CN113111950B - Wheat rust classification method based on ensemble learning - Google Patents
Wheat rust classification method based on ensemble learning Download PDFInfo
- Publication number
- CN113111950B CN113111950B CN202110416254.1A CN202110416254A CN113111950B CN 113111950 B CN113111950 B CN 113111950B CN 202110416254 A CN202110416254 A CN 202110416254A CN 113111950 B CN113111950 B CN 113111950B
- Authority
- CN
- China
- Prior art keywords
- model
- rust
- wheat
- training
- optimal
- 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.)
- Active
Links
- JEIPFZHSYJVQDO-UHFFFAOYSA-N iron(III) oxide Inorganic materials O=[Fe]O[Fe]=O JEIPFZHSYJVQDO-UHFFFAOYSA-N 0.000 title claims abstract description 73
- 238000000034 method Methods 0.000 title claims abstract description 61
- 241000209140 Triticum Species 0.000 title claims abstract description 54
- 235000021307 Triticum Nutrition 0.000 title claims abstract description 54
- 238000012549 training Methods 0.000 claims abstract description 52
- 230000010354 integration Effects 0.000 claims abstract description 24
- 230000004927 fusion Effects 0.000 claims abstract description 13
- 238000005516 engineering process Methods 0.000 claims abstract description 5
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 238000012795 verification Methods 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 238000013527 convolutional neural network Methods 0.000 abstract description 4
- 201000010099 disease Diseases 0.000 description 22
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 22
- 241000221535 Pucciniales Species 0.000 description 5
- 238000011160 research Methods 0.000 description 5
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000001575 pathological effect Effects 0.000 description 2
- 235000013339 cereals Nutrition 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
- 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
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A40/00—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production
- Y02A40/10—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production in agriculture
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Business, Economics & Management (AREA)
- Marine Sciences & Fisheries (AREA)
- Animal Husbandry (AREA)
- Mining & Mineral Resources (AREA)
- Agronomy & Crop Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a wheat rust classification method based on ensemble learning, which comprises the following steps of 1: collecting wheat rust images with a plurality of devices after the wheat turning green period; step 2: preprocessing the collected images, and performing sample expansion by using a data enhancement technology; and step 3: constructing an integrated model, integrating various convolutional neural networks with different structures, and training the model by using a snapshot integration method; and 4, step 4: selecting an optimal model and model parameters according to the loss and the used fusion algorithm; and 5: and using the optimal model for the identification and classification of the wheat rust image. The wheat rust images can be classified by combining the advantages of different convolutional neural network architectures, the best model can be selected more quickly by utilizing the bagging integration algorithm, the snapshot integration method and the fusion algorithm, the accuracy of wheat rust classification is improved, and the problem that leaf rust, stem rust and stripe rust are difficult to distinguish is solved.
Description
Technical Field
The invention belongs to the field of crop disease classification, and particularly relates to a wheat rust classification method based on ensemble learning.
Background
Wheat rust is a disease with extremely strong harmfulness, can reduce the total wheat yield, influence the livelihood of farmers, threaten the grain safety of the whole country under severe conditions, and is difficult to monitor, control and eradicate the wheat rust on a large scale. The three wheat rusts, namely stem rust, leaf rust and stripe rust, have similar disease characteristics, the disease positions are crossed, and people in more professionals are needed to distinguish the wheat rusts from each other, but much time and labor are consumed, so that the three wheat rusts are classified in a time-saving and labor-saving mode, the identification accuracy can be guaranteed, and then the wheat rusts are treated with medicines according to the diseases, so that the yield loss is avoided. With the continuous development of machine learning and deep learning, more and more researchers in agriculture adopt the method for research, but the diseases with higher similarity cannot be well distinguished.
Aiming at the problems, the research integrates various mainstream convolutional neural network models by utilizing ensemble learning instead of relying on one network model, trains the integrated model by using a snapshot integration mode, and screens the optimal model by using a fusion algorithm twice. Finally, the method can better classify the three rust diseases of wheat. At present, few scholars study the classification problem of similar diseases in crops, and no method for specially and quantitatively studying three rust classifications of wheat exists.
Disclosure of Invention
In order to overcome the defects of the technology, the invention provides a wheat rust classification method based on ensemble learning.
In order to solve the technical problems, the invention adopts the technical scheme that:
a wheat rust classification method based on ensemble learning is characterized by comprising the following steps:
step 1: after the wheat is in the green turning stage, leaf rust, stem rust and stripe rust appear in succession, the onset characteristics of the three diseases are similar, and images of the wheat rust are collected by various devices, such as: unmanned aerial vehicles, cameras, mobile phones, and the like;
step 2: preprocessing an image, and specifically comprises the following steps:
step 201: making labels on the collected images, and dividing the labels into three categories of leaf rust wheat, stem rust wheat and stripe rust wheat;
step 202: sample expansion is carried out on the three types of images by utilizing various data enhancement technologies, the data volume required by training is increased, and the generalization capability and robustness of the model are improved;
and step 3: constructing an integrated model, and training the model by using a snapshot integration method, wherein the method comprises the following specific steps:
step 301: selecting four models of VGG, ResNet, DenseNet and inclusion V3 with different convolutional layer connection modes for integration;
step 302: the four models for integration are placed in a model library, a bagging integration mode is adopted, one model is randomly selected for training by utilizing hyper-parameter optimization search during each training, and enough iteration times N are set to ensure that each model is selected. The formula of the Bagging integration algorithm is as follows:
wherein, yn(x) Representing the output of the model, this integration is a random sampling with a put back;
step 303: the method comprises the steps of training a model by adopting a snapshot integration mode, storing the model and parameters of each point reaching a local optimal point, and selecting the optimal model in the current training process through the loss of a verification set after one training is finished;
and 4, step 4: selecting the optimal model by adopting a fusion algorithm according to the loss in the training process, and specifically comprising the following steps of:
step 401: the loss magnitude is calculated using a weighted cross entropy loss function, as follows:
wherein N represents the number of categories, here 3, representing respectively stem rust, leaf rust and stripe rust; p is a radical ofiRepresenting the probability of the sample belonging to the ith class; y ═ y0,…,yN-1]Is the onehot representation of the sample label, when the sample belongs to category i, yi1, the rest is 0; w ═ W0,…,wN-1]A weight representing the category i;
step 402: after the optimal model in the current training process is selected after one training, a fusion algorithm is generated by utilizing the 'replacement' idea, and the results of all models obtained by the current training are processed. Multiplying the verification set loss of the optimal model in the current training process with the verification set losses of the other models in the current training process, and then performing evolution, if the obtained loss value is smaller, taking the smaller value as an optimal value, and obtaining the corresponding model weight through the same calculation, and taking the model as a new optimal model in the current training process;
step 403: after the iteration times reach the upper limit, the optimal model obtained by each iteration is fused and screened again by using the fusion algorithm to obtain the optimal model in the whole training process;
and 5: and using the obtained optimal model for identifying and classifying the wheat rust images.
The invention has the following beneficial effects: the three rust diseases of the wheat can be classified from the wheat rust images obtained by different equipment, so that the different rust diseases of the wheat can be prevented and controlled in time, the yield reduction is avoided, and the technical support is provided for further pathological research on the different rust diseases of the wheat.
Drawings
FIG. 1 is a flow chart of an embodiment of the present invention.
FIG. 2 is a comparison of classification results of different methods according to an embodiment of the present invention.
FIG. 3 is a partial classification effect display according to an embodiment of the present invention; (a) stem rust disease; (b) leaf rust disease; (c) stripe rust.
Detailed Description
The present invention will be described in detail with reference to specific examples.
Referring to fig. 1, the invention is a flow chart of a wheat rust classification method based on ensemble learning, and the method comprises the following steps:
step 1: the classification process of wheat rust is shown in fig. 1, firstly, after the wheat is in the green turning stage, images of leaf rust, stem rust and stripe rust begin to be collected, the onset characteristics of the three diseases are similar, and the images of wheat rust are collected by paying attention to distinguishing and supporting the collection of the images of wheat rust by using various devices, such as: unmanned aerial vehicles, cameras, mobile phones, and the like;
step 2: the image is preprocessed in order to increase the data amount required by training, and even if the number of images is too small due to the problems of insufficient professionals and the like when the images are collected, the training amount is not necessary to be worried about, and the image preprocessing method comprises the following specific steps:
step 201: making labels on the collected images, and dividing the labels into three categories of leaf rust wheat, stem rust wheat and stripe rust wheat;
step 202: sample expansion is carried out on the three types of images by sequentially utilizing five data enhancement technologies of random cutting, horizontal turning, vertical turning, contrast enhancement and aliasing enhancement, the data quantity required by training is increased, and the generalization capability and robustness of the model are improved;
and 3, step 3: the method comprises the following steps of constructing an integrated model, and training the model by using a snapshot integration method, wherein the method can obtain a plurality of models reaching local optimal points in one training process, and the local optimal points contain useful information, and the method comprises the following specific steps:
step 301: selecting four models of VGG, ResNet, DenseNet and inclusion V3 with different convolutional layer connection modes for integration;
step 302: putting four models for integration into a model library, adopting a bagging integration mode, randomly selecting one model for training by utilizing hyper-parameter optimization search during each training, and setting enough iteration times N to ensure that each model is selected; the Bagging integration algorithm is as follows:
wherein, yn(x) Representing the output of the model, and the integration mode is a random sampling with a replace;
step 303: the model training is carried out by adopting a snapshot integration mode, the method saves the model and parameters of each point reaching the local optimum point, and selects the optimum model in the current training process through the loss of a verification set after one training is finished, FIG. 2 shows the accuracy of wheat rust classification by using four single models, namely VGG, ResNet, DenseNet and inclusion V3 in the existing research and the classification accuracy of the method, and the classification accuracy of the method reaches 96 percent as can be seen from FIG. 2 and Table 1, thereby showing the effectiveness of the method;
TABLE 1
And 4, step 4: selecting an optimal model by adopting a fusion algorithm according to the loss in the training process, wherein the fusion algorithm is used twice, one time is used for screening the optimal model in the current training process in each training process, and the other time is used for selecting the optimal model in the whole training process after the number of arrival iterations ensures that all convolutional neural network models are randomly selected, and the method comprises the following specific steps:
step 401: the loss magnitude is calculated using a weighted cross entropy loss function, as follows:
wherein N represents the number of categories, here 3, representing respectively stem rust, leaf rust and stripe rust; pi represents the probability that the sample belongs to class i; y ═ y0,…,yN-1]Is the onehot representation of the sample label, when the sample belongs to category i, yi1, the rest is 0; w ═ W0,…,wN-1]A weight representing a category i;
step 402: after the optimal model in the current training process is selected after one training, a fusion algorithm is generated by utilizing the 'replacement' idea, and the results of all models obtained by the current training are processed. Multiplying the verification set loss of the optimal model in the current training process with the verification set losses of the other models in the current training process, and then performing evolution, if the obtained loss value is smaller, taking the smaller value as an optimal value, and obtaining the corresponding model weight through the same calculation, and taking the model as a new optimal model in the current training process;
step 403: after the iteration times reach the upper limit, the optimal model obtained by each iteration is fused and screened again by using the fusion algorithm to obtain the optimal model in the whole training process;
and 5: and using the obtained optimal model for identifying and classifying the wheat rust images. After the optimal model is provided, the wheat images containing stem rust, leaf rust and stripe rust can be identified in any wheat image, and partial classification effects of three types of rust wheat are shown in fig. 3, which shows that the method can better extract disease characteristics of different rust for identification, thereby completing effective classification.
The invention has the following beneficial effects: the three rust diseases of the wheat can be well identified and classified, so that the model can be used for classifying stem rust diseases, leaf rust diseases and stripe rust diseases of wheat rust disease images obtained from different devices, and timely prevention and control of the different rust diseases of the wheat are researched according to the classification, and yield reduction is avoided. In addition, the method can be further used for providing technical support for the pathological research of different wheat rusts.
It will be understood that modifications and variations can be made by persons skilled in the art in light of the above teachings and all such modifications and variations are intended to be included within the scope of the invention as defined in the appended claims.
Claims (3)
1. A wheat rust classification method based on ensemble learning is characterized by comprising the following steps:
step 1: collecting an image of wheat rust after the wheat is in a green turning period;
step 2: preprocessing the image;
and step 3: constructing an integrated model, and training the model by using a snapshot integration method, wherein the method comprises the following specific steps:
step 301: selecting four models of VGG, ResNet, DenseNet and inclusion V3 with different convolutional layer connection modes for integration;
step 302: putting four models for integration into a model library, adopting a bagging integration mode, randomly selecting one model for training by utilizing hyper-parameter optimization search during each training, and setting enough iteration times N to ensure that each model is selected;
step 303: the method comprises the steps of training a model by adopting a snapshot integration mode, storing the model and parameters of each point reaching a local optimal point, and selecting the optimal model in the current training process through the loss of a verification set after one training is finished;
and 4, step 4: selecting the optimal model by adopting a fusion algorithm according to the loss in the training process, and specifically comprising the following steps of:
step 401: the loss magnitude is calculated using a weighted cross entropy loss function, as follows:
wherein N represents the number of categories, here 3, representing respectively stem rust, leaf rust and stripe rust; p is a radical ofiRepresenting the probability of the sample belonging to the ith class; y ═ y0,…,yN-1]Is a sample markOnehot representation of the label, y when the sample belongs to category ii1, the rest is 0; w ═ W0,…,wN-1]A weight representing a category i;
step 402: after the optimal model in the current training process is selected after one-time training is finished, a fusion algorithm is generated by using a 'replacement' thought, and results of all models obtained by current training are processed; multiplying the verification set loss of the optimal model in the current training process with the verification set losses of the other models in the current training process, and then performing evolution, if the obtained loss value is smaller, taking the smaller value as an optimal value, and obtaining the corresponding model weight through the same calculation, and taking the model as a new optimal model in the current training process;
step 403: after the iteration times reach the upper limit, the optimal model obtained by each iteration is fused and screened again by using the fusion algorithm to obtain the optimal model in the whole training process;
and 5: and using the obtained optimal model for identifying and classifying the wheat rust images.
2. The method according to claim 1, wherein the step 2 comprises the following specific steps:
step 201: making labels on the collected images, and dividing the labels into three categories of leaf rust wheat, stem rust wheat and stripe rust wheat;
step 202: and sample expansion is carried out on the three types of images by utilizing various data enhancement technologies, so that the data volume required by training is increased, and the generalization capability and robustness of the model are improved.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110416254.1A CN113111950B (en) | 2021-04-19 | 2021-04-19 | Wheat rust classification method based on ensemble learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110416254.1A CN113111950B (en) | 2021-04-19 | 2021-04-19 | Wheat rust classification method based on ensemble learning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113111950A CN113111950A (en) | 2021-07-13 |
CN113111950B true CN113111950B (en) | 2022-05-31 |
Family
ID=76718403
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110416254.1A Active CN113111950B (en) | 2021-04-19 | 2021-04-19 | Wheat rust classification method based on ensemble learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113111950B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113920358A (en) * | 2021-09-18 | 2022-01-11 | 广州番禺职业技术学院 | Automatic classification method, device and system for wheat diseases |
CN114549939B (en) * | 2022-04-26 | 2022-09-23 | 汉谷云智(武汉)科技有限公司 | Tree disease and pest image classification method based on image feature extraction |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109145939A (en) * | 2018-07-02 | 2019-01-04 | 南京师范大学 | A kind of binary channels convolutional neural networks semantic segmentation method of Small object sensitivity |
CN111476713A (en) * | 2020-03-26 | 2020-07-31 | 中南大学 | Intelligent weather image identification method and system based on multi-depth convolution neural network fusion |
WO2020194077A1 (en) * | 2019-03-22 | 2020-10-01 | International Business Machines Corporation | Unification of models having respective target classes with distillation |
CN112381116A (en) * | 2020-10-21 | 2021-02-19 | 福州大学 | Self-supervision image classification method based on contrast learning |
WO2021056046A1 (en) * | 2019-09-25 | 2021-04-01 | Presagen Pty Ltd | Method and system for performing non-invasive genetic testing using an artificial intelligence (ai) model |
Family Cites Families (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9443192B1 (en) * | 2015-08-30 | 2016-09-13 | Jasmin Cosic | Universal artificial intelligence engine for autonomous computing devices and software applications |
CN109165672A (en) * | 2018-07-16 | 2019-01-08 | 华南理工大学 | A kind of Ensemble classifier method based on incremental learning |
CN109978134B (en) * | 2019-02-26 | 2020-11-24 | 华中科技大学 | Fault prediction method based on fast integrated convolutional neural network |
CN111652238B (en) * | 2019-04-16 | 2023-06-02 | 上海铼锶信息技术有限公司 | Multi-model integration method and system |
CN110378380B (en) * | 2019-06-17 | 2023-09-29 | 江苏大学 | Image classification method based on multi-core integrated classification learning |
US20210049473A1 (en) * | 2019-08-14 | 2021-02-18 | The Board Of Trustees Of The Leland Stanford Junior University | Systems and Methods for Robust Federated Training of Neural Networks |
CN110532920B (en) * | 2019-08-21 | 2023-12-29 | 长江大学 | Face recognition method for small-quantity data set based on FaceNet method |
US11726492B2 (en) * | 2019-10-02 | 2023-08-15 | Zoox, Inc. | Collision avoidance perception system |
-
2021
- 2021-04-19 CN CN202110416254.1A patent/CN113111950B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109145939A (en) * | 2018-07-02 | 2019-01-04 | 南京师范大学 | A kind of binary channels convolutional neural networks semantic segmentation method of Small object sensitivity |
WO2020194077A1 (en) * | 2019-03-22 | 2020-10-01 | International Business Machines Corporation | Unification of models having respective target classes with distillation |
WO2021056046A1 (en) * | 2019-09-25 | 2021-04-01 | Presagen Pty Ltd | Method and system for performing non-invasive genetic testing using an artificial intelligence (ai) model |
CN111476713A (en) * | 2020-03-26 | 2020-07-31 | 中南大学 | Intelligent weather image identification method and system based on multi-depth convolution neural network fusion |
CN112381116A (en) * | 2020-10-21 | 2021-02-19 | 福州大学 | Self-supervision image classification method based on contrast learning |
Non-Patent Citations (1)
Title |
---|
Kohler, M等.Statistical theory for image classification using deep convolutional neural networks with cross-entropy loss.《arXiv:2011.13602》.2020,第1-27页. * |
Also Published As
Publication number | Publication date |
---|---|
CN113111950A (en) | 2021-07-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107016405B (en) | A kind of pest image classification method based on classification prediction convolutional neural networks | |
Sun et al. | Deep learning classification for crop types in north dakota | |
CN113111950B (en) | Wheat rust classification method based on ensemble learning | |
Zhou et al. | Strawberry maturity classification from UAV and near-ground imaging using deep learning | |
CN107392901A (en) | A kind of method for transmission line part intelligence automatic identification | |
CN113538390B (en) | Quick identification method for shaddock diseases and insect pests | |
CN113392775A (en) | Sugarcane seedling automatic identification and counting method based on deep neural network | |
CN107169556A (en) | stem cell automatic counting method based on deep learning | |
CN115272828B (en) | Intensive target detection model training method based on attention mechanism | |
Sharma et al. | Geospatial technology in agroforestry: status, prospects, and constraints | |
Yalcin | An approximation for a relative crop yield estimate from field images using deep learning | |
Yu et al. | A recognition method of soybean leaf diseases based on an improved deep learning model | |
CN114627411A (en) | Crop growth period identification method based on parallel detection under computer vision | |
Ariza-Sentís et al. | Object detection and tracking on UAV RGB videos for early extraction of grape phenotypic traits | |
CN111814563A (en) | Method and device for classifying planting structures | |
Ärje et al. | Automatic flower detection and classification system using a light-weight convolutional neural network | |
Zhang et al. | Appearance quality classification method of Huangguan pear under complex background based on instance segmentation and semantic segmentation | |
CN118097465A (en) | Unmanned aerial vehicle image and deep learning-based field rice spike positioning and identifying method | |
Chadwick et al. | Transferability of a mask R–CNN model for the delineation and classification of two species of regenerating tree crowns to untrained sites | |
Bharath et al. | Detection of plant leaf diseases using CNN | |
Li et al. | Automatic counting of rapeseed inflorescences using deep learning method and UAV RGB imagery | |
CN116503740A (en) | Unmanned aerial vehicle vision system capable of accurately identifying crop types | |
CN114511850B (en) | Method for identifying size particle image of sunlight rose grape fruit | |
Lozano-Garzon et al. | Remote sensing and machine learning modeling to support the identification of sugarcane crops | |
Zhu et al. | Exploring soybean flower and pod variation patterns during reproductive period based on fusion deep learning |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |