Metabee all-generation monitoring system and method

Abstract

The invention relates to the technical field of beehive monitoring, and particularly discloses a beehive all-generation monitoring system and method, and the system comprises a data collection module, a data transmission module, a data storage module, a data analysis module, and a user interface module. According to the method, multi-modal data fusion and deep learning are carried out by using an enhanced AI algorithm, the population number, variety and development stage of the megapes can be accurately identified, the high-precision monitoring can effectively reduce false alarm and missing alarm, the accuracy of the monitoring data is ensured, and by analyzing the environmental data and megape population dynamics in real time, the megapes can be accurately identified. The system can predict the growth trend and the potential outbreak time of a megacebee population in advance, and a scientific basis is provided for taking prevention and control measures in time; the system can dynamically adjust the monitoring frequency, the sensor power and other resource allocation according to the real-time monitoring data and the AI analysis result, unnecessary energy waste is avoided, and the overall efficiency and sustainability of the system are improved.

Description

Technical Field

[0001] The invention belongs to the technical field of sawfly monitoring, and in particular relates to a sawfly all-generation monitoring system and method. Background Art

[0002] In traditional agricultural management, the population monitoring and prevention of sawflies has always been a challenging task. Due to the wide variety of sawfly species, complex life histories, and susceptibility to environmental factors, traditional manual monitoring methods are not only inefficient, but also often have high false alarm and missed alarm rates, making it difficult to meet the needs of modern agricultural precision management. Due to the lack of real-time, comprehensive environmental data and population dynamics analysis, the formulation of prevention and control measures often lags behind the actual changes in sawfly populations, exacerbating the risk of crop damage.

[0003] In this regard, the inventors propose a sawfly all-generation monitoring system and method to solve the above problems. Summary of the invention

[0004] The object of the present invention is to provide a system and method for monitoring all generations of sawflies to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A sawfly all-generation monitoring system, comprising:

[0007] Data acquisition module, where environmental sensors are arranged in the monitoring area to update the environmental conditions in real time, and high-resolution cameras are deployed to continuously collect images and videos of sawflies and their larvae to ensure comprehensive monitoring at different time periods;

[0008] The data transmission module uses wireless transmission technology to transmit the data of sensors and image capture devices to the central control unit. Edge computing devices are deployed near the data acquisition equipment to perform preliminary data processing and filtering, reduce the load of data transmission and improve the response speed. Combined with the deep learning model, the transmitted images are used to identify the sawfly population, analyze its development stage, and integrate it with environmental data to make a comprehensive population dynamics assessment. Dynamic monitoring is carried out based on the assessment results to optimize monitoring efficiency and resource utilization;

[0009] The data storage module uses a data storage server to centrally store all environmental and image data and back up AI analysis results to ensure data security and traceability;

[0010] The data analysis module uses a big data analysis platform to integrate multi-dimensional data, predict the trend of sawfly population changes through machine learning algorithms, and give corresponding management suggestions;

[0011] The user interface module provides users with real-time monitoring data, AI analysis results, and early warning information through the Web platform and mobile applications. It graphically displays the system's working status and monitoring results, and generates customized analysis reports based on user needs, covering the distribution of sawfly populations, development trends, and environmental conditions analysis.

[0012] Preferably, the environmental sensors include temperature sensors, humidity sensors, light intensity sensors, air pressure sensors, soil moisture sensors, wind speed and direction sensors, and noise sensors. The network composed of environmental sensors collects environmental data in real time and provides multi-dimensional information for establishing a more accurate sawfly population dynamics model.

[0013] Preferably, the wireless transmission technology is one or more of LoRa, Zigbee or 5G.

[0014] Preferably, the deep learning model used in the sawfly population identification is a convolutional neural network model CNN, and the specific expression is:

[0015] y=f(W·x+b)

[0016] Where x: data vector of the input image, representing the collected sawfly image;

[0017] W: convolution kernel weight matrix, used to extract features from the image;

[0018] b: bias vector, which helps the model fit more complex patterns;

[0019] f(): activation function, ReLU or Sigmoid, which introduces nonlinear features to enhance the model's expressiveness;

[0020] y: output result, i.e., the feature map after multiple layers of convolution and pooling, and finally the classification result obtained by the fully connected layer, indicating the species or developmental stage of the sawfly;

[0021] By extracting and classifying features from sawfly images using a CNN model, it is possible to accurately identify the species and developmental stages of sawflies, improving the accuracy of species identification and enabling the system to more effectively track the life cycle of sawflies and detect pest outbreaks in a timely manner.

