Early warning method of forest cultivation diseases and insect pests combined with multi-source data analysis

Abstract

The present application provides a forest cultivation pest and disease early warning method combined with multi-source data analysis, which relates to the field of pest and disease early warning technology, including: determining a list of pest and disease types and a list of pest and disease distribution position weights in a preset forest cultivation area; classifying the pest and disease type list according to the way pests and diseases spread in vegetation; determining the activity area of ​​each type of pest and disease in the first type pest and disease list and the second type pest and disease list; collecting a first monitoring data set, analyzing the internal damage characteristics of the vegetation, and generating a first diffusion situation list; analyzing the surface damage characteristics of the vegetation, and generating a second diffusion situation list; and providing pest and disease early warning based on the first diffusion situation list and the second diffusion situation list. The present application can solve the technical problems existing in the prior art due to the difficulty of multi-sensor data fusion, realize the efficient fusion and accurate analysis of multi-sensor data, and achieve the technical effect of improving the accuracy of pest and disease monitoring.

Description

Technical Field

[0001] The present application relates to the technical field of early warning of plant diseases and insect pests, and in particular to an early warning method for plant diseases and insect pests in forest cultivation combined with multi-source data analysis. Background Art

[0002] With global climate change and the deterioration of the ecological environment, pests and diseases are posing an increasingly serious threat to agriculture, forestry and natural ecosystems. Early detection and accurate prediction of pests and diseases are crucial to preventing their spread and reducing losses.

[0003] At present, although some advanced multi-sensor fusion technologies and artificial intelligence algorithms have been applied to pest and disease monitoring, practical applications still face some technical challenges. Among them, how to effectively fuse data from different sensors and accurately analyze them through intelligent algorithms, and how to achieve multi-level and multi-dimensional monitoring and early warning of different types of pests and diseases are still difficult technical implementation problems. In addition, the diversity and complexity of pests and diseases make it more difficult to accurately predict their spread. The spread mode and damage manifestation of different pests and diseases vary greatly. The accuracy and efficiency of existing technologies in dealing with these complex situations still need to be improved.

[0004] In summary, the existing technology has technical problems such as the difficulty of multi-sensor data fusion and insufficient algorithm processing precision, which leads to insufficient accuracy and real-time performance of pest and disease monitoring and early warning, further affecting the effective prevention and control and management of pests and diseases, and being unable to respond to the spread of pests and diseases and the complex and changeable damage manifestations in a timely and accurate manner. Summary of the invention

[0005] The purpose of this application is to provide a forest cultivation pest and disease early warning method combined with multi-source data analysis, in order to solve the technical problems in the prior art that the accuracy and real-time performance of pest and disease monitoring and early warning are insufficient due to the difficulty of multi-sensor data fusion and insufficient algorithm processing precision, which further affects the effective prevention and control and management of pests and diseases, and cannot respond to the spread of pests and diseases and the complex and changeable damage manifestations in a timely and accurate manner.

[0006] In view of the above problems, the present application provides a forest cultivation pest and disease early warning method combined with multi-source data analysis, including: determining a list of pest and disease types and a list of pest and disease distribution position weights in a preset forest cultivation area; classifying the pest and disease type list according to the spread mode of pests in vegetation, generating a first type of pest and disease list corresponding to a first spread mode and a second type of pest and disease list corresponding to a second spread mode, wherein the first spread mode is a spread mode from the inside to the outside of the vegetation, and the second spread mode is a spread mode from the outside to the inside of the vegetation; based on the pest and disease distribution position weight list, the activity area of ​​each type of pest and disease in the first type of pest and disease list and the second type of pest and disease list is determined, and a first type of monitoring module and a second type of monitoring module are configured; a first monitoring data set is collected by the first type of monitoring module, and the internal damage characteristics of the vegetation are analyzed to generate a first diffusion situation list corresponding to the first type of pest and disease list; a second monitoring data set is collected by the second type of monitoring module, and the surface damage characteristics of the vegetation are analyzed to generate a second diffusion situation list corresponding to the second type of pest and disease list; and pest and disease early warning is performed according to the first diffusion situation list and the second diffusion situation list.

[0007] The technical solution provided in the present application has at least the following technical effects or advantages: by determining a list of pest and disease types and a list of pest and disease distribution position weights in a preset forest cultivation area; classifying the pest and disease type list according to the way pests and disease spread in vegetation, generating a first type of pest and disease list corresponding to a first diffusion mode and a second type of pest and disease list corresponding to a second diffusion mode, wherein the first diffusion mode is a diffusion mode from the inside to the outside of the vegetation, and the second diffusion mode is a diffusion mode from the outside to the inside of the vegetation; based on the pest and disease distribution position weight list, the activity area of ​​each type of pest and disease in the first type of pest and disease list and the second type of pest and disease list is determined, and a first type of monitoring module and a second type of monitoring module are configured. Monitoring module group; collecting a first monitoring data set through the first type of monitoring module, analyzing the internal damage characteristics of vegetation, and generating a first diffusion situation list corresponding to the first type of pests and diseases list; collecting a second monitoring data set through the second type of monitoring module, analyzing the surface damage characteristics of vegetation, and generating a second diffusion situation list corresponding to the second type of pests and diseases list; performing pest and disease early warning according to the first diffusion situation list and the second diffusion situation list, that is, by realizing efficient fusion and precise analysis of multi-sensor data, using intelligent algorithms to perform multi-level and multi-dimensional monitoring and early warning of different types of pests and diseases, so as to achieve the technical effect of improving the accuracy of pest and disease monitoring, enhancing the early warning response speed, and enhancing the accuracy of prevention and control decisions.

