CN109670509A - Winter wheat seedling stage growing way parameter evaluation method and system based on convolutional neural networks - Google Patents
Winter wheat seedling stage growing way parameter evaluation method and system based on convolutional neural networks Download PDFInfo
- Publication number
- CN109670509A CN109670509A CN201910007712.9A CN201910007712A CN109670509A CN 109670509 A CN109670509 A CN 109670509A CN 201910007712 A CN201910007712 A CN 201910007712A CN 109670509 A CN109670509 A CN 109670509A
- Authority
- CN
- China
- Prior art keywords
- seedling stage
- winter wheat
- neural networks
- convolutional neural
- target
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/32—Normalisation of the pattern dimensions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Computational Mathematics (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Pure & Applied Mathematics (AREA)
- Evolutionary Computation (AREA)
- Mathematical Optimization (AREA)
- Computational Linguistics (AREA)
- Mathematical Analysis (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Evolutionary Biology (AREA)
- Databases & Information Systems (AREA)
- Algebra (AREA)
- Probability & Statistics with Applications (AREA)
- Operations Research (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention provides a kind of winter wheat seedling stage growing way parameter evaluation method and system based on convolutional neural networks, comprising: the seedling stage image of continuous acquisition target winter wheat, and according to the corresponding input data set of seedling stage picture construction of the target winter wheat；The input data set is inputted into trained convolutional neural networks model, obtains the estimation result of the target winter wheat seedling stage growing way parameter；Wherein, the trained convolutional neural networks model is with training dataset for input, is obtained using gradient descent method training.High degree of automation of the present invention and recognition efficiency height, the intermediate link and manual intervention of winter wheat seedling stage growing way parameter estimation can be effectively reduced, the application cost and complexity for reducing identification process, effectively improve the accuracy and real-time for setting winter wheat seedling stage growing way parameter estimation.
Description
Technical field
The present embodiments relate to depth learning technology field, more particularly, to a kind of based on convolutional neural networks
Winter wheat seedling stage growing way parameter evaluation method and system.
Background technique
Leaf area index and ground biomass are two important parameters for characterizing winter wheat growing way.The leaf area of Field Scale
Index and ground biomass estimation have great importance for winter wheat seedling stage Growing state survey and field precision management.Traditional
Leaf area index and geodyte measuring method need the sampling of field destructiveness and manual measurement analysis, and there are low efficiencys, work
The problems such as work amount is big is not able to satisfy the Plant phenotypic analysis demand of high-throughput automation.Remote sensing is the long potential parameter of current winter wheat
One of main method of nondestructive measurement, using the canopy of winter wheat spectroscopic data of acquisition, by calculating vegetation index and and growing way
Parameter measured data carries out regression analysis, can be realized the nondestructive measurement of leaf area index and ground biomass.But due to spectrum
Data acquisition is needed using dedicated equipment, and this method exists certain insufficient in terms of use cost and convenience.
Visible images have many advantages, such as that at low cost, data acquisition is convenient.Based on computer vision technique, from visible light figure
Numerical characteristic is extracted as in, accurate Fitting Analysis can be carried out to leaf area index and ground biomass.Although this method takes
Certain effect was obtained, but there are problems that 2, comprising largely by light in the winter wheat image that (1) is acquired vulnerable to noise jamming, field
According to the noise that uneven and complex background generates, there is serious influence to the accuracy rate of winter wheat image segmentation and feature extraction；
(2) higher to the degree of dependence of characteristics of image, but the characteristics of image generalization ability of usually engineer is limited, causes this method difficult
To expand application.Convolutional neural networks are one of current most effective deep learning methods, can directly using image as inputting,
It with high accuracy for examination, is identified in weeds, pest, the multiple fields such as plant disease, stress diagnosis, Agriculture Image segmentation
It is widely used.
And in the winter wheat seedling stage image obtained in the actual environment of field, due to existing largely by illumination condition unevenness
The reasons such as the noise that even, complex background environment generates, certainly will will affect the accuracy rate of feature extraction and optimization, and increase at image
The calculation amount of reason method, so can pair winter wheat seedling stage growing way parameter estimation real-time and accuracy impact, cause
This method is difficult to promote in practical applications.
Therefore, how to design it is a kind of low cost, effectively reduce intermediate link and the winter wheat seedling stage growing way of manual intervention
Parameter Method of fast estimating, is a problem to be solved.
Summary of the invention
The embodiment of the invention provides it is a kind of overcome the above problem or at least be partially solved the above problem based on volume
The winter wheat seedling stage growing way parameter evaluation method and system of product neural network.
The embodiment of the invention provides a kind of that the winter wheat seedling stage growing way parameter based on convolutional neural networks is estimated for first aspect
Calculation method characterized by comprising
The seedling stage image of continuous acquisition target winter wheat, and it is corresponding according to the seedling stage picture construction of the target winter wheat
Input data set；
The input data set is inputted into trained convolutional neural networks model, it is long to obtain the target winter wheat seedling stage
The estimation result of potential parameter；Wherein, the trained convolutional neural networks model is to utilize ladder with training dataset for input
Degree descent method training obtains.
On the other hand the embodiment of the invention provides a kind of, and the winter wheat seedling stage growing way parameter based on convolutional neural networks is estimated
Calculation system, comprising:
Seedling stage image acquisition units, for the seedling stage image of continuous acquisition target winter wheat, and it is small according to the target winter
The corresponding input data set of seedling stage picture construction of wheat；
Growing way parameter estimation unit is obtained for the input data set to be inputted trained convolutional neural networks model
To the estimation result of the target winter wheat seedling stage growing way parameter；Wherein, the trained convolutional neural networks model be with
Training dataset is input, is obtained using gradient descent method training.
The embodiment of the invention provides include processor, communication interface, memory and bus for the third aspect, wherein processing
Device, communication interface, memory complete mutual communication by bus, and processor can call the logical order in memory,
To execute the winter wheat seedling stage growing way parameter evaluation method based on convolutional neural networks of first aspect offer.
