CN114022717A - Blade disease identification method and system based on image identification - Google Patents

Blade disease identification method and system based on image identification Download PDF

Info

Publication number
CN114022717A
CN114022717A CN202111609169.3A CN202111609169A CN114022717A CN 114022717 A CN114022717 A CN 114022717A CN 202111609169 A CN202111609169 A CN 202111609169A CN 114022717 A CN114022717 A CN 114022717A
Authority
CN
China
Prior art keywords
image
blade
leaf
original
disease
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
Application number
CN202111609169.3A
Other languages
Chinese (zh)
Inventor
张达平
宫庆涛
杨鸿钰
李淼
李桂祥
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shandong Institute of Pomology
Original Assignee
Shandong Institute of Pomology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shandong Institute of Pomology filed Critical Shandong Institute of Pomology
Priority to CN202111609169.3A priority Critical patent/CN114022717A/en
Publication of CN114022717A publication Critical patent/CN114022717A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • 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
    • 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
    • G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
    • 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
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/13—Edge detection
    • 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
    • G06T2207/30168—Image quality inspection

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Computing Systems (AREA)
  • Molecular Biology (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Probability & Statistics with Applications (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Health & Medical Sciences (AREA)
  • Quality & Reliability (AREA)
  • Image Analysis (AREA)

Abstract

The application discloses a blade disease identification method and system based on image identification, which are used for solving the problems of low cost, low efficiency and insufficient accuracy of the existing blade disease identification. The method comprises the steps of collecting an original blade image to be processed; screening the original blade image based on the image definition; preprocessing the screened original blade image to obtain a blade area image; and classifying the leaf area images according to a pre-trained leaf disease identification model based on a convolutional neural network, and identifying and determining the disease types of the leaves corresponding to the leaf area images. According to the method, the health condition of the leaves is monitored in an automatic image identification mode, the accuracy rate is high, the disease condition of the leaves can be accurately judged, the monitoring efficiency is improved, and the labor cost is reduced.