[0008] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0010] Figure 1 A flowchart of a forest cultivation pest and disease early warning method combined with multi-source data analysis for this application;

[0011] Figure 2 This is a flow chart of configuring the first type of monitoring module and the second type of monitoring module in the forest cultivation pest early warning method combined with multi-source data analysis in this application. DETAILED DESCRIPTION

[0012] This application provides a forest cultivation pest warning method combined with multi-source data analysis, which solves the existing technical problems that the accuracy and real-time performance of pest monitoring and warning are insufficient due to the difficulty of multi-sensor data fusion and insufficient algorithm processing accuracy, further affecting the effective prevention and control and management of pests and diseases, and the inability to respond to the spread of pests and diseases and complex and changeable damage manifestations in a timely and accurate manner. The efficient fusion and accurate analysis of multi-sensor data is achieved, and intelligent algorithms are used to monitor and warn different types of pests and diseases at multiple levels and dimensions, so as to achieve the technical effect of improving the accuracy of pest monitoring, improving the warning response speed, and enhancing the accuracy of prevention and control decisions.

[0013] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0014] Please refer to the attached Figure 1 This application provides a forest cultivation pest and disease early warning method combined with multi-source data analysis, which specifically includes the following steps:

[0015] Step 1: Determine the list of pest and disease types and the list of pest and disease distribution location weights in the preset forest cultivation area.

[0016] Specifically, the pre-set forest cultivation area refers to the area that has been planned for forest resource management. By collecting and analyzing historical monitoring data, the list of pest and disease types and the list of pest and disease distribution location weights in the pre-set forest cultivation area are determined. The list of pest and disease types refers to a list of all pest and disease types that may affect plants in the area, and the list of pest and disease distribution location weights refers to a list of each pest and disease ranked according to the risk level of occurrence in different locations.

[0017] For example, by analyzing historical pest and disease occurrence data, the probability of occurrence of various pests and diseases in different forest cultivation areas is counted. Among them, the tree age and tree species distribution in different forest cultivation areas are different. The weight list of pest and disease distribution locations is shown in Table 1 below:

[0018] Table 1: Pest and disease distribution location weight list

[0019]

[0020] Step 2: Classify the pest and disease type list according to the way pests and diseases spread in the vegetation, and generate a first type pest and disease list corresponding to the first spread method and a second type pest and disease list corresponding to the second spread method, wherein the first spread method is a spread from the inside to the outside of the vegetation, and the second spread method is a spread from the outside to the inside of the vegetation.

[0021] Specifically, the diffusion mode refers to the path and mode of pest spread, including the first diffusion mode and the second diffusion mode. The first diffusion mode is diffusion from the inside of the vegetation to the outside, and the second diffusion mode is diffusion from the outside of the vegetation to the inside. The first diffusion mode and the second diffusion mode will affect the spread speed and range of pests and diseases. According to the diffusion mode of pests and diseases in the vegetation, the pest and disease type list is classified, and the pests and diseases in the pest and disease type list are classified to generate the first type of pest and disease list corresponding to the diffusion from the inside of the vegetation to the outside and the second type of pest and disease list corresponding to the diffusion from the outside of the vegetation to the inside. For example, the first type of pests and diseases may include trunk rot (such as spruce trunk rot), in which pathogenic fungi infect the internal tissues of the tree trunk and gradually spread outward, causing the bark to crack and the trunk to decay; root diseases caused by fungi affect the water and nutrient absorption of the plant, and the disease gradually spreads to the above-ground part of the root rot (such as pine root rot); and vascular diseases (such as wilt) in which pathogens invade the vascular tissue of the plant through the roots or wounds and then spread to the branches and leaves, causing the plant to wilt and die; the first type of pests and diseases Medium insect pests may include longhorn beetle larvae (such as pine brown longhorn beetle), whose larvae feed under the bark or xylem, gradually destroying the trunk structure and eventually causing the tree to die. After the adult insects emerge, they further spread to other plants. Trunk-boring pests (such as jewel beetles) whose larvae bore into the tree, causing the trunk to deform or crack. As the larvae grow, the pest gradually spreads outward. Nematode diseases (such as pine wood nematode disease) are caused by nematodes parasitizing in the vascular bundles of the trunk, destroying the conductive tissues, and eventually spreading to the entire plant and further spread by vector insects. The diseases of the second type of pests and diseases may include leaf spot diseases (such as pine needle brown spot disease) caused by fungal spores spread through the air, first infecting the leaf surface, then developing into the deep layer of the mesophyll, and eventually causing the leaves to fall off, and powdery mildew (such as locust tree powdery mildew) caused by fungal spores spread through the air, initially affecting the leaf surface, gradually spreading to the inside of the leaf tissue, affecting photosynthesis, and rust diseases (such as pine rust) caused by fungal spores spread by the wind, initially infecting the leaf or bark surface, and then invading the inside of the plant, affecting tree growth; the pests of the second type of pests and diseases may include piercing-sucking pests (such as aphids and psyllids) that attach to the surface of vegetation and feed on sap, then reproduce and spread, causing the leaves to curl and turn yellow, and leaf-eating pests (such as loopers and pine caterpillars) whose larvae feed on leaves or twigs, gradually spreading to the inside of the crown, affecting the photosynthesis of the entire tree, and wood borers (such as the light-shouldered star beetle) whose adults lay eggs on the bark surface, and after the larvae hatch, they gradually bore into the trunk, forming holes and expanding inward.

