A hive deslagging intelligent early warning method and system based on an internet of things

Abstract

The application discloses a kind of based on thing networking's hive slag intelligent early warning method and system, belong to intelligent monitoring early warning technical field, it includes obtaining the vibration signal and weight data of hive bottom, and the weight data obtained is corrected, generates net slag discharge amount;Obtain environmental monitoring data and adjacent hive slag discharge data, while dynamically adjusting the preset basic slag discharge threshold, generate dynamic slag discharge threshold;When the net slag discharge amount exceeds the dynamic slag discharge threshold, generate primary early warning signal;Response the primary early warning signal triggers video acquisition instruction;Execute the video acquisition instruction to obtain the honeybee activity video of preset time length;The honeybee activity video is analyzed and feature extraction, when identifying honeybee abnormal behavior characteristics, generate secondary early warning signal.The application adopts vibration trigger weighing mechanism and environment self-adaptive dynamic threshold combination, can realize low energy consumption, high precision, traceable bee colony health early warning.

Description

Technical Field

[0001] This invention relates to the field of intelligent monitoring and early warning technology, and in particular to an intelligent early warning method and system for beehive slag discharge based on the Internet of Things. Background Technology

[0002] Currently, monitoring beehive waste disposal is a crucial aspect of bee health management, but traditional methods relying on manual inspections are insufficient for the needs of large-scale beekeeping. In recent years, IoT technology has been applied to beehive monitoring, enabling automated data collection by deploying sensors within the hive.

[0003] Current mainstream solutions use load cells to detect changes in weight at the bottom of the hive, triggering an alarm when the weight increase exceeds a preset threshold. Some improved solutions add temperature and humidity sensors to assist in the judgment, or use cameras to periodically capture images of the hive entrance.

[0004] Single weighing solutions are susceptible to interference from bee swarm activity, wind swaying, etc., leading to false alarms; fixed thresholds cannot adapt to seasonal changes and regional climate differences; independent video surveillance and data monitoring lack linkage, making it impossible to trace the cause of anomalies; and the high power consumption of continuously operating sensors restricts field deployment. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides an intelligent early warning method and system for beehive waste disposal based on the Internet of Things. It combines a vibration-triggered weighing mechanism with an environmentally adaptive dynamic threshold and establishes a closed-loop diagnostic chain that uses data anomaly-driven video verification. This enables low-energy, high-precision, and traceable early warning of bee colony health.

[0006] The above objectives can be achieved through the following approach:

[0007] A smart early warning method and system for beehive waste removal based on the Internet of Things (IoT) includes: acquiring vibration signals and weight data from the bottom of the beehive; activating the weight data acquisition process based on the vibration signals and correcting the obtained weight data to generate a net waste removal amount; acquiring environmental monitoring data and waste removal data from neighboring beehives; dynamically adjusting a preset basic waste removal threshold based on the environmental monitoring data and neighboring beehive waste removal data to generate a dynamic waste removal threshold; performing a comprehensive analysis based on the net waste removal amount and the dynamic waste removal threshold to generate a primary early warning signal; triggering a video acquisition command in response to the primary early warning signal; executing the video acquisition command to acquire a bee activity video of a preset duration; analyzing and extracting features from the bee activity video to identify abnormal bee behavior characteristics; and when abnormal bee behavior characteristics are identified, performing system analysis and generating a secondary early warning signal.

[0008] Optionally, generating the net slag discharge amount includes: detecting whether the intensity of the vibration signal exceeds a preset vibration intensity threshold, and activating the power supply of the weight sensor when it exceeds the threshold; collecting the original weight data output by the activated weight sensor; and performing compensation calculations based on the original weight data to generate the net slag discharge amount.

[0009] Optionally, generating the dynamic slag discharge threshold includes: determining an environmental baseline value based on a preset environmental parameter-slag discharge volume correlation curve; obtaining a dataset of slag discharge from neighboring beehives within a preset geographical range; calculating the statistical average value of the dataset of slag discharge from neighboring beehives; correcting the environmental baseline value based on the difference between the statistical average value and the environmental baseline value, and outputting it as the dynamic slag discharge threshold.

[0010] Optionally, the environmental parameter-slag discharge volume correlation curve includes: collecting historical environmental parameter datasets and corresponding historical slag discharge volume data; grouping and statistically analyzing the historical environmental parameter datasets and the historical slag discharge volume data, and calculating the slag discharge volume distribution characteristic value for each group of data; fitting the slag discharge volume distribution characteristic value and the corresponding relationship curve of the historical environmental parameter dataset to generate the environmental parameter-slag discharge volume correlation curve.

[0011] Optionally, generating the primary warning signal includes: obtaining the rate of change of net slag discharge within a preset time period based on the net slag discharge volume; analyzing and calculating based on the dynamic slag discharge threshold to generate a corresponding dynamic change rate threshold; and generating a primary warning signal based on the comparison result between the net slag discharge volume and the dynamic slag discharge threshold when the rate of change of net slag discharge volume exceeds the dynamic change rate threshold.

[0012] Optionally, the analysis and feature extraction of the bee activity video includes: extracting the bee's flight trajectory data from the bee activity video; calculating the motion feature parameters of the flight trajectory data and identifying the visual feature parameters of the objects attached to the bee's body surface; and combining the motion feature parameters and the visual feature parameters to analyze and extract features from the bee activity video to generate behavior analysis results.

