Forestry intelligent pest control system based on multi-sensor cooperation

Abstract

The invention discloses a forestry intelligent pest control system based on multi-sensor cooperation, and relates to the technical field of intelligent pest control. In order to solve the problems of low diagnosis precision and poor cooperation efficiency in the prior art, the invention provides a hypothesis-verification cooperation strategy: a system firstly generates a preliminary diagnosis hypothesis by using macroscopic sensor data, and adaptively regulates and controls a microscopic sensor according to the preliminary diagnosis hypothesis to carry out targeted verification so as to realize high-confidence diagnosis; the system further establishes a disease and insect pest spatio-temporal diffusion risk map for trend prediction, and formulates graded and partitioned precise prevention and control strategies. According to the invention, through prevention and treatment effect evaluation and closed-loop self-optimization learning of the model, intelligent prevention and treatment from passive response to active prediction are realized, and the diagnosis accuracy and the resource utilization efficiency are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent pest control, and in particular to an intelligent forestry pest control system based on multi-sensor collaboration. Background Technology

[0002] Traditional forestry pest and disease control relies on inefficient and delayed manual inspections. While macro-monitoring technologies such as satellite remote sensing and drones have been introduced to improve efficiency, they can only identify abnormal vegetation stress and cannot accurately diagnose specific pest and disease types, resulting in poor diagnostic specificity, high false alarm rates, and the need for extensive manual verification. Existing technologies attempt to integrate information from multiple sensors to improve accuracy, but generally employ simple parallel data acquisition and "stacked" fusion, i.e., indiscriminately collecting and analyzing all data. This approach lacks the collaborative logic for targeted detection based on preliminary diagnoses, leading to high system energy consumption and enormous costs, and failing to establish an efficient and interpretable diagnostic verification chain. Furthermore, existing control decision-making functions are relatively weak, mostly stopping at disease diagnosis and generally lacking the ability to dynamically predict the spatiotemporal spread trends of pests and diseases, thus failing to formulate truly graded and zoned precision control strategies. More importantly, the entire "monitoring-diagnosis-control" process is often an open-loop system, lacking feedback links for control effect evaluation and model self-optimization, making it difficult for the system to achieve self-evolution and continuous learning, and limiting its level of intelligence. Summary of the Invention

[0003] To address the aforementioned problems in existing technologies, the present invention aims to provide an intelligent forestry pest and disease control system based on multi-sensor collaboration, comprising: The multi-dimensional sensing module is configured to collect multimodal sensing information of forestry target areas, including at least first-type sensor data and second-type sensor data; the first-type sensor data is macroscopic data reflecting the physiological state of vegetation, and the second-type sensor data is microscopic data characterizing the direct or indirect features of pests and diseases.

[0004] The central processing module, electrically connected to the multi-dimensional sensing module, is used to: identify abnormal vegetation areas and generate at least one preliminary diagnostic hypothesis based on the data from the first type of sensor and through a preset vegetation anomaly analysis model; the preliminary diagnostic hypothesis includes candidate pest and disease types and corresponding initial confidence levels; adaptively adjust the operating parameters or data acquisition priority of the second type of sensor according to the preliminary diagnostic hypothesis; fuse the adjusted data from the second type of sensor and the data from the first type of sensor to construct a collaborative verification model, verify or correct the preliminary diagnostic hypothesis, and output a final diagnostic conclusion with verified confidence levels.

[0005] The intelligent decision-making and control module is electrically connected to the central processing module and is configured to generate and execute targeted pest and disease control instructions based on the final diagnostic conclusion.

[0006] Preferably, the first type of sensor data is selected from at least one of satellite remote sensing imagery, UAV multispectral / hyperspectral data, and chlorophyll fluorescence data; the second type of sensor data is selected from at least one of high-resolution image data, forest acoustic vibration signals, pest pheromone concentration data, or specific volatile organic compound concentration data.

[0007] Preferably, the central processing module generates a preliminary diagnostic hypothesis by: inputting the first type of sensor data into a pre-trained feature matching knowledge base, which stores the macroscopic data feature spectra of vegetation under different pest and disease stresses; and generating the preliminary diagnostic hypothesis and initial confidence level by calculating the similarity between the input data and the feature spectra in the knowledge base.

