A remote sensing monitoring and early warning system for crop diseases based on time-series spectral information analysis

Abstract

This invention discloses a remote sensing monitoring and early warning system for crop diseases based on time-series spectral information analysis, belonging to the field of agricultural pest and disease monitoring technology. It utilizes computer graphics software to divide cornfields into proportionally equal areas and deploys sensors to monitor changes in the growth environment within each grid in real time, generating microenvironmental data. Based on this data, a microenvironmental difference index Xbd is calculated, identifying areas of concentrated microenvironmental differences and issuing risk analysis commands. Upon receiving these commands, a drone and multispectral sensors capture relevant spectral data of these areas. By analyzing the relevant spectral data, a multispectral fluctuation coefficient Xgp is constructed. Combined with the microenvironmental difference index Xbd and a trained risk assessment model, the calculated rust disease risk assessment index Zpg is compared with a preset threshold W, and a corresponding level of early warning is obtained based on the degree of rust disease risk.

Description

Technical Field

[0001] This invention relates to the field of agricultural pest and disease monitoring technology, specifically to a remote sensing monitoring and early warning system for crop diseases based on time-series spectral information analysis. Background Technology

[0002] In the field of agricultural pest and disease monitoring, with the advancement of modernization and intelligentization in global agricultural production, the application of precision monitoring and early warning technologies is receiving increasing attention. Early detection and rapid response to pests and diseases are crucial for ensuring crop yield and quality. This is especially true in the field of corn pest and disease monitoring. As corn is one of the main food crops, pests and diseases during its growth process have a significant impact on yield and quality. Therefore, conducting pest and disease monitoring for corn is particularly important. Southern corn rust, as a major disease affecting corn production, causes rust spots on corn leaves, water loss, and reduced photosynthetic capacity after infection, thereby reducing corn yield. Moreover, the disease has complex development conditions and spreads rapidly, seriously threatening the safety of corn cultivation. Therefore, research on monitoring and early warning of southern corn rust is urgently needed to improve control efficiency and reduce economic losses.

[0003] Traditional methods for monitoring and early warning of maize rust in southern China mainly rely on manual inspections and ground sample testing. This approach is not only time-consuming and labor-intensive, but also suffers from low monitoring frequency and insufficient spatial coverage. Ground sample testing requires laboratory analysis of the samples, a cumbersome and time-consuming process that cannot quickly provide feedback on disease changes. Furthermore, subjective judgments can lead to misdiagnosis and missed diagnoses, making it difficult to capture subtle changes in rust at different growth stages of maize. This results in early detection of rust, allowing it to spread. In addition, changes in the microenvironment, leaf color, and water content caused by rust are often overlooked, making it difficult to comprehensively assess the risk of rust occurrence. To address these shortcomings, a remote sensing monitoring and early warning system for crop diseases based on time-series spectral information analysis has been developed. This system integrates microenvironmental change information and spectral data using remote sensing technology, based on the impact of rust on maize growth, to achieve precise monitoring and dynamic early warning of maize fields. This helps disease control personnel take scientific and effective control measures, thereby increasing maize yield. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a remote sensing monitoring and early warning system for crop diseases based on time-series spectral information analysis, thus solving the problems mentioned in the background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solution: a remote sensing monitoring and early warning system for crop diseases based on time-series spectral information analysis, comprising a grid division module, a microenvironment monitoring module, a difference analysis module, a remote sensing capture module, a rust disease risk assessment module, and an early warning module;

[0006] The grid division module is used to draw the target corn planting field at a scale using computer drawing software, and to obtain several sets of planar grid units by combining the grid division function of the computer drawing software with the pre-set grid size and density, and to deploy sensors on the obtained several sets of planar grid units.

[0007] The microenvironment monitoring module is used to monitor the changes in the corn growth microenvironment within the several sets of planar grid units in real time based on the impact of rust on corn leaf photosynthesis, so as to construct microenvironment data information.

[0008] The difference analysis module is used to construct a microenvironment difference index Xbd within several sets of planar grid cells based on the microenvironment data information, determine the microenvironment difference clustering area in the corn planting field based on the value of the microenvironment difference index Xbd within several sets of planar grid cells, and issue a risk analysis command.

[0009] The remote sensing capture module is used to capture the spectral reflectance information of corn in the microenvironment difference aggregation area in real time by combining UAV remote sensing technology and multispectral sensors after receiving the risk analysis command, so as to monitor the chlorophyll content level and leaf moisture status in the microenvironment difference aggregation area and obtain relevant spectral data information.

[0010] The rust risk assessment module is used to analyze relevant spectral data information for each time period to construct a multispectral fluctuation coefficient Xgp, and by associating it with the microenvironmental difference index Xbd, and combining it with the trained rust risk assessment model, to fit and obtain the rust risk assessment index Zpg.

[0011] The early warning module is used to pre-set the assessment threshold W. By comparing and analyzing the rust risk assessment index Zpg with the assessment threshold W, the corresponding level of early warning is obtained based on the degree of rust risk in the current microenvironment difference aggregation area.

