Harmful insect detection system based on vibration frequency recognition

Abstract

The invention discloses a harmful insect detection system based on vibration frequency recognition, which belongs to the technical field of insect pest detection and comprises an intelligent beehive built-in sensor module, an intelligent beehive data processing module, a computer processing module and a beehive early warning module. Wherein the intelligent beehive built-in sensor module collects vibration frequency data in the intelligent beehive; a data processing module of the intelligent beehive compresses and transmits the vibration frequency data; the computer processing module extracts vibration frequency characteristics based on the vibration frequency data, and performs insect type identification based on a pre-trained SVM model and a random forest model; and the beehive early warning module sends an alarm notification based on the identification result. Through accurate vibration frequency analysis, real-time monitoring and efficient management of potential threats of the bee colony are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of insect pest detection, and more particularly to a harmful insect detection system based on vibration frequency recognition. Background Art

[0002] As an important pollinating insect, bees play a key role in agricultural production and the balance of the ecosystem. However, the number of bee colonies has dropped sharply in recent years, which has aroused widespread concern. In response to this challenge, various technical means have been introduced into bee breeding and management, including traditional beehive management, bee colony behavior monitoring, and disease prevention and control measures. At present, there are some intelligent beehive products on the market. These products usually monitor the environmental parameters inside the intelligent beehive through temperature and humidity sensors, cameras, GPS positioning and other equipment to help beekeepers understand the health status of the bee colony.

[0003] However, although the existing smart beehive technology has improved beekeepers' management of bee colonies to a certain extent, there are still some problems that need to be solved. First, the monitoring functions of most existing equipment are relatively single, and it is difficult to fully reflect the dynamic changes of bee colonies, especially in the key aspect of detecting natural enemy invasions (harmful insects). Secondly, data processing and real-time early warning mechanisms are still imperfect, resulting in beekeepers often not responding in time when faced with emergencies. In addition, the adaptability of existing technologies under different climatic conditions and bee colony sizes needs to be improved.

[0004] Therefore, how to provide a harmful insect detection system based on vibration frequency recognition that can solve the above problems is an issue that technical personnel in this field urgently need to solve. Summary of the invention

[0005] In view of this, the present invention provides a harmful insect detection system based on vibration frequency recognition to solve the technical problems existing in the above-mentioned prior art.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] A harmful insect detection system based on vibration frequency recognition, comprising:

[0008] The smart beehive has a built-in sensor module, a smart beehive data processing module, a computer processing module, and a beehive early warning module;

[0009] Wherein, the built-in sensor module of the smart beehive collects vibration frequency data in the smart beehive;

[0010] The data processing module of the intelligent beehive compresses and transmits the vibration frequency data;

[0011] The computer processing module extracts the vibration frequency features based on the vibration frequency data, and performs insect species identification based on the pre-trained SVM model and random forest model respectively;

[0012] The beehive early warning module sends an alarm notification based on the identification result.

[0013] Optionally, the smart beehive built-in sensor module includes: a vibration sensor and a filter;

[0014] Among them, the vibration sensor collects the vibration frequency data in the smart beehive, and the filter is used to process the influence of environmental noise on the vibration frequency data in the smart beehive collected by the vibration sensor. The expression of the filtering process is:

[0015] y[n] = Filter(x[n]);

[0016] Where x[n] is the original frequency-time domain signal, and y[n] is the filtered frequency-time domain signal.

[0017] Optionally, in the data processing module of the smart beehive, extracting the vibration frequency features based on the vibration frequency data includes:

[0018]

[0019] Where Y(k) is the frequency domain representation, y[n] is the filtered frequency time domain signal, and N is the number of samples.

[0020] Optionally, in the data processing module of the intelligent beehive, insect species identification is performed based on a pre-trained SVM model and a random forest model, including:

[0021] Obtain annotated data sets and use them to train SVM models and random forest models respectively;

[0022] The extracted frequency features are standardized to make the dimensions of different features consistent;

[0023] For the standardized frequency features, the random forest model combines the prediction results of multiple decision trees and uses expert scoring to obtain the final classification result as the first detection result;

[0024] For the standardized vibration frequency features, the SVM model calculates the probability of belonging to each harmful insect category, and takes the category with the highest probability as the second detection result;

[0025] The first detection result and the second detection result are combined to output the final insect species recognition result.

