A forestry pest information integrated management system and method

Abstract

The present invention discloses an information-based integrated management system and method for forest pests, which relates to the field of forestry information technology. The system includes the following steps: constructing a fuzzy cognitive graph model, predicting the management priority score of each forest area based on time-aligned multi-source data of the forest area, and generating a management priority heat map; obtaining the maximum photochemical efficiency within the forest farm and recording the distribution coordinates based on the management priority heat map, and generating an ecological baseline data set through time-space matching; generating pesticide application instructions based on the priority heat map using a multi-criteria decision-making algorithm, and dispatching drones to perform differentiated pesticide application based on the pesticide application instructions; obtaining a pesticide application correction coefficient and combining it with the ecological baseline data set, calculating the ecological restoration score using the ecological resonance index method, and adjusting the pesticide application instructions in real time based on the ecological restoration score. By constructing the fuzzy cognitive graph model, the present invention realizes a dynamic coupling analysis of the pest risk index and ecological sensitivity, thereby improving the accuracy and adaptability of risk assessment.

Description

Technical Field

[0001] The present invention relates to the field of forestry information technology, and in particular to an information-based integrated management system and method for forestry harmful organisms. Background Art

[0002] Information-based management of forest pests is an important technical field for forest resource protection and ecological security maintenance. Currently, the technical system in this field is mainly based on multi-source data collection, spatial analysis, and decision support systems. Conventional methods usually use remote sensing monitoring technology to obtain vegetation index data, combined with ground sensor networks to collect environmental parameters, and achieve preliminary identification of pest areas through GIS spatial analysis. In terms of data processing, existing technologies generally use time series analysis and spatial interpolation methods to pre-process monitoring data and establish rule-based pest risk assessment models. In recent years, with the development of machine learning technology, some advanced systems have been able to combine historical pest data to train prediction models and achieve a certain degree of prediction of pest occurrence probability. In terms of prevention and control decision-making, existing methods often use hierarchical analysis to determine prevention and control priorities and use visualization technology to display analysis results, providing decision-making references for forestry management.

[0003] However, existing technologies still have room for improvement in risk assessment models and prevention and control decision-making mechanisms. On the one hand, conventional pest risk assessment models often use static weight allocation methods, which are difficult to adapt to the differences in ecological characteristics of different forest areas, resulting in deviations between assessment results and the actual degree of damage. On the other hand, existing prevention and control decision-making systems lack dynamic monitoring and feedback mechanisms for ecological recovery capacity, making it impossible to adjust prevention and control strategies in real time based on the ecological response after pesticide application, affecting the sustainability and accuracy of prevention and control effects. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides an information-based integrated management method for forestry pests to solve the problems of insufficient assessment accuracy and delayed prevention and control decision-making in the existing technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, the present invention provides an information-based integrated management method for forestry pests, which includes dividing a forest farm into multiple forest areas, collecting multi-source data from each forest area and performing time alignment; constructing a fuzzy cognitive graph model, predicting the governance priority score of each forest area based on the time-aligned multi-source data of the forest areas, and generating a governance priority heat map; based on the governance priority heat map, obtaining the maximum photochemical efficiency in the forest farm and recording the distribution coordinates, and generating an ecological baseline data set through time-space matching; according to the priority heat map, using a multi-criteria decision algorithm to generate pesticide application instructions, and dispatching drones to perform differentiated pesticide application according to the pesticide application instructions; obtaining a pesticide application correction coefficient and combining it with the ecological baseline data set, calculating the ecological restoration score by the ecological resonance index method, and adjusting the pesticide application instructions in real time according to the ecological restoration score.

[0008] As a preferred solution of the forestry pest information integrated management method of the present invention, wherein: the forest farm is divided into multiple forest areas by using a spatial cluster analysis method;

[0009] The GIS spatial index is used to spatially align the discretized multi-source data of each forest area, and clock drift compensation is used to dynamically calibrate the space to generate the spatiotemporally aligned multi-source data of the forest area.

[0010] As a preferred solution of the forestry pest information integrated management method of the present invention, wherein: the steps of constructing the fuzzy cognitive graph model are as follows:

[0011] define pest risk index and ecological sensitivity;

[0012] The causal relationship between the pest risk index and ecological sensitivity and multi-source data of forest areas was established through partial least squares path analysis, and the fuzzy cognitive map weight matrix was constructed using the Bayesian network inference algorithm.

[0013] Define the state transition rules between the pest risk index and ecological sensitivity, and iterate and update them through Sigmoid to generate a dynamic state transition mechanism;

[0014] Through the fuzzy dynamic collaborative algorithm, the fuzzy cognitive graph weight matrix is combined with the dynamic state transfer mechanism to form a fuzzy cognitive graph model.

[0015] As a preferred solution of the forestry pest information integrated management method of the present invention, wherein: the steps of predicting the management priority score of each forest area and generating a management priority heat map are as follows:

[0016] The fuzzy cognitive graph model predicts the governance priority score of each forest area based on the aligned multi-source data of the forest area, the pest risk index and the ecological sensitivity;

[0017] The K-means algorithm was used to cluster the pest spread rate in historical forest areas and define low priority thresholds and high priority thresholds.