[0022] Preferably, the dynamic monitoring adjusts the data acquisition frequency, sensor power, and camera shooting interval parameters in real time, which is expressed as:

[0023] F(t)=F0·e -λ·R(t)

[0024] Where F(t): data acquisition frequency or sensor power at time t;

[0025] F0: initial data acquisition frequency or sensor power;

[0026] λ: adjustment coefficient, used to control the adjustment speed, obtained based on experience or through training;

[0027] R(t): Population risk assessment value generated at time t, predicted sawfly population growth rate;

[0028] The working parameters of the monitoring system can be dynamically adjusted according to the real-time risk assessment results to optimize resource utilization, reduce unnecessary monitoring costs, and improve the response efficiency of the system. In particular, the monitoring density can be increased in a timely manner when the sawfly population is growing rapidly, and resource consumption can be reduced during low-risk periods.

[0029] Preferably, the machine learning algorithm is time series prediction, expressed as:

[0030]

[0031] in The predicted sawfly population at time t+k in the future;

[0032] y(t-i+1): the sawfly population data at the historical time t-i+1;

[0033] wi: weight coefficient obtained from model training, reflecting the influence of each historical data point on future predictions;

[0034] ∈: error term, used for model calibration;

[0035] Predicting future population trends based on historical monitoring data can help identify population outbreak risks in advance, provide more reliable forecasts of future population dynamics, support early intervention, and reduce the damage of sawfly pests to agriculture.

[0036] A method for monitoring the entire generation of sawflies comprises the following steps:

[0037] S1. Initial data collection: deploy environmental sensors and image capture equipment, start monitoring, collect initial data, input the initial data into the enhanced AI algorithm model, conduct preliminary training, and improve the accuracy of sawfly population identification and development stage judgment;

[0038] S2, real-time data processing, through sensors and cameras, continuously collects environmental data and images of sawfly activity areas, performs preliminary data cleaning and processing at edge computing nodes, filters out redundant data, and optimizes transmission efficiency;

[0039] S3, enhanced AI algorithm analysis, through deep learning, accurately identifies the collected images, determines the developmental stage of sawflies and their larvae, uses multimodal data fusion, and AI algorithms fuse and analyze environmental data with sawfly population data to generate a population dynamics model and predict future development trends;

[0040] S4. Dynamically adjust monitoring strategies. Based on the analysis results of AI algorithms, dynamically adjust monitoring strategies, such as increasing or decreasing data collection frequency, adjusting sensor power, changing camera shooting intervals, and intelligently allocate system resources according to population dynamics and environmental conditions to reduce unnecessary energy consumption and costs and improve the overall efficiency of the monitoring system.

[0041] S5. Early warning and response. When the AI ​​analysis results predict that the sawfly population is about to experience explosive growth or environmental conditions cause abnormal changes, the system automatically generates early warning information when the early warning threshold is triggered, notifies relevant personnel through the user interface, and provides specific management suggestions to reduce the impact of sawfly pests on agriculture.

[0042] S6. Long-term monitoring and model optimization: Continuously monitor the dynamic changes of sawfly populations and feed new data back to the AI ​​model for continuous learning and optimization. Regularly evaluate monitoring strategies and AI algorithm performance, and adjust parameters to ensure that the system can maintain efficient operation in different environments.

[0043] Preferably, the expression of the multimodal data fusion is:

[0044] z=α·g1(x1)+β·g2(x2)+γ·g3(x3)

[0045] x1, x2, x3: represent different types of data input, such as environmental data, image data, and historical monitoring data;

[0046] g1(), g2(), g3(): functions for feature extraction and transformation of different input data, including convolution, aggregation or dimensionality reduction operations;

[0047] α, β, γ: weight parameters, indicating the contribution of different data modalities to the fusion results, usually obtained through training;

[0048] z: the output of multimodal data fusion, representing the population dynamics assessment results after comprehensive analysis;

[0049] The integration of multiple data sources, including environmental conditions, images, and historical data, can improve the comprehensiveness and accuracy of the analysis, help to more accurately predict changes in sawfly populations, enhance the system's comprehensive analysis capabilities, reduce the bias caused by a single data modality, and improve the monitoring system's responsiveness to complex environments.