Description

Blade disease identification method and system based on image identification
Technical Field
The application relates to the technical field of image recognition, in particular to a blade disease recognition method and system based on image recognition.
Background
In the process of plant growth, a series of morphological and physiological pathological changes exist, which can hinder the normal growth and development of plants, thereby affecting the economic benefit. The diseases of plants are externally shown as pathological changes of leaves, roots and the like, such as discoloration, necrosis, rot and the like.
At present, diseases of plants can be monitored and treated in a manual monitoring mode.
Based on this, a scheme for more accurately and efficiently identifying the leaf diseases through an image identification mode is needed.
Disclosure of Invention
The embodiment of the application provides a blade disease identification method and system based on image identification, and aims to solve the problems of low cost, low efficiency and insufficient accuracy of the conventional blade disease identification.
The blade disease identification method based on image identification provided by the embodiment of the application comprises the following steps:
collecting an original blade image to be processed;
screening the original blade image based on the image definition;
preprocessing the screened original blade image to obtain a blade area image;
classifying the leaf area images according to a pre-trained leaf disease identification model based on a convolutional neural network, and identifying and determining the disease types of the leaves corresponding to the leaf area images.
In one example, screening the original blade image based on image sharpness specifically includes: evaluating the definition of the original blade image according to a gradient function; and screening to obtain the original blade image larger than the definition threshold value according to the preset definition threshold value.
In one example, screening the original blade image based on image sharpness specifically includes: extracting gradient values according to a Laplacian operator; according to the formula
Figure 604951DEST_PATH_IMAGE001
Figure 472544DEST_PATH_IMAGE002
Calculating to obtain a definition value of the original blade image; wherein D (f) represents the image definition calculation result, G (x, y) represents the convolution of Laplacian operators at pixel points (x, y), and G (x, y) represents the convolution of Laplacian operators at pixel points (x, y) represents the image definition calculation resultxRepresents the convolution of Laplacian horizontal direction edge detection operator at pixel point (x, y), GyExpressing the convolution of Laplacian vertical direction edge detection operators at pixel points (x, y), and T expressing a preset edge detection threshold value; and determining the preset definition threshold value as 200, and screening to obtain the original blade image with the definition value larger than 200.
In one example, preprocessing the screened original leaf image to obtain a leaf region image specifically includes: randomly turning over the screened original leaf image and carrying out average brightness processing; obtaining a blade area image containing the blade through identification and interception; and carrying out normalization processing on the blade area image.
In one example, the leaf disease identification model is trained by: collecting original leaf images, screening and preprocessing to obtain leaf area images serving as training data sets; determining a classification label corresponding to the training data set; and training a ResNet34 convolutional neural network model according to the training data set and the classification labels to obtain a trained leaf disease identification model.
In one example, the classifying the leaf area image according to a pre-trained leaf disease identification model based on a convolutional neural network, and identifying and determining the disease type of a leaf corresponding to the leaf area image specifically includes: inputting the leaf area image into a pre-trained leaf disease identification model based on a convolutional neural network, and calculating an average value corresponding to each preset classification label; and selecting the classification label corresponding to the maximum probability from a plurality of probabilities corresponding to the classification labels as the disease type of the blade corresponding to the blade area image which is identified and determined.
In one example, inputting the leaf area image into a pre-trained leaf disease recognition model based on a convolutional neural network, and calculating an average value corresponding to each preset classification label, specifically including: according to the formula
Figure 167967DEST_PATH_IMAGE003
Calculating the average value corresponding to each preset classification label; wherein, rankiIndicates the serial number of the ith sample, M indicates the number of positive samples, and N indicates the number of negative samples.
In one example, acquiring a raw blade image to be processed specifically includes: determining acquisition modes aiming at different planting areas according to the plant density and the leaf density corresponding to each plant; and respectively acquiring the original blade images to be processed in the different planting areas according to the determined acquisition modes.
In one example, the method further comprises: determining first position information of the diseased leaf area image in the corresponding original leaf image; determining image acquisition equipment corresponding to the original blade image; and determining the plant to which the diseased leaf area image belongs according to the second position information of the image acquisition equipment and the first position information.
The blade disease identification system based on image recognition that this application embodiment provided includes:
the image acquisition equipment is used for acquiring an original blade image to be processed;
and the processor is used for screening the original blade images based on the image definition, preprocessing the screened original blade images to obtain blade area images, classifying the blade area images according to a pre-trained blade disease identification model based on a convolutional neural network, and identifying and determining the disease types of the blades corresponding to the blade area images.
The embodiment of the application provides a blade disease identification method and system based on image identification, which are based on screening and preprocessing of original blade images, extract blade area images with more information content, and enhance the characteristic expression of the images, so that corresponding disease types can be identified more accurately when the blade area images are identified by a blade disease identification model in the follow-up process. The health condition of the leaves is monitored in an automatic image identification mode, the accuracy rate is high, the disease condition of the leaves can be accurately judged, meanwhile, a convolutional neural network classification method is adopted, the classification accuracy of the diseases of the leaves is improved, reproducibility and instantaneity are achieved, related personnel can timely treat the diseases of plants, and yield reduction caused by information stagnation is effectively avoided. Moreover, the method preprocesses the picture, effectively avoids the influence of outdoor sunlight change on image exposure, and is also beneficial to improving the accuracy of leaf disease identification.