[0018] Each forest area is marked as hazardous based on the comparison of its governance priority score with the low priority threshold and the high priority threshold;

[0019] The probability density of each marked forest area is interpolated through spatial kernel density estimation to generate a heat map of governance priority.

[0020] As a preferred solution of the information-based integrated management method for forest pests of the present invention, the steps of obtaining the maximum photochemical efficiency in the forest farm and recording the distribution coordinates, and generating an ecological baseline data set through spatiotemporal matching are as follows:

[0021] Obtain the spatial mask of the high-risk forest area in the treatment priority heat map, and use a stratified random sampling method to measure the initial and maximum chlorophyll fluorescence of dark-adapted leaves within the spatial mask of the high-risk forest area using an AM fluorometer. The maximum photochemical efficiency is obtained using the Genty formula, and the distribution coordinates are recorded using GPS.

[0022] The Kriging spatial interpolation method was used to spatially match the NDVI vegetation index, maximum photochemical efficiency, and distribution coordinates within the forest farm, and the improved dynamic time warping algorithm was used to align the time series to generate an ecological baseline dataset.

[0023] As a preferred solution of the forestry pest information integrated management method of the present invention, wherein: the multi-criteria decision algorithm is used to generate the pesticide application instruction, the steps are as follows:

[0024] The spatial mask of the high-hazard forest area was aligned with the ecological baseline dataset using the GDAL library to generate a multivariate raster stack.

[0025] Band splitting method was used to separate the three criteria layers from the multivariate raster stack, and the weights were assigned by analytic hierarchy process.

[0026] According to the weight of each criterion layer, MCDA was used to conduct spatial weighted overlay analysis to obtain the coordinates of the pesticide-applied forest areas and the amount of pesticide applied, and the governance path was generated according to the governance priority score of each forest area.

[0027] The coordinates of the pesticide application forest area, the amount of pesticide applied and the treatment path are mapped into pesticide application instructions through multi-criteria decision mapping.

[0028] As a preferred solution of the information-based integrated management method for forest pests of the present invention, the steps are as follows: calculating the ecological restoration score by the ecological resonance index method and adjusting the pesticide application instructions in real time according to the ecological restoration score.

[0029] Based on the application rate in the application instructions and the maximum photochemical efficiency in the ecological baseline dataset, the dose-response surface method was used to calculate the application rate correction coefficient. Combined with the ecological baseline dataset, the ecological recovery score of each forest area was calculated using the ecological resonance index method.

[0030] Based on the ecological restoration score of each forest area, dynamic clustering analysis is used to identify and mark the ecological recovery deviation status, and the pesticide application correction coefficient is updated through an incremental learning algorithm, and the pesticide application instructions are optimized according to the updated pesticide application correction coefficient.

[0031] In a second aspect, the present invention provides an information-based integrated management system for forest pests, including a data acquisition module, a priority division module, a baseline generation module, an instruction generation module, and an ecological restoration assessment module;

[0032] The data collection module is used to divide the forest farm into multiple forest areas, collect multi-source data from each forest area, and perform time alignment;

[0033] The priority classification module is used to build a fuzzy cognitive graph model, predict the governance priority score of each forest area based on the time-aligned multi-source data of the forest area, and generate a governance priority heat map;

[0034] The baseline generation module is used to obtain the maximum photochemical efficiency within the forest farm and record the distribution coordinates based on the governance priority heat map, and generate an ecological baseline dataset through spatiotemporal matching;

[0035] The instruction generation module is used to generate spraying instructions based on the priority heat map and a multi-criteria decision-making algorithm, and dispatch drones to perform differentiated spraying according to the spraying instructions;

[0036] The ecological restoration assessment module is used to obtain the pesticide application correction coefficient and combine it with the ecological baseline data set to calculate the ecological restoration score through the ecological resonance index method, and make real-time adjustments to the pesticide application instructions based on the ecological restoration score.

[0037] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the method for integrated information management of forestry pests as described in the first aspect of the present invention is implemented.

[0038] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the method for integrated information management of forestry pests as described in the first aspect of the present invention is implemented.

[0039] The beneficial effects of the present invention are as follows: by constructing a fuzzy cognitive graph model, a dynamic coupling analysis of the pest risk index and ecological sensitivity is realized, thereby improving the accuracy and adaptability of risk assessment; by calculating the ecological restoration score of each forest area through the ecological resonance index method, a dynamic correlation evaluation of the pesticide application effect and the ecological restoration status is realized, thereby enhancing the pertinence and timeliness of prevention and control measures. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0041] Figure 1 This is a flow chart of the integrated information management method for forestry pests.

[0042] Figure 2 This is a schematic diagram of the integrated information management system for forestry pests.

[0043] Figure 3 Flowchart constructed for the fuzzy cognitive map model.

[0044] Figure 4 Flowchart generated for the governance priority heat map. DETAILED DESCRIPTION

[0045] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0046] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0047] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0048] Reference Figures 1 to 4 , is an embodiment of the present invention, which provides an information-based integrated management method for forestry pests, comprising the following steps:

[0049] S1. Divide the forest farm into multiple forest areas, collect multi-source data from each forest area and perform time alignment;

[0050] Multi-source data of forest areas include NDVI vegetation index, soil temperature, leaf surface electrical conductivity and soil moisture content;

[0051] Furthermore, the NDVI vegetation index is obtained by aerial survey using a multispectral sensor mounted on an unmanned aerial vehicle (UAV); soil temperature and moisture content are monitored in real time using embedded soil sensors; and leaf conductivity is measured at fixed points within a typical sample plot using a handheld plant physiological monitor.