[0008] Preferably, the central processing module adaptively adjusts the operating parameters of the second type of sensor according to the preliminary diagnostic hypothesis, including: if the preliminary diagnostic hypothesis is that the pest is a borer, then the acquisition frequency and analysis sensitivity of the acoustic sensor in the forest are increased; if the preliminary diagnostic hypothesis is that the pest is a leaf disease, then the UAV is instructed to conduct a detailed investigation of the target area using high-resolution visible light or thermal imaging.

[0009] Preferably, the collaborative verification model is a dynamic weighted fusion model; when the central processing module performs fusion, it uses the verification results of the second type of sensor data as an adjustment factor to dynamically adjust the weight coefficient of the first type of sensor data in generating the final diagnostic conclusion; when the second type of data supports the preliminary diagnostic hypothesis, it is used to increase the weight of the corresponding first type of data features.

[0010] Preferably, the procedure includes the following steps: Among them, the first type of sensor data is acquired from the forestry target area. The first type of sensor data is macroscopic data reflecting the physiological state of vegetation.

[0011] Specifically, the first type of sensor data is processed to identify abnormal vegetation areas through a preset vegetation anomaly analysis model, and at least one preliminary diagnostic hypothesis is generated. The preliminary diagnostic hypothesis includes candidate pest and disease types and corresponding initial confidence levels.

[0012] In response to the preliminary diagnostic hypothesis, the second type of sensor is adaptively adjusted to collect second type of sensor data, which is microscopic data characterizing the direct or indirect features of pests and diseases.

[0013] Specifically, the second type of sensor data and the first type of sensor data are input into a collaborative verification model, which verifies or corrects the preliminary diagnostic hypothesis to generate a final diagnostic conclusion with verified confidence.

[0014] Based on the final diagnostic conclusion, a prevention and control instruction is generated, and the prevention and control equipment is controlled to execute the prevention and control instruction.

[0015] Preferably, it further includes: based on the final diagnostic conclusion, and combined with the topographic data, meteorological data and biological transmission model of the diagnosed pests and diseases in the target area, establishing a spatiotemporal spread risk map of pests and diseases to predict the spread trend and impact range of pests and diseases in the future.

[0016] Step S5 includes, The spatiotemporal diffusion risk map is automatically divided into multiple functional areas, including at least a core prevention and control area, a prevention buffer zone, and a monitoring and observation area, according to a preset risk value threshold.

[0017] Specifically, for the core prevention and control area and the prevention buffer zone, the priority of prevention and control measures and the precise dosage of pesticides or natural enemies are dynamically calculated based on the specific risk level, vegetation density data and biological models of the diagnosed pests and diseases within the area.

[0018] The method uses topographic data and energy consumption models of prevention and control equipment as constraints, with the optimization objectives of shortest operation time, lowest energy consumption and highest prevention and control coverage rate, to plan an optimal execution path for the prevention and control equipment to traverse the core prevention and control area and the prevention buffer zone.

[0019] Preferably, the collaborative verification model in step S4 adopts a Bayesian inference network, wherein the preliminary diagnostic hypothesis generated by the data of the first type of sensor is used as the prior probability, the features collected by the second type of sensor are used as new evidence, and the confidence level for various pest and disease types is updated by calculating the posterior probability to obtain the final diagnostic conclusion.

[0020] Preferably, the method further includes: after executing the prevention and control command, collecting the first type of sensor data and the second type of sensor data again, evaluating the prevention and control effect, and feeding back the evaluation results for iterative optimization of the vegetation anomaly analysis model and the collaborative verification model.

[0021] Compared to existing technologies, the advantages of this invention are as follows: By constructing an innovative "hypothesis-verification" closed-loop logic, this invention significantly improves the accuracy and efficiency of pest and disease diagnosis while reducing operating costs. Unlike the parallel data fusion of existing technologies, this invention first uses macroscopic data to generate preliminary diagnostic hypotheses, and then uses targeted modulation of microscopic sensors based on these hypotheses for verification. This hierarchical and progressive collaborative model avoids the waste of sensor resources and upgrades diagnosis from vague anomaly detection to high-confidence etiological confirmation, providing a solid foundation for precision prevention and control.

[0022] This invention elevates pest and disease control from passive diagnosis to proactive prediction and dynamic planning, achieving intelligent closed-loop management. After accurate diagnosis, the system can establish a spatiotemporal spread risk map to predict pest and disease spread trends, and automatically generate tiered and zoned precise control strategies and optimal execution paths based on this, ensuring efficient utilization of control resources. More importantly, this invention introduces a feedback mechanism for control effect evaluation and model self-optimization, enabling the system to continuously learn and improve, demonstrating excellent adaptability and long-term application value. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the system module structure of the present invention.

