Comprehensive ecological prevention and control method for diaphorina citri

Abstract

The invention relates to the technical field of citrus insect pest prevention and control, and discloses a comprehensive ecological prevention and control method for diaphorina citri. The method comprises the following steps: acquiring current insect pest state information based on real-time monitoring data of a citrus planting area; extracting diaphorina citri biological characteristics and associated ecological environment characteristics through the ecological characteristic library to form an initial ecological characteristic set; fusing the real-time monitoring data with the initial ecological feature set to generate a dynamic ecological feature set; calculating the priority score of each prevention and control strategy according to the spatial and temporal distribution density, migration trend intensity, natural enemy inhibition coefficient, host plant resistance index and environmental stress factor of each feature in the dynamic set; generating a strategy sorting table in a descending order according to the scores, sequentially selecting to-be-executed strategies, and judging whether the to-be-executed strategies complete the target population density reduction amount or not in a preset period. According to the method, real-time data and ecological characteristics are combined, the strategy is dynamically adjusted, and the prevention and control accuracy and effectiveness are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of citrus pest control, and in particular to a comprehensive ecological control method for citrus psyllids. Background Art

[0002] Citrus, an important cash crop, is widely cultivated worldwide, and its development significantly impacts regional economies and agricultural structures. However, the citrus psyllid, a major pest in the citrus industry, not only directly sucks the sap from young citrus shoots, causing plant weakness and deformed new shoots, but more importantly, it is the primary vector for citrus Huanglongbing (Huanglongbing), a devastating disease with no effective cure. Once it strikes, it inflicts significant losses on the citrus industry.

[0003] Currently, the primary control method for citrus psyllids is chemical control, which relies on frequent use of insecticides to suppress population density. However, long-term reliance on chemical control can lead to a series of problems. Citrus psyllids are prone to developing resistance to insecticides, leading to increasing pesticide usage and control costs, while also exacerbating ecological pollution. Furthermore, while chemical pesticides kill citrus psyllids, they can also accidentally harm their natural enemies, disrupting the natural balance of the ecosystem and making the citrus psyllid population more likely to rebound, creating a vicious cycle.

[0004] Although ecological control methods such as biological control and agricultural control are environmentally friendly, their current application often has limitations. In biological control, the introduction and use of natural enemies are greatly affected by environmental conditions, climatic factors, etc., making it difficult to achieve stable and effective control; agricultural control measures such as reasonable pruning and cleaning of orchards, although they can reduce the source of insects to a certain extent, lack precise integration with the dynamics of insect pest occurrence, and the control effect is not ideal. In addition, most existing control methods rely on empirical operations and lack effective integration and analysis of real-time monitoring data in citrus planting areas. It is impossible to adjust the control strategy in time according to the dynamic changes of insect pests, resulting in insufficient pertinence and timeliness of control. Summary of the Invention

[0005] The object of the present invention is to provide a comprehensive ecological control method for citrus psyllids to solve the problems raised in the above background technology.

[0006] To achieve the above objectives, the present invention provides a comprehensive ecological control method for citrus psyllids, comprising:

[0007] Obtain current pest status information based on real-time monitoring data from citrus growing areas;

[0008] The biological characteristics and related ecological environment characteristics of citrus psyllids were extracted through the ecological characteristic database to form an initial ecological characteristic set;

[0009] Perform feature fusion processing on real-time monitoring data and initial ecological feature set to generate dynamic ecological feature set;

[0010] Calculate the priority score of each ecological control strategy based on the spatiotemporal distribution density, migration trend intensity, natural enemy inhibition coefficient, host plant resistance index and environmental stress factor of each feature in the dynamic ecological feature set;

[0011] A prevention and control strategy ranking table is generated by arranging the strategies in descending order according to their priority scores. The strategies to be executed are selected in turn to determine whether they can achieve the target insect population density reduction within the preset period.

[0012] Preferably, the extraction of biological characteristics and associated ecological environment characteristics of citrus psyllids through an ecological characteristic library includes:

[0013] The individual behavior characteristics, population dispersal characteristics, natural enemy restriction characteristics, plant volatile characteristics and microclimate response characteristics of citrus psyllids were extracted hierarchically from the ecological characteristic database.

[0014] The individual behavior characteristics and population diffusion characteristics of the same level are averaged to obtain the primary fusion characteristics;

[0015] Perform weighted superposition of primary fusion features and microclimate response features to generate secondary fusion features;

[0016] Hierarchical feature fusion is iteratively performed until the highest-level feature processing is completed, and a dynamic ecological feature set containing multi-level correlation relationships is output.

[0017] Preferably, the calculation of the priority score of each ecological prevention and control strategy includes:

[0018] Obtain the proportion of spatiotemporal distribution density, migration trend intensity, natural enemy inhibition coefficient, host plant resistance index and environmental stress factor corresponding to each strategy;

[0019] Multiply the migration trend intensity ratio by the first weight coefficient, the natural enemy inhibition coefficient ratio by the second weight coefficient, the host plant resistance index ratio by the third weight coefficient, the environmental stress factor ratio by the fourth weight coefficient, and the spatiotemporal distribution density ratio by the fifth weight coefficient;

[0020] The priority score is calculated by summing the five product results and deducting the resource consumption cost ratio.

[0021] Preferably, the step of sequentially selecting the strategies to be executed and determining whether the strategies can achieve the target insect population density reduction within a preset period includes:

[0022] Calculate the insect population density reduction that can be achieved by a single strategy within a preset period;

[0023] Comparing the achievable insect population reduction with the target insect population reduction;

[0024] If the achievable amount is greater than or equal to the target amount, it is determined that the single strategy can be completed;

[0025] If the achievable amount is less than the target amount, the number of strategies that need to be supplemented is calculated.

