CN109446987A - Method based on PCA and PNN algorithm detection rice pest grade - Google Patents
Method based on PCA and PNN algorithm detection rice pest grade Download PDFInfo
- Publication number
- CN109446987A CN109446987A CN201811266086.7A CN201811266086A CN109446987A CN 109446987 A CN109446987 A CN 109446987A CN 201811266086 A CN201811266086 A CN 201811266086A CN 109446987 A CN109446987 A CN 109446987A
- Authority
- CN
- China
- Prior art keywords
- grade
- pest
- sprinkling
- pca
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/188—Vegetation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Mining
-
- 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
-
- 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
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/194—Terrestrial scenes using hyperspectral data, i.e. more or other wavelengths than RGB
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Business, Economics & Management (AREA)
- Marine Sciences & Fisheries (AREA)
- Animal Husbandry (AREA)
- Agronomy & Crop Science (AREA)
- Mining & Mineral Resources (AREA)
- Multimedia (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Catching Or Destruction (AREA)
Abstract
This application discloses a kind of methods based on PCA and PNN algorithm detection rice pest grade, comprising steps of the original high-spectral data of acquisition rice, remove the abnormal point in original high-spectral data, wave band data dimensionality reduction and data compression are carried out to the original high-spectral data in adjacent band by principal component analysis PCA method, the judgement of pest and disease damage grade is carried out to sampled point according to training set using probabilistic neural network PNN method, sprinkling grade is carried out to sampled point using neural network sprinkling model to judge, pesticide is carried out according to sprinkling grade and is quantitatively sprayed.The automatic Prediction of pest and disease damage grade may be implemented in the present invention, and automatic identification pest and disease damage grade can substantially reduce cost of labor and time cost, provides foundation for the accurate sprinkling of pesticide, is of great significance for environmental protection and human health.
Description
Technical field
The present invention relates to nerual network techniques and EO-1 hyperion to detect pest and disease damage field, specifically, being to be related to EO-1 hyperion inspection
The method based on PCA and PNN algorithm detection rice pest grade in pest and disease damage is surveyed, to realize that pest and disease damage grade judges.
Background technique
Diseases and pests of agronomic crop is always an important factor for restricting agricultural production, to cause to the yield and quality of crop biggish
It influences.Estimate according to FAO (Food and Agriculture Organization of the United Nation) (FAO), world food yield is throughout the year because of pest and disease damage loss 10%, because disease is lost
14% or so.Rice prevents from having the important meaning as China's staple food crop, pest and disease damage.Principal component analysis
(principal components analysis, PCA) is that statistical nature basis is established in the smallest situation of root-mean-square error
On best orthogonal linear transformation, it is therefore an objective to multiple indexs are reduced to a kind of statistical method of a few composite target.
High-spectral data wave band can achieve several hundred or even thousands of, and information is increased simultaneously as there are very high adjacent to wave band
Correlation leads to a large amount of redundancy of high-spectral data.So that Data Dimensionality Reduction and data compression are carried out using PCA, so as to subsequent place
Reason.Probabilistic neural network (probabilistic neural network, PNN) was mentioned in 1989 by doctor Specht first
It out, is the new neural network of a kind of Radial Basis Function neural member and competition neurons common combination.Its, instruction simple with mode
Practice the features such as quick, it is very widely used, it is particularly suitable for solving classification problem.Probabilistic neural network generally has following four layers: defeated
Enter layer, mode layer, summation layer and output layer.Input layer is responsible for for feature vector being passed to network, and input layer number is sample characteristics
Number.Mode layer is connect by connection weight with input layer.Calculate of each mode in input feature value and training set
With degree, that is, similarity, its distance is sent into Gaussian function and obtains the output of mode layer.The number of the neuron of mode layer
It is the number of input sample vector, that is, how many sample, the layer is with regard to how many neuron.Summation layer, is just responsible for
The pattern layer units of each class are connected, the neuron number of this layer is the class number of sample.If output layer, just
It is responsible for that one kind of highest scoring in output summation layer.
Summary of the invention
The technical problem to be solved by the present invention is to carry out the prediction of pest and disease damage severity based on high-spectral data, provide
A method of rice pest grade being detected based on PCA and PNN algorithm, to carry out the essence of pesticide in practical plant protection operation
Quasi- sprinkling.