[0024] Figure 2 This is an example flowchart of the pest and disease control method of the present invention.

[0025] Figure 3 This is an exemplary flowchart of the propagation prediction planning method of the present invention. Detailed Implementation

[0026] The present invention will be further described below with reference to specific embodiments.

[0027] like Figure 1 The diagram shown illustrates the module structure of the intelligent forestry pest and disease control system based on multi-sensor collaboration provided in this embodiment, including: The multi-dimensional sensing module is configured to collect multimodal sensing information from forestry target areas, including at least first-type sensor data and second-type sensor data. The first-type sensor data is macroscopic data reflecting the physiological state of vegetation, and the second-type sensor data is microscopic data characterizing the direct or indirect features of pests and diseases.

[0028] The central processing module, electrically connected to the multi-dimensional sensing module, is used to identify abnormal vegetation areas and generate at least one preliminary diagnostic hypothesis based on data from the first type of sensors and a preset vegetation anomaly analysis model. The preliminary diagnostic hypothesis includes candidate pest and disease types and their corresponding initial confidence levels. Based on the preliminary diagnostic hypothesis, the module adaptively adjusts the operating parameters or data acquisition priority of the second type of sensors. It then fuses the adjusted data from the second type of sensors with the data from the first type of sensors to construct a collaborative verification model, which verifies or corrects the preliminary diagnostic hypothesis and outputs a final diagnostic conclusion with verified confidence levels.

[0029] The intelligent decision-making and control module is electrically connected to the central processing module and is configured to generate and execute targeted pest and disease control instructions based on the final diagnostic conclusions.

[0030] It should be noted that, in this embodiment, the multi-dimensional sensing module preferably consists of an airborne unmanned aerial vehicle (UAV) platform, a ground-based fixed sensor network, and a forest-based mobile node. The UAV platform is equipped with hyperspectral and thermal infrared cameras for periodic aerial surveys of macroscopic data; the ground-based sensor network includes a chlorophyll fluorescence detector, a pheromone trap, and a VOC detector; the forest-based mobile node includes a small automated inspection vehicle equipped with a high-resolution camera and acoustic sensors. The different sensors are networked via LoRa, NB-IoT, or 4G / 5G communication methods, and all collected data is timestamped and aggregated to the central processing module.

[0031] For example, in a pine forest monitoring scenario, hyperspectral imagery showed a 40% decrease in NDVI values ​​in the target area. The central processing module generated a preliminary diagnostic hypothesis of "suspected pine wilt disease" with an initial confidence level of 0.61. Subsequently, the module issued instructions to increase the sampling rate of the acoustic sensors from 2kHz to 6kHz to capture the typical activity acoustic signatures of stem-boring pests. Finally, after a co-validation model fused the image and acoustic data, the diagnostic conclusion was "pine wilt disease, confidence level 0.83".

[0032] The first type of sensor data is selected from at least one of satellite remote sensing imagery, UAV multispectral / hyperspectral data, and chlorophyll fluorescence data; the second type of sensor data is selected from at least one of high-resolution image data, forest acoustic vibration signals, pest pheromone concentration data, or specific volatile organic compound concentration data.

[0033] In this embodiment, the acquisition cycle for the first type of macroscopic data is typically 7–10 days. For example, UAV hyperspectral data covers the band 400–1000 nm with a resolution of 0.2 m; satellite remote sensing data uses Sentinel-2 NDVI products (10 m resolution, acquisition interval of 5 days); chlorophyll fluorescence arrays record changes in photosynthetic efficiency at 10-minute intervals. The acquisition cycle for the second type of microscopic data is typically several hours to real-time. For example, forest acoustic sensors capture xylem micro-vibration signals with a bandwidth of 2–8 kHz; VOC sensors sample at 5-minute intervals to detect the concentrations of key volatile organic compounds such as methanol, acetaldehyde, and pinene, with a resolution ≤5 ppb.

[0034] For example, in one monitoring mission, the NDVI value of the UAV in the first type of data decreased to 0.35. Combined with the decreasing trend of chlorophyll fluorescence Fv / Fm (from 0.82 to 0.65), the system generated a preliminary diagnosis of leaf disease. Subsequently, the high-resolution images in the second type of data showed typical spot textures, and the methanol concentration in the VOC concentration increased to 12 ppb, thereby enhancing the diagnostic confidence of "leaf spot disease".