[0022] Step three: determining the activity area of ​​each type of pests and diseases in the first type of pests and diseases list and the second type of pests and diseases list based on the pests and diseases distribution position weight list, and configuring the first type of monitoring module and the second type of monitoring module.

[0023] Specifically, the activity areas of each type of pests in the first type of pests list and the second type of pests list are determined based on the risk level of different pests in each area in the pest distribution location weight list to determine the most likely area for each pest. For example, some pests may only grow in humid areas, while others prefer warm and dry environments.

[0024] Next, based on the determination of the activity area, select and arrange appropriate monitoring equipment for real-time monitoring, and configure the first type of monitoring module and the second type of monitoring module. The configuration of the first type of monitoring module and the second type of monitoring module are respectively for different types of pests and diseases, and different sensors and equipment may be used. For example, the first type of monitoring module monitors pests that spread from the inside. The pests are inside the trunk, roots or soil, and cannot be monitored from the outside; the second type of monitoring module monitors the spread of pests from the outside to the inside. The pests are active on the external surface of the vegetation. For example, cameras can be set up to monitor the surface damage characteristics of the vegetation and the characteristics of pest activity to ensure that the occurrence of different types of pests and diseases in different areas can be effectively monitored, and monitoring points and sensors can be arranged more efficiently to ensure that the monitoring system can play a role in areas where pests and diseases are most active.

[0025] Step 4: Collect a first monitoring data set through the first type of monitoring module, analyze the internal damage characteristics of the vegetation, and generate a first diffusion situation list corresponding to the first type of pest and disease list.

[0026] Specifically, the first monitoring data set is collected through the first type of monitoring module, including signals provided by sound waves, thermal imaging or gas sensors, to analyze the internal damage characteristics of vegetation, generate a first diffusion situation list corresponding to the first type of pest list, and obtain the diffusion of pests and diseases at different time nodes, reflecting the spread speed and impact range of pests and diseases. Among them, the internal damage characteristics of vegetation refer to the impact of pests and diseases on the internal structure or physiological condition of plants, such as changes in plant tissues.

[0027] Step 5: Collect a second monitoring data set through the second type of monitoring module, analyze the surface damage characteristics of vegetation, and generate a second diffusion situation list corresponding to the second type of pest and disease list.

[0028] Specifically, the second monitoring data set is collected through the second type of monitoring module to analyze the surface damage characteristics of vegetation, obtain the impact of pests and diseases on the plant surface, such as spots and withering on leaves, and generate a second diffusion situation list corresponding to the second type of pest and disease list. The spread of the second type of pest and disease at different time points is recorded, thereby forming a comprehensive description of the spread trend of the pest and disease.

[0029] Step six: Issue pest and disease warning based on the first diffusion situation list and the second diffusion situation list.

[0030] Specifically, by combining the first diffusion trend list and the second diffusion trend list, early warning of pests and diseases is carried out, and the possible impact of pests and diseases in the future is predicted according to the diffusion trend of pests and diseases. If the diffusion trend of a pest and disease significantly intensifies, an early warning is issued to prompt relevant departments to take prevention and control measures, thereby avoiding the large-scale spread of pests and diseases.

[0031] The forest cultivation pest and disease early warning method combined with multi-source data analysis can achieve efficient fusion and precise analysis of multi-sensor data, and use intelligent algorithms to conduct multi-level and multi-dimensional monitoring and early warning of different types of pests and diseases, thereby achieving the technical effect of improving the accuracy of pest and disease monitoring, enhancing the early warning response speed, and enhancing the accuracy of prevention and control decisions.

[0032] Further, if Figure 2 As shown, the present application also includes: extracting the pest distribution position weights corresponding to each type of pest based on the pest distribution position weight list; matching each first type of pest in the first type of pest list with the pest distribution position weight, and determining the first monitoring point density distribution according to the first matching result; matching each second type of pest in the second type of pest list with the pest distribution position weight, and determining the second monitoring point density distribution according to the second matching result; configuring the first type of sensor based on the first monitoring point density distribution to generate the first type of monitoring module, and configuring the second type of sensor based on the second monitoring point density distribution to generate the second type of monitoring module.

[0033] Specifically, the pest distribution location weight list is a list sorted according to the risk level of pests and diseases in different areas. Based on the pest distribution location weight list, the pest distribution location weights corresponding to each type of pests and diseases are extracted, including the weights of each location of any pest in the preset area. The pest distribution location weight is the probability of each type of pest and disease at different locations. The higher the pest distribution location weight, the higher the possibility of pest damage, and the higher the required monitoring point density, and vice versa.

[0034] Through the weighted calculation method, each first type of pest and disease in the first type of pest and disease list is matched with the distribution position weight of each pest and disease, that is, a one-to-one match is performed according to the correspondence between the pest and disease type and the distribution position weight. The result obtained is the first matching result, and then the monitoring point density is configured according to the distribution position weight in the first matching result. The monitoring point density is proportional to the size of the distribution position weight, that is, the larger the distribution position weight, the larger the monitoring point density. For example, the distribution position weight of the longhorn beetle in a certain area is 0.8, and more monitoring points for the longhorn beetle can be set in this area. Specifically, it can be allocated according to the distribution position weight by professional and technical personnel in this field in combination with the actual monitoring equipment resources. The density of monitoring points required for each pest and disease in the first type of pest and disease list corresponding to the area can be determined, that is, the first monitoring point density distribution is determined.

