AU2020102885A4 - Disease recognition method of winter jujube based on deep convolutional neural network and disease image - Google Patents
Disease recognition method of winter jujube based on deep convolutional neural network and disease image Download PDFInfo
- Publication number
- AU2020102885A4 AU2020102885A4 AU2020102885A AU2020102885A AU2020102885A4 AU 2020102885 A4 AU2020102885 A4 AU 2020102885A4 AU 2020102885 A AU2020102885 A AU 2020102885A AU 2020102885 A AU2020102885 A AU 2020102885A AU 2020102885 A4 AU2020102885 A4 AU 2020102885A4
- Authority
- AU
- Australia
- Prior art keywords
- disease
- image
- color
- dcnn
- layer
- 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.)
- Ceased
Links
- 201000010099 disease Diseases 0.000 title claims abstract description 121
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 title claims abstract description 121
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 15
- 240000008866 Ziziphus nummularia Species 0.000 title 1
- 241001247821 Ziziphus Species 0.000 claims abstract description 55
- 230000003902 lesion Effects 0.000 claims abstract description 26
- 238000012549 training Methods 0.000 claims abstract description 15
- 238000003064 k means clustering Methods 0.000 claims abstract description 4
- 238000011176 pooling Methods 0.000 claims description 12
- 238000012360 testing method Methods 0.000 claims description 10
- 238000007781 pre-processing Methods 0.000 claims description 8
- 238000001514 detection method Methods 0.000 claims description 6
- 210000002569 neuron Anatomy 0.000 claims description 6
- 238000009499 grossing Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 2
- 230000002708 enhancing effect Effects 0.000 claims description 2
- 238000001914 filtration Methods 0.000 claims description 2
- 238000003062 neural network model Methods 0.000 claims description 2
- 230000002776 aggregation Effects 0.000 claims 1
- 238000004220 aggregation Methods 0.000 claims 1
- 238000012544 monitoring process Methods 0.000 abstract description 3
- 230000006870 function Effects 0.000 description 5
- 230000004913 activation Effects 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 3
- 235000013399 edible fruits Nutrition 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 230000002265 prevention Effects 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 239000000654 additive Substances 0.000 description 1
- 230000000996 additive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000006806 disease prevention Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- JEIPFZHSYJVQDO-UHFFFAOYSA-N iron(III) oxide Inorganic materials O=[Fe]O[Fe]=O JEIPFZHSYJVQDO-UHFFFAOYSA-N 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30096—Tumor; Lesion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
- G06T2207/30188—Vegetation; Agriculture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/68—Food, e.g. fruit or vegetables
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The winter jujube disease identification method is proposed based on deep convolutional
neural network and disease image. Firstly convert the original winter jujube disease RGB
color image collected by IOT into a YUV color model, and then preprocess it, then obtain the
disease rectangular area of interest, segment the spot image by K-means clustering algorithm
to obtain YUV color lesion image. Construct a three-channel hierarchical convolutional
neural network model, use the training data to train the model, and finally input the winter
jujube disease image into the trained model to identify the disease category. The invention
can be applied to the greenhouse winter jujube disease based on IOT. In the monitoring
system, higher disease identification results can be obtained.
-1/3
S Color components Hierarchical
Deep
Convolutional Neural Network
Color Hierarchical Deep Classif- Output
Disease-- * Color components Convolutional Neural Network ication recognition
image Flayer result
S Color component 3 Hierarchical
Deep
Convolutional Neural Network
Fig. 1
Description
-1/3
S Color components Hierarchical Deep Convolutional Neural Network
Color Hierarchical Deep Classif- Output Disease-- * Color components Convolutional Neural Network ication recognition image Flayer result
S Color component 3 Hierarchical Deep Convolutional Neural Network
Fig. 1
PATENTS ACT 1990
Disease recognition method of winter jujube based on deep convolutional neural network and disease image
The invention is described in the following statement:-
Disease recognition method of winter jujube based on deep convolutional neural network and disease image
The invention relates to the technical field of fruit disease image processing and machine learning, in particular to a winter jujube disease recognition method based on deep convolutional neural networks (DCNN) and disease images.