[0050] Preferably, the setting and triggering expression of the warning threshold is:

[0051]

[0052] The predicted sawfly population size or density at time t, obtained through time series prediction or deep learning models;

[0053] T(t): basic warning threshold at time t, which is set based on historical data, sawfly life cycle model and expert experience, indicating the population threshold at which the system should trigger an early warning under normal circumstances;

[0054] Δ(t): Dynamic adjustment factor, used to adjust the basic warning threshold T(t) according to the current environmental conditions, uncertainty of the prediction model or other external factors. This factor is dynamically adjusted over time or as the environment changes to improve the accuracy of the warning;

[0055] Alert(t): warning signal, the value is 1 or 0, 1 means triggering the warning and notifying relevant personnel to take action; 0 means not triggering the warning;

[0056] Set a reasonable warning threshold. When the predicted population size exceeds the threshold, the system automatically issues a warning to notify relevant personnel to take action, thereby improving the accuracy of the warning, avoiding excessive false alarms or missed alarms, and ensuring that the system can effectively remind users at critical moments.

[0057] Preferably, the management suggestions include taking timely prevention and control measures and adjusting crop planting strategies.

[0058] Compared with the prior art, the present invention has the following beneficial effects:

[0059] (1) The present invention uses an enhanced AI algorithm for multimodal data fusion and deep learning, which can accurately identify the population size, species and developmental stage of sawflies. This high-precision monitoring can effectively reduce false positives and false negatives, ensure the accuracy of monitoring data, and through real-time analysis of environmental data and sawfly population dynamics, the system can predict the growth trend and potential outbreak time of sawfly populations in advance, providing a scientific basis for timely prevention and control measures; the system can dynamically adjust the monitoring frequency, sensor power and other resource allocations according to real-time monitoring data and AI analysis results, avoid unnecessary energy waste, and improve the overall efficiency and sustainability of the system.

[0060] (2) The system of the present invention can adapt to a variety of environmental conditions, including different climates, soil moisture and light conditions, by dynamically adjusting the monitoring strategy, ensuring effective operation in various agricultural ecosystems; continuous learning and optimization, the AI ​​algorithm continuously learns and optimizes during long-term operation, and by continuously updating and feeding back monitoring data, the system can improve itself and make more accurate predictions about future dynamic changes of sawflies.

[0061] (3) When the sawfly population increases or environmental conditions change significantly, the present invention can issue an early warning in a timely manner to reduce the potential harm of sawflies to crops. The accuracy of the early warning triggering is due to the efficient prediction and real-time data processing of the AI ​​algorithm. The early warning system not only provides alarms, but also gives specific management suggestions based on the analysis results, such as when to take chemical control or biological control measures, thereby reducing the losses caused by sawfly outbreaks. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 A block diagram of a sawfly all-generation monitoring system of the present invention;

[0063] Figure 2 It is a block diagram of the environmental sensor of the present invention;

[0064] Figure 3 It is a block diagram of the wireless transmission technology of the present invention;

[0065] Figure 4 A flow chart of a method for monitoring all generations of sawflies according to the present invention;

[0066] Figure 5 It is a block diagram of the analysis report composition of the present invention;

[0067] Figure 6 A block diagram of the management proposal composition of the present invention. DETAILED DESCRIPTION

[0068] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0069] Embodiment 1:

[0070] See also Figures 1 to 6 As shown, a sawfly full-generation monitoring system comprises:

[0071] Data acquisition module, where environmental sensors are arranged in the monitoring area to update the environmental conditions in real time, and high-resolution cameras are deployed to continuously collect images and videos of sawflies and their larvae to ensure comprehensive monitoring at different time periods;

[0072] The data transmission module uses wireless transmission technology to transmit the data of sensors and image capture devices to the central control unit. Edge computing devices are deployed near the data acquisition equipment to perform preliminary data processing and filtering, reduce the load of data transmission and improve the response speed. Combined with the deep learning model, the transmitted images are used to identify the sawfly population, analyze its development stage, and integrate it with environmental data to make a comprehensive population dynamics assessment. Dynamic monitoring is carried out based on the assessment results to optimize monitoring efficiency and resource utilization;

[0073] The data storage module uses a data storage server to centrally store all environmental and image data and back up AI analysis results to ensure data security and traceability;

[0074] The data analysis module uses a big data analysis platform to integrate multi-dimensional data, predict the trend of sawfly population changes through machine learning algorithms, and give corresponding management suggestions;

[0075] The user interface module provides users with real-time monitoring data, AI analysis results, and early warning information through the Web platform and mobile applications. It graphically displays the system's working status and monitoring results, and generates customized analysis reports based on user needs, covering the distribution of sawfly populations, development trends, and environmental conditions analysis.

[0076] Specifically, the environmental sensors include temperature sensors, humidity sensors, light intensity sensors, air pressure sensors, soil moisture sensors, wind speed and direction sensors, and noise sensors. The network of environmental sensors collects environmental data in real time and provides multi-dimensional information for establishing a more accurate sawfly population dynamics model.