[0035] Next, each second type of pest in the second type of pest list is matched with each pest distribution location weight, that is, each second type of pest is multiplied by the corresponding pest distribution location weight, and each product result is added to obtain a second matching result, and then the monitoring point density required for each pest corresponding area in the second type of pest list can be determined, that is, the second monitoring point density distribution is determined.

[0036] Next, the first type of sensors are configured based on the density distribution of the first monitoring point to generate the first type of monitoring module; the second type of sensors are configured based on the density distribution of the second monitoring point to generate the second type of monitoring module. The configuration of the first type of sensors and the second type of sensors is carried out according to the density of their respective monitoring points, ensuring that more sensors are set up in areas with higher risks for real-time monitoring.

[0037] By extracting the weights of the distribution locations of pests and diseases, the risk level of each pest and disease is analyzed; then the density distribution of monitoring points is determined based on these risk data, and corresponding sensors are configured for different types of pests and diseases.

[0038] Furthermore, the present application also includes: the first category of sensors includes acoustic wave sensors, thermal imaging sensors and gas sensors, and the second category of sensors includes visual sensors.

[0039] Specifically, the first category of sensors includes acoustic sensors, thermal imaging sensors, and gas sensors. The invasion process will produce obvious characteristics, such as the sound of pest activity. Acoustic wave sensors are sensors that detect the presence of pests and diseases by detecting sound waves or vibrations. For example, by detecting changes in the vibration of plant leaves or tree trunks, the activity of pests can be detected.

[0040] The activities of plants or pests and diseases may cause temperature changes, so thermal imaging sensors can help detect areas with abnormal temperatures and further discover pest and disease problems. Among them, the temperature distribution of the aboveground part changes due to abnormal water transport caused by the invasion of the interior or roots of trees. Plants evaporate water through the stomata of the leaves, absorbing heat during the transpiration process, thereby reducing the surface temperature of the leaves. Thermal imaging sensors monitor the health of the environment or vegetation by capturing temperature changes on the surface of objects. For example, when transpiration is vigorous, the surface temperature of the leaves will be lower than the ambient temperature. If the water supply is insufficient (such as drought conditions or invasion of the interior or roots of trees, resulting in abnormal water transport), transpiration is weakened and the surface temperature of the leaves may increase.

[0041] Gas sensors identify pest and disease activity by monitoring changes in the concentration of certain specific gases in the air. For example, pests and diseases may cause plants to release certain specific gases, and gas sensors can capture these changes.

[0042] Then, the second category of sensors includes visual sensors. Visual sensors refer to devices that use image or video capture technology to monitor pests and diseases. For example, a high-definition camera or multispectral camera can take pictures or videos of plants, and use image recognition technology to analyze changes in the appearance of plants, such as spots, discoloration, or traces of pests on leaves, to determine whether there is an infestation of pests and diseases. The advantage of visual sensors is that they can provide intuitive and clear image data, which is easy to analyze by manual or automated systems.

[0043] Through different types of sensors, pests and diseases can be comprehensively monitored from multiple dimensions. By combining these data, it is possible to more accurately determine whether there are pest and disease problems in the area.

[0044] Furthermore, the present application also includes: extracting the first acoustic wave data, the first thermal imaging data and the first gas sensor data corresponding to each first type of pest and disease in the first monitoring data set; filtering the first acoustic wave data, the first thermal imaging data and the first gas sensor data to identify the degree of internal damage to the vegetation, and continuously monitoring and analyzing the changing trend of the degree of internal damage to the vegetation, generating each first diffusion situation corresponding to each first type of pest and disease, and adding them to the first diffusion situation list.

[0045] Specifically, the first acoustic wave data, the first thermal imaging data and the first gas sensor data corresponding to each first type of pest in the first monitoring data set are extracted, that is, the pest monitoring data diffused from the inside.

[0046] Since the first acoustic wave data, the first thermal imaging data and the first gas sensor data may be affected by noise during the collection process, the first acoustic wave data, the first thermal imaging data and the first gas sensor data are filtered, that is, the interference components in the data are removed by technical means to make the data more accurate. After filtering, the degree of internal damage to the vegetation is identified, that is, the damage caused by pests and diseases to the internal structure of the plant, and then the depth of the impact of pests and diseases on the plant is reflected, such as the traces of the disease left on the surface or inside the plant.