Because winter jujube is grown in the large greenhouse and the greenhouse environment is suitable for the occurrence of winter jujube diseases, there are many kinds of diseases and frequent occurrence. The main diseases include jujube rust, jujube anthracnose, jujube fruit shrinking disease, jujube scorch leaf disease, etc. The disease of winter jujube seriously affected the yield and quality of winter jujube. There are many ways to prevent and control winter jujube diseases. Some authors designed a Web GIS based information service platform for winter jujube diseases. A lot of disease prevention and control problems have been investigated in 10 townships in Dali County, Shaanxi, and summarized the occurrence of common diseases of winter jujube and the corresponding control measures. We summarized the comprehensive prevention and control methods of common winter jujube diseases, and provided references for the detection and control of winter jujube diseases, and focus on the prevalence of the main diseases of winter jujube and comprehensive prevention and control techniques, which can provide references for the majority of grassroots agricultural technicians and jujube farmers. Although there are many methods to prevent and control winter jujube diseases, most of the existing methods still rely on manual observation and comparison of disease symptoms for disease identification. Due to the complex causes of winter jujube diseases and the large environmental differences in various winter jujube greenhouses, it is difficult for most fruit farmers to correctly diagnose their winter jujube diseases according to existing methods.
Research on crop disease recognition methods based on crop disease images has always been an important research topic in the field of image processing and machine learning. However, due to the ever-changing disease images of winter jujube and the complicated background, traditional crop disease identification methods cannot meet the actual needs of the winter jujube disease monitoring system based on the Internet of Things (IOT). In recent years, image recognition methods based on deep learning have achieved outstanding classification results. The rapid development of computer science, IOT technology and deep learning and their applications provides the possibility to realize remote intelligent identification of winter jujube diseases. IOT can collect winter jujube disease images in real time, and deep learning can automatically learn representative features from the collected massive and complex disease images. These features can quickly and accurately identify disease types, but the above technologies have not been discovered yet. Integrated disease identification method of winter jujube.
In order to overcome the above-mentioned shortcomings of the classical methods, the purpose of the invention is to provide a method for identifying diseases of winter jujube based on DCNN and disease images, combining DCNN and the disease images to recognize the disease of winter jujube in the greenhouse. Prevention and treatment provide accurate disease information, and realize automatic detection and identification of diseases in greenhouse winter jujube based on IOT. In order to achieve the above objective, the invention adopts the following technical solutions.
Disease recognition method of winter jujube based on DCNN and disease image,
Step 1: After filtering, enhancing, smoothing and normalizing the disease image of winter jujube collected by IOT, the binary disease image is obtained after binary preprocessing, and then the binary disease image is multiplied with the original image to obtain the preprocessing after the color disease image.
Step 2: Combine the color characteristics of the preprocessed color disease image with the histogram threshold method to obtain the pixel value range of the color disease image, and then estimate the center of the color disease image after the preprocessing, and then take the center of the disease spots as the center. Extract a region of interest rectangle containing the color disease image by the region of interest detection method.
Step 3: Use the K-means clustering algorithm to segment the rectangle of the region of interest to obtain the color lesion image.
Step 4: Divide all the segmented color lesion images into training set and test set.
Step 5: Construct a three-channel hierarchical DCNN.
Step 6: Three color components of each color lesion image in the training set are input into three channels of three-channel hierarchical DCNN, and the representative features are output after layer-by-layer calculation, the model parameters are continuously adjusted, Iteratively train three-channel hierarchical DCNN until the recognition difference between two previous iterations is less than a given threshold, the iteration is terminated, and a stable trained DCNN model is obtained.