[0026] Preferably, the calculation of the number of policies to be supplementally executed includes:

[0027] Divide the target insect population density reduction by the achievable amount of a single strategy, and round up to get the number of strategies implemented;

[0028] Detect the total amount of available ecological resources based on the number of strategies implemented;

[0029] When sufficient resources are available, start the corresponding number of policy instances;

[0030] When available resources are insufficient, instances ranked lower than the current policy are selected from the activated policies and terminated to release resources until resource requirements are met or there are no more instances to terminate.

[0031] Preferably, the method further comprises:

[0032] Analyze pest spread paths based on geographic information system topology structure;

[0033] Determine the insect source sinks and migration corridors through gradient field calculations;

[0034] Extract key diffusion features that meet the insect population density threshold and duration of continuous infection;

[0035] Key diffusion characteristics were input into pre-trained ecological models to generate enhanced control fields.

[0036] Preferably, inputting key diffusion characteristics into a pre-trained ecological model to generate an enhanced prevention and control field includes:

[0037] Introducing topological consistency constraints into the loss function of the ecological model;

[0038] Through adversarial training, the matching degree between the control field and the real ecological gradient field is optimized and strengthened;

[0039] The enhanced control field is spatially coupled with basic environmental factors to generate a composite control field containing biotic and abiotic factors.

[0040] Preferably, the method further comprises:

[0041] Simulate the spatiotemporal evolution of the composite control field;

[0042] Separate the biological field of action and the environmental field of action after evolution;

[0043] Reconstruct features of biological interaction fields and real-time monitoring data, and output optimized prevention and control instruction sets.

[0044] Preferably, the method further comprises:

[0045] Continuously monitor the implementation status of executed strategies;

[0046] When new monitoring data triggers an insect situation change event, the dormant strategy instance is awakened;

[0047] If the policy instance does not match a valid instruction after waking up, it will enter the dormant waiting state;

[0048] If a valid instruction is matched, the implementation status is changed to executing, and the execution result data is fed back to the dynamic ecological feature set.

[0049] Preferably, after outputting the optimized control instruction set, the method further includes:

[0050] Verify the actual insect population density change curve after the instruction set is executed;

[0051] When the deviation between the actual reduction amount and the predicted value exceeds a threshold, the key diffusion features are re-extracted;

[0052] Iteratively update the parameters of the enhanced prevention and control field until the deviation value falls into the acceptable range.

[0053] Compared with the prior art, the present invention has the following beneficial effects:

[0054] By obtaining current pest status information based on real-time monitoring data from citrus growing areas, control measures can be closely aligned with actual pest occurrences, avoiding the inappropriate measures that can be implemented in traditional control measures due to information lags. The use of real-time monitoring data allows for a more precise understanding of citrus psyllid occurrence dynamics and population changes, providing a reliable and realistic basis for the development of subsequent control strategies.

[0055] By extracting the biological characteristics of the citrus psyllid and its associated ecological and environmental characteristics from the ecological feature library, we formed an initial ecological feature set. This combined the biological characteristics of the citrus psyllid with the ecological and environmental factors in which it exists, comprehensively considering all factors influencing the occurrence of the pest. This multi-dimensional feature extraction breaks through the limitations of traditional prevention and control methods that focus on a single factor, enabling a more systematic analysis of the root causes and patterns of pest occurrence, laying a comprehensive foundation for subsequent feature integration and strategy development.

[0056] Real-time monitoring data is fused with the initial ecological feature set to generate a dynamic ecological feature set, effectively integrating real-time information with basic feature information. This dynamic ecological feature set evolves with updated monitoring data, reflecting the dynamic characteristics of pests under varying temporal and spatial conditions. This enables more timely and comprehensive understanding of pests and enables dynamic adjustments to prevention and control strategies.

[0057] Based on the spatial and temporal distribution density, migration trend intensity, natural enemy suppression coefficient, host plant resistance index, and environmental stress factors of each dynamic ecological feature set, a priority score is calculated for each ecological control strategy, making the selection of control strategies more scientific and reasonable. By comprehensively considering these multiple characteristic factors, the applicability and potential effectiveness of different control strategies in the current situation can be accurately assessed, avoiding the blind selection of strategies based on experience in traditional control measures and improving the targetedness of control measures.

[0058] A control strategy ranking table is generated by sorting the strategies in descending order of priority. The currently executed strategies are then selected and evaluated to determine whether they can achieve the target population density reduction within the preset period, ensuring the orderly implementation and effectiveness evaluation of control measures. This orderly execution and effectiveness evaluation mechanism allows for the timely identification of potential problems during the control process. If a strategy fails to achieve the expected results, subsequent strategies can be adjusted promptly, ensuring the flexibility and effectiveness of the control process and contributing to a continuously optimized control cycle, thereby achieving better ecological control of citrus psyllids. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is a time sequence diagram of the comprehensive ecological control method for citrus psyllids according to the present invention;

[0060] Figure 2 Flowchart for ecological feature extraction and fusion;

[0061] Figure 3 A flowchart for policy execution judgment;

[0062] Figure 4 Flowcharts generated for pest spread analysis and strengthening of control sites;

[0063] Figure 5 Flowchart of dynamic monitoring and awakening of strategies. DETAILED DESCRIPTION

[0064] The technical solutions in the embodiments of the present invention will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present invention.