[0047] Next, collect and update relevant data on internal damage to vegetation regularly or in real time, analyze the changing trend of the degree of internal damage to vegetation, and track the spread of pests and diseases and changes in the degree of damage. In the process of analyzing the changing trend of the degree of internal damage to vegetation, track the spread of pests and diseases and the degree of damage by constructing a neural network model. Specifically, an acoustic wave analysis model, a thermal imaging analysis model, and a gas analysis model for analyzing the first acoustic wave data, the first thermal imaging data, and the first gas sensor data can be constructed respectively, wherein the outputs of the acoustic wave analysis model, the thermal imaging analysis model, and the gas analysis model are all vegetation internal damage data, such as the degree of damage, the input of the acoustic wave analysis model is the acoustic wave data, the input of the thermal imaging analysis model is the thermal imaging data, and the input of the gas analysis model is the gas sensor data, the acoustic wave analysis model, the thermal imaging analysis model, and the gas analysis model are all existing BP neural network models, and the BP neural network model is trained by collecting temperature difference samples and corresponding internal damage degrees through historical data to construct an acoustic wave analysis model. The BP neural network model is trained by collecting acoustic wave sample data and corresponding internal damage degrees through historical data to construct an acoustic wave analysis model. The BP neural network model is trained by collecting gas sensor sample data and the corresponding internal damage degree through historical data, and the acoustic wave analysis model is constructed. Then, the first acoustic wave data, the first thermal imaging data and the first gas sensor data are input into the acoustic wave analysis model, the thermal imaging analysis model and the gas analysis model respectively, and the damage degrees output by the three models are weighted and calculated to obtain the comprehensive internal damage degree of the vegetation.

[0048] Furthermore, a certain pest may cause small-scale damage to the plant surface in the early stage, but as time goes by, the scope and depth of the damage may increase. By analyzing these changing trends and generating various first diffusion trends corresponding to each first-type pest, the spread speed and impact range of the pest can be predicted, so as to take corresponding prevention and control measures. For example, a curve is drawn with the time node as the horizontal coordinate and the comprehensive degree of internal damage to the vegetation as the vertical coordinate, thereby describing the development and expansion of the first-type pest over time, intuitively displaying the diffusion trend or change trend of the first-type pest, and generating various first diffusion trends, reflecting the changes in damage caused by the first-type pest at different time points. Each first diffusion trend is added to the first diffusion trend list, which is used to summarize and record the diffusion process of the pest. In other words, the first diffusion trend list is a list composed of the diffusion curves of the degree of damage corresponding to each first-type pest.

[0049] Then, each first diffusion situation corresponding to each first type of pest is added to the first diffusion situation list, and a corresponding diffusion situation is generated for each first type of pest, so as to form a record set of pest diffusion trends.

[0050] By filtering and cleaning the acoustic wave, thermal imaging and gas sensor data of different types of pests and diseases, the internal damage of the plants can be identified through these data, and the damage changes can be continuously monitored. By analyzing the trend of the damage degree change, the spread of pests and diseases can be generated and added to the spread situation list, providing comprehensive pest and disease spread information.

[0051] Furthermore, the present application also includes: collecting real-time soil moisture and real-time weather conditions of the preset forest cultivation area; collecting temperature difference samples between the vegetation surface temperature and the ambient temperature with the real-time soil moisture and the real-time weather conditions as constraints; using the temperature difference samples to identify anomalies in the filtered first thermal imaging data to generate a first internal damage degree parameter; constructing sound wave sample data to identify anomalies in the filtered first sound wave data to generate a second internal damage degree parameter; constructing gas sensor sample data to identify anomalies in the filtered first gas sensor data to generate a third internal damage degree parameter; weighting the first internal damage degree parameter, the second internal damage degree parameter and the third internal damage degree parameter to generate a comprehensive internal damage degree parameter caused by each of the first types of pests and diseases.

[0052] Specifically, the preset forest cultivation area refers to a pre-determined monitoring area where forest management or cultivation is ongoing. The real-time soil moisture and weather conditions of the preset forest cultivation area are collected to obtain the soil moisture content and meteorological data such as temperature, humidity, wind speed, precipitation, etc. in the area, providing important environmental parameters for subsequent pest and disease monitoring.

[0053] Next, with real-time soil moisture and real-time weather conditions as constraints, the real-time soil moisture and weather conditions are used as limiting factors to guide the collection work, compare the temperature of the plant surface with the temperature of the surrounding environment, and collect the temperature difference sample of the plant surface temperature and the ambient temperature. Specifically, a historical database is constructed by the historical soil moisture, historical weather conditions, and the corresponding historical temperature difference data of the plant surface and the surrounding environment collected by professional and technical personnel in the past time, that is, the historical database contains multiple data, each of which contains historical soil moisture, historical weather conditions and historical temperature difference data. The real-time soil moisture and real-time weather conditions are the data collected in real time at the current moment. Based on this, the historical temperature difference data corresponding to the historical soil moisture and historical weather conditions that are the same as the real-time soil moisture and real-time weather conditions are screened in the historical database as the temperature difference samples. Among them, the plant itself has a transpiration effect. The plant evaporates water through the stomata of the leaves and absorbs heat during the transpiration process, thereby reducing the surface temperature of the leaves. The temperature difference can reflect the health of the plant, especially in the presence of pests and diseases, the temperature of the plant surface may change abnormally. For example, healthy plants usually exhibit a stable transpiration cooling effect. When affected by disease or water stress, the cooling effect is significantly weakened and the leaf surface temperature increases.

[0054] Then, the temperature difference samples are used to perform abnormal identification on the filtered first thermal imaging data, identify abnormal conditions in the thermal imaging data, and generate a first internal damage degree parameter. If the temperature difference is obvious, it may indicate that the plant is damaged or has pests and diseases, and then the degree of damage caused by pests and diseases inside the plant is obtained.

