CN112800877A - Pest control method based on agricultural Internet of things - Google Patents

Pest control method based on agricultural Internet of things Download PDF

Info

Publication number
CN112800877A
CN112800877A CN202110051665.5A CN202110051665A CN112800877A CN 112800877 A CN112800877 A CN 112800877A CN 202110051665 A CN202110051665 A CN 202110051665A CN 112800877 A CN112800877 A CN 112800877A
Authority
CN
China
Prior art keywords
data
pest
things
svm
layer
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202110051665.5A
Other languages
Chinese (zh)
Inventor
江煜
杨忠
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jinling Institute of Technology
Original Assignee
Jinling Institute of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Jinling Institute of Technology filed Critical Jinling Institute of Technology
Priority to CN202110051665.5A priority Critical patent/CN112800877A/en
Publication of CN112800877A publication Critical patent/CN112800877A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01G—HORTICULTURE; CULTIVATION OF VEGETABLES, FLOWERS, RICE, FRUIT, VINES, HOPS OR SEAWEED; FORESTRY; WATERING
    • A01G13/00—Protection of plants
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/02—Agriculture; Fishing; Forestry; Mining
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16Y—INFORMATION AND COMMUNICATION TECHNOLOGY SPECIALLY ADAPTED FOR THE INTERNET OF THINGS [IoT]
    • G16Y10/00—Economic sectors
    • G16Y10/05—Agriculture
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Business, Economics & Management (AREA)
  • General Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • General Business, Economics & Management (AREA)
  • Economics (AREA)
  • Data Mining & Analysis (AREA)
  • Agronomy & Crop Science (AREA)
  • Artificial Intelligence (AREA)
  • Primary Health Care (AREA)
  • Animal Husbandry (AREA)
  • Marine Sciences & Fisheries (AREA)
  • Mining & Mineral Resources (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Multimedia (AREA)
  • Human Resources & Organizations (AREA)
  • Marketing (AREA)
  • General Engineering & Computer Science (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Toxicology (AREA)
  • Environmental Sciences (AREA)
  • Accounting & Taxation (AREA)
  • Development Economics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Medical Informatics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种基于农业物联网的害虫防治方法，该方法包括以下步骤：步步骤1，搭建以STM32为基础的害虫防治物联网系统；步骤2：确定网络层通信协议；步骤3，获取模拟噪声环境数据；步骤4，训练SVM害虫分类模型；步骤5，实现基于农业物联网的害虫防治。本发明模拟了传感器在噪声环境下采集地数据，提出了一种基于农业物联网的害虫防治方法，提高了物联网平台层对害虫分类的稳定性和鲁棒性，同时物联网平台层结合感知层采集的各类数据信息有针对性地对害虫进行防治。

Figure 202110051665

A pest control method based on the Agricultural Internet of Things, the method comprising the following steps: Step 1, build a pest control Internet of Things system based on STM32; Step 2: Determine a network layer communication protocol; Step 3: Acquire simulated noise environment data ; Step 4, train the SVM pest classification model; Step 5, realize the pest control based on the Agricultural Internet of Things. The invention simulates the data collected by the sensor in the noise environment, and proposes a pest control method based on the agricultural Internet of Things, which improves the stability and robustness of the Internet of Things platform layer for pest classification, and at the same time, the Internet of Things platform layer combines perception Various types of data and information collected by the layers can be used to control pests in a targeted manner.