[0077] Temperature sensor, used to measure the air or soil temperature in the monitoring area. Temperature is an important factor affecting the life cycle of sawflies. By monitoring temperature, the activity patterns and development speed of sawflies can be analyzed and predicted.

[0078] Humidity sensors are used to measure relative humidity in the air. Air humidity has an important impact on the survival and reproduction of sawflies. Humidity data can help analyze the environmental suitability of sawfly activities.

[0079] Light intensity sensors are used to monitor the intensity of sunlight or light in the environment. Light affects the activity and reproduction of sawflies, especially when it changes during the day and night cycle. Monitoring light intensity helps predict the peak activity periods of sawflies;

[0080] Air pressure sensors, which measure atmospheric pressure. Changes in air pressure may affect sawfly activity patterns and flight capabilities. Combining air pressure data can more accurately predict sawfly behavior.

[0081] Soil moisture sensors are used to measure the moisture content in the soil. Soil moisture affects the health of plants, which in turn affects the food source of sawflies. By monitoring soil moisture, the living conditions of sawflies can be indirectly assessed.

[0082] Wind speed and direction sensors are used to measure wind speed and direction. Wind speed and direction have a direct impact on the flight and spread of sawflies. Monitoring these parameters can help predict the spread path and range of sawfly populations.

[0083] Noise sensors are used to measure environmental noise levels. Environmental noise may affect the behavioral patterns of sawflies. Noise data analysis is used to understand its potential impact on sawfly activities.

[0084] Specifically, the wireless transmission technology is LoRa.

[0085] Specifically, the deep learning model used in sawfly population identification is a convolutional neural network model CNN, and the specific expression is:

[0086] y=f(W·x+b)

[0087] Where x: data vector of the input image, representing the collected sawfly image;

[0088] W: convolution kernel weight matrix, used to extract features from the image;

[0089] b: bias vector, which helps the model fit more complex patterns;

[0090] f(): activation function, ReLU or Sigmoid, which introduces nonlinear features to enhance the model's expressiveness;

[0091] y: output result, i.e., the feature map after multiple layers of convolution and pooling, and finally the classification result obtained by the fully connected layer, indicating the species or developmental stage of the sawfly;

[0092] By extracting and classifying features from sawfly images using a CNN model, it is possible to accurately identify the species and developmental stages of sawflies, improving the accuracy of species identification and enabling the system to more effectively track the life cycle of sawflies and detect pest outbreaks in a timely manner.

[0093] Specifically, dynamic monitoring and real-time adjustment of data collection frequency, sensor power, and camera shooting interval parameters are expressed as:

[0094] F(t)=F0·e -λ·R(t)

[0095] Where F(t): data acquisition frequency or sensor power at time t;

[0096] F0: initial data acquisition frequency or sensor power;

[0097] λ: adjustment coefficient, used to control the adjustment speed, obtained based on experience or through training;

[0098] R(t): Population risk assessment value generated at time t, predicted sawfly population growth rate;

[0099] The working parameters of the monitoring system can be dynamically adjusted according to the real-time risk assessment results to optimize resource utilization, reduce unnecessary monitoring costs, and improve the response efficiency of the system. In particular, the monitoring density can be increased in a timely manner when the sawfly population is growing rapidly, and resource consumption can be reduced during low-risk periods.

[0100] Specifically, the machine learning algorithm is time series prediction, expressed as:

[0101]

[0102] in The predicted sawfly population at time t+k in the future;

[0103] y(t-i+1): the sawfly population data at the historical time t-i+1;

[0104] wi: weight coefficient obtained from model training, reflecting the influence of each historical data point on future predictions;

[0105] ∈: error term, used for model calibration;

[0106] Predicting future population trends based on historical monitoring data can help identify population outbreak risks in advance, provide more reliable forecasts of future population dynamics, support early intervention, and reduce the damage of sawfly pests to agriculture.

[0107] As can be seen above, the use of enhanced AI algorithms for multimodal data fusion and deep learning can accurately identify the population size, species, and developmental stage of sawflies. This high-precision monitoring can effectively reduce false positives and false negatives, ensure the accuracy of monitoring data, and through real-time analysis of environmental data and sawfly population dynamics, the system can predict the growth trend and potential outbreak time of sawfly populations in advance, providing a scientific basis for timely prevention and control measures; the system can dynamically adjust the monitoring frequency, sensor power, and other resource allocations based on real-time monitoring data and AI analysis results, avoid unnecessary energy waste, and improve the overall efficiency and sustainability of the system;

[0108] By dynamically adjusting monitoring strategies, the system can adapt to a variety of environmental conditions, including different climates, soil moisture and light conditions, to ensure effective operation in various agricultural ecosystems; through continuous learning and optimization, the AI ​​algorithm continuously learns and optimizes during long-term operation. By constantly updating and feeding back monitoring data, the system can improve itself and make more accurate predictions about future dynamic changes in sawflies.