[0012] Preferably, the mesh partitioning module includes a planar partitioning unit and a deployment unit;

[0013] The planar division unit is used to draw the target cornfield to scale using computer graphics software based on high-resolution satellite images, determining the area and shape of the target cornfield. It also combines the grid division function of the computer graphics software with the pre-set grid size and density to generate a two-dimensional model grid. Through uniform grid division, several sets of planar grid units are obtained. The grid size and density are set according to the area of ​​the target cornfield, crop growth characteristics, and monitoring requirements.

[0014] The deployment unit is used to deploy sensors on several sets of planar grid cells. It requires that at least one non-dispersive infrared sensor be deployed at the center of each planar grid cell, and that the deployment positions be evenly distributed. The non-dispersive infrared sensors transmit data via wireless communication and establish a cloud server for data storage and processing.

[0015] Preferably, the microenvironment monitoring module includes a monitoring unit and a data preprocessing unit;

[0016] The monitoring unit is used to monitor the changes in the corn growth microenvironment within several planar grid units in real time, based on the impact of rust on corn leaf photosynthesis and in conjunction with non-dispersive infrared sensors deployed within these units, to construct microenvironmental data. This data includes carbon dioxide concentration values ​​(Cyt) within the corresponding monitoring time periods within the several planar grid units. Feature extraction is performed on these Cyt values ​​to obtain the minimum Cyt value for the corresponding monitoring time period within each of the several planar grid units. min and the maximum carbon dioxide concentration Cyt max ;

[0017] The data preprocessing unit is used to preprocess relevant data information within the microenvironment data information. The preprocessing includes noise removal, missing value filling, and data smoothing operations. The missing value filling methods include mean filling, median filling, interpolation filling, and regression filling. Dimensionless processing technology is used to scale the relevant data information within the disease data set so that the relevant data information within the microenvironment data information falls within the range of [0,1].

[0018] Preferably, the difference analysis module is used to construct a number of microenvironment difference indices Xbd within planar grid cells based on the microenvironment data information, using the microenvironment difference index Xbd within the i-th group of planar grid cells as an example. i For example, it is obtained in the following way:

[0019]

[0020] In the formula, This represents the maximum carbon dioxide concentration within the i-th group of planar grid cells. This represents the minimum carbon dioxide concentration within the i-th group of planar grid cells;

[0021] Based on the above, the microenvironment difference index Xbd within the i-th group of planar grid cells is obtained. i The method involves acquiring the microenvironmental difference index Xbd within several groups of planar grid cells, and then calculating the mean of the microenvironmental difference index using a statistical averaging algorithm.

[0022] By comparing the microenvironmental difference index Xbd within several groups of planar grid cells with the mean... Perform size comparisons and obtain values ​​exceeding the mean. The corresponding planar grid cell contains the microenvironmental difference index Xbd, and these are all marked as microenvironmental difference clusters. When the microenvironmental difference clusters exceed 5% of the total number of planar grid cells, a risk analysis command is issued through the cloud server.

[0023] Preferably, after receiving the risk analysis command from the cloud server, the remote sensing capture module combines UAV remote sensing technology and uses the multispectral sensor carried by the UAV to capture the spectral reflectance information of corn in the microenvironment difference aggregation area in real time, so as to monitor the chlorophyll content level and leaf moisture status in the microenvironment difference aggregation area and construct relevant spectral data information. The relevant spectral data information includes the visible light band Bkj and the near-infrared light band Bjh for the corresponding time period.

[0024] Preferably, the rust risk assessment module includes a time-series spectral analysis unit, a model training unit, and a comprehensive rust risk assessment unit;

[0025] The time-series spectral analysis unit is used to extract features from the relevant spectral data information. After linear normalization, the visible light fluctuation coefficient Xkj and the near-infrared light fluctuation coefficient Xjh are obtained respectively. The visible light fluctuation coefficient Xkj and the near-infrared light fluctuation coefficient Xjh are obtained by the following formulas respectively.

[0026]

[0027] In the formula, Bkj j The visible light band at time period j represents the region of microenvironmental difference aggregation. Bjh represents the average visible light band value of the region where microenvironmental differences are concentrated. j This represents the near-infrared band of the j-th time period, representing the region of microenvironmental difference aggregation. The near-infrared light band average value represents the area of ​​microenvironmental difference aggregation, j = 1, 2, 3, ..., n, where n represents the monitoring period.

[0028] Preferably, the obtained visible light fluctuation coefficient Xkj and near-infrared light fluctuation coefficient Xjh are correlated to construct a multispectral fluctuation coefficient Xgp, which is obtained by the following formula;

[0029]

[0030] In the formula, γ and δ represent the weight values ​​of the visible light wave coefficient Xkj and the near-infrared light wave coefficient Xjh, respectively, and A represents the first correction constant.