In addition, based on automatic identification and classification of the blade disease identification model, the efficiency of monitoring the blade diseases is effectively improved, the labor cost is reduced, and the method has the characteristics of low cost and wide application area.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:
fig. 1 is a flowchart of a method for identifying a leaf disease based on image identification according to an embodiment of the present application;
FIG. 2 is a flowchart of training a leaf disease identification model according to an embodiment of the present disclosure;
fig. 3 is a schematic structural diagram of a blade disease identification system based on image identification according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be described in detail and completely with reference to the following specific embodiments of the present application and the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
Fig. 1 is a flowchart of a method for identifying a leaf disease based on image identification, which includes the following steps:
s101: and acquiring an original blade image to be processed.
Aiming at the plants to be monitored, the images of the leaves can be collected and used as the original leaf images to be processed for subsequent processing and identification. One original leaf image may only contain the leaf image of one plant, and may also contain the leaf images of a plurality of different plants.
S102: and screening the original blade image based on the image definition.
The acquired original blade image may have poor image definition, difficult identification and poor identification effect due to plant shaking, bad weather and other reasons. Based on the method, the original blade image needs to be screened according to the image definition so as to obtain an effective image capable of performing subsequent identification operation, so that the efficiency of blade disease identification is improved, useless operation is saved, and the accuracy of subsequent blade disease identification is guaranteed.
In one embodiment, when the original leaf images are screened, the definition of the original leaf images can be evaluated according to a gradient function, and then the original leaf images larger than a definition threshold are screened according to a preset definition threshold to serve as images for subsequent identification. By determining the screening standard of the original leaf image, the operability of image screening is guaranteed, the definition of the screened image is guaranteed, and the accuracy of leaf disease identification is improved subsequently.
Specifically, the screening process of the original blade image may include:
firstly, extracting gradient values according to a Laplacian operator;
Figure 878434DEST_PATH_IMAGE004
second, according to the formula
Figure 747033DEST_PATH_IMAGE005
，
Figure 593766DEST_PATH_IMAGE006
Calculating to obtain the definition value of the original blade image; wherein D (f) represents the image definition calculation result, G (x, y) represents the convolution of Laplacian operators at pixel points (x, y), and G (x, y) represents the convolution of Laplacian operators at pixel points (x, y) represents the image definition calculation resultxRepresents the convolution of Laplacian horizontal direction edge detection operator at pixel point (x, y), GyExpressing the convolution of Laplacian vertical direction edge detection operators at pixel points (x, y), and T expressing a preset edge detection threshold value;
thirdly, determining the preset definition threshold value as 200, and screening to obtain the original blade image with the definition value larger than 200.
The scheme is favorable for accurately determining the definition value of the original blade image and is convenient for image screening.
S103: and preprocessing the screened original blade image to obtain a blade area image.
Because the acquired original blade image may contain useless and interfering information, such as soil, sky, small animals, insects and other interferents, in order to improve the accuracy of image identification, the original blade image needs to be preprocessed to obtain a specific blade area image, so as to enhance the attention of a subsequent model to the blade to be identified and ensure the accuracy of blade disease identification. Wherein, only the blade to be identified is generally contained in the blade area image.
In one embodiment, the pre-processing of the raw leaf image may include:
firstly, randomly turning an original blade image vertically, horizontally, 90 degrees and the like at a certain probability (such as 50%);
secondly, setting the uniform exposure and contrast, and calculating the brightness value of the original blade image; if the brightness value is slightly bright, the exposure of the image is reduced; if the brightness value is darker, the exposure of the image is increased so as to realize the average brightness processing of the original leaf image; the threshold values of the brightness and the darkness can be determined through presetting;
thirdly, after the blade area image containing the blade is obtained through identification and interception, normalization processing is carried out on the blade area image. In particular by a normalizing function
Figure 194512DEST_PATH_IMAGE007
Implementation, where x represents an image pixel point value, min (x) is an image pixel minimum value, and max (x) is an image pixel maximum value.
S104: classifying the leaf area images according to a pre-trained leaf disease identification model based on a convolutional neural network, and identifying and determining the disease types of the leaves corresponding to the leaf area images.
The obtained leaf area images are identified and classified through pre-training a leaf disease identification model based on a convolutional neural network, so that the disease types corresponding to the leaf area images are determined, and the leaf diseases are identified and monitored.
In one embodiment, as shown in fig. 2, the training mode of the leaf lesion identification model includes the following steps:
firstly, collecting an original blade image, screening and preprocessing the original blade image, identifying the original blade image to obtain a blade area image, and normalizing the blade area image to be used as a training data set;
secondly, determining a classification label corresponding to the training data set, wherein the classification label comprises a disease-free type and a plurality of disease types corresponding to plants;
thirdly, training a ResNet34 convolutional neural network model according to the training data set and the classification labels, training and obtaining the optimal learning rate, and obtaining a trained leaf disease identification model.