Figure 202110051665

Description

Pest control method based on agricultural Internet of things
Technical Field
The invention relates to the field of agricultural Internet of things, in particular to a pest control method based on the agricultural Internet of things.
Background
Agriculture is the basic industry of national economy in China, and the pest control effect is directly related to agricultural production benefit, agricultural product quality safety, agricultural ecological environment safety and the like. In recent years, global climate is abnormally changeable, farming systems and production modes are changed, crop multiple cropping indexes are improved, and crop insect pests are frequently and massively infected. The insect pest detection method mainly depends on the experience of growers, agricultural technicians and agricultural experts are invited to guide the plants at home or remotely in some areas, and the traditional method usually misses the best prevention and control opportunity, and has low efficiency, high cost and poor timeliness.
In recent years, with the development of machine learning and deep learning, the occurrence of agricultural insect pests can be greatly reduced by recognizing agricultural insect pests based on machine vision recognition and convolutional neural network deep learning, when the characteristics of pest bodies appear, the rapid analysis and judgment of images can be realized, the insect pest types can be accurately positioned, early warning is timely sent out, and the pests are killed in germination. Meanwhile, with the wide deployment of sensors in the real world, the internet technology gradually permeates into the physical entity world, more and more physical entities are connected to the internet through the sensors to realize information sharing, and the internet of things is applied to the background. The new wave of information development of the Internet of Things (Internet of Things) has attracted great attention from the industry and academia. The thing networking has contained 4 parts: the system comprises physical entities in the real world, sensors for sensing state information of the physical entities, a transmission network and an intelligent processing system.
Disclosure of Invention
In order to solve the problems, the invention simulates the data collected by the sensor in the noise environment, provides the pest control method based on the agricultural Internet of things, improves the stability and robustness of the Internet of things platform layer for classifying the pests, and meanwhile, the Internet of things platform layer combines various data information collected by the sensing layer to pertinently control the pests. In addition, in order to enhance the separability of farmland image data, the invention adopts an SVM classifier. To achieve the purpose, the invention provides a pest control method based on an agricultural Internet of things, which comprises the following specific steps:
step 1, building a pest control Internet of things system based on STM 32: the STM32 is used as a controller of a network layer of the Internet of things, and a WiFi communication module is used as a network layer to be in wireless communication with a perception layer and a platform layer respectively;
step 2: determining a network layer communication protocol: the communication data format comprises a frame head, an address, command codes, effective data, check bits and a frame tail, different command codes are used for sending a working instruction to the STM32, and address coding is carried out on different sensing layer sensors so as to distinguish data;
step 3, acquiring simulated noise environment data: adding salt and pepper noise to the agricultural image data acquired by the platform in the step 2, and simulating interference on an acquired signal in a noise environment by adding the salt and pepper noise;
step 4, training an SVM pest classification model: dividing the agricultural image data obtained in the step 3 into training samples and testing samples, taking the characteristics of the training samples as input, taking the labels of the training samples as output, and obtaining an SVM pest classification model;
and 5, realizing pest control based on the agricultural Internet of things: and sending a corresponding command code to an STM32 controller according to a pest classification result output by the SVM pest classification model, and killing pests in a targeted manner through an external pest control system.
Further, the pest control internet of things system based on STM32 built in the step 1 can be expressed as follows:
the system is divided into: perception layer, network layer, the data of perception layer are connected with the network layer through STM 32's wiFi module, and the perception layer includes: the temperature data that temperature sensor gathered, the humidity data that humidity transducer gathered, the PH value data that PH value sensor gathered, the light intensity data that light intensity sensor gathered, the farmland image data that CCD sensor gathered, every sensor node is independent each other. The STM32 controller passes through drive circuit, filtering amplifier circuit, AD converting circuit and gathers the all kinds of data on perception layer to utilize the WIFI module to send the data transmission who gathers to the thing networking platform, the thing networking platform layer is to the processing analysis of further step of data that detect.
Further, the acquisition of the simulated noise environment data in step 3 can be represented as:
before data are processed and analyzed on an Internet of things platform layer, a pest classification model based on the agricultural Internet of things is constructed. Thing networking platform layer distinguishes the data that the perception layer was gathered through network communication protocol's command code to obtain the farmland image of real-time collection, after STM32 detected all kinds of data, can be with data packing into data format and be: frame head, address, command code, valid data, check digit, frame tail, platform layer receive farmland image data after, establish farmland image data set, add salt and pepper noise simulation environmental noise on the basis of image data set, the salt and pepper noise model that adds is:
Figure BDA0002899286420000021
wherein, ImaxAnd IminIs the maximum value and the minimum value of the image pixel points, p is the probability of noise occurrence of the image, the value range of p is 20 percent to 30 percent, ixyThe actual value of the collected image pixel point (x, y) is obtained, and the f (x, y) is the image pixel point value after the salt and pepper noise is added.
Further, the process of training the SVM pest classification model in step 4 can be expressed as:
the agricultural image dataset constitutes a sample set (x) of k datai,yi)jWherein j is 1,2, the., k, i is 1,2, the., n, n is the number of training samples, y is the pest category label, and the optimal hyperplane of k classifications obtained by learning through an interval maximization learning strategy is as follows:
ω·x+b＝0 (2)
where ω is the normal vector and b is the displacement, the general form of the corresponding linear classification function is:
f(x)＝ω·x+b (3)
while the optimal hyperplane makes all sample points satisfy:
f(x)|≥1 (4)
the support vector of the SVM is a sample point for establishing the formula 4, and after the support vector is found, the data can be divided into two types; the SVM model adopts a Gaussian radial basis kernel function, wherein the Gaussian radial basis kernel function is as follows:
Figure BDA0002899286420000031
during classification, the SVM algorithm maps sample data from a low-dimensional feature space to a high-dimensional feature space in a nonlinear way through a kernel function, an optimal hyperplane with maximized interval is found in the high-dimensional feature space, and the nonlinear classification function is as follows:
Figure BDA0002899286420000032
the SVM algorithm is a two-classification model, a plurality of two-classification models are constructed at the same time, the multi-classification problem can be solved, when n pest categories exist, a group of SVM classifiers need to design n (n-1)/2 SVM models, and finally k SVM pest classifiers are trained.
Further, the process of implementing pest control based on the agricultural internet of things in step 5 can be expressed as:
after obtaining SVM pest classifier in step 4 training, imbed the classifier in the platform layer of thing networking, the temperature data is monitored at each monitoring point to the perception layer, humidity data, PH value data, light intensity data, behind the farmland image data, utilize the WIFI module to send to the platform layer with data packing through STM32, the platform layer distinguishes various data through the command code in the data, independently carry out the analysis, show real-time temperature data on the platform, humidity data, PH value data, light intensity data, and input farmland image data to SVM pest classifier, output farmland pest type, combine all kinds of data to have pertinence ground infrastructure, adjust planting structure, reasonable fertilization, in time water, prevention and cure measures such as chemical prevention and cure send to the platform layer, remind the grower to prevent and cure.
The pest control method based on the agricultural Internet of things has the beneficial effects that: the invention has the technical effects that:
1. the agricultural data acquisition system utilizes the sensing layer to acquire various agricultural data, and independently transmits various data to the platform layer through the driving circuit, the filtering and amplifying circuit, the A/D conversion circuit, the WIFI module and the like;
2. the method effectively simulates the interference of the agricultural image in the noise environment when the data are collected, and the data simulating the noise are used for training the SVM classification model, so that the robustness and stability of the model in the noise environment are enhanced;
drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
The invention is described in further detail below with reference to the following detailed description and accompanying drawings:
the invention provides a pest control method based on an agricultural Internet of things, and aims to improve the robustness of data decoupling of a multi-dimensional force sensor in a noise environment and improve the stability and accuracy of data decoupling. FIG. 1 is a flow chart of the present invention. The steps of the present invention will be described in detail with reference to the flow chart.
Step 1, building a pest control Internet of things system based on STM 32: the STM32 is used as a controller of a network layer of the Internet of things, and a WiFi communication module is used as a network layer to be in wireless communication with a perception layer and a platform layer respectively;
the pest control internet of things system based on STM32 built in the step 1 can be expressed as follows:
the system is divided into: perception layer, network layer, the data of perception layer are connected with the network layer through STM 32's wiFi module, and the perception layer includes: the temperature data that temperature sensor gathered, the humidity data that humidity transducer gathered, the PH value data that PH value sensor gathered, the light intensity data that light intensity sensor gathered, the farmland image data that CCD sensor gathered, every sensor node is independent each other. The STM32 controller passes through drive circuit, filtering amplifier circuit, AD converting circuit and gathers the all kinds of data on perception layer to utilize the WIFI module to send the data transmission who gathers to the thing networking platform, the thing networking platform layer is to the processing analysis of further step of data that detect.