Step 7: Input three color components of each color lesion image in the test set to the trained DCNN, and get the disease recognition result in the classification layer of DCNN.
The training set of step 4 is used to train and recognize three-channel hierarchical DCNN to obtain stable parameters of three-channel layered DCNN, and the test set is used to test three-channel hierarchical DCNN performance.
The three-channel hierarchical DCNN of Step 5 is composed of an input layer, three hierarchical DCNNs, and a classification layer. Three components of the color lesion image are used as the input data of the input layer. The classification layer is a classifier, the output features of each layered DCNN are used as the input data of the classification layer, and the output of the classification layer is the classification result of the lesion image. Each layered DCNN consists of 3 convolutional layers, 2 pooling layers and 3 fully connected layers. Each layer has multiple feature maps, each feature map has multiple neurons, and each feature map uses a convolution filter to extract a feature of the input lesion image component. Convolutional layer is used for feature extraction of input image. The pooling layer uses the maximum pooling algorithm to aggregate and statistics the features obtained from different positions, and calculates the maximum value of a feature on the selected area of the lesion image as the feature value. The fully connected layer is used to calculate the dot product between the input data and the weight, and to further propose representative features through the multiple connected neurons.
The representative features in step 6 include the combination of edge features, basic shape features and color feature combinations of disease images.
The training process of step 6 uses the unlabeled disease images for unsupervised learning. When three components of a color disease image are input into three channels of the constructed three-channel hierarchical DCNN, The three-channel hierarchical DCNN can quickly extract features, abstract layer by layer until the concept of disease image is formed, and classify at the classification layer.
The classification layer in step 7 is a classifier, and the classifier uses the feature combination obtained in step 6 to classify disease images.
The beneficial effects of the invention, three-channel hierarchical deep learning model proposed in the invention is a novel and efficient feature extraction method, which can automatically learn effective identification features from a large number of winter jujube training lesion images, thereby avoiding the process of artificial feature extraction and improving the disease of winter jujube. Recognition rate. The winter jujube disease identification model proposed in the invention can be applied to a video monitoring system for winter jujube diseases in a greenhouse based on OT, thereby improving the intelligence and automation of winter jujube disease detection and recognition. The invention has the advantages of high real-time performance, high recognition accuracy, stable recognition effect and strong practicability. It can be implemented on ordinary PC computers without special requirements for operating systems.
Figure 1 is a three-channel hierarchical DCNN.
Figure 2 is a hierarchical DCNN.
Figure 3 shows the winter jujube disease recognition model based on DCNN.
The invention will be described in detail below with reference to the figures and embodiments.
The winter jujube disease recognition method based on DCNN and disease images mainly includes the following steps.
Step 1: Image preprocessing. During the preprocessing of the winter jujube disease image, each pixel of the RGB color video image collected by IOT has three components of R, G, and B, that is, each pixel needs 3 bytes to store, so that a color disease can be stored images require larger storage space and are more complicated to process. Therefore, first convert the RGB image of the winter jujube disease into the YUV color space, where Y is brightness, and U and V are chromaticity, which are different between the components R and Y and different between the components B and Y, respectively. The advantages of the image are represented by YUV. Its luminance component (Y) and chrominance component (U, V) are independent of each other. Only two components of U and V can be used to express color characteristic, which can be encoded separately. It is easy to achieve compression and convenient for transmission and subsequent processing. Perform grayscale processing on the diseased YUV image of winter jujube, then use Gaussian filter to filter the diseased image, and then perform enhancement, smoothing, normalization and gray preprocessing to obtain the binary disease image. The image is multiplied by the original image to obtain the preprocessed color YUV disease image.
Step 2: Extract the regions of interest in winter jujube disease images, combine the preprocessed color YUV disease image with the histogram threshold method to obtain the pixel value range of the color spot image, thereby determine the gray range of the disease spots in the disease image, estimate the disease spots of the color YUV disease image center, take the center of the lesion as the center, and then use the region of interest detection method to extract a rectangular image of the region of interest that contains the color spot image. The length and width of the rectangle should be determined according to the parameters of the IOT video device.
