 | Genetic algorithm optimized BP neural network GF-2 image forest classification method |
| CN111982822A (en) * | 2020-09-28 | 2020-11-24 | 武汉工程大学 | Long-time sequence high-precision vegetation index improvement algorithm |
| CN111982822B (en) * | 2020-09-28 | 2022-10-18 | 武汉工程大学 | Long-time sequence high-precision vegetation index improvement algorithm |
| CN114612896A (en) * | 2022-03-07 | 2022-06-10 | 广东省科学院广州地理研究所 | Rice yield prediction method, device and equipment based on remote sensing image |
| CN114782829A (en) * | 2022-06-22 | 2022-07-22 | 浙江甲骨文超级码科技股份有限公司 | Method and system for constructing yield estimation model and yield estimation method and system |
| CN115471760A (en) * | 2022-10-31 | 2022-12-13 | 吉林高分遥感应用研究院有限公司 | Corn growth monitoring method and system, electronic equipment and computer storage medium |
| CN115471760B (en) * | 2022-10-31 | 2023-02-28 | 吉林高分遥感应用研究院有限公司 | Corn growth monitoring method and system, electronic equipment and computer storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110428107A (en) | A kind of corn yield remote sensing prediction method and system | |
| CN105608293B (en) | Merge Biomass retrieval method and system on the woodland of spectrum and textural characteristics | |
| CN105004320B (en) | A kind of high score satellite data land table vegetation coverage inversion method and system | |
| CN110414738A (en) | A kind of crop yield prediction technique and system | |
| CN109213964A (en) | A kind of satellite AOD product bearing calibration for merging multi-source feature geographic factor | |
| CN108876021B (en) | Medium-and-long-term runoff forecasting method and system | |
| CN112163375B (en) | Long-time sequence near-ground ozone inversion method based on neural network | |
| CN106373022B (en) | BP-GA-based greenhouse crop planting efficiency condition optimization method and system | |
| CN110569605B (en) | NSGA 2-ELM-based non-glutinous rice leaf nitrogen content inversion model method | |
| CN111783987A (en) | Farmland reference crop evapotranspiration prediction method based on improved BP neural network | |
| CN101482929B (en) | Remote-sensing image processing method and system | |
| CN111855591A (en) | Rice overground part carbon-nitrogen ratio remote sensing inversion model and method | |
| CN104915982A (en) | Canopy layer illumination distribution prediction model construction method and illumination distribution detection method | |
| CN109214591B (en) | Method and system for predicting aboveground biomass of woody plant | |
| CN110969282A (en) | Runoff stability prediction method based on LSTM composite network | |
| CN108872508A (en) | A kind of eutrophy quality evaluation method of GA-BP optimization TSFNN | |
| CN116662753B (en) | Rice ear fertilizer nitrogen accurate recommendation method based on crop model and remote sensing coupling | |
| CN114581791A (en) | Inversion method and system for atmospheric water vapor content based on MODIS data | |
| CN110197308A (en) | A kind of crop monitoring system and method for agriculture Internet of Things | |
| CN105184400A (en) | Tobacco field soil moisture prediction method | |
| CN115859789A (en) | Method for improving inversion accuracy of polar atmosphere temperature profile | |
| CN109543911A (en) | A kind of solar radiation prediction technique and system | |
| CN114460015A (en) | Method and model for predicting chlorophyll content of canopy of rice in cold region | |
| CN116629453B (en) | Remote sensing yield estimation method suitable for whole crop growth period | |
| CN117035066A (en) | Ground surface temperature downscaling method coupling geographic weighting and random forest |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191108 |