CN112381156B - Chinese medicinal material pest and disease damage detection system based on image recognition - Google Patents
Chinese medicinal material pest and disease damage detection system based on image recognition Download PDFInfo
- Publication number
- CN112381156B CN112381156B CN202011294151.4A CN202011294151A CN112381156B CN 112381156 B CN112381156 B CN 112381156B CN 202011294151 A CN202011294151 A CN 202011294151A CN 112381156 B CN112381156 B CN 112381156B
- Authority
- CN
- China
- Prior art keywords
- pest
- disease
- identification
- module
- image
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Catching Or Destruction (AREA)
Abstract
The invention relates to the technical field of image processing and deep learning, and particularly discloses a traditional Chinese medicine disease and insect pest detection system based on image recognition, which comprises the following components: the database is used for storing various pest control methods; the image acquisition module is used for acquiring a pest and disease damage image; the interactive module is used for generating an online pest and disease damage identification request according to the pest and disease damage image and sending the online pest and disease damage identification request to the server; the pest and disease identification system is used for analyzing the received pest and disease online identification request to obtain a pest and disease image, and performing pest and disease identification on the analyzed pest and disease image through a pest and disease identification model to obtain a pest and disease identification result; the prevention and control pushing module is used for inquiring the prevention and control method corresponding to the pest and disease damage from the database according to the pest and disease damage identification result and feeding the inquired prevention and control method back to the user terminal. By adopting the technical scheme of the invention, the plant diseases and insect pests can be quickly and efficiently identified and prevented.
Description
Technical Field
The invention relates to the technical field of image processing and deep learning, in particular to a traditional Chinese medicine disease and insect pest detection system based on image recognition.
Background
The prevention and control of the plant diseases and insect pests of the traditional Chinese medicinal materials are restricted by factors such as multiple types of plant diseases and insect pests, unstable occurrence, few professional technicians, weak basic research, no professional technical popularization system and the like of the traditional Chinese medicinal materials, and the traditional Chinese medicinal materials are the weakest link in the production of the traditional Chinese medicinal materials accepted in the industry. The industry has widely recognized that: pests such as sick cordyceps, rats and the like are one of the main factors influencing the production and the quality of the traditional Chinese medicinal materials, and almost all planting bases have the problems of diseases and pests. The yield of medicinal materials caused by plant diseases and insect pests is reduced by 50%, and the annual economic loss is about 400 hundred million yuan. However, the phenomenon of pesticide abuse is serious, pesticide residues of traditional Chinese medicinal materials exceed the standard, and the phenomenon of traditional Chinese medicine becomes one of the social hotspots due to the lack of a scientific and effective prevention and treatment method for plant diseases and insect pests for planting personnel.
Traditional Chinese medicinal materials are special agricultural products, like agricultural products, the yield needs to be investigated, but the quality (effective component content) is the guarantee of the clinical effectiveness, so that the planting technology of the traditional Chinese medicinal materials, particularly the pest control technology of the traditional Chinese medicinal materials, needs specialized technical personnel to develop specialized research. Besides strengthening the basic research of the traditional Chinese medicine disease and pest prevention and control technology, the key for solving the problem is to construct a professional and efficient popularization system of the traditional Chinese medicine disease and pest prevention and control technology. However, the agricultural technology popularization in China is a part of administrative functions of an agricultural director, is mainly responsible for the agricultural technology popularization of grain crops, almost no personnel specialized in the agricultural technology popularization of traditional Chinese medicinal materials exist, and currently, the number of personnel specialized in the technical research of the pest control of the traditional Chinese medicinal materials in China is less than 100. How to rapidly and efficiently popularize the research results of the pest control technology of the traditional Chinese medicinal materials into the vast market of planting the traditional Chinese medicinal materials is a technical problem to be solved urgently at present.
Disclosure of Invention
In order to solve the technical problem of how to quickly and efficiently identify, prevent and treat plant diseases and insect pests, the invention provides a traditional Chinese medicine plant disease and insect pest detection system based on image identification.
The basic scheme of the invention is as follows:
chinese-medicinal material plant diseases and insect pests detecting system based on image recognition, including user terminal and server, user terminal includes image acquisition module and interactive module, and the server includes database, plant diseases and insect pests identification system and prevention and cure propelling movement module, wherein:
the database is used for storing various pest control methods;
the image acquisition module is used for acquiring a pest image;
the interactive module is used for generating an online pest and disease damage identification request according to the pest and disease damage image and sending the online pest and disease damage identification request to the server;
the pest and disease identification system is used for analyzing the received pest and disease online identification request to obtain a pest and disease image, and performing pest and disease identification on the analyzed pest and disease image through a pest and disease identification model to obtain a pest and disease identification result;
the prevention and control pushing module is used for inquiring the prevention and control method corresponding to the pest and disease damage from the database according to the pest and disease damage identification result and feeding the inquired prevention and control method back to the user terminal.