Step 3: Use the K-means clustering algorithm to segment the rectangular image of the region of interest and obtain the color lesion image, K is the number of clusters. Its default value of K is 3.
Step 4: Divide all segmented lesion images into the training set and test set.
Step 5: Construct a three-channel hierarchical DCNN. The three-channel hierarchical DCNN shown in Figure 1 includes an input layer, three hierarchical DCNNs, and a classification layer. Three components (Y, U and V) of the YUV color lesion image are used as the input data of the input layer of the model. The classification layer is generally selected as a BP network, the output feature of each layered DCNN is used as the input of the BP network, and its output is the classification result of the lesion image. The hierarchical DCNN model shown in Figure 2 consists of 3 convolutional layers, 2 pooling layers, and 3 fully connected layers. Each layer has multiple feature maps, and each feature map has multiple neurons, each feature map uses a convolution filter to extract the feature of the input lesion image component. The convolutional layer is used to extract the features of the input image, the pooling layer uses the maximum pooling algorithm to aggregate and evaluate the features of different locations, and calculate the maximum value of a feature on the selected area of the disease image as the feature value. The fully connected layer realizes the dot product between the input data and the weight, and further specializes the features through multiple connected neurons.
The training process of three-channel layered DCNN is divided into two stages: forward propagation and back propagation. In the forward propagation stage, the output of the previous layer is the input of the current layer, and is passed layer by layer through the activation function. Take a sample X labeled y from the sample set, input the
sample into three-channel hierarchical DCNN, and calculate the corresponding actual output o, . The actual output of the entire three-channel layered DCNN is expressed as
0,=f,( ... (f2 (f,(XWI)W2)...),) K(1) where f0 is the activation function, and W(i=1,2,...,n) represents the trained mapping weight matrix.
The output of the current layer is represented as
x' = f(W'x'- 1 +b') (2)
where I is the number of layers of three-channel layered DCNN, W' represents the mapping weight matrix of the current layer that has been trained, and b' is the additive bias of the current three-channel layered DCNN. The three-channel layered DCNN adjusts and compresses the output through the activation function after completing the convolution mapping.
In the back-propagation stage, the error function is back-propagated. The error is regarded as the sensitivity of the base of each neuron. Generally, the stochastic gradient descent method is used to optimize the convolution parameters and bias. The main operation of each layered DCNN is convolution and pooling. In the first layer, a trainable filter F' convolution is used to calculate the component of an input disease image plus a bias b' to obtain the volume LaminatedC'. Use pooling technology to integrate feature points in small neighborhoods to obtain new features, for example, the sum of 4 pixels in each neighborhood integrate to a pixel. In the 1+1layer, weighting by w"',increasing the bias b'', and then using the activation function to generate a feature map S"' reduced by 4 times, and abstract the feature map layer by layer until the essential features of the disease image are formed. The pooling layer can be regarded as a fuzzy filter, plays the role of secondary feature extraction, and can reduce overfitting.
Step 6: As shown in Figure 3, in the winter jujube disease recognition model based on the DCNN and disease images, the training set is used to train three-channel layered convolutional deep neural network model. Three components of each YUV color lesion image in the training set are respectively input to three channels of three-channel hierarchical DCNN, and iterative training until the error of the recognition rate of the overall model is less than a given threshold, The iteration is terminated, and a stable DCNN is obtained.
Step 7: Three components of each color lesion image in the test set are input to the trained three-channel hierarchical DCNN, and then the extracted classification features are input to the BP network. The BP network uses these feature combination to analyze the disease image class.
Claims (6)
1. A winter jujube disease recognition method based on IOT and disease images, its characteristics include the following steps.