It should be noted that, if some steps in the blade defect identification model are similar to some steps in the blade defect identification method, the parts that are not described in detail in this embodiment may specifically refer to corresponding parts in the embodiment of the blade defect identification method, and this is not described in detail in this embodiment.
In one embodiment, since the different types of plants have different leaf diseases and different image characteristics, the corresponding leaf disease identification models can be trained for the different types of plants respectively, so that the leaf diseases of the corresponding types of plants can be identified by using a specific model, and the accuracy of identifying the leaf diseases of the different types of plants can be improved.
In one embodiment, when the leaf area image is processed by using the leaf disease identification model, the leaf area image is input into a pre-trained leaf disease identification model based on a convolutional neural network, and an average value corresponding to each preset classification label is calculated. And selecting a classification label corresponding to the maximum probability from a plurality of probabilities corresponding to the classification labels, wherein the classification label represents that the probability that the blade area image is matched with the classification label is maximum, and then taking the disease type represented by the classification label as the disease type of the blade corresponding to the identified blade area image.
Specifically, the leaf area image is input into a pre-trained leaf disease identification model based on a convolutional neural network, and when an average value corresponding to each preset classification label is calculated, the average value can be calculated according to a formula
Figure 2062DEST_PATH_IMAGE008
Calculating the average value corresponding to each preset classification label; wherein, rankiIndicates the serial number of the ith sample, M indicates the number of positive samples, and N indicates the number of negative samples.
In the embodiment of the application, the leaf area image with more information content is extracted based on screening and preprocessing of the original leaf image, and the feature expression of the image is enhanced, so that the corresponding disease type can be more accurately identified when the leaf area image is identified by adopting a leaf disease identification model subsequently. In addition, based on automatic identification and classification of the blade disease identification model, the efficiency of monitoring the blade diseases is effectively improved, and the labor cost is reduced.
In one embodiment, since the plants have different growth conditions and the density of the plants in different planting areas are different, when the original leaf image to be processed is collected, the collection mode for different planting areas can be determined according to the plant density and the leaf density corresponding to each plant. The acquisition mode comprises an acquisition range, the number of acquisition points, an acquisition angle and the like. For example, in a dense plant area, more acquisition points are arranged, and the acquisition height and angle are adjusted for multiple times, so that all leaf images of all plants are acquired.
And respectively acquiring original leaf images to be processed in different planting areas according to the determined acquisition modes. This is favorable to all plants to planting the region all to carry out effective control, prevents to lead to some plants to be covered and the condition of lack of control to take place because of the plant is too close.
In one embodiment, if it is determined that the leaf disease exists in the leaf area image, it is determined which plant corresponds to the leaf disease. First position information of the diseased leaf area image in the corresponding original leaf image, which represents the position of the leaf area image in the original leaf image before image segmentation, can be determined. Secondly, determining image acquisition equipment corresponding to the acquisition of the original blade image and state information including acquisition height, angle and the like of the image acquisition equipment maintained at that time. Finally, the corresponding actual position of the blade area image when being collected can be determined according to the second position information of the image collecting device, the state information of the image collecting device and the first position information of the blade area image, and the plant to which the diseased blade area image belongs can be determined.
It should be noted that the scheme can be applied to identification of leaf diseases, and also can be applied to identification of other diseases with similar principles, such as identification of diseases of plant roots and stems.
In one embodiment, peaches are one of the important fruits native to our country, and as the world's largest producing countries of peaches and nectarines, the yield reaches over 1400 million tons per year. However, because of the existence of some peach diseases and insect pests (such as peach bacterial perforation disease and cladosporium cucumerinum), the peaches are easily affected by serious yield reduction. Therefore, the scheme is applied to prevention and treatment of peach diseases, good growth of peach trees is facilitated, and economic benefits are improved.
Based on the same inventive concept, the blade disease identification method based on image identification provided by the embodiment of the present application further provides a corresponding blade disease identification system based on image identification, as shown in fig. 3.
Fig. 3 is a schematic structural diagram of a blade disease identification system based on image identification according to an embodiment of the present application, which specifically includes:
the image acquisition device 31 is used for acquiring original blade images to be processed and can be a high-definition camera. Wherein, the place of placing of image acquisition equipment, place quantity can be confirmed according to the planting density of plant to make the collection picture can cover most plant leaf. The image acquisition device is provided with independent position information, which can be stored in the memory 33.
And the processor 32 is configured to receive the image acquired by the image acquisition device 31, screen the original blade image based on image sharpness, pre-process the screened original blade image to obtain a blade area image, classify the blade area image according to a pre-trained convolutional neural network-based blade disease identification model, and identify and determine a disease type of a blade corresponding to the blade area image.
The system may also include a display 34 for displaying plant leaf disease information.
The embodiments in the present application are described in a progressive manner, and the same and similar parts among the embodiments can be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims (8)

