Forestry pest early warning model construction system and method

Abstract

The invention discloses a forestry pest early-warning model construction system and method, and relates to the technical field of forestry early-warning model construction.An influence factor is output for each piece, the initial monitoring force of each piece is revised through the influence factor, and after real-time data of all pieces are subjected to comprehensive analysis regularly, the real-time data of all pieces are subjected to early-warning model construction; and obtaining an overall performance index of the monitored forest region, combining the overall performance index with a dynamic judgment rule, analyzing whether the current monitoring strength of all the pieces needs to be optimized and adjusted, dynamically generating a personalized scheme for each piece based on an analysis result, and mapping the personalized scheme to the monitored forest region early warning model for display. According to the construction system, monitoring strength can be dynamically adjusted according to actual conditions of different slices, monitoring pertinence and precision are improved, real-time data of the slices are periodically and comprehensively analyzed, an overall performance index is generated, and a dynamic judgment rule and a dynamic adjustment monitoring strategy are combined, so that the system can adapt to environment and risk changes.

Description

Technical Field

[0001] The present invention relates to the technical field of forestry early warning model construction, and in particular to a forestry harmful organism early warning model construction system and method. Background Art

[0002] With climate change and global warming, the distribution, life cycle and reproduction rate of forest pests may change. New pest populations may invade, such as the gypsy moth, and the spread of existing populations may intensify. The instability factors brought about by climate change make the monitoring and early warning of forest pests particularly important. The background of building a system is mainly to respond to and prevent the threats of forest pests (such as pests, pathogens, etc.) to forest ecosystems. The invasion of forest pests will not only destroy forest resources and lead to a decrease in timber production, but may also cause forest fires, affect biodiversity, etc., and thus have a wide range of impacts on the ecological environment and social economy.

[0003] The prior art has the following defects:

[0004] 1. The existing construction system usually uses a large forest area as the monitoring unit, lacking detailed analysis of the micro-environmental differences within the region. Factors such as topography, climate, and vegetation conditions vary significantly in different areas, but the unified monitoring plan ignores these factors, resulting in the monitoring results not being able to fully reflect the local situation, and it is difficult to conduct accurate monitoring and prevention and control for specific risk areas;

[0005] 2. The monitoring strategies in existing technologies are often based on fixed monitoring intensity and frequency, and cannot be dynamically adjusted according to real-time environmental changes and risk levels. For example, when the pest risk suddenly increases, the built system may not be able to respond quickly and still perform monitoring according to the established plan, resulting in insufficient attention to high-risk areas or waste of resources in low-risk areas.

[0006] Based on this, the present invention proposes a forest pest early warning model construction system and method. By refining the monitoring units, the monitoring intensity can be dynamically adjusted according to the actual conditions of different blocks, thereby improving the pertinence and accuracy of monitoring. By regularly and comprehensively analyzing the real-time data of the blocks, an overall performance index is generated and combined with dynamic judgment rules, the monitoring strategy can be dynamically adjusted to enable the system to adapt to environmental and risk changes. Summary of the invention

[0007] The purpose of the present invention is to provide a forestry pest early warning model construction system and method to solve the shortcomings of the background technology.

[0008] In order to achieve the above object, the present invention provides the following technical solution: a method for constructing a forestry pest early warning model, the construction method comprising the following steps:

[0009] The construction system divides the monitored forest area into multiple blocks, and after building a three-dimensional model for each block, the three-dimensional models of multiple blocks are combined into a monitoring forest area early warning model;

[0010] Generate an initial monitoring intensity for each patch, obtain historical pest data for each patch, the historical pest data includes the amplitude of harmful biodiversity and the amplitude of population dynamic change trend, substitute the amplitude of harmful biodiversity and the amplitude of population dynamic change trend into the evaluation model, the evaluation model outputs an impact factor for each patch, and re-corrects the initial monitoring intensity of each patch based on the impact factor;

[0011] After regularly conducting comprehensive analysis of the real-time data of all patches, the overall performance index of the monitored forest area is obtained, and the overall performance index is combined with dynamic judgment rules to analyze whether it is necessary to optimize and adjust the current monitoring intensity of all patches. After dynamically generating personalized plans for each patch based on the analysis results, the personalized plans are mapped to the monitored forest area early warning model for display.