Rice pest grade is detected based on PCA and PNN algorithm in order to solve the above technical problems, the present invention provides one kind
Method, which is characterized in that comprising steps of
The original high-spectral data of rice is acquired in sampled point by the bloom spectrometer of UAV flight；
The abnormal point in original high-spectral data is removed, the terrestrial object information of 1024 wave bands is obtained；
Wave band data dimensionality reduction and data are carried out to the original high-spectral data in adjacent band by principal component analysis PCA method
Compression, the spectral value after obtaining dimensionality reduction, the corresponding image of each of them spectral value；
The judgement of pest and disease damage grade is carried out to sampled point according to training set using probabilistic neural network PNN method, comprising: will be described
Each image corresponding to each spectral value after dimensionality reduction is as training set, and the training set is true value library, according to described
Image judges true value of the pest and disease damage grade as each training spectrum, carries out neural network model training, obtains neural network
Spray model；
Feature learning is carried out to the spectral value after the dimensionality reduction, sampled point is carried out using neural network sprinkling model
Spray grade judgement；
It carries out pesticide according to sprinkling grade quantitatively to spray, quantitative sprinkling calculates in accordance with the following methods:
Q=Max × Y,
Wherein, as 1≤X≤4, Y=0.25 × (X-1)；As X=5, Y=1；
Q is fountain height, and X is sprinkling grade, and Y is sprinkling ratio.
Preferably, the principal component analysis PCA method comprising steps of
Centralization is carried out to all original high-spectral datas, that is, sample data, goes mean value；
Solve the covariance matrix of sample；
Eigenvalues Decomposition is carried out to covariance, including the use of Eigenvalues Decomposition or singular value decomposition, solve characteristic value and
Feature vector；
The corresponding n feature vector of maximum n characteristic value is taken out, after feature vector is standardized, forms projection matrix；
Using projection matrix, the data of dimensionality reduction are obtained.
Preferably, the pest and disease damage grade has 5 grades altogether, and 1 grade is health, and 2-5 grades of pest and disease damage severity gradually increase.
Preferably, the spectroscopic data that the unmanned plane is obtained in takeoff and landing is abnormal point.
Preferably, described that true value of the pest and disease damage grade as each training spectrum is judged according to described image, carry out mind
Through network model training, neural network sprinkling model is obtained, further for unmanned plane is by bloom spectrometer in sample farmland
The growing state of crops is monitored, and the high-spectral data of acquisition is carried out as sample high-spectral data by high-definition image
It identifies, after the growth information for recognizing crop, the big data and agronomy expert appraisal obtained is tested according to pest and disease damage, based on every
The growth information for opening the crop that high-definition image is reacted, obtains sprinkling grade corresponding to every high-definition image, wherein sprinkling etc.
Grade include pesticide spraying grade and fertilizer spray grade, obtain high-definition image and spray grade corresponding relationship after, with
Sample high-spectral data corresponding to high-definition image is as the input vector in training set, to spray grade as in training set
Output vector carries out sample training to neural network model, constructs neural network sprinkling model.
Preferably, the sprinkling grade is divided into 5 grades, and 1 to 5 grade more high required dose or fertilizer are higher.
Currently, rice pest application is also rested on and is rule of thumb administered, the excessive use of pesticide generally can be all caused, is endangered
Environmental and human health impacts are done harm to.The present invention predicts pest and disease damage severity by the training of PCA and PNN algorithm and is administered on demand, power
It strives and controls formulation rate to best.
Compared with prior art, the method for the present invention based on PCA and PNN algorithm detection rice pest grade,
Reach following effect:
The automatic Prediction of pest and disease damage grade may be implemented by the training of given data collection in the present invention, by PCA by wave band
PNN training speed can be improved in dimensionality reduction.Automatic identification pest and disease damage grade of the invention can substantially reduce cost of labor and time
Cost provides foundation for the accurate sprinkling of pesticide, is of great significance for environmental protection and human health.