[0031] Preferably, the model training unit is used to construct a basic model using convolutional neural network technology, and to train and test the basic model with microenvironmental data information and relevant spectral data information. The trained basic model is used as a related rust disease association model. Feature information within the related rust disease association model is obtained, and the obtained feature information is used to train and test the related rust disease association model. Combined with the risk analysis command issued by the cloud server, the trained related rust disease association model is used as a rust disease risk assessment model.

[0032] Preferably, the rust risk comprehensive assessment unit is used to correlate the multispectral fluctuation coefficient Xgp of the microenvironment difference aggregation area with the corresponding microenvironment difference index Xbd, and combine it with the trained rust risk assessment model to fit and obtain the rust risk assessment index Zpg, which is obtained by the following formula;

[0033]

[0034] In the formula, Both σ and B represent weight values, and B represents the second correction constant.

[0035] Preferably, the early warning module is used to pre-set an assessment threshold W, and by comparing and analyzing the rust risk assessment index Zpg with the assessment threshold W, obtain a corresponding level of early warning based on the degree of rust risk in the current microenvironment difference aggregation area, as detailed below;

[0036] If the rust risk assessment index Zpg > assessment threshold W, it indicates that the corn in the current microenvironment difference cluster area is suffering from rust, generating a first risk warning and notifying disease control personnel that the corn in the current area is suffering from rust. At the same time, it requires on-site investigation of the corn in the microenvironment difference cluster area to confirm the actual situation and the extent of spread, and to take remedial measures such as spraying fungicides and applying fertilizers to the microenvironment difference cluster area to prevent the rust from spreading further.

[0037] If the rust risk assessment index Zpg ≤ assessment threshold W, it indicates that the corn in the current microenvironment difference cluster area is not infected with rust, a second risk warning is generated, and the current area is continuously monitored.

[0038] This invention provides a remote sensing monitoring and early warning system for crop diseases based on time-series spectral information analysis, which has the following beneficial effects:

[0039] (1) By integrating grid division, microenvironment monitoring, differential analysis, remote sensing capture, rust risk assessment and early warning modules, accurate monitoring and early dynamic early warning of maize rust in southern China have been achieved, overcoming the shortcomings of traditional monitoring methods. First, efficient grid division and sensor deployment ensured coverage of maize fields and real-time data collection, thereby enhancing the monitoring frequency and spatial resolution. Second, the microenvironment monitoring module, combined with differential analysis, can identify potential risk areas in a timely manner, providing early warnings and reducing the possibility of disease spread. In addition, the rust risk assessment module targets areas with concentrated microenvironmental differences and combines multispectral data and microenvironmental data information, using a training model for risk fitting, which improves the accuracy and reliability of rust risk assessment. In summary, this series of improvements effectively identifies rust risk areas, provides timely early warning information, and achieves accurate monitoring and early dynamic early warning of maize rust in southern China. It improves the accuracy and response speed of monitoring, helps disease control personnel take scientific and effective control measures, prevents further spread of rust, and thus increases maize yield, reduces economic losses and ensures food security.

[0040] (2) Through precise grid division and sensor deployment, real-time monitoring and data acquisition of the microenvironment of maize growth are realized, overcoming the problems of insufficient coverage and response delay in traditional monitoring methods. The microenvironment monitoring module uses carbon dioxide concentration information obtained by non-dispersive infrared sensors, and ensures the accuracy and reliability of the data through feature extraction and data preprocessing. By obtaining the microenvironment difference index Xbd, the microenvironment difference aggregation area is identified. When the microenvironment difference aggregation area exceeds 5% of the total number of planar grid units, a risk analysis command is issued. This systematic and automated monitoring method can not only identify potential risk areas in a timely manner and provide early warnings to reduce the possibility of disease spread, but also effectively prevent and mitigate the risk of maize rust spreading in the south.

[0041] (3) By using multispectral sensors and time-series spectral information analysis carried by UAVs, efficient monitoring and assessment of maize rust in southern China was achieved. The advantage of this system is that it can capture spectral data of microenvironmental difference aggregation areas in real time, extract chlorophyll content and leaf moisture status, and then calculate the visible light fluctuation coefficient Xkj and near-infrared light fluctuation coefficient Xjh, construct the multispectral fluctuation coefficient Xgp, and improve the accuracy of disease risk assessment. The rust risk assessment model constructed by using convolutional neural networks makes risk prediction more scientific and reliable, and can generate risk warnings in real time to guide disease control personnel to take timely measures, thereby effectively preventing the further spread of the disease. This systematic monitoring and early warning method solves the problems of slow response and insufficient coverage in traditional methods, and also improves the efficiency of maize rust control in southern China, ensuring food production security. Attached Figure Description

[0042] Figure 1 This is a block diagram of a remote sensing monitoring and early warning system for crop diseases based on time-series spectral information analysis, according to the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] Example 1

[0045] Please see Figure 1 This invention provides a remote sensing monitoring and early warning system for crop diseases based on time-series spectral information analysis, including a grid division module, a microenvironment monitoring module, a difference analysis module, a remote sensing capture module, a rust disease risk assessment module, and an early warning module;

[0046] The grid division module is used to draw the target corn planting field at a scale using computer drawing software, and to obtain several sets of planar grid units by combining the grid division function of the computer drawing software with the pre-set grid size and density, and to deploy sensors on the obtained several sets of planar grid units.

