CN113159060B - Crop pest detection method and system - Google Patents
Crop pest detection method and system Download PDFInfo
- Publication number
- CN113159060B CN113159060B CN202110199295.XA CN202110199295A CN113159060B CN 113159060 B CN113159060 B CN 113159060B CN 202110199295 A CN202110199295 A CN 202110199295A CN 113159060 B CN113159060 B CN 113159060B
- Authority
- CN
- China
- Prior art keywords
- network
- pest
- layer
- training
- crop
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- 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
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Catching Or Destruction (AREA)
Abstract
The invention provides a crop pest detection method and a crop pest detection system, wherein the method comprises the following steps: s1, acquiring and processing crop pest images to be detected; s2, extracting features of the crop pest images to be detected, extracting feature parameters of the pests, and obtaining classification features according to the feature parameters of the pests; s3, inputting the classification features into a trained deep learning detection model, and identifying the insect pest type by the deep learning detection model; s4, judging whether the type of the insect pest meets the preset alarm type; if yes, executing step S5; s5, giving an alarm, and outputting the name of the insect pest and corresponding prevention and treatment measures. The invention can accurately identify and detect the insect damage of crops, and provides corresponding insect pest control measures to effectively control the damage of insect pests to crops, thereby improving the yield of agricultural products.
Description
Technical Field
The invention relates to the field of acquisition and processing of computer vision images, in particular to a crop pest detection method and system.
Background
The control principle of crop insect pest is 'prevention as main and comprehensive control'. According to the occurrence rule of insect pests, the weak links of the growth and development and the key moment of prevention and control are grasped, an effective and feasible method is adopted, and the insect pests are effectively controlled before a large amount of insect pests occur or cause damage, so that the insect pests cannot occur or spread on a large scale, and the crops are protected from damage. Meanwhile, the application of the cultivation technology is enhanced, and according to the rules of occurrence and development of the plant diseases and insect pests, biological, physical, chemical and other control measures are reasonably coordinated and applied according to time and local conditions, so that conditions unfavorable for the occurrence and damage of the plant diseases and insect pests are created, the aim of economically, safely and effectively controlling the occurrence of the plant diseases and insect pests is fulfilled, and the damage caused by the plant diseases and insect pests is reduced to the minimum level.
The basic agricultural workers are often overwhelmed by the lack of specialized knowledge in meeting the insect pests, and the detection of the insect pests at present mainly depends on human eye observation, including detection, consulting and comparison of insect pest patterns and consulting related specialists through the existing experience. For common insect pest types, basic-level farmers can directly distinguish the insect pest types, but some insect pest types are quite special, and have different shape characteristics in different growth periods, and the like, if the insect pest types are judged only by comparing the patterns according to the word description of the insect pests, artificial misjudgment is often caused. In the face of insect pest, the insect pest cannot be treated timely or wrongly.
In view of the foregoing, there is an urgent need in the industry to develop a method or system that can scientifically and efficiently detect the type and control measures of crop pests.
Disclosure of Invention
The present invention is directed to solving the problems of the prior art or related art. Therefore, one object of the present invention is to provide a method and a system for crop pest detection.
A crop pest detection method comprising: s1, acquiring and processing crop pest images to be detected; s2, extracting features of the crop pest images to be detected, extracting feature parameters of the pests, and obtaining classification features according to the feature parameters of the pests; s3, inputting the classification features into a trained deep learning detection model, and identifying the insect pest type by the deep learning detection model; s4, judging whether the type of the insect pest meets the preset alarm type; if yes, executing step S5; s5, giving an alarm, and outputting the name of the insect pest and corresponding prevention and treatment measures. The invention can accurately identify and detect the insect damage of crops, and provides corresponding insect pest control measures to effectively control the damage of insect pests to crops, thereby improving the yield of agricultural products.
Preferably, processing the crop pest image to be detected includes: and converting the acquired insect pest image to be detected into RGB image data and storing the RGB image data.
Preferably, the pest characteristic parameters include color, area, lines and texture.