[0012] In a preferred embodiment, after obtaining the historical pest data of each patch, the pest data is substituted into the evaluation model, and the evaluation model outputs the impact factor for each patch, including the following steps:

[0013] The historical pest data of each patch is obtained. The historical pest data includes the amplitude of harmful biodiversity and the amplitude of population dynamic change trend. The amplitude of harmful biodiversity and the amplitude of population dynamic change trend are substituted into the evaluation model. The evaluation model expression is:

[0014] , where is the impact factor, is the magnitude of harmful biodiversity, is the amplitude of population dynamics change trend, , are the adjustment coefficients of the amplitude of harmful biodiversity and the amplitude of population dynamic change trend, respectively, and , Both are greater than 0.

[0015] In a preferred embodiment, re-correcting the initial monitoring intensity of each block by the influencing factor comprises the following steps:

[0016] After obtaining the impact factor of each patch, the initial monitoring intensity of each patch is corrected by the impact factor. The expression is: , where is the initial data collection frequency, is the corrected data collection frequency, is the impact factor.

[0017] In a preferred embodiment, after regularly analyzing the real-time data of all patches, the overall performance index of the monitored forest area is obtained, including the following steps:

[0018] Regularly obtain real-time data of the block, the real-time data includes vegetation anomaly assignment and pest monitoring frequency, normalize the vegetation anomaly assignment and pest monitoring frequency, map the value range of the vegetation anomaly assignment and pest monitoring frequency to [0,1], obtain the normalized value of the vegetation anomaly assignment and the normalized value of the pest monitoring frequency, and sum the normalized value of the vegetation anomaly assignment and the normalized value of the pest monitoring frequency to obtain the performance factor of the block;

[0019] The vegetation coverage rate of each block is obtained, the vegetation coverage rates of all blocks are summed up to obtain the total coverage rate, the vegetation coverage rate is divided by the total coverage rate to obtain the weight coefficient of the block, and the performance factors of all blocks are weighted and calculated to obtain the overall performance index of the monitored forest area.

[0020] In a preferred embodiment, the overall performance index is combined with the dynamic judgment rule to analyze whether it is necessary to optimize and adjust the current monitoring intensity of all blocks, including the following steps:

[0021] The overall performance index is substituted into the dynamic judgment rule. The rule logic is: compare the obtained overall performance index with the preset performance threshold. The performance threshold is used to determine whether the monitored forest area needs to be managed, that is, dynamically adjust the monitoring intensity of each block. If the overall performance index is greater than or equal to the performance threshold, it is determined that the monitored forest area needs to be managed. If the overall performance index is less than the performance threshold, it is determined that the monitored forest area does not need to be managed.

[0022] In a preferred embodiment, after dynamically generating a personalized solution for each block based on the analysis results, the personalized solution is mapped to the monitoring forest area early warning model display, including the following steps:

[0023] Compare the performance factor of the block with the first gradient threshold and the second gradient threshold, the first gradient threshold is used to determine whether it is necessary to reduce the monitoring intensity of the block, and the second gradient threshold is used to determine whether it is necessary to increase the monitoring intensity of the block;

[0024] If the performance factor of the patch is less than the first gradient threshold, it is determined that the monitoring intensity of the patch needs to be reduced. The adjustment algorithm is: , where is the data collection frequency after secondary adjustment, is the corrected data collection frequency, is the performance factor;

[0025] If the performance factor of the block is greater than or equal to the first gradient threshold, and the performance factor is less than or equal to the second gradient threshold, it is determined that there is no need to adjust the monitoring intensity of the block;

[0026] If the performance factor of the patch is greater than the second gradient threshold, it is determined that the monitoring intensity of the patch needs to be increased, and the adjustment algorithm is: , where is the data collection frequency after secondary adjustment, is the corrected data collection frequency, For performance factors.

[0027] In a preferred embodiment, the logic for obtaining the vegetation anomaly assignment is: collect remote sensing images of patches through drones, identify the normal color and abnormal color of vegetation in the area through existing image recognition technology, divide the remote sensing image into several 1×1 grids, obtain the number of abnormal color grids, and then divide the number of abnormal color grids by the total number of grids in the remote sensing image to obtain the vegetation anomaly assignment.

[0028] In a preferred embodiment, the calculation logic of the harmful biodiversity amplitude is: based on the database, the number of harmful species, the growth rate of harmful species, and the reduction rate of harmful species in the block are obtained, and the harmful biodiversity amplitude is calculated and obtained, and the expression is: , where is the number of pest species, Indicates at time The growth rate of pest species at a given moment, Indicates at time Pest species reduction rate at each moment, Indicates the monitoring duration. , , are the information gains, and .

