Dynamic early warning and prevention and control system based on forestry pest identification

Abstract

The application relates to the technical field of pest and disease identification, in particular to a dynamic early warning and prevention and control system based on forestry pest and disease identification, which obtains crown layer leaf attribute data of target forest area tree species and excitation source physical characteristic parameters of a patrol unmanned aerial vehicle, determines aerodynamic pressure required for turning over of the crown layer leaves according to the leaf attribute data, determines aerodynamic excitation control parameters in combination with the excitation source physical characteristic parameters and a final parameter mapping model, controls the unmanned aerial vehicle to force the leaves to turn over by using a downward airflow to obtain leaf back image data, and adaptively corrects the control parameters according to an effective leaf back exposure rate; image recognition is performed on the leaf back image data to obtain pest and disease characteristic data, and a leaf stiffness index is calculated through time domain analysis; pest and disease severity is generated according to the pest and disease characteristic data and the leaf stiffness index, and a graded early warning is performed; the application can effectively obtain leaf back image information, and realizes early and accurate identification and dynamic early warning of pests and diseases.

Description

Technical Field

[0001] This invention relates to the field of pest and disease identification technology, specifically to a dynamic early warning and control system based on forest pest and disease identification. Background Technology

[0002] In forestry production, early monitoring and precise control of pests and diseases are crucial for ensuring the health of forest ecosystems and the economic value of timber. UAV-based optical inspection is a widely used monitoring method in forestry pest and disease monitoring. It uses visible light or multispectral cameras to take aerial images of the tree canopy and uses image recognition algorithms to detect pest and disease characteristics such as leaf color changes and texture abnormalities. It can achieve high monitoring efficiency and automation and is suitable for continuous online monitoring tasks in large-area forest areas.

[0003] However, in actual forestry pest and disease monitoring environments, many forestry pests mainly feed and lay eggs on the underside of leaves, and typical symptoms of pests and diseases are also concentrated on the underside of leaves. The natural growth posture of tree canopy leaves usually means that the front of the leaves faces the light source, while the underside of the leaves is in a shaded state. Existing technologies usually analyze information based on images of the front of the leaves or fixed observation angles, making it difficult to fully obtain image information of the underside of the leaves. Therefore, it is difficult to fully reflect the pest and disease characteristics of the underside of the leaves during the monitoring process, which leads to the failure of forestry pest and disease identification and early warning to achieve the expected results. Summary of the Invention

[0004] In view of the problems in related technologies, the present invention provides a dynamic early warning and control system based on forest pest and disease identification to overcome the above-mentioned technical problems existing in the existing related technologies.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a dynamic early warning and control system based on forestry pest and disease identification, comprising:

[0006] The data acquisition module is used to acquire canopy and leaf attribute data of tree species in the target forest area and physical characteristic parameters of the excitation source of the inspection drone;

[0007] The control parameter solving module is used to determine the aerodynamic pressure required to cause the canopy blades to flip based on the canopy blade attribute data, and to determine the aerodynamic excitation control parameters required to characterize the UAV to achieve blade flipping through the downwash airflow by combining the physical characteristic parameters of the excitation source and the pre-built final parameter mapping model.

[0008] An adaptive acquisition module is used to control the UAV to perform acquisition operations according to the aerodynamic excitation control parameters, obtain leaf back image data, calculate the effective leaf back exposure rate based on the leaf back image data, determine whether the effective leaf back exposure rate is greater than a preset exposure rate threshold, if not, then correct the aerodynamic excitation control parameters in the direction of increasing the effective leaf back exposure rate to exceed the exposure rate threshold, and regenerate the leaf back image data based on the corrected aerodynamic excitation control parameters;

[0009] The feature extraction module is used to perform image recognition on the leaf back image data to obtain pest and disease feature data; at the same time, it performs time domain analysis on the leaf back image data to obtain vibration frequency data, and calculates the leaf stiffness index based on the vibration frequency data.

[0010] The pest and disease early warning and control module is used to analyze the pests and diseases of the target tree species based on the pest and disease characteristic data and the leaf stiffness index, and generate the severity of pests and diseases of the target tree species.

[0011] Preferably, determining the aerodynamic excitation control parameters required for the UAV to achieve blade flipping via downwash airflow includes the following steps:

[0012] The gravitational torque of the blade and the elastic restoring torque required for the blade to flip to a preset angle are extracted from the canopy blade attribute data.

[0013] The aerodynamic pressure required to cause the canopy blades to flip is calculated based on the gravitational torque and the elastic restoring torque.

[0014] The aerodynamic pressure and the physical characteristic parameters of the excitation source are input into the final parameter mapping model for mapping, thereby generating UAV flight parameters that satisfy the aerodynamic pressure.

[0015] The deflection angle of the image acquisition device carried by the UAV is calculated based on the UAV flight parameters and aerodynamic pressure; the deflection angle represents the deflection angle of the optical axis of the image acquisition device relative to the vertical downward direction;

[0016] The flight parameters of the UAV and the deflection angle are combined to obtain the aerodynamic excitation control parameters.

[0017] Preferably, obtaining the leaf underside image data includes the following steps:

[0018] The system calls a preset inspection flight speed and controls the UAV to perform a leaf back image acquisition task at the preset flight path to obtain leaf back image data. During the leaf back image acquisition task, the UAV is controlled to adjust its flight state and the observation attitude of the image acquisition device according to the aerodynamic excitation control parameters.

[0019] The task of acquiring images of the back of the leaf includes the following steps:

[0020] The drone is controlled to fly at the inspection flight speed along a preset route. The downwash airflow is used to force the canopy leaves of the target tree species ahead of the route to flip. The image acquisition device on the drone continuously captures frames of the area of ​​the canopy of the target tree species that is disturbed by the airflow, and obtains the raw video stream data.

[0021] Optical flow field analysis is performed on the raw video stream data to calculate the optical flow vector between adjacent frames, identify dynamic regions where the magnitude of the optical flow vector exceeds a preset vector magnitude threshold, and mark the dynamic regions as aerodynamic disturbance regions.