[0047] The microenvironment monitoring module is used to monitor the changes in the corn growth microenvironment within the several sets of planar grid units in real time based on the impact of rust on corn leaf photosynthesis, so as to construct microenvironment data information.

[0048] The difference analysis module is used to construct a microenvironment difference index Xbd within several sets of planar grid cells based on the microenvironment data information, determine the microenvironment difference clustering area in the corn planting field based on the value of the microenvironment difference index Xbd within several sets of planar grid cells, and issue a risk analysis command.

[0049] The remote sensing capture module is used to capture the spectral reflectance information of corn in the microenvironment difference aggregation area in real time by combining UAV remote sensing technology and multispectral sensors after receiving the risk analysis command, so as to monitor the chlorophyll content level and leaf moisture status in the microenvironment difference aggregation area and obtain relevant spectral data information.

[0050] The rust risk assessment module is used to analyze relevant spectral data information for each time period to construct a multispectral fluctuation coefficient Xgp, and by associating it with the microenvironmental difference index Xbd, and combining it with the trained rust risk assessment model, to fit and obtain the rust risk assessment index Zpg.

[0051] The early warning module is used to pre-set the assessment threshold W. By comparing and analyzing the rust risk assessment index Zpg with the assessment threshold W, the corresponding level of early warning is obtained based on the degree of rust risk in the current microenvironment difference aggregation area.

[0052] In this embodiment, the comprehensive application of multiple modules enables accurate monitoring and early dynamic warning of maize rust in southern China. Firstly, the grid division module ensures refined monitoring of the planting field, allowing real-time capture of microenvironmental changes in each area and enhancing data accuracy. The microenvironmental monitoring module tracks key growth parameters in real time, including carbon dioxide concentration (Cyt), facilitating timely identification of crop health status. The differential analysis module, by constructing a microenvironmental differential index (Xbd), can quickly identify potential disease risk areas and issue risk analysis commands, ensuring rapid response. Subsequently, the remote sensing capture module utilizes drones and multispectral imaging... Sensors monitor spectral reflectance information in real time, providing crucial data on chlorophyll content and moisture status, laying the foundation for subsequent risk assessment. The rust risk assessment module combines the microenvironmental difference index Xbd with multispectral data to generate a rust risk assessment index Zpg. By comparing this index with a set threshold, it can provide clear early warning information. This series of measures not only enables disease control personnel to take timely and effective control measures to reduce economic losses, but also improves the scientific and intelligent level of monitoring and early warning for maize rust in southern China. Especially in the context of addressing climate change and frequent disease outbreaks, it demonstrates significant practical application value.

[0053] Example 2

[0054] Please refer to Figure 1 Specifically: the mesh partitioning module includes planar partitioning units and deployment units;

[0055] The planar division unit is used to draw the target cornfield to scale using computer graphics software based on high-resolution satellite images, determining the area and shape of the target cornfield. It also combines the grid division function of the computer graphics software with the pre-set grid size and density to generate a two-dimensional model grid. Through uniform grid division, several sets of planar grid units are obtained. The grid size and density are set according to the area of ​​the target cornfield, crop growth characteristics, and monitoring requirements.

[0056] The deployment unit is used to deploy sensors on several sets of planar grid cells. It requires that at least one non-dispersive infrared sensor be deployed at the center of each planar grid cell, and that the deployment positions be evenly distributed. The non-dispersive infrared sensors transmit data via wireless communication and establish a cloud server for data storage and processing.

[0057] The microenvironment monitoring module includes a monitoring unit and a data preprocessing unit;

[0058] Specifically, the monitoring unit is used to monitor the changes in the maize growth microenvironment within several planar grid units in real time, based on the impact of rust disease on maize leaf photosynthesis and in conjunction with non-dispersive infrared sensors deployed within these planar grid units. This enables the construction of microenvironmental data, which includes carbon dioxide concentration values ​​(Cyt) within the corresponding monitoring time periods within the several planar grid units. By extracting features from the Cyt values ​​within the corresponding monitoring time periods within the several planar grid units, the minimum Cyt value of carbon dioxide concentration within the corresponding monitoring time periods within the several planar grid units is obtained. min and the maximum carbon dioxide concentration Cyt max ;

[0059] It should be noted that the microenvironment in which corn grows refers to environmental conditions within a small area, including temperature, humidity, light, soil properties, and gas concentration. These factors can differ from the surrounding environment in a small space. In particular, the difference in carbon dioxide concentration in the corn's microenvironment can reflect the corn's rust disease status. This is because carbon dioxide is an important raw material for plant photosynthesis. Healthy corn plants absorb carbon dioxide and release oxygen during photosynthesis. Rust infection leads to leaf damage, reducing photosynthetic efficiency and thus reducing carbon dioxide absorption. As the disease worsens, the plant's metabolic activities are severely inhibited. Therefore, monitoring differences in carbon dioxide concentration can provide early warning signals, help identify potential rust risks, and provide a basis for timely prevention and control measures.