Preferably, training of the deep learning detection model includes: acquiring and screening images related to insect pests through a network, and acquiring crop insect pest images through an image acquisition subsystem; randomly mixing the screened insect pest images with the insect pest images acquired by the image acquisition subsystem to obtain a data set; dividing the data set into a training set and a testing set; the training set is used for training the deep learning detection model, and the testing set is used for evaluating the generalization capability of the trained deep learning detection model; improving a classical deep learning one-stage detection network ssd, and providing a RESssd network; the RESssd network is to change the backbone network vgg of the ssd network into a resnet50, and extract the conv4 layer of the resnet50 and the front network layer before the conv4 layer; and 5 feature layers are added in addition; the structure of each characteristic layer is composed of 6 parts, namely a convolution layer, a Batch Normalization layer, a Relu layer, a convolution layer, a BN layer and a Relu layer; of the first 3 additional feature layers, the stride is 2, the offset is 1, and the last 2 additional feature layers, the stride is 1, the offset is 0. The BN structure can make the data more balanced, and is beneficial to extracting the characteristics. Training the RESssd network by using a training set, training the RESssd network for K rounds in total, testing the map of the deep learning detection model with the test set from the io-0.5 to the io-0.95 after each round training, and taking a network weight file corresponding to the model with the best map effect as a final model; k is more than or equal to 100.
Preferably, the alerting includes alerting the user by mail or short message.
A crop pest detection system comprising: the image acquisition subsystem, the image processing device and the front-end alarm display device are sequentially connected; the image acquisition subsystem is arranged on a crop growing field;
the image acquisition subsystem is used for acquiring and processing crop insect pest images;
the image processing device is used for analyzing and extracting characteristics according to the collected crop pest images and identifying pests according to the characteristics;
and the front-end alarm display device is used for displaying the identification result in real time and sending out an alarm according to the identification result.
Preferably, the image acquisition subsystem comprises a support column, a controller, a network camera, a light source for inducing insect pests and a carrying platform for carrying the induced insect pests; the lower extreme of carrying the thing platform sets up the support, and webcam, light source are all fixed through the pillar, and the light source is located carrying the top of thing platform, and webcam aims at carrying the thing platform, and webcam, controller, image processing device connect gradually.
Preferably, the image acquisition subsystem further comprises a solar panel for powering the light source, the solar panel being fixed to the top end of the post.
Preferably, the image processing device is a server, and the server is internally provided with a trained deep learning detection model.
Compared with the prior art, the invention has the following advantages and beneficial effects:
the present invention preselects that, starting from the collection of a dataset, the dataset has 2 sources: and (3) crawling insect pest images from the Internet by using a web crawler, collecting insect pest data images by using an image collecting subsystem, and training an insect pest detection model with higher map and stronger generalization capability by adopting a deep learning related algorithm to the well-organized data set. And then a set of systems integrating image acquisition, image detection and early warning (alarming) are developed by taking the insect pest detection model (training deep learning detection model) as a core. The method can automatically, real-timely and efficiently acquire the type and the prevention and control measures of insect pests, and has important practical significance for promoting the modern development of agriculture. In addition, the invention has the advantages and beneficial effects that:
1. in the process of training the deep learning detection model, the improved RESssd target detection network is trained by using the training set, and 5 prediction feature layers are additionally added to the RESssd network, so that the imaged features of each feature are different as much as possible, the detection effect on a small target is very obvious, and the robustness and generalization capability of the deep learning recognition model are improved.
2. According to the invention, through automatic acquisition of insect pest images, the computer vision recognition and CNN convolutional neural network model are integrated into the detection system, so that the bandwidth occupation of the images is reduced, the network resources are optimized, CNN convolutional neural network model detection can be rapidly performed, and the detection efficiency is greatly improved.
Drawings
Fig. 1 is a block diagram of a crop pest detection system of the present embodiment.
Fig. 2 is a schematic flow chart of the crop pest detection method of the present embodiment.
Fig. 3 is a ressd network configuration diagram of the present embodiment.