[0065] See also Figure 1 The present invention provides a comprehensive ecological control method for citrus psyllids, comprising:

[0066] This approach is achieved through digital integration of real-time monitoring data, ecological feature analysis, and strategy optimization. The method first obtains current pest status information based on real-time monitoring data from citrus growing areas. This monitoring data is derived from a sensor network installed in the growing area, including infrared cameras, soil moisture sensors, and anemometers. The data is sampled every 30 minutes and covers parameters such as insect population density, temperature, and humidity. The system receives sensor data through a pre-set data interface, filters out noise and processes outliers, and outputs structured pest status information. The system then accesses a pre-built ecological feature library to extract biological characteristics of citrus psyllids and associated ecological features, forming an initial ecological feature set. This ecological feature library, stored in a distributed database, contains historical datasets and predefined ecological model parameters. The extracted features include insect behavior data, plant volatile concentrations, and microclimate factors. The system then fuses the real-time monitoring data with the initial ecological feature set. Using a multi-source data fusion algorithm, the data is aligned and normalized based on spatiotemporal coordinates to generate a dynamic ecological feature set. This set contains updated feature values, such as revised population distribution locations. Next, the system quantifies the spatiotemporal distribution density, migration trend intensity, natural enemy suppression coefficient, host plant resistance index, and environmental stress factors of each feature in the dynamic ecological feature set. These factors are digitized and input into the priority calculation module. This module calculates a priority score for each ecological control strategy based on pre-set scoring rules. The priority score is used to rank the strategies, generating a control strategy ranking table and sorting the strategies in descending order. After ranking, the system selects the current strategy to be executed and uses a predictive model to determine whether it can achieve the target insect population reduction within a preset period. The preset period is set to 7 days, based on a target reduction threshold of 50%. If the strategy is achievable, the system executes it immediately; otherwise, it transitions to the supplementary strategy processing process. The entire process is managed by a central control unit and implemented through cloud computing resource scheduling. This automated loop mechanism ensures dynamic response to changes in insect pests.

[0067] Example 1: See Figure 2During the ecological feature extraction phase, a distributed database stores multi-level structured data. This database utilizes a B+ tree indexing mechanism. The base layer stores individual behavioral characteristic datasets, including spatial coordinate sequences and timestamps of feeding trajectories, time-series statistics of mating frequency, and histograms of resting time distribution. The middle layer contains a dataset of population dispersal characteristics, specifically composed of a probability distribution model of migration distance, a coordinate matrix of distribution range contour maps, and a matrix representation of population age structure. The higher-level layer integrates natural enemy constraint feature vectors (correlations between the number of natural enemy species and predation rates), plant volatile concentration gradient fields (centered on chromatographic analysis of volatile organic compounds), and microclimate response feature matrices (tensor representations of temperature and humidity correlation coefficients).

[0068] After the feature fusion operation is initiated, the system loads data in hierarchical order. At the basic level processing unit, the algorithm automatically matches the spatiotemporal labels of individual behavioral characteristics and population diffusion characteristics. For each paired coordinate point, the system performs data alignment: the feeding trajectory coordinate points and migration distance data are matched through spatial interpolation, and the mating frequency time series values ​​and distribution range data are aligned through time window sliding average. After alignment, the averaging calculation engine is implemented: the arithmetic mean is calculated for the matching points. For example, the feeding intensity value (0.7) and the migration intensity value (0.5) of a certain spatial point generate a primary fusion feature value (0.6). This calculation covers all paired points to form the initial feature plane.

[0069] Microclimate response characteristics are input through independent channels. The system collects environmental sensor data streams in real time and updates the temperature and humidity response coefficient table every five minutes. This coefficient table establishes a spatial mapping relationship with the primary fusion feature plane. When the weighted overlay module is activated, the environmental monitoring unit dynamically generates weight factors: when the temperature data rises for three consecutive hours, the microclimate weight is automatically increased to 0.65; when the relative humidity falls below the critical value of 40%, the humidity-related weight is set to 0.35. The weighting process performs point-by-point matrix multiplication: the primary fusion feature value of each spatial point is multiplied by the corresponding microclimate weight and then superimposed with the original microclimate response value. For example, for a point with coordinates (X, Y), the calculation formula is (0.6×0.65)+0.8 (microclimate value)=1.19, generating a secondary fusion feature data layer.

[0070] The hierarchical iterative processing adopts a loop control mechanism. During the intermediate layer processing, the secondary fusion feature data is used as the input source for channel fusion with the high-level layer features. The fusion console creates a double buffer area: the front buffer stores the secondary fusion data, and the back buffer loads the natural enemy restriction feature vector. The system establishes an index mapping through a feature matching algorithm, specifically resampling the predation rate curve to the same spatial resolution. Weighted superposition reuses the same algorithm kernel, but the weight parameters are switched to the hierarchical weight coefficient table. When the processing is completed, the new feature plane is cyclically input to the next level. In the final level processing, the high-level plant volatile features are generated into a three-dimensional concentration field through a gas chromatography simulator and fused with the input features in three-dimensional space. The entire iterative process performs integrity checks: a tensor dimension check is performed on each layer output, and a data reconstruction process is triggered when the dimension does not match.

[0071] Once feature fusion is complete and a dynamic ecological feature set is formed, the priority scoring module is immediately activated. Scoring parameters are obtained through five independent collectors: a spatiotemporal density collector scans the monitoring area grid and divides the current insect population density by the five-year historical peak; a migration trend analyzer processes the motion vector field and calculates the consistency strength of movement direction; a natural enemy biodiversity counter counts the number of natural enemy species per unit area; a plant resistance detector measures leaf antibody protein concentration; and an environmental stress monitoring station integrates soil salinity and heavy metal content data. Each collector outputs a normalized percentage value (ranging from 0 to 1).