[0055] Next, the acoustic wave sample data is constructed to identify abnormalities in the filtered first acoustic wave data, and abnormal fluctuations in the sound data can be found to generate a second internal damage degree parameter, which reflects the degree of damage caused by pests and diseases inside the plant. The acoustic wave sample data is established by analyzing the vibration signals generated inside or on the surface of the plant.

[0056] Then, if the plant is affected by pests and diseases, it may release specific gases, such as volatile organic compounds. Constructing gas sensing sample data to perform abnormal identification on the filtered first gas sensing data can detect abnormal fluctuations in the gas sensor data, thereby discovering the damage of pests and diseases to the plant, and generating a third internal damage degree parameter, which is used to measure the abnormal changes in the gas inside the plant and its degree of damage. Among them, the gas sensing sample data is constructed by monitoring the gas changes in the air around the plant.

[0057] Finally, the first internal damage degree parameter, the second internal damage degree parameter and the third internal damage degree parameter are weighted according to the coefficient of variation method to generate the comprehensive internal damage degree parameter caused by each first type of pests and diseases. Weighting refers to the weighted average based on the influence or importance of each damage degree parameter to generate a comprehensive damage parameter. By weighting the damage data from three different sources (thermal imaging, acoustic wave and gas sensor data), the comprehensive evaluation results obtained reflect the overall internal damage of the first type of pests and diseases to the plant.

[0058] The damage caused by pests and diseases to plants is comprehensively assessed by combining multiple data sources. Finally, a comprehensive damage parameter is obtained by weighted calculation, which can accurately assess the damage caused by pests and diseases to plants under different environmental conditions and provide data support for pest and disease control.

[0059] Furthermore, the present application also includes: continuous monitoring to generate multiple comprehensive internal damage degree parameters for multiple time nodes corresponding to each first type of pest and disease; using the time node as the horizontal axis and the comprehensive internal damage degree parameter as the vertical axis, drawing a curve based on the multiple comprehensive internal damage degree parameters to generate the first diffusion trends.

[0060] Specifically, the activities of pests and diseases are continuously observed and data collected over a period of time to generate multiple comprehensive internal damage degree parameters for multiple time nodes corresponding to each first type of pest and disease. The activities of each pest and disease will have different degrees of impact at different time points. At multiple time nodes, the degree of damage is quantified through multiple parameters such as sound waves, thermal imaging, and gas sensors to obtain the overall degree of damage to plants caused by pest and disease attacks. For example, over a period of one month, a certain pest may gradually increase its impact on plants, resulting in different degrees of damage at different time nodes.

[0061] Next, with the time node as the horizontal axis and the comprehensive internal damage degree parameter as the vertical axis, a curve is drawn based on multiple comprehensive internal damage degree parameters, and the data at different time points can be presented in the form of a chart to generate each first diffusion trend, that is, the damage development of pests and diseases at each time node. The trend of the curve can reflect the spread speed and degree of damage of pests and diseases. For example, the damage of pests and diseases is relatively light at the beginning, but as time goes by, the curve may gradually rise, indicating that the damage of pests and diseases is gradually increasing.

[0062] By continuously monitoring the activities of pests and diseases and recording their internal damage degree parameters at multiple time points, a pest and disease spread curve can be drawn, thereby intuitively showing the spread trend of pests and diseases at different time points and helping to predict their future development trends.

[0063] Furthermore, the present application also includes: extracting the monitoring images in the second monitoring data set, and inputting the pre-trained surface damage feature analysis model for analysis, generating each comprehensive external damage degree parameter corresponding to each second type of pest and disease in the second type of pest and disease list, wherein the surface damage feature analysis model is constructed based on the training of the sample invasion image data corresponding to the second type of pest and disease list; continuous monitoring, generating multiple comprehensive external damage degree parameters for multiple time nodes corresponding to each second type of pest and disease; using the time node as the horizontal axis and the comprehensive external damage degree parameter as the vertical axis, drawing curves based on the multiple comprehensive external damage degree parameters, generating each second diffusion situation, and adding the second diffusion situation list.

[0064] Specifically, the pre-trained surface damage feature analysis model refers to a model that has been trained through historical sample images, which can identify and classify the features of plant surface damage. Among them, the surface damage feature analysis model is an existing convolutional neural network model, whose input is a vegetation surface image, and the output is the features of plant surface damage, including the area, depth, and morphology of the damage. Specifically, the convolutional neural network model is trained with historical sample images and identification information marked with the area, depth, and morphology of the damage to obtain a surface damage feature analysis model. The training of the convolutional neural network model is a common technical means for those skilled in the art, so it will not be expanded here. The monitoring images in the second monitoring data set are extracted, including images of plant surfaces damaged by pests and diseases. The monitoring images are input into the pre-trained surface damage feature analysis model for analysis, and each comprehensive external damage degree parameter corresponding to each second type of pest and disease in the second type of pest and disease list is generated to indicate the degree of damage caused by the pest and disease to the plant surface, such as information including the area, depth, morphology, etc. of the damage.

[0065] Then, the second type of pests and diseases are monitored regularly or in real time over a period of time in order to capture the changes in surface damage during the development of pests and diseases at different time points, and generate multiple comprehensive external damage degree parameters for multiple time points corresponding to each second type of pests and diseases, reflecting the degree of damage to the plant surface by pests and diseases at different time points, providing dynamic and time-varying information for subsequent analysis.

