Remote sensing-based pest and disease damage monitoring and early warning method and system

Abstract

The invention provides a disease and pest monitoring and early warning method and system based on remote sensing, and belongs to the technical field of disease and pest monitoring. Disease and pest information is collected and analyzed regularly, the hazard level is predicted in combination with historical data, and accurate early warning and timely response are achieved; meanwhile, remote sensing data processing levels are dynamically divided according to the estimated hazard level, different feature extraction methods are adopted, high efficiency of data processing and accurate extraction of key information are ensured, the data processing amount is reduced, the processing efficiency is improved, and a feature data set is matched with a pest and disease damage database, so that the accuracy of remote sensing data processing is improved. Intelligent identification and decision support of diseases and insect pests are realized, the monitoring cost is reduced, and the monitoring efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of pest monitoring, and in particular to a pest monitoring and early warning method based on remote sensing and a system thereof. Background Art

[0002] In agricultural production, pests and diseases are important factors that threaten the healthy growth of crops and affect the quality and yield of agricultural products. In traditional pest and disease monitoring, managers or farmers are often required to go to the fields to observe and record in person. This process is not only time-consuming, but also requires a lot of manpower. Especially during the high incidence of pests and diseases, the workload of monitoring work will increase sharply, which brings a great burden to managers or farmers. This is not only time-consuming and laborious, but also easily affected by human factors. Secondly, a large amount of monitoring data needs to be collected and processed. However, if a unified processing method is adopted, it is not only inefficient, but also difficult to meet the processing requirements of different pest and disease data. This makes the data processing process lengthy and cumbersome, affecting the timeliness and accuracy of monitoring work. When processing and analyzing pest and disease data, if an overly simple or unified method is adopted, it is easy to cause the loss of key information.

[0003] Therefore, it is necessary to provide a pest and disease monitoring and early warning method based on remote sensing to solve the above technical problems. Summary of the invention

[0004] To solve the above technical problems, the present invention provides a remote sensing-based pest monitoring and early warning method and system for solving the problems of tedious manual data collection and analysis and complex and low-precision processing of monitoring data in the process of crop pest monitoring.

[0005] The present invention provides a pest monitoring and early warning method based on remote sensing, the monitoring and early warning method comprising the following steps:

[0006] Obtain historical data on pests and diseases of crops in the current planting area, and evaluate and record the initial damage level of each pest and disease;

[0007] Obtain information on various pests and diseases of crops in the current planting area according to the preset cycle, and analyze the destructive changes of each pest and disease respectively to obtain the destructive change results of each pest and disease in the cycle;

[0008] Based on the destructive changes of each pest and disease during the period and the initial damage level of each pest and disease, the final damage level of each pest and disease is estimated respectively;

[0009] Acquire remote sensing data of crops in real time, and classify the remote sensing data processing levels corresponding to each pest based on the estimated final damage level of each pest;

[0010] Based on the processing levels of the divided remote sensing data, the features of the remote sensing data are extracted in stages to obtain feature data sets;

[0011] The feature data set is input into the constructed pest and disease database for matching, the pest and disease matching results are output, and the corresponding early warning notification is generated.

[0012] Preferably, the steps of obtaining the historical data of pests and diseases of crops in the current planting area, and evaluating and recording the initial damage level of each pest and disease include:

[0013] Retrieve the historical records of crop diseases and insect pests from the current planting area database, including the occurrence time, type and impact range data;

[0014] Conduct quantitative analysis on the degree of damage of each type of pests and diseases in historical data and set the initial damage level, where the damage levels include level one, level two and level three;

[0015] The initial hazard level is associated with the corresponding pest and disease type and recorded and stored in the local database.