[0029] In a preferred embodiment, the calculation logic of the amplitude of the population dynamic change trend is: based on the database, the amount of harmful organisms migrating out of the current block is obtained, and when harmful organisms in the current block are migrating out, the amount of harmful organisms migrating in other blocks around the block is obtained, the amount of harmful organisms migrating out of the current block and the amount of harmful organisms migrating in other blocks are summed and divided by two to obtain the migrating out impact index, the amount of harmful organisms migrating in the current block is obtained, and when harmful organisms in the current block are migrating in, the amount of harmful organisms migrating out of other blocks around the block is obtained, the amount of harmful organisms migrating in the current block and the amount of harmful organisms migrating out of other blocks are summed and divided by two to obtain the migrating in impact index;

[0030] After obtaining the out-migration impact index and the in-migration impact index of multiple time points of the current block, calculate the mean of the out-migration impact index and the standard deviation of the out-migration impact index, calculate the mean of the in-migration impact index and the standard deviation of the in-migration impact index, divide the mean of the out-migration impact index by the standard deviation of the out-migration impact index to obtain the coefficient of variation of the out-migration impact, and divide the mean of the in-migration impact index by the standard deviation of the in-migration impact index to obtain the coefficient of variation of the in-migration impact;

[0031] The coefficient of variation of outbound impact and the coefficient of variation of inbound impact are normalized so that their value ranges are mapped to [0,1]. The normalized coefficient of variation of outbound impact and the coefficient of variation of inbound impact are summed to obtain the amplitude of population dynamic change trend. The amplitude of population dynamic change trend fully considers the dynamic change trend of pests in the current area in time and space.

[0032] A forestry pest early warning model construction system, including a model construction module, a monitoring intensity correction module, and a dynamic optimization module;

[0033] Model building module: divide the monitored forest area into multiple blocks, build a three-dimensional model for each block, and then combine the three-dimensional models of multiple blocks into a monitored forest area early warning model;

[0034] Monitoring intensity correction module: Generates initial monitoring intensity for each patch, obtains historical pest data for each patch, substitutes the pest data into the evaluation model, the evaluation model outputs an impact factor for each patch, and re-corrects the initial monitoring intensity of each patch based on the impact factor;

[0035] Dynamic optimization module: After regularly conducting comprehensive analysis of the real-time data of all patches, the overall performance index of the monitored forest area is obtained, and the overall performance index is combined with dynamic judgment rules to analyze whether it is necessary to optimize and adjust the current monitoring intensity of all patches. After dynamically generating personalized plans for each patch based on the analysis results, the personalized plans are mapped to the monitored forest area early warning model for display.

[0036] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0037] After acquiring the historical pest data of each patch, the present invention substitutes the pest data into the evaluation model, and the evaluation model outputs the impact factor for each patch, and re-corrects the initial monitoring intensity of each patch through the impact factor. After regularly conducting a comprehensive analysis of the real-time data of all patches, the overall performance index of the monitored forest area is obtained, and the overall performance index is combined with dynamic judgment rules to analyze whether it is necessary to optimize and adjust the current monitoring intensity of all patches. After dynamically generating a personalized plan for each patch based on the analysis results, the personalized plan is mapped to the monitoring forest area early warning model for display. The construction system can dynamically adjust the monitoring intensity according to the actual situation of different patches by refining the monitoring units, improve the pertinence and accuracy of monitoring, and generate an overall performance index by regularly comprehensively analyzing the real-time data of patches and combining it with dynamic judgment rules to dynamically adjust the monitoring strategy, so that the system can adapt to environmental and risk changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0039] Figure 1 The present invention is a flow chart of the method. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] Example 1: Please refer to Figure 1 As shown, the forest pest early warning model construction method described in this embodiment includes the following steps:

[0042] The construction system divides the monitored forest area into multiple blocks, and after establishing a three-dimensional model for each block, the three-dimensional models of multiple blocks are combined into a monitoring forest area early warning model, and an initial monitoring intensity is generated for each block. After obtaining the historical pest data of each block, the pest data is substituted into the evaluation model. The evaluation model outputs an impact factor for each block, and the initial monitoring intensity of each block is re-corrected according to the impact factor. After regularly conducting a comprehensive analysis of the real-time data of all blocks, the overall performance index of the monitored forest area is obtained, and the overall performance index is combined with dynamic judgment rules to analyze whether it is necessary to optimize and adjust the current monitoring intensity of all blocks. After dynamically generating a personalized plan for each block based on the analysis results, the personalized plan is mapped to the monitoring forest area early warning model for display.