The embodiment of the invention provides a kind of non-transient computer readable storage medium, the non-transient calculating for fourth aspect
Machine readable storage medium storing program for executing stores computer instruction, the computer instruction make that the computer executes that first aspect provides based on
The winter wheat seedling stage growing way parameter evaluation method of convolutional neural networks.
As shown from the above technical solution, a kind of winter wheat seedling stage growing way ginseng based on convolutional neural networks provided by the invention
Number evaluation method and system, evaluation method includes automatic collection target winter wheat seedling stage image, and constructs input data set；By institute
The input that the corresponding input data set of seedling stage image is model is stated, is established for estimating the target winter wheat seedling stage growing way parameter
Convolutional neural networks model；And the convolutional neural networks model is carried out according to gradient descent algorithm and test data set
Model training, verifying and test obtain the estimation result to the target winter wheat seedling stage growing way parameter.The present invention automates journey
Degree is high and estimation accuracy rate is high, can effectively reduce the intermediate link and manual intervention of winter wheat seedling stage growing way parameter estimation, drop
The application cost and complexity of low leaf area index and ground biomass estimation process effectively improve winter wheat seedling stage growing way ginseng
The accuracy and real-time of number estimation, also the correlative study for winter wheat seedling stage Growing state survey provides reliable and accurate data
Basis.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is this hair
Bright some embodiments for those of ordinary skill in the art without creative efforts, can be with root
Other attached drawings are obtained according to these attached drawings.
Fig. 1 is a kind of winter wheat seedling stage growing way parameter estimation side based on convolutional neural networks provided in an embodiment of the present invention
Method flow chart；
Fig. 2 is step 100 in the winter wheat seedling stage growing way parameter evaluation method of the invention based on convolutional neural networks
Flow chart；
Fig. 3 be in the winter wheat seedling stage growing way parameter evaluation method of the invention based on convolutional neural networks step 200 it
The preceding flow chart for establishing convolutional neural networks model；
Fig. 4 is the structural schematic diagram of convolutional neural networks model of the invention；
Fig. 5 is step 200 in the winter wheat seedling stage growing way parameter evaluation method of the invention based on convolutional neural networks
Flow chart；
Fig. 6 is the application example of the winter wheat seedling stage growing way parameter evaluation method of the invention based on convolutional neural networks
Flow diagram；
Fig. 7 is a kind of winter wheat seedling stage growing way parameter estimation system based on convolutional neural networks provided in an embodiment of the present invention
The structural block diagram of system；
Fig. 8 is the structural schematic diagram of a kind of electronic equipment provided in an embodiment of the present invention.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical solution in the embodiment of the present invention is explicitly described, it is clear that described embodiment is the present invention
A part of the embodiment, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not having
Every other embodiment obtained under the premise of creative work is made, shall fall within the protection scope of the present invention.
Fig. 1 is a kind of winter wheat seedling stage growing way parameter estimation side based on convolutional neural networks provided in an embodiment of the present invention
Method flow chart, as shown in Figure 1, comprising:
Step 100, the seedling stage image of continuous acquisition target winter wheat, and according to the seedling stage image structure of the target winter wheat
Build corresponding input data set.
In step 100, the seedling stage in the winter wheat seedling stage growing way parameter estimation system based on convolutional neural networks
The target winter wheat seedling stage image that image acquisition units collect automatically, and acquire the specific of target winter wheat seedling stage image
Mode can by target winter wheat test field image acquisition region be arranged image capture device, the image capture device according to
Preset control parameters carry out automatic collection to mesh winter wheat seedling stage image, and collected seedling stage image is sent to the seedling stage
Image acquisition units.It is understood that the winter wheat is the wheat planted under field condition.It is understood that institute
Input data set is stated to be made of the image data of the seedling stage image.
For example, described image acquisition equipment is specifically as follows a kind of industrial camera；It is set in the winter wheat test Tanaka
Multiple industrial cameras are equipped with, each industrial camera is connect with a controller, and controller is used for according to default control
State modulator industrial camera acquires winter wheat seedling stage image, and the industrial camera will each collected winter wheat seedling stage figure
As the seedling stage Image Acquisition list being sent in the winter wheat seedling stage growing way parameter estimation system based on convolutional neural networks
Member, so that the seedling stage image acquisition units receive winter wheat seedling stage image automatically.
It is understood that the seedling stage image is specially RGB color image, RGB therein is by red channel
(R), green channel (G), blue channel (B) composition, blue=white of the green of most bright red+most bright+most bright；It is most dark
Red+most dark green+most dark blue=black；And between most bright and most dark, the red of identical shading value+identical bright
The green of darkness+identical shading value blue=grey.In any one channel of RGB, white and black indicate this face
The shading value of color.So having white or linen place, tri- channels R, G, B are impossible to be black, because necessary
There are tri- channels R, G, B to constitute these colors.
Step 200, the input data set is inputted into trained convolutional neural networks model, it is small obtains the target winter
The estimation result of wheat seeding phase long potential parameter；Wherein, it is defeated that the trained convolutional neural networks model, which is with training dataset,
Enter, is obtained using gradient descent method training.
In step 200, the seedling stage image acquisition units are in the target winter wheat seedling stage image collected automatically
Afterwards, meeting multiple target winter wheat seedling stage images after preset quantity, to be sent to the winter based on convolutional neural networks small
Convolutional neural networks model foundation unit in wheat seeding phase growing way parameter estimation system, the convolutional neural networks model foundation list
Member receives the multiple target winter wheat seedling stage image, and using the corresponding input data set of the seedling stage image as the defeated of model
Enter, according to characteristics such as the sizes of the corresponding input data set of the seedling stage image, establishes dedicated for estimating that the target winter is small
The convolutional neural networks model of wheat seeding phase long potential parameter.It is understood that the convolutional neural networks model is in addition to successively connecting
Outside input layer, full articulamentum and the output layer connect, the convolutional layer that is connected between input layer and full articulamentum
The set-up mode and the number of plies of (convolutional layer) and pond layer (pooling layer) are schemed depending on the seedling stage
The characteristics such as the size as corresponding input data set.It is understood that convolutional neural networks CNN (Convolutional
Neural Network) it is a kind of feedforward neural network, its artificial neuron can respond the week in a part of coverage area
Unit is enclosed, has outstanding performance suitable for image procossing.