[0016] Preferably, the information of various pests and diseases of crops in the current planting area is obtained according to a preset period, and the destructive changes of various pests and diseases are analyzed respectively to obtain the destructive change results of various pests and diseases in the period, and the specific steps include:

[0017] Collect pest and disease information of the planting area according to a pre-set fixed period, where the pest and disease information of the planting area includes insect population density, disease spot spread rate and crop damage area;

[0018] Compare the pest and disease information of the current period with that of the previous period, calculate the destructive growth rate of each pest and disease, and obtain the destructive change results of each pest and disease;

[0019] Record the destructive changes of each pest and disease, and generate a dynamic change trend chart.

[0020] Preferably, the final hazard level of each pest and disease is estimated based on the destructive change results of each pest and disease during the period and the initial hazard level of each pest and disease, and the specific steps include:

[0021] The initial hazard level and destructive change results are input into the linear weighted prediction model to output the estimated values ​​of each pest and disease at present;

[0022] Based on the preset multiple threshold segments, by determining the threshold segment to which the estimated value of each current pest and disease belongs, the final hazard level corresponding to each current pest and disease and the threshold segment is divided.

[0023] Preferably, the real-time acquisition of crop remote sensing data and the classification of various remote sensing data processing levels corresponding to various pests and diseases based on the estimated final damage levels of various pests and diseases at present include:

[0024] Real-time remote sensing data of crops can be obtained through drone multispectral sensors, including multispectral, hyperspectral and radar data;

[0025] According to the estimated final damage level of each current pest and disease, the multispectral, hyperspectral and radar data are dynamically divided into different processing levels, among which the processing levels specifically include low, medium and high.

[0026] Preferably, the steps of extracting the features of various types of remote sensing data based on the divided processing levels of various types of remote sensing data and obtaining feature data sets include:

[0027] According to different processing levels, the remote sensing data of crops acquired in real time are processed accordingly. Among them, the high-level remote sensing data uses convolutional neural network to extract the texture features of disease spots, and the intermediate and low-level remote sensing data adopts fast feature extraction;

[0028] The extracted feature data are integrated to form a feature data set.

[0029] Preferably, the step of inputting the feature data set into the constructed pest and disease database for matching, outputting the pest and disease matching result, and generating a corresponding early warning notification comprises:

[0030] Input the feature data set into the constructed pest and disease database, and use the similarity measurement method to match the pest and disease characteristics in the database to obtain the matching results;

[0031] According to the matching results, determine the specific types of pests and diseases suffered by the crops in the current planting area and their degree of damage;

[0032] The specific types of pests and diseases suffered by the crops in the current planting area and their degree of damage will be determined and sent to managers or farmers via remote notification.

[0033] A pest monitoring and early warning system based on remote sensing, the monitoring and early warning system comprising:

[0034] The initial assessment module is used to obtain the historical data of pests and diseases of crops in the current planting area, and to assess and record the initial damage level of each pest and disease;

[0035] The data analysis module is used to obtain information on various pests and diseases of crops in the current planting area according to a preset period, and analyze the destructive changes of each pest and disease respectively to obtain the destructive change results of each pest and disease in the period;

[0036] The analysis and estimation module is used to estimate the final damage level of each pest and disease based on the destructive change results of each pest and disease during the period and the initial damage level of each pest and disease;

[0037] The data classification module is used to obtain remote sensing data of crops in real time and classify the remote sensing data processing levels corresponding to each pest based on the estimated final damage level of each pest;

[0038] A feature extraction module is used to extract the features of various types of remote sensing data based on the processing levels of the divided types of remote sensing data, and obtain a feature data set;

[0039] The matching and warning module is used to input the feature data set into the constructed pest and disease database for matching, output the pest and disease matching results, and generate corresponding warning notifications.

[0040] Compared with the related art, the pest monitoring and early warning method based on remote sensing provided by the present invention has the following beneficial effects:

[0041] The present invention realizes accurate early warning and timely response by regularly collecting and analyzing information on pests and diseases and predicting the hazard level in combination with historical data. At the same time, the remote sensing data processing level is dynamically divided according to the estimated hazard level and different feature extraction methods are adopted to ensure the high efficiency of data processing and the accurate extraction of key information, reduce the amount of data processing, improve processing efficiency, and match the feature data set with the pest and disease database to realize intelligent identification and decision support of pests and diseases, reduce monitoring costs and improve monitoring efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A schematic diagram of a process of a pest monitoring and early warning method based on remote sensing of the present invention;

[0043] Figure 2 The present invention is a structural block diagram of a pest monitoring and early warning system based on remote sensing. DETAILED DESCRIPTION

[0044] The present invention will be further described below in conjunction with the accompanying drawings and implementation modes.