[0035] Example data is shown in the table below:

[0036] For example, in one monitoring mission, the NDVI value of the UAV in the first type of data decreased to 0.35. Combined with the decreasing trend of chlorophyll fluorescence Fv / Fm (from 0.82 to 0.65), the system generated a preliminary diagnosis of leaf disease. Subsequently, the high-resolution images in the second type of data showed typical spot textures, and the methanol concentration in the VOC concentration increased to 12 ppb, thereby enhancing the diagnostic confidence of "leaf spot disease".

[0037] In this embodiment, the central processing module generates a preliminary diagnostic hypothesis by inputting the first type of sensor data into a pre-trained feature matching knowledge base. The knowledge base stores the macroscopic data feature spectrum of vegetation under different pest and disease stresses. By calculating the similarity between the input data and the feature spectrum in the knowledge base, a preliminary diagnostic hypothesis and an initial confidence level are generated.

[0038] For example, when the similarity between the UAV hyperspectral data and the "leaf spot / early stage" template is 0.73 and the similarity between the UAV hyperspectral data and the "rust / mid stage" template is 0.58, the system outputs a preliminary diagnostic hypothesis of "leaf spot" with an initial confidence level of 0.62.

[0039] The central processing module adaptively adjusts the operating parameters of the second type of sensor based on the preliminary diagnostic hypothesis, including: if the preliminary diagnosis hypothesis is that the insects are borers, the acquisition frequency and analysis sensitivity of the acoustic sensors in the forest are increased; if the preliminary diagnosis hypothesis is that the insects are leaf diseases, the UAV is instructed to conduct a detailed investigation of the target area using high-resolution visible light or thermal imaging.

[0040] The collaborative verification model is a dynamic weighted fusion model. During fusion, the central processing module uses the verification results of the second type of sensor data as an adjustment factor to dynamically adjust the weight coefficients of the first type of sensor data in generating the final diagnostic conclusion. When the second type of data supports the preliminary diagnostic hypothesis, it is used to increase the weight of the corresponding first type of data features. For example, when the diagnostic hypothesis is "trunk borer," the sampling rate of the tree trunk acoustic sensor is increased from 2kHz to 6kHz, and a 9dB amplification gain is added to capture pest activity signals in the 2–8kHz range. When the diagnostic hypothesis is "leaf disease," the UAV is instructed to acquire 0.2m resolution RGB images and collect leaf surface temperature distribution data for subsequent heat stress identification.

[0041] For example, in a pine forest inspection mission, the central processing module issued acoustic sensor parameter patches based on the hypothesis of suspected pine wilt disease and completed the detailed data collection of the area within 30 minutes.

[0042] like Figure 2 As shown, this embodiment includes the following steps: Step S1: Obtain the first type of sensor data for the forestry target area. The first type of sensor data is macroscopic data reflecting the physiological state of vegetation.

[0043] Step S2: Process the first type of sensor data, identify abnormal vegetation areas through a preset vegetation anomaly analysis model, and generate at least one preliminary diagnostic hypothesis, which includes candidate pest and disease types and corresponding initial confidence levels.

[0044] Step S3: In response to the preliminary diagnostic hypothesis, the second type of sensor is adaptively adjusted to collect second type of sensor data, which is microscopic data characterizing the direct or indirect features of pests and diseases.

[0045] Step S4: Input the second type of sensor data and the first type of sensor data into a co-validation model. The co-validation model verifies or corrects the preliminary diagnostic hypothesis to generate a final diagnostic conclusion with verified confidence.

[0046] Step S5: Based on the final diagnostic conclusion, generate a prevention and control instruction and control the prevention and control equipment to execute the instruction.

[0047] It should be noted that in one example workflow: Step S1 acquires the UAV hyperspectral image matrix M ∈ ℝ^{1024×1024×64}; Step S2 the model outputs an anomaly area mask, accounting for 12%, generating the diagnostic hypothesis "leaf spot disease" with an initial confidence level of 0.58; Step S3 triggers the UAV to collect 10 high-resolution images and infrared temperature distribution; Step S4 after co-validation, the conclusion "leaf spot disease (moderate)" is obtained with a confidence level of 0.87; Step S5 generates the task script, sends the UAV to perform a 0.3 L / acre dosage spray in the core prevention and control area, and uploads the flight trajectory and dosage record after the operation is completed.

[0048] After step S4, the method further includes: based on the final diagnostic conclusion, and combined with the topographic data, meteorological data and biological transmission model of the diagnosed pests and diseases in the target area, establishing a spatiotemporal spread risk map of pests and diseases to predict the spread trend and impact range of pests and diseases in the future.