The basic principle of the scheme is as follows: the image acquisition module acquires a disease and insect pest image needing disease and insect pest identification, the interaction module and the server send the disease and insect pest image to the server, and a disease and insect pest identification system of the server carries out disease and insect pest identification on the disease and insect pest image through a disease and insect pest identification model to obtain a disease and insect pest identification result. And then inquiring a corresponding prevention and control method from the database according to the pest and disease identification result, and feeding back to the user terminal.
Compare in current mode of relying on artifical plant diseases and insect pests discernment, have following advantage:
a user only needs to collect a pest image to obtain a pest identification result, and pest identification is faster and more efficient.
The user does not need to have the experience of preventing and controlling diseases and pests, the scheme can automatically feed back the corresponding prevention and control method according to the recognition result of the diseases and pests, and the acquisition of the disease and pest prevention and control method is simpler.
Furthermore, the user terminal also comprises a positioning module, the server also comprises an epidemic situation statistic module,
the positioning module is used for acquiring position information when a pest and disease damage image is acquired;
the interactive module is also used for generating an online pest and disease damage identification request according to the pest and disease damage image and the position information and sending the online pest and disease damage identification request to the server;
the epidemic situation statistic module is used for counting the pest and disease identification result obtained by identification, the correspondingly received pest and disease image and the position information and generating epidemic situation statistic information.
Has the beneficial effects that: by counting the pest and disease identification results of each area, the epidemic situation of the pests and diseases can be analyzed, the pest and disease conditions of each area can be mastered, the pests and diseases can be conveniently prevented and controlled, and the pests and diseases can be effectively treated.
Furthermore, the user terminal also comprises an identification mode selection module, and the server also comprises an offline system deployment module;
the identification mode selection module is used for generating an identification instruction according to the fed-back identification mode, and the identification instruction comprises an online identification instruction and an offline identification instruction;
the interactive terminal is also used for analyzing the identification instruction, and if the identification instruction is analyzed to be an online identification instruction, the acquired pest and disease damage image is sent to the server; if the command is analyzed to be an offline identification command, whether a pest and disease identification system exists in the user terminal is searched, and if the command does not exist, an offline identification request is sent to the server;
and the offline system deployment module is used for deploying the pest and disease damage identification system at the user terminal when receiving the offline identification request.
Has the advantages that: the offline pest and disease identification of the user terminal is realized by deploying the pest and disease identification system at the user terminal.
Furthermore, the user terminal also comprises a performance evaluation module,
the offline system deployment module is also used for sending a hardware performance evaluation instruction to the user terminal when receiving the offline identification request;
the performance evaluation module is used for evaluating the hardware performance of the user terminal when receiving the hardware performance evaluation instruction to obtain a hardware performance evaluation result and sending the hardware performance evaluation result to the server, wherein the hardware performance evaluation result comprises high-performance hardware and low-performance hardware;
the offline system deployment module is also used for analyzing the received hardware performance evaluation result, and directly deploying the pest and disease damage identification system at the user terminal if the analysis result is high-performance hardware; and if the analysis result is low-performance hardware, deploying the compressed pest and disease damage identification system at the user terminal.
Has the advantages that: the pest and disease identification model based on the deep neural network has larger requirements on memory and calculation power, so that a user terminal (such as a mobile phone) with low-performance hardware possibly cannot meet the requirements of the pest and disease identification model. By compressing the pest and disease identification system, the demand of the pest and disease identification system on hardware resources of the user terminal is reduced, so that the user terminal with low-performance hardware can also realize off-line pest and disease identification.
Further, the offline system deployment module compresses the pest and disease identification system by adopting a deep neural network compression technology.
Has the advantages that: the deep neural network compression technology is adopted, the neural network structure is simplified, the requirements of user terminals (such as mobile phones) on hardware resources are reduced, and the problems that the pest and disease identification system has high requirements on memory and computing power, and the hardware resources of some user terminals cannot meet the requirements are solved.
Further, the offline system deployment module is further configured to check whether the pest identification system of the user terminal is the current latest version of pest identification system after detecting that the pest identification system of the server is updated, and if not, update and update the pest identification system of the user terminal according to the latest version of pest identification system of the server.
Has the beneficial effects that: and the disease and insect identification system of the user terminal is kept to be the latest version, so that the offline disease and insect identification result of the user terminal is the same as the online disease and insect identification result on the server, and the accuracy of the offline disease and insect identification result is ensured.
Further, the performance evaluation module is also used for evaluating the network connection state of the user terminal, and if the evaluation result of the network connection state is that the network is not available, an offline identification instruction is generated; and if the evaluation result of the network connection state is that the network is good, generating an online identification instruction.