Detailed description of the invention
The drawings described herein are used to provide a further understanding of the present application, constitutes part of this application, this Shen
Illustrative embodiments and their description please are not constituted an undue limitation on the present application for explaining the application.In the accompanying drawings:
Fig. 1 is the flow chart for detecting the method for rice pest grade in the embodiment of the present invention 1 based on PCA and PNN algorithm；
Fig. 2 is the PCA projection for detecting the method for rice pest grade in the embodiment of the present invention 3 based on PCA and PNN algorithm
Matrix diagram；
Fig. 3 is the two dimensional PCA for detecting the method for rice pest grade in the embodiment of the present invention 3 based on PCA and PNN algorithm
Projection theory figure；
Fig. 4 is the PNN principle for detecting the method for rice pest grade in the embodiment of the present invention 3 based on PCA and PNN algorithm
Figure.
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 should be noted that described embodiment only actually is a part of the embodiment of the present invention, rather than whole realities
Example is applied, and is actually merely illustrative, never as to the present invention and its application or any restrictions used.The guarantor of the application
Protect range as defined by the appended claims.
Below in conjunction with attached drawing, invention is further described in detail, but not as a limitation of the invention.
Embodiment 1:
Shown in Figure 1 is the specific of the herein described method that rice pest grade is detected based on PCA and PNN algorithm
Embodiment, this method comprises:
Step 101, the original high-spectral data for acquiring rice in sampled point by the bloom spectrometer of UAV flight.
Abnormal point in step 102, the original high-spectral data of removal, obtains the terrestrial object information of 1024 wave bands.
Step 103 carries out wave band data drop to the original high-spectral data in adjacent band by principal component analysis PCA method
Peacekeeping data compression, the spectral value after obtaining dimensionality reduction, the corresponding image of each of them spectral value.
Step 104 carries out the judgement of pest and disease damage grade, packet to sampled point according to training set using probabilistic neural network PNN method
Include: using each image corresponding to each spectral value after the dimensionality reduction as training set, the training set is true value library,
True value of the pest and disease damage grade as each training spectrum is judged according to described image, is carried out neural network model training, is obtained
Neural network sprays model.
Step 105 carries out feature learning to the spectral value after the dimensionality reduction, using neural network sprinkling model to adopting
Sampling point carries out sprinkling grade judgement.
Step 106 is quantitatively sprayed according to sprinkling grade progress pesticide, and quantitative sprinkling calculates in accordance with the following methods:
Q=Max × Y,
Wherein, as 1≤X≤4, Y=0.25 × (X-1)；As X=5, Y=1；
Q is fountain height, and X is sprinkling grade, and Y is sprinkling ratio.
Embodiment 2:
The method based on PCA and PNN algorithm detection rice pest grade is present embodiments provided, this method comprises:
Step 201, the original high-spectral data for acquiring rice in sampled point by the bloom spectrometer of UAV flight.
Abnormal point in step 202, the original high-spectral data of removal, obtains the terrestrial object information of 1024 wave bands.
Step 203 carries out wave band data drop to the original high-spectral data in adjacent band by principal component analysis PCA method
Peacekeeping data compression, the spectral value after obtaining dimensionality reduction, the corresponding image of each of them spectral value.
Step 204 carries out the judgement of pest and disease damage grade, packet to sampled point according to training set using probabilistic neural network PNN method
Include: using each image corresponding to each spectral value after the dimensionality reduction as training set, the training set is true value library,
True value of the pest and disease damage grade as each training spectrum is judged according to described image, is carried out neural network model training, is obtained
Neural network sprays model.
Step 205 carries out feature learning to the spectral value after the dimensionality reduction, using neural network sprinkling model to adopting
Sampling point carries out sprinkling grade judgement.
Step 206 is quantitatively sprayed according to sprinkling grade progress pesticide, and quantitative sprinkling calculates in accordance with the following methods:
Q=Max × Y,
Wherein, as 1≤X≤4, Y=0.25 × (X-1)；As X=5, Y=1；
Q is fountain height, and X is sprinkling grade, and Y is sprinkling ratio.
In above-mentioned steps 203, the principal component analysis PCA method comprising steps of
Centralization is carried out to all original high-spectral datas, that is, sample data, goes mean value；
Solve the covariance matrix of sample；
Eigenvalues Decomposition is carried out to covariance, including the use of Eigenvalues Decomposition or singular value decomposition, solve characteristic value and
Feature vector；
The corresponding n feature vector of maximum n characteristic value is taken out, after feature vector is standardized, forms projection matrix；
Using projection matrix, the data of dimensionality reduction are obtained.
In above-mentioned steps 204, the pest and disease damage grade has 5 grades altogether, and 1 grade is health, and 2-5 grades of pest and disease damage severity are gradually
Increase.