[0049] like Figure 3 The diagram shows the sub-steps included in step S5 of this embodiment. Step S501: The spatiotemporal diffusion risk map is automatically divided into multiple functional areas, including at least a core prevention and control area, a prevention buffer zone, and a monitoring and observation area, according to a preset risk value threshold.

[0050] Step S502: For the core prevention and control area and the prevention buffer zone, based on the specific risk level, vegetation density data and biological model of the diagnosed pests and diseases in the area, dynamically calculate the priority of prevention and control measures and the precise dosage of pesticides or natural enemies.

[0051] For example, the division criteria are: core area R ≥ 0.8, buffer area 0.6 ≤ R < 0.8, and observation area 0.4 ≤ R < 0.6. Drug dosage is determined using the formula... ,in Host density is represented by Sev, and disease severity is represented by Sev. The priority score is calculated by weighting the risk value, severity, and wind direction exposure coefficient.

[0052] For example, the dose was calculated to be 0.30 L / acre in core area A1 (host density 0.75, severity 0.6); the dose was 0.18 L / acre in buffer zone B3, and a pheromone trapping zone with a width of 30m was set up.

[0053] Step S503: Using topographic data and the energy consumption model of the prevention and control equipment as constraints, and with the shortest operation time, lowest energy consumption and highest prevention and control coverage as optimization objectives, an optimal execution path is planned for the prevention and control equipment to traverse the core prevention and control area and the prevention buffer zone.

[0054] For example, the objective function for path optimization is: Where T is time, E is energy consumption, and Coverage is coverage rate. Constraints include no-fly zones, detours with slopes greater than 35°, and battery capacity not less than 20%. The solution method is an A* heuristic search combined with genetic algorithm optimization.

[0055] For example, in a drone operation mission, the optimal path length is 3.6km, the estimated time is 18.5 minutes, the coverage rate is 97.8%, and the remaining battery power is 23% after execution.

[0056] In step S4, the collaborative verification model uses a Bayesian inference network. The preliminary diagnostic hypothesis generated from the data of the first type of sensor is used as the prior probability, and the features collected by the second type of sensor are used as new evidence. The confidence level for various types of pests and diseases is updated by calculating the posterior probability to obtain the final diagnostic conclusion.

[0057] After step S5, the process further includes: after executing the prevention and control instructions, collecting data from the first type of sensor and the second type of sensor again, evaluating the prevention and control effect, and feeding back the evaluation results to iteratively optimize the vegetation anomaly analysis model and the collaborative verification model.

[0058] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A forestry intelligent pest and disease control system based on multi-sensor collaboration, characterized in that, include: The multi-dimensional sensing module is configured to collect multi-modal sensing information, including at least data from the first type of sensor and data from the second type of sensor, for forestry target areas. The first type of sensor data is macroscopic data reflecting the physiological state of vegetation, while the second type of sensor data is microscopic data characterizing the direct or indirect features of pests and diseases. The central processing module, electrically connected to the multi-dimensional sensing module, is used to: identify abnormal vegetation areas and generate at least one preliminary diagnostic hypothesis based on the data from the first type of sensors and a preset vegetation anomaly analysis model. The preliminary diagnostic hypothesis includes candidate pest and disease types and corresponding initial confidence levels. Based on the preliminary diagnostic hypothesis, the module adaptively adjusts the operating parameters or data acquisition priority of the second type of sensors. It then fuses the adjusted data from the second type of sensors with the data from the first type of sensors to construct a collaborative verification model, verifies or corrects the preliminary diagnostic hypothesis, and outputs a final diagnostic conclusion with verified confidence levels. The intelligent decision-making and control module is electrically connected to the central processing module and is configured to generate and execute targeted pest and disease control instructions based on the final diagnostic conclusion.

2. The intelligent forestry pest and disease control system based on multi-sensor collaboration according to claim 1, characterized in that: The first type of sensor data is selected from at least one of satellite remote sensing imagery, UAV multispectral / hyperspectral data, and chlorophyll fluorescence data; the second type of sensor data is selected from at least one of high-resolution image data, forest acoustic vibration signals, pest pheromone concentration data, or specific volatile organic compound concentration data.