[0066] Next, with the time node as the horizontal axis and the comprehensive external damage degree parameter as the vertical axis, a curve is drawn based on multiple comprehensive external damage degree parameters to describe the development and expansion of pests and diseases over time. The spread trend or change situation of pests and diseases can be intuitively displayed, and each second diffusion situation is generated, reflecting the damage changes caused by pests and diseases at different time points. Each second diffusion situation is added to the second diffusion situation list to summarize and record the spread process of pests and diseases.

[0067] The second type of pests and diseases is monitored over a long period of time through image data, and the parameters of the degree of damage caused by pests and diseases to the plant surface are extracted, and the spread curve of pests and diseases is drawn based on these data. Through monitoring at multiple time points, a curve can be formed to represent the spread trend of pests and diseases, and the curve can be added to the spread trend list.

[0068] Furthermore, the present application also includes: collecting vegetation distribution information and regional geographic location information within the preset forest cultivation area; mining pest and disease types based on the vegetation distribution information and the regional geographic location information to generate a pest and disease type list; mining historical monitoring data for each pest and disease type in the pest and disease type list, analyzing the susceptibility of various vegetation characteristics based on historical monitoring data, and generating a pest and disease distribution location weight list in combination with the vegetation distribution information.

[0069] Specifically, vegetation distribution information and regional geographic location information within the preset forest cultivation area are collected through sensors, satellite images or other tools. The preset forest cultivation area refers to a pre-set area used for forest management or research, such as a specific area used in forest ecological restoration, disease control and other projects. Vegetation distribution information refers to the distribution location and density of different types of plants in the area, and each plant is susceptible to different pests and diseases. Regional geographic location information includes location-related environmental information such as the area's geographic coordinates, altitude, and climatic conditions. Different geographical locations have different climates, which have an impact on pests and diseases.

[0070] Next, based on the vegetation distribution information and regional geographic location information, pest and disease type mining is carried out to identify the types of pests and diseases that may cause harm to different types of plants and generate a list of pest and disease types, that is, a list of all pest and disease types related to plants in a specific area is generated.

[0071] Then, historical monitoring data mining is performed for each pest type in the pest type list, such as past pest outbreak records, meteorological data, etc. The susceptibility of various vegetation characteristics is analyzed based on historical monitoring data, and the susceptibility of different plant species to different pests and diseases in history is analyzed. Susceptibility refers to the susceptibility of plants to specific pests and diseases, which is related to factors such as the physiological characteristics and disease resistance of plants. Various vegetation characteristics include vegetation type, height, growth stage, etc. The susceptibility of pests and diseases is combined with the distribution location of vegetation to generate a weight list of pest and disease distribution locations, reflecting the risk level of different regions when encountering pests and diseases.

[0072] By mining the types of pests and diseases and analyzing the susceptibility of plants in combination with historical monitoring data, a weight list is finally generated to indicate the risk level of pests and diseases faced by different areas. For example, if a certain area has a high density of vegetation and has been susceptible to certain pests and diseases in history, the weight of the distribution location of the pests and diseases in the area will be higher, which means that the area needs more prevention and control measures.

[0073] In summary, the forest cultivation pest and disease early warning method combined with multi-source data analysis provided by the present application has the following technical effects: by determining a list of pest and disease types and a list of pest and disease distribution position weights in a preset forest cultivation area; classifying the pest and disease type list according to the way pests and diseases spread in vegetation, generating a first type of pest and disease list corresponding to a first diffusion mode and a second type of pest and disease list corresponding to a second diffusion mode, wherein the first diffusion mode is a diffusion mode from the inside to the outside of the vegetation, and the second diffusion mode is a diffusion mode from the outside to the inside of the vegetation; based on the pest and disease distribution position weight list, the activity areas of each type of pest and disease in the first type of pest and disease list and the second type of pest and disease list are determined, and a first type of monitoring model is configured. group and a second type of monitoring module; collecting a first monitoring data set through the first type of monitoring module, analyzing the internal damage characteristics of vegetation, and generating a first diffusion situation list corresponding to the first type of pests and diseases list; collecting a second monitoring data set through the second type of monitoring module, analyzing the surface damage characteristics of vegetation, and generating a second diffusion situation list corresponding to the second type of pests and diseases list; performing pest and disease early warning according to the first diffusion situation list and the second diffusion situation list, that is, by realizing efficient fusion and precise analysis of multi-sensor data, using intelligent algorithms to perform multi-level and multi-dimensional monitoring and early warning of different types of pests and diseases, so as to achieve the technical effect of improving the accuracy of pest and disease monitoring, enhancing the early warning response speed, and enhancing the accuracy of prevention and control decisions.