In above-mentioned steps 202, the spectroscopic data that the unmanned plane is obtained in takeoff and landing is abnormal point.
It is described that true value of the pest and disease damage grade as each training spectrum is judged according to described image in above-mentioned steps 204,
Neural network model training is carried out, neural network sprinkling model is obtained；The specific steps are unmanned plane is by bloom spectrometer to sample
The growing state of crops in farmland is monitored, and the high-spectral data of acquisition passes through high definition as sample high-spectral data
Image is identified, after the growth information for recognizing crop, the big data and agronomy expert obtained according to pest and disease damage experiment is reflected
Fixed, the growth information based on the crop that every high-definition image is reacted obtains sprinkling grade corresponding to every high-definition image,
In, sprinkling grade includes pesticide spraying grade and fertilizer spray grade, in the corresponding pass for obtaining high-definition image with spraying grade
After system, using with sample high-spectral data corresponding to high-definition image as the input vector in training set, using spray grade as
Output vector in training set carries out sample training to neural network model, constructs neural network sprinkling model.
In above-mentioned steps 206, the sprinkling grade is divided into 5 grades, and 1 to 5 grade more high required dose or fertilizer are higher.
Embodiment 3:
Another embodiment of the present invention provides the methods based on PCA and PNN algorithm detection rice pest grade, should
Method includes:
Step 301, the original high-spectral data for acquiring rice in sampled point by the bloom spectrometer of UAV flight.
Abnormal point in step 302, the original high-spectral data of removal, obtains the terrestrial object information of 1024 wave bands.
Step 303 carries out wave band data drop to the original high-spectral data in adjacent band by principal component analysis PCA method
Peacekeeping data compression, the spectral value after obtaining dimensionality reduction, the corresponding image of each of them spectral value, specific PCA algorithm steps
Are as follows:
(1) centralization is carried out to all original high-spectral datas, that is, sample data, goes mean value, and carry out normalizing as needed
Change processing；
(2) covariance matrix of sample is solved；
(3) Eigenvalues Decomposition is carried out to covariance, including the use of Eigenvalues Decomposition or singular value decomposition, solve characteristic value with
And feature vector；
(4) the corresponding n feature vector of maximum n characteristic value is taken out, after feature vector is standardized, composition projection square
Battle array, referring to fig. 2；
(5) projection matrix is utilized, obtains the data of dimensionality reduction,
It wherein, is to find a vector u for the reduction process of 2D to 1D1, u1It indicates a direction, then will own
Original sample is mapped on this direction, and a vector indicates a sample space, obtains a new samples in new dimensional space
Dimensionality reduction data in space；
For the reduction process of 3D to 2D, that is, find two vector u1And u2, (u1, u2), the two vectors define one
Then the sample in original sample space is mapped in this new sample space data after obtaining dimensionality reduction by new feature space, ginseng
See Fig. 3；
For the reduction process of nD to kD, that is, find k vector, u1,u2,…,uk, this k vector define one it is new
Then the sample in original sample space is mapped in this new sample space data after obtaining dimensionality reduction by vector space.
Step 304, referring to fig. 4 carries out pest and disease damage grade to sampled point according to training set using probabilistic neural network PNN method
Judgement, comprising: be using each image corresponding to each spectral value after the dimensionality reduction as training set, the training set
True value library is the premise of training pattern, judges true value of the pest and disease damage grade as each training spectrum according to described image, into
The training of row neural network model obtains neural network sprinkling model.
Step 305 carries out feature learning to the spectral value after the dimensionality reduction, is input with sample high-spectral data, utilizes
The neural network model calculates sprinkling grade, and sprinkling grade is divided into 5 grades, 1 to 5 grade more high required dose or fertilizer
It is higher.Specific formula for calculation:
Wherein, lgIndicate the quantity of g class；N indicates the number (being n=5 in this embodiment) of feature；σ is smoothing parameter, model
It encloses for 0-1；xijIndicate j-th of data of i-th of neuron of g class；xjIndicate j-th of characteristic dimension；ygIndicate classification.
Step 306 is quantitatively sprayed according to sprinkling grade progress pesticide, and quantitative sprinkling calculates in accordance with the following methods:
Q=Max × Y,
Wherein, as 1≤X≤4, Y=0.25 × (X-1)；As X=5, Y=1；
Q is fountain height, and X is sprinkling grade, and Y is sprinkling ratio.