[0072] The weight allocation unit has a preset fixed coefficient set: migration trend 0.20, natural enemy suppression 0.25, plant resistance 0.15, environmental stress 0.30, and spatiotemporal density 0.10. The computation core performs five parallel multiplications: the proportion of migration trend intensity; the proportion of natural enemy suppression coefficient; the proportion of plant resistance index; the proportion of environmental stress factors; and the proportion of spatiotemporal density.

[0073] The product is input into the adder accumulator: 0.17 + 0.175 + 0.09 + 0.135 + 0.09 = 0.66. The resource cost accounting module operates simultaneously: it reads the policy resource registration table, extracts labor hours, drug consumption, and equipment occupancy time, and converts them into total cost percentages using the resource cost conversion table. Finally, a subtraction operation is performed to generate a priority score. This score is automatically associated with the policy number and stored in the priority queue for sorting and recall.

[0074] All computing processes are completed on dedicated hardware accelerators: tensor processors are used in the feature extraction stage to fuse three-dimensional feature fields; priority calculation is implemented using a five-channel floating-point operation unit for parallel computing. The error handling mechanism includes three layers of verification: feature value range verification (resampling if it exceeds the 0-1 range), weight and verification (the total weight coefficient is always equal to 1.0), and calculation result verification (trigger manual review if the score value is out of limit). The entire process execution time is allocated by the central controller to ensure that all processing tasks are completed under the time constraint.

[0075] Example 2: see Figure 3 In the policy execution evaluation stage, the system calls preset parameters and prediction models to judge the effectiveness of the policy. The preset period is defined by a fixed time window, and the standard setting is a seven-day period; the target reduction amount is initialized by the configuration module and stored in the system core database, and the typical setting is to reduce the original insect population density by fifty percent. The policy effect prediction model accesses the data stream of the dynamic ecological feature set, continuously receives the population size change rate, the natural enemy activity frequency matrix, and the host plant response curve as dynamic input parameters. The model architecture is built on a time series prediction algorithm, which contains three layers of recurrent neural network processing time dependence, and performs full-amount prediction operation every sixty minutes.

[0076] When starting the single-policy effect prediction process, the simulation engine loads the feature parameter cluster of the target policy. For example, when evaluating the "release predatory natural enemies" policy, the natural enemy species number, unit area release amount, and release point geographic coordinate set need to be input. The model initializes the baseline environment scenario: input the insect population density distribution field, temperature gradient map, and wind speed vector field at the current time into the simulation space. During the operation process, the time stepping mechanism is executed, and the state calculation is advanced every hour: first, update the biological interaction sub-model, calculate the natural enemy predation amount based on the Lotka-Volterra equation; simultaneously update the environmental factor influence sub-module, and apply the temperature and humidity coupling algorithm to correct the insect activity rate. At the end of each time step, record the insect population density change value, and after continuously advancing one hundred and sixty-eight steps (corresponding to a seven-day period), terminate the simulation. The final output is a quantitative result value, typically presented as the insect population density reduction percentage at the end of the prediction period, such as reducing from five hundred units to three hundred units, which indicates that the achievable reduction amount is forty percent.

[0077] The system then activates the reduction comparator. The comparator's core algorithm performs numerical comparison: extracting the achievable reduction output from the simulation and performing a scalar comparison with the target reduction stored in the database. The comparison logic uses a precise threshold judgment criterion: a Boolean true value is generated when the achievable value is greater than or equal to the target value; otherwise, a Boolean false value is generated. The judgment result triggers the corresponding action instruction. If a true value is returned, an activation instruction is sent to the policy executor, which contains the policy execution code, a resource allocation list, and a set of spatial coordinates. The executor calls the hardware interface to control the drone swarm to deploy natural enemies or activate the spray equipment array according to the coordinate sequence. The entire implementation process is recorded in an operation log, including operation timestamps and mechanical status codes.

[0078] If the test returns a false value, the system switches to policy replenishment mode. This mode incorporates dual calculation and resource coordination mechanisms. First, the policy quantity requirement is calculated: the target reduction value is divided by the single policy achievable value to obtain the theoretical multiple. The math processor applies a ceiling function to this multiple. The standard library function ceil(1.25) outputs the integer value 2. This result is stored as the policy implementation quantity parameter.

[0079] The resource scheduling module then begins coordinated operation. The resource database maintains a multidimensional resource inventory in real time: the biological resource pool records the inventory of available natural enemy species; the chemical resource library stores the remaining amount of pesticides and their expiration dates; the equipment resource table records the number of drones available and their battery status; and the human resource pool counts the number of technicians on call. The resource demand estimator calculates total consumption based on the quantity parameters for strategy implementation: if a single strategy requires 50 natural enemy units, the total required is 100 units; if a single strategy consumes 20 liters of pesticide, the total required is 40 liters; and equipment utilization time is calculated as 12 hours, based on a single strategy's six-hour utilization time. The resource sufficiency analysis engine performs parallel checks: comparing whether the available biological resource quantity exceeds the total demand, whether the available pesticide quantity exceeds the calculated total demand, and whether the idle equipment hours meet the total utilization demand. A resource confirmation signal is generated when all three tests return a passing result.

[0080] When sufficient resources are available, the execution controller invokes the policy cloning function: It creates multiple parallel execution instances based on the policy execution quantity parameter. Instance launch utilizes a distributed architecture, with each instance assigned a unique process ID and resource slot. For example, if the execution quantity is two, two instance identifiers, S001A and S001B, are generated, binding execution resources in the north and south regions, respectively. Status monitoring is maintained throughout the execution process, with each instance reporting progress data to the central coordinator every 30 minutes.