[0052] The forest farm was divided into several forest areas through spatial cluster analysis method;

[0053] Furthermore, the DBSCAN density clustering algorithm was used to spatially cluster the multi-source data from each forest area, forming similar forest units. The silhouette coefficient was then used to evaluate clustering effectiveness, and forest units with ambiguous boundaries were subdivided. Finally, a GIS spatial index was used to map forest units to geographic coordinates, completing the division of the forest farm into multiple forest areas. The silhouette coefficient is calculated as the ratio of the average distance between each spatial coordinate point in the forest area and other spatial coordinate points in the same cluster (cohesion) to the average distance to the nearest cluster point (separation). Its value range is [-1, 1], with values closer to 1 indicating better clustering effectiveness.

[0054] The GIS spatial index is used to spatially align the discretized multi-source data of each forest area, and clock drift compensation is used to dynamically calibrate the space to generate the spatiotemporally aligned multi-source data of the forest area.

[0055] Furthermore, the spatial location index of multi-source data in the forest area is first established based on the R-tree index structure, and multi-source data such as NDVI vegetation index, soil temperature, leaf surface conductivity and soil moisture are matched according to geographic coordinates through spatial join operations; then the NTP protocol is used to synchronize the timestamps of each data acquisition device, and the linear interpolation compensation algorithm is applied to the data records with clock drift for time calibration; finally, the spatially aligned multi-source data are fused with the time calibration results through spatiotemporal association rules to generate spatiotemporally aligned multi-source data of the forest area.

[0056] S2. Construct a fuzzy cognitive graph model to predict the governance priority score of each forest area based on the time-aligned multi-source data of the forest area and generate a governance priority heat map;

[0057] Based on the aligned multi-source data of the forest area, define the pest risk index and ecological sensitivity;

[0058] Furthermore, the pest risk index is calculated by a weighted combination of normalized NDVI vegetation index anomalies, soil temperature deviations, and sudden increases in leaf conductivity. The value range is [0, 1], with larger values indicating higher pest risk. The ecological sensitivity is determined by the ratio of the soil moisture gradient change rate to the historical maximum photochemical efficiency. The value range is [0.1, 0.9], with larger values indicating stronger ecological sensitivity.

[0059] The causal relationship between the pest risk index and ecological sensitivity and multi-source data of forest areas was established through partial least squares path analysis, and the fuzzy cognitive map weight matrix was constructed using the Bayesian network inference algorithm.

[0060] Furthermore, the method first used spatiotemporally aligned multi-source forest data as exogenous latent variables, and the pest risk index and ecological sensitivity as endogenous latent variables. An iterative weighting algorithm was used to calculate the path coefficients between the latent variables, and bootstrap sampling was used to verify path significance. For significant paths that passed the test, standardized regression coefficients were used to quantify causal relationships, such as the positive impact of the NDVI vegetation index on the pest risk index and the negative impact of soil moisture on ecological sensitivity. Based on causal relationships, a Bayesian network inference algorithm was used to construct a fuzzy cognitive map weight matrix. The following process was used: the verified path coefficients were converted into a conditional probability table of the Bayesian network, and the transition probabilities between nodes were estimated through Markov chain Monte Carlo sampling. The expectation-maximization algorithm was used to optimize the network parameters, ultimately generating a fuzzy cognitive map weight matrix that reflects the nonlinear relationship between the multi-source data and the evaluation indicators. The matrix elements were restricted to the interval [-1, 1], with positive values indicating a promoting relationship and negative values indicating an inhibitory relationship.

[0061] Based on the changing patterns of multi-source data in historical forest areas, the state transition rules between the pest risk index and ecological sensitivity are defined, and iterative updates are performed through Sigmoid to generate a dynamic state transition mechanism.

[0062] Furthermore, the authors first analyzed the temporal variation patterns of pest risk index and ecological sensitivity in historical data, identifying typical scenarios where the pest risk index increases when the NDVI vegetation index continuously decreases and soil temperature continues to rise. Furthermore, they identified key conditions for improved ecological sensitivity when soil moisture recovers and leaf conductivity stabilizes (for example, an NDVI decline for three consecutive months and a 0.5°C increase in the monthly mean soil temperature triggers an escalation in pest risk). These key conditions were converted into state transition rules, and a rule base consisting of "IF-THEN" conditional statements was established. The rule base was iteratively updated using a Sigmoid function: the conditional portion of each rule was input into the Sigmoid function to calculate the activation value. The slope parameter of the Sigmoid function was then adjusted by backpropagating the error between the actual observed values and the rule predictions from the current multi-source data in the forest area. After multiple iterations, a dynamic state transition mechanism with adaptive capabilities was developed.

[0063] It should be noted that the variation patterns of historical multi-source forest data refer to typical scenarios and correlation patterns before and after an insect infestation. For example, the NDVI vegetation index shows a continuous drop of more than 0.15 three weeks before an insect infestation outbreak, soil temperatures remain above 28°C for five consecutive days during the insect infestation period, and leaf conductivity in ecologically sensitive areas often fluctuates by more than two standard deviations of the historical mean.