Wherein, the pest and disease damage grade has 5 grades altogether, and 1 grade is health, and 2-5 grades of pest and disease damage severity gradually increase.
Compared with prior art, the method for the present invention based on PCA and PNN algorithm detection rice pest grade,
Reach following effect:
The automatic Prediction of pest and disease damage grade may be implemented by the training of given data collection in the present invention, by PCA by wave band
PNN training speed can be improved in dimensionality reduction.Automatic identification pest and disease damage grade of the invention can substantially reduce cost of labor and time
Cost provides foundation for the accurate sprinkling of pesticide, is of great significance for environmental protection and human health.
Although some specific embodiments of the invention are described in detail by example, the skill of this field
Art personnel it should be understood that example above merely to being illustrated, the range being not intended to be limiting of the invention.Although referring to before
Stating embodiment, invention is explained in detail, for those skilled in the art, still can be to aforementioned reality
Technical solution documented by example is applied to modify or equivalent replacement of some of the technical features.It is all of the invention
Within spirit and principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.
The scope of the present invention is defined by the appended claims.
Claims (6)
1. a kind of method based on PCA and PNN algorithm detection rice pest grade, which is characterized in that comprising steps of
The original high-spectral data of rice is acquired in sampled point by the bloom spectrometer of UAV flight；
The abnormal point in original high-spectral data is removed, the terrestrial object information of 1024 wave bands is obtained；
Wave band data dimensionality reduction and data pressure are carried out to the original high-spectral data in adjacent band by principal component analysis PCA method
Contracting, the spectral value after obtaining dimensionality reduction, the corresponding image of each of them spectral value；
The judgement of pest and disease damage grade is carried out to sampled point according to training set using probabilistic neural network PNN method, comprising: by the dimensionality reduction
Each image corresponding to each spectral value afterwards is as training set, and the training set is true value library, according to described image
Judge true value of the pest and disease damage grade as each training spectrum, carry out neural network model training, obtains neural network sprinkling
Model；
Feature learning is carried out to the spectral value after the dimensionality reduction, sampled point is sprayed using neural network sprinkling model
Grade judgement；
It carries out pesticide according to sprinkling grade quantitatively to spray, quantitative sprinkling calculates in accordance with the following methods:
Q=Max × Y,
Wherein, as 1≤X≤4, Y=0.25 × (X-1)；As X=5, Y=1；
Q is fountain height, and X is sprinkling grade, and Y is sprinkling ratio.
2. the method according to claim 1 based on PCA and PNN algorithm detection rice pest grade, which is characterized in that
The principal component analysis PCA method comprising steps of
Centralization is carried out to all original high-spectral datas, that is, sample data, goes mean value；
Solve the covariance matrix of sample；
Eigenvalues Decomposition is carried out to covariance, including the use of Eigenvalues Decomposition or singular value decomposition, solves characteristic value and feature
Vector；
The corresponding n feature vector of maximum n characteristic value is taken out, after feature vector is standardized, forms projection matrix；
Using projection matrix, the data of dimensionality reduction are obtained.
3. the method according to claim 1 based on PCA and PNN algorithm detection rice pest grade, which is characterized in that
The pest and disease damage grade has 5 grades altogether, and 1 grade is health, and 2-5 grades of pest and disease damage severity gradually increase.
4. the method according to claim 1 based on PCA and PNN algorithm detection rice pest grade, which is characterized in that
The spectroscopic data that the unmanned plane is obtained in takeoff and landing is abnormal point.
5. the method according to claim 1 based on PCA and PNN algorithm detection rice pest grade, which is characterized in that
It is described that true value of the pest and disease damage grade as each training spectrum is judged according to described image, neural network model training is carried out,
Neural network sprinkling model is obtained, further for unmanned plane is by bloom spectrometer to the growth feelings of the crops in sample farmland
Condition is monitored, and the high-spectral data of acquisition is identified by high-definition image as sample high-spectral data, recognizes crop
Growth information after, the big data and agronomy expert appraisal that obtain are tested according to pest and disease damage, it is anti-based on every high-definition image institute
The growth information of the crop answered obtains sprinkling grade corresponding to every high-definition image, wherein sprinkling grade includes pesticide spraying
Grade and fertilizer spray grade, obtain high-definition image and spray grade corresponding relationship after, with corresponding to high-definition image
Sample high-spectral data as the input vector in training set, to spray grade as the output vector in training set, to mind
Sample training is carried out through network model, constructs neural network sprinkling model.