[0060] The data preprocessing unit is used to preprocess relevant data information within the microenvironment data information. The preprocessing includes noise removal, missing value filling, and data smoothing operations. The missing value filling methods include mean filling, median filling, interpolation filling, and regression filling. Dimensionless processing technology is used to scale the relevant data information within the disease data set so that the relevant data information within the microenvironment data information falls within the range of [0,1].

[0061] In this embodiment, the planar division unit utilizes high-definition satellite imagery and computer graphics software to generate a uniform and reasonable two-dimensional grid model based on the actual conditions of the target planting field. This ensures that each planar grid unit covers the crop's growth characteristics and health status. The deployment unit ensures that each grid is equipped with at least one non-dispersive infrared sensor, guaranteeing the comprehensiveness and uniformity of data collection. The microenvironment monitoring module provides important information on crop photosynthesis and growth status by tracking the carbon dioxide concentration (Cyt) value in the microenvironment data in real time. This is crucial for timely detection of potential risks of rust disease. In addition, the data preprocessing unit ensures the accuracy and reliability of the monitoring data by removing noise and filling in missing values, improving the scientific nature of subsequent analysis. Through these methods, not only can the health status of crops be effectively monitored, but disease risks can also be quickly identified and responded to, providing scientific and accurate decision support for disease control personnel. This efficient monitoring and early warning mechanism reduces the labor costs and information lag problems commonly found in traditional monitoring methods, making crop disease control more intelligent and systematic, thereby improving the yield and quality of corn and protecting the economic interests of farmers.

[0062] Example 3

[0063] Please refer to Figure 1 Specifically: the difference analysis module is used to construct several sets of microenvironment difference indices Xbd within planar grid cells based on the microenvironment data information, using the microenvironment difference index Xbd within the i-th set of planar grid cells as an example. i For example, it is obtained in the following way:

[0064]

[0065] In the formula, This represents the maximum carbon dioxide concentration within the i-th group of planar grid cells. This represents the minimum carbon dioxide concentration within the i-th group of planar grid cells;

[0066] Based on the above, the microenvironment difference index Xbd within the i-th group of planar grid cells is obtained. i The method involves acquiring the microenvironmental difference index Xbd within several groups of planar grid cells, and then calculating the mean of the microenvironmental difference index using a statistical averaging algorithm.

[0067] By comparing the microenvironmental difference index Xbd within several groups of planar grid cells with the mean... Perform size comparisons and obtain values ​​exceeding the mean. The corresponding planar grid cell contains the microenvironmental difference index Xbd, and these are all marked as microenvironmental difference clusters. When the microenvironmental difference clusters exceed 5% of the total number of planar grid cells, a risk analysis command is issued through the cloud server.

[0068] In this embodiment, through in-depth analysis of microenvironment data, a microenvironmental difference index Xbd was constructed within a planar grid cell, thereby achieving precise monitoring of crop growth status. Specifically, the microenvironmental difference index Xbd was obtained by calculating and analyzing the differences in carbon dioxide concentration within each planar grid cell, and the mean value of the microenvironmental difference index was further obtained. This process enables the system to identify which planar grid cells have microenvironmental conditions that deviate from the mean, thereby marking areas of concentrated microenvironmental differences. When the area of ​​concentrated microenvironmental differences exceeds 5% of the total number of planar grid cells, a risk analysis command is automatically issued. This process not only improves the early warning capability for potential risks of disease occurrence, but this systematic and automated monitoring method can also identify potential risk areas in a timely manner, provide early warnings, reduce the possibility of disease spread, and effectively prevent and mitigate the risk of the spread of corn rust in southern China.

[0069] Example 4

[0070] Please refer to Figure 1 Specifically: After receiving the risk analysis command from the cloud server, the remote sensing capture module combines UAV remote sensing technology and uses the multispectral sensor carried by the UAV to capture the spectral reflectance information of corn in the microenvironment difference aggregation area in real time, so as to monitor the chlorophyll content level and leaf moisture status in the microenvironment difference aggregation area and construct relevant spectral data information. The relevant spectral data information includes the visible light band Bkj and the near-infrared light band Bjh for the corresponding time period.

[0071] In this embodiment, by combining UAV remote sensing technology with multispectral sensors, it is possible to efficiently and in real time acquire spectral reflectance information of maize within areas of microenvironmental difference aggregation, especially monitoring chlorophyll content and leaf moisture status. This precise data acquisition not only improves the ability to assess the impact of rust on maize health but also enables timely identification of potential growth problems. By acquiring visible light band Bkj and near-infrared light band Bjh at specific time periods, disease control personnel can make more accurate decisions based on scientific data. This real-time monitoring method effectively supplements the shortcomings of traditional monitoring methods, reduces labor costs and time consumption, and enhances the monitoring capability of rust's impact on maize health.