Has the beneficial effects that: on-line pest and disease identification needs to upload pest and disease images to a server, so when network connection is not smooth, the efficiency of off-line pest and disease identification is superior to that of on-line pest and disease identification, and therefore a high-efficiency pest and disease identification mode is intelligently selected by evaluating the network connection state.
The server further comprises a pest sample acquisition module, and the pest sample acquisition module is used for learning the artificially marked pest image sample by adopting an antagonistic generation network picture generation technology to generate a pest training sample.
Has the advantages that: the manual marking of the pest image sample requires professional knowledge for distinguishing the pest, and the manual marking of the pest image sample is labor-intensive, so that the marking cost is high. By learning on the artificially marked pest and disease image samples by adopting a confrontation generation network picture generation technology, more high-quality training samples can be generated, so that the problems of high marking cost and insufficient pest and disease image samples of the pest and disease image samples are solved.
Further, the pest and disease identification system is also used for carrying out target detection on the pest and disease image to obtain the pest and disease image to be identified.
Has the advantages that: the scanning detection of the image is time-consuming, and the efficiency of the image scanning detection seriously influences the system performance. By carrying out target detection on the image, the candidate area can be effectively reduced, so that the image scanning detection speed is improved.
Further, the server further comprises a disease and pest model generation module, wherein the disease and pest model generation module is used for generating a neural network model by training the disease and pest training samples through a model fusing a residual error net ResNet and an IceptionNet.
Has the advantages that: influenced by the pest training samples, the defects of difficult convergence of the model, low generalization of the model and the like can be presented in the characteristic learning process of the model. According to the scheme, the model with the residual error net ResNet and IceptionNet integrated is adopted, the depth network structure can be optimized according to the characteristics of various plant diseases and insect pests, the number of only parameters of the model is reduced, and the training difficulty of the network is reduced.
Drawings
FIG. 1 is a logic block diagram of a first embodiment of a Chinese medicinal material pest detection system based on image recognition;
FIG. 2 is a logic block diagram of a second embodiment of a Chinese medicinal material pest detection system based on image recognition;
FIG. 3 is a logic block diagram of a third embodiment of a Chinese medicinal material pest detection system based on image recognition;
FIG. 4 is a logic block diagram of a fourth embodiment of a Chinese medicinal material pest detection system based on image recognition;
fig. 5 is a logic block diagram of a fifth embodiment of a traditional Chinese medicine pest detection system based on image recognition.
Detailed Description
The following is further detailed by way of specific embodiments:
example one
As shown in fig. 1, the system for detecting plant diseases and insect pests of traditional Chinese medicinal materials based on image recognition comprises a user terminal and a server, wherein in this embodiment, the user terminal is a smart phone equipped with a system, and the server is a Web server.
The server comprises a pest sample acquisition module and a pest model generation module; wherein:
and the pest and disease sample acquisition module is used for learning the artificially marked pest and disease image sample by adopting a confrontation generation network picture generation technology to generate a pest and disease training sample.
And the disease and insect pest model generation module is used for training the disease and insect pest training sample to generate a disease and insect pest identification model by adopting a model in which a residual error net ResNet and an IceptionNet are fused. According to the characteristics of various plant diseases and insect pests, the deep neural network structure is optimized through the model, the number of free parameters of the model is reduced, and the training difficulty of the grid is reduced. Meanwhile, a data enhancement technology is introduced, and a pest model generation module performs scale change enhancement, rotation enhancement and color space enhancement on pest training samples to increase the number of pest training samples and improve the stability of the training process.
The user terminal comprises an image acquisition module, an interaction module and an identification mode selection module, the server further comprises a database, a pest and disease identification system, a prevention and control pushing module and an off-line system deployment module, wherein:
the database is used for storing various pest control methods and pest training samples.
The image acquisition module is used for acquiring a pest image.
The identification mode selection module is used for generating an identification instruction according to the fed-back identification mode, and the identification instruction comprises an online identification instruction and an offline identification instruction. In this embodiment, the user terminal displays a selectable identification mode, the user selects a required identification mode at the user terminal, and the identification mode selection module acquires the identification mode selected by the user at the user terminal.
The interactive module is used for generating an online pest and disease damage identification request according to the pest and disease damage image and sending the online pest and disease damage identification request to the server; the method comprises the following specific steps: the interactive terminal analyzes the identification instruction, and sends the acquired pest and disease damage image to a server if the identification instruction is analyzed to be an online identification instruction; and if the command is analyzed to be an offline identification command, searching whether the pest and disease damage identification system exists in the user terminal, and if not, sending an offline identification request to the server.