[0081] Resource reorganization is triggered in resource shortage scenarios. The priority queue scanner searches the sorted table for subsequent entries of the current policy. If the priority score of the policy to be launched is 85, it searches for already running policies with scores between 70 and 80. The instance status analyzer selects instances with an "active" status and non-core protection policies. The instance terminator sends a termination command to the selected instances: first, it saves intermediate execution data to persistent storage; then, it removes the resource occupation flag and returns the released resource value to the resource pool. The released resource amount is updated in real time in the resource database; for example, if 10 liters of reagent are released, the chemical resource table is immediately updated. The resource reorganization cycle continues: after terminating an instance, resource sufficiency checks are rerun. If the requirement is met, the replenishment policy is initiated. If not, the next-level policy instance is terminated until the requirement is met or no more instances are available for termination. If no instances are available for termination, a resource shortage alert code "ERR-507" is generated, and the current policy launch request is suspended and placed in a standby queue. The standby queue has a periodic wake-up mechanism that retries the resource reorganization process every six hours until resource conditions are met or administrator intervention terminates.

[0082] The resource reorganization process uses atomic transaction management to ensure data consistency: instance termination and resource release operations are encapsulated within transactions, and any step failure automatically rolls back to the state before the operation. Concurrency conflicts are resolved using a timestamp ordering protocol to avoid resource miscalculation caused by multiple terminations. All operation details are recorded in the audit log, including the terminated instance ID, released resource type, operation time, and change in remaining resource amounts. Log entries are transmitted via encrypted channels to a remote backup server for persistent storage.

[0083] Example 3: See Figure 4 When the spatial analysis phase starts, the geographic information platform loads a multidimensional environmental dataset. The elevation raster data uses a digital terrain model, with a resolution set to five square meters per pixel representing actual terrain. The vegetation distribution polygon layer contains tree species classification codes and canopy cover attribute values. The road network dataset consists of line features, with each road recording its width and pavement material code. The topology modeling engine converts monitoring point coordinates into network nodes, and the distance between adjacent nodes is automatically calculated using the Euclidean distance algorithm. Path connection weights are determined by three parameters: the vegetation cover influence coefficient is a normalized value between 0 and 1 (weighted 0.3 for broadleaf forest areas and 0.7 for coniferous forest areas), the slope resistance coefficient is calculated using an inverse function (weighting decreases by 0.1 for every five degrees of slope angle increase), and the road guidance factor is fixed at 0.9. The system constructs a weighted adjacency matrix to store migration resistance between nodes, with the matrix dimension matching the number of monitoring points.

[0084] Migration path simulation is implemented using a modified Dijkstra algorithm. The algorithm initially sets all node distances to infinity, selecting the insect source core area as the starting point set. During iteration, relaxation is performed: when the shortest path to the current node is updated, its adjacent nodes are traversed and temporary distances are calculated. A dynamic weight parameter adjustment mechanism is activated: when the temperature sensor returns a value exceeding 32°C, the path weight is temporarily multiplied by a temperature correction factor of 1.2; rainfall exceeding 10 mm triggers a humidity correction factor of 0.8. Ultimately, a set of minimum resistance paths is generated, each path annotated with the cumulative resistance value and node sequence.

[0085] The potential energy field model is introduced for gradient field calculations. The input basic fields include the insect population density spatial distribution matrix (resolution 20m×20m) and the wind speed vector field (updated at 0.5 seconds). The core of the physical modeling solves the following partial differential equations:

[0086]

[0087] in: represents the migration potential energy at the coordinate (x, y) at time t, is the population density gradient operator (partial derivative in two-dimensional space), is the wind speed vector (including direction and velocity components), is the temperature time rate of change. Take 0.75 for density weight, Take 0.55 to control the wind field impact intensity, A value of -0.35 indicates that increasing temperature reduces migration potential (the temperature parameter is derived from network weather station data). The solver uses the finite difference method for discretization, with a grid step size matching the spatial resolution and a time step interval of ten minutes. After twenty consecutive iterations, a stable potential energy field is output. Local minima are marked as insect source sinks (coordinates are output in GeoJSON format). Zones where the potential energy gradient in the direction of decrease exceeds a threshold are identified as migration corridors.

[0088] The key diffusion feature extraction unit is configured with a dual-condition filter. The spatial density condition detection module scans each grid cell: a candidate area is generated when the insect population density sensor returns values ​​consistently above the preset threshold of 100 insects / square meter. A temporal persistence verifier connects to a historical database to check whether the density condition has been met for three consecutive days. The area where the two conditions intersect is marked as a red alert zone, and the system extracts its spatial characteristic parameters: a set of core coordinate points, an area expansion rate (square meters / day), and a list of boundary curvature radii. A snapshot of associated environmental parameters is also collected: the average temperature, humidity fluctuation range, and chlorophyll index value corresponding to the red alert zone.

[0089] The ecological model adopts a three-dimensional convolutional neural network architecture. The input layer receives the key diffusion feature tensor with dimensions of 64x64x18 (the first two dimensions correspond to the spatial grid, and the third dimension contains thirteen biological parameters and five environmental parameters). The convolution kernel size is designed as 5x5x5, and the features are extracted by three-dimensional sliding on the feature tensor . The loss function improvement contains a topological constraint term:

[0090]

[0091] wherein is the mean square error of the predicted field and the measured field, takes 0.5 as the constraint strength coefficient, represents the output feature map of the i-th layer of the model, corresponds to the down-sampled version of the real topological feature map at the same scale. Adversarial training operates under the generator-discriminator framework: the generator inputs the key features to generate the predicted prevention and control field, and the discriminator inputs the actual ecological gradient field data (collected from the sensor network). Training iterations are performed for fifteen rounds, each round containing three generator optimizations and two discriminator updates. The optimization algorithm uses adaptive moment estimation, with an initial learning rate of 0.001 decaying by 20% every five rounds. After convergence, the enhanced prevention and control field data is generated, with a format of three-channel grid: the first channel stores the biological inhibition intensity value (0-1 normalized), the second channel records the environmental adjustment parameter, and the third channel is the spatial confidence index.