Detailed Description
For a better understanding of the technical solution of the present invention, examples provided by the present invention are described in detail below with reference to the accompanying drawings, but embodiments of the present invention are not limited thereto.
Referring to fig. 1, a crop pest detection system comprising: the image acquisition subsystem, the image processing device and the front-end alarm display device 9 are connected in sequence; the image acquisition subsystem is arranged on a crop growing field; the image acquisition subsystem is used for acquiring and processing crop insect pest images; the image processing device is used for analyzing and extracting characteristics according to the collected crop pest images and identifying pests according to the characteristics; the front-end alarm display device 9 is used for displaying the identification result in real time and sending out an alarm according to the identification result.
In this embodiment, the image acquisition subsystem comprises a support column 1, a controller 7, a network camera 3, a light source 4 for inducing insect pests and a carrying platform 5 for carrying the induced insect pests; the lower extreme of year thing platform 5 sets up support 6, and webcam 3, light source 4 are all fixed through pillar 1, and light source 4 is located the top of year thing platform 5, and webcam 3 aims at year thing platform 5, and webcam 3, controller 7, image processing apparatus connect gradually. The image acquisition subsystem further comprises a solar panel 2 for powering the light source 4, the solar panel 2 being fixed to the top end of the column 1.
In this embodiment, the image processing apparatus is a server 8, and a trained deep learning detection model is built in the server 8. The front-end alarm display device 9 is a PC.
The crop pest detection system can accurately detect pests and provide corresponding pest control measures, effectively control damage of the pests to crops, and comprises the following specific processes:
the network camera 3 collects data, the data are stored in the controller 7, meanwhile, the controller 7 transmits the data to the server background, the server background decodes the data, the deep learning model is called to detect the resolved image, a real-time detection result is displayed on the front-end alarm display device 9, whether the detection result meets the preset alarm category is judged, if yes, the front-end alarm display device 9 pushes the result to a user in a mode of short message, mail and the like, and the user can conduct accurate control according to the specific content of pushing. If not, the detection is continued.
In this embodiment, the network camera 3 collects the data stream, decodes the data stream into a frame-by-frame image, and finally parses the data stream into pixel data in RGB format (RGB format is a method for encoding colors, which is collectively called "color space" or "color gamut"), and transmits and stores the pixel data on a server, and at this time, a trained network model is loaded to monitor the data in real time, and at the same time, the detected implementation effect is displayed on the front end. The user can log in the website of the early warning system through the mobile phone end or the pc end, so that the implementation effect can be checked.
As shown in fig. 2, a crop pest detection method includes the steps of:
s1, acquiring and processing an image of crop insect damage to be detected; specifically, insect pest is induced to an image acquisition subsystem arranged on a crop growing field, and the image acquisition subsystem acquires an image of the insect pest of the crop to be detected. Processing the crop pest image to be detected includes: and converting the acquired insect pest image to be detected into RGB image data and storing the RGB image data. After conversion, a true color original pest growth image of 24-bit R, G, B color space is obtained.
S2, extracting features of the crop pest images to be detected, extracting feature parameters of the pests, and obtaining classification features according to the feature parameters of the pests;
s3, inputting the classification features into a trained deep learning detection model, and identifying the insect pest type by the deep learning detection model; training of a deep learning detection model is carried out without constructing a data set, and the construction of the pest data set comprises the following steps:
and (5) compiling a web crawler to crawl images related to insect pests and screening.
And the image acquisition subsystem is used for acquiring crop pest images, and the resolution ratio of the acquired pest images is adjusted so as to reduce the storage space and improve the model training speed. The images of the insect pests screened after the crawler and the images collected by the image device are taken as data sets. For training a deep learning network model.
After the pest data set is constructed, the pest data set is divided into a training set and a testing set, wherein the proportion is 8:2 (four groups of experiments are carried out, and the dividing proportion of the training set and the testing set is 9:1, 8:2, 7.5:2.5 and 7:3 respectively, wherein the 8:2 effect is the best).