[0013] Optionally, identifying abnormal bee behavior features includes: comparing the motion feature parameters with a preset range of normal flight parameters; detecting whether the visual feature parameters exceed a preset attachment threshold; and marking abnormal bee behavior features when the motion feature parameters exceed the range of normal flight parameters or the visual feature parameters exceed the attachment threshold.

[0014] Optionally, generating a secondary warning signal includes: generating a comprehensive score for abnormal behavior based on the abnormal bee behavior characteristics, combined with the net amount of waste and the dynamic waste threshold; generating a corresponding secondary warning signal level when the comprehensive score for abnormal behavior falls within a preset warning level threshold range; and generating a secondary warning signal by associating the secondary warning signal level with a preset bee colony health status database.

[0015] Optionally, the method further includes: extracting features based on the current environmental monitoring data and the abnormal bee behavior characteristics to generate a comprehensive feature vector; performing feature matching on a preset disease feature database based on the comprehensive feature vector to generate an abnormality cause category; and generating a structured early warning report by combining the generated abnormality cause category.

[0016] Based on the same inventive concept, this invention also provides an IoT-based intelligent early warning system for beehive waste removal. The system includes: a signal acquisition module for acquiring vibration signals and weight data from the bottom of the beehive, activating the weight data acquisition process based on the vibration signals, and correcting the obtained weight data to generate a net waste removal amount; a threshold dynamic adjustment module for acquiring environmental monitoring data and waste removal data from neighboring beehives, dynamically adjusting a preset basic waste removal threshold based on the environmental monitoring data and neighboring beehive waste removal data to generate a dynamic waste removal threshold; a threshold comparison early warning module for performing a comprehensive analysis based on the net waste removal amount and the dynamic waste removal threshold to generate a primary early warning signal; a video triggering module for triggering a video acquisition command in response to the primary early warning signal; a video acquisition module for executing the video acquisition command to acquire a preset duration of bee activity video; a behavior analysis module for analyzing and extracting features from the bee activity video to identify abnormal bee behavior characteristics; and a secondary early warning generation module for performing system analysis and generating a secondary early warning signal when abnormal bee behavior characteristics are identified.

[0017] Compared with the prior art, the present invention has the following advantages:

[0018] 1. This invention uses a coordinated triggering mechanism of vibration signal and weight acquisition to significantly reduce equipment energy consumption while ensuring the accuracy of slag discharge monitoring, solving the power loss problem caused by traditional continuous weighing, and using weight data to verify the authenticity of vibration events, significantly suppressing false alarms caused by environmental interference.

[0019] 2. This invention integrates dual-path dynamic threshold correction technology with environmental parameters and neighboring bee colony data, breaking through the fixed threshold early warning mode, enabling the system to automatically adapt to regional climate fluctuations and differences in bee colony behavior, and improving the reliability of early warning in complex environments.

[0020] 3. This invention establishes a closed-loop verification chain for targeted video evidence collection triggered by data anomalies, deeply coupling the monitoring of ash discharge with the identification of bee behavior characteristics, realizing the leap from simple anomaly alarms to disease cause diagnosis, and providing beekeepers with an operable intervention basis.

[0021] 4. This invention adopts a stability adaptive learning mechanism for multi-day slag discharge sequences to dynamically optimize early warning sensitivity during long-term operation, balance the contradiction between false alarm rate and false alarm rate, and reduce the frequency of invalid manual verification.

[0022] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating an intelligent early warning method for beehive slag discharge based on the Internet of Things, according to an embodiment of the present invention.

[0025] Figure 2 This is a timing diagram of the vibration-triggered weighing mechanism according to an embodiment of the present invention.

[0026] Figure 3 This is a dynamic threshold dual correction diagram according to an embodiment of the present invention.

[0027] Figure 4 This is an early warning chart of the rate of change in slag discharge volume according to an embodiment of the present invention.

[0028] Figure 5 This is a schematic diagram of the structure of an IoT-based intelligent early warning system for beehive slag discharge, according to an embodiment of the present invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] Reference Figure 1 One embodiment of the present invention proposes an intelligent early warning method for beehive slag discharge based on the Internet of Things. It adopts a combination of vibration-triggered weighing mechanism and environmentally adaptive dynamic threshold, and establishes a closed-loop diagnostic chain for data anomaly-driven video verification, which can achieve low-energy consumption, high precision and traceability of bee colony health early warning.

[0031] The method described in this embodiment specifically includes:

[0032] The vibration signal and weight data at the bottom of the beehive are obtained. The weight data acquisition process is activated based on the vibration signal, and the obtained weight data is corrected to generate the net amount of ash discharged.

[0033] Acquire environmental monitoring data and neighboring beehive waste discharge data, and dynamically adjust the preset basic waste discharge threshold based on the environmental monitoring data and neighboring beehive waste discharge data to generate a dynamic waste discharge threshold;

[0034] Based on a comprehensive analysis of the net slag discharge volume and the dynamic slag discharge threshold, a primary early warning signal is generated.