[0045] Embodiment 1

[0046] like Figure 1 As shown, a pest monitoring and early warning method based on remote sensing includes the following steps:

[0047] S1. Obtain the historical data of pests and diseases of crops in the current planting area, and evaluate and record the initial damage level of each pest and disease.

[0048] S2. Obtain information on various diseases and pests of crops in the current planting area according to a preset period, and analyze the destructive changes of each disease and pest respectively to obtain the destructive change results of each disease and pest in the period.

[0049] S3. Based on the destructive changes of each pest and disease during the period and the initial damage level of each pest and disease, the final damage level of each pest and disease is estimated respectively.

[0050] S4. Acquire remote sensing data of crops in real time, and classify the remote sensing data processing levels corresponding to each pest and disease based on the estimated final damage level of each pest and disease.

[0051] S5. Based on the divided processing levels of various types of remote sensing data, the features of various types of remote sensing data are extracted hierarchically to obtain feature data sets.

[0052] S6. Input the feature data set into the constructed pest and disease database for matching, output the pest and disease matching result, and generate a corresponding early warning notification.

[0053] In the specific implementation process, the specific steps of step S1 are:

[0054] S101. Retrieve historical records of crop diseases and insect pests from the current planting site database, including data on occurrence time, type and impact range.

[0055] Specifically, in this embodiment, a connection is first established with the current planting site database to ensure smooth access and query of data, and the data is filtered and retrieved from the database according to a set time range such as the past year and the type of pests and diseases.

[0056] S102. Quantify the degree of damage of each type of pests and diseases in the historical data and set an initial damage level, wherein the damage levels specifically include level one, level two, and level three.

[0057] Specifically, taking the occurrence frequency, spread rate and affected area as indicators, set the coefficient threshold sections for evaluation respectively; obtain the occurrence frequency, spread rate and affected area of ​​each type of pests and diseases from the historical data of pests and diseases, judge the coefficient threshold section to which the three indicators of occurrence frequency, spread rate and affected area of ​​each type of pests and diseases belong, assign corresponding coefficients to the three indicators of occurrence frequency, spread rate and affected area of ​​each type of pests and diseases, and use the weighted summation method to calculate the comprehensive hazard coefficient of each type of pests and diseases, and set the initial hazard level of each type of pests and diseases according to the preset comprehensive hazard coefficient sections. Among them, in this embodiment, the occurrence frequency is divided into three sections: low (0-30%), medium (31%-60%), and high (61%-100%); the diffusion rate is divided into three sections: low (0-30%), medium (31%-60%), and high (61%-100%); the affected area is divided into three sections: low (0-30%), medium (31%-60%), and high (61%-100%); among them, the coefficients corresponding to the low, medium, and high sections are 0.2, 0.5, and 0.8, respectively, and the preset comprehensive hazard coefficient sections are: 0-0.4 for the first level, 0.41-0.7 for the second level, and 0.71-1 for the third level. The calculation formula for weighted summation is:

[0058]

[0059] For example, if the occurrence frequency, diffusion speed, and affected area are all in the high range, the coefficients assigned to them are 0.8, and the sum of the coefficients is 2.4. The weighted summation formula is used to calculate the value. is 0.8, and its initial hazard level is level 3; if two of the three indicators of occurrence frequency, diffusion speed and affected area are in the high range and one is in the medium range, the sum of the coefficients is 0.8+0.8+0.5=2.1, and the weighted summation formula is used to calculate It is 0.7, and its initial hazard level is level two.

[0060] S103, recording and storing the initial hazard level and the corresponding pest type in a local database.