As shown in fig. 3, the proposed RESssd network structure based on the classical one-stage deep learning target detection network ssd includes the following specific ideas and steps:
the backbone network vgg of the ssd network is extracted and replaced by the resnet50, while Conv4 of the resnet50 and its preceding layers are extracted and the rest of the structure is discarded, after which an additional 5 prediction feature layers are added.
A BN layer is added to each prediction feature layer, and the BN layer is actually normalized to the output of each layer, i.e. the input of each next layer. The loss function space of the neural network is smoother, and the robustness of the system is improved.
The Conv4 layer of the resnet50 in fig. 3 is a first layer prediction feature layer (FeatureMap 1), the first residual block1 of the Conv4 layer is changed for layer fixing before the Conv4 layer, the original step size of the block1 is set to be 1 (steps s=2 of the second convolution kernels 3x3 and 256 of the first residual block1 of the Conv4 layer of the original resnet50, steps s=2 of 1x1 and 256 of the shortcut convolution), and the purpose of setting s=1 is to obtain the same output size as the Conv 3. The trend of halving the picture size is reserved as far as possible in the shallow characteristic network, and rich and extractable information is reserved for the characteristic network in the rear deep layer.
The outputs of the additional 5 feature layers correspond to FeatureMap2 to FeatureMap6, respectively. The first two prediction feature layers are shallow layers, more information is reserved, relatively smaller targets are detected, and accordingly the network abstraction degree is deepened, and relatively larger targets are predicted towards the later network feature layers.
The deep learning recognition model adopts a PyTorch deep learning framework, and the hardware equipment is configured as follows: CPU adopts
Core TM i5-8400 CPU@2.80GHz ×6, 16GB memory, geForce GTX 1080Ti memory 11G for GPU, 535.21 NVIDIA drive version, 9.0.176,CUDNN 7.0.5 neural network acceleration library for CUDA version, linux Ubuntu18.04 LTS for operation system, and 500G mechanical hard disk for storage.
Each time the image is transmitted, the size of the image is a RGB color map of 300x300x3 channels; the momentum factor is set to 0.9; the attenuation coefficient is set to 0.0005; the learning rate is initially set to 0.003; training was stopped for 500 batches, with a training strategy that was 10-fold reduced for every 50 batches of learning rate to obtain less loss. The recognition accuracy of the RESssd network and the basic network ssd proposed by the invention is shown in table 1:
TABLE 1
Table 1 shows that: the improved algorithm has better effect than the classical ssd network, and the RESssd is 4.4 percent higher than the ssd. The RESssd network structure is shown in fig. 2:
s4, judging whether the type of the insect pest meets the preset alarm type; if yes, executing step S5;
s5, giving an alarm, and outputting the name of the insect pest and corresponding prevention and treatment measures.
By the embodiment of the invention, the pest control is effectively detected, identified and timely reminded to relevant personnel, and the problems of crop yield reduction, agricultural product quality reduction, economic loss and the like caused by the pest are avoided.
The above examples are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above examples, and any other changes, modifications, substitutions, combinations, and simplifications that do not depart from the spirit and principle of the present invention should be made in the equivalent manner, and the embodiments are included in the protection scope of the present invention.