[0092] The spatial coupling processor performs field data integration operations. The basic environmental factor database outputs temperature grid, humidity distribution map, and soil type classification map. The registration engine uses an affine transformation algorithm to align the enhanced prevention and control field with the basic environmental map at the pixel level. The coupling operation uses pixel-by-pixel weight superposition: for a pixel point with coordinates (x, y), the composite value

[0093]

[0094] wherein: is the biological inhibition intensity value, is the environmental adjustment parameter, represents the environmental factor value. The weight coefficients = 0.6 (biological weight), = 0.25 (adjustment coefficient), = 0.15 (environmental basis). Sigmoid activation function is used to limit the output range. A composite prevention and control field data structure is finally generated, each storage unit contains thirty feature dimensions, and the spatial coverage range completely matches the monitoring area. The timestamp is marked as the UTC time when it is generated. The data compression module uses the Zstandard algorithm for compression storage, and the compression level is set to 12 to balance efficiency and resource consumption. The composite field is stored by a distributed storage system, and each piece of data is attached with a check code for transmission integrity verification.

[0095] Example 4: The system processes the composite prevention and control field data, performs spatio-temporal evolution simulation, field separation, feature reconstruction and iterative verification operations. The entire process relies on a cloud computing platform for operation, and the central processing unit coordinates the interaction of each module. Take a specific example: select a typical citrus planting area as the implementation object. The area is divided into 10,000 grid units (each unit is 10 meters x 10 meters), and the simulation period is set to seven days (168 hours) with a fixed time step of one hour. The initial input data is the composite prevention and control field data, which is a multi-dimensional grid array: each unit stores biological inhibition intensity value (normalized 0-1), environmental regulation parameters (temperature, humidity comprehensive index), and associated confidence value, totaling 30 feature dimensions. The system loads the composite prevention and control field through the data interface, and the format uses HDF5 binary files with a capacity of about 5GB, and transmission is completed through high-speed optical fiber network.

[0096] The spatio-temporal evolution simulation stage starts the model calculation engine. The simulation core is based on the cellular automata architecture: the model initializes the starting state data of the composite prevention and control field (time point T0), and the parameter configuration includes biological action coefficient and environmental factor parameters. The simulation rule is defined as a neighborhood interaction mechanism: the state change of each unit depends on the biological and environmental parameters of its eight adjacent units. For example, when calculating the evolution of grid unit (50, 60), the system scans the neighborhood unit data: obtains the temperature mean value, natural enemy density value, and insect population density baseline value; then applies the rule set to judge the migration behavior, such as the migration probability decreases by 20% when the temperature is above 30°C, and the predation rate increases by 15% when the natural enemy density is high. The model performs a time step loop: the state is updated every hour, and 168 steps are calculated. The system records the state data of each unit at key time points (such as every hour), including the predicted insect population density change value (in percentage form). The simulation outputs the evolved composite field data, including the complete spatio-temporal sequence, and the format is converted to GeoJSON and NetCDF hybrid format for subsequent processing. The simulation process takes about 45 minutes, and GPU cluster is used for accelerated calculation.

[0097] Next, the post-evolution field separation operation is performed. The system inputs the completed composite field data to the separation module. The separation uses a principal component analysis algorithm: the input data is first standardized, reducing the multi-dimensional features to the core dimension. Specifically for the 30 dimensions of the composite field, the covariance matrix is calculated and the eigenvectors are solved; the first three principal components are selected according to the eigenvalue sorting: the first principal component explains biological action, the second principal component explains environmental action, and the third principal component represents noise. The system outputs two independent data fields: the biological action field contains the reduced biological feature vector (core parameters such as natural enemy distribution density); the environmental action field contains the reduced environmental factor vector (core parameters such as heat stress index). The data format is converted to a raster layer, with each cell associated with geographic coordinates and a timestamp. The numerical snapshot of part of the cells after separation (time point is the 24th hour of simulation), see Table 1.

[0098] Table 1: Separated field numerical snapshot table.

[0099] Grid coordinates (x,y) Biological field strength Environmental field value Predicted change in insect population density (%) (50,60) 0.82 0.58 -12.5 (51,60) 0.75 0.63 -11.8 (50,61) 0.78 0.61 -12.0 (51,61) 0.80 0.59 -12.3

[0100] This table is generated by database query and stored in the memory cache area for real-time access.

[0101] The feature reconstruction phase integrates the separated biological action field with real-time monitoring data. Real-time monitoring data comes from the sensor network within the region, with a sampling frequency of once per minute, including real-time insect population density values, temperature and humidity readings. The reconstruction algorithm calls the Kalman filter: input the predicted value of the biological action field and the real-time sensor stream; perform time series alignment, apply the data fusion function after matching to the same space-time coordinates. For example, at the point (50, 60), integrate the biological action value 0.82, the real-time insect population density measurement value 550 per square meter, and the temperature value 28°C; output the reconstructed feature set, including the updated biological activity curve and the environmental impact matrix. After reconstruction, an optimized prevention and control instruction set is generated: the instruction generator analyzes the feature set and maps it to a specific operation command sequence. The instruction format is a JSON array: each instruction contains the operation type, target location (coordinate list), execution time window, and resource requirements. Output example: the instruction set contains 20 commands, which are stored in the central queue after priority sorting and issued to the execution device through the API interface.