[0022] Extract the chromaticity feature data of the current frame and the previous frame in the aerodynamic disturbance area from the original video stream data, and calculate the chromaticity difference between the current frame and the previous frame in the aerodynamic disturbance area based on the chromaticity feature data;

[0023] Determine whether the chromaticity difference is greater than a preset chromaticity deviation threshold and whether the mean of the chromaticity feature data is within a preset leaf back chromaticity range;

[0024] If the chromaticity difference is greater than the preset chromaticity deviation threshold and the mean of the chromaticity feature data is within the preset leaf back chromaticity range, then the blade in the aerodynamic disturbance area is determined to be in an active flipping state.

[0025] When the blades in the aerodynamic disturbance area are in an active flipping state, the proportion of pixels in the original video stream data whose chromaticity values ​​are in the characteristic chromaticity range of the leaf back is continuously calculated, and a leaf back exposure change curve is generated based on the proportion of pixels.

[0026] Select the moment when the percentage of pixels reaches its peak from the curve of exposure on the back of the blade, and take the image frame corresponding to that moment as the best target frame on the back of the blade for the aerodynamic disturbance region.

[0027] Based on the optimal target frame on the back of the leaf and its corresponding aerodynamic disturbance region coordinates, the image region containing the flipped leaf is cropped and extracted from the original video stream data and aggregated to obtain the image data of the back of the leaf.

[0028] By calculating the aerodynamic pressure required for the leaves to flip, and combining the physical characteristic parameters of the excitation source with the final parameter mapping model to determine the aerodynamic excitation control parameters, the UAV is controlled to use the downwash airflow to force the canopy leaves to flip. This achieves effective exposure of the leaf underside area, which is naturally shaded, so that the typical symptoms of pests and diseases on the leaf underside can be fully observed and identified, providing a reliable data foundation for the accurate diagnosis of pests and diseases.

[0029] Preferably, the step of regenerating the blade back image data based on the corrected aerodynamic excitation control parameters includes the following steps:

[0030] Extract the vegetation coverage area containing all leaves from the leaf back image data, and count the total number of pixels in the vegetation coverage area, which is recorded as the total number of pixels in the whole leaf area.

[0031] Traverse all pixels within the vegetation coverage area and count the percentage of pixels whose chromaticity values ​​fall within the preset leaf back feature chromaticity range, which is recorded as the leaf back feature pixel count.

[0032] The effective leaf back exposure rate is calculated based on the number of feature pixels on the leaf back and the number of pixels in the entire leaf area.

[0033] Determine whether the effective blade back exposure rate is greater than a preset exposure rate threshold. If not, correct the aerodynamic excitation control parameters in the direction of increasing the effective blade back exposure rate to exceed the exposure rate threshold, and regenerate the blade back image data based on the corrected aerodynamic excitation control parameters.

[0034] The step of correcting the aerodynamic excitation control parameters in the direction of increasing the effective blade back exposure rate to exceed the exposure rate threshold includes the following steps:

[0035] Set the current iteration number to And the maximum number of iterations is ; Define an optimization space for aerodynamic excitation control parameters, and randomly generate parameters within that space. Each parameter optimization data point corresponds to a set of aerodynamic excitation control parameters, resulting in a parameter optimization dataset.

[0036] An objective function is constructed based on the effective leaf back exposure rate and the exposure rate threshold. The fitness value of each parameter optimization data is calculated based on the objective function, and the parameter optimization data with the highest fitness value is selected as the current optimal solution.

[0037] Update the parameters to optimize the behavior control factors of the dataset;

[0038] Each parameter optimization data selects a position update strategy in the aerodynamic excitation control parameter optimization space based on the behavior control factor, and performs position updates according to the selected position update strategy;

[0039] The fitness value of each parameter after position update is calculated according to the objective function. If the parameter optimization data with the highest fitness value is reselected as the current optimal solution, and if the fitness value of the parameter optimization data after position update is greater than the original fitness value, the new position is used to replace the original position; otherwise, the original position is retained.

[0040] Determine the Is it greater than or equal to the stated If the above Greater than or equal to the If the condition is met, the current optimal solution is output as the corrected aerodynamic excitation control parameters; otherwise, the iteration continues until the condition is met. Greater than or equal to the .

[0041] By calculating the effective leaf back exposure rate and determining whether it reaches the preset threshold, an intelligent optimization algorithm is used to adaptively correct the aerodynamic excitation control parameters. The flight parameters of the UAV and the observation attitude of the image acquisition device are dynamically adjusted according to the canopy structure and leaf characteristics of different tree species. This ensures that leaf back image data that meets the recognition requirements can be obtained in different scenarios, and improves the adaptability of the pest and disease monitoring system to complex forest environments and the stability of data acquisition.

[0042] Preferably, obtaining the pest and disease characteristic data includes the following steps:

[0043] Construct a pest and disease type feature database; the pest and disease type feature database contains standard feature vectors corresponding to different types of pests and diseases;

[0044] The leaf back image data is segmented to identify and segment each non-overlapping leaf region in the image.

[0045] Feature extraction is performed on the image data corresponding to each leaf region in the leaf back image data to generate the feature vector to be measured for each leaf region.

[0046] Select any leaf region as the target leaf region, and calculate the similarity between the feature vector to be tested corresponding to the target leaf region and the standard feature vector in each pest and disease type feature data in the pest and disease type feature database in turn, so as to obtain the similarity matrix of the target leaf region.

[0047] The maximum similarity value is selected from the similarity matrix. It is then determined whether the maximum similarity value is greater than a preset similarity threshold. If the maximum similarity value is greater than the preset similarity threshold, the pest and disease type feature data corresponding to the maximum similarity value is labeled to obtain real-time pest and disease type data; otherwise, no operation is performed.

[0048] After traversing all leaf regions, if the maximum similarity value in all similarity matrices is less than the preset similarity threshold, the output pest and disease feature data is "no pests or diseases have appeared"; otherwise, the output pest and disease feature data is "pests or diseases have appeared". All real-time pest and disease type data are then summarized to obtain a real-time pest and disease type dataset.