6. the method according to claim 1 or 5 based on PCA and PNN algorithm detection rice pest grade, feature exist
In the sprinkling grade is divided into 5 grades, and 1 to 5 grade more high required dose or fertilizer are higher.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811266086.7A CN109446987A (en) | 2018-10-29 | 2018-10-29 | Method based on PCA and PNN algorithm detection rice pest grade |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811266086.7A CN109446987A (en) | 2018-10-29 | 2018-10-29 | Method based on PCA and PNN algorithm detection rice pest grade |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109446987A true CN109446987A (en) | 2019-03-08 |
Family
ID=65549013
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811266086.7A Pending CN109446987A (en) | 2018-10-29 | 2018-10-29 | Method based on PCA and PNN algorithm detection rice pest grade |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109446987A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110956132A (en) * | 2019-11-29 | 2020-04-03 | 苏州经贸职业技术学院 | Method for intelligently constructing weed control model based on unmanned aerial vehicle cooperation |
| CN111027380A (en) * | 2019-11-05 | 2020-04-17 | 北京海益同展信息科技有限公司 | Spray head control method and device, computer equipment and storage medium |
| CN111427377A (en) * | 2020-04-08 | 2020-07-17 | 海南耐威科技系统技术研究院有限公司 | Special flight control system for plant protection unmanned aerial vehicle |
| CN113255894A (en) * | 2021-06-02 | 2021-08-13 | 华南农业大学 | Training method of BP neural network model, pest and disease damage detection method and electronic equipment |
| CN113298289A (en) * | 2021-04-14 | 2021-08-24 | 北京市燃气集团有限责任公司 | Method and device for predicting gas consumption of gas user |
| CN113326841A (en) * | 2021-05-31 | 2021-08-31 | 山东深蓝智谱数字科技有限公司 | Hyperspectrum-based crop pest monitoring method and device |
| CN113468964A (en) * | 2021-05-31 | 2021-10-01 | 山东省邮电工程有限公司 | Hyperspectrum-based agricultural disease and pest monitoring method and device |
| CN113836999A (en) * | 2021-08-16 | 2021-12-24 | 山东大学 | Tunnel construction risk intelligent identification method and system based on ground penetrating radar |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106023065A (en) * | 2016-05-13 | 2016-10-12 | 中国矿业大学 | Tensor hyperspectral image spectrum-space dimensionality reduction method based on deep convolutional neural network |
| CN108684282A (en) * | 2018-04-11 | 2018-10-23 | 北京麦飞科技有限公司 | A kind of agricultural, which is examined, beats integrated machine system and sprinkling control method |
| CN108693119A (en) * | 2018-04-20 | 2018-10-23 | 北京麦飞科技有限公司 | Pest and disease damage based on unmanned plane high-spectrum remote-sensing intelligently examines the system of beating |
-
2018
- 2018-10-29 CN CN201811266086.7A patent/CN109446987A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106023065A (en) * | 2016-05-13 | 2016-10-12 | 中国矿业大学 | Tensor hyperspectral image spectrum-space dimensionality reduction method based on deep convolutional neural network |
| CN108684282A (en) * | 2018-04-11 | 2018-10-23 | 北京麦飞科技有限公司 | A kind of agricultural, which is examined, beats integrated machine system and sprinkling control method |
| CN108693119A (en) * | 2018-04-20 | 2018-10-23 | 北京麦飞科技有限公司 | Pest and disease damage based on unmanned plane high-spectrum remote-sensing intelligently examines the system of beating |
Non-Patent Citations (4)
| Title |
|---|
| JASTER_WISDOM: "概率神经网络(PNN)", 《HTTPS://BLOG.CSDN.NET/JASTER_WISDOM/ARTICLE/DETAILS/69558088?READLOG》 * |
| 乔园园，张明涛: "《简明计算机化学教程》", 31 August 2005, 天津：南开大学出版社 * |
| 是一地和的店: "概率神经网络讲解", 《HTTPS://WENKU.BAIDU.COM/VIEW/9E26773C777F5ACFA1C7AA00B52ACFC789EB9FF2.HTML》 * |
| 陈雯柏: "《人工神经网络原理与实践》", 31 January 2016 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111027380A (en) * | 2019-11-05 | 2020-04-17 | 北京海益同展信息科技有限公司 | Spray head control method and device, computer equipment and storage medium |
| CN110956132A (en) * | 2019-11-29 | 2020-04-03 | 苏州经贸职业技术学院 | Method for intelligently constructing weed control model based on unmanned aerial vehicle cooperation |