[0061] Specifically, the initial hazard level obtained by quantitative analysis is associated with the corresponding pest and disease type to form a complete data record and store it in a local database for query and use in subsequent steps.

[0062] In the specific implementation process, the specific steps of step S2 are:

[0063] S201. Collect pest and disease information of the planting site at a preset fixed period, wherein the pest and disease information of the planting site includes insect population density, disease spot spread rate and crop damage area.

[0064] Specifically, the fixed period is set as weekly, monthly or quarterly. During the set collection period, remote sensing technology, such as drone remote sensing or satellite remote sensing, is used to collect pest and disease information in the planting area, including key indicators such as insect population density, disease spot spread rate and crop damage area, and the collected pest and disease information is stored in the database in a timely manner.

[0065] Drones are equipped with high-resolution cameras, multispectral sensors, radars and other equipment to monitor crop fields according to predetermined routes. Their equipment can capture detailed images and data of crop canopies, leaves, stems and other parts. Satellites carry optical and radar sensors to continuously monitor the earth's surface from high altitudes. These sensors can capture information such as the reflectivity, temperature and vegetation index of the surface. For insect population density extraction: use existing image recognition technology, such as deep learning models pre-trained with image data, to identify and count insects in the image, and combine the number of identified insects with the geographical location information in the image for spatial analysis to obtain a distribution map of insect population density; for lesion spread rate extraction: use existing image processing technology threshold segmentation or morphological processing to identify and extract lesions in the image, distinguish the boundaries between lesions and normal leaves, and calculate the area of ​​lesions. Then, perform time series analysis on the lesion area data at different time points, and calculate the growth rate of the lesion area to obtain the lesion spread rate; for crop damage area extraction: use the vegetation index (NDVI or GNDVI) in remote sensing data to evaluate the growth status of crops, compare the vegetation index data with the vegetation index of normal crops to identify the damaged area, and finally, use the existing geographic information system (GIS) technology to calculate the area of ​​the identified damaged area to obtain the crop damage area.

[0066] In this embodiment, a fixed cycle of once a week is set. On the weekly collection day, a drone is used to conduct remote sensing monitoring of a crop planting site to obtain pest and disease information such as insect population density, disease spot spread rate, and crop damage area.

[0067] S202: Compare the pest information of the current period with the pest information of the previous period, calculate the destructive growth rate of each pest, and obtain the destructive change result of each pest.

[0068] Specifically, extract the pest and disease information of the current cycle and the previous cycle from the database to ensure the integrity and consistency of the data, use professional data analysis software such as SPSS, Excel, etc. to compare and analyze the pest and disease information of the current cycle and the previous cycle, and calculate the destructive growth rate of each pest and disease, including the growth rate of insect population density or the growth rate of disease spot spread rate.

[0069] S203, recording the destructive change results of each pest and disease, and generating a dynamic change trend graph.

[0070] Specifically, the destructive changes of each pest and disease are recorded in the database in chronological order to ensure the integrity and traceability of the data; and professional drawing tools such as Visio and PowerPoint are used to generate dynamic change trend charts based on the recorded pest and disease information.

[0071] In the specific implementation process, the specific steps of step S3 are:

[0072] S301, input the initial hazard level and the destructive change results into a linear weighted prediction model, and output the estimated values ​​of the current pests and diseases.

[0073] Specifically, the linear weighted prediction model is: final level = initial level × α + destructive growth rate × β, where α and β are weight coefficients, α represents the weight of the initial hazard level, and β represents the weight of the destructive growth rate; it should be noted that the initial values ​​of α and β are set to α=0.5, β=0.5, and are adjusted based on the type of pests and diseases, specifically: for pests and diseases other than locust plagues, α is adjusted to 0.6 and β is adjusted to 0.4; for locust plagues, α is adjusted to 0.3 and β is adjusted to 0.7.

[0074] S302: Based on a plurality of preset threshold segments, by determining the threshold segment to which the estimated value of each current pest or disease belongs, the final hazard level corresponding to each current pest or disease and the threshold segment is divided.