Step 1: After filtering, enhancing, smoothing and normalizing the disease image of winter jujube collected by IOT, the binary disease image is obtained after binary preprocessing, and then the binary disease image is multiplied with preprocessing of the original image.
Step 2: Combine the color characteristics of the preprocessed color disease image with the histogram threshold method to obtain the pixel value range of the color disease image, thereby obtaining the disease spot center of the preprocessed color disease image, and then take the disease spot center as the center, extract a region of interest rectangle containing the color disease image by the region of interest detection method.
Step 3: Use the K-means clustering algorithm to segment the rectangle of the region of interest to obtain the color lesion image.
Step 4: Divide all the segmented color lesion images into training set and test set.
Step 5: Construct a three-channel hierarchical DCNN.
Step 6: Three color components of each color lesion image in the training set are input to three channels of three-channel hierarchical DCNN, and the representative features are output after layer-by-layer calculation, the model parameters are continuously adjusted. Iteratively train three-channel hierarchical DCNN until the difference in the recognition rate of disease recognition using the classification features obtained from the previous two iterations is less than a given threshold, the iteration is terminated, and a stable deep convolution is obtained neural network model.
Step 7: Input three color components of each color lesion image in the test set to the trained DCNN, and get the disease recognition result in the classification layer of DCNN.
2. The winter jujube disease recognition method based on DCNN and disease images according to claim 1, wherein the training set of step 4 is used to train and recognize a three-channel hierarchical DCNN, obtain stable parameters of three-channel layered DCNN, the test set is used to test the performance of three-channel layered DCNN.
3. The method for identifying winter jujube diseases based on DCNN and disease images according to claim 1, wherein three-channel hierarchical DCNN in step 5 consists of one input layer, three The hierarchical DCNN and a classification layer are composed, three components of the color lesion image are used as the input data of the input layer, the classification layer is a classifier, and the output feature of each layered DCNN is used as the input data of the classification layer, the output of the classification layer is the classification result of the lesion image. Each layered DCNN consists of 3 convolutional layers, 2 pooling layers and 3 fully connected layers, each layer of the layered DCNN has multiple feature maps, each feature map has multiple neurons, and each feature map uses a convolution filter to extract a feature of the input lesion image component, the convolution layer is used to extract the features of the input image, the pooling layer uses the maximum pooling algorithm to perform aggregation statistics on the features obtained from the different positions, and calculate the maximum value of a feature on the selected area of the lesion image as the feature value, the fully connected layer is used to calculate the dot product between the input data and the weight.
4. The winter jujube disease recognition method based on DCNN and disease images according to claim 1, characterized in that: the representative features of step 6 include the combination of edge features, basic shape features and color feature combination.
5. The winter jujube disease recognition method based on DCNN and disease images according to claim 1, characterized in that: the training process of step 6 uses unlabeled disease images for unsupervised learning. When three components of the color disease image are input into three channels of the constructed three-channel layered DCNN, the DCNN model can quickly extract features, and the abstract layer by layer until the disease image is formed concept and classified in the classification layer.
6. The winter jujube disease recognition method based on DCNN and disease images according to claim 1, characterized in that: the classification layer of step 7) is a classifier, and the classifier uses the step 6). The resulting feature combination classifies disease images.