[0064] Through the fuzzy dynamic collaborative algorithm, the fuzzy cognitive graph weight matrix is combined with the dynamic state transfer mechanism to form a fuzzy cognitive graph model.

[0065] Furthermore, a fuzzy association mapping method is used to establish a mapping relationship between the fuzzy cognitive map weight matrix and the dynamic state transition mechanism, and the causal nodes in the fuzzy cognitive map weight matrix are mapped to the conditional variables of the state transition rule. Then, a fuzzy inference engine is used to synchronously process the causal influence of the weight matrix and the temporal constraints of the state transition rule, and the activation degree of each factor is quantified through the membership function. Finally, a collaborative optimization algorithm is applied to balance the static weight relationship and the intensity of the dynamic transition rule. When the difference between the predicted value output by the weight matrix and the result produced by the state transition mechanism exceeds a critical value (usually in the range of [0, 1]), the contribution weights of the two are automatically adjusted, so that the fuzzy cognitive map model can not only reflect the causal relationship between multi-source data, but also adapt to the dynamic characteristics of environmental changes.

[0066] The fuzzy cognitive graph model predicts the governance priority score of each forest area based on the aligned multi-source data of the forest area, the pest risk index, and the ecological sensitivity. The expression is:

[0067] ;

[0068] in, It is The governance priority score of each forest area, is the Sigmoid function, is the pest risk index, is the ecological sensitivity index, It is Multi-source data of forest areas, is the index variable of the forest area, is the total number of forest areas;

[0069] Furthermore, firstly, the pest risk index Ecological Sensitivity Index The interaction of the two groups is normalized by product to reflect the basic threat level; for the multi-source data of each forest area , respectively evaluate the dynamic matching degree between pest risk and multi-source data of forest areas, and the coordination between ecological sensitivity and multi-source data of forest areas. These two complementary indicators quantify the actual impact through nonlinear adjustment functions; finally, integrate the local impact assessment of all forest areas to form a comprehensive governance priority judgment. Finally, through the Sigmoid function Perform normalization and output the Evaluation of the governance priorities of forest areas The scoring value range is 0 to 1, and the larger the value, the higher the governance priority.

[0070] It should be noted that the pest risk index H is a quantitative value obtained by weighted summation of historical pest occurrence frequency, insect population density monitoring data, and host plant distribution; the ecological sensitivity index is a quantitative value obtained by the hierarchical analysis method through the rarity of vegetation types, species diversity, and soil erosion sensitivity, which characterizes the vulnerability of forest ecosystems to interference; both are standardized using a 0-1 range, and larger values indicate higher pest risks or stronger ecological sensitivity, respectively.

[0071] The K-means algorithm is used to cluster the spread rate of insect pests in historical forest areas and define the low priority threshold P1 and the high priority threshold P2;

[0072] Furthermore, after the K-means algorithm performed three-category clustering on the pest spread rate in the historical forest area, the upper quartile of the lowest risk category in the clustering results was selected as P1 (value range 0.2-0.4), representing the safety critical value of the pest spread rate; the lower quartile of the highest risk category was selected as P2 (value range 0.6-0.8), marking the dangerous critical value of the pest spread rate.

[0073] when When <P1, the current forest area is marked as a low-hazard forest area; for example: the priority score of a forest area governance =0.35, lower than P1=0.4.

[0074] When P1≤ <P2, the current forest area is marked as a medium-hazard forest area; for example: the priority score of a forest area governance =0.55, which is between P1=0.4 and P2=0.7.

[0075] when When ≥P2, the current forest area is marked as a high-hazard forest area; for example: the priority score of a forest area governance =0.8, exceeding P2=0.7.

[0076] The probability density of each marked forest area is interpolated through spatial kernel density estimation to generate a heat map of governance priority.

[0077] Furthermore, based on the spatial distribution coordinates of the marked low-hazard forest areas, medium-hazard forest areas and high-hazard forest areas, the Gaussian kernel function is used to calculate the probability density of the hazard level of each geographical location; a regular grid is established within the forest farm, and the kernel function contribution value of the marking results of the surrounding forest areas is superimposed at each regular grid node; the smoothness of the density estimation is controlled by the bandwidth parameter, so that the low-hazard forest area forms a blue cold area, the high-hazard forest area forms a red hot area, and the medium-hazard forest area forms a yellow transition area; finally, a continuous probability density surface reflecting the spatial distribution characteristics of governance priorities is output, namely, the governance priority heat map.

[0078] S3. Based on the governance priority heat map, obtain the maximum photochemical efficiency in the forest farm and record the distribution coordinates, and generate an ecological baseline dataset through spatiotemporal matching;

[0079] Use the inRange function in the OpenCV library to obtain the spatial mask of the high-risk forest area in the governance priority heat map;

[0080] It should be noted that the spatial mask refers to a binary image matrix generated based on the coordinates of the marked high-hazard forest areas and used to identify the spatial distribution range of high-hazard forest areas; for example, the grid cells corresponding to all high-hazard forest areas within the coordinate range (X1, Y1) to (X2, Y2) in the northwest area of a certain forest farm are assigned a value of 1 in the mask, and the remaining areas are assigned a value of 0, forming a high-hazard area identification map with clear spatial boundaries.