Step 2: determining a network layer communication protocol: the communication data format comprises a frame head, an address, command codes, effective data, check bits and a frame tail, different command codes are used for sending a working instruction to the STM32, and address coding is carried out on different sensing layer sensors so as to distinguish data;
step 3, acquiring simulated noise environment data: adding salt and pepper noise to the agricultural image data acquired by the platform in the step 2, and simulating interference on an acquired signal in a noise environment by adding the salt and pepper noise;
the acquisition of the simulated noise environment data in step 3 can be represented as:
before data are processed and analyzed on an Internet of things platform layer, a pest classification model based on the agricultural Internet of things is constructed. Thing networking platform layer distinguishes the data that the perception layer was gathered through network communication protocol's command code to obtain the farmland image of real-time collection, after STM32 detected all kinds of data, can be with data packing into data format and be: frame head, address, command code, valid data, check digit, frame tail, platform layer receive farmland image data after, establish farmland image data set, add salt and pepper noise simulation environmental noise on the basis of image data set, the salt and pepper noise model that adds is:
Figure BDA0002899286420000041
wherein, ImaxAnd IminIs the maximum value and the minimum value of the image pixel points, p is the probability of noise occurrence of the image, the value range of p is 20 percent to 30 percent, ixyThe actual value of the collected image pixel point (x, y) is obtained, and the f (x, y) is the image pixel point value after the salt and pepper noise is added.
Step 4, training an SVM pest classification model: dividing the agricultural image data obtained in the step 3 into training samples and testing samples, taking the characteristics of the training samples as input, taking the labels of the training samples as output, and obtaining an SVM pest classification model;
the process of training the SVM pest classification model in step 4 can be expressed as:
the agricultural image dataset constitutes a sample set (x) of k datai,yi)jWhere j is 1, 2.. times, k, i is 1, 2.. times, n, n is the number of training samples, y ∈ { -1, 1} isPest category labels, and learning by an interval maximization learning strategy to obtain the optimal hyperplane of k classifications as follows:
ω·x+b＝0 (2)
where ω is the normal vector and b is the displacement, the general form of the corresponding linear classification function is:
f(x)＝ω·x+b (3)
while the optimal hyperplane makes all sample points satisfy:
f(x)|≥1 (4)
the support vector of the SVM is a sample point for establishing the formula 4, and after the support vector is found, the data can be divided into two types; the SVM model adopts a Gaussian radial basis kernel function, wherein the Gaussian radial basis kernel function is as follows:
Figure BDA0002899286420000051
during classification, the SVM algorithm maps sample data from a low-dimensional feature space to a high-dimensional feature space in a nonlinear way through a kernel function, an optimal hyperplane with maximized interval is found in the high-dimensional feature space, and the nonlinear classification function is as follows:
Figure BDA0002899286420000052
the SVM algorithm is a two-classification model, a plurality of two-classification models are constructed at the same time, the multi-classification problem can be solved, when n pest categories exist, a group of SVM classifiers need to design n (n-1)/2 SVM models, and finally k SVM pest classifiers are trained.
And 5, realizing pest control based on the agricultural Internet of things: and sending a corresponding command code to an STM32 controller according to a pest classification result output by the SVM pest classification model, and killing pests in a targeted manner through an external pest control system.
The process of realizing pest control based on the agricultural internet of things in the step 5 can be expressed as follows:
after obtaining SVM pest classifier in step 4 training, imbed the classifier in the platform layer of thing networking, the temperature data is monitored at each monitoring point to the perception layer, humidity data, PH value data, light intensity data, behind the farmland image data, utilize the WIFI module to send to the platform layer with data packing through STM32, the platform layer distinguishes various data through the command code in the data, independently carry out the analysis, show real-time temperature data on the platform, humidity data, PH value data, light intensity data, and input farmland image data to SVM pest classifier, output farmland pest type, combine all kinds of data to have pertinence ground infrastructure, adjust planting structure, reasonable fertilization, in time water, prevention and cure measures such as chemical prevention and cure send to the platform layer, remind the grower to prevent and cure.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the present invention in any way, but any modifications or equivalent variations made according to the technical spirit of the present invention are within the scope of the present invention as claimed.