And the offline system deployment module is used for deploying the pest and disease damage identification system at the user terminal when receiving the offline identification request. The offline system deployment module is further used for checking whether the pest and disease identification system of the user terminal is the current latest version of pest and disease identification system after detecting that the pest and disease identification system of the server is updated, and if not, updating and upgrading the pest and disease identification system of the user terminal according to the latest version of pest and disease identification system of the server.
And the pest and disease identification system is used for analyzing the received pest and disease online identification request to obtain a pest and disease image, and performing pest and disease identification on the analyzed pest and disease image through a pest and disease identification model to obtain a pest and disease identification result. In order to improve the scanning and detecting speed of the pest and disease image, the pest and disease identification system is also used for carrying out target detection on the pest and disease image to obtain the pest and disease image to be identified. In the embodiment, the pest identification system performs target detection on the pest image by adopting a Mask R-CNN target detection model.
The prevention and control pushing module is used for inquiring the prevention and control method corresponding to the pest and disease damage from the database according to the pest and disease damage identification result and feeding the inquired prevention and control method back to the user terminal.
The specific implementation process comprises the following steps: firstly, a calibrated pest and disease damage image sample is collected manually and uploaded to a server. And the pest and disease sample acquisition module learns the artificially marked pest and disease image sample by adopting a confrontation generation network picture generation technology to generate a pest and disease training sample. And then, carrying out scale change enhancement, rotation enhancement and color space enhancement on the pest training samples of the pest model generation module to increase the number of the pest training samples, and training the pest training samples by adopting a model formed by fusing a residual error net ResNet and an IceptionNet to generate a pest identification model.
When a user needs to identify the plant diseases and insect pests, the smart phone carrying the system is used for shooting the plant disease and insect pest images needing to identify the plant diseases and insect pests, an identification mode is selected, and if the identification mode is online identification, the interaction module generates a plant disease and insect pest online identification request according to the shot plant disease and insect pest images and sends the plant disease and insect pest online identification request to the server. And the pest and disease identification system analyzes the received pest and disease identification request to obtain a pest and disease image, performs target detection on the pest and disease image obtained by analysis, reduces candidate domains without pest and disease targets, obtains a pest and disease image to be identified, and then performs pest and disease identification on the pest and disease image to be identified through a pest and disease identification model to obtain a pest and disease identification result. And the prevention and control push module inquires a prevention and control method corresponding to the plant diseases and insect pests from the database according to the plant disease and insect pest identification result and feeds the inquired prevention and control method back to the user terminal.
If the selected identification mode is offline identification, whether a pest identification system exists in the user terminal is searched, if yes, the pest identification system of the user terminal carries out target detection on the shot pest image, candidate domains without pest targets are reduced, a pest image to be identified is obtained, then pest identification is carried out on the pest image to be identified through a pest identification model, a pest identification result is obtained, and then the interactive terminal sends the pest identification result to the server. And the prevention and control push module inquires a prevention and control method corresponding to the plant diseases and insect pests from the database according to the plant disease and insect pest identification result and feeds the inquired prevention and control method back to the user terminal.
If the user terminal does not have the pest and disease identification system, the interaction module sends an offline identification request to the server, and the offline system deployment module is used for deploying the pest and disease identification system at the user terminal when receiving the offline identification request. After deployment is completed, the pest and disease identification system of the user terminal performs target detection and pest and disease identification on the shot pest and disease image to obtain a pest and disease identification result. And then the interactive terminal sends the pest and disease identification result to the server. And the prevention and control pushing module inquires a prevention and control method corresponding to the pest and disease damage from the database according to the pest and disease damage identification result and feeds the inquired prevention and control method back to the user terminal.
Example two
The difference from the first embodiment is that: as shown in fig. 2, the user terminal further includes a positioning module, the server further includes an epidemic situation statistics module,
the positioning module is used for acquiring position information when the pest and disease damage image is acquired. In this embodiment, a GPS positioning mode is adopted.
The interactive module is also used for generating an online pest and disease damage identification request according to the pest and disease damage image and the position information and sending the online pest and disease damage identification request to the server;
the epidemic situation statistic module is used for counting the pest and disease identification result obtained by identification, the correspondingly received pest and disease image and the position information and generating epidemic situation statistic information.
By counting the pest and disease conditions of each region, the epidemic situation analysis can be carried out according to the epidemic situation statistical information formed by counting, the construction of Chinese medicinal plant disease and disease damage big data and rural informatization construction are facilitated, the basic data of the Chinese medicinal plant disease and disease damage can be effectively accumulated, and the Chinese medicinal plant disease and disease damage can be predicted and prevented by combining the meteorological big data of each region.