[0102] The verification process continuously monitors the execution results of the instruction set. After the system activates the actuator, it collects real-time data on changes in insect population density. Sensors upload density values ​​every 10 minutes, and a time series curve is constructed in the database. The execution cycle covers seven days (168 hours). At the end, the actual reduction curve is calculated by comparing the difference between the starting and ending insect population densities. The system has a preset deviation threshold of 10% (adjustable in the configuration file) and uses the absolute deviation formula: the absolute value of the actual reduction minus the predicted reduction divided by the predicted value. If the deviation exceeds 10%, the system triggers an iterative update process: the key diffusion feature extraction module is re-invoked to generate an updated feature set based on the new sensor data. Subsequently, the ecological model is input and parameter adjustments are performed, modifying the loss coefficient or weight matrix. After the enhanced control field is regenerated, it is fed back to the starting point of this embodiment, and the spatiotemporal evolution simulation, field separation, and reconstruction are repeated. The iteration limit is set to five: a parameter change log is recorded after each loop; the process terminates when the deviation value drops below 10%, the final instruction set is locked, and persistent storage is performed. The loop is managed through event triggers: each round of verification and testing takes 20 minutes; an alarm code is output when resources are exceeded.

[0103] Resource consumption throughout the entire process is recorded using monitoring tools: peak CPU usage reached 85% and memory usage reached approximately 20GB. Data is stored in a distributed system to ensure redundancy. The system performs atomic transactions: any iteration failure automatically rolls back to the previous stable state. Finally, an optimized instruction set is output and delivered to the execution unit, forming a closed-loop control system.

[0104] Example 5: See Figure 5 During system operation, a policy status monitoring and management process is implemented. The monitoring module is configured with a periodic polling mechanism, preset to automatically scan the status data of all active policy instances every 60 minutes. Status data types include execution progress percentage, resource consumption records, and operation interruption codes. This data is transmitted via a structured protocol, formatted as a sequence of key-value pairs, and stored in a log database. The database adopts a distributed architecture, with partitioned storage of policy instance identifiers and timestamps, supporting fast indexed queries. For example, when the status of the instance with identifier "STRAT-001" is updated, the record includes the progress value, resource usage value, and UTC timestamp. The monitoring process is designed based on an event-driven model: a core event trigger monitors sensor network streams in real time, including insect population density detectors and environmental parameter samplers. When new monitoring data triggers an insect status change event, the trigger activates the state change engine. An event is defined as the rate of change of a key monitoring indicator exceeding a preset threshold, typically set as an average daily fluctuation of the raw data value greater than 10%. The detection algorithm uses a differential calculation method: it reads sensor stream data in real time (refreshed every five minutes), calculates the absolute difference between the current value and the value of the previous hour, and divides it by the baseline mean; if the calculated value is greater than 0.10, an event code "EVENT-507" is generated and sent to the sleep strategy processing queue.

[0105] Dormant policy instance management utilizes a memory-resident mechanism. Dormant instances are stored in a dedicated memory pool structured as a priority queue, sorted by policy priority score. Each instance is associated with a metadata set: a policy type code, a resource reservation flag, and a wakeup wait counter. When an event trigger outputs a bug status change event, the system executes the instance wakeup operation. The wakeup function invokes a kernel-level restart routine: the first step is to unfreeze the dormant instance state, extracting instance data from the memory pool and loading it into the active process; the second step is to reset the instance's internal timer to its initial value; and the third step is to change the instance status flag to "awakening." For example, after loading policy instance "SLP-002," the original configuration parameters, such as the target coordinate set and operation time window, are restored. After the wakeup operation is complete, the system immediately checks the instruction matching conditions.

[0106] The instruction matching check is performed through the buffer pool retrieval module. The instruction buffer pool is built on the message queue system and stores the latest generated optimization prevention and control instruction set. The buffer pool structure is aligned with the policy type code: each policy type corresponds to an independent sub-queue, and the instruction format is a JSON object, which contains an operation command string, a geographic coordinate array, and a validity timestamp. The matching process performs double verification: the first verification checks whether the instruction validity timestamp covers the current time; the second verification compares the consistency of the policy type code and the instruction command type. If any verification fails, the system determines that no valid instruction is matched. At this time, the instance state automatically switches to the dormant waiting state: the waiting state manager sets the periodic detection cycle to poll the buffer pool every thirty minutes; at the same time, the instance resource occupancy is frozen; the waiting timeout counter accumulates, and the upper limit is set to twenty-four hours. The timeout triggers automatic re-entry into the dormant program, retaining the original state data.

[0107] When the match is successful, the system activates the policy execution process. First, the instance status mark is modified to "executing", the log record is updated, and a status broadcast message is sent to the coordinator. The instruction execution unit then receives the parsed command parameters: extracts the operation command string, coordinate point set, and resource requirement list. The executor calls the underlying hardware interface: releases the resource reservation tag to bind the physical device, and starts the operation sequence. For example, the execution coordinate array drives the drone to fly along the path, releasing a specified amount of biological agent at each point. During the execution process, the system collects feedback data every fifteen minutes: the execution result data unit contains the actual completion progress, resource consumption increment, and effect indicators. Data collection uses a transaction locking mechanism to ensure write integrity.