[0043] This application obtains the historical pest data of each patch, substitutes the pest data into the evaluation model, and the evaluation model outputs the impact factor for each patch. The initial monitoring intensity of each patch is re-corrected by the impact factor. After regular comprehensive analysis of the real-time data of all patches, the overall performance index of the monitored forest area is obtained, and the overall performance index is combined with dynamic judgment rules to analyze whether it is necessary to optimize and adjust the current monitoring intensity of all patches. After dynamically generating a personalized plan for each patch based on the analysis results, the personalized plan is mapped to the monitoring forest area early warning model for display. By refining the monitoring units, the construction system can dynamically adjust the monitoring intensity according to the actual situation of different patches, improve the pertinence and accuracy of monitoring, and generate an overall performance index by regularly comprehensively analyzing the real-time data of the patches and combining it with dynamic judgment rules to dynamically adjust the monitoring strategy, so that the system can adapt to environmental and risk changes.

[0044] Embodiment 2: A construction system divides a monitored forest area into multiple blocks, establishes a three-dimensional model for each block, combines the three-dimensional models of multiple blocks into a monitored forest area early warning model, and generates an initial monitoring intensity for each block, including the following steps:

[0045] Obtain basic data on the topography, vegetation, and meteorology of the monitored forest area, and collect high-definition images and mapping data through remote sensing, drone photography, or ground surveys. Formulate rules for dividing blocks based on the topography, climate conditions, vegetation types, and historical distribution of pests in the forest area. It can be divided by geographical area (such as valleys and slopes). Divide by risk level (high, medium, and low). Divide by management unit (such as administrative boundaries or natural divisions). Use spatial analysis algorithms (such as Voronoi diagrams or clustering algorithms) to automatically divide the forest area to ensure that the divided blocks have a balanced area or specific geographical characteristics.

[0046] Using remote sensing data, LiDAR scanning data and ground survey data, detailed terrain, vegetation structure and environmental characteristics data are generated for each patch. Use professional 3D modeling software (such as ArcGIS 3D Analyst, Pix4D) to build a 3D model of each patch, including the following details: terrain characteristics (slope, aspect, elevation). Vegetation levels (crown height, density). Key environmental elements such as water sources and roads.

[0047] Combine the 3D models of each patch into an overall forest model, and use data fusion algorithms (such as Poisson surface reconstruction or spatial interpolation algorithms) to ensure smooth transition and integrity between patches. Calibrate the combined early warning model to eliminate boundary errors and improve the continuity and authenticity of the model.

[0048] Generate an initial monitoring intensity for each patch. The initial monitoring intensity is the data collection frequency. The initial monitoring intensity of each patch is the same. Map the initial monitoring intensity of the patch to the three-dimensional early warning model, and generate an initial monitoring plan. The plan includes: a monitoring intensity distribution map for each patch. The layout plan of monitoring resources and equipment. The initial monitoring schedule and execution plan. It can achieve precise monitoring intensity allocation for the monitored forest area and provide a basis for subsequent dynamic adjustment and optimization.

[0049] After obtaining the historical pest data of each patch, the pest data is substituted into the evaluation model, and the evaluation model outputs the impact factor for each patch, including the following steps:

[0050] The historical pest data of each patch is obtained. The historical pest data includes the amplitude of harmful biodiversity and the amplitude of population dynamic change trend. The amplitude of harmful biodiversity and the amplitude of population dynamic change trend are substituted into the evaluation model. The evaluation model expression is:

[0051] , where is the impact factor, is the magnitude of harmful biodiversity, is the amplitude of population dynamics change trend, , are the adjustment coefficients of the amplitude of harmful biodiversity and the amplitude of population dynamic change trend, respectively, and , Both are greater than 0.

[0052] In this application:

[0053] The calculation logic of the harmful biodiversity amplitude is: based on the database, the number of harmful species, the growth rate of harmful species, and the reduction rate of harmful species in the block are obtained, and the harmful biodiversity amplitude is calculated. The expression is: , where is the number of pest species, Indicates at time The growth rate of pest species at a given moment, Indicates at time Pest species reduction rate at each moment, Indicates the monitoring duration. , , are the information gains, and .

[0054] The harmful biodiversity amplitude measures the species types and their dynamic changes within the patch. The more species and the more drastic the changes, the higher the ecological complexity of the patch, and more data collection is needed to capture the changes. The larger the harmful biodiversity amplitude, the richer the species in the patch, and the significant rate of species growth or decrease. The monitoring demand increases accordingly, and the data collection frequency (monitoring intensity) needs to be improved.

[0055] The calculation logic of the amplitude of the population dynamic change trend is as follows: based on the database, the outward migration of harmful organisms in the current block is obtained, and when harmful organisms in the current block are migrated out, the inward migration of harmful organisms in other blocks around the block is obtained, the outward migration of harmful organisms in the current block and the inward migration of harmful organisms in other blocks are summed and divided by two to obtain the outward migration impact index; the inward migration of harmful organisms in the current block is obtained, and when harmful organisms in the current block are migrated in, the outward migration of harmful organisms in other blocks around the block is obtained, the inward migration of harmful organisms in the current block and the outward migration of harmful organisms in other blocks are summed and divided by two to obtain the inward migration impact index;

[0056] After obtaining the out-migration impact index and the in-migration impact index of multiple time points of the current block, calculate the mean of the out-migration impact index and the standard deviation of the out-migration impact index, calculate the mean of the in-migration impact index and the standard deviation of the in-migration impact index, divide the mean of the out-migration impact index by the standard deviation of the out-migration impact index to obtain the coefficient of variation of the out-migration impact, and divide the mean of the in-migration impact index by the standard deviation of the in-migration impact index to obtain the coefficient of variation of the in-migration impact;

[0057] In this application, the standard deviation of the out-migration impact index and the standard deviation of the in-migration impact index are obtained through the general standard deviation calculation formula, which is expressed as follows: , where is the parameter standard deviation, Indicates the number of time points, Indicates parameter values, represents the parameter mean.

[0058] The calculation expressions of the coefficient of variation of out-migration effect and the coefficient of variation of in-migration effect are:

[0059] , where is the coefficient of variation of the out-migration effect, is the coefficient of variation of immigration impact, is the mean of the out-migration impact index, is the standard deviation of the out-migration impact index, is the mean value of the immigration impact index, is the standard deviation of the immigration impact index.

[0060] The coefficient of variation of outbound impact and the coefficient of variation of inbound impact are normalized so that their value ranges are mapped to [0,1]. The normalized coefficient of variation of outbound impact and the coefficient of variation of inbound impact are summed to obtain the amplitude of population dynamic change trend. The amplitude of population dynamic change trend fully considers the dynamic change trend of pests in the current area in time and space.

[0061] There is a positive correlation between the monitoring intensity (i.e., the frequency of data collection) and the amplitude of the population dynamic change trend. That is, when the amplitude of the population dynamic change trend increases, the required monitoring intensity will also increase accordingly. When the amplitude of the population dynamic change is large, it indicates that the migration of pests within the block is more complicated, and a higher frequency of monitoring may be required to capture this change. If the migration and dynamic changes of pests in the block are significant, then building a system requires frequent acquisition of higher resolution data to analyze and predict changes in pest behavior.

[0062] When the in-and-out migration of pests in a patch is relatively stable and the fluctuation of the migration impact index is small, the amplitude of the population dynamics trend is low. At this time, the monitoring intensity is maintained at the basic level and the collection frequency will not be too high. If the in-and-out migration of pests in a patch is more drastic, the amplitude of the population dynamics trend is large, indicating that the ecological changes in the patch are more complex and changeable, and the monitoring frequency needs to be significantly increased. When the amplitude of the population dynamics change of a patch is close to the maximum, it means that the pests in the patch have undergone extremely complex migration patterns, which may involve frequent species flow or drastic population changes between different patches. At this time, the monitoring intensity may increase rapidly in an exponential manner to ensure sufficient data collection and change capture.

[0063] The initial monitoring intensity of each block is revised based on the impact factor, including the following steps:

[0064] The larger the impact factor, the more severe the overall situation of pests in the patch, and the more it is necessary to increase the monitoring intensity of the patch. After obtaining the impact factor of each patch, the initial monitoring intensity of each patch is re-corrected by the impact factor. The expression is:

[0065] , where is the initial data collection frequency, is the corrected data collection frequency, is the impact factor.

[0066] Regularly analyze the real-time data of all patches to obtain the overall performance index of the monitored forest area, including the following steps:

[0067] Regularly obtain real-time data of the block, the real-time data includes vegetation anomaly assignment and pest monitoring frequency, normalize the vegetation anomaly assignment and pest monitoring frequency, map the value range of the vegetation anomaly assignment and pest monitoring frequency to [0,1], obtain the normalized value of the vegetation anomaly assignment and the normalized value of the pest monitoring frequency, and sum the normalized value of the vegetation anomaly assignment and the normalized value of the pest monitoring frequency to obtain the performance factor of the block;

[0068] The vegetation coverage rate of each patch is obtained, and the vegetation coverage rates of all patches are summed up to obtain the total coverage rate. The weight coefficient of the patch is obtained by dividing the vegetation coverage rate by the total coverage rate. The overall performance index of the monitored forest area is obtained by weighted calculation of the performance factors of all patches. The expression is:

[0069] , where is the overall performance index, For the The weight coefficient of each patch, For the The performance factor of each piece.