[0049] Preferably, calculating the blade stiffness index based on the vibration frequency data includes the following steps:

[0050] Select the leaf region corresponding to the peak of the maximum similarity in all similarity matrices, and extract the time index and spatial coordinate index of the leaf region in the leaf back image data;

[0051] Based on the time index and spatial coordinate index, the corresponding local dynamic frame sequence is extracted from the original video stream data;

[0052] Feature point tracking and spectrum analysis are performed on the local dynamic frame sequence to obtain vibration frequency data characterizing the transient dynamic response of the blade;

[0053] The blade stiffness index, which characterizes the structural integrity and physiological health of the blade, is calculated based on the vibration frequency data.

[0054] Preferably, the generation of the severity of pests and diseases in the target tree species includes the following steps:

[0055] The severity of pests and diseases in the target tree species was analyzed based on the pest and disease characteristic data and the leaf stiffness index.

[0056] If the pest and disease characteristic data indicates the presence of pests and diseases, and the leaf stiffness index is less than the preset health stiffness threshold, then the severity of pests and diseases of the target tree species is output as severely damaged, a first-level early warning signal is generated, and the real-time pest and disease type dataset is pushed to the pest and disease early warning and control platform for display.

[0057] If the pest and disease characteristic data indicates the presence of pests and diseases, and the leaf stiffness index is greater than or equal to the preset health stiffness threshold, then the severity of pests and diseases in the target tree species is output as the initial stage of infection, a secondary warning signal is generated, and the real-time pest and disease type dataset is pushed to the pest and disease early warning and control platform for display.

[0058] If the pest and disease characteristic data indicates that no pests or diseases have appeared, and the leaf stiffness index is less than the preset health stiffness threshold, then the severity of pests and diseases of the target tree species is output as physiological sub-health, and a three-level warning signal is generated.

[0059] If the pest and disease characteristic data indicates that no pests or diseases have appeared, and the leaf stiffness index is greater than or equal to the preset health stiffness threshold, then the severity of pests and diseases of the target tree species will be output as healthy, and no warning will be issued.

[0060] By performing time-domain analysis on leaf underside image data to extract vibration frequency data and calculating the leaf stiffness index, the structural integrity and physiological health of the leaves can be quantified. In the early stages when changes in leaf appearance due to pest and disease infection are not yet obvious, abnormal changes in the leaf stiffness index can detect the state of physiological damage. Combined with the visual characteristics of pests and diseases, graded early warning can be achieved, providing a basis for timely control measures.

[0061] By employing the above technical solution, the present invention provides a dynamic early warning and control system based on forestry pest and disease identification, which has at least the following beneficial effects:

[0062] 1. This invention acquires leaf underside image data by using downwash airflow from a drone to flip the leaves. Combined with an adaptive parameter optimization mechanism and a comprehensive analysis method that integrates visual feature recognition of pests and diseases with dynamic assessment of leaf physiological status, it achieves effective acquisition and accurate identification of pest and disease characteristics in the canopy leaf underside area. Furthermore, by quantitatively assessing physiological health status through leaf stiffness index, it enables early warning response in the early stages of pest and disease infection, thereby improving the accuracy and timeliness of forestry pest and disease monitoring.

[0063] 2. This invention calculates the aerodynamic pressure required for the leaves to flip over, and determines the aerodynamic excitation control parameters by combining the physical characteristic parameters of the excitation source and the final parameter mapping model. It controls the UAV to use the downwash airflow to force the canopy leaves to flip over, thereby achieving effective exposure of the leaf underside area that is naturally covered. This allows the typical symptoms of pests and diseases on the leaf underside to be fully observed and identified, providing a reliable data foundation for the accurate diagnosis of pests and diseases.

[0064] 3. This invention calculates the effective leaf underside exposure rate and determines whether it reaches a preset threshold. It then uses an intelligent optimization algorithm to adaptively correct the aerodynamic excitation control parameters. Based on the canopy structure and leaf characteristics of different tree species, it dynamically adjusts the UAV flight parameters and the observation attitude of the image acquisition device. This ensures that leaf underside image data that meets the recognition requirements can be acquired in different scenarios, thereby improving the adaptability of the pest and disease monitoring system to complex forest environments and the stability of data acquisition.

[0065] 4. This invention extracts vibration frequency data and calculates leaf stiffness index by performing time-domain analysis on leaf back image data, thereby quantifying the structural integrity and physiological health of the leaf. In the early stages when changes in leaf appearance due to pest and disease infection are not yet obvious, abnormal changes in leaf stiffness index can detect the damaged state of physiological function. Combined with the visual characteristics of pests and diseases, a graded early warning can be achieved, providing a decision-making basis for timely prevention and control measures. Attached Figure Description

[0066] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain the application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0067] Figure 1 This is a schematic diagram of the modules of the dynamic early warning and control system based on forestry pest and disease identification provided by the present invention. Detailed Implementation

[0068] 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.

[0069] Exemplary system:

[0070] Existing technologies typically rely on images of the front of leaves or fixed observation angles for analysis, making it difficult to fully reflect the characteristics of pests and diseases on the underside of leaves during monitoring. This results in ineffective identification and early warning systems for forestry pests and diseases. This embodiment proposes a dynamic early warning and control system based on forestry pest and disease identification. Figure 1 As shown, the system includes:

[0071] The data acquisition module is used to acquire canopy and leaf attribute data of tree species in the target forest area and physical characteristic parameters of the excitation source of the inspection drone;

[0072] The control parameter solving module is used to determine the aerodynamic pressure required to cause the canopy blades to flip based on the canopy blade attribute data, and to determine the aerodynamic excitation control parameters required to characterize the UAV to achieve blade flipping through the downwash airflow by combining the physical characteristic parameters of the excitation source and the pre-built final parameter mapping model.

[0073] An adaptive acquisition module is used to control the UAV to perform acquisition operations according to the aerodynamic excitation control parameters, obtain leaf back image data, calculate the effective leaf back exposure rate based on the leaf back image data, determine whether the effective leaf back exposure rate is greater than a preset exposure rate threshold, if not, then correct the aerodynamic excitation control parameters in the direction of increasing the effective leaf back exposure rate to exceed the exposure rate threshold, and regenerate the leaf back image data based on the corrected aerodynamic excitation control parameters;

[0074] The feature extraction module is used to perform image recognition on the leaf back image data to obtain pest and disease feature data; at the same time, it performs time domain analysis on the leaf back image data to obtain vibration frequency data, and calculates the leaf stiffness index based on the vibration frequency data.