[0075] Specifically, in this embodiment, the estimated values ​​1-2 are set as low hazard levels, the estimated values ​​2-3 are set as medium hazard levels, and the estimated values ​​3-4 are set as high hazard levels. The estimated values ​​of the current pests and diseases output in step S301 are compared with the preset threshold segments to determine which threshold segment these values ​​belong to. Finally, based on the judgment result of the value attribution, the current pests and diseases are divided into the final hazard levels corresponding to the threshold segments.

[0076] In the specific implementation process, the specific steps of step S4 are:

[0077] S401. Obtain remote sensing data of crops in real time through UAV multispectral sensors, including multispectral, hyperspectral and radar data.

[0078] Specifically, the drone takes off according to a predetermined route and continuously photographs crops through a multispectral sensor during the flight. It uses the function of multispectral sensors to cover multiple spectral bands such as visible light and near-infrared to obtain multispectral, hyperspectral and radar data of crops.

[0079] S402. Dynamically divide the multispectral, hyperspectral and radar data into different processing levels according to the estimated final damage levels of the current pests and diseases, wherein the processing levels specifically include low, medium and high levels.

[0080] Specifically, the processing levels include low, medium and high, which correspond to the low hazard level, medium hazard level and high hazard level of the final hazard level respectively.

[0081] In the specific implementation process, the specific steps of step S5 are:

[0082] S501. According to different processing levels, the remote sensing data of crops acquired in real time are processed accordingly, wherein the high-level remote sensing data uses a convolutional neural network to extract the texture features of disease spots, and the intermediate and low-level remote sensing data use fast feature extraction.

[0083] Specifically, the multispectral, hyperspectral and radar data of crop remote sensing data acquired in real time are first preprocessed, including radiation correction, geometric correction, image denoising, etc., to improve the data quality and analysis accuracy; then, for high-level remote sensing data, the trained convolutional neural network feature extraction is used, specifically, the texture features of the lesions are extracted through the convolutional layer, pooling layer, fully connected layer and other structures of the convolutional neural network; and for intermediate and low-level remote sensing data, fast feature extraction is used, among which the fast feature extraction methods include simple image processing techniques such as color space conversion, texture analysis, and shape analysis. In this embodiment, the training process of the convolutional neural network is as follows: image data containing pest and disease characteristics are screened out from historical remote sensing data sets, the collected image data are manually annotated, the type, location, size and other information of the pest and disease are marked, and the marked results are used as supervision information in the training process. Then, the input layer of the convolutional neural network receives the preprocessed remote sensing image data, the convolutional layer includes multiple convolution kernels for extracting local features in the image, the pooling layer downsamples the feature map output by the convolutional layer, the fully connected layer converts the feature vectors output by the convolutional layer and the pooling layer into feature vectors of fixed length, and the output layer outputs the type and degree of damage of the pest and disease. Finally, the image data of the pest and disease characteristics are input into the convolutional neural network as training data, and the output result is obtained through calculation of the convolutional layer, the pooling layer and the fully connected layer, and the trained convolutional neural network is obtained after training.

[0084] S502: Integrate the extracted feature data to form a feature data set.

[0085] Specifically, the extracted feature data is cleaned to remove redundant, erroneous or invalid data, and then standardized and normalized to ensure the consistency and comparability of the data. The processed feature data is integrated to form a feature data set.

[0086] For example, if it is determined that the processing level of leaf spot disease is advanced, a convolutional neural network is used to process the advanced remote sensing data to extract the texture features of leaf spot disease spots.

[0087] In the specific implementation process, the specific steps of step S6 are:

[0088] S601, input the feature data set into the constructed pest and disease database, and use the similarity measurement method to match the pest and disease features in the database to obtain a matching result.