-1/3-
Hierarchical Deep Color component1 Convolutional Neural Network 2020102885
Color Classif- Output Hierarchical Deep Disease Color component 2 ication recognition Convolutional Neural Network image layer result
Hierarchical Deep Color component 3 Convolutional Neural Network
Fig. 1
-2/3-
Fig. 2
-3/3-
Fig. 3
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| AU2020102885A AU2020102885A4 (en) | 2020-10-20 | 2020-10-20 | Disease recognition method of winter jujube based on deep convolutional neural network and disease image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| AU2020102885A AU2020102885A4 (en) | 2020-10-20 | 2020-10-20 | Disease recognition method of winter jujube based on deep convolutional neural network and disease image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| AU2020102885A4 true AU2020102885A4 (en) | 2020-12-17 |
Family
ID=73746684
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| AU2020102885A Ceased AU2020102885A4 (en) | 2020-10-20 | 2020-10-20 | Disease recognition method of winter jujube based on deep convolutional neural network and disease image |
Country Status (1)
| Country | Link |
|---|---|
| AU (1) | AU2020102885A4 (en) |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112801119A (en) * | 2021-03-09 | 2021-05-14 | 中国农业科学院果树研究所 | Pear variety identification method based on image identification |
| CN112801942A (en) * | 2020-12-31 | 2021-05-14 | 广西慧云信息技术有限公司 | Citrus huanglongbing image identification method based on attention mechanism |
| CN112837805A (en) * | 2021-01-12 | 2021-05-25 | 浙江大学 | Deep learning-based eyelid topological morphology feature extraction method |
| CN112884025A (en) * | 2021-02-01 | 2021-06-01 | 安徽大学 | Tea disease classification system based on multi-feature sectional type training |
| CN113158754A (en) * | 2021-02-04 | 2021-07-23 | 安徽农业大学 | Tomato disease image identification method |
| CN113192026A (en) * | 2021-04-28 | 2021-07-30 | 北京林业大学 | Walnut maturity detection and prediction method based on visible light image |
| CN113267993A (en) * | 2021-04-22 | 2021-08-17 | 上海大学 | Network training method and device based on collaborative learning |
| CN113409253A (en) * | 2021-06-02 | 2021-09-17 | 南京公诚节能新材料研究院有限公司 | Agricultural condition monitoring key technical method |
| CN114359727A (en) * | 2021-12-31 | 2022-04-15 | 华南农业大学 | Tea disease identification method and system based on lightweight optimization Yolo v4 |
| CN114494261A (en) * | 2022-04-18 | 2022-05-13 | 陕西易合交通科技有限公司 | Underwater structure disease data processing method |
| CN114511732A (en) * | 2021-12-31 | 2022-05-17 | 广西慧云信息技术有限公司 | Citrus spotted disease and insect pest fine-grained image identification method |
| CN114549939A (en) * | 2022-04-26 | 2022-05-27 | 汉谷云智(武汉)科技有限公司 | Tree pest and disease picture classification method based on image feature extraction |
| CN114663386A (en) * | 2022-03-21 | 2022-06-24 | 东南大学 | Water film removing method for airport pavement disease image |
| CN114764887A (en) * | 2022-04-12 | 2022-07-19 | 中国农业大学 | Crop disease identification method, device, equipment and storage medium |
| CN115245122A (en) * | 2021-12-21 | 2022-10-28 | 重庆三峡学院 | Accurate cultivation system suitable for thunberg fritillary bulb is planted |
| CN115294405A (en) * | 2022-09-29 | 2022-11-04 | 浙江天演维真网络科技股份有限公司 | Method, device, equipment and medium for constructing crop disease classification model |
| CN116246176A (en) * | 2023-05-12 | 2023-06-09 | 山东建筑大学 | Crop disease detection method and device, electronic equipment and storage medium |
| CN116740650A (en) * | 2023-08-10 | 2023-09-12 | 青岛农业大学 | Crop breeding monitoring method and system based on deep learning |
| CN117392552A (en) * | 2023-12-13 | 2024-01-12 | 江西农业大学 | Blade disease identification method and system based on dual-path convolutional neural network |
| CN117456214A (en) * | 2023-11-06 | 2024-01-26 | 江苏省农业科学院 | Tomato leaf spot identification method, system and electronic equipment |