[0026] Optionally, the labeled data set includes: the vibration frequency characteristics of various insects and their corresponding category labels.

[0027] Optionally, training a random forest model using a labeled dataset includes:

[0028] Construct multiple decision trees, each trained on a random subset of the labeled dataset, so that the decision tree can make classification decisions based on the feature vector.

[0029] Optionally, training an SVM model using a labeled dataset includes:

[0030] Divide the labeled data set into a training set and a test set;

[0031] The test set is input into the SVM model, and the penalty parameter c and the parameter g of the kernel function are calculated respectively;

[0032] Gradually expand the range of the penalty parameter c and the kernel function parameter g and reduce the step size, saving the penalty parameter c and the kernel function parameter g each time;

[0033] Randomly select the SVM model corresponding to the penalty parameter c and kernel function parameter g saved in the test set from the test set and perform multiple tests to obtain the optimal penalty parameter c and kernel function parameter g;

[0034] Save the optimal penalty parameter c, the kernel function parameter g and the corresponding SVM model.

[0035] Optionally, the first detection result and the second detection result are fused to output a final insect species identification result, including:

[0036]

[0037]

[0038] Wherein, W represents the final insect species recognition result; F() represents the symbol of the function; W1 represents the first detection result of the output of the random forest model; W2 represents the first detection result of the output of the SVM model; β1 represents the error value of the random forest model; β2 represents the error value of the SVM model; ω1 represents the weight parameter of the first detection result of the output of the random forest model; ω2 represents the weight parameter of the first detection result of the output of the SVM model; θ represents the parameter.

[0039] Optionally, it also includes: a data storage and analysis module, which is used to store all collected vibration frequency data and recognition results, and perform behavior analysis and trend analysis.

[0040] Optionally, it also includes: a regular system calibration module, which is used to perform regular self-inspection and calibration on the smart beehive built-in sensor module, the smart beehive data processing module and the computer processing module set in the smart beehive.

[0041] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a harmful insect detection system based on frequency recognition, which realizes the identification and early warning of harmful insects by detecting the vibration frequency (frequency signal) of insect wings. The vibration sensor built into the intelligent beehive can collect the frequency data of insect wings passing through the entrance of the beehive in real time. The computer processing module extracts features from the collected frequency data, and matches the extracted characteristic frequency with the built-in harmful insect frequency database, thereby identifying harmful insects such as wasps and wasps. Once harmful insects are identified, the system will immediately trigger the early warning mechanism and promptly notify the beekeeper to take protective measures. The present invention realizes real-time monitoring and efficient management of potential threats to the bee colony through precise frequency analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0043] Figure 1 It is a structural schematic diagram of the present invention. DETAILED DESCRIPTION

[0044] 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.

[0045] The purpose of the present invention is to provide a harmful insect detection system based on vibration frequency recognition, comprising: a smart beehive built-in sensor module, a smart beehive data processing module, a computer processing module and a beehive early warning module; wherein the smart beehive built-in sensor module collects vibration frequency data in the smart beehive; the smart beehive data processing module compresses and transmits the vibration frequency data; the computer processing module extracts vibration frequency features based on the vibration frequency data, and performs insect species recognition based on pre-trained SVM models and random forest models; the beehive early warning module sends an alarm notification based on the recognition result. Provides a solution to the monitoring blind spot and early warning lag problems in the prior art.

[0046] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0047] See also Figure 1 The embodiment of the present invention discloses a harmful insect detection system based on vibration frequency recognition, comprising:

[0048] The smart beehive has a built-in sensor module, a smart beehive data processing module, a computer processing module, and a beehive early warning module;

[0049] Wherein, the built-in sensor module of the smart beehive collects vibration frequency data in the smart beehive;

[0050] The data processing module of the intelligent beehive compresses and transmits the vibration frequency data;

[0051] The computer processing module extracts the vibration frequency features based on the vibration frequency data, and performs insect species identification based on the pre-trained SVM model and random forest model respectively;

[0052] The beehive early warning module sends an alarm notification based on the identification result.