EXAMPLE III
The difference from the first embodiment is that: as shown in fig. 3, the user terminal further includes a performance evaluation module,
the offline system deployment module is also used for sending a hardware performance evaluation instruction to the user terminal when receiving the offline identification request;
the performance evaluation module is used for evaluating the hardware performance of the user terminal when receiving the hardware performance evaluation instruction to obtain a hardware performance evaluation result and sending the hardware performance evaluation result to the server, wherein the hardware performance evaluation result comprises high-performance hardware and low-performance hardware; the performance evaluation module is also used for evaluating the network connection state of the user terminal, and generating an offline identification instruction if the evaluation result of the network connection state is that the network is not normal; and if the evaluation result of the network connection state is that the network is good, generating an online identification instruction.
The offline system deployment module is also used for analyzing the received hardware performance evaluation result, and directly deploying the pest and disease damage identification system at the user terminal if the analysis result is high-performance hardware; and if the analysis result is low-performance hardware, deploying the compressed pest and disease damage identification system at the user terminal. In this embodiment, the offline system deployment module compresses the pest identification system by using a deep neural network compression technology.
The pest identification system has higher requirements on memory and computing power, hardware resources of the user terminals are uneven, and some user terminals may not meet the hardware performance requirements of the pest identification system. Through a deep neural network compression technology, a portable pest and disease identification system is optimally designed and deployed in a user terminal, and offline automatic identification of pests and diseases is realized. Meanwhile, the Web server is adopted to provide online identification service in a Web service mode. And the two modes of off-line pest and disease damage identification and on-line pest and disease damage identification can be automatically selected according to hardware performance resources and networking states of the user terminal.
Example four
The difference from the first embodiment is that: as shown in fig. 4, the user terminal further includes a prevention and control revision module, the server further includes a prevention and control verification module, the prevention and control revision module is configured to send a prevention and control method revision request to the server, the prevention and control method revision request includes a prevention and control method to be revised and expected recovery condition information at each stage, the prevention and control verification module is configured to parse the received prevention and control method revision request to obtain the prevention and control method to be revised and the expected recovery condition information, then, according to the prevention and control method to be revised, control implementation reminding information and reminding time information are generated, and corresponding prevention and control implementation reminding information is sent to the user terminal according to the reminding time information. And after receiving the prevention implementation reminding information, the prevention revision module pops up the prevention implementation reminding information, starts the camera, shoots the pest and disease damage image as prevention implementation evidence and uploads the pest and disease damage image to the server.
The control verification module is also used for carrying out comparative analysis on control implementation evidences received at adjacent time to obtain recovery condition information of plant diseases and insect pests, carrying out comparative analysis on the recovery condition information and expected recovery condition information, if the recovery condition information is the same as the expected recovery condition information or the recovery condition is better, continuously sending corresponding control implementation reminding information to the user terminal according to the reminding time information until no next control implementation reminding information exists according to the reminding time information, and storing the control mode to be revised and the corresponding plant disease and insect pest identification result in a database in a correlation manner; if the recovery condition information does not reach the expected recovery condition information, the process of sending corresponding prevention and control implementation reminding information to the user terminal according to the reminding time information is stopped, and prevention and control mode stop reminding information is sent to the user terminal.
Some Chinese medicinal material planting users accumulate Chinese medicinal material pest and disease management experience which is possibly more effective than a prevention and treatment method given by system experts. Through the system, the users can upload more effective pest control methods, so that the more effective pest control methods are popularized. And the effective prevention method is analyzed by verifying the recovery condition of each stage of the plant diseases and insect pests in the implementation process of the prevention method to be revised.
EXAMPLE five
The difference from the first embodiment is that: as shown in the fifth drawing, the image acquisition module is a camera and a controller, and in order to worry about the problem that the traditional Chinese medicinal materials are stolen, a fixing column is arranged on the planting field for planting the traditional Chinese medicinal materials, and the camera is arranged on the fixing column. The user terminal further comprises a wind detection module, the wind detection module is arranged at the free end of the fixed column, the wind detection module is used for generating a wind signal when detecting that wind blows and sending the wind signal to the controller, and after the controller receives the wind signal, the controller controls the camera to move downwards along the fixed column until the ground surface and then controls the camera to shoot the pest and disease damage image.
The abnormal conditions of leaves generated by some pests and diseases only appear on the backs or the roots of the leaves, such as aphids, only attach the leaves to the backs or stems, and the pests and the diseases cannot be detected by conventional shooting from top to bottom. When wind blows, the blades can be blown, the back of the leaves can be exposed, and the leaf surfaces sheltered from each other can be separated, so that the back of the leaves and the pest and disease damage images of the sheltered leaves can be shot, and the detection of the pest and disease damage is more comprehensive and accurate.