[0075] The pest and disease early warning and control module is used to analyze the pests and diseases of the target tree species based on the pest and disease characteristic data and the leaf stiffness index, and generate the severity of pests and diseases of the target tree species.

[0076] The canopy blade attribute data includes at least the canopy blade's gravitational torque, elastic restoring torque, average force-bearing area, average lever arm, and morphological distribution index; the excitation source physical characteristic parameters include at least the rotor geometry parameters, rotor power conversion parameters, and inherent flow field morphological parameters.

[0077] Determining the aerodynamic excitation control parameters required for the UAV to achieve blade flipping via downwash airflow includes the following steps:

[0078] The gravitational torque of the blade and the elastic restoring torque required for the blade to flip to a preset angle are extracted from the canopy blade attribute data.

[0079] The aerodynamic pressure required for the canopy blades to flip over is calculated based on the gravitational torque and elastic restoring torque; the calculation formula is as follows:

[0080] ,

[0081] in, This indicates the aerodynamic pressure required to cause the canopy blades to flip over. and These represent the expected values ​​of the gravitational torque and elastic restoring torque of the canopy leaves, respectively. and These represent the standard deviations of the gravitational torque and the elastic restoring torque of the canopy leaves, respectively. This indicates the coverage confidence factor. , as well as These represent the average force-bearing area, average lever arm, and morphological distribution index of the canopy leaves, respectively.

[0082] The morphological distribution index characterizes the ability of the canopy leaves to effectively intercept downwash airflow in the three-dimensional spatial posture under natural growth conditions.

[0083] Preferably, for broad-leaved tree species with horizontally spread leaves, such as poplar, the leaf inclination angle is small, and the interception section of the downwash airflow is large, so the morphological distribution index is set to a range of 0.7-0.9; for tree species with drooping or needle-like leaves, such as willow or pine, the airflow transmittance is high, so the morphological distribution index is set to a range of 0.3-0.6.

[0084] The coverage confidence factor represents an additional strength redundancy coefficient applied on the basis of theoretically calculated air pressure in order to overcome individual differences in leaves within the canopy and environmental uncertainties.

[0085] Preferably, when in normal inspection mode, the preset flip ratio is set to 60%-80%, and the coverage confidence factor is set to 1.1-1.2 to reduce the energy consumption of the drone; when in precision prevention and control or severe suspected mode, the preset flip ratio is set to 95% or more, and the coverage confidence factor is set to 1.3-1.5 to provide forced aerodynamic redundancy by increasing the downwash air pressure.

[0086] The aerodynamic pressure and the physical characteristic parameters of the excitation source are input into the final parameter mapping model for mapping to generate UAV flight parameters that satisfy the aerodynamic pressure; the UAV flight parameters include at least the excitation altitude and rotor speed;

[0087] The deflection angle of the image acquisition device carried by the UAV is calculated based on the UAV's flight parameters and aerodynamic pressure; the deflection angle represents the deflection angle of the optical axis of the image acquisition device relative to the vertically downward direction; the calculation formula is as follows:

[0088] ,

[0089] in, Represents the jet diffusivity coefficient. This indicates the rotor radius of the inspection drone. Represents the rotor thrust conversion constant. and These represent the excitation altitude and rotor speed in the flight parameters of the UAV, respectively. Indicates the distance from the virtual origin;

[0090] The jet diffusivity coefficient and the distance to the virtual origin are inherent aerodynamic constants obtained by numerical fitting of the axial velocity decay law and radial diffusion characteristics of the flow field based on measured or simulated data of the downwash airflow field of the UAV rotor configuration.

[0091] The flight parameters of the UAV and the deflection angle are combined to obtain the aerodynamic excitation control parameters.

[0092] Furthermore, the final parameter mapping model can adopt the cINN model; the model construction process includes the following steps:

[0093] A historical mission dataset is obtained by acquiring several sets of historical mission data of drones performing flight missions in forest areas through the Internet of Things; the historical mission data includes at least the flight parameters of the drones performing flight missions in forest areas, the physical characteristic parameters of their own excitation sources, and the aerodynamic pressure experienced by the canopy leaves of tree species when they flip over.

[0094] Construct an initial parameter mapping model, set the training data ratio, such as 8:2 or 7:3, which can be reasonably adjusted according to the actual situation. Divide the historical task dataset according to the training data ratio to obtain the training dataset and the test dataset.

[0095] Set a training error threshold, such as 5%-10%, which can be adjusted reasonably according to the actual situation. Input the training data in the training dataset into the initial parameter mapping model for training. Continuously adjust the parameters of the initial parameter mapping model according to the training results until the training error is less than the training error threshold, and obtain the trained parameter mapping model.

[0096] Set the test precision, such as 90%-95%, which can be adjusted reasonably according to the actual situation. Input the test data in the test dataset into the trained parameter mapping model for testing, and calculate the accuracy of the test results. If the accuracy of the test results is greater than the test precision, the final parameter mapping model is obtained; otherwise, retrain until the accuracy of the test results is greater than the test precision.

[0097] The structure of the initial parameter mapping model can be seen in Table 1 below:

[0098]

[0099] Table 1

[0100] Among them, the canopy leaves represent the group of branches and leaves located at the top and outer edge of the tree canopy in the forest stand and within the effective flow field coverage of the downwash airflow of the UAV rotor; this area usually has a large number of new tender leaves due to apical dominance, and is a high-frequency activity area for forest pests to feed and lay eggs.

[0101] The process of obtaining the leaf underside image data includes the following steps:

[0102] The system calls a preset inspection flight speed and controls the UAV to perform a leaf back image acquisition task at the preset flight path to obtain leaf back image data. During the leaf back image acquisition task, the UAV is controlled to adjust its flight state and the observation attitude of the image acquisition device according to the aerodynamic excitation control parameters.

[0103] Furthermore, the task of acquiring images of the leaf's back includes the following steps:

[0104] The drone is controlled to fly at the inspection flight speed along a preset route. The downwash airflow is used to force the canopy leaves of the target tree species ahead of the route to flip. The image acquisition device on the drone continuously captures frames of the area of ​​the canopy of the target tree species that is disturbed by the airflow, and obtains the raw video stream data.