-
2020
- 2020-10-20 AU AU2020102885A patent/AU2020102885A4/en not_active Ceased
Cited By (33)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112801942A (en) * | 2020-12-31 | 2021-05-14 | 广西慧云信息技术有限公司 | Citrus huanglongbing image identification method based on attention mechanism |
| CN112801942B (en) * | 2020-12-31 | 2023-10-13 | 广西慧云信息技术有限公司 | Citrus yellow dragon disease image identification method based on attention mechanism |
| CN112837805B (en) * | 2021-01-12 | 2024-03-29 | 浙江大学 | Eyelid topological morphology feature extraction method based on deep learning |
| CN112837805A (en) * | 2021-01-12 | 2021-05-25 | 浙江大学 | Deep learning-based eyelid topological morphology feature extraction method |
| CN112884025A (en) * | 2021-02-01 | 2021-06-01 | 安徽大学 | Tea disease classification system based on multi-feature sectional type training |
| CN112884025B (en) * | 2021-02-01 | 2022-11-04 | 安徽大学 | Tea disease classification system based on multi-feature sectional type training |
| CN113158754A (en) * | 2021-02-04 | 2021-07-23 | 安徽农业大学 | Tomato disease image identification method |
| CN112801119A (en) * | 2021-03-09 | 2021-05-14 | 中国农业科学院果树研究所 | Pear variety identification method based on image identification |
| CN112801119B (en) * | 2021-03-09 | 2023-10-31 | 中国农业科学院果树研究所 | Pear variety identification method based on image identification |
| CN113267993A (en) * | 2021-04-22 | 2021-08-17 | 上海大学 | Network training method and device based on collaborative learning |
| CN113192026A (en) * | 2021-04-28 | 2021-07-30 | 北京林业大学 | Walnut maturity detection and prediction method based on visible light image |
| CN113192026B (en) * | 2021-04-28 | 2023-08-04 | 北京林业大学 | Walnut maturity detection and prediction method based on visible light image |
| CN113409253A (en) * | 2021-06-02 | 2021-09-17 | 南京公诚节能新材料研究院有限公司 | Agricultural condition monitoring key technical method |
| CN115245122A (en) * | 2021-12-21 | 2022-10-28 | 重庆三峡学院 | Accurate cultivation system suitable for thunberg fritillary bulb is planted |
| CN114359727A (en) * | 2021-12-31 | 2022-04-15 | 华南农业大学 | Tea disease identification method and system based on lightweight optimization Yolo v4 |
| CN114511732A (en) * | 2021-12-31 | 2022-05-17 | 广西慧云信息技术有限公司 | Citrus spotted disease and insect pest fine-grained image identification method |
| CN114359727B (en) * | 2021-12-31 | 2024-06-04 | 华南农业大学 | Tea disease identification method and system based on lightweight optimization Yolo v4 |
| CN114511732B (en) * | 2021-12-31 | 2024-05-14 | 广西慧云信息技术有限公司 | Orange spot disease and insect pest fine-granularity image identification method |
| CN114663386A (en) * | 2022-03-21 | 2022-06-24 | 东南大学 | Water film removing method for airport pavement disease image |
| CN114764887A (en) * | 2022-04-12 | 2022-07-19 | 中国农业大学 | Crop disease identification method, device, equipment and storage medium |
| CN114494261B (en) * | 2022-04-18 | 2023-04-25 | 陕西易合交通科技有限公司 | Disease data processing method for underwater structure |
| CN114494261A (en) * | 2022-04-18 | 2022-05-13 | 陕西易合交通科技有限公司 | Underwater structure disease data processing method |
| CN114549939A (en) * | 2022-04-26 | 2022-05-27 | 汉谷云智(武汉)科技有限公司 | Tree pest and disease picture classification method based on image feature extraction |
| CN115294405B (en) * | 2022-09-29 | 2023-01-10 | 浙江天演维真网络科技股份有限公司 | Method, device, equipment and medium for constructing crop disease classification model |
| CN115294405A (en) * | 2022-09-29 | 2022-11-04 | 浙江天演维真网络科技股份有限公司 | Method, device, equipment and medium for constructing crop disease classification model |
| CN116246176A (en) * | 2023-05-12 | 2023-06-09 | 山东建筑大学 | Crop disease detection method and device, electronic equipment and storage medium |
| CN116246176B (en) * | 2023-05-12 | 2023-09-19 | 山东建筑大学 | Crop disease detection method and device, electronic equipment and storage medium |