[0109] Embodiment 2:

[0110] refer to Figures 1 to 6 As shown, a method for monitoring the whole generation of sawflies comprises the following steps:

[0111] S1. Initial data collection: deploy environmental sensors and image capture equipment, start monitoring, collect initial data, input the initial data into the enhanced AI algorithm model, conduct preliminary training, and improve the accuracy of sawfly population identification and development stage judgment;

[0112] S2, real-time data processing, through sensors and cameras, continuously collects environmental data and images of sawfly activity areas, performs preliminary data cleaning and processing at edge computing nodes, filters out redundant data, and optimizes transmission efficiency;

[0113] S3, enhanced AI algorithm analysis, through deep learning, accurately identifies the collected images, determines the developmental stage of sawflies and their larvae, uses multimodal data fusion, and AI algorithms fuse and analyze environmental data with sawfly population data to generate a population dynamics model and predict future development trends;

[0114] S4. Dynamically adjust monitoring strategies. Based on the analysis results of AI algorithms, dynamically adjust monitoring strategies, such as increasing or decreasing data collection frequency, adjusting sensor power, changing camera shooting intervals, and intelligently allocate system resources according to population dynamics and environmental conditions to reduce unnecessary energy consumption and costs and improve the overall efficiency of the monitoring system.

[0115] S5. Early warning and response. When the AI ​​analysis results predict that the sawfly population is about to experience explosive growth or environmental conditions cause abnormal changes, the system automatically generates early warning information when the early warning threshold is triggered, notifies relevant personnel through the user interface, and provides specific management suggestions, including timely prevention and control measures and adjustment of crop planting strategies to reduce the impact of sawfly pests on agriculture.

[0116] S6. Long-term monitoring and model optimization: Continuously monitor the dynamic changes of sawfly populations and feed new data back to the AI ​​model for continuous learning and optimization. Regularly evaluate monitoring strategies and AI algorithm performance, and adjust parameters to ensure that the system can maintain efficient operation in different environments.

[0117] Specifically, the expression of the multimodal data fusion is:

[0118] z=α·g1(x1)+β·g2(x2)+γ·g3(x3)

[0119] x1, x2, x3: represent different types of data input, such as environmental data, image data, and historical monitoring data;

[0120] g1(), g2(), g3(): functions for feature extraction and transformation of different input data, including convolution, aggregation or dimensionality reduction operations;

[0121] α, β, γ: weight parameters, indicating the contribution of different data modalities to the fusion results, usually obtained through training;

[0122] z: the output of multimodal data fusion, representing the population dynamics assessment results after comprehensive analysis;

[0123] The integration of multiple data sources, including environmental conditions, images, and historical data, can improve the comprehensiveness and accuracy of the analysis, help to more accurately predict changes in sawfly populations, enhance the system's comprehensive analysis capabilities, reduce the bias caused by a single data modality, and improve the monitoring system's responsiveness to complex environments.

[0124] Specifically, the setting and triggering expressions of the warning threshold are:

[0125]

[0126] The predicted sawfly population size or density at time t, obtained through time series prediction or deep learning models;

[0127] T(t): basic warning threshold at time t, which is set based on historical data, sawfly life cycle model and expert experience, indicating the population threshold at which the system should trigger an early warning under normal circumstances;

[0128] Δ(t): Dynamic adjustment factor, used to adjust the basic warning threshold T(t) according to the current environmental conditions, uncertainty of the prediction model or other external factors. This factor is dynamically adjusted over time or as the environment changes to improve the accuracy of the warning;

[0129] Alert(t): warning signal, the value is 1 or 0, 1 means triggering the warning and notifying relevant personnel to take action; 0 means not triggering the warning;

[0130] By adding the adjustment factor Δ(t), the system can flexibly adjust the warning threshold according to real-time environmental conditions or prediction uncertainties, avoiding the problem of false alarms or missed alarms caused by fixed thresholds. The above formula ensures that the warning signal is triggered only when the sawfly population may reach a dangerous level, reducing the false alarm rate of the system and ensuring that users can take appropriate prevention and control measures at critical moments. The warning threshold is adjusted according to environmental changes, which enhances the system's adaptability to different environmental conditions, so that it can operate effectively in a variety of agricultural environments.