[0105] Optical flow field analysis is performed on the raw video stream data to calculate the optical flow vector between adjacent frames, identify dynamic regions where the magnitude of the optical flow vector exceeds a preset vector magnitude threshold, and mark the dynamic regions as aerodynamic disturbance regions.

[0106] Extract the chromaticity feature data of the current frame and the previous frame in the aerodynamic disturbance area from the original video stream data, and calculate the chromaticity difference between the current frame and the previous frame in the aerodynamic disturbance area based on the chromaticity feature data;

[0107] Determine whether the chromaticity difference is greater than a preset chromaticity deviation threshold and whether the mean of the chromaticity feature data is within a preset leaf back chromaticity range;

[0108] If the chromaticity difference is greater than the preset chromaticity deviation threshold and the mean of the chromaticity feature data is within the preset leaf back chromaticity range, then the blade in the aerodynamic disturbance area is determined to be in an active flipping state.

[0109] When the blades in the aerodynamic disturbance area are in an active flipping state, the proportion of pixels in the original video stream data whose chromaticity values ​​are in the characteristic chromaticity range of the leaf back is continuously calculated, and a leaf back exposure change curve is generated based on the proportion of pixels.

[0110] Select the moment when the percentage of pixels reaches its peak from the curve of exposure on the back of the blade, and take the image frame corresponding to that moment as the best target frame on the back of the blade for the aerodynamic disturbance region.

[0111] Based on the optimal target frame on the back of the leaf and its corresponding aerodynamic disturbance region coordinates, the image region containing the flipped leaf is cropped and extracted from the original video stream data and aggregated to obtain the image data of the back of the leaf.

[0112] Wherein, the chromaticity feature data represents the set of chromaticity component values ​​of all pixels within the aerodynamic disturbance region in the current frame and the previous frame image; the set of chromaticity component values ​​represents a feature matrix containing the chromaticity data of all pixels within the aerodynamic disturbance region;

[0113] The chromaticity difference value represents the statistical mean of the chromaticity change of corresponding pixels in the current frame and the previous frame within the aerodynamic disturbance area.

[0114] The characteristic chromaticity range of the leaf underside is defined based on the inherent spectral reflectance characteristics and chromaticity distribution statistical laws determined by the microstructure of the spongy tissue on the underside of the target tree species' leaves.

[0115] By calculating the aerodynamic pressure required for the leaves to flip, and combining the physical characteristic parameters of the excitation source with the final parameter mapping model to determine the aerodynamic excitation control parameters, the UAV is controlled to use the downwash airflow to force the canopy leaves to flip. This achieves effective exposure of the leaf underside area, which is naturally shaded, so that the typical symptoms of pests and diseases on the leaf underside can be fully observed and identified, providing a reliable data foundation for the accurate diagnosis of pests and diseases.

[0116] The process of regenerating the blade back image data based on the corrected aerodynamic excitation control parameters includes the following steps:

[0117] Extract the vegetation coverage area containing all leaves from the leaf back image data, and count the total number of pixels in the vegetation coverage area, which is recorded as the total number of pixels in the whole leaf area.

[0118] Traverse all pixels within the vegetation coverage area and count the percentage of pixels whose chromaticity values ​​fall within the preset leaf back feature chromaticity range, which is recorded as the leaf back feature pixel count.

[0119] The effective leaf back exposure rate is calculated based on the number of feature pixels on the leaf back and the number of pixels in the entire leaf area.

[0120] Determine whether the effective blade back exposure rate is greater than a preset exposure rate threshold. If not, correct the aerodynamic excitation control parameters in the direction of increasing the effective blade back exposure rate to exceed the exposure rate threshold, and regenerate the blade back image data based on the corrected aerodynamic excitation control parameters.

[0121] Furthermore, the step of correcting the aerodynamic excitation control parameters in the direction of increasing the effective blade back exposure rate to exceed the exposure rate threshold includes the following steps:

[0122] Set the current iteration number to And the maximum number of iterations is ; Define an optimization space for aerodynamic excitation control parameters, and randomly generate parameters within that space. Each parameter optimization data point corresponds to a set of aerodynamic excitation control parameters, resulting in a parameter optimization dataset.

[0123] An objective function is constructed based on the effective leaf underside exposure rate and the exposure rate threshold. The fitness value of each parameter optimization data point in the parameter optimization dataset is calculated based on the objective function, and the parameter optimization data point with the highest fitness value is selected as the current optimal solution. The objective function expression is as follows:

[0124] ,

[0125] in, Indicates the first Each parameter optimizes the fitness value of the data. Indicates according to the first The effective back exposure rate is calculated from the back image data generated by the aerodynamic excitation control parameters corresponding to the optimized parameter data. This indicates the preset exposure rate threshold. Indicates the bias value;

[0126] The parameter optimization dataset's behavior control factor is updated; the update formula is as follows:

[0127] ,

[0128] in, Indicates behavioral control factors. This indicates the maximum sensitivity value. This represents a random number that follows a uniform distribution between (0,1);

[0129] like Then, the optimized data of each parameter in the parameter optimization dataset is updated in the aerodynamic excitation control parameter optimization space using a global search strategy; the position update formula is as follows:

[0130] ,

[0131] in, Indicates the first The position is updated after optimizing the data with each parameter. Indicates the position of the current optimal solution. Indicates the first Each parameter optimizes the current position of the data. This represents a random number that follows a uniform distribution between (0,1). Indicates the first Each parameter optimizes the sensitivity sensing range of the data. = ,in, This represents a random number that follows a uniform distribution between (0,1);

[0132] like Then, the optimized data of each parameter in the parameter optimization dataset is updated in the aerodynamic excitation control parameter optimization space using a local search strategy; the position update formula is as follows:

[0133] ,

[0134] in, This represents a random number that follows a uniform distribution between (0,1). This indicates that the roulette strategy is used for the first time. The parameter optimization data is randomly selected from angles between 0 degrees and 360 degrees;

[0135] The fitness value of each parameter optimization data in the parameter optimization dataset is calculated according to the objective function after position update. If the parameter optimization data with the highest fitness value is reselected as the current optimal solution, and if the fitness value of the parameter optimization data after position update is greater than the original fitness value, the new position is used to replace the original position; otherwise, the original position is retained.