In addition, a moving mechanism which can be controlled to move up and down along the fixed column is arranged on the fixed column, the camera is fixed on the moving mechanism, the control module is also used for receiving a wind signal to control the moving mechanism to move down along the fixed column, and the moving mechanism is controlled to verify that the fixed column moves up after the camera shoots the pest and disease image. The pest and disease identification system further comprises a soil identification module and a soil condition reminding module, wherein the soil identification module is used for analyzing and identifying the soil condition of the collected pest and disease images to obtain soil condition information, and the soil condition information comprises soil humidity information; the soil condition reminding module is used for generating reminding information according to the soil condition information and the pre-acquired Chinese medicinal material variety information and sending the reminding information to the user terminal.
When the wind is strong, the wind can also blow up all the leaves of the traditional Chinese medicinal materials, the roots of the traditional Chinese medicinal materials are exposed, the earth surface soil is displayed, the camera moves downwards at the moment, the picture of the soil can be shot, the soil condition information is obtained by analyzing and identifying the pest and disease damage images of the existing soil, and the watering reminding required to be watered is carried out according to the soil condition information and the pre-acquired traditional Chinese medicinal material variety information.
The foregoing is merely an example of the present invention and common general knowledge of known specific structures and features of the embodiments is not described herein in any greater detail. It should be noted that, for those skilled in the art, without departing from the structure of the present invention, several variations and modifications can be made, which should also be considered as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the utility of the patent. The scope of the claims of the present application shall be defined by the claims, and the description of the embodiments and the like in the specification shall be used to explain the contents of the claims.
Claims (10)
1. Chinese-medicinal material plant diseases and insect pests detecting system based on image recognition, its characterized in that, including user terminal and server, user terminal includes image acquisition module and interactive module, and the server includes database, plant diseases and insect pests identification system and prevention and cure propelling movement module, wherein:
the database is used for storing various pest control methods;
the image acquisition module is used for acquiring a pest and disease damage image;
the interactive module is used for generating an online pest and disease damage identification request according to the pest and disease damage image and sending the online pest and disease damage identification request to the server;
the pest and disease identification system is used for analyzing the received pest and disease online identification request to obtain a pest and disease image, and performing pest and disease identification on the analyzed pest and disease image through a pest and disease identification model to obtain a pest and disease identification result;
the prevention and control pushing module is used for inquiring the prevention and control method corresponding to the pest and disease damage from the database according to the pest and disease damage identification result and feeding the inquired prevention and control method back to the user terminal;
the image acquisition module is a camera and a controller, a fixing column is arranged on a planting field for planting the traditional Chinese medicinal materials, and the camera is arranged on the fixing column; the user terminal also comprises a wind detection module, the wind monitoring module is arranged at the free end of the fixed column and used for generating a wind signal when detecting that wind blows and sending the wind signal to the controller, and the controller is used for controlling the camera to move downwards along the fixed column to the ground surface after receiving the wind signal and controlling the camera to shoot a pest image;
the control module is also used for receiving a wind signal to control the moving mechanism to move downwards along the fixed column and controlling the moving mechanism to move upwards along the fixed column after the camera shoots a pest and disease damage image; the soil recognition module is used for analyzing and recognizing the soil condition of the collected pest and disease images to obtain soil condition information, and the soil condition information comprises soil humidity information; the soil condition reminding module is used for generating reminding information according to the soil condition information and the pre-acquired Chinese medicinal material variety information and sending the reminding information to the user terminal;
the user terminal further comprises a prevention and control revision module, the server further comprises a prevention and control verification module, the prevention and control revision module is used for sending a prevention and control method revision request to the server, the prevention and control method revision request comprises a prevention and control method to be revised and expected recovery condition information of each stage, the prevention and control verification module is used for analyzing the received prevention and control method revision request to obtain the prevention and control method to be revised and the expected recovery condition information, then prevention and control implementation reminding information and reminding time information are generated according to the prevention and control method to be revised, and corresponding prevention and control implementation reminding information is sent to the user terminal according to the reminding time information; after receiving the prevention implementation reminding information, the prevention revision module pops up the prevention implementation reminding information, starts a camera, shoots an insect disease image as a prevention implementation evidence and uploads the insect disease image to the server;
the control verification module is also used for carrying out comparative analysis on control implementation evidences received at adjacent time to obtain recovery condition information of plant diseases and insect pests, carrying out comparative analysis on the recovery condition information and expected recovery condition information, if the recovery condition information is the same as the expected recovery condition information or the recovery condition is better, continuously sending corresponding control implementation reminding information to the user terminal according to the reminding time information until no next control implementation reminding information exists according to the reminding time information, and storing the control mode to be revised and the corresponding plant disease and insect pest identification result in a database in a correlation manner; if the recovery condition information does not reach the expected recovery condition information, the process of sending corresponding prevention and control implementation reminding information to the user terminal according to the reminding time information is terminated, and prevention and control mode termination reminding information is sent to the user terminal.