The convolutional neural networks model foundation unit is completed for estimating the target winter wheat seedling stage growing way parameter
Convolutional neural networks model after, the winter in the winter wheat seedling stage growing way parameter estimation system based on convolutional neural networks is small
The evaluation unit of wheat seeding phase long potential parameter is according to gradient descent algorithm and the test data set to the convolutional neural networks mould
Type carries out model training, verifying and test, obtains to the target winter wheat seedling stage growing way parameter estimation result.
It is understood that the gradient descent algorithm is an optimization algorithm, also commonly referred to as steepest descent method.Most
Fast descent method is to solve for most simple and most ancient one of the method for unconstrained optimization problem, although not had now practical
Property, but many efficient algorithms are all obtained from being improved and corrected based on it.Steepest descent method is to use negative gradient
Direction is the direction of search, and for steepest descent method closer to target value, step-length is smaller, is advanced slower.
A kind of winter wheat seedling stage growing way parameter evaluation method based on convolutional neural networks provided in an embodiment of the present invention, from
Dynamicization degree height and estimation accuracy rate height can effectively reduce the intermediate link of winter wheat seedling stage growing way parameter estimation and manually do
In advance, the application cost and complexity for reducing leaf area index and ground biomass estimation process, effectively improve winter wheat seedling stage
The accuracy and real-time of growing way parameter estimation, also the correlative study for winter wheat seedling stage Growing state survey provides reliable and accurate
Data basis.
In the above-described embodiments, as shown in Fig. 2, the seedling stage picture construction according to the target winter wheat is corresponding defeated
Enter data set, specifically include:
Step A00 pre-processes the seedling stage image of the target winter wheat, obtains initial data set；
Step B00 expands using rotation and overturns the data enhancement method expanded, the data concentrated to the primary data
Amount is expanded, and the corresponding input data set of the target winter wheat seedling stage image is obtained.
In step A00, the winter wheat seedling stage growing way parameter estimation system based on convolutional neural networks is to the seedling
Phase image is pre-processed, described to carry out pretreatment including image denoising to the seedling stage image and adjust the seedling stage image
Size, such as the pixel of collected multiple target seedling stage images is uniformly adjusted to 128 × 128 × 3.
It is understood that in step 200 to the convolutional neural networks model carry out model training, verifying and
The process of test, initial data set is divided are as follows: training dataset, validation data set and test data set, wherein training data
The data volume ratio that collection, validation data set and test data are concentrated can be different, and the ratio of the data volume of training data concentration
Highest.
In step B00, the winter wheat seedling stage growing way parameter estimation system based on convolutional neural networks is using rotation
Expand and/or the data enhancement method of overturning expansion, the data volume concentrated to the primary data expand, obtains the seedling
The corresponding input data set of phase image.It is understood that described be rotated or turned over expansion as the seedling stage image after duplication
After rotating a certain angle or overturn along outer profile, the new seedling stage image collectively constituted with former seedling stage image, new seedling
Phase image is the seedling stage image after expanding；Wherein, the number rotated and turn over determines the expansion of the seedling stage image
Multiple.Picture size when according to the seedling stage Image Acquisition and subsequent image ruler required when modeling processing is carried out to it
It is very little, the seedling stage image can be extended to 2 to 100 times.
As can be seen from the above description, the winter wheat seedling stage growing way ginseng based on convolutional neural networks that the embodiment of the present invention provides
Number evaluation method, can effectively expand target winter wheat seedling stage image, and improve the estimation target Winter Wheat Seedling
The reliability of the data basis of phase long potential parameter.
In the above-described embodiments, as shown in figure 3, the input data set is inputted trained convolutional neural networks mould
Type, before obtaining the estimation result of the target winter wheat seedling stage growing way parameter, further includes:
Step 201, using the corresponding input data set of the target winter wheat seedling stage image as the convolutional neural networks
Input layer input；
Step 202, the picture size for the image data concentrated according to the input data establishes the convolutional neural networks
The processing module of model；
Step 203, and, the input layer, processing module, the output for preventing the processing module are sequentially connected
The Dropout layer of fitting, for converting the full articulamentum of one-dimensional vector for the output of the processing module and being used to export institute
The regression forecasting output layer for stating target winter wheat seedling stage growing way parameter estimation result, completes building for the convolutional neural networks model
It is vertical.
In above-mentioned steps 201 to 203, the integral frame of the convolutional neural networks model is as shown in figure 4, the processing
Sequentially connected five processing units are included at least in module, and first processing unit includes into the 4th processing unit
Sequentially connected convolutional layer, crowd normalization layer (Batch Normalization Layer, BN layer), the linear unit R eLU of amendment
Layer and pond layer, the 5th processing unit include the convolutional layer, batch BN layers of standardization, the linear unit R eLU of amendment of connection
Layer；
Wherein, the size constancy and convolution kernel of the convolution kernel in each convolutional layer in sequentially connected five processing units
Quantity it is incremented by successively, the size of the convolution kernel of the convolutional layer in each processing unit is all larger than described in same processing unit
The size of convolution kernel in the layer of pond.
As can be seen from the above description, the winter wheat seedling stage growing way ginseng based on convolutional neural networks that the embodiment of the present invention provides
Number evaluation method, provides a kind of for estimating the effective and reliable of the convolutional neural networks model of winter wheat seedling stage growing way parameter
Method for building up, the structure therein for estimating the convolutional neural networks model of winter wheat seedling stage growing way parameter is accurate and is directed to
Property is strong, can accurately estimate winter wheat seedling stage growing way parameter.