[0081] By using a stratified random sampling method, the initial and maximum chlorophyll fluorescence of dark-adapted leaves were measured using an AM fluorometer within a spatial mask in a high-hazard forest area. The maximum photochemical efficiency was obtained using the Genty formula, and the distribution coordinates were recorded using GPS.

[0082] Furthermore, the spatial mask was first divided into several sampling units of equal area, and 3-5 sampling points were randomly selected in each unit; the initial chlorophyll fluorescence and maximum chlorophyll fluorescence of dark-adapted leaves at the sampling points were measured using an AM fluorometer; the maximum photochemical efficiency of each sampling point was calculated based on Genty; and the precise coordinates of each sampling point were recorded using a GPS device; finally, a data set containing the maximum photochemical efficiency measurement values and their corresponding spatial positions was obtained.

[0083] It should be noted that dark-adapted leaves refer to plant leaves that have reached a stable physiological state after being fully treated with darkness; for example, the leaves to be measured are shielded with a light-shielding clip for 30 minutes before measurement to allow their photosynthetic apparatus to completely relax, and then the chlorophyll fluorescence is measured.

[0084] The Kriging spatial interpolation method was used to spatially match the NDVI vegetation index, maximum photochemical efficiency, and distribution coordinates within the forest farm, and an improved dynamic time warping algorithm was used to align the time series to generate an ecological baseline dataset.

[0085] It should be noted that the improvement process of the dynamic time warping algorithm is as follows: first, the single time bending path of the dynamic time warping algorithm is expanded to a spatiotemporal coupled warping path (spatial adjacent nodes are forced to share a time offset of >60%); secondly, a dual-channel distance measurement function is designed, the LSTM channel calculates the cosine similarity of seasonal trends, and the Transformer channel detects local mutation points (such as the week of insect pest outbreak) through self-attention, and the two are dynamically fused through spatial entropy weights; finally, an improved dynamic time warping algorithm is formed to improve the spatiotemporal alignment accuracy and ecological event detection capabilities.

[0086] S4. Based on the priority heat map, a multi-criteria decision-making algorithm is used to generate spraying instructions, and the drone is dispatched to perform differentiated spraying according to the spraying instructions;

[0087] The spatial mask of the high-hazard forest area was aligned with the ecological baseline dataset using the GDAL library to generate a multivariate raster stack.

[0088] Furthermore, the Warper tool of the GDAL library was first used to read the raster data of the high-hazard forest area spatial mask and the raster layers such as the NDVI vegetation index and maximum photochemical efficiency in the ecological baseline dataset; the nearest neighbor resampling method was used to unify the layers to the same spatial resolution; the geographic coordinate conversion was used to ensure that the pixel positions of all layers were strictly aligned; finally, the aligned spatial mask and the layers of the ecological baseline data were superimposed in band order to form a multivariate raster stack containing spatial location information and multidimensional ecological parameters.

[0089] Band unpacking was used to separate three criterion layers from the multivariate grid stack; the three criterion layers included the criterion layer of hazard degree, the criterion layer of ecological sensitivity and the criterion layer of pesticide feasibility.

[0090] Furthermore, the maximum photochemical efficiency was first used as the first band, the treatment priority heat map as the second band, the surface slope data obtained based on DEM as the third band, and the distance between the pixel and the water source in each multivariate raster stack obtained through hydrological analysis as the fourth band; then, the R channel intensity value data (range 0-255) in the treatment priority heat map of the second band was extracted by the band index extraction method as the hazard degree criterion layer; then, based on the maximum photochemical efficiency data of the first band, the ecological sensitivity threshold (range 0.1-0.64) was defined based on the historical maximum photochemical efficiency and binary segmentation was performed to generate the ecological sensitivity criterion layer; finally, dual conditional judgment was performed on the surface slope data of the third band and the distance between the pixel and the water source of the fourth band (for example: slope <25° and water source distance >50m), and the application feasibility criterion layer was generated through pixel-by-pixel logical and Boolean operations.

[0091] The weights of the hazard degree criterion layer, ecological sensitivity criterion layer and pesticide application feasibility criterion layer were allocated by using the analytic hierarchy process.

[0092] Furthermore, a judgment matrix was first constructed, including the hazard level criterion layer, the ecological sensitivity criterion layer, and the pesticide application feasibility criterion layer. The relative importance of each criterion layer was determined through pairwise comparison. The eigenvector method was used to calculate the maximum eigenvalue of the judgment matrix and its corresponding normalized eigenvector to obtain the initial weights of each criterion layer. A consistency check was performed to ensure the logical rationality of the judgment matrix, and the comparison scale was adjusted if necessary. Finally, a weight distribution result for the criterion layers that met the consistency requirements was output, with the sum of the weights of the three criterion layers strictly equal to 1. For example, the hazard level criterion layer had a weight of 0.5, the ecological sensitivity criterion layer had a weight of 0.3, and the pesticide application feasibility criterion layer had a weight of 0.2.

[0093] According to the weight of each criterion layer, MCDA was used to conduct spatial weighted overlay analysis to obtain the coordinates of the pesticide-applied forest areas and the amount of pesticide applied, and the governance path was generated according to the governance priority score of each forest area.