[0070] The logic for obtaining the vegetation anomaly value is as follows: collect remote sensing images of blocks through drones, identify the normal color of vegetation in the area (taking green leaf plants as an example, the normal color is green) and abnormal color (taking green leaf plants as an example, abnormal colors may be yellow (such as withering), gray (such as branches without leaves), etc.) through existing image recognition technology, divide the remote sensing image into several 1×1 grids, obtain the number of abnormal color grids, and then divide the number of abnormal color grids by the total number of grids in the remote sensing image to obtain the vegetation anomaly value;

[0071] The calculation logic of pest monitoring frequency is: obtain the number of pests monitored during the monitoring period, and divide the number of pests by the monitoring duration to obtain the pest monitoring frequency.

[0072] Combine the overall performance index with the dynamic judgment rules to analyze whether it is necessary to optimize and adjust the current monitoring intensity of all blocks, including the following steps:

[0073] The overall performance index is substituted into the dynamic judgment rule. The rule logic is: compare the obtained overall performance index with the preset performance threshold. The performance threshold is used to determine whether the monitored forest area needs to be managed, that is, dynamically adjust the monitoring intensity of each block. If the overall performance index is greater than or equal to the performance threshold, it is determined that the monitored forest area needs to be managed. If the overall performance index is less than the performance threshold, it is determined that the monitored forest area does not need to be managed. By making an overall judgment on the monitored forest area, it is beneficial to improve the management efficiency of the monitored forest area and avoid excessive management.

[0074] After dynamically generating a personalized plan for each patch based on the analysis results, the personalized plan is mapped to the monitoring forest area early warning model display, including the following steps:

[0075] When it is determined that the monitored forest area needs to be managed, the larger the performance factor of the block, the worse the vegetation health of the block, and the more frequent the occurrence of pests (this may be caused by insufficient monitoring efforts), and the smaller the performance factor of the block, the better the vegetation health of the block, and the less frequent the occurrence of pests (in this case, there is no need for excessive monitoring). The personalized solutions for each block are as follows:

[0076] Therefore, the performance factor of the block is compared with the first gradient threshold and the second gradient threshold. The first gradient threshold is used to determine whether the monitoring intensity of the block needs to be reduced, and the second gradient threshold is used to determine whether the monitoring intensity of the block needs to be increased.

[0077] If the performance factor of the patch is less than the first gradient threshold, it is determined that the monitoring intensity of the patch needs to be reduced. The adjustment algorithm is: , where is the data collection frequency after secondary adjustment, is the corrected data collection frequency, is the performance factor;

[0078] If the performance factor of the block is greater than or equal to the first gradient threshold, and the performance factor is less than or equal to the second gradient threshold, it is determined that there is no need to adjust the monitoring intensity of the block;

[0079] If the performance factor of the patch is greater than the second gradient threshold, it is determined that the monitoring intensity of the patch needs to be increased, and the adjustment algorithm is: , where is the data collection frequency after secondary adjustment, is the corrected data collection frequency, is the performance factor;

[0080] Construction of the overall forest area model: Based on the personalized solutions of all patches, the overall monitoring model of the forest area is constructed. Each patch will correspond to a specific monitoring plan, and the early warning model integrates the personalized needs of all patches.

[0081] Establish a mapping relationship between the personalized plan of each patch and the location and characteristics of its corresponding area in the early warning model. Map the patch data, monitoring intensity, monitoring requirements, meteorological information, etc. to the entire forest area model through algorithms (such as spatial data processing, regional division, topological model). The mapped data not only reflects the personalized monitoring strategy, but also shows its relationship with other patches, the affected area, etc.

[0082] Use visualization tools (such as GIS platforms, 3D modeling tools, etc.) to display the forest area early warning model, showing the implementation of personalized plans for each block, monitoring intensity, early warning information, etc. The model needs to be updated in real time to reflect the adjustment of each block's monitoring plan, execution effect, and real-time data changes. For example, the construction system can display information such as monitoring intensity, key monitoring areas, and population changes through heat maps, 3D views, and color block changes. Users can interact with the model to view detailed information on each block, adjust monitoring strategies, view environmental change trends, and respond to emergencies.