| CN116740650B (en) * | 2023-08-10 | 2023-10-20 | 青岛农业大学 | Crop breeding monitoring method and system based on deep learning |
| CN116740650A (en) * | 2023-08-10 | 2023-09-12 | 青岛农业大学 | Crop breeding monitoring method and system based on deep learning |
| CN117456214A (en) * | 2023-11-06 | 2024-01-26 | 江苏省农业科学院 | Tomato leaf spot identification method, system and electronic equipment |
| CN117456214B (en) * | 2023-11-06 | 2024-05-31 | 江苏省农业科学院 | Tomato leaf spot identification method, system and electronic equipment |
| CN117392552A (en) * | 2023-12-13 | 2024-01-12 | 江西农业大学 | Blade disease identification method and system based on dual-path convolutional neural network |
| CN117392552B (en) * | 2023-12-13 | 2024-02-20 | 江西农业大学 | Blade disease identification method and system based on dual-path convolutional neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| AU2020102885A4 (en) | Disease recognition method of winter jujube based on deep convolutional neural network and disease image | |
| Waghmare et al. | Detection and classification of diseases of grape plant using opposite colour local binary pattern feature and machine learning for automated decision support system | |
| Tm et al. | Tomato leaf disease detection using convolutional neural networks | |
| Dhingra et al. | Study of digital image processing techniques for leaf disease detection and classification | |
| JP6935377B2 (en) | Systems and methods for automatic inference of changes in spatiotemporal images | |
| Jaisakthi et al. | Grape leaf disease identification using machine learning techniques | |
| Ganatra et al. | A survey on diseases detection and classification of agriculture products using image processing and machine learning | |
| CN111598001B (en) | Identification method for apple tree diseases and insect pests based on image processing | |
| Anandhakrishnan et al. | Deep Convolutional Neural Networks for image based tomato leaf disease detection | |
| CN107563389A (en) | A kind of corps diseases recognition methods based on deep learning | |
| CN111860330A (en) | Apple leaf disease identification method based on multi-feature fusion and convolutional neural network | |
| CN109344699A (en) | Winter jujube disease recognition method based on depth of seam division convolutional neural networks | |
| Sabrol et al. | Fuzzy and neural network based tomato plant disease classification using natural outdoor images | |
| CN111127423B (en) | Rice pest and disease identification method based on CNN-BP neural network algorithm | |
| Lv et al. | A visual identification method for the apple growth forms in the orchard | |
| Patil et al. | Enhanced radial basis function neural network for tomato plant disease leaf image segmentation | |
| Pinto et al. | Crop disease classification using texture analysis | |
| CN109086823B (en) | Automatic statistical method for wheat scab ear disease rate | |
| CN115376008A (en) | Method and device for identifying plant diseases and insect pests, electronic equipment and storage medium | |
| Liu et al. | Recognition of pyralidae insects using intelligent monitoring autonomous robot vehicle in natural farm scene | |
| Rai et al. | Classification of diseased cotton leaves and plants using improved deep convolutional neural network | |
| Patil | Pomegranate fruit diseases detection using image processing techniques: a review | |
| Wakhare et al. | Development of automated leaf disease detection in pomegranate using Alexnet algorithm | |
| Sutha et al. | Plant disease detection using fuzzy classification | |
| CN113221913A (en) | Agriculture and forestry disease and pest fine-grained identification method and device based on Gaussian probability decision-level fusion |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| FGI | Letters patent sealed or granted (innovation patent) | ||
| MK22 | Patent ceased section 143a(d), or expired - non payment of renewal fee or expiry |