In the above-described embodiments, as shown in figure 5, described input trained convolutional neural networks for the input data set
Model obtains the estimation result of the target winter wheat seedling stage growing way parameter, specifically includes:
Step 301, gradient descent algorithm is based in five processing units in the processing module successively to the input
Data set is handled and is transmitted, until exporting target signature in five processing units；Wherein, the processing module
In the convolutional layer extract the characteristic pattern of the received input data set；Described BN layers to the characteristic pattern received into
Row normalized；Described ReLU layers carries out non-linear transfer to the characteristic pattern after the normalized received；It is described
Pond layer carries out size reduction processing to the characteristic pattern after received non-linear transfer.
In step 301, the convolutional layer in the processing module extracts the feature of the received input data set
Figure, comprising:
The convolutional layer extracts the characteristic pattern of the received input data set according to formula one:
In formula one, xijFor the characteristic pattern of i-th of output of j-th of convolutional layer in convolutional neural networks model；M is volume
The quantity of j-th of convolutional layer input feature vector figure in product neural network model；
bijFor i-th of bias term of j-th of convolutional layer；φ () is nonlinear activation function；
Corresponding, the pond layer carries out size reduction processing to the characteristic pattern after received non-linear transfer, comprising:
Characteristic pattern after received non-linear transfer is carried out size reduction processing according to formula two by the pond layer:
In formula two, down () is down-sampled function；F is desampling fir filter size；S is down-sampled step-length.
Step 302, the target signature is converted one-dimensional vector by the full articulamentum, comprising:
The full articulamentum converts one-dimensional vector for the target signature according to formula three:
In formula three, vjFor the output one-dimensional vector of j-th of full articulamentum；wjFor the weight square of j-th of full articulamentum
Battle array；bjThe bias term of j-th of full articulamentum；φ () is nonlinear activation function.
Step 303, the regression forecasting layer exports the target winter wheat seedling stage growing way parameter according to the one-dimensional vector
Estimation result.
As can be seen from the above description, high degree of automation of the present invention and estimation accuracy rate height, can effectively reduce Winter Wheat Seedling
The intermediate link and manual intervention of phase growing way parameter estimation, reduction leaf area index and ground biomass estimation process are applied to
Sheet and complexity, effectively improve the accuracy and real-time of winter wheat seedling stage growing way parameter estimation.
For further instruction this programme, the winter wheat seedling stage growing way based on convolutional neural networks that the present invention also provides a kind of
The application example of parameter evaluation method specifically includes following content referring to Fig. 6:
S101, winter wheat seedling stage image is obtained, and the seedling stage image is pre-processed；
Specifically, the pretreatment includes image denoising and the size for adjusting the seedling stage image, such as it is adjusted to 128 ×
128 × 3 pixels.
S102, according to the pretreated seedling stage picture construction initial data set, using data enhancement methods, to described
The data volume of initial data set is expanded；
Specifically, it is based on the pretreated seedling stage image, constructs initial data set, and use data enhancement methods,
Initial data set is expanded.The data enhancement methods that the present invention uses are as follows:
Seedling stage image clockwise is rotated 90,180 and 270 degree, and carries out flip horizontal and flip vertical.Rotation and
On the basis of overturning, the color space of the image data set rotated and turn over is changed to HSV from RGB, by the conversion face
Image after the colour space carries out brightness enhancing and reduced adjustment, the brightness adjustment amplitude be respectively brightness enhancing 10%,
20%, reduce by 10%, 20%.The data volume of data set is 26 times of initial data set after the expansion.
Data set is divided into 3 part of training dataset, validation data set and test data set, ratio is respectively as follows: 60%,
15% and 25%.
S103, building convolutional neural networks model structure are right according to the data set and gradient descent method after the expansion
The convolutional neural networks model is trained, verifies and tests；
Specifically, convolutional neural networks model of the present invention is using winter wheat seedling stage RGB color space figure picture as input, to estimate
The winter wheat seedling stage growing way parameter of calculation includes 6 modules as output altogether.First module to the 4th module includes 1
Convolutional layer, 1 BN layers, 1 RELU layers and 1 pond layer.The convolution kernel size of convolutional layer is 5 × 5, for obtaining input picture
Feature.The RELU layers of characteristic pattern for generating convolutional layer carry out non-linear transfer, realize the sparse of network, reduce the mutual of parameter
Dependence alleviates the generation of overfitting problem.The RELU layers of size that will not change characteristic pattern.The convolution kernel of pond layer is big
Small is 2 × 2, step-length 2, is maximum pond, and characteristic pattern size is reduced to original 1/2.It is rolled up in four module convolutional layers
The size of product core gradually increases, and is followed successively by 32,64,128,256.The size of 5th module convolution kernel and the function of each layer are with before
Four modules are identical, including 1 convolutional layer, and 1 BN layers, 1 RELU layers, convolution kernel number is 512 in convolutional layer.The last one
Module includes 2 dropout layers, 2 full articulamentums, 1 recurrence output layer, and 2 dropout layers of loss ratios are 0.5,2
First full articulamentum contains 500 neurons in full articulamentum, and second full articulamentum contains 1 neuron.Return output
The winter wheat seedling stage growing way parameter of layer estimation includes leaf area index, biomass.
S104, according to the convolutional neural networks model, growing way parameter estimation is carried out to the seedling stage image of input；
Specifically, the convolutional neural networks model to the seedling stage image of input into long potential parameter include leaf area index,
The estimation of growth parameter(s).
Winter wheat seedling stage growing way parameter evaluation method provided by the invention based on convolutional neural networks, can be with winter wheat
Seedling stage image is enhanced by data directly as the input of appraising model and expands input data amount, and convolutional neural networks mould is constructed
Type is simultaneously trained, verifies and tests, and realizes the quick estimation to winter wheat seedling stage growing way parameter, improves the accuracy of estimation,
Realize the practical application of winter wheat seedling stage growing way parameter evaluation method.
On the basis of the various embodiments described above, further, convolution kernel in convolutional neural networks model convolutional layer of the present invention
The specific implementation of feature extraction are as follows:
In formula one, xijFor the characteristic pattern of i-th of output of j-th of convolutional layer in convolutional neural networks model；M is volume
The quantity of j-th of convolutional layer input feature vector figure in product neural network model；bijFor i-th of bias term of j-th of convolutional layer；φ
() is nonlinear activation function；
Corresponding, the pond layer carries out size reduction processing to the characteristic pattern after received non-linear transfer, comprising:
Characteristic pattern after received non-linear transfer is carried out size reduction processing according to formula two by the pond layer:
In formula two, down () is down-sampled function；F is desampling fir filter size；S is down-sampled step-length.
The target signature is converted one-dimensional vector by the full articulamentum, comprising:
The full articulamentum converts one-dimensional vector for the target signature according to formula three:
In formula three, vjFor the output one-dimensional vector of j-th of full articulamentum；wjFor the weight square of j-th of full articulamentum
Battle array；bjThe bias term of j-th of full articulamentum；φ () is nonlinear activation function.
On the basis of the various embodiments described above, further, convolutional neural networks model of the present invention is calculated using gradient decline
Method is trained.
As can be seen from the above description, the winter wheat seedling stage growing way ginseng based on convolutional neural networks that the embodiment of the present invention provides
Number evaluation method, high degree of automation and recognition efficiency height, can effectively reduce the centre of winter wheat seedling stage growing way parameter estimation
Link and manual intervention reduce the application cost and complexity of estimation process, effectively improve winter wheat seedling stage growing way parameter and estimate
The accuracy and real-time of calculation, also the correlative study for winter wheat seedling stage Growing state survey provides reliable and accurate data base
Plinth.
Fig. 7 is a kind of winter wheat seedling stage growing way parameter estimation system based on convolutional neural networks provided in an embodiment of the present invention
The structural block diagram of system, as shown in fig. 7, comprises: seedling stage image acquisition units 10 and growing way parameter estimation unit 20.Wherein:
Seedling stage image acquisition units 10 are used for the seedling stage image of continuous acquisition target winter wheat, and small according to the target winter
The corresponding input data set of seedling stage picture construction of wheat.Growing way parameter estimation unit 20 is used to input the input data set and instruct
The convolutional neural networks model perfected obtains the estimation result of the target winter wheat seedling stage growing way parameter；Wherein, the training
Good convolutional neural networks model is with training dataset for input, is obtained using gradient descent method training.
Specifically, the available winter wheat seedling stage image of seedling stage image acquisition units 10, and the seedling stage image is carried out
Pretreatment, the pretreatment includes image denoising and the size for adjusting the seedling stage image, such as is adjusted to 128 × 128 × 3 pictures
Element.
Further, expanded just according to pretreated seedling stage picture construction initial data set using data enhancement methods
The size of beginning data set.For example, by seedling stage image clockwise rotate 90,180 and 270 degree, and carry out flip horizontal and vertically
Overturning.Based on data enhancement methods of the present invention, the data volume of data set is extended for 26 times of initial data set after expansion.By data
Collection is divided into 3 part of training dataset, validation data set and test data set, and ratio is respectively as follows: 60%, 15% and 25%.
Growing way parameter estimation unit 20 is for constructing convolutional neural networks structure of the present invention.Convolutional neural networks mould of the present invention
Type is using the RGB color image in winter wheat seedling stage as input, using the winter wheat seedling stage growing way parameter of estimation as output, altogether
Including 6 modules.First module to the 4th module includes 1 convolutional layer, 1 BN layers, 1 RELU layers and 1 pond
Layer.The convolution kernel size of convolutional layer is 5 × 5, for obtaining the feature of input picture.The RELU layers of characteristic pattern for generating convolutional layer
Non-linear transfer is carried out, the sparse of network is realized, reduces the relation of interdependence of parameter, alleviate the generation of overfitting problem.
The RELU layers of size that will not change characteristic pattern.The convolution kernel size of pond layer is 2 × 2, and step-length 2 is maximum pond,
Characteristic pattern size is reduced to original 1/2.The size of convolution kernel gradually increases in four module convolutional layers, be followed successively by 32,64,
128,256.The size of 5th module convolution kernel and the function of each layer are identical as first four module, including 1 convolutional layer, and 1
BN layers, 1 RELU layers, the convolution kernel number of convolutional layer is 512.The last one module includes 2 dropout layers, 2 full connections
Layer, 1 recurrence output layer, 2 dropout layers of loss ratios be in 0.5,2 full articulamentums first full articulamentum contain 500
A neuron, second full connection survey and contain 1 neuron.Returning output layer estimation winter wheat seedling stage growing way parameter includes blade face
Product index, biomass.The growing way parameter estimation unit 20 carries out convolution based on the data set and gradient descent algorithm after expansion
Training, verifying and the test of neural network, and based on the convolutional neural networks model after test, to the winter wheat seedling stage of input
Image carries out the estimation of long potential parameter.
On the basis of the various embodiments described above, further, convolution kernel feature in 20 convolutional layer of growing way parameter estimation unit
The specific implementation of extraction are as follows:
In formula one, xijFor the characteristic pattern of i-th of output of j-th of convolutional layer in convolutional neural networks model；M is volume
The quantity of j-th of convolutional layer input feature vector figure in product neural network model；bijFor i-th of bias term of j-th of convolutional layer；φ
() is nonlinear activation function；
Corresponding, the pond layer carries out size reduction processing to the characteristic pattern after received non-linear transfer, comprising:
Characteristic pattern after received non-linear transfer is carried out size reduction processing according to formula two by the pond layer:
In formula two, down () is down-sampled function；F is desampling fir filter size；S is down-sampled step-length.
The target signature is converted one-dimensional vector by the full articulamentum, comprising:
The full articulamentum converts one-dimensional vector for the target signature according to formula three:
In formula three, vjFor the output one-dimensional vector of j-th of full articulamentum；wjFor the weight square of j-th of full articulamentum
Battle array；bjThe bias term of j-th of full articulamentum；φ () is nonlinear activation function.
The embodiment of winter wheat seedling stage growing way parameter estimation provided by the invention based on convolutional neural networks specifically can be with
For executing the process flow of the embodiment of the above-mentioned winter wheat seedling stage growing way parameter evaluation method based on convolutional neural networks,
Details are not described herein for function, is referred to the detailed description of above method embodiment.
As can be seen from the above description, the winter wheat seedling stage growing way ginseng based on convolutional neural networks that the embodiment of the present invention provides
Number estimation can be enhanced by data with winter wheat seedling stage image directly as the input of identification model and expand input data amount,
Building convolutional neural networks model is simultaneously trained, verifies and tests, and realizes the estimation of winter wheat seedling stage growing way parameter, improves
The accuracy of estimation realizes the practical application of winter wheat seedling stage growing way parameter estimation.
Fig. 8 is the structural schematic diagram of a kind of electronic equipment provided in an embodiment of the present invention, as shown in figure 8, electronic equipment packet
It includes: processor (processor) 801, communication interface (Communications Interface) 802, memory (memory)
803 and bus 804, wherein processor 801, communication interface 802, memory 803 complete mutual communication by bus 804.
Processor 801 can call the logical order in memory 803, to execute following method, for example, the continuous acquisition target winter
The seedling stage image of wheat, and according to the corresponding input data set of seedling stage picture construction of the target winter wheat；By the input
Data set inputs trained convolutional neural networks model, obtains the estimation result of the target winter wheat seedling stage growing way parameter；
Wherein, the trained convolutional neural networks model is with training dataset for input, is obtained using gradient descent method training
's.
Logical order in above-mentioned memory 802 can be realized and as independent by way of SFU software functional unit
Product when selling or using, can store in a computer readable storage medium.Based on this understanding, of the invention
Substantially the part of the part that contributes to existing technology or the technical solution can be produced technical solution in other words with software
The form of product embodies, which is stored in a storage medium, including some instructions are used so that one
Platform computer equipment (can be personal computer, server or the network equipment etc.) executes described in each embodiment of the present invention
The all or part of the steps of method.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (ROM, Read-
Only Memory), random access memory (RAM, Random Access Memory), magnetic or disk etc. are various can be with
Store the medium of program code.
The embodiment of the present invention provides a kind of non-transient computer readable storage medium, the non-transient computer readable storage
Medium storing computer instruction, the computer instruction make the computer execute side provided by above-mentioned each method embodiment
Method, for example, the seedling stage image of continuous acquisition target winter wheat, and according to the seedling stage picture construction pair of the target winter wheat
The input data set answered；The input data set is inputted into trained convolutional neural networks model, it is small to obtain the target winter
The estimation result of wheat seeding phase long potential parameter；Wherein, it is defeated that the trained convolutional neural networks model, which is with training dataset,
Enter, is obtained using gradient descent method training.
Those of ordinary skill in the art will appreciate that: realize that all or part of the steps of above method embodiment can pass through
The relevant hardware of program instruction is completed, and program above-mentioned can be stored in a computer readable storage medium, the program
When being executed, step including the steps of the foregoing method embodiments is executed；And storage medium above-mentioned includes: ROM, RAM, magnetic disk or light
The various media that can store program code such as disk.
The embodiments such as communication equipment described above are only schematical, wherein unit as illustrated by the separation member
It may or may not be physically separated, component shown as a unit may or may not be physics list
Member, it can it is in one place, or may be distributed over multiple network units.It can be selected according to the actual needs
In some or all of the modules achieve the purpose of the solution of this embodiment.Those of ordinary skill in the art are not paying creativeness
Labour in the case where, it can understand and implement.
Through the above description of the embodiments, those skilled in the art can be understood that each embodiment can
It realizes by means of software and necessary general hardware platform, naturally it is also possible to pass through hardware.Based on this understanding, on
Stating technical solution, substantially the part that contributes to existing technology can be embodied in the form of software products in other words, should
Computer software product may be stored in a computer readable storage medium, such as ROM/RAM, magnetic disk, CD, including several fingers
It enables and using so that a computer equipment (can be personal computer, server or the network equipment etc.) executes each implementation
The method of certain parts of example or embodiment.
Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although
Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: it still may be used
To modify the technical solutions described in the foregoing embodiments or equivalent replacement of some of the technical features；
And these are modified or replaceed, technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution spirit and
Range.
Claims (10)
1. a kind of winter wheat seedling stage growing way parameter evaluation method based on convolutional neural networks characterized by comprising
The seedling stage image of continuous acquisition target winter wheat, and according to the corresponding input of seedling stage picture construction of the target winter wheat
Data set；
The input data set is inputted into trained convolutional neural networks model, obtains the target winter wheat seedling stage growing way ginseng
Several estimation results；Wherein, the trained convolutional neural networks model is with training dataset for input, using under gradient
The training of drop method obtains.
2. method according to claim 1, which is characterized in that the seedling stage picture construction pair according to the target winter wheat
The input data set answered, specifically includes:
The seedling stage image of the target winter wheat is pre-processed, initial data set is obtained；
Expanded using rotation and the data enhancement method of overturning expansion, the data volume concentrated to the primary data expand,
Obtain the corresponding input data set of the target winter wheat seedling stage image.