[0053] In a specific embodiment, the smart beehive built-in sensor module includes: a vibration sensor and a filter;

[0054] Among them, the vibration sensor installed at the entrance of the beehive collects the vibration frequency data of the insect wings in the environment, and the filter is used to process the influence of environmental noise on the vibration frequency data in the intelligent beehive collected by the vibration sensor. The expression of the filtering process is:

[0055] y[n] = Filter(x[n]);

[0056] Where x[n] is the original frequency-time domain signal, and y[n] is the filtered frequency-time domain signal.

[0057] Specifically, the sensor module built into the smart beehive monitors and records the vibration frequency of all insects passing through the entrance of the beehive in real time. The main function of this module is to obtain accurate vibration frequency data to provide a basis for subsequent analysis.

[0058] Specifically, the data processing module of the smart beehive transmits all the collected frequency data to the data processing module of the smart beehive in real time through wireless communication. The data processing module of the smart beehive is responsible for receiving and preliminarily processing (compressing) these data. Among them, a data compression algorithm (such as Huffman coding or Lempel-Ziv-Welch compression) can be used to compress the frequency data to further reduce the amount of data transmission.

[0059] In a specific embodiment, the data processing module of the intelligent beehive performs feature extraction on the received vibration frequency data to extract characteristic vibration frequencies (f) representing different insects. The key function of this step is to convert complex vibration frequency data into characteristic frequencies for subsequent identification and classification.

[0060] Specifically, extracting frequency features based on frequency data includes:

[0061]

[0062] Where Y(k) is the frequency domain representation, y[n] is the filtered frequency time domain signal, and N is the number of samples.

[0063] Specifically, the construction of the feature vector, f = [f1, amp1, energy1, ..., f M ,amp M ,energy M ], where M is the number of different frequency components detected.

[0064] In a specific embodiment, in the data processing module of the intelligent beehive, insect species recognition is performed based on the pre-trained SVM model and the random forest model, including

[0065] Obtain annotated data sets and use them to train SVM models and random forest models respectively;

[0066] The extracted frequency features are standardized to make the dimensions of different features consistent;

[0067] For the standardized frequency features, the random forest model combines the prediction results of multiple decision trees and uses expert scoring to obtain the final classification result as the first detection result;

[0068] For the standardized vibration frequency features, the SVM model calculates the probability of belonging to each harmful insect category, and takes the category with the highest probability as the second detection result;

[0069] The first detection result and the second detection result are combined to output the final insect species recognition result.

[0070] In a specific embodiment, the labeled data set includes: the vibration frequency characteristics of various insects and their corresponding category labels. The various insects include: harmful insects such as wasps and wasps, and the data set stores the vibration frequency characteristics of common harmful insects.

[0071] In a specific embodiment, it also includes:

[0072] Matching function: define a matching function Match(f,d i ), where di is the feature vector of the i-th insect.

[0073] Decision rule: If max i Match(f,d i ) exceeds the set threshold, it is considered to be that insect species.

[0074] In a specific embodiment, training a random forest model using a labeled data set includes:

[0075] Construct multiple decision trees, each trained on a random subset of the labeled dataset, so that the decision tree can make classification decisions based on the feature vector.

[0076] Specifically, the construction of the decision tree:

[0077] (1) Select the best split point: At each node, the algorithm will traverse all features and their possible split points (for classification problems, it is usually based on indicators such as information gain and Gini impurity) and select the best split method. Mathematically, this can be expressed as maximizing information gain IG(D,a), where D is the data set of the current node and a is a feature and its split point:

[0078]

[0079] Among them, Entropy(D) is the entropy of data set D, D v It is a subset divided according to the value v of feature a.