[0094] Furthermore, based on the weights of the hazard severity, ecological sensitivity, and application feasibility criteria layers determined by the analytic hierarchy process (AHP), the ecological parameters in the multivariate raster stack were standardized to eliminate dimensional differences. Using a weighted linear combination method, the spatial data corresponding to each criterion layer were multiplied by their weights and then superimposed to generate a pesticide application priority distribution surface. The continuous priority values were divided into several application rate levels using the natural breaks classification method. Taking into account the temporal dimension of the management priority scores, a shortest path algorithm was used to plan a management path network that balanced spatial and temporal efficiency. The final output was a decision result set containing the geographic coordinates of the treated forest area, the application rate level, and the path sequence. For example, a high-hazard forest area (coordinates X:123.45, Y:67.89) was assigned an application rate level of 3 and was assigned to the third operation point on the second management path.

[0095] Through multi-criteria decision mapping, the coordinates of the pesticide application forest area, the amount of pesticide application and the treatment path are mapped into pesticide application instructions, and drones are dispatched according to the pesticide application instructions to perform differentiated pesticide application.

[0096] Furthermore, the coordinate set of the spraying forest area is converted into a sequence of waypoints recognizable by the drone, and a mapping relationship between spatial location and operating parameters is established. The spraying rate level is mapped to specific operating parameters such as spray flow rate, flight altitude, and spray width through preset conversion rules, and the priority of the treatment path is converted into the waypoint access sequence and time scheduling plan. By integrating spatial location information, operating parameter settings, and time scheduling requirements, a structured spraying instruction set is generated to ensure that the drone can accurately perform differentiated spraying operations. For example, the spraying instruction corresponding to the waypoint (X:123.45, Y:67.89) includes parameters such as a spray flow rate of 250ml / min, a flight altitude of 4.5m, and a spray width of 3.2m, and is marked as the third operation task to be executed at 10:00 am the next day.

[0097] S5. Obtain the pesticide application rate correction coefficient and combine it with the ecological baseline data set to calculate the ecological restoration score through the ecological resonance index method, and adjust the pesticide application instructions in real time based on the ecological restoration score.

[0098] Based on the application rate in the application instruction and the maximum photochemical efficiency in the ecological baseline data set, the dosage correction coefficient is calculated by the dose-response surface method, and the expression is:

[0099] ;

[0100] in, It is Forest District The dosage correction factor for each period, is the dose-response curvature parameter, is the photochemical efficiency threshold (range 0.75-0.85), is the maximum photochemical efficiency, It is Forest District The amount of pesticide applied during the period, is the base of natural logarithms;

[0101] Furthermore, based on the maximum photochemical efficiency in the ecological baseline dataset and the dosage in the application instructions , construct an S-type dose-response function relationship; the maximum photochemical efficiency With the preset photochemical efficiency threshold By comparing the curvature parameters Controls the steepness of the response curve; when the maximum photochemical efficiency Deviation threshold When the amount of pesticide applied Correction factor The influence of the change shows nonlinear characteristics; the correction coefficient of the final output The value range is between 0 and 1, reflecting the adjustment ratio of the dosage under the current photosynthetic efficiency level. For example, the maximum photochemical efficiency measured in a forest area during the monitoring period is =0.82, the corresponding dosage correction coefficient =0.86.

[0102] It should be noted that the photochemical efficiency threshold It is defined by statistical analysis based on the maximum photochemical efficiency of historical vegetation in the forest farm using the percentile method, with a value range of 0.75-0.85;

[0103] Based on the dosage correction coefficient and the ecological baseline data set, the ecological restoration score of each forest area was calculated using the ecological resonance index method, which is expressed as follows:

[0104] ;

[0105] in, It is Ecological restoration scores of forest areas, It is The first forest area The value of ecological baseline data, It is The historical average value of ecological baseline data (same season and same forest area), It is The historical standard deviation of the ecological baseline data, is the total number of ecological baseline data, Is a symbolic function (used to determine The difference can be positive, negative, or zero).

[0106] Further, get the Various ecological baseline data of forest areas , compared with the historical average value of the same forest area in the same season and historical standard deviation Conduct standardized comparisons; the degree of deviation of each ecological baseline data Divide by Get the standardized difference and multiply it by the corresponding dosage correction coefficient The sign function is used to retain the deviation direction information and distinguish between positive improvement and negative degradation. Finally, the average of the adjusted results of all ecological baseline data is calculated to obtain a score reflecting the overall ecological recovery status of the forest area. For example, after the application of pesticides in a certain forest area, the NDVI vegetation index increased by 1.2 standard deviations compared with the historical average, and the soil moisture content decreased by 0.8 standard deviations. The ecological recovery score was calculated. =0.45.

[0107] Based on the ecological recovery scores of each forest area, dynamic cluster analysis is used to identify and mark ecological recovery deviation states; ecological recovery deviation states include normal, warning, and abnormal;

[0108] Furthermore, we first collected the ecological restoration scores of each forest area for consecutive periods. , a density-based clustering algorithm was used to analyze the score distribution characteristics; the silhouette coefficient was used to determine the optimal number of clusters, and the forest area was divided into three categories: "normal" state with concentrated ecological restoration scores, "warning" state with high scores, and "abnormal" state with low scores; the time continuity constraint was considered in the clustering process to ensure that the state labeling results had temporal consistency. For example, the ecological restoration scores of a forest area for three consecutive monitoring cycles were It is stable in the range of 0.4-0.6 and is clustered as a "normal" state.