2. The traditional Chinese medicine material pest and disease damage detection system based on image recognition as claimed in claim 1, wherein: the user terminal also comprises a positioning module, the server also comprises an epidemic situation statistic module,
the positioning module is used for acquiring position information when a pest and disease damage image is acquired;
the interactive module is also used for generating an online pest and disease damage identification request according to the pest and disease damage image and the position information and sending the online pest and disease damage identification request to the server;
and the epidemic situation statistics module is used for counting the pest and disease identification result obtained by identification, the correspondingly received pest and disease image and the position information and generating epidemic situation statistics information.
3. The traditional Chinese medicine material pest and disease detection system based on image recognition as claimed in claim 1, wherein: the user terminal also comprises an identification mode selection module, and the server also comprises an offline system deployment module;
the identification mode selection module is used for generating an identification instruction according to the fed-back identification mode, and the identification instruction comprises an online identification instruction and an offline identification instruction;
the interactive terminal is also used for analyzing the identification instruction, and if the identification instruction is analyzed to be an online identification instruction, the acquired pest and disease damage image is sent to the server; if the command is analyzed to be an offline identification command, whether a pest and disease identification system exists in the user terminal is searched, and if the command does not exist, an offline identification request is sent to the server;
and the offline system deployment module is used for deploying the pest and disease damage identification system at the user terminal when receiving the offline identification request.
4. The traditional Chinese medicine material pest and disease damage detection system based on image recognition as claimed in claim 3, wherein: the user terminal further comprises a performance evaluation module,
the offline system deployment module is also used for sending a hardware performance evaluation instruction to the user terminal when receiving the offline identification request;
the performance evaluation module is used for evaluating the hardware performance of the user terminal when receiving the hardware performance evaluation instruction to obtain a hardware performance evaluation result and sending the hardware performance evaluation result to the server, wherein the hardware performance evaluation result comprises high-performance hardware and low-performance hardware;
the offline system deployment module is also used for analyzing the received hardware performance evaluation result, and directly deploying the pest and disease damage identification system at the user terminal if the analysis result is high-performance hardware; and if the analysis result is low-performance hardware, deploying the compressed pest and disease damage identification system at the user terminal.
5. The traditional Chinese medicine material pest and disease damage detection system based on image recognition as claimed in claim 4, wherein: and the offline system deployment module compresses the pest and disease identification system by adopting a deep neural network compression technology.
6. The traditional Chinese medicine pest and disease detection system based on image recognition as claimed in claim 4, wherein: the offline system deployment module is further used for checking whether the pest identification system of the user terminal is the current latest version of pest identification system after detecting that the pest identification system of the server is updated, and if not, updating and updating the pest identification system of the user terminal according to the latest version of pest identification system of the server.
7. The system for detecting plant diseases and insect pests of traditional Chinese medicinal materials based on image recognition as claimed in claim 5, wherein the performance evaluation module is further used for evaluating the network connection state of the user terminal, and generating an offline recognition instruction if the network connection state evaluation result indicates that the network is not normal; and if the evaluation result of the network connection state is that the network is good, generating an online identification instruction.
8. The traditional Chinese medicine material pest and disease detection system based on image recognition as claimed in claim 1, wherein: the server further comprises a pest sample acquisition module, and the pest sample acquisition module is used for learning the artificially marked pest image sample by adopting a confrontation generation network picture generation technology to generate a pest training sample.
9. The traditional Chinese medicine material pest and disease detection system based on image recognition as claimed in claim 1, wherein: and the pest and disease identification system is also used for carrying out target detection on the pest and disease image to obtain the pest and disease image to be identified.