[0035] Specifically, this method first acquires vibration signals from the bottom of the beehive using a vibration sensor installed at the bottom. When the vibration signal reaches a preset activation threshold, a weight sensor is triggered to collect the weight data of the beehive. The collected weight data may contain noise introduced by bee colony activity or environmental factors, thus requiring correction. The correction process uses a baseline calibration method, which records a baseline weight under no-debris conditions and subtracts the baseline weight from the real-time weight data to obtain the net debris amount. Simultaneously, environmental monitoring data such as temperature and humidity are acquired through environmental sensors, and debris data from neighboring beehives is received. Based on this data, a dynamic weighted algorithm is used to adjust the preset basic debris threshold. By long-term monitoring of the debris amount of healthy bee colonies in different seasons and statistically analyzing its distribution range (e.g., median ± standard deviation), an initial threshold is set in conjunction with the experience of beekeeping experts. For example, in high-temperature and high-humidity environments or when the debris amount of neighboring beehives surges, the basic debris threshold is appropriately increased to generate a dynamic debris threshold. Finally, by combining the net ash discharge volume, the rate of change of net ash discharge volume within a preset time period is obtained, and the corresponding dynamic change rate threshold is calculated based on the dynamic ash discharge threshold. When the rate of change of net ash discharge volume exceeds the dynamic change rate threshold, a primary warning signal is generated by combining the comparison results of net ash discharge volume and dynamic ash discharge threshold, indicating that there may be abnormal bee colony activity or pest and disease risks. Figure 2 As shown, the solid black line represents the vibration signal intensity at the bottom of the beehive, the dashed black line represents the response signal after the weight sensor is activated, and the gray dashed line represents the preset vibration intensity threshold. When the vibration signal exceeds the preset signal intensity threshold, the weight sensor is activated. The preset signal intensity threshold is the minimum effective signal value that triggers the weight sensor, set by actual measured signal intensity distribution (such as bee colony activity and wind interference).

[0036] The video acquisition command is triggered in response to the primary warning signal;

[0037] Specifically, once the system generates a primary warning signal, this signal automatically initiates the video acquisition process. The video acquisition command is transmitted via an IoT communication module to cameras deployed around the beehives. Upon receiving the command, the cameras immediately activate and acquire video of bee activity according to the preset shooting duration and resolution. During video acquisition, the system monitors video quality indicators in real time, including image clarity and lighting conditions. If blurry images or insufficient light are detected, the system automatically adjusts the camera focus or activates supplementary lighting. The acquired video data is transmitted wirelessly to a cloud server for storage and analysis. This automated triggering mechanism ensures timely acquisition of bee activity images when abnormal waste discharge occurs, providing a reliable data source for subsequent behavioral analysis.

[0038] Execute the video acquisition command to obtain a video of bee activity of a preset duration;

[0039] Specifically, upon receiving a video acquisition command, the system first reads preset video acquisition parameters, including key indicators such as acquisition duration, resolution, and frame rate. The acquisition duration is set to a fixed value based on the characteristics of bee colony activity to ensure complete recording of bee behavior patterns. After the camera is activated, it begins recording according to the set technical parameters, while simultaneously monitoring ambient light intensity in real time using built-in sensors. If the light is insufficient, it automatically switches to night vision mode. During video acquisition, the system continuously monitors storage space and network transmission status to ensure complete preservation of video data. After acquisition, the system automatically generates a video file of bee activity containing a timestamp and beehive number, and uploads it to the cloud analysis platform via an encrypted channel. This process, through standardized acquisition procedures, ensures the integrity and comparability of video data, providing a standardized data source for subsequent analysis.

[0040] The video of bee activity was analyzed and features were extracted to identify abnormal bee behavior characteristics;

[0041] When abnormal bee behavior is detected, the system performs analysis and generates a secondary warning signal.

[0042] Specifically, this method first captures video of bee activity using a camera and then analyzes each frame of the video using computer vision algorithms. In the feature extraction stage, optical flow is used to calculate the bee's movement trajectory, combined with a convolutional neural network to identify abnormal bee behavior characteristics, such as violent shaking, disordered flight, or prolonged stillness. The movement trajectory is calculated based on the instantaneous speed of a single bee and the average speed of the bee colony; a higher trajectory value indicates a higher probability of an anomaly. When the trajectory value exceeds a preset threshold, the system performs a comprehensive analysis of the abnormal behavior characteristics, including the duration, frequency, and spatial distribution of the behavior. If the comprehensive score meets the warning standard, a secondary warning signal is generated. This method improves recognition accuracy through multi-dimensional feature fusion and avoids misjudgment based on a single indicator.

[0043] Optionally, the net slag discharge volume includes:

[0044] Detect whether the intensity of the vibration signal exceeds a preset vibration intensity threshold; if it does, activate the power supply of the weight sensor.

[0045] Collect the raw weight data output by the activated weight sensor;

[0046] Compensation calculations are performed based on the original weight data to generate the net slag discharge volume.

[0047] Specifically, this method uses a vibration sensor to monitor the vibration signal at the bottom of the beehive in real time and compares the detected vibration signal intensity with a preset vibration intensity threshold. The difference in signal intensity between normal bee colony activities (such as bees entering and leaving, foraging for nectar) and abnormal vibrations (such as predator intrusion) is measured experimentally, and the value with the highest discriminative power is selected as the preset vibration intensity threshold. If the vibration signal intensity exceeds this threshold, the power supply to the weight sensor is activated, causing it to begin collecting raw weight data. The raw weight data may include the weight of the beehive itself, the weight of the bees, and environmental interference, therefore compensation calculations are needed to generate the net amount of waste. The specific method of compensation calculation is to record the baseline weight data of the beehive in a state without waste collection, and subtract the baseline weight data from the real-time collected raw weight data; the difference is the net amount of waste. For calculating the net amount of waste... ,have:

[0048] ,

[0049] in, Raw weight data collected in real time by the weight sensor; The baseline weight data for the beehive under no-debris conditions is obtained by averaging the weight data under stable conditions over a long period of time. This process ensures the accuracy of the net debris discharge volume and eliminates interference from non-debris discharge factors.