[0074] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0075] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. A forest cultivation pest and disease early warning method combining multi-source data analysis, characterized in that: include: Determine the list of pest and disease types and the list of pest and disease distribution weights in the pre-set forest cultivation area; Classify the pest type list according to the spreading mode of pests in vegetation, and generate a first type pest list corresponding to a first spreading mode and a second type pest list corresponding to a second spreading mode, wherein the first spreading mode is a spreading mode from inside to outside of the vegetation, and the second spreading mode is a spreading mode from outside to inside of the vegetation; Determine the activity area of ​​each type of pests and diseases in the first type of pests and diseases list and the second type of pests and diseases list based on the pests and diseases distribution position weight list, and configure a first type of monitoring module and a second type of monitoring module; Collecting a first monitoring data set through the first type of monitoring module, analyzing the internal damage characteristics of vegetation, and generating a first diffusion situation list corresponding to the first type of pest and disease list; Collecting a second monitoring data set through the second type of monitoring module, analyzing the vegetation surface damage characteristics, and generating a second diffusion situation list corresponding to the second type of pest and disease list; Providing early warning of pests and diseases according to the first diffusion situation list and the second diffusion situation list; Determine the list of pest and disease types and the weighted list of pest and disease distribution locations in the pre-set forest cultivation area, including: Collecting vegetation distribution information and regional geographical location information within the preset forest cultivation area; Mining pest and disease types based on the vegetation distribution information and the regional geographical location information to generate a pest and disease type list; Historical monitoring data mining is performed on each pest type in the pest type list, the susceptibility of various vegetation characteristics is analyzed based on the historical monitoring data, and a pest distribution location weight list is generated in combination with the vegetation distribution information.

2. The forest cultivation pest and disease early warning method combined with multi-source data analysis as claimed in claim 1, characterized in that: Determining the activity area of ​​each type of pests and diseases in the first type of pests and diseases list and the second type of pests and diseases list based on the pests and diseases distribution position weight list, and configuring a first type of monitoring module and a second type of monitoring module, including: Extracting the pest distribution location weights corresponding to each type of pest based on the pest distribution location weight list; Matching each first type of pests and diseases in the first type of pests and diseases list with the weight of the distribution position of each pest and disease, and determining the density distribution of the first monitoring point according to the first matching result; Matching each second type of pests and diseases in the second type of pests and diseases list with the weight of the distribution position of each pest and disease, and determining the density distribution of the second monitoring point according to the second matching result; Based on the first monitoring point density distribution, a first type of sensor is configured to generate the first type of monitoring module, and based on the second monitoring point density distribution, a second type of sensor is configured to generate the second type of monitoring module.

3. The forest cultivation pest and disease early warning method combined with multi-source data analysis as claimed in claim 2, characterized in that: The first type of sensors include acoustic wave sensors, thermal imaging sensors and gas sensors, and the second type of sensors include visual sensors.

4. The forest cultivation pest and disease early warning method combined with multi-source data analysis as claimed in claim 1, characterized in that: The first type of monitoring module collects a first monitoring data set, analyzes the internal damage characteristics of vegetation, and generates a first diffusion situation list corresponding to the first type of pest list, including: Extracting first acoustic wave data, first thermal imaging data, and first gas sensor data corresponding to each first type of pest in the first monitoring data set; After filtering the first acoustic wave data, the first thermal imaging data and the first gas sensor data, the degree of internal damage to the vegetation is identified, and continuous monitoring is performed to analyze the changing trend of the degree of internal damage to the vegetation, and each first diffusion situation corresponding to each first type of pest and disease is generated and added to the first diffusion situation list.

5. The forest cultivation pest and disease early warning method combined with multi-source data analysis as claimed in claim 4, characterized in that: The first acoustic wave data, the first thermal imaging data, and the first gas sensor data are filtered to identify the degree of damage inside the vegetation, including: Collecting real-time soil moisture and real-time weather conditions in the preset forest cultivation area; Taking the real-time soil moisture and the real-time weather conditions as constraints, collecting temperature difference samples between vegetation surface temperature and ambient temperature; Using the temperature difference sample to perform abnormality recognition on the filtered first thermal imaging data to generate a first internal damage degree parameter; Constructing acoustic wave sample data to perform abnormality recognition on the filtered first acoustic wave data and generate a second internal damage degree parameter; Constructing gas sensor sample data to perform abnormality recognition on the first gas sensor data after filtering, and generating a third internal damage degree parameter; The first internal damage degree parameter, the second internal damage degree parameter and the third internal damage degree parameter are weighted to generate comprehensive internal damage degree parameters caused by each of the first types of pests and diseases.

6. The forest cultivation pest early warning method combined with multi-source data analysis as claimed in claim 5, characterized in that: Analyze the changing trend of the degree of damage inside vegetation and generate the first diffusion trends corresponding to each first type of pests and diseases, including: Continuously monitor and generate multiple comprehensive internal damage degree parameters at multiple time points corresponding to each first type of pests and diseases; The time node is used as the horizontal coordinate and the comprehensive internal damage degree parameter is used as the vertical coordinate. Curves are drawn based on the multiple comprehensive internal damage degree parameters to generate the first diffusion situations.

7. The forest cultivation pest and disease early warning method combined with multi-source data analysis as claimed in claim 1, characterized in that: The second type of monitoring module collects a second monitoring data set, analyzes the vegetation surface damage characteristics, and generates a second diffusion situation list corresponding to the second type of pest list, including: Extracting monitoring images from the second monitoring data set and inputting them into a pre-trained surface damage feature analysis model for analysis to generate each comprehensive external damage degree parameter corresponding to each second type of pests and diseases in the second type of pests and diseases list, wherein the surface damage feature analysis model is constructed based on training of sample invasion image data corresponding to the second type of pests and diseases list; Continuously monitor and generate multiple comprehensive external damage degree parameters at multiple time points corresponding to each second type of pests and diseases; With the time node as the horizontal axis and the comprehensive external damage degree parameter as the vertical axis, a curve is drawn based on the multiple comprehensive external damage degree parameters to generate each second diffusion situation, and the second diffusion situation list is added.

Images