Claims (5)

1.一种基于农业物联网的害虫防治方法，具体步骤如下，其特征在于：1. a pest control method based on agricultural internet of things, concrete steps are as follows, it is characterized in that: 步骤1，搭建以STM32为基础的害虫防治物联网系统：使用STM32作为物联网网络层的控制器，利用WiFi通讯模块作为网络层分别与感知层和平台层的无线通信方式；Step 1, build a pest control Internet of Things system based on STM32: use STM32 as the controller of the Internet of Things network layer, and use the WiFi communication module as the wireless communication mode of the network layer and the perception layer and the platform layer respectively; 步骤2：确定网络层通信协议：通讯数据格式有帧头、地址、命令码、有效数据、校验位、帧尾，利用不同命令码对STM32发送工作指令，并对不同的感知层传感器进行地址编码，以做数据辨别；Step 2: Determine the network layer communication protocol: The communication data format includes frame header, address, command code, valid data, check digit, and frame end. Use different command codes to send work instructions to STM32, and address different sensing layer sensors. encoding for data identification; 步骤3，获取模拟噪声环境数据：向步骤2中平台获取的农业图像数据添加椒盐噪声，通过添加椒盐噪声模拟噪声环境下采集信号所受到的干扰；Step 3, obtaining simulated noise environment data: adding salt and pepper noise to the agricultural image data obtained by the platform in step 2, and simulating the interference of the collected signals under the noise environment by adding salt and pepper noise; 步骤4，训练SVM害虫分类模型：将步骤3获得的农业图像数据化分为训练样本和测试样本，将训练样本的特征作为输入，训练样本的标签作为输出，获得SVM害虫分类模型；Step 4, train the SVM pest classification model: Divide the agricultural image data obtained in step 3 into training samples and test samples, take the features of the training samples as input, and use the labels of the training samples as the output to obtain the SVM pest classification model; 步骤5，实现基于农业物联网的害虫防治：针对SVM害虫分类模型输出的害虫分类结果，发送对应命令码至STM32控制器中，通过外接的害虫防治系统，有针对性的灭除害虫。Step 5, implement pest control based on the agricultural Internet of Things: according to the pest classification results output by the SVM pest classification model, send the corresponding command code to the STM32 controller, and use the external pest control system to eliminate pests in a targeted manner. 2.根据权利要求1所述的一种基于农业物联网的害虫防治方法，其特征在于：步骤1中搭建以STM32为基础的害虫防治物联网系统可以表示为：2. a kind of pest control method based on agricultural internet of things according to claim 1, is characterized in that: in step 1, build the pest control internet of things system based on STM32 can be expressed as: 该系统分为：感知层、网络层，感知层的数据通过STM32的WiFi模块与网络层相连接，感知层包括：温度传感器采集的温度数据、湿度传感器采集的湿度数据、PH值传感器采集的PH值数据、光强传感器采集的光强数据、CCD传感器采集的农田图像数据，每个传感器节点互相独立；STM32控制器通过驱动电路、滤波放大电路、A/D转换电路采集感知层的各类数据，并利用WIFI模块将采集到的数据发送至物联网平台，物联网平台层对检测到的数据做近一步的处理分析。The system is divided into: perception layer and network layer. The data of the perception layer is connected to the network layer through the WiFi module of STM32. The perception layer includes: the temperature data collected by the temperature sensor, the humidity data collected by the humidity sensor, and the pH value collected by the PH sensor. Value data, light intensity data collected by light intensity sensor, farmland image data collected by CCD sensor, each sensor node is independent of each other; STM32 controller collects various data of perception layer through drive circuit, filter amplifier circuit, and A/D conversion circuit , and use the WIFI module to send the collected data to the IoT platform, and the IoT platform layer will further process and analyze the detected data. 3.