3. The intelligent forestry pest and disease control system based on multi-sensor collaboration according to claim 1, characterized in that: The central processing module generates a preliminary diagnostic hypothesis by inputting the first type of sensor data into a pre-trained feature matching knowledge base. The knowledge base stores the macroscopic data feature spectra of vegetation under different pest and disease stresses. By calculating the similarity between the input data and the feature spectra in the knowledge base, the preliminary diagnostic hypothesis and initial confidence level are generated.

4. The intelligent forestry pest and disease control system based on multi-sensor collaboration according to claim 1, characterized in that: The central processing module adaptively adjusts the operating parameters of the second type of sensor based on the preliminary diagnostic hypothesis, including: if the preliminary diagnostic hypothesis is that the pest is a stem borer, then the acquisition frequency and analysis sensitivity of the acoustic sensor in the forest are increased; if the preliminary diagnostic hypothesis is that the pest is a leaf disease, then the UAV is instructed to conduct a detailed investigation of the target area using high-resolution visible light or thermal imaging.

5. The intelligent forestry pest and disease control system based on multi-sensor collaboration according to claim 1, characterized in that: The collaborative verification model is a dynamic weighted fusion model. When the central processing module performs fusion, it uses the verification results of the second type of sensor data as an adjustment factor to dynamically adjust the weight coefficient of the first type of sensor data in generating the final diagnostic conclusion. When the second type of data supports the preliminary diagnostic hypothesis, it is used to increase the weight of the corresponding first type of data features.

6. The intelligent forestry pest and disease control system based on multi-sensor collaboration according to claim 1, characterized in that: Implementing a prevention and control method includes the following steps: Step S1: Obtain the first type of sensor data of the forestry target area. The first type of sensor data is macroscopic data reflecting the physiological state of vegetation. Step S2: Process the first type of sensor data, identify abnormal vegetation areas through a preset vegetation anomaly analysis model, and generate at least one preliminary diagnostic hypothesis, which includes candidate pest and disease types and corresponding initial confidence levels. Step S3: In response to the preliminary diagnostic hypothesis, the second type of sensor is adaptively adjusted to collect second type of sensor data, which is microscopic data characterizing the direct or indirect features of pests and diseases. Step S4: Input the second type of sensor data and the first type of sensor data into a collaborative verification model. The collaborative verification model verifies or corrects the preliminary diagnostic hypothesis to generate a final diagnostic conclusion with verified confidence. Step S5: Based on the final diagnostic conclusion, generate a prevention and control instruction and control the prevention and control equipment to execute the prevention and control instruction.

7. The intelligent forestry pest and disease control system based on multi-sensor collaboration according to claim 6, characterized in that: Following step S4, the method further includes: based on the final diagnostic conclusion, and combined with the topographic data, meteorological data, and biological transmission model of the diagnosed pests and diseases in the target area, establishing a spatiotemporal spread risk map of pests and diseases to predict the spread trend and impact range of pests and diseases in the future.

8. The intelligent forestry pest and disease control system based on multi-sensor collaboration according to claim 6, characterized in that: Step S5 includes, Step S501: The spatiotemporal diffusion risk map is automatically divided into multiple functional areas, including at least a core prevention and control area, a prevention buffer zone, and a monitoring and observation area, according to a preset risk value threshold. Step S502: For the core prevention and control area and the prevention buffer zone, based on the specific risk level, vegetation density data and biological model of the diagnosed pests and diseases in the area, dynamically calculate the priority of prevention and control measures and the precise dosage of pesticides or natural enemies. Step S503: Using topographic data and the energy consumption model of the prevention and control equipment as constraints, and with the shortest operation time, lowest energy consumption and highest prevention and control coverage as optimization objectives, an optimal execution path is planned for the prevention and control equipment to traverse the core prevention and control area and the prevention buffer zone.

9. The intelligent forestry pest and disease control system based on multi-sensor collaboration according to claim 6, characterized in that: The collaborative verification model in step S4 uses a Bayesian inference network, where the preliminary diagnostic hypothesis generated from the data of the first type of sensor is used as the prior probability, and the features collected by the second type of sensor are used as new evidence. The confidence level for various types of pests and diseases is updated by calculating the posterior probability to obtain the final diagnostic conclusion.

10. The intelligent forestry pest and disease control system based on multi-sensor collaboration according to claim 6, characterized in that: After step S5, the method further includes: after executing the prevention and control command, collecting the first type of sensor data and the second type of sensor data again, evaluating the prevention and control effect, and feeding back the evaluation results to iteratively optimize the vegetation anomaly analysis model and the collaborative verification model.

Images