[0050] For example, assuming a beehive's preset vibration intensity threshold is 50 units, when the vibration sensor detects a vibration signal intensity of 60 units, the system automatically activates the weight sensor and collects the current raw weight data as 15.8 kg. The system then calls the pre-stored baseline weight data of 15.5 kg and calculates the net waste discharge as 0.3 kg through compensation. The beneficial effect of this verification example is that by dynamically activating the weight sensor and using compensation calculation, misjudgments caused by fluctuations in the beehive's own weight or bee activity are effectively avoided, ensuring the accuracy and reliability of waste discharge detection, while also reducing power consumption and extending the sensor's lifespan.

[0051] Optionally, the generation of the dynamic slag discharge threshold includes:

[0052] The environmental baseline value is determined based on the preset environmental parameter-slag discharge volume correlation curve;

[0053] Obtain a dataset of neighboring beehive waste within a preset geographical range;

[0054] Calculate the statistical average of the neighboring beehive waste disposal dataset;

[0055] The environmental baseline value is corrected based on the difference between the statistical average value and the environmental baseline value, and then output as a dynamic slag discharge threshold.

[0056] Specifically, the system first calls a pre-established correlation curve between environmental parameters and beehive discharge volume. By collecting historical environmental data (temperature and humidity) and corresponding discharge volumes, the system statistically analyzes the distribution of discharge volumes by temperature and humidity intervals (as shown in a box plot). Cubic spline interpolation is used to fit a continuous surface to obtain the correlation curve between environmental parameters and discharge volume. This curve, through historical data analysis, determines the normal discharge volume range of beehives under different temperature and humidity conditions. Based on the currently collected environmental temperature and relative humidity, the system queries the curve to obtain the environmental baseline value. Simultaneously, the system uses the Internet of Things to obtain a dataset of discharge volumes from 10 neighboring beehives within a preset geographical range (a radius of 500 meters). After removing outliers, the statistical average is calculated. This process is then used to calculate the statistical average discharge volume of neighboring beehives. ,have:

[0057] ,

[0058] in, For the first The amount of waste discharged from a neighboring beehive. To ensure sufficient data, environmental baseline values ​​are compared with statistical averages. If the difference exceeds 15%, the weighted average of the two values ​​is used as the dynamic slag discharge threshold. The environmental baseline value has a weight of 60%, and the statistical average has a weight of 40%. The final output is the dynamic slag discharge threshold. For details on calculating the dynamic slag discharge threshold... ,have:

[0059] ,

[0060] in, This is to retrieve the environmental baseline value. This method combines environmental factors and group behavioral characteristics to make the threshold setting more realistic, such as... Figure 3 As shown, the light gray bars represent the theoretical baseline values ​​considering only environmental parameters, the dark gray bars represent the statistical average of the amount of slag discharged from neighboring hives, and the black bars represent the final dynamic threshold, which is 60% of the environmental baseline and 40% of the neighboring data.

[0061] For example, a bee farm collected environmental parameters in spring, with a temperature of 22 degrees Celsius. ℃ With a humidity level of 70%, the baseline value can be obtained by referring to the curve. The average amount of waste discharged from the eight neighboring beehives was... If the difference exceeds 20%, the system automatically calculates a weighted value to determine the dynamic threshold. Environmental curves provide a theoretical basis, while data from neighboring hives reflect the actual situation. This dual reference makes threshold setting both scientific and practical, effectively avoiding misjudgments caused by a single environmental factor and improving the accuracy and adaptability of early warnings.

[0062] Optionally, the environmental parameter-slag discharge volume correlation curve includes:

[0063] Collect historical environmental parameter datasets and corresponding historical slag discharge data;

[0064] The historical environmental parameter dataset and the historical slag discharge data are grouped and statistically analyzed, and the slag discharge distribution characteristic value of each group is calculated.

[0065] Fit the curve showing the relationship between the characteristic value of the slag discharge volume distribution and the historical environmental parameter dataset to generate the correlation curve between the environmental parameters and the slag discharge volume.

[0066] Specifically, the system first collects historical data from at least one complete beekeeping cycle, including daily environmental parameters such as temperature and humidity, as well as corresponding records of beehive waste discharge. Temperature is divided into multiple intervals of 5°C, and humidity is divided into intervals of 10%. Within each temperature and humidity combination interval, the median and interquartile range of waste discharge are calculated. Cubic spline interpolation is used to smooth these discrete statistical points, constructing a continuous environmental parameter-waste discharge correlation surface. For each temperature and humidity combination, the correlation surface outputs a corresponding baseline waste discharge prediction value. The system then calculates the baseline waste discharge prediction value. ,have:

[0067] ,

[0068] in, This represents the correlation function between pre-calibrated environmental parameters and slag discharge volume. This represents the number of basis functions in the temperature dimension. The number of basis functions in the humidity dimension. The spline coefficients are determined by least squares fitting. ， For three B Spline basis functions. After the surface is built, it is stored in the system database for querying during real-time monitoring. This method of grouped statistics and surface fitting can accurately reflect the normal range of bee colony waste discharge under different environmental conditions.

[0069] For example, a bee farm collected environmental monitoring data for the entire year of 2024, dividing the temperature into... , Humidity is divided into equal intervals. , The same interval. At temperature... ,humidity The median amount of slag discharged within the interval was obtained statistically. Interquartile range is The system uses these data points to fit a smooth, correlated surface. When the current environment is detected as... , At humidity, the baseline slag discharge volume is automatically obtained by querying the curved surface. The association model built on large sample historical data is reliable, grouped statistics avoid interference from extreme values, and surface fitting enables continuous querying, making the benchmark setting more accurate and reasonable, and providing a scientific basis for judging abnormal slag discharge.