[0089] Specifically, a database containing characteristic information of various pests and diseases is pre-established based on the pest and disease standards. Then, the extracted characteristic data set is input into the pest and disease database, and matched with the pest and disease characteristics in the database using a similarity measurement method, specifically Euclidean distance or cosine similarity. By calculating the similarity between the characteristic data set and the characteristics of each pest and disease in the database, the most similar pest and disease characteristic is found to obtain a matching result.

[0090] For example, a spectral reflectance feature data set of crops in a certain planting area is extracted through remote sensing technology and input into a pest and disease database. The database contains spectral feature information of a specific pest and disease. By calculating the cosine similarity between the feature data set and the spectral feature of the pest and disease, it is found that the similarity is as high as 0.95, where the similarity ranges from 0 to 1, and the larger the value, the higher the similarity.

[0091] S602: Determine the specific types of diseases and insect pests suffered by the crops in the current planting area and their degree of damage based on the matching results.

[0092] S603: Determine the specific types of pests and diseases suffered by the crops in the current planting area and their degree of damage, and send them to the management personnel or the planting farmers through remote notification.

[0093] Specifically, according to the result determined in step S602, prepare notification information including the specific type and degree of damage of pests and diseases, and select a remote notification method, including SMS, email or APP push, to send the notification content to management personnel or farmers, to ensure that the notification content can be conveyed to relevant personnel in a timely and accurate manner.

[0094] Embodiment 2

[0095] like Figure 2 As shown, a pest monitoring and early warning system based on remote sensing for a pest monitoring and early warning method based on remote sensing specifically includes:

[0096] The initial assessment module is used to obtain the historical data of pests and diseases of crops in the current planting area, and to assess and record the initial damage level of each pest and disease;

[0097] The data analysis module is used to obtain information on various pests and diseases of crops in the current planting area according to a preset period, and analyze the destructive changes of each pest and disease respectively to obtain the destructive change results of each pest and disease in the period;

[0098] The analysis and estimation module is used to estimate the final damage level of each pest and disease based on the destructive change results of each pest and disease during the period and the initial damage level of each pest and disease;

[0099] The data classification module is used to obtain remote sensing data of crops in real time and classify the remote sensing data processing levels corresponding to each pest based on the estimated final damage level of each pest;

[0100] A feature extraction module is used to extract the features of various types of remote sensing data based on the processing levels of the divided types of remote sensing data, and obtain a feature data set;

[0101] The matching and warning module is used to input the feature data set into the constructed pest and disease database for matching, output the pest and disease matching results, and generate corresponding warning notifications.

[0102] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0103] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0104] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

Claims

1. A pest monitoring and early warning method based on remote sensing, characterized in that: The monitoring and early warning method comprises the following steps: Obtain historical data on pests and diseases of crops in the current planting area, and evaluate and record the initial damage level of each pest and disease; Obtain information on various pests and diseases of crops in the current planting area according to the preset cycle, and analyze the destructive changes of each pest and disease respectively to obtain the destructive change results of each pest and disease in the cycle; Based on the destructive changes of each pest and disease during the period and the initial damage level of each pest and disease, the final damage level of each pest and disease is estimated respectively; Acquire remote sensing data of crops in real time, and classify the remote sensing data processing levels corresponding to each pest based on the estimated final damage level of each pest; Based on the processing levels of the divided remote sensing data, the features of the remote sensing data are extracted in stages to obtain feature data sets; The feature data set is input into the constructed pest and disease database for matching, the pest and disease matching results are output, and the corresponding early warning notification is generated.

2. A pest monitoring and early warning method based on remote sensing according to claim 1, characterized in that: The specific steps of obtaining the historical data of pests and diseases of crops in the current planting area and evaluating and recording the initial damage level of each pest and disease include: Retrieve the historical records of crop diseases and insect pests from the current planting area database, including the occurrence time, type and impact range data; Conduct quantitative analysis on the degree of damage of each type of pests and diseases in historical data and set the initial damage level, where the damage levels include level one, level two and level three; The initial hazard level is associated with the corresponding pest and disease type and recorded and stored in the local database.