3. method according to claim 1, which is characterized in that the input data set is being inputted trained convolutional Neural
Network model, before obtaining the estimation result of the target winter wheat seedling stage growing way parameter, further includes:
Using the corresponding input data set of the target winter wheat seedling stage image as the defeated of the input layer of the convolutional neural networks
Enter；
The picture size for the image data concentrated according to the input data establishes the processing mould of the convolutional neural networks model
Block；
And it is sequentially connected the input layer, processing module, the output over-fitting for preventing the processing module
Dropout layers, for converting the full articulamentum of one-dimensional vector for the output of the processing module and being used to export the target
The regression forecasting output layer of winter wheat seedling stage growing way parameter estimation result, completes the foundation of the convolutional neural networks model.
4. method according to claim 3, which is characterized in that included at least in the processing module at five once connected
Manage unit, and first processing unit into the 4th processing unit include sequentially connected convolutional layer, batch standardization BN layers,
Correct linear unit R eLU layers and pond layer, the 5th processing unit include the convolutional layer of connection, batch BN layers of standardization and
Correct linear unit R eLU layers；
Wherein, the number of the size constancy of the convolution kernel in each convolutional layer in sequentially connected five processing units and convolution kernel
Measure incremented by successively, the size of the convolution kernel of the convolutional layer in each processing unit is all larger than pond layer described in same processing unit
In convolution kernel size.
5. method according to claim 4, which is characterized in that described that the input data set is inputted trained convolution mind
Through network model, the estimation result of the target winter wheat seedling stage growing way parameter is obtained, is specifically included:
In five processing units in the processing module based on gradient descent algorithm successively to the input data set at
Reason and transmitting, until exporting target signature in five processing units；Wherein, the convolution in the processing module
Layer extracts the characteristic pattern of the received input data set；Described BN layers is normalized place to the characteristic pattern received
Reason；Described ReLU layers carries out non-linear transfer to the characteristic pattern after the normalized received；The pond layer docking
Characteristic pattern after the non-linear transfer of receipts carries out size reduction processing；
Described Dropout layers by the neuron in the target signature according to certain probability random drop；
The target signature is converted one-dimensional vector by the full articulamentum；
And the regression forecasting layer exports the target winter wheat seedling stage growing way parameter estimation knot according to the one-dimensional vector
Fruit.
6. method according to claim 5, which is characterized in that the convolutional layer in the processing module extracts received institute
The characteristic pattern for stating input data set specifically includes:
The convolutional layer extracts the characteristic pattern of the received input data set according to formula one:
In formula one, xijFor the characteristic pattern of i-th of output of j-th of convolutional layer in convolutional neural networks model；M is convolution mind
Quantity through j-th of convolutional layer input feature vector figure in network model；bijFor i-th of bias term of j-th of convolutional layer；φ () is
Nonlinear activation function；
Corresponding, the pond layer carries out size reduction processing to the characteristic pattern after received non-linear transfer, comprising:
Characteristic pattern after received non-linear transfer is carried out size reduction processing according to formula two by the pond layer:
In formula two, down () is down-sampled function；F is desampling fir filter size；S is down-sampled step-length.
7. method according to claim 5, which is characterized in that the full articulamentum converts the target signature to one-dimensional
Vector, comprising:
The full articulamentum converts one-dimensional vector for the target signature according to formula three:
In formula three, vjFor the output one-dimensional vector of j-th of full articulamentum；wjFor the weight matrix of j-th of full articulamentum；bj
The bias term of j-th of full articulamentum；φ () is nonlinear activation function.
8. a kind of winter wheat seedling stage growing way parameter estimation system based on convolutional neural networks characterized by comprising
Seedling stage image acquisition units, for the seedling stage image of continuous acquisition target winter wheat, and according to the target winter wheat
The corresponding input data set of seedling stage picture construction；
Growing way parameter estimation unit obtains institute for the input data set to be inputted trained convolutional neural networks model
State the estimation result of target winter wheat seedling stage growing way parameter；Wherein, the trained convolutional neural networks model is with training
Data set is input, is obtained using gradient descent method training.
9. a kind of electronic equipment, which is characterized in that including processor, communication interface, memory and bus, wherein processor leads to
Believe that interface, memory complete mutual communication by bus, processor can call the logical order in memory, to execute
Winter wheat seedling stage growing way parameter evaluation method as described in any one of claim 1 to 7 based on convolutional neural networks.