[0083] Example 3: Please refer to Figure 1 As shown, a forestry pest early warning model construction system described in this embodiment includes a model construction module, a monitoring intensity correction module, and a dynamic optimization module;

[0084] Model building module: divide the monitored forest area into multiple blocks, and after building a three-dimensional model for each block, combine the three-dimensional models of multiple blocks into a monitoring forest area early warning model, send the monitoring forest area early warning model to the dynamic optimization module, and send the block division result to the monitoring intensity correction module;

[0085] Monitoring intensity correction module: Generates initial monitoring intensity for each patch, obtains historical pest data for each patch, substitutes the pest data into the evaluation model, the evaluation model outputs an impact factor for each patch, and re-corrects the initial monitoring intensity of each patch based on the impact factor. The initial monitoring intensity correction information is sent to the dynamic optimization module;

[0086] Dynamic optimization module: After regularly conducting comprehensive analysis of the real-time data of all patches, the overall performance index of the monitored forest area is obtained, and the overall performance index is combined with dynamic judgment rules to analyze whether it is necessary to optimize and adjust the current monitoring intensity of all patches. After dynamically generating personalized plans for each patch based on the analysis results, the personalized plans are mapped to the monitored forest area early warning model for display.

[0087] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0088] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0089] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for constructing a forestry pest early warning model, characterized in that: The construction method comprises the following steps: The construction system divides the monitored forest area into multiple blocks, and after building a three-dimensional model for each block, the three-dimensional models of multiple blocks are combined into a monitoring forest area early warning model; Generate an initial monitoring intensity for each patch, obtain historical pest data for each patch, the historical pest data includes the amplitude of harmful biodiversity and the amplitude of population dynamic change trend, substitute the amplitude of harmful biodiversity and the amplitude of population dynamic change trend into the evaluation model, the evaluation model outputs an impact factor for each patch, and re-corrects the initial monitoring intensity of each patch based on the impact factor; After regularly conducting comprehensive analysis of the real-time data of all patches, the overall performance index of the monitored forest area is obtained, and the overall performance index is combined with dynamic judgment rules to analyze whether it is necessary to optimize and adjust the current monitoring intensity of all patches. After dynamically generating personalized plans for each patch based on the analysis results, the personalized plans are mapped to the monitored forest area early warning model for display.

2. A forestry pest early warning model construction method according to claim 1, characterized in that: After obtaining the historical pest data of each patch, the pest data is substituted into the evaluation model, and the evaluation model outputs the impact factor for each patch, including the following steps: Substituting the amplitude of harmful biodiversity and the amplitude of population dynamic change trend into the assessment model, the assessment model expression is: , where is the impact factor, is the magnitude of harmful biodiversity, is the amplitude of population dynamics change trend, , are the adjustment coefficients of the amplitude of harmful biodiversity and the amplitude of population dynamic change trend, respectively, and , Both are greater than 0.

3. A forestry pest early warning model construction method according to claim 2, characterized in that: The initial monitoring intensity of each block is revised based on the impact factor, including the following steps: After obtaining the impact factor of each patch, the initial monitoring intensity of each patch is corrected by the impact factor. The expression is: , where is the initial data collection frequency, is the corrected data collection frequency, is the impact factor.

4. A forestry pest early warning model construction method according to claim 3, characterized in that: Regularly analyze the real-time data of all patches to obtain the overall performance index of the monitored forest area, including the following steps: Regularly obtain real-time data of the block, the real-time data includes vegetation anomaly assignment and pest monitoring frequency, normalize the vegetation anomaly assignment and pest monitoring frequency, map the value range of the vegetation anomaly assignment and pest monitoring frequency to [0,1], obtain the normalized value of the vegetation anomaly assignment and the normalized value of the pest monitoring frequency, and sum the normalized value of the vegetation anomaly assignment and the normalized value of the pest monitoring frequency to obtain the performance factor of the block; The vegetation coverage rate of each block is obtained, the vegetation coverage rates of all blocks are summed up to obtain the total coverage rate, the vegetation coverage rate is divided by the total coverage rate to obtain the weight coefficient of the block, and the performance factors of all blocks are weighted and calculated to obtain the overall performance index of the monitored forest area.

5. The method for constructing a forestry pest early warning model according to claim 4, characterized in that: Combine the overall performance index with the dynamic judgment rules to analyze whether it is necessary to optimize and adjust the current monitoring intensity of all blocks, including the following steps: The overall performance index is substituted into the dynamic judgment rule. The rule logic is: compare the obtained overall performance index with the preset performance threshold. The performance threshold is used to determine whether the monitored forest area needs to be managed, that is, dynamically adjust the monitoring intensity of each block. If the overall performance index is greater than or equal to the performance threshold, it is determined that the monitored forest area needs to be managed. If the overall performance index is less than the performance threshold, it is determined that the monitored forest area does not need to be managed.