10. The traditional Chinese medicine material pest and disease detection system based on image recognition as claimed in claim 1, wherein: the server further comprises a disease and insect pest model generation module, and the disease and insect pest model generation module is used for generating a disease and insect pest identification model by training the disease and insect pest training samples through a model fusing a residual error net ResNet and an IceptionNet.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011294151.4A CN112381156B (en) | 2020-11-18 | 2020-11-18 | Chinese medicinal material pest and disease damage detection system based on image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011294151.4A CN112381156B (en) | 2020-11-18 | 2020-11-18 | Chinese medicinal material pest and disease damage detection system based on image recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112381156A CN112381156A (en) | 2021-02-19 |
| CN112381156B true CN112381156B (en) | 2023-04-18 |
Family
ID=74585139
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011294151.4A Active CN112381156B (en) | 2020-11-18 | 2020-11-18 | Chinese medicinal material pest and disease damage detection system based on image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112381156B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112949796A (en) * | 2021-03-23 | 2021-06-11 | 青岛农业大学 | Hybrid pest and disease identification device and method based on deep learning |
| CN113303151A (en) * | 2021-05-19 | 2021-08-27 | 苏州三润景观工程有限公司 | Green plant disease and insect pest repairing method based on big data |
| CN113887595A (en) * | 2021-09-23 | 2022-01-04 | 四川飘绿植物保护有限公司 | Garden pest identification method, system, computer equipment and storage medium |
| CN115119669A (en) * | 2022-05-30 | 2022-09-30 | 嘉兴职业技术学院 | Method for preventing and treating crocus sativus corm rot disease by using medicament and application |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109379557A (en) * | 2018-09-30 | 2019-02-22 | 田东县文设芒果专业合作社 | Mango insect pest intelligent monitor system based on image recognition |
| CN209805978U (en) * | 2019-06-26 | 2019-12-17 | 大安市中科佰澳格霖农业发展有限公司 | Plant growth monitoring device based on agricultural internet of things |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103810361A (en) * | 2012-11-09 | 2014-05-21 | 上海斐讯数据通信技术有限公司 | Diagnosis system and diagnosis method |
| CN106682617B (en) * | 2016-12-28 | 2019-11-05 | 电子科技大学 | A kind of image clearly based on frequency spectrum segment information and fuzzy discrimination feature extracting method |
| CN106991619A (en) * | 2017-06-05 | 2017-07-28 | 河北哲瀚网络科技有限公司 | A kind of diseases and pests of agronomic crop intelligent diagnosis system and diagnostic method |
| CN210376158U (en) * | 2019-08-05 | 2020-04-21 | 重庆电子工程职业学院 | Portable intelligent crop disease and pest detector |
| CN111767849A (en) * | 2020-06-29 | 2020-10-13 | 京东数字科技控股有限公司 | Crop pest and disease identification method and device and storage medium |
| CN111884345A (en) * | 2020-08-17 | 2020-11-03 | 安徽南瑞继远电网技术有限公司 | Power transmission line image monitoring equipment with intelligent image analysis function |
-
2020
- 2020-11-18 CN CN202011294151.4A patent/CN112381156B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109379557A (en) * | 2018-09-30 | 2019-02-22 | 田东县文设芒果专业合作社 | Mango insect pest intelligent monitor system based on image recognition |
| CN209805978U (en) * | 2019-06-26 | 2019-12-17 | 大安市中科佰澳格霖农业发展有限公司 | Plant growth monitoring device based on agricultural internet of things |
Non-Patent Citations (1)
| Title |
|---|
| 房若民 ; 沈凯文 ; 李浩伟 ; .MobileNet算法的嵌入式农业病虫害识别系统.单片机与嵌入式系统应用.2020,(第05期),61-63. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112381156A (en) | 2021-02-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112381156B (en) | Chinese medicinal material pest and disease damage detection system based on image recognition | |
| RU2764872C2 (en) | Weed detection in natural environment | |
| WO2019228267A1 (en) | Short video synthesis method and apparatus, and device and storage medium | |
| CN111767802B (en) | Method and device for detecting abnormal state of object | |
| CN114972843B (en) | Agricultural pest diagnosis and early warning system based on big data | |
| CN110046586A (en) | A kind of data processing method, equipment and storage medium | |
| CN113657469B (en) | Automatic observation method and system for woody plant waiting period based on image recognition | |
| JP6704148B1 (en) | Crop yield forecast program and crop quality forecast program | |
| CN112067616A (en) | Real-time detection device and method for grain quality | |
| Tumang | Pests and diseases identification in mango using MATLAB | |
| CN107180056A (en) | The matching process and device of fragment in video | |
| Mann et al. | Automatic flower detection and phenology monitoring using time‐lapse cameras and deep learning | |
| CN111985724A (en) | Crop yield estimation method, device, equipment and storage medium | |
| CN112733958A (en) | Greenhouse ozone concentration control method and system | |
| CN115292334A (en) | Intelligent planting method and system based on vision, electronic equipment and storage medium | |
| CN111213900A (en) | Intelligent image analysis automatic control system and method for tobacco leaf baking | |
| CN113435345A (en) | Growth stage determination method and device, agricultural system, equipment and storage medium | |
| CN117197595A (en) | Fruit tree growth period identification method, device and management platform based on edge calculation | |
| CN117218534A (en) | Crop leaf disease identification method | |
| CN115379150B (en) | System and method for automatically generating dynamic video of rice growth process in remote way | |
| Victorino et al. | Grapevine yield prediction using image analysis-improving the estimation of non-visible bunches | |
| CN115690521A (en) | Cabbage mustard maturity identification method | |
| CN115082706A (en) | Method, device and equipment for identifying plant diseases and insect pests and readable storage medium | |
| CN112001290B (en) | Rice planthopper migration path prediction method based on YOLO algorithm | |
| EP3973767A1 (en) | Information processing device and index value calculation method |
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 |