[0131] Set a reasonable warning threshold. When the predicted population size exceeds the threshold, the system automatically issues a warning to notify relevant personnel to take action, thereby improving the accuracy of the warning, avoiding excessive false alarms or missed alarms, and ensuring that the system can effectively remind users at critical moments.

[0132] From the above, we can see that this method can issue early warnings in a timely manner when the sawfly population grows or environmental conditions change significantly, thereby reducing the potential harm of sawflies to crops. The accuracy of the early warning triggering is due to the efficient prediction and real-time data processing of the AI ​​algorithm. The early warning system not only provides alarms, but also gives specific management suggestions based on the analysis results, such as when to take chemical or biological control measures, thereby reducing losses caused by sawfly outbreaks.

[0133] Embodiment three:

[0134] This design is applied to large farms or orchards. The system and method can monitor the population dynamics of sawflies in real time. Once it is predicted that the population growth trend may cause harm, the early warning mechanism is immediately triggered to guide farmers to take precise prevention and control measures at the best time, such as spraying biological pesticides or releasing natural enemy insects, effectively reducing the use of pesticides and improving the quality and safety of agricultural products. By intelligently adjusting the monitoring frequency and sensor power, the system can significantly reduce energy consumption and optimize the allocation of agricultural resources. The monitoring intensity is reduced during the inactive period of sawflies, and the monitoring is strengthened during the key growing season to ensure the effective use of resources, thereby effectively addressing the problems of low efficiency of manual monitoring, and the high false alarm and missed alarm rates of manual monitoring, which are difficult to meet the needs of modern agricultural precision management.

[0135] Specifically, by real-time monitoring of environmental data and sawfly population dynamics, combined with advanced prediction models of AI algorithms, the system can predict the growth trend and potential outbreak time of sawfly populations in advance, and provide farmers with timely and scientific prevention and control guidance. At the same time, the system can also intelligently adjust resource allocation such as monitoring frequency and sensor power based on monitoring data and prediction results, optimize energy consumption, and improve the overall efficiency and sustainability of the system.

[0136] Furthermore, a certain agricultural area entered the peak period of sawfly infestation in early May. Through the sawfly monitoring system, the environmental data and historical data collected in real time, as well as the sawfly population predicted by the AI ​​model, were used to determine whether an early warning was needed.

[0137] Time period: May 1 to May 10, 2024;

[0138] Environmental conditions and data collection:

[0139] Temperature: 20℃ to 28℃;

[0140] Humidity: 55% to 75%

[0141] Light intensity: 12000 to 22000 Lux;

[0142] Wind speed: 3 to 5 m / s;

[0143] Rainfall: 0 to 10 mm (most of the time no rainfall);

[0144] Historical data (before May 1):

[0145] Normal sawfly population threshold T(t): The average number of sawflies observed per day is 100 to 150.

[0146] Number of sawflies in the past month: From the beginning of April to the end of April, the sawfly population gradually increased, reaching about 120 per day at the end of April.

[0147] Prediction data

[0148] Through the time series prediction model, the predicted values ​​of sawfly population from May 1 to May 10, 2024 are shown in Table 1:

[0149]

[0150] Table 1

[0151] Analysis Notes:

[0152] Basic threshold T(t);

[0153] Based on historical data and expert experience, the basic threshold is set at 150 sawflies per day. This threshold means that once the predicted number of sawflies exceeds 150, an early warning will be triggered.

[0154] Adjustment factor Δ(t);

[0155] The adjustment factor Δ(t) was appropriately increased or decreased on different dates based on environmental conditions (e.g., higher temperatures and increased sawfly activity), uncertainty in the forecast model, and historical growth trends. For example, on May 5 and May 6, the adjustment factor was set to 15 because the forecast sawfly population continued to rise and the environment was suitable for sawfly activity.

[0156] Total threshold T(t)+Δ(t);

[0157] The total daily threshold is the sum of the basic threshold and the adjustment factor. The system compares the predicted value and the total threshold to determine whether an early warning needs to be triggered.

[0158] Warning trigger:

[0159] From May 3 to May 8, the predicted number of sawflies exceeded the total threshold for the day, and the system triggered an early warning signal (Alert = 1), reminding relevant managers to take countermeasures (such as spraying pesticides or introducing biological control measures).

[0160] System reactions and effects:

[0161] Early prevention: Between May 3 and May 8, the system issued an early warning signal, prompting managers to take preventive measures. This early intervention strategy effectively reduced the damage that sawfly populations could cause to crops.