[0136] Determine the Is it greater than or equal to the stated If the above Greater than or equal to the If the optimal parameter is found to have the highest fitness value, then the optimized parameter data is output as the corrected aerodynamic excitation control parameter; otherwise, the iteration continues until the optimal parameter is found. Greater than or equal to the .

[0137] By calculating the effective leaf back exposure rate and determining whether it reaches the preset threshold, an intelligent optimization algorithm is used to adaptively correct the aerodynamic excitation control parameters. The flight parameters of the UAV and the observation attitude of the image acquisition device are dynamically adjusted according to the canopy structure and leaf characteristics of different tree species. This ensures that leaf back image data that meets the recognition requirements can be obtained in different scenarios, and improves the adaptability of the pest and disease monitoring system to complex forest environments and the stability of data acquisition.

[0138] Meanwhile, the intelligent optimization algorithm has good robustness, which can ensure the stability of the algorithm iteration process and effectively avoid the correction result from getting trapped in a local optimum.

[0139] The process of obtaining pest and disease characteristic data includes the following steps:

[0140] Construct a pest and disease type feature database; the pest and disease type feature database contains standard feature vectors corresponding to different types of pests and diseases;

[0141] The leaf back image data is segmented to identify and segment each non-overlapping leaf region in the image.

[0142] Feature extraction is performed on the image data corresponding to each leaf region in the leaf back image data to generate the feature vector to be measured for each leaf region.

[0143] Select any leaf region as the target leaf region, and calculate the similarity between the feature vector to be tested corresponding to the target leaf region and the standard feature vector in each pest and disease type feature data in the pest and disease type feature database in turn, so as to obtain the similarity matrix of the target leaf region.

[0144] The maximum similarity value is selected from the similarity matrix. It is then determined whether the maximum similarity value is greater than a preset similarity threshold. If the maximum similarity value is greater than the preset similarity threshold, the pest and disease type feature data corresponding to the maximum similarity value is labeled to obtain real-time pest and disease type data; otherwise, no operation is performed.

[0145] After traversing all leaf regions, if the maximum similarity value in all similarity matrices is less than the preset similarity threshold, the output pest and disease feature data is "no pests or diseases have appeared"; otherwise, the output pest and disease feature data is "pests or diseases have appeared". All real-time pest and disease type data are then summarized to obtain a real-time pest and disease type dataset.

[0146] Image segmentation refers to separating the foreground leaves from the background and blocking the adhering areas based on the color or texture gradient features of the image. Specifically, it can be achieved by using OTSU threshold segmentation combined with the watershed algorithm or gradient-based edge detection to extract non-overlapping independent leaf connected components for single leaf feature calculation.

[0147] Feature extraction refers to numerical dimensionality reduction encoding of the texture structure, geometric shape, or color distribution of leaf regions. Specifically, it can be implemented using gray-level co-occurrence matrix, local binary mode, color moment, or invariant moment algorithms to generate multidimensional feature vectors with rotation and scale invariance, so as to quantitatively characterize the visual changes on the underside of leaves caused by pests and diseases.

[0148] The calculation of the blade stiffness index based on the vibration frequency data includes the following steps:

[0149] Select the leaf region corresponding to the peak of the maximum similarity in all similarity matrices, and extract the time index and spatial coordinate index of the leaf region in the leaf back image data;

[0150] Based on the time index and spatial coordinate index, the corresponding local dynamic frame sequence is extracted from the original video stream data;

[0151] Feature point tracking and spectrum analysis are performed on the local dynamic frame sequence to obtain vibration frequency data characterizing the transient dynamic response of the blade;

[0152] The blade stiffness index, which characterizes the structural integrity and physiological health of the blade, is calculated based on the vibration frequency data; the calculation formula is as follows:

[0153] ,

[0154] in, Indicates the blade stiffness index. Indicates the equivalent mass of the blade. This represents the principal oscillation frequency with the largest amplitude in the vibration frequency data. This indicates the length of the characteristic lever arm, taken as the petiole length.

[0155] Furthermore, the feature point tracking and spectral analysis of the local dynamic frame sequence includes the following steps:

[0156] Temporal analysis is performed on the local dynamic frame sequence to calculate the centroid of the blade region. The maximum point of Euclidean distance on the blade profile relative to the centroid is extracted. The pixel coordinates in the neighborhood of the maximum point are Gaussian smoothed to obtain the feature tracking point that characterizes the maximum amplitude of the blade's aeroelastic response.

[0157] The displacement vector changes of the feature tracking points in the local dynamic frame sequence are extracted using an optical flow tracing algorithm to generate a time-domain vibration displacement signal;

[0158] The time-domain vibration displacement signal is subjected to a fast Fourier transform to obtain vibration frequency data characterizing the transient dynamic response of the blade.

[0159] The process of determining the severity of pests and diseases affecting the target tree species includes the following steps:

[0160] The severity of pests and diseases in the target tree species was analyzed based on the pest and disease characteristic data and the leaf stiffness index.

[0161] If the pest and disease characteristic data indicates the presence of pests and diseases, and the leaf stiffness index is less than the preset health stiffness threshold, then the severity of pests and diseases of the target tree species is output as severely damaged, a first-level early warning signal is generated, and the real-time pest and disease type dataset is pushed to the pest and disease early warning and control platform for display.

[0162] If the pest and disease characteristic data indicates the presence of pests and diseases, and the leaf stiffness index is greater than or equal to the preset health stiffness threshold, then the severity of pests and diseases in the target tree species is output as the initial stage of infection, a secondary warning signal is generated, and the real-time pest and disease type dataset is pushed to the pest and disease early warning and control platform for display.

[0163] If the pest and disease characteristic data indicates that no pests or diseases have appeared, and the leaf stiffness index is less than the preset health stiffness threshold, then the severity of pests and diseases of the target tree species is output as physiological sub-health, and a three-level warning signal is generated.