[0108] Execution result data is fed back to the dynamic ecological feature set in real time. This feedback channel is implemented through a data flow pipeline: the pipeline input is connected to the execution result unit, and the output is connected to the feature fusion processing module. The data transmission protocol uses binary encoding and processes approximately one thousand records per second. Upon receiving the data, the fusion module executes the feature update algorithm: comparing the new result data with the original set values ​​and applying a differential update function to replace the old parameters. This includes updating the natural enemy suppression coefficient (based on actual natural enemy activity data) and revising the host plant resistance index (based on plant response sensor values). After the update is complete, the dynamic ecological feature set triggers a system reentry mechanism: if the feature change rate exceeds a limit, the prevention and control strategy evaluation loop is automatically restarted. The entire process forms a closed-loop control loop: monitoring state changes drives wake-up and feedback, and feedback data drives feature iteration, maintaining system adaptive operation. Resource management uses a rollback mechanism: in the event of any state anomaly, the system reverts to the previous stable snapshot to ensure continuous operation. During implementation, CPU load was controlled within a threshold of 70%, and network bandwidth usage was capped at a preset 50Mbps.

[0109] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

Claims

1. A comprehensive ecological control method for citrus psyllids, characterized in that: The following steps are involved: Obtain current pest status information based on real-time monitoring data from citrus growing areas; The biological characteristics and related ecological environment characteristics of citrus psyllids were extracted through the ecological characteristic database to form an initial ecological characteristic set; Perform feature fusion processing on real-time monitoring data and initial ecological feature set to generate dynamic ecological feature set; Calculate the priority score of each ecological control strategy based on the spatiotemporal distribution density, migration trend intensity, natural enemy inhibition coefficient, host plant resistance index and environmental stress factor of each feature in the dynamic ecological feature set; A prevention and control strategy ranking table is generated by arranging the strategies in descending order according to their priority scores. The strategies to be executed are selected in turn to determine whether they can achieve the target insect population density reduction within the preset period.

2. The method according to claim 1, characterized in that The extraction of biological characteristics and associated ecological environment characteristics of citrus psyllids through the ecological characteristic library includes: The individual behavior characteristics, population dispersal characteristics, natural enemy restriction characteristics, plant volatile characteristics and microclimate response characteristics of citrus psyllids were extracted hierarchically from the ecological characteristic database. The individual behavior characteristics and population diffusion characteristics of the same level are averaged to obtain the primary fusion characteristics; Perform weighted superposition of primary fusion features and microclimate response features to generate secondary fusion features; Hierarchical feature fusion is iteratively performed until the highest-level feature processing is completed, and a dynamic ecological feature set containing multi-level correlation relationships is output.

3. The method according to claim 1, characterized in that The calculation of the priority score of each ecological prevention and control strategy includes: Obtain the proportion of spatiotemporal distribution density, migration trend intensity, natural enemy inhibition coefficient, host plant resistance index and environmental stress factor corresponding to each strategy; Multiply the migration trend intensity ratio by the first weight coefficient, the natural enemy inhibition coefficient ratio by the second weight coefficient, the host plant resistance index ratio by the third weight coefficient, the environmental stress factor ratio by the fourth weight coefficient, and the spatiotemporal distribution density ratio by the fifth weight coefficient; The priority score is calculated by summing the five product results and deducting the resource consumption cost ratio.

4. The method according to claim 1, characterized in that The step of sequentially selecting the strategies to be executed and determining whether the strategies can achieve the target insect population density reduction within a preset period includes: Calculate the insect population density reduction that can be achieved by a single strategy within a preset period; Comparing the achievable insect population reduction with the target insect population reduction; If the achievable amount is greater than or equal to the target amount, it is determined that the single strategy can be completed; If the achievable amount is less than the target amount, the number of strategies that need to be supplemented is calculated.

5. The method according to claim 4, characterized in that: The calculation of the number of strategies that need to be supplemented includes: Divide the target insect population density reduction by the achievable amount of a single strategy, and round up to get the number of strategies implemented; Detect the total amount of available ecological resources based on the number of strategies implemented; When sufficient resources are available, start the corresponding number of policy instances; When available resources are insufficient, instances ranked lower than the current policy are selected from the activated policies and terminated to release resources until resource requirements are met or there are no more instances to terminate.

6. The method according to claim 1, characterized in that Also includes: Analyze pest spread paths based on geographic information system topology structure; Determine the insect source sinks and migration corridors through gradient field calculations; Extract key diffusion features that meet the insect population density threshold and duration of continuous infection; Key diffusion characteristics were input into pre-trained ecological models to generate enhanced control fields.

7. The method according to claim 6, characterized in that The step of inputting key diffusion characteristics into a pre-trained ecological model to generate an enhanced prevention and control field includes: Introducing topological consistency constraints into the loss function of the ecological model; Through adversarial training, the matching degree between the control field and the real ecological gradient field is optimized and strengthened; The enhanced control field is spatially coupled with basic environmental factors to generate a composite control field containing biotic and abiotic factors.

8. The method according to claim 7, characterized in that: Also includes: Simulate the spatiotemporal evolution of the composite control field; Separate the biological field of action and the environmental field of action after evolution; Reconstruct features of biological interaction fields and real-time monitoring data, and output optimized prevention and control instruction sets.

9. The method according to claim 1, characterized in that: Also includes: Continuously monitor the implementation status of executed strategies; When new monitoring data triggers an insect situation change event, the dormant strategy instance is awakened; If the policy instance does not match a valid instruction after waking up, it will enter the dormant waiting state; If a valid instruction is matched, the implementation status is changed to executing, and the execution result data is fed back to the dynamic ecological feature set.

10. The method according to claim 8, characterized in that: The output optimization control instruction set also includes: Verify the actual insect population density change curve after the instruction set is executed; When the deviation between the actual reduction amount and the predicted value exceeds a threshold, the key diffusion features are re-extracted; Iteratively update the parameters of the enhanced prevention and control field until the deviation value falls into the acceptable range.

Images