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, it is as described in any one of claim 1 to 7 based on convolution that the computer instruction executes the computer
The winter wheat seedling stage growing way parameter evaluation method of neural network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910007712.9A CN109670509A (en) | 2019-01-04 | 2019-01-04 | Winter wheat seedling stage growing way parameter evaluation method and system based on convolutional neural networks |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910007712.9A CN109670509A (en) | 2019-01-04 | 2019-01-04 | Winter wheat seedling stage growing way parameter evaluation method and system based on convolutional neural networks |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109670509A true CN109670509A (en) | 2019-04-23 |
Family
ID=66149245
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910007712.9A Pending CN109670509A (en) | 2019-01-04 | 2019-01-04 | Winter wheat seedling stage growing way parameter evaluation method and system based on convolutional neural networks |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109670509A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109975250A (en) * | 2019-04-24 | 2019-07-05 | 中国科学院遥感与数字地球研究所 | A kind of inversion method of leaf area index and device |
| CN111523781A (en) * | 2020-04-14 | 2020-08-11 | 四川大学 | Method for comprehensively evaluating growth vigor of fir regale under forest line environment |
| CN112131954A (en) * | 2020-08-26 | 2020-12-25 | 航天信德智图（北京）科技有限公司 | China northern winter wheat remote sensing estimated yield based on convolutional neural network |
| CN112149351A (en) * | 2020-09-22 | 2020-12-29 | 吉林大学 | Microwave circuit physical dimension estimation method based on deep learning |
| CN113537140A (en) * | 2021-08-03 | 2021-10-22 | 电子科技大学 | Weed detection method based on deep neural network |
| CN113643248A (en) * | 2021-08-04 | 2021-11-12 | 河南农业大学 | Wheat growth process monitoring method based on improved convolutional neural network |
| CN114708505A (en) * | 2022-04-08 | 2022-07-05 | 中国科学院合肥物质科学研究院 | Wheat yield estimation method based on unmanned aerial vehicle low-altitude visible light image and target detection network |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107516061A (en) * | 2016-06-17 | 2017-12-26 | 北京市商汤科技开发有限公司 | A kind of image classification method and system |
| CN107784305A (en) * | 2017-09-29 | 2018-03-09 | 中国农业科学院农业环境与可持续发展研究所 | Facilities vegetable disease recognition method and device based on convolutional neural networks |
-
2019
- 2019-01-04 CN CN201910007712.9A patent/CN109670509A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107516061A (en) * | 2016-06-17 | 2017-12-26 | 北京市商汤科技开发有限公司 | A kind of image classification method and system |
| CN107784305A (en) * | 2017-09-29 | 2018-03-09 | 中国农业科学院农业环境与可持续发展研究所 | Facilities vegetable disease recognition method and device based on convolutional neural networks |
Non-Patent Citations (4)
| Title |
|---|
| HAO LU等: "TasselNet: counting maize tassels in the wild via local counts regression network", 《PLANT METHODS》 * |
| 徐旭，陈国庆，王良，叶桂香，王振林，李勇: "基于敏感光谱波段图像特征的冬小麦LAI和地上部生物量监测", 《农业工程学报》 * |
| 陈怀亮，李颖，张红卫: "农作物长势遥感监测业务化应用与研究进展", 《气象与环境科学》 * |
| 高林，杨贵军，李红军，李振海，冯海宽，王磊，董锦绘，贺鹏: "基于无人机数码影像的冬小麦叶面积指数探测研究", 《中国生态农业学报》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109975250A (en) * | 2019-04-24 | 2019-07-05 | 中国科学院遥感与数字地球研究所 | A kind of inversion method of leaf area index and device |
| CN111523781A (en) * | 2020-04-14 | 2020-08-11 | 四川大学 | Method for comprehensively evaluating growth vigor of fir regale under forest line environment |
| CN111523781B (en) * | 2020-04-14 | 2023-04-25 | 四川大学 | Method for comprehensively evaluating growth vigor of Minjiang fir seedlings in forest environment |
| CN112131954A (en) * | 2020-08-26 | 2020-12-25 | 航天信德智图（北京）科技有限公司 | China northern winter wheat remote sensing estimated yield based on convolutional neural network |
| CN112149351A (en) * | 2020-09-22 | 2020-12-29 | 吉林大学 | Microwave circuit physical dimension estimation method based on deep learning |
| CN113537140A (en) * | 2021-08-03 | 2021-10-22 | 电子科技大学 | Weed detection method based on deep neural network |
| CN113537140B (en) * | 2021-08-03 | 2023-05-02 | 电子科技大学 | Weed detection method based on deep neural network |
| CN113643248A (en) * | 2021-08-04 | 2021-11-12 | 河南农业大学 | Wheat growth process monitoring method based on improved convolutional neural network |
| CN113643248B (en) * | 2021-08-04 | 2024-02-20 | 河南农业大学 | Wheat fertility process monitoring method based on improved convolutional neural network |
| CN114708505A (en) * | 2022-04-08 | 2022-07-05 | 中国科学院合肥物质科学研究院 | Wheat yield estimation method based on unmanned aerial vehicle low-altitude visible light image and target detection network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109670509A (en) | Winter wheat seedling stage growing way parameter evaluation method and system based on convolutional neural networks | |
| US10977494B2 (en) | Recognition of weed in a natural environment | |
| US11048940B2 (en) | Recognition of weed in a natural environment | |
| CN109359681B (en) | Field crop pest and disease identification method based on improved full convolution neural network | |
| CN106796716B (en) | For providing the device and method of super-resolution for low-resolution image | |
| CN107784305A (en) | Facilities vegetable disease recognition method and device based on convolutional neural networks | |
| CN109325495A (en) | A kind of crop image segmentation system and method based on deep neural network modeling | |
| CN109740721B (en) | Wheat ear counting method and device | |
| CN112464766B (en) | Automatic farmland land identification method and system | |
| CN107067043A (en) | A kind of diseases and pests of agronomic crop detection method | |
| CN109344738A (en) | The recognition methods of crop diseases and pest crop smothering and device | |
| EP3279831A1 (en) | Recognition of weed in a natural environment using a digital image | |
| CN109344699A (en) | Winter jujube disease recognition method based on depth of seam division convolutional neural networks | |
| CN112931150B (en) | Irrigation system and method based on spectral response of citrus canopy | |
| CN117036088A (en) | Data acquisition and analysis method for identifying growth situation of greening plants by AI | |
| CN110321774B (en) | Crop disaster situation evaluation method, device, equipment and computer readable storage medium | |
| Zhang et al. | Improving wheat yield estimates using data augmentation models and remotely sensed biophysical indices within deep neural networks in the Guanzhong Plain, PR China | |
| CN115661544B (en) | Spinach seedling water stress grade classification system and method based on N-MobileNetXt | |
| CN108629289A (en) | The recognition methods in farmland and system, applied to the unmanned plane of agricultural | |
| CN115687850A (en) | Method and device for calculating irrigation water demand of farmland | |
| CN117876823B (en) | Tea garden image detection method and model training method and system thereof | |
| CN110532935A (en) | A kind of high-throughput reciprocity monitoring system of field crop phenotypic information and monitoring method | |
| CN113989181A (en) | Remote sensing monitoring method and electronic equipment for growing period of crops in plastic greenhouse | |
| CN110555343B (en) | Method and system for extracting three elements of forest, shrub and grass in typical resource elements | |
| CN110111263A (en) | A kind of tobacco planting based on image procossing instructs system |
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: 20190423 |