[0072] Example 5

[0073] Please refer to Figure 1Specifically: the rust risk assessment module includes a time-series spectral analysis unit, a model training unit, and a comprehensive rust risk assessment unit;

[0074] The time-series spectral analysis unit is used to extract features from the relevant spectral data information. After linear normalization, the visible light fluctuation coefficient Xkj and the near-infrared light fluctuation coefficient Xjh are obtained respectively. The visible light fluctuation coefficient Xkj and the near-infrared light fluctuation coefficient Xjh are obtained by the following formulas respectively.

[0075]

[0076] In the formula, Bkj j The visible light band at time period j represents the region of microenvironmental difference aggregation. Bjh represents the average visible light band value of the region where microenvironmental differences are concentrated. j This represents the near-infrared band of the j-th time period, representing the region of microenvironmental difference aggregation. The near-infrared light band average value represents the area of ​​microenvironmental difference aggregation, j = 1, 2, 3, ..., n, where n represents the monitoring period.

[0077] Specifically, the obtained visible light fluctuation coefficient Xkj and near-infrared light fluctuation coefficient Xjh are correlated to construct a multispectral fluctuation coefficient Xgp, which is obtained by the following formula;

[0078]

[0079] In the formula, γ and δ represent the weight values ​​of the visible light wave coefficient Xkj and the near-infrared light wave coefficient Xjh, respectively, and A represents the first correction constant.

[0080] Specifically, the model training unit is used to construct a basic model using convolutional neural network technology, and to train and test the basic model with microenvironmental data information and relevant spectral data information. The trained basic model is used as a related rust disease association model. Feature information within the related rust disease association model is obtained, and the obtained feature information is used to train and test the related rust disease association model. Combined with the risk analysis command issued by the cloud server, the trained related rust disease association model is used as a rust disease risk assessment model.

[0081] Specifically, the rust disease risk comprehensive assessment unit is used to correlate the multispectral fluctuation coefficient Xgp of the microenvironment difference aggregation area with the corresponding microenvironment difference index Xbd, and combine it with the trained rust disease risk assessment model to fit and obtain the rust disease risk assessment index Zpg. The rust disease risk assessment index Zpg is obtained by the following formula.

[0082]

[0083] In the formula, Both σ and B represent weight values, and B represents the second correction constant.

[0084] In this embodiment, the synergistic effect of three units—time-series spectral analysis, model training, and comprehensive evaluation—improves the monitoring and early warning capabilities for maize rust. First, the time-series spectral analysis unit extracts features from relevant spectral data to obtain the visible light fluctuation coefficient Xkj and near-infrared light fluctuation coefficient Xjh, thereby capturing the chlorophyll content level and leaf moisture status in areas of concentrated microenvironmental differences. This precise data processing enables timely identification of potential disease risks, providing a data foundation for subsequent decision-making. Next, the model training unit uses convolutional neural network technology to construct a rust-related model, combining microenvironmental data and spectral information for in-depth analysis. This not only… This process improves the model's predictive accuracy and enhances its adaptability. Finally, the comprehensive rust risk assessment unit combines the multispectral fluctuation coefficient Xgp with the corresponding regional microenvironmental difference index Xbd. Through scientific weight allocation and correction, it fits a more reliable rust risk assessment index Zpg. This comprehensive assessment method not only provides real-time risk analysis instructions to help disease control personnel take timely control measures, but also effectively reduces economic losses caused by diseases, achieving refined agricultural management. In summary, the implementation of this process enhances the early warning and response capabilities for crop diseases such as rust, promoting the sustainable development of agricultural production and increasing yields.

[0085] Example 6

[0086] Please refer to Figure 1 Specifically: The early warning module is used to pre-set the assessment threshold W, and by comparing and analyzing the rust risk assessment index Zpg with the assessment threshold W, it obtains the corresponding level of early warning based on the degree of rust risk in the current microenvironment difference aggregation area. The specific content is as follows:

[0087] If the rust risk assessment index Zpg > assessment threshold W, it indicates that the corn in the current microenvironment difference cluster area is suffering from rust, generating a first risk warning and notifying disease control personnel that the corn in the current area is suffering from rust. At the same time, it requires on-site investigation of the corn in the microenvironment difference cluster area to confirm the actual situation and the extent of spread, and to take remedial measures such as spraying fungicides and applying fertilizers to the microenvironment difference cluster area to prevent the rust from spreading further.

[0088] If the rust risk assessment index Zpg ≤ assessment threshold W, it indicates that the corn in the current microenvironment difference cluster area is not infected with rust, a second risk warning is generated, and the current area is continuously monitored.