Claims (6)
1. A method for crop pest detection comprising:
s1, acquiring and processing crop pest images to be detected, wherein the processing of the crop pest images to be detected comprises the following steps: converting the acquired insect pest image to be detected into RGB image data and storing the RGB image data;
s2, extracting features of the crop pest images to be detected, extracting feature parameters of the pests, and obtaining classification features according to the feature parameters of the pests;
s3, inputting the classification features into a trained deep learning detection model, and identifying the insect pest type by the deep learning detection model;
training of the deep learning detection model includes:
acquiring and screening images related to insect pests through a network, and acquiring crop insect pest images through an image acquisition subsystem; randomly mixing the screened insect pest images with the insect pest images acquired by the image acquisition subsystem to obtain a data set;
dividing the data set into a training set and a testing set;
improving a classical deep learning one-stage detection network ssd, and providing a RESssd network; the RESssd network is to change the backbone network vgg of the ssd network into a resnet50, and extract the conv4 layer of the resnet50 and the front network layer before the conv4 layer; and 5 feature layers are added in addition; the structure of each characteristic layer is a convolution layer, a Batch Normalization layer, a Relu layer, a convolution layer, a BN layer and a Relu layer respectively; the first 3 additional added feature layers have a stride of 2, a bias of 1, and the second 2 additional added feature layers have a stride of 1, a bias of 0;
training the RESssd network by using a training set, training the RESssd network for K rounds in total, testing the map of the deep learning detection model with the test set from the io-0.5 to the io-0.95 after each round training, and taking a network weight file corresponding to the model with the best map effect as a final model; k is more than or equal to 100;
training the RESssd network using a training set, comprising: each time the size of the image is 300x300x3 channel RGB color map, the attenuation coefficient is set to 0.0005, and the learning rate is initially set to 0.003;
s4, judging whether the type of the insect pest meets the preset alarm type; if yes, executing step S5;
s5, giving an alarm, and outputting the name of the insect pest and corresponding prevention and treatment measures.
2. A method of crop pest detection as claimed in claim 1, wherein the pest characteristic parameters include colour, area, line and texture.
3. The method of claim 1, wherein alerting comprises alerting the user by mail or short message alert.
4. A crop pest detection system, comprising: the image acquisition subsystem, the image processing device and the front-end alarm display device are sequentially connected; the image acquisition subsystem is arranged on a crop growing field;
the image acquisition subsystem is used for acquiring and processing crop insect pest images;
the image processing device is used for analyzing and extracting characteristics according to the collected crop pest images and identifying pests according to the characteristics; processing the crop pest image to be detected includes: converting the acquired insect pest image to be detected into RGB image data and storing the RGB image data;
the image processing device is a server, and a trained deep learning detection model is built in the server;
training of the deep learning detection model includes:
acquiring and screening images related to insect pests through a network, and acquiring crop insect pest images through an image acquisition subsystem; randomly mixing the screened insect pest images with the insect pest images acquired by the image acquisition subsystem to obtain a data set;
dividing the data set into a training set and a testing set;
improving a classical deep learning one-stage detection network ssd, and providing a RESssd network; the RESssd network is to change the backbone network vgg of the ssd network into a resnet50, and extract the conv4 layer of the resnet50 and the front network layer before the conv4 layer; and 5 feature layers are added in addition; the structure of each characteristic layer is a convolution layer, a Batch Normalization layer, a Relu layer, a convolution layer, a BN layer and a Relu layer respectively; the first 3 additional added feature layers have a stride of 2, a bias of 1, and the second 2 additional added feature layers have a stride of 1, a bias of 0;
training the RESssd network by using a training set, training the RESssd network for K rounds in total, testing the map of the deep learning detection model with the test set from the io-0.5 to the io-0.95 after each round training, and taking a network weight file corresponding to the model with the best map effect as a final model; k is more than or equal to 100;
training the RESssd network using a training set, comprising: each time the size of the image is 300x300x3 channel RGB color map, the attenuation coefficient is set to 0.0005, and the learning rate is initially set to 0.003;
and the front-end alarm display device is used for displaying the identification result in real time and sending out an alarm according to the identification result.
5. The crop pest detection system of claim 4, wherein the image acquisition subsystem includes a support column, a controller, a webcam, a light source for inducing pests, and a carrier platform for carrying the induced pests; the lower extreme of carrying the thing platform sets up the support, and webcam, light source are all fixed through the pillar, and the light source is located carrying the top of thing platform, and webcam aims at carrying the thing platform, and webcam, controller, image processing device connect gradually.