1. A blade disease identification method based on image identification is characterized by comprising the following steps:
collecting an original blade image to be processed;
screening the original blade image based on the image definition;
preprocessing the screened original blade image to obtain a blade area image;
classifying the leaf area images according to a pre-trained leaf disease identification model based on a convolutional neural network, and identifying and determining the disease types of the leaves corresponding to the leaf area images;
based on the image definition, screening the original blade image, which specifically comprises the following steps:
extracting gradient values according to a Laplacian operator;
according to the formula
Figure 979604DEST_PATH_IMAGE001
Figure 71057DEST_PATH_IMAGE002
Calculating to obtain a definition value of the original blade image; wherein D (f) represents the image definition calculation result, G (x, y) represents the convolution of Laplacian operators at pixel points (x, y), and G (x, y) represents the convolution of Laplacian operators at pixel points (x, y) represents the image definition calculation resultxRepresents the convolution of Laplacian horizontal direction edge detection operator at pixel point (x, y), GyExpressing the convolution of Laplacian vertical direction edge detection operators at pixel points (x, y), and T expressing a preset edge detection threshold value;
and determining the preset definition threshold value as 200, and screening to obtain the original blade image with the definition value larger than 200.
2. The method according to claim 1, wherein preprocessing the screened original leaf image to obtain a leaf region image comprises:
randomly turning over the screened original leaf image and carrying out average brightness processing;
obtaining a blade area image containing the blade through identification and interception;
and carrying out normalization processing on the blade area image.
3. The method according to claim 1, wherein the leaf disease identification model is trained by:
collecting original leaf images, screening and preprocessing to obtain leaf area images serving as training data sets;
determining a classification label corresponding to the training data set;
and training a ResNet34 convolutional neural network model according to the training data set and the classification labels to obtain a trained leaf disease identification model.
4. The method according to claim 1, wherein the leaf area images are classified according to a pre-trained leaf disease identification model based on a convolutional neural network, and the identification and determination of the disease types of the leaves corresponding to the leaf area images specifically include:
inputting the leaf area image into a pre-trained leaf disease identification model based on a convolutional neural network, and calculating an average value corresponding to each preset classification label;
and selecting the classification label corresponding to the maximum probability from a plurality of probabilities corresponding to the classification labels as the disease type of the blade corresponding to the blade area image which is identified and determined.
5. The method according to claim 4, wherein the leaf area image is input into a pre-trained leaf disease recognition model based on a convolutional neural network, and an average value corresponding to each preset classification label is calculated, specifically comprising:
according to the formula
Figure 45966DEST_PATH_IMAGE003
Calculating the average value corresponding to each preset classification label; wherein, rankiIndicates the serial number of the ith sample, M indicates the number of positive samples, and N indicates the number of negative samples.
6. The method according to claim 1, wherein acquiring raw blade images to be processed specifically comprises:
determining acquisition modes aiming at different planting areas according to the plant density and the leaf density corresponding to each plant;
and respectively acquiring the original blade images to be processed in the different planting areas according to the determined acquisition modes.
7. The method of claim 1, further comprising:
determining first position information of the diseased leaf area image in the corresponding original leaf image;
determining image acquisition equipment corresponding to the original blade image;
and determining the plant to which the diseased leaf area image belongs according to the second position information of the image acquisition equipment and the first position information.
8. A leaf disease identification system based on image identification is characterized by comprising:
the image acquisition equipment is used for acquiring an original blade image to be processed;
and the processor is used for screening the original blade images based on the image definition, preprocessing the screened original blade images to obtain blade area images, classifying the blade area images according to a pre-trained blade disease identification model based on a convolutional neural network, and identifying and determining the disease types of the blades corresponding to the blade area images.
CN202111609169.3A 2021-12-27 2021-12-27 Blade disease identification method and system based on image identification Pending CN114022717A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111609169.3A CN114022717A (en) 2021-12-27 2021-12-27 Blade disease identification method and system based on image identification