[0070] Optionally, generating the primary warning signal includes:

[0071] Based on the net slag discharge volume, obtain the rate of change of net slag discharge volume within a preset time period;

[0072] Based on the dynamic slag discharge threshold, an analysis and calculation are performed to generate the corresponding dynamic change rate threshold;

[0073] When the rate of change of net slag discharge exceeds the dynamic rate of change threshold, a primary warning signal is generated by combining the comparison results of net slag discharge and dynamic slag discharge threshold.

[0074] Specifically, the system first calculates the trend of net slag discharge over the past 3 hours, and then uses a linear regression method to obtain the hourly rate of change of net slag discharge. ,have:

[0075] ,

[0076] in, and This represents the net slag discharge at adjacent time points. The time interval is defined as follows. Simultaneously, based on the current dynamic slag discharge threshold, an allowable rate of change threshold is calculated according to a preset proportional relationship. This is used to calculate the dynamic rate of change threshold. ,have:

[0077] ，

[0078] in, To establish a preset proportional coefficient, historical data analysis is used to statistically determine the correlation between the rate of change in bee colony ash discharge under different environmental conditions and the dynamic ash discharge threshold. A linear relationship is fitted, and the slope is calculated as the proportional coefficient. If the actual rate of change exceeds the threshold, the current net ash discharge volume is further compared with the dynamic threshold. When both conditions are met—that the net ash discharge rate of change is greater than the dynamic rate of change threshold and the current net ash discharge volume is greater than the dynamic threshold—the system generates a primary warning signal. Figure 4 As shown, the black stepped lines represent the time variation of net slag discharge, the dashed stepped lines represent dynamically adjusted thresholds, and the dark-filled areas represent warning zones where slag discharge exceeds the thresholds. As illustrated, if the above conditions are met within 3-5 hours, a primary warning is triggered. This method achieves a dual-verification warning mechanism by monitoring both the rate and absolute value of slag discharge changes, thus improving the accuracy of the warning.

[0079] For example, a beehive's net ash discharge between 8:00 AM and 11:00 AM was from Increase to The rate of change in net slag discharge was calculated. for Current dynamic slag discharge threshold for , corresponding to the dynamic rate of change threshold for .because Not exceeding Even if the net slag discharge at this point in time is It has exceeded The system still cannot generate a primary early warning signal. Simultaneously considering a dual judgment mechanism of both rate of change and absolute value, it can capture both sudden abnormal growth and sustained, slow exceedances, avoiding missed or false alarms that may result from judging by a single indicator, thus making the early warning results more comprehensive and reliable.

[0080] Optionally, the analysis and feature extraction of the bee activity video includes:

[0081] Extract the flight trajectory data of the bees from the bee activity video;

[0082] Calculate the motion characteristic parameters of the flight trajectory data and identify the visual characteristic parameters of the substances attached to the bee's body surface;

[0083] The video of bee activity is analyzed and features are extracted by combining the motion feature parameters and the visual feature parameters to generate behavior analysis results.

[0084] Specifically, the system uses computer vision technology to process the collected videos of bee activity. First, it extracts individual bees from each frame using background subtraction. Then, it tracks the continuous motion trajectory of the bees using optical flow to extract flight trajectory data. For each tracked bee, it calculates three motion characteristic parameters from the flight trajectory data: flight speed, angular velocity, and trajectory tortuosity. For calculating flight speed... angular velocity Curvature of the motion trajectory ,have:

[0085] ,

[0086] in, The time interval between video frames. This represents the displacement of the bee within a time interval. To change the angle of flight direction, This is the total length of the actual flight path. The distance is the straight-line distance from the starting point to the ending point. Simultaneously, image segmentation algorithms are used to detect attachments on the bee's body surface, calculating two visual feature parameters: the proportion of the attachment coverage area and the color abnormality index. These five feature parameters are input into a pre-defined behavioral analysis model, which outputs a judgment of whether the bee's behavior is normal or abnormal. The behavioral analysis model is built using machine learning methods (such as support vector machines and random forests) to train behavioral data such as bee flight trajectories, thereby labeling normal and abnormal samples, and then optimizing the model parameters through cross-validation. This method comprehensively assesses the health status of bees through multi-dimensional feature analysis.

[0087] For example, the system analyzes a 30-second video and tracks the average flight speed of 15 bees. The average angular velocity is The average trajectory tortuosity was 1.8. Simultaneously, three bees were detected with over 15% of their bodies covered in exfoliated material, exhibiting a high color abnormality index. The behavioral analysis model, integrating these characteristics, determined that the bee colony exhibited abnormal behavior. Multi-parameter joint analysis can accurately identify subtle behavioral changes in bees, detect potential pest and disease threats early, provide beekeepers with timely and effective management decision-making support, and prevent further deterioration of the bee colony's health.

[0088] Optionally, the identification of abnormal bee behavior features includes:

[0089] Compare the motion characteristic parameters with a preset range of normal flight parameters;

[0090] Detect whether the visual feature parameters exceed a preset attachment threshold;

[0091] When the motion characteristic parameters exceed the normal flight parameter range or the visual characteristic parameters exceed the attachment threshold, abnormal bee behavior characteristics are marked.

[0092] Specifically, the system compares the collected bee movement characteristic parameters with preset normal ranges and determines whether the visual characteristic parameters exceed preset attachment thresholds. The normal range for flight speed is... The normal range of angular velocity is The normal range for trajectory tortuosity is: For visual feature parameters, the threshold for the coverage ratio of the attachment is set to 10%, and the threshold for the color anomaly index is set to 15. If the flight speed of any bee is detected to be lower than... or higher Or the angular velocity value exceeds The range, or the trajectory tortuosity value, is not within the range. If the bee's movement is abnormal, it is determined that the bee's body surface is abnormal. Additionally, if the percentage of bee covered by attached material exceeds 10% or the color abnormality index exceeds 15, the bee is considered to have abnormal behavior. When more than 20% of the individuals in the colony are marked as abnormal, the system determines that the entire colony exhibits abnormal behavioral characteristics. This method achieves an objective assessment of bee health by establishing quantitative judgment criteria.

[0093] For example, in one detection, the system found that 25% of the bees in the colony had a flight speed lower than [a certain value]. 15% of the bees had a 12% coverage of their body surface. Since both the percentage of individual abnormalities exceeding 20% ​​and visual parameters exceeding standards were met, the system determined that the bee colony exhibited abnormal behavioral characteristics. By setting strict dual judgment criteria, false alarms caused by individual bee abnormalities can be avoided, while timely detection of abnormal signs in the colony can be achieved, providing beekeepers with reliable early warnings of bee colony health and facilitating targeted prevention and control measures.

[0094] Optionally, generating the secondary warning signal includes:

[0095] Based on the abnormal bee behavior characteristics, a weighted calculation is performed using the net slag discharge volume and the dynamic slag discharge threshold to generate a comprehensive score for abnormal behavior.

[0096] When the comprehensive score of the abnormal behavior falls into the preset warning level threshold range, a corresponding secondary warning signal level is generated;

[0097] The secondary warning signal level is correlated with a preset bee colony health status database to generate a secondary warning signal.

[0098] Specifically, combining the characteristics of abnormal bee behavior, net waste discharge volume, and dynamic waste discharge threshold, a weighted algorithm is used to calculate a comprehensive score for abnormal behavior. ,have:

[0099] ,

[0100] in, Indicates the intensity of abnormal bee behavior characteristics. This indicates the deviation between the net slag discharge volume and the dynamic slag discharge threshold. and These are the weighting coefficients of the two, and + =1. The initial setting of the weighting coefficients is mainly based on expert experience and historical data analysis to ensure that each indicator (such as the intensity of abnormal bee behavior characteristics and the deviation of waste excrement) contributes reasonably to the overall score. For example, during peak pest and disease seasons, the weight of the intensity of abnormal bee behavior characteristics should be appropriately increased. When environmental factors change drastically, the weight of the deviation in slag discharge volume can be appropriately increased. Adaptive weight adjustment can be achieved by collecting data over a period of time (e.g., one year), using machine learning regression algorithms to fit the comprehensive score with the beekeeper's manual diagnostic results, and then back-calculating the optimal weight coefficients. and The system presets three warning level threshold ranges. Y ,have:

[0101] ,

[0102] After matching the calculated comprehensive score of abnormal behavior with the range of warning levels, a corresponding secondary warning signal level is generated. The system retrieves the health assessment report and treatment suggestions corresponding to the warning level from the preset bee colony health status database, and maps them to the secondary warning signal level to generate a secondary warning signal containing a specific risk description and response measures. The preset bee colony health status database is obtained by integrating historical disease cases, literature, and expert knowledge, and structurally storing disease characteristics, warning levels, and treatment measures. Data sources include laboratory research, apiary records, and industry standards, such as data on typical diseases included in the "Handbook for the Prevention and Control of Bee Diseases and Pests," which are associated with their symptoms and treatment plans. This method achieves accurate graded warning of bee colony health risks through quantitative assessment and multi-dimensional data fusion.

[0103] For example, a bee colony detected the intensity of abnormal behavioral characteristics. slag discharge deviation ,Pick , Calculated comprehensive score The system determines that the value falls within the mild warning range, automatically generating a secondary warning signal of "suspected risk of parasitic infection," and recommends the treatment plan of "immediately check the bee colony and prepare for drug treatment." The comprehensive scoring mechanism takes into account both behavioral abnormalities and abnormal waste discharge, with clear and operable warning level classifications. The associated database provides targeted treatment suggestions, enabling beekeepers to take appropriate prevention and control measures based on the warning level, effectively improving the level of bee colony health management.

[0104] Optionally, the method further includes:

[0105] Based on the current environmental monitoring data and the abnormal bee behavior characteristics, feature extraction is performed to generate a comprehensive feature vector;

[0106] Based on the comprehensive feature vector, feature matching is performed on the preset disease feature library to generate abnormal cause categories;

[0107] Based on the aforementioned anomaly cause categories, a structured early warning report is generated.