[0162] Reduce false alarms: When the sawfly population drops below the threshold (such as on May 9 and May 10), the system automatically stops issuing warnings, avoiding unnecessary prevention and control actions and saving resources.

[0163] Implementation effect:

[0164] Precise control: By dynamically adjusting thresholds, the system can accurately identify high-risk periods and issue early warnings, thus avoiding explosive growth in sawfly populations.

[0165] Resource optimization: Reasonable warning threshold settings and flexible adjustment factors help optimize the allocation of prevention and control resources, improve overall prevention and control efficiency, and reduce unnecessary operations. The above examples show how warning threshold settings and trigger formulas can help manage sawfly populations in actual applications and optimize monitoring and prevention strategies through dynamic adjustments.

[0166] From the above, we can see that through accurate monitoring and early warning, and timely implementation of effective sawfly control measures, crop losses can be significantly reduced and crop yields and quality can be improved. Intelligent monitoring and control strategies can help farmers invest resources more reasonably, avoid excessive use of pesticides or other control measures, and reduce production costs.

[0167] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A sawfly all-generation monitoring system, characterized in that: include: Data acquisition module, where environmental sensors are arranged in the monitoring area to update the environmental conditions in real time, and high-resolution cameras are deployed to continuously collect images and videos of sawflies and their larvae to ensure comprehensive monitoring at different time periods; The data transmission module uses wireless transmission technology to transmit the data of sensors and image capture devices to the central control unit. Edge computing devices are deployed near the data acquisition equipment to perform preliminary data processing and filtering, reduce the load of data transmission and improve the response speed. Combined with the deep learning model, the transmitted images are used to identify the sawfly population, analyze its development stage, and integrate it with environmental data to make a comprehensive population dynamics assessment. Dynamic monitoring is carried out based on the assessment results to optimize monitoring efficiency and resource utilization; The data storage module uses a data storage server to centrally store all environmental and image data and back up AI analysis results to ensure data security and traceability; The data analysis module uses a big data analysis platform to integrate multi-dimensional data, predict the trend of sawfly population changes through machine learning algorithms, and give corresponding management suggestions; The user interface module provides users with real-time monitoring data, AI analysis results, and early warning information through the Web platform and mobile applications. It graphically displays the system's working status and monitoring results, and generates customized analysis reports based on user needs, covering the distribution of sawfly populations, development trends, and environmental conditions analysis.

2. The sawfly all-generation monitoring system according to claim 1, characterized in that: The environmental sensors include temperature sensors, humidity sensors, light intensity sensors, air pressure sensors, soil moisture sensors, wind speed and direction sensors, and noise sensors. The network composed of environmental sensors collects environmental data in real time and provides multi-dimensional information for establishing a more accurate sawfly population dynamic model.

3. The sawfly all-generation monitoring system according to claim 1, characterized in that: The wireless transmission technology is one or more of LoRa, Zigbee or 5G.

4. The sawfly all-generation monitoring system according to claim 1, characterized in that: The deep learning model used in the sawfly population identification is a convolutional neural network model CNN, and the specific expression is: y=f(W·x+b) Where x: data vector of the input image, representing the collected sawfly image; W: convolution kernel weight matrix, used to extract features from the image; b: bias vector, which helps the model fit more complex patterns; f(): activation function, ReLU or Sigmoid, which introduces nonlinear features to enhance the model's expressiveness; y: output result, i.e., the feature map after multiple layers of convolution and pooling, and finally the classification result obtained by the fully connected layer, indicating the species or developmental stage of the sawfly; By extracting and classifying features from sawfly images using a CNN model, it is possible to accurately identify the species and developmental stages of sawflies, improving the accuracy of species identification and enabling the system to more effectively track the life cycle of sawflies and detect pest outbreaks in a timely manner.

5. The sawfly all-generation monitoring system according to claim 1, characterized in that: The dynamic monitoring adjusts the data acquisition frequency, sensor power, and camera shooting interval parameters in real time, which is expressed as: F(t)=F0·e -λ·R(t) Where F(t): data acquisition frequency or sensor power at time t; F0: initial data acquisition frequency or sensor power; λ: adjustment coefficient, used to control the adjustment speed, obtained based on experience or through training; R(t): Population risk assessment value generated at time t, predicted sawfly population growth rate; The working parameters of the monitoring system can be dynamically adjusted according to the real-time risk assessment results to optimize resource utilization, reduce unnecessary monitoring costs, and improve the response efficiency of the system. In particular, the monitoring density can be increased in a timely manner when the sawfly population is growing rapidly, and resource consumption can be reduced during low-risk periods.