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111609169.3A CN114022717A (en) 2021-12-27 2021-12-27 Blade disease identification method and system based on image identification

Publications (1)

Publication Number Publication Date
CN114022717A true CN114022717A (en) 2022-02-08

Family

ID=80069207

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111609169.3A Pending CN114022717A (en) 2021-12-27 2021-12-27 Blade disease identification method and system based on image identification

Country Status (1)

Country Link
CN (1) CN114022717A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115471747A (en) * 2022-08-30 2022-12-13 广东省农业科学院环境园艺研究所 AI (artificial intelligence) rapid identification method for camellia diseases and insect pests and physiological diseases and application

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115471747A (en) * 2022-08-30 2022-12-13 广东省农业科学院环境园艺研究所 AI (artificial intelligence) rapid identification method for camellia diseases and insect pests and physiological diseases and application

Similar Documents

Publication Publication Date Title
Palacios et al. Automated grapevine flower detection and quantification method based on computer vision and deep learning from on-the-go imaging using a mobile sensing platform under field conditions
CN110926430B (en) Air-ground integrated mangrove forest monitoring system and control method
CN109086826B (en) Wheat drought identification method based on image deep learning
CN110046631A (en) System and method for inferring the variation of time-space image automatically
Sarkate et al. Application of computer vision and color image segmentation for yield prediction precision
Jafari et al. Weed detection in sugar beet fields using machine vision
CN112733958B (en) Greenhouse ozone concentration control method and system
US11880981B2 (en) Method and system for leaf age estimation based on morphological features extracted from segmented leaves
CN111832448A (en) Disease identification method and system for grape orchard
Pereira et al. Pixel-based leaf segmentation from natural vineyard images using color model and threshold techniques
Zhou et al. Research on cucumber downy mildew detection system based on SVM classification algorithm
CN115292334A (en) Intelligent planting method and system based on vision, electronic equipment and storage medium
Hacking et al. Vineyard yield estimation using 2-D proximal sensing: A multitemporal approach
CN115620151A (en) Method and device for identifying phenological period, electronic equipment and storage medium
CN114022717A (en) Blade disease identification method and system based on image identification
Loresco et al. Viola-jones method of marker detection for scale-invariant calculation of lettuce leaf area
CN110544237B (en) Tea-oil tree plant disease and insect pest model training method and identification method based on image analysis
CN112257730A (en) Plant pest image identification method, device, equipment and storage medium
CN116524279A (en) Artificial intelligent image recognition crop growth condition analysis method for digital agriculture
Miranda et al. Pest identification using image processing techniques in detecting image pattern through neural network
CN110135481B (en) Crop lesion detection method and detection device
Rony et al. BottleNet18: Deep Learning-Based Bottle Gourd Leaf Disease Classification
Santhosh et al. Detection of leaf diseases and classifying them using multiclass SVM
Godse et al. Detecting jute plant disease using image processing and machine learning
CN116311232B (en) Imperfect grain type identification method and device and computer equipment

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