[0109] Based on the deviation state of ecological restoration, the pesticide application rate correction coefficient is updated through the incremental learning algorithm, and the pesticide application instructions are optimized according to the updated pesticide application rate correction coefficient.

[0110] Furthermore, based on the ecological recovery deviation status marking results obtained by dynamic cluster analysis, the curvature parameters in the dose-response surface were adjusted using the online gradient descent method for forest areas in the "warning" and "abnormal" states. ; By calculating the current dosage correction coefficient The error with the target recovery score is back-propagated to update the model parameters; the updated dose-response surface is used to recalculate the dosage correction coefficient of each forest area. , and then generate optimized pesticide application instructions that take into account the dynamic characteristics of ecological restoration. For example, the pesticide application correction coefficient for a certain "abnormal" forest area Adjust from 0.7 to 0.8, and increase the application amount by 15% accordingly.

[0111] This embodiment also provides an information-based integrated management system for forest pests, including: a data acquisition module, a priority classification module, a baseline generation module, an instruction generation module, and an ecological restoration assessment module;

[0112] The data collection module is used to divide the forest farm into multiple forest areas, collect multi-source data from each forest area, and perform time alignment;

[0113] The priority classification module is used to build a fuzzy cognitive graph model, predict the governance priority score of each forest area based on the time-aligned multi-source data of the forest area, and generate a governance priority heat map;

[0114] The baseline generation module is used to obtain the maximum photochemical efficiency within the forest farm and record the distribution coordinates based on the governance priority heat map, and generate an ecological baseline dataset through spatiotemporal matching;

[0115] The instruction generation module is used to generate spraying instructions based on the priority heat map and a multi-criteria decision-making algorithm, and dispatch drones to perform differentiated spraying according to the spraying instructions;

[0116] The ecological restoration assessment module is used to obtain the pesticide application correction coefficient and combine it with the ecological baseline data set to calculate the ecological restoration score through the ecological resonance index method, and make real-time adjustments to the pesticide application instructions based on the ecological restoration score.

[0117] This embodiment also provides a computer device suitable for the information-based integrated management method of forestry pests, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the information-based integrated management method of forestry pests proposed in the above embodiment.

[0118] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0119] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for realizing the information-based integrated management of forestry pests as proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0120] In summary, the present invention achieves dynamic coupling analysis of pest risk index and ecological sensitivity through the steps of constructing a fuzzy cognitive graph model, thereby improving the accuracy and adaptability of risk assessment; calculates the ecological restoration score of each forest area through the ecological resonance index method, thereby achieving dynamic correlation evaluation between pesticide application effect and ecological restoration status, and enhancing the pertinence and timeliness of prevention and control measures.

[0121] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for comprehensive information management of forest pests, characterized by: include, Divide the forest farm into multiple forest areas, collect multi-source data from each forest area and perform time alignment; the multi-source data for each forest area includes NDVI vegetation index, soil temperature, leaf surface conductivity, and soil moisture content; A fuzzy cognitive graph model was constructed to predict the governance priority scores of each forest area based on time-aligned multi-source forest area data, and a governance priority heat map was generated. To build a fuzzy cognitive graph model, the steps are as follows: Based on the aligned multi-source data of the forest area, define the pest risk index and ecological sensitivity; The pest risk index is calculated by a weighted combination of normalized NDVI vegetation index anomalies, soil temperature deviations, and leaf conductivity spikes. A larger value indicates a higher pest risk. Ecological sensitivity is determined by the ratio of the soil moisture gradient change rate to the historical maximum photochemical efficiency. A larger value indicates a stronger ecological sensitivity. The causal relationship between the pest risk index and ecological sensitivity and multi-source data of forest areas was established through partial least squares path analysis, and the fuzzy cognitive map weight matrix was constructed using the Bayesian network inference algorithm. Define the state transition rules between the pest risk index and ecological sensitivity, and iterate and update them through Sigmoid to generate a dynamic state transition mechanism; Through the fuzzy dynamic collaborative algorithm, the fuzzy cognitive map weight matrix is combined with the dynamic state transfer mechanism to form a fuzzy cognitive map model; Based on the heat map of governance priorities, the maximum photochemical efficiency within the forest farm was obtained and the distribution coordinates were recorded. The ecological baseline dataset was generated through spatiotemporal matching. To predict the governance priority scores of each forest area and generate a governance priority heat map, the steps are as follows: The fuzzy cognitive graph model predicts the governance priority score of each forest area based on the aligned multi-source data of the forest area, the pest risk index and the ecological sensitivity; The K-means algorithm was used to cluster the pest spread rate in historical forest areas and define low-priority thresholds and high-priority thresholds. Each forest area is marked as hazardous based on the comparison of its governance priority score with the low priority threshold and the high priority threshold; The probability density of each marked forest area is interpolated through spatial kernel density estimation to generate a heat map of governance priorities; Based on the treatment priority heat map, a multi-criteria decision-making algorithm is used to generate pesticide application instructions, and drones are dispatched according to the pesticide application instructions to perform differentiated pesticide application. Based on the application rate in the application instruction and the maximum photochemical efficiency in the ecological baseline data set, the dose-response surface method was used to calculate the application rate correction factor; Maximum photochemical efficiency based on the ecological baseline dataset and the dosage in the application instructions , construct an S-type dose-response function relationship; the maximum photochemical efficiency With the preset photochemical efficiency threshold By comparing the curvature parameters Controls the steepness of the response curve; When the maximum photochemical efficiency Deviation threshold When the amount of pesticide applied Correction factor The influence of shows nonlinear change characteristics; Correction factor of final output The value range is between 0 and 1, reflecting the adjustment ratio of the pesticide dosage under the current level of photosynthesis efficiency; Obtain the pesticide application rate correction coefficient and combine it with the ecological baseline data set to calculate the ecological restoration score using the ecological resonance index method, and adjust the pesticide application instructions in real time based on the ecological restoration score; Based on the dosage correction coefficient and the ecological baseline data set, the ecological restoration score of each forest area was calculated using the ecological resonance index method, which is expressed as follows: ； in, It is Ecological restoration scores of forest areas, It is The first forest area The value of ecological baseline data, It is The historical average value of the same ecological baseline data in the same season and forest area, It is The historical standard deviation of the ecological baseline data, is the total number of ecological baseline data, Is a symbolic function, which is used to judge The difference is positive, negative, or zero.