6. The sawfly all-generation monitoring system according to claim 1, characterized in that: The machine learning algorithm is time series prediction, expressed as: in The predicted sawfly population at time t+k in the future; y(t-i+1): the sawfly population data at the historical time t-i+1; wi: weight coefficient obtained from model training, reflecting the influence of each historical data point on future predictions; ∈: error term, used for model calibration; Predicting future population trends based on historical monitoring data can help identify population outbreak risks in advance, provide more reliable forecasts of future population dynamics, support early intervention, and reduce the damage of sawfly pests to agriculture.

7. A method for monitoring the entire generation of sawflies, characterized in that: The following steps are involved: S1. Initial data collection: deploy environmental sensors and image capture equipment, start monitoring, collect initial data, input the initial data into the enhanced AI algorithm model, conduct preliminary training, and improve the accuracy of sawfly population identification and development stage judgment; S2, real-time data processing, through sensors and cameras, continuously collects environmental data and images of sawfly activity areas, performs preliminary data cleaning and processing at edge computing nodes, filters out redundant data, and optimizes transmission efficiency; S3, enhanced AI algorithm analysis, through deep learning, accurately identifies the collected images, determines the developmental stage of sawflies and their larvae, uses multimodal data fusion, and AI algorithms fuse and analyze environmental data with sawfly population data to generate a population dynamics model and predict future development trends; S4. Dynamically adjust monitoring strategies. Based on the analysis results of AI algorithms, dynamically adjust monitoring strategies, including increasing or decreasing data collection frequency, adjusting sensor power, changing camera shooting intervals, and intelligently allocating system resources according to population dynamics and environmental conditions to reduce energy consumption and costs and improve the overall efficiency of the monitoring system. S5. Early warning and response. When the AI ​​analysis results predict that the sawfly population is about to experience explosive growth or that environmental conditions are causing abnormal changes, triggering the early warning threshold, the system automatically generates early warning information, notifies relevant personnel through the user interface, and provides specific management suggestions to reduce the impact of sawfly pests on agriculture. S6. Long-term monitoring and model optimization: Continuously monitor the dynamic changes of sawfly populations and feed new data back to the AI ​​model for continuous learning and optimization. Regularly evaluate monitoring strategies and AI algorithm performance, and adjust parameters to ensure that the system can maintain efficient operation in different environments.

8. The method for monitoring all generations of sawflies according to claim 7, characterized in that: The expression of the multimodal data fusion is: z=α·g1(x1)+β·g2(x2)+γ·g3(x3) x1, x2, x3: represent different types of data input, including environmental data, image data, and historical monitoring data; g1(), g2(), g3(): functions for feature extraction and transformation of different input data, including convolution, aggregation or dimensionality reduction operations; α, β, γ: weight parameters, indicating the contribution of different data modalities to the fusion results, obtained through training; z: the output of multimodal data fusion, representing the population dynamics assessment results after comprehensive analysis; The integration of multiple data sources, including environmental conditions, images, and historical data, can improve the comprehensiveness and accuracy of the analysis, help to more accurately predict changes in sawfly populations, enhance the system's comprehensive analysis capabilities, reduce the bias caused by a single data modality, and improve the monitoring system's responsiveness to complex environments.

9. The method for monitoring all generations of sawflies according to claim 7, characterized in that: The setting and triggering expressions of the warning threshold are: The predicted sawfly population size or density at time t, obtained through time series prediction or deep learning models; T(t): basic warning threshold at time t, which is set based on historical data, sawfly life cycle model and expert experience, indicating the population threshold at which the system should trigger an early warning under normal circumstances; Δ(t): Dynamic adjustment factor, used to adjust the basic warning threshold T(t) according to the current environmental conditions, uncertainty of the prediction model or other external factors. This factor is dynamically adjusted over time or as the environment changes to improve the accuracy of the warning; Alert(t): warning signal, the value is 1 or 0, 1 means triggering the warning and notifying relevant personnel to take action; 0 means not triggering the warning; Set a reasonable warning threshold. When the predicted population size exceeds the threshold, the system automatically issues a warning to notify relevant personnel to take action, thereby improving the accuracy of the warning, avoiding excessive false alarms or missed alarms, and ensuring that the system can effectively remind users at critical moments.

10. The method for monitoring all generations of sawflies according to claim 7, characterized in that: The management recommendations include taking timely prevention and control measures and adjusting crop planting strategies.

Images