6. The method for constructing a forestry pest early warning model according to claim 5, characterized in that: After dynamically generating a personalized plan for each patch based on the analysis results, the personalized plan is mapped to the monitoring forest area early warning model display, including the following steps: Compare the performance factor of the block with the first gradient threshold and the second gradient threshold, the first gradient threshold is used to determine whether it is necessary to reduce the monitoring intensity of the block, and the second gradient threshold is used to determine whether it is necessary to increase the monitoring intensity of the block; If the performance factor of the patch is less than the first gradient threshold, it is determined that the monitoring intensity of the patch needs to be reduced. The adjustment algorithm is: , where is the data collection frequency after secondary adjustment, is the corrected data collection frequency, is the performance factor; If the performance factor of the block is greater than or equal to the first gradient threshold, and the performance factor is less than or equal to the second gradient threshold, it is determined that there is no need to adjust the monitoring intensity of the block; If the performance factor of the patch is greater than the second gradient threshold, it is determined that the monitoring intensity of the patch needs to be increased, and the adjustment algorithm is: , where is the data collection frequency after secondary adjustment, is the corrected data collection frequency, For performance factors.

7. The method for constructing a forestry pest early warning model according to claim 6, characterized in that: The logic for obtaining the vegetation anomaly value is as follows: collect remote sensing images of patches through drones, identify the normal color and abnormal color of vegetation in the area through existing image recognition technology, divide the remote sensing image into several 1×1 grids, obtain the number of abnormal color grids, and then divide the number of abnormal color grids by the total number of grids in the remote sensing image to obtain the vegetation anomaly value.

8. The method for constructing a forestry pest early warning model according to claim 7, characterized in that: The calculation logic of the harmful biodiversity amplitude is: based on the database, the number of harmful species, the growth rate of harmful species and the reduction rate of harmful species in the block are obtained, and the harmful biodiversity amplitude is calculated and obtained. The expression is: , where is the number of pest species, Indicates at time The growth rate of pest species at a given moment, Indicates at time Pest species reduction rate at each moment, Indicates the monitoring duration. , , are the information gains, and .

9. The method for constructing a forestry pest early warning model according to claim 8, characterized in that: The calculation logic of the amplitude of the population dynamic change trend is as follows: based on the database, the outflow of harmful organisms in the current block is obtained, and when harmful organisms in the current block are emigrated, the inflow of harmful organisms in other blocks around the block is obtained, the outflow of harmful organisms in the current block and the inflow of harmful organisms in other blocks are summed and divided by two to obtain the outflow impact index, the inflow of harmful organisms in the current block is obtained, and when harmful organisms in the current block are emigrated, the outflow of harmful organisms in other blocks around the block is obtained, the inflow of harmful organisms in the current block and the outflow of harmful organisms in other blocks are summed and divided by two to obtain the inflow impact index; After obtaining the out-migration impact index and the in-migration impact index of multiple time points of the current block, calculate the mean of the out-migration impact index and the standard deviation of the out-migration impact index, calculate the mean of the in-migration impact index and the standard deviation of the in-migration impact index, divide the mean of the out-migration impact index by the standard deviation of the out-migration impact index to obtain the coefficient of variation of the out-migration impact, and divide the mean of the in-migration impact index by the standard deviation of the in-migration impact index to obtain the coefficient of variation of the in-migration impact; The coefficient of variation of outbound impact and the coefficient of variation of inbound impact are normalized so that their value ranges are mapped to [0,1]. The normalized coefficient of variation of outbound impact and the coefficient of variation of inbound impact are summed to obtain the amplitude of population dynamic change trend. The amplitude of population dynamic change trend fully considers the dynamic change trend of pests in the current area in time and space.

10. A forest pest early warning model construction system, used to implement the construction method according to any one of claims 1 to 9, characterized in that: Including model building module, monitoring intensity correction module, and dynamic optimization module; Model building module: divide the monitored forest area into multiple blocks, build a three-dimensional model for each block, and then combine the three-dimensional models of multiple blocks into a monitored forest area early warning model; Monitoring intensity correction module: Generates initial monitoring intensity for each patch, obtains historical pest data for each patch, substitutes the pest data into the evaluation model, the evaluation model outputs an impact factor for each patch, and re-corrects the initial monitoring intensity of each patch based on the impact factor; Dynamic optimization module: After regularly conducting comprehensive analysis of the real-time data of all patches, the overall performance index of the monitored forest area is obtained, and the overall performance index is combined with dynamic judgment rules to analyze whether it is necessary to optimize and adjust the current monitoring intensity of all patches. After dynamically generating personalized plans for each patch based on the analysis results, the personalized plans are mapped to the monitored forest area early warning model for display.

Images