根据权利要求1所述的一种基于农业物联网的害虫防治方法，其特征在于：步骤3中获取模拟噪声环境数据可以表示为：3. a kind of pest control method based on agricultural internet of things according to claim 1, is characterized in that: in step 3, obtain simulated noise environment data can be expressed as: 在物联网平台层对数据进行处理分析前，先构建基于农业物联网的害虫分类模型，物联网平台层通过网络通讯协议的命令码区分感知层采集到的数据，并获得实时采集的农田图像，当STM32检测到各类数据后，会将数据打包成数据格式为：帧头、地址、命令码、有效数据、校验位、帧尾，平台层收到农田图像数据后，建立农田图像数据集，在图像数据集的基础上添加椒盐噪声模拟环境噪声，所添加的椒盐噪声模型为：Before processing and analyzing the data at the IoT platform layer, a pest classification model based on the agricultural IoT is first constructed. The IoT platform layer distinguishes the data collected by the perception layer through the command code of the network communication protocol, and obtains the farmland images collected in real time. When STM32 detects all kinds of data, it will pack the data into the data format: frame header, address, command code, valid data, check digit, frame end. After receiving the farmland image data, the platform layer establishes the farmland image data set , on the basis of the image data set, salt and pepper noise is added to simulate environmental noise, and the added salt and pepper noise model is:
Figure FDA0002899286410000011
Figure FDA0002899286410000011
 其中，Imax和Imin是图像像素点的最大值和最小值，p为图像出现噪声的概率，p的取值范围为20％～30％，ixy是采集图像像素点(x,y)的实际值，f(x,y)是添加椒盐噪声后的图像像素点值。Among them, I max and I min are the maximum and minimum values of image pixel points, p is the probability of noise in the image, the value range of p is 20% to 30%, and i xy is the acquisition image pixel point (x, y) The actual value of , f(x,y) is the image pixel value after adding salt and pepper noise. 4.根据权利要求1所述的一种基于农业物联网的害虫防治方法，其特征在于：步骤4中训练SVM害虫分类模型的过程可以表示为：4. a kind of pest control method based on agricultural internet of things according to claim 1, is characterized in that: the process of training SVM pest classification model in step 4 can be expressed as: 农业图像数据集组成k个数据的样本集(xi,yi)j，其中j＝1,2,...,k，i＝1,2,...,n，n是训练样本的个数，y∈{-1，1}是害虫类别标号，通过“间隔最大化”学习策略学习得到k个分类的最优超平面为：The agricultural image dataset consists of a sample set of k data (x i , y i ) j , where j=1,2,...,k, i=1,2,...,n, n is the training sample The number, y∈{-1, 1} is the label of the pest category, and the optimal hyperplane of k categories obtained by learning through the "interval maximization" learning strategy is: ω·x+b＝0 (2)ω·x+b=0 (2) 式中ω是法向量，b是位移量，相应的线性分类函数的一般形式为：where ω is the normal vector, b is the displacement, and the general form of the corresponding linear classification function is: f(x)＝ω·x+b (3)f(x)=ω·x+b (3) 同时最优超平面使所有样本点满足：At the same time, the optimal hyperplane makes all sample points satisfy: f(x)|≥1 (4)f(x)|≥1 (4) SVM的支持向量是使式4成立的样本点，找到支持向量后，可将数据分为两类；SVM模型采用高斯径向基核函数，其中高斯径向基核函数函数如下：The support vector of SVM is the sample point that makes Equation 4 true. After finding the support vector, the data can be divided into two categories; the SVM model adopts the Gaussian radial basis kernel function, and the Gaussian radial basis kernel function function is as follows:
Figure FDA0002899286410000021
Figure FDA0002899286410000021
 在分类时，SVM算法通过核函数将样本数据从低维特征空间非线性的映射到高维特征空间中，在高维特征空间里找出间隔最大化的最优超平面，非线性分类函数为：During classification, the SVM algorithm non-linearly maps the sample data from the low-dimensional feature space to the high-dimensional feature space through the kernel function, and finds the optimal hyperplane that maximizes the interval in the high-dimensional feature space. The nonlinear classification function is :
Figure FDA0002899286410000022
Figure FDA0002899286410000022
 SVM算法本身是个二分类模型，同时构造多个二分类模型可解决多分类问题，当有n个害虫类别时，一组SVM分类器需要设计n(n-1)/2个SVM模型，最终训练k个SVM害虫分类器。The SVM algorithm itself is a binary classification model, and multiple binary classification models can be constructed at the same time to solve the multi-classification problem. When there are n pest categories, a set of SVM classifiers needs to design n(n-1)/2 SVM models, and finally train k SVM pest classifiers. 5.根据权利要求1所述的一种基于农业物联网的害虫防治方法，其特征在于：步骤5中实现基于农业物联网的害虫防治的过程表示为：5. a kind of pest control method based on agricultural internet of things according to claim 1 is characterized in that: the process of realizing the pest control based on agricultural internet of things in step 5 is expressed as: 在步骤4训练得到SVM害虫分类器后，将分类器嵌入到物联网的平台层中，感知层在各个监测点监测到温度数据、湿度数据、PH值数据、光强数据、农田图像数据后，通过STM32将数据打包利用WIFI模块发送至平台层，平台层通过数据中的命令码区分各类数据，独立进行分析，在平台上显示实时的温度数据、湿度数据、PH值数据、光强数据，并将农田图像数据输入到SVM害虫分类器中，输出农田害虫类型，结合各类数据有针对性地将调整农田基本建设、调整种植结构、合理施肥、适时灌水、化学防治等防治措施发送至平台层，提醒种植人员进行防治。After training the SVM pest classifier in step 4, the classifier is embedded in the platform layer of the Internet of Things. The data is packaged through STM32 and sent to the platform layer using the WIFI module. The platform layer distinguishes various types of data through the command code in the data, analyzes it independently, and displays real-time temperature data, humidity data, PH value data, and light intensity data on the platform. The farmland image data is input into the SVM pest classifier, and the farmland pest types are output. Combined with various data, the control measures such as adjusting the farmland infrastructure, adjusting the planting structure, rational fertilization, timely irrigation, and chemical control are sent to the platform. layer, and remind planters to carry out control.
CN202110051665.5A 2021-01-15 2021-01-15 Pest control method based on agricultural Internet of things Pending CN112800877A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110051665.5A CN112800877A (en) 2021-01-15 2021-01-15 Pest control method based on agricultural Internet of things