3. The method for monitoring and early warning of pests and diseases based on remote sensing according to claim 1, characterized in that: The specific steps of obtaining information on various pests and diseases of crops in the current planting area according to a preset period, and analyzing the destructive changes of each pest and disease respectively to obtain the destructive change results of each pest and disease in the period include: Collect pest and disease information of the planting area according to a pre-set fixed period, where the pest and disease information of the planting area includes insect population density, disease spot spread rate and crop damage area; Compare the pest and disease information of the current period with that of the previous period, calculate the destructive growth rate of each pest and disease, and obtain the destructive change results of each pest and disease; Record the destructive changes of each pest and disease, and generate a dynamic change trend chart.

4. The method for monitoring and early warning of pests and diseases based on remote sensing according to claim 1, characterized in that: The method of estimating the final hazard level of each pest and disease based on the destructive change results of each pest and disease during the period and the initial hazard level of each pest and disease respectively comprises the following specific steps: The initial hazard level and destructive change results are input into the linear weighted prediction model to output the estimated values ​​of each pest and disease at present; Based on the preset multiple threshold segments, by determining the threshold segment to which the estimated value of each current pest and disease belongs, the final hazard level corresponding to each current pest and disease and the threshold segment is divided.

5. The method for monitoring and early warning of pests and diseases based on remote sensing according to claim 1, characterized in that: The real-time acquisition of crop remote sensing data and the classification of various remote sensing data processing levels corresponding to various pests and diseases based on the estimated final damage level of each pest and disease at present include: Real-time remote sensing data of crops can be obtained through drone multispectral sensors, including multispectral, hyperspectral and radar data; According to the estimated final damage level of each current pest and disease, the multispectral, hyperspectral and radar data are dynamically divided into different processing levels, among which the processing levels specifically include low, medium and high.

6. The method for monitoring and early warning of pests and diseases based on remote sensing according to claim 1, characterized in that: The processing levels of the divided remote sensing data are used to extract the features of the remote sensing data in a hierarchical manner to obtain the feature data set, and the specific steps include: According to different processing levels, the remote sensing data of crops acquired in real time are processed accordingly. Among them, the high-level remote sensing data uses convolutional neural network to extract the texture features of disease spots, and the intermediate and low-level remote sensing data adopts fast feature extraction; The extracted feature data are integrated to form a feature data set.

7. The method for monitoring and early warning of pests and diseases based on remote sensing according to claim 1, characterized in that: The specific steps of inputting the feature data set into the constructed pest and disease database for matching, outputting the pest and disease matching result, and generating the corresponding early warning notification include: Input the feature data set into the constructed pest and disease database, and use the similarity measurement method to match the pest and disease characteristics in the database to obtain the matching results; According to the matching results, determine the specific types of pests and diseases suffered by the crops in the current planting area and their degree of damage; The specific types of pests and diseases suffered by the crops in the current planting area and their degree of damage will be determined and sent to managers or farmers via remote notification.

8. A pest monitoring and early warning system based on remote sensing, applied to a pest monitoring and early warning method based on remote sensing according to any one of claims 1 to 7, characterized in that: The monitoring and early warning system comprises: The initial assessment module is used to obtain the historical data of pests and diseases of crops in the current planting area, and to assess and record the initial damage level of each pest and disease; The data analysis module is used to obtain information on various pests and diseases of crops in the current planting area according to a preset period, and analyze the destructive changes of each pest and disease respectively to obtain the destructive change results of each pest and disease in the period; The analysis and estimation module is used to estimate the final damage level of each pest and disease based on the destructive change results of each pest and disease during the period and the initial damage level of each pest and disease; The data classification module is used to obtain remote sensing data of crops in real time and classify the remote sensing data processing levels corresponding to each pest based on the estimated final damage level of each pest; A feature extraction module is used to extract the features of various types of remote sensing data based on the processing levels of the divided types of remote sensing data, and obtain a feature data set; The matching and warning module is used to input the feature data set into the constructed pest and disease database for matching, output the pest and disease matching results, and generate corresponding warning notifications.

Images