[0089] In this embodiment, an assessment threshold W is preset. By comparing and analyzing the rust disease risk assessment index Zpg with the assessment threshold W, the real-time response capability and accuracy of rust disease early warning are enhanced. When the rust disease risk assessment index Zpg exceeds the assessment threshold W, the system immediately generates a first risk warning, promptly notifying disease control personnel, thereby prompting rapid remedial measures to effectively prevent the further spread of rust and reduce potential economic losses. When the rust disease risk assessment index Zpg does not exceed the assessment threshold W, the system generates a second risk warning to ensure continuous monitoring and provide safety assurance. This dual early warning mechanism not only improves the efficiency of monitoring the health status of corn crops but also provides scientific basis for disease control personnel, helping to formulate more reasonable disease control strategies and ultimately achieving higher agricultural yields and quality.

[0090] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A crop disease remote sensing monitoring and early warning system based on time series spectral information analysis, characterized by: The grid division module, the micro-environment monitoring module, the difference analysis module, the remote sensing capturing module, the rust disease risk assessment module and the early warning module are included. The grid division module is used for drawing the target corn planting field in equal proportion by computer drawing software, combining the grid division function of the computer drawing software and the pre-set grid size and density, obtaining a plurality of groups of plane grid units, and deploying sensors on the obtained plurality of groups of plane grid units. The micro-environment monitoring module is used for monitoring the micro-environment change of the corn in the plurality of groups of plane grid units in real time according to the influence of the rust disease on the photosynthesis of the corn leaves, so as to construct micro-environment data information. The difference analysis module is used for constructing a micro-environment difference index Xbd in the plurality of groups of plane grid units according to the micro-environment data information, determining the micro-environment difference aggregation area in the corn planting field according to the numerical value of the micro-environment difference index Xbd in the plurality of groups of plane grid units, and issuing a risk analysis instruction. The remote sensing capturing module is used for capturing the corn spectral reflection information in the micro-environment difference aggregation area in real time by using a multi-spectral sensor in combination with the unmanned aerial vehicle remote sensing technology after receiving the risk analysis instruction, so as to monitor the chlorophyll content level and the leaf water condition in the micro-environment difference aggregation area and obtain related spectral data information. The rust disease risk assessment module is used for analyzing the related spectral data information at each time period, constructing a multi-spectral fluctuation coefficient Xgp, and fitting a rust disease risk assessment index Zpg by associating the rust disease risk assessment index Zpg with the micro-environment difference index Xbd in combination with the trained rust disease risk assessment model. The early warning module is used for pre-setting an evaluation threshold W, comparing and analyzing the rust disease risk assessment index Zpg with the evaluation threshold W, obtaining a corresponding grade warning according to the rust disease risk degree in the current micro-environment difference aggregation area. The difference analysis module is configured to construct microenvironment difference indexes Xbd in a plurality of groups of planar grid units according to the microenvironment data information, and the microenvironment difference index Xbd in the i-th group of planar grid units is calculated according to the following formula: i For example, the microenvironment difference index Xbd is obtained in the following manner: ； wherein represents a maximum value of the carbon dioxide concentration within the i-th group of planar grid cells, represents a minimum value of the carbon dioxide concentration within the i-th group of planar grid cells. According to the above-mentioned way of obtaining the microenvironment difference index Xbd in the i-th group of planar grid cells i , the microenvironment difference indexes Xbd in several groups of planar grid cells are obtained respectively, and the mean value of the microenvironment difference indexes is calculated according to a statistical mean value algorithm . By comparing the microenvironmental difference index Xbd within several groups of planar grid cells with the mean... Perform size comparisons and obtain values ​​exceeding the mean. The corresponding planar grid cell contains the microenvironmental difference index Xbd, and all of them are marked as microenvironmental difference cluster areas. When the microenvironmental difference cluster area exceeds 5% of the total number of planar grid cells, a risk analysis command is issued through the cloud server. The rust disease risk assessment module includes a time series spectral analysis unit, a model training unit and a rust disease risk comprehensive assessment unit. The time series spectral analysis unit is used for extracting features from the related spectral data information, obtaining a visible light fluctuation coefficient Xkj and a near-infrared light fluctuation coefficient Xjh after linear normalization, and obtaining the visible light fluctuation coefficient Xkj and the near-infrared light fluctuation coefficient Xjh by the following formulas respectively. ； In the formula, a visible light band of the jth period of the microenvironment difference aggregation area, a visible light band mean of the microenvironment difference aggregation area, a near-infrared light band of the jth period of the microenvironment difference aggregation area, a near-infrared light band mean of the microenvironment difference aggregation area, j=1, 2, 3,..., n, n represents a monitoring period; The obtained visible light fluctuation coefficient Xkj and near-infrared light fluctuation coefficient Xjh are associated to construct a multi-spectral fluctuation coefficient Xgp, and the multi-spectral fluctuation coefficient Xgp is obtained by the following formula. ； In the formula, and respectively represent the weight values of the visible light fluctuation coefficient Xkj and the near-infrared light fluctuation coefficient Xjh, and A represents a first correction constant; The rust disease risk comprehensive assessment unit is used for associating the multi-spectral fluctuation coefficient Xgp of the micro-environment difference aggregation area with the micro-environment difference index Xbd of the corresponding area, fitting a rust disease risk assessment index Zpg in combination with the trained rust disease risk assessment model, and obtaining the rust disease risk assessment index Zpg by the following formula. ； wherein and are represented as weight values, B is represented as a second normalizing constant. 2.The crop disease remote sensing monitoring and early warning system based on time-series spectral information analysis according to claim 1, characterized in that: The grid division module includes a plane division unit and a deployment unit. The plane division unit is used for drawing a satellite high-definition image of the target corn planting field by computer drawing software in proportion, determining the area and shape of the target corn planting field, and combining the grid division function of the computer drawing software and the pre-set grid size and density to generate a grid of a two-dimensional model, and obtaining a plurality of groups of plane grid units through uniform grid division; wherein the grid size and density are set according to the area size of the target corn planting field, the growth characteristics of the crops and the monitoring requirements; The deployment unit is used for deploying sensors on the obtained plurality of groups of plane grid units, requiring at least one non-dispersive infrared sensor to be deployed at the center position of each plane grid unit, and the deployed positions are uniformly distributed, the non-dispersive infrared sensor transmits data through wireless communication, and a cloud server is established to store and process data. 3.The crop disease remote sensing monitoring and early warning system based on time-series spectral information analysis according to claim 2, characterized in that: The micro-environment monitoring module includes a monitoring unit and a data preprocessing unit; The monitoring unit is used for real-time monitoring of the change of the corn growth microenvironment in the plurality of groups of planar grid units according to the influence of the rust on the photosynthesis of the corn leaves, in combination with the non-dispersed infrared sensors arranged in the plurality of groups of planar grid units, to construct microenvironment data information, the microenvironment data information including carbon dioxide concentration values Cyt of corresponding monitoring time periods in the plurality of groups of planar grid units, and the carbon dioxide concentration values Cyt of the corresponding monitoring time periods in the plurality of groups of planar grid units are subjected to feature extraction to obtain carbon dioxide concentration minimum values Cyt min and carbon dioxide concentration maximum values Cyt max of the corresponding monitoring time periods in the plurality of groups of planar grid units. The data preprocessing unit is configured to preprocess the relevant data information in the microenvironment data information, the preprocessing including removing noise, filling in missing values, and data smoothing operation, wherein the method for filling in missing values includes mean filling, median filling, interpolation filling, and regression filling; and using dimensionless processing technology to scale the relevant data information in the disease data set, so that the range of the relevant data information in the microenvironment data information falls between 0 and 1. between 0 and 1.