[0080] (2) Recursive splitting: This process is recursively applied to each child node until a stopping condition is met (e.g., the number of samples in the node is less than a certain threshold, all samples belong to the same category, or there are no more features to split);

[0081] In a random forest, each decision tree makes a classification prediction based on the input feature vector x. The final classification decision is usually based on the "voting" results of all trees:

[0082] Specifically, voting mechanism: For classification problems, the most common strategy is majority voting. That is, each tree outputs a class label, and the final prediction is the one that appears the most times among these labels.

[0083] Specifically, multiple decision trees are constructed by maximizing criteria such as information gain, and the prediction results of multiple decision trees are combined using strategies such as majority voting. The model performance is evaluated by comparing the prediction results with the true labels.

[0084] In a specific embodiment, training the SVM model using the labeled data set includes:

[0085] Divide the labeled data set into a training set and a test set;

[0086] The test set is input into the SVM model, and the penalty parameter c and the parameter g of the kernel function are calculated respectively;

[0087] Gradually expand the range of the penalty parameter c and the kernel function parameter g and reduce the step size, saving the penalty parameter c and the kernel function parameter g each time;

[0088] Randomly select the SVM model corresponding to the penalty parameter c and kernel function parameter g saved in the test set from the test set and perform multiple tests to obtain the optimal penalty parameter c and kernel function parameter g;

[0089] Save the optimal penalty parameter c, the kernel function parameter g and the corresponding SVM model.

[0090] In a specific embodiment, the first detection result and the second detection result are fused to output a final insect species identification result, including:

[0091]

[0092] Wherein, W represents the final insect species recognition result; F() represents the symbol of the function; W1 represents the first detection result of the output of the random forest model; W2 represents the first detection result of the output of the SVM model; β1 represents the error value of the random forest model; β2 represents the error value of the SVM model; ω1 represents the weight parameter of the first detection result of the output of the random forest model; ω2 represents the weight parameter of the first detection result of the output of the SVM model; θ represents the parameter.

[0093] Specifically, an integrated learning model combining random forest (RF) and support vector machine (SVM) is constructed to identify insect species. This integration method usually combines the prediction results of different models through weighted average or voting mechanism to improve the overall recognition accuracy and robustness.

[0094] In a specific embodiment, the beehive warning module sets different types of insect invasion thresholds and adjusts the alarm level according to the severity. When the computer identifies harmful insects, it immediately triggers the early warning system in the beehive and sends an alarm notification to the beekeeper. The specific alarm information includes the type of insect, invasion time and frequency, etc. The purpose of this step is to promptly remind the beekeeper to take countermeasures to protect the safety of the bee colony.

[0095] In a specific embodiment, it also includes: a data storage and analysis module for storing all collected frequency data and recognition results, and performing behavior analysis and trend analysis. The long-term accumulation of data helps to improve the accuracy of the recognition algorithm and provide data support for future bee behavior research.

[0096] In a specific embodiment, it also includes: a regular system calibration module, which is used to perform regular self-inspection and calibration on the smart beehive built-in sensor module, the smart beehive data processing module and the computer processing module set in the smart beehive, so as to ensure the recognition accuracy and reliability of the system, ensure the long-term stable operation of the equipment, and reduce the possibility of false alarms and missed alarms.

[0097] Example 2

[0098] This embodiment provides an intelligent monitoring beehive designed for bee breeding. The intelligent monitoring beehive uses any of the harmful insect detection systems based on vibration frequency recognition provided in Embodiment 1 of the present invention to monitor the activity status of bee colonies in real time and detect harmful insects such as wasps and hornets. At the same time, the intelligent monitoring beehive monitors the weight change of honey combs and the temperature and humidity in the box. The beehive transmits the collected data to the mobile terminal of the beekeeper through the Internet of Things technology to achieve real-time early warning and remote management.

[0099] Specifically, the main structure of the beehive includes:

[0100] Hive Shell: The main structure of the beehive, made of waterproof and anti-corrosion materials to protect the internal components from the external environment.

[0101] Beehive entrance: Located at the front of the beehive, it is used by bees to enter and exit the beehive. The smart beehive built-in sensor module is installed at the entrance. The smart beehive built-in sensor module is responsible for detecting the vibration frequency