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110051665.5A CN112800877A (en) 2021-01-15 2021-01-15 Pest control method based on agricultural Internet of things

Publications (1)

Publication Number Publication Date
CN112800877A true CN112800877A (en) 2021-05-14

Family

ID=75809425

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110051665.5A Pending CN112800877A (en) 2021-01-15 2021-01-15 Pest control method based on agricultural Internet of things

Country Status (1)

Country Link
CN (1) CN112800877A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113741597A (en) * 2021-09-03 2021-12-03 安徽中昆绿色防控科技有限公司 Intelligent control system for insect trapping, measuring and reporting in agriculture and forestry
CN116472812A (en) * 2023-03-21 2023-07-25 湖南容源农业开发有限公司 Soil improvement method for high-standard farmland construction
CN119004122A (en) * 2024-10-24 2024-11-22 南京信息工程大学 Bearing life prediction method based on amplitude-phase characteristics and time-frequency domain self-supervision

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106210043A (en) * 2016-07-11 2016-12-07 太原理工大学 A kind of industrialized agriculture environmental information remote intelligent monitoring system based on Internet of Things
CN110913012A (en) * 2019-12-05 2020-03-24 金陵科技学院 High-speed parallel data processing method based on agricultural Internet of things
CN112085032A (en) * 2020-09-14 2020-12-15 衢州学院 Dynamic monitoring model of Huyou Pomelo diseases and insect pests based on the combination of Internet of Things image processing and SVM

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106210043A (en) * 2016-07-11 2016-12-07 太原理工大学 A kind of industrialized agriculture environmental information remote intelligent monitoring system based on Internet of Things
CN110913012A (en) * 2019-12-05 2020-03-24 金陵科技学院 High-speed parallel data processing method based on agricultural Internet of things
CN112085032A (en) * 2020-09-14 2020-12-15 衢州学院 Dynamic monitoring model of Huyou Pomelo diseases and insect pests based on the combination of Internet of Things image processing and SVM

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
江 煜 等: "基于 Ｍ２Ｍ 平台的智能农业物联网监控系统设计", 《金陵科技学院学报》, pages 52 - 54 *
江 煜 等: "物联网工程中 Ｍ２Ｍ 技术研究", 《金陵科技学院学报》, pages 17 - 21 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113741597A (en) * 2021-09-03 2021-12-03 安徽中昆绿色防控科技有限公司 Intelligent control system for insect trapping, measuring and reporting in agriculture and forestry
CN113741597B (en) * 2021-09-03 2022-04-12 安徽中昆绿色防控科技有限公司 Intelligent control system for insect trapping, measuring and reporting in agriculture and forestry
CN116472812A (en) * 2023-03-21 2023-07-25 湖南容源农业开发有限公司 Soil improvement method for high-standard farmland construction
CN119004122A (en) * 2024-10-24 2024-11-22 南京信息工程大学 Bearing life prediction method based on amplitude-phase characteristics and time-frequency domain self-supervision

Similar Documents

Publication Publication Date Title
CN112800877A (en) Pest control method based on agricultural Internet of things
Maduranga et al. Machine learning applications in IoT based agriculture and smart farming: A review
CN110765916B (en) Farmland seedling ridge identification method and system based on semantics and example segmentation
Wani et al. An appropriate model predicting pest/diseases of crops using machine learning algorithms
Ramana et al. Leaf disease classification in smart agriculture using deep neural network architecture and IoT
Yashaswini et al. Smart automated irrigation system with disease prediction
CN113948220B (en) A detection method for anthracnose pathogen infection stage with pre-analysis ability
CN112461828A (en) Intelligent pest and disease damage forecasting and early warning system based on convolutional neural network
CN118261488B (en) A smart management system based on digital farm
Naidu et al. Mango leaf disease detection using ultrasonic sensor
Abdulla et al. Agriculture based on internet of things and deep learning
Zahan et al. A deep learning-based approach for mushroom diseases classification
Nalwanga et al. Design of an embedded machine learning based system for an environmental-friendly crop prediction using a sustainable soil fertility management
CN113591990A (en) Meteorological disaster early warning method based on agricultural Internet of things
Kar et al. IoT and drone-based field monitoring and surveillance system
CN107741444A (en) A crop VOCs real-time monitoring device for crop monitoring and early warning
CN112019607A (en) Smart farm comprehensive control method and system based on visual cloud platform
Diya et al. IoT-based precision agriculture: A review
CN110084145A (en) A kind of multiple dimensioned identifying system of pest and disease damage time-frequency domain and operating method based on TensorFlow
CN116560946A (en) Soil pollution data pushing system based on cloud computing
Huang et al. Application of data augmentation and migration learning in identification of diseases and pests in tea trees
Thakre et al. UAV Based System For Detection in Integrated Insect Management for Agriculture Using Deep Learning
CN114359708A (en) Crop identification method and device
Sathiya et al. Plant disease identification using IoT and deep learning algorithms
Kannale et al. A Review on Smart Irrigation System: Sugarcane Diseases Detection Using Raspberry-pi

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20210514

RJ01 Rejection of invention patent application after publication