6. The crop pest detection system of claim 5, wherein the image acquisition subsystem further includes a solar panel for powering the light source, the solar panel being secured to the top end of the post.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110199295.XA CN113159060B (en) | 2021-02-23 | 2021-02-23 | Crop pest detection method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110199295.XA CN113159060B (en) | 2021-02-23 | 2021-02-23 | Crop pest detection method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113159060A CN113159060A (en) | 2021-07-23 |
| CN113159060B true CN113159060B (en) | 2023-04-28 |
Family
ID=76883247
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110199295.XA Active CN113159060B (en) | 2021-02-23 | 2021-02-23 | Crop pest detection method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113159060B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109919239A (en) * | 2019-03-15 | 2019-06-21 | 尹显东 | A kind of diseases and pests of agronomic crop intelligent detecting method based on deep learning |
| CN111914951A (en) * | 2020-08-21 | 2020-11-10 | 安徽省农业科学院农业经济与信息研究所 | Crop pest intelligent diagnosis system and method based on image real-time identification |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109472252A (en) * | 2018-12-28 | 2019-03-15 | 华南农业大学 | A kind of field crops insect pest automatic identification and job management system |
| CN111783685A (en) * | 2020-05-08 | 2020-10-16 | 西安建筑科技大学 | Target detection improved algorithm based on single-stage network model |
| CN111611889B (en) * | 2020-05-12 | 2023-04-18 | 安徽大学 | Miniature insect pest recognition device in farmland based on improved convolutional neural network |
-
2021
- 2021-02-23 CN CN202110199295.XA patent/CN113159060B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109919239A (en) * | 2019-03-15 | 2019-06-21 | 尹显东 | A kind of diseases and pests of agronomic crop intelligent detecting method based on deep learning |
| CN111914951A (en) * | 2020-08-21 | 2020-11-10 | 安徽省农业科学院农业经济与信息研究所 | Crop pest intelligent diagnosis system and method based on image real-time identification |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113159060A (en) | 2021-07-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10430533B2 (en) | Method for automatic behavioral phenotyping | |
| CN109964235A (en) | For carrying out the prediction model of vision sorter to insect | |
| Dakos | Identifying best-indicator species for abrupt transitions in multispecies communities | |
| CN111048214A (en) | Early warning method and device for spreading situation of foreign livestock and poultry epidemic diseases | |
| Couret et al. | Delimiting cryptic morphological variation among human malaria vector species using convolutional neural networks | |
| CN116030343A (en) | Crop pest monitoring system based on machine vision identification | |
| Singh et al. | Performance evaluation of plant leaf disease detection using deep learning models | |
| Martinez et al. | A machine learning model for the diagnosis of coffee diseases | |
| Ranjan et al. | MortCam: An Artificial Intelligence-aided fish mortality detection and alert system for recirculating aquaculture | |
| Guo et al. | Pigeon cleaning behavior detection algorithm based on light-weight network | |
| CN114898405A (en) | Portable broiler chicken abnormity monitoring system based on edge calculation | |
| Dandekar et al. | Weed Plant Detection from Agricultural Field Images using YOLOv3 Algorithm | |
| Doan | An efficient system for real-time mobile smart device-based insect detection | |
| CN113159060B (en) | Crop pest detection method and system | |
| US11069197B2 (en) | Method and system of drawing random numbers via sensors for gaming applications | |
| Kalfas et al. | Towards in-field insect monitoring based on wingbeat signals: The importance of practice oriented validation strategies | |
| CN112507770B (en) | Rice disease and insect pest identification method and system | |
| CN113256567B (en) | Banana leaf area index detection method and system | |
| Santo et al. | Potato skin defect detection and classification through image processing | |
| Hall et al. | Automated, non-invasive Varroa mite detection by vibrational measurements of gait combined with machine learning | |
| Petrellis | A Windows Phone Application for Plant Disease Diagnosis. | |
| Kumari et al. | Plant leaf disease identification using machine learning | |
| Wanninayake et al. | IoT-Enabled Smart Solution for Rice Disease Detection, Yield Prediction, and Remediation | |
| Sharma et al. | Cotton Crop Disease Detection System Using Deep Learning Approach | |
| Rozikin et al. | Benchmarking the LGBM, Random Forest, and XGBoost Models Based on Accuracy in Classifying Melon Leaf Disease |
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 |