[0164] If the pest and disease characteristic data indicates that no pests or diseases have appeared, and the leaf stiffness index is greater than or equal to the preset health stiffness threshold, then the severity of pests and diseases of the target tree species will be output as healthy, and no warning will be issued.

[0165] By performing time-domain analysis on leaf underside image data to extract vibration frequency data and calculating the leaf stiffness index, the structural integrity and physiological health of the leaves can be quantified. In the early stages when changes in leaf appearance due to pest and disease infection are not yet obvious, abnormal changes in the leaf stiffness index can detect the state of physiological damage. Combined with the visual characteristics of pests and diseases, graded early warning can be achieved, providing a basis for timely control measures.

[0166] Exemplary computer-readable media:

[0167] Embodiments of this application may also be computer-readable storage media storing computer program instructions thereon, which, when executed by a processor, cause the processor to perform the steps described in the "Exemplary Methods" section above according to the various embodiments of this application.

[0168] Computer-readable storage media may take the form of any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may, for example, include, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0169] The basic principles of this application have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this application are merely examples and not limitations, and should not be considered as essential features of each embodiment of this application. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the application to the necessity of employing the aforementioned specific details for implementation.

[0170] The block diagrams of devices, apparatuses, devices, and systems involved in this application are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0171] It should also be noted that in the apparatus, equipment, and methods of this application, the components or steps can be disassembled and / or recombined. These disassemblies and / or recombinations should be considered as equivalent solutions of this application.

[0172] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of this application. Therefore, this application is not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0173] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this application to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A dynamic early warning and control system based on forestry pest and disease identification, characterized in that, include: The data acquisition module is used to acquire canopy and leaf attribute data of tree species in the target forest area and physical characteristic parameters of the excitation source of the inspection drone; The control parameter solving module is used to determine the aerodynamic pressure required to cause the canopy blades to flip based on the canopy blade attribute data, and to determine the aerodynamic excitation control parameters required to characterize the UAV to achieve blade flipping through the downwash airflow by combining the physical characteristic parameters of the excitation source and the pre-built final parameter mapping model. An adaptive acquisition module is used to control the UAV to perform acquisition operations according to the aerodynamic excitation control parameters, obtain leaf back image data, calculate the effective leaf back exposure rate based on the leaf back image data, determine whether the effective leaf back exposure rate is greater than a preset exposure rate threshold, if not, then correct the aerodynamic excitation control parameters in the direction of increasing the effective leaf back exposure rate to exceed the exposure rate threshold, and regenerate the leaf back image data based on the corrected aerodynamic excitation control parameters; The feature extraction module is used to perform image recognition on the leaf back image data to obtain pest and disease feature data; at the same time, it performs time domain analysis on the leaf back image data to obtain vibration frequency data, and calculates the leaf stiffness index based on the vibration frequency data. The pest and disease early warning and control module is used to analyze the pests and diseases of the target tree species based on the pest and disease characteristic data and the leaf stiffness index, and generate the severity of pests and diseases of the target tree species.

2. The dynamic early warning and control system based on forestry pest and disease identification according to claim 1, characterized in that, The canopy blade attribute data includes at least the extracted gravitational torque of the blade and the elastic restoring torque required for the blade to flip to a preset angle. Determining the aerodynamic excitation control parameters required for the UAV to achieve blade flipping through the downwash airflow includes the following steps: The aerodynamic pressure required to cause the canopy blades to flip is calculated based on the gravitational torque and the elastic restoring torque. The aerodynamic pressure and the physical characteristic parameters of the excitation source are input into the final parameter mapping model for mapping, thereby generating UAV flight parameters that satisfy the aerodynamic pressure. The deflection angle of the image acquisition device carried by the UAV is calculated based on the UAV flight parameters and aerodynamic pressure; the deflection angle represents the deflection angle of the optical axis of the image acquisition device relative to the vertical downward direction; The flight parameters of the UAV and the deflection angle are combined to obtain the aerodynamic excitation control parameters.

3. The dynamic early warning and control system based on forestry pest and disease identification according to claim 1, characterized in that, The process of obtaining the leaf underside image data includes the following steps: The system calls a preset inspection flight speed and controls the UAV to perform a leaf back image acquisition task at the preset flight path to obtain leaf back image data. During the leaf back image acquisition task, the UAV is controlled to adjust its flight state and the observation attitude of the image acquisition device according to the aerodynamic excitation control parameters. The task of acquiring images of the back of the leaf includes the following steps: The drone is controlled to fly at the inspection flight speed along a preset route. The downwash airflow is used to force the canopy leaves of the target tree species ahead of the route to flip. The image acquisition device on the drone continuously captures frames of the area of ​​the canopy of the target tree species that is disturbed by the airflow, and obtains the raw video stream data. Optical flow field analysis is performed on the raw video stream data to calculate the optical flow vector between adjacent frames, identify dynamic regions where the magnitude of the optical flow vector exceeds a preset vector magnitude threshold, and mark the dynamic regions as aerodynamic disturbance regions. Extract the chromaticity feature data of the current frame and the previous frame in the aerodynamic disturbance area from the original video stream data, and calculate the chromaticity difference between the current frame and the previous frame in the aerodynamic disturbance area based on the chromaticity feature data; Determine whether the chromaticity difference is greater than a preset chromaticity deviation threshold and whether the mean of the chromaticity feature data is within a preset leaf back chromaticity range; If the chromaticity difference is greater than the preset chromaticity deviation threshold and the mean of the chromaticity feature data is within the preset leaf back chromaticity range, then the blade in the aerodynamic disturbance area is determined to be in an active flipping state. When the blades in the aerodynamic disturbance area are in an active flipping state, the proportion of pixels in the original video stream data whose chromaticity values ​​are in the characteristic chromaticity range of the leaf back is continuously calculated, and a leaf back exposure change curve is generated based on the proportion of pixels. Select the moment when the percentage of pixels reaches its peak from the curve of exposure on the back of the blade, and take the image frame corresponding to that moment as the best target frame on the back of the blade for the aerodynamic disturbance region. Based on the optimal target frame on the back of the leaf and its corresponding aerodynamic disturbance region coordinates, the image region containing the flipped leaf is cropped and extracted from the original video stream data and aggregated to obtain the image data of the back of the leaf.