4. The crop disease remote sensing monitoring and early warning system based on time series spectral information analysis according to claim 1, characterized in that: The remote sensing capture module, after receiving the risk analysis instruction issued by the cloud server, combines the unmanned aerial vehicle remote sensing technology, uses the multispectral sensor carried by the unmanned aerial vehicle to capture the corn spectral reflectance information in the micro-environment difference aggregation area in real time, monitors the chlorophyll content level and leaf water condition in the micro-environment difference aggregation area, and constructs related spectral data information, the related spectral data information includes visible light band Bkj and near-infrared light band Bjh in the corresponding period.

5. The crop disease remote sensing monitoring and early warning system based on time series spectral information analysis according to claim 1, characterized in that: The model training unit is used for constructing a basic model using convolutional neural network technology, training and testing the basic model with micro-environment data information and related spectral data information, and taking the trained basic model as a related rust correlation model, respectively acquiring feature information in the related rust correlation model, and training and testing the related rust correlation model with the acquired feature information, combining the risk analysis instruction issued by the cloud server, and taking the trained related rust correlation model as a rust risk assessment model. 6.The crop disease remote sensing monitoring and early warning system based on time-series spectral information analysis according to claim 1, characterized in that: The early warning module is used for pre-setting an evaluation threshold W, comparing and analyzing the rust risk assessment index Zpg with the evaluation threshold W, acquiring a corresponding grade warning according to the rust disease risk degree in the current micro-environment difference aggregation area, and the specific content is as follows; If the rust risk assessment index Zpg is greater than the evaluation threshold W, it indicates that the corn in the current micro-environment difference aggregation area is affected by rust, a first risk warning is generated, and the disease control personnel is notified that the corn in the current area is affected by rust, and is required to conduct on-site investigation on the corn in the micro-environment difference aggregation area, confirm the actual situation and the spread degree, and take remedial measures such as spraying fungicides and fertilizing on the corn in the micro-environment difference aggregation area to prevent the rust from further spreading; If the rust risk assessment index Zpg is less than or equal to the evaluation threshold W, it indicates that the corn in the current micro-environment difference aggregation area is not affected by rust, a second risk warning is generated, and the current area is continuously monitored.

Images