[0108] Specifically, the system first normalizes environmental monitoring data, including parameters such as temperature and humidity, with abnormal bee behavioral characteristics, including the proportion of abnormal flight speeds and the coverage of body surface attachments, to construct a multi-dimensional comprehensive feature vector. ,have:

[0109] ,

[0110] in, Indicates the first A normalized environmental characteristic component, Indicates the first Each normalized behavioral feature component Represents the dimensions of environmental characteristics. This represents the behavioral characteristic dimension. A cosine similarity algorithm is used to match this comprehensive feature vector with typical disease features in a pre-set disease feature database to calculate a similarity score. Through collaboration with agricultural experts, continuous monitoring can be conducted on bee colonies diagnosed with specific diseases (such as bee microsporidiasis). Environmental parameters, waste excretion, and abnormal bee behavior characteristics (such as the proportion of individuals with slow flight and the coverage of body surface attachments) under typical disease stages can be recorded. The statistical values ​​(such as mean and variance) of these parameters are combined to form the feature vector of the disease and stored in a pre-set disease feature database. For example, the entry for "bee microsporidiasis": "{Feature vector: [Temperature: 25-30℃, Humidity: 70-80%, Proportion of slow flight: >30%, Waste excretion deviation: 1.2-1.5 times], Symptom description: 'Abnormal behavior: Bees fly slowly, crawl more, some individuals in the hive...'" Hesitant at the entrance or unable to take off; Physical characteristics: swollen abdomen; Waste excretion: significantly increased waste excretion containing undigested pollen grains; Colony impact: decreased overall colony vitality, reduced brood rearing, and in severe cases, colony collapse. Prevention and control recommendations: Isolate infected colonies: immediately move suspected infected hives to an isolation area to avoid cross-infection; Drug treatment: administer 0.1g / hive of fumonisin mixed with syrup for 5 consecutive days; Hygiene management: replace combs and thoroughly disinfect the hive (acetic acid fumigation or flame burning can be used); Environmental control: reduce humidity inside the hive to below 60% and avoid high temperature and high humidity environments; Follow-up monitoring: continuously observe waste excretion and bee behavior for 1 week after treatment to ensure no recurrence. (Regarding the calculation of similarity scores...) ,have:

[0111] ,

[0112] in, The first in the disease feature database Feature vectors of disease types For vector dot product, For modulo operations, the top three disease categories with the highest similarity are selected as candidate anomaly cause categories. A second screening is then performed, considering current seasonal factors and regional characteristics, to finally determine the most likely anomaly cause category. Based on this category, corresponding symptom descriptions, severity levels, and control recommendations are extracted from the knowledge base, automatically generating a structured early warning report containing cause analysis, risk level, and treatment plans. This method achieves accurate diagnosis of bee colony anomalies through multi-source data fusion and intelligent matching.

[0113] For example, the system detected the current environmental characteristics as a temperature of 28℃ and humidity of 75%, and bee behavior characteristics as 30% of individuals flying slowly and 20% having white spots on their bodies. After matching the constructed feature vector with the disease database, the similarity with the feature of "bee microsporidiasis" reached 0.92. Considering the current spring season, the system determined that the abnormality was caused by microsporidiasis infection and generated a report recommending "immediate isolation of the infected colony and treatment with fumonisin." Intelligent feature matching quickly identifies potential disease types, and the structured report provides a complete treatment chain, helping beekeepers accurately identify the root cause of the problem and providing highly actionable solutions, significantly improving the efficiency of bee colony disease control.

[0114] Based on the same inventive concept, such as Figure 5 As shown, the present invention also provides an intelligent early warning system for beehive slag discharge based on the Internet of Things, the system comprising:

[0115] The signal acquisition module is used to acquire vibration signals and weight data at the bottom of the beehive, activate the weight data acquisition process based on the vibration signals, and correct the obtained weight data to generate the net ash discharge amount.

[0116] The threshold dynamic adjustment module is used to acquire environmental monitoring data and neighboring beehive ash discharge data, and dynamically adjust the preset basic ash discharge threshold according to the environmental monitoring data and neighboring beehive ash discharge data to generate a dynamic ash discharge threshold;

[0117] The threshold comparison and early warning module is used to perform a comprehensive analysis based on the net slag discharge volume and the dynamic slag discharge threshold to generate a primary early warning signal.

[0118] The video triggering module is used to trigger a video acquisition command in response to the primary warning signal;

[0119] The video acquisition module is used to execute the video acquisition command to acquire a video of bee activity of a preset duration;

[0120] The behavior analysis module is used to analyze and extract features from the bee activity videos to identify abnormal bee behavior characteristics;

[0121] The secondary early warning generation module is used to perform system analysis and generate secondary early warning signals when abnormal bee behavior is detected.

[0122] It should be noted that the formulas described above, through the principle of dimensional consistency and mathematical standardization methods (such as normalization, dimensionless parameter conversion, or unit system unification), can translate physical quantities with different properties into unitless standard values ​​or parameters that can be superimposed in the same dimension. This eliminates the interference of different dimensions on the computational logic, allowing the formulas to retain the original data distribution characteristics while possessing mathematical rationality and adaptability to objective laws. These are conventional technical methods and will not be elaborated further. The electrical connections between the various units described above do not necessarily represent direct or indirect connections; any indirect connection method is applicable to the embodiments of this invention as long as it achieves the purpose of this invention. The above descriptions are merely exemplary embodiments of this invention and should not be construed as limiting the scope of this invention.

[0123] All equivalent changes and modifications made in accordance with the teachings of this invention are still within the scope of this invention. Those skilled in the art will readily conceive of other embodiments of this invention upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this invention that follow the general principles of this invention and include common knowledge or conventional techniques in the art not described herein.

Claims

1. A smart early warning method for beehive slag discharge based on the Internet of Things, characterized in that, The method includes: The vibration signal and weight data at the bottom of the beehive are obtained. The weight data acquisition process is activated based on the vibration signal, and the obtained weight data is corrected to generate the net amount of ash discharged. Acquire environmental monitoring data and neighboring beehive waste discharge data, and dynamically adjust the preset basic waste discharge threshold based on the environmental monitoring data and neighboring beehive waste discharge data to generate a dynamic waste discharge threshold; Based on a comprehensive analysis of the net slag discharge volume and the dynamic slag discharge threshold, a primary early warning signal is generated. The video acquisition command is triggered in response to the primary warning signal; Execute the video acquisition command to obtain a video of bee activity of a preset duration; The video of bee activity was analyzed and features were extracted to identify abnormal bee behavior characteristics; When abnormal bee behavior is detected, the system performs analysis and generates a secondary warning signal.

2. The intelligent early warning method for beehive slag discharge based on the Internet of Things according to claim 1, characterized in that, The net slag discharge volume includes: Detect whether the intensity of the vibration signal exceeds a preset vibration intensity threshold; if it does, activate the power supply of the weight sensor. Collect the raw weight data output by the activated weight sensor; Compensation calculations are performed based on the original weight data to generate the net slag discharge volume.

3. The intelligent early warning method for beehive slag discharge based on the Internet of Things according to claim 1, characterized in that, The dynamic slag discharge threshold includes: The environmental baseline value is determined based on the preset environmental parameter-slag discharge volume correlation curve; Obtain a dataset of neighboring beehive waste within a preset geographical range; Calculate the statistical average of the neighboring beehive waste disposal dataset; The environmental baseline value is corrected based on the difference between the statistical average value and the environmental baseline value, and then output as a dynamic slag discharge threshold.

4. The intelligent early warning method for beehive slag discharge based on the Internet of Things according to claim 3, characterized in that, The environmental parameter-slag discharge volume correlation curve includes: Collect historical environmental parameter datasets and corresponding historical slag discharge data; The historical environmental parameter dataset and the historical slag discharge data are grouped and statistically analyzed, and the slag discharge distribution characteristic value of each group is calculated. Fit the curve showing the relationship between the characteristic value of the slag discharge volume distribution and the historical environmental parameter dataset to generate the correlation curve between the environmental parameters and the slag discharge volume.

5. The intelligent early warning method for beehive slag discharge based on the Internet of Things according to claim 1, characterized in that, The generation of the primary early warning signal includes: Based on the net slag discharge volume, obtain the rate of change of net slag discharge volume within a preset time period; Based on the dynamic slag discharge threshold, an analysis and calculation are performed to generate the corresponding dynamic change rate threshold; When the rate of change of net slag discharge exceeds the dynamic rate of change threshold, a primary warning signal is generated by combining the comparison results of net slag discharge and dynamic slag discharge threshold.

6. The intelligent early warning method for beehive slag discharge based on the Internet of Things according to claim 1, characterized in that, The analysis and feature extraction of the bee activity video includes: Extract the flight trajectory data of the bees from the bee activity video; Calculate the motion characteristic parameters of the flight trajectory data and identify the visual characteristic parameters of the substances attached to the bee's body surface; The video of bee activity is analyzed and features are extracted by combining the motion feature parameters and the visual feature parameters to generate behavior analysis results.

7. The intelligent early warning method for beehive slag discharge based on the Internet of Things according to claim 6, characterized in that, The identification of abnormal bee behavior characteristics includes: Compare the motion characteristic parameters with a preset range of normal flight parameters; Detect whether the visual feature parameters exceed a preset attachment threshold; When the motion characteristic parameters exceed the normal flight parameter range or the visual characteristic parameters exceed the attachment threshold, abnormal bee behavior characteristics are marked.

8. The intelligent early warning method for beehive slag discharge based on the Internet of Things according to claim 1, characterized in that, The generation of secondary early warning signals includes: Based on the abnormal bee behavior characteristics, a weighted calculation is performed using the net slag discharge volume and the dynamic slag discharge threshold to generate a comprehensive score for abnormal behavior. When the comprehensive score of the abnormal behavior falls into the preset warning level threshold range, a corresponding secondary warning signal level is generated; The secondary warning signal level is correlated with a preset bee colony health status database to generate a secondary warning signal.

9. The intelligent early warning method for beehive slag discharge based on the Internet of Things according to claim 8, characterized in that, The method further includes: Based on the current environmental monitoring data and the abnormal bee behavior characteristics, feature extraction is performed to generate a comprehensive feature vector; Based on the comprehensive feature vector, feature matching is performed on the preset disease feature library to generate abnormal cause categories; Based on the aforementioned anomaly cause categories, a structured early warning report is generated.

10. An IoT-based intelligent early warning system for beehive ash discharge, applied to the IoT-based intelligent early warning method for beehive ash discharge as described in any one of claims 1-9, characterized in that, The system includes: The signal acquisition module is used to acquire vibration signals and weight data at the bottom of the beehive, activate the weight data acquisition process based on the vibration signals, and correct the obtained weight data to generate the net ash discharge amount. The threshold dynamic adjustment module is used to acquire environmental monitoring data and neighboring beehive ash discharge data, and dynamically adjust the preset basic ash discharge threshold according to the environmental monitoring data and neighboring beehive ash discharge data to generate a dynamic ash discharge threshold; The threshold comparison and early warning module is used to perform a comprehensive analysis based on the net slag discharge volume and the dynamic slag discharge threshold to generate a primary early warning signal. The video triggering module is used to trigger a video acquisition command in response to the primary warning signal; The video acquisition module is used to execute the video acquisition command to acquire a video of bee activity of a preset duration; The behavior analysis module is used to analyze and extract features from the bee activity videos to identify abnormal bee behavior characteristics; The secondary early warning generation module is used to perform system analysis and generate secondary early warning signals when abnormal bee behavior is detected.

Images