2. The method for comprehensive information management of forest pests according to claim 1, wherein: The forest farm was divided into several forest areas through spatial cluster analysis method; The GIS spatial index is used to spatially align the discretized multi-source data of each forest area, and clock drift compensation is used to dynamically calibrate the space to generate the spatiotemporally aligned multi-source data of the forest area.

3. The method for comprehensive information management of forest pests according to claim 1, wherein: The steps of obtaining the maximum photochemical efficiency in the forest farm and recording the distribution coordinates, and generating the ecological baseline dataset through spatiotemporal matching are as follows: Obtain the spatial mask of the high-risk forest area in the treatment priority heat map, and use a stratified random sampling method to measure the initial and maximum chlorophyll fluorescence of dark-adapted leaves within the spatial mask of the high-risk forest area using an AM fluorometer. The maximum photochemical efficiency is obtained using the Genty formula, and the distribution coordinates are recorded using GPS. The Kriging spatial interpolation method was used to spatially match the NDVI vegetation index, maximum photochemical efficiency, and distribution coordinates within the forest farm, and the improved dynamic time warping algorithm was used to align the time series to generate an ecological baseline dataset.

4. The method for comprehensive information management of forest pests according to claim 3, wherein: The steps of using the multi-criteria decision-making algorithm to generate the pesticide application instruction are as follows: The spatial mask of the high-hazard forest area was aligned with the ecological baseline dataset using the GDAL library to generate a multivariate raster stack. Band splitting method was used to separate the three criteria layers from the multivariate raster stack, and the weights were assigned by analytic hierarchy process. According to the weight of each criterion layer, MCDA was used to conduct spatial weighted overlay analysis to obtain the coordinates of the pesticide-applied forest areas and the amount of pesticide applied, and the governance path was generated according to the governance priority score of each forest area. The coordinates of the pesticide application forest area, the amount of pesticide applied and the treatment path are mapped into pesticide application instructions through multi-criteria decision mapping.

5. The method for comprehensive information management of forest pests according to claim 1, wherein: The ecological restoration score is calculated by the ecological resonance index method, and the pesticide application instructions are adjusted in real time according to the ecological restoration score. The steps are as follows: Based on the application rate in the application instructions and the maximum photochemical efficiency in the ecological baseline dataset, the dose-response surface method was used to calculate the application rate correction coefficient. Combined with the ecological baseline dataset, the ecological recovery score of each forest area was calculated using the ecological resonance index method. Based on the ecological restoration score of each forest area, dynamic clustering analysis is used to identify and mark the ecological recovery deviation status, and the pesticide application correction coefficient is updated through an incremental learning algorithm, and the pesticide application instructions are optimized according to the updated pesticide application correction coefficient.

6. An information-based integrated management system for forest pests, based on the information-based integrated management method for forest pests according to any one of claims 1 to 5, characterized in that: Including data collection module, priority division module, baseline generation module, instruction generation module and ecological restoration assessment module; The data collection module is used to divide the forest farm into multiple forest areas, collect multi-source data from each forest area, and perform time alignment; The priority classification module is used to build a fuzzy cognitive graph model, predict the governance priority score of each forest area based on the time-aligned multi-source data of the forest area, and generate a governance priority heat map; The baseline generation module is used to obtain the maximum photochemical efficiency within the forest farm and record the distribution coordinates based on the governance priority heat map, and generate an ecological baseline dataset through spatiotemporal matching; The instruction generation module is used to generate pesticide application instructions based on the treatment priority heat map and adopt a multi-criteria decision-making algorithm, and dispatch drones to perform differentiated pesticide application according to the pesticide application instructions; The ecological restoration assessment module is used to obtain the pesticide application correction coefficient and combine it with the ecological baseline data set to calculate the ecological restoration score through the ecological resonance index method, and make real-time adjustments to the pesticide application instructions based on the ecological restoration score.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the forestry pest information integrated management method according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for comprehensive information management of forestry pests according to any one of claims 1 to 5 are implemented.

Images