| CN110956132B (en) * | 2019-11-29 | 2023-12-29 | 苏州经贸职业技术学院 | Method for constructing weed control model based on unmanned aerial vehicle cooperation intelligence |
| CN111427377A (en) * | 2020-04-08 | 2020-07-17 | 海南耐威科技系统技术研究院有限公司 | Special flight control system for plant protection unmanned aerial vehicle |
| CN113298289A (en) * | 2021-04-14 | 2021-08-24 | 北京市燃气集团有限责任公司 | Method and device for predicting gas consumption of gas user |
| CN113326841A (en) * | 2021-05-31 | 2021-08-31 | 山东深蓝智谱数字科技有限公司 | Hyperspectrum-based crop pest monitoring method and device |
| CN113468964A (en) * | 2021-05-31 | 2021-10-01 | 山东省邮电工程有限公司 | Hyperspectrum-based agricultural disease and pest monitoring method and device |
| CN113255894A (en) * | 2021-06-02 | 2021-08-13 | 华南农业大学 | Training method of BP neural network model, pest and disease damage detection method and electronic equipment |
| CN113255894B (en) * | 2021-06-02 | 2022-12-06 | 华南农业大学 | Training method of BP neural network model, pest and disease damage detection method and electronic equipment |
| CN113836999A (en) * | 2021-08-16 | 2021-12-24 | 山东大学 | Tunnel construction risk intelligent identification method and system based on ground penetrating radar |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109446987A (en) | Method based on PCA and PNN algorithm detection rice pest grade | |
| Liu et al. | Weed detection for selective spraying: a review | |
| Ampatzidis et al. | Citrus rootstock evaluation utilizing UAV-based remote sensing and artificial intelligence | |
| Oerke | Remote sensing of diseases | |
| Jhuria et al. | Image processing for smart farming: Detection of disease and fruit grading | |
| Hong et al. | Detection system of smart sprayers: Status, challenges, and perspectives | |
| Kukreja et al. | Automatic classification of wheat rust diseases using deep convolutional neural networks | |
| Pádua et al. | Very high resolution aerial data to support multi-temporal precision agriculture information management | |
| Chandel et al. | Apple powdery mildew infestation detection and mapping using high-resolution visible and multispectral aerial imaging technique | |
| Walter et al. | Flourish-a robotic approach for automation in crop management | |
| Zhang et al. | Opportunities of UAVs in orchard management | |
| Muppala et al. | Machine vision detection of pests, diseases and weeds: A review | |
| Kumar et al. | Automatic leaf disease detection and classification using hybrid features and supervised classifier | |
| Poornam et al. | Image based Plant leaf disease detection using Deep learning | |
| CN108846370B (en) | Method for automatically analyzing severity of disease in middle and late stages of wheat powdery mildew | |
| Zhou et al. | UAV remote sensing: An innovative tool for detection and management of rice diseases | |
| Hruška et al. | Deep learning-based methodological approach for vineyard early disease detection using hyperspectral data | |
| Moshou et al. | Multisensor fusion of remote sensing data for crop disease detection | |
| CN115761542A (en) | Winter wheat ear stage pest and disease degree evaluation method based on unmanned aerial vehicle remote sensing image | |
| Sankararao et al. | Water stress detection in pearl millet canopy with selected wavebands using UAV based hyperspectral imaging and machine learning | |
| Alkhasawneh | Olive oil ripping time prediction model based on image processing and neural network | |
| Widiyanto et al. | Monitoring the growth of tomatoes in real time with deep learning-based image segmentation | |
| Li et al. | Determination of Wheat Heading Stage Using Convolutional Neural Networks on Multispectral UAV Imaging Data | |
| Sharma et al. | Predict growth stages of wheat crop using digital image processing | |
| Lopes et al. | Plantplotgan: A physics-informed generative adversarial network for plant disease prediction |
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: 20190308 |