4. The dynamic early warning and control system based on forestry pest and disease identification according to claim 1, characterized in that, The process of regenerating the blade back image data based on the corrected aerodynamic excitation control parameters includes the following steps: Extract the vegetation coverage area containing all leaves from the leaf back image data, and count the total number of pixels in the vegetation coverage area, which is recorded as the total number of pixels in the whole leaf area. Traverse all pixels within the vegetation coverage area and count the percentage of pixels whose chromaticity values ​​fall within the preset leaf back feature chromaticity range, which is recorded as the leaf back feature pixel count. The effective leaf back exposure rate is calculated based on the number of feature pixels on the leaf back and the number of pixels in the entire leaf area. Determine whether the effective blade back exposure rate is greater than a preset exposure rate threshold. If not, correct the aerodynamic excitation control parameters in the direction of increasing the effective blade back exposure rate to exceed the exposure rate threshold, and regenerate the blade back image data based on the corrected aerodynamic excitation control parameters.

5. The dynamic early warning and control system based on forestry pest and disease identification according to claim 4, characterized in that, The step of correcting the aerodynamic excitation control parameters in the direction of increasing the effective blade back exposure rate to exceed the exposure rate threshold includes the following steps: Set the current iteration number to And the maximum number of iterations is ; Define an optimization space for aerodynamic excitation control parameters, and randomly generate parameters within that space. Each parameter optimization data point corresponds to a set of aerodynamic excitation control parameters, resulting in a parameter optimization dataset. An objective function is constructed based on the effective leaf back exposure rate and the exposure rate threshold. The fitness value of each parameter optimization data is calculated based on the objective function, and the parameter optimization data with the highest fitness value is selected as the current optimal solution. Update the parameters to optimize the behavior control factors of the dataset; Each parameter optimization data selects a position update strategy in the aerodynamic excitation control parameter optimization space based on the behavior control factor, and performs position updates according to the selected position update strategy; The fitness value of each parameter after position update is calculated according to the objective function. If the parameter optimization data with the highest fitness value is reselected as the current optimal solution, and if the fitness value of the parameter optimization data after position update is greater than the original fitness value, the new position is used to replace the original position; otherwise, the original position is retained. Determine the Is it greater than or equal to the stated If the above Greater than or equal to the If the condition is met, the current optimal solution is output as the corrected aerodynamic excitation control parameters; otherwise, the iteration continues until the condition is met. Greater than or equal to the .

6. The dynamic early warning and control system based on forestry pest and disease identification according to claim 1, characterized in that, The process of obtaining pest and disease characteristic data includes the following steps: Construct a pest and disease type feature database; the pest and disease type feature database contains standard feature vectors corresponding to different types of pests and diseases; The leaf back image data is segmented to identify and segment each non-overlapping leaf region in the image. Feature extraction is performed on the image data corresponding to each leaf region in the leaf back image data to generate the feature vector to be measured for each leaf region. Select any leaf region as the target leaf region, and calculate the similarity between the feature vector to be tested corresponding to the target leaf region and the standard feature vector in each pest and disease type feature data in the pest and disease type feature database in turn, so as to obtain the similarity matrix of the target leaf region. The maximum similarity value is selected from the similarity matrix. It is then determined whether the maximum similarity value is greater than a preset similarity threshold. If the maximum similarity value is greater than the preset similarity threshold, the pest and disease type feature data corresponding to the maximum similarity value is labeled to obtain real-time pest and disease type data; otherwise, no operation is performed. After traversing all leaf regions, if the maximum similarity value in all similarity matrices is less than the preset similarity threshold, the output pest and disease feature data is "no pests or diseases have appeared"; otherwise, the output pest and disease feature data is "pests or diseases have appeared". All real-time pest and disease type data are then summarized to obtain a real-time pest and disease type dataset.

7. The dynamic early warning and control system based on forestry pest and disease identification according to claim 6, characterized in that, The calculation of the blade stiffness index based on the vibration frequency data includes the following steps: Select the leaf region corresponding to the peak of the maximum similarity in all similarity matrices, and extract the time index and spatial coordinate index of the leaf region in the leaf back image data; Based on the time index and spatial coordinate index, the corresponding local dynamic frame sequence is extracted from the original video stream data; Feature point tracking and spectrum analysis are performed on the local dynamic frame sequence to obtain vibration frequency data characterizing the transient dynamic response of the blade; The blade stiffness index, which characterizes the structural integrity and physiological health of the blade, is calculated based on the vibration frequency data.

8. The dynamic early warning and control system based on forestry pest and disease identification according to claim 1, characterized in that, The process of determining the severity of pests and diseases affecting the target tree species includes the following steps: The severity of pests and diseases in the target tree species was analyzed based on the pest and disease characteristic data and the leaf stiffness index. If the pest and disease characteristic data indicates the presence of pests and diseases, and the leaf stiffness index is less than the preset health stiffness threshold, then the severity of pests and diseases of the target tree species is output as severely damaged, a first-level early warning signal is generated, and the real-time pest and disease type dataset is pushed to the pest and disease early warning and control platform for display. If the pest and disease characteristic data indicates the presence of pests and diseases, and the leaf stiffness index is greater than or equal to the preset health stiffness threshold, then the severity of pests and diseases in the target tree species is output as the initial stage of infection, a secondary warning signal is generated, and the real-time pest and disease type dataset is pushed to the pest and disease early warning and control platform for display. If the pest and disease characteristic data indicates that no pests or diseases have appeared, and the leaf stiffness index is less than the preset health stiffness threshold, then the severity of pests and diseases of the target tree species is output as physiological sub-health, and a three-level warning signal is generated. If the pest and disease characteristic data indicates that no pests or diseases have appeared, and the leaf stiffness index is greater than or equal to the preset health stiffness threshold, then the severity of pests and diseases of the target tree species will be output as healthy, and no warning will be issued.

9. An electronic device comprising a memory and a processor, characterized in that: The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions, which, when executed by the processor, implement the system as described in any one of claims 1-8.

10. A computer storage medium storing computer-executable instructions thereon, characterized in that: When the computer-executable instructions are executed by a processor, they implement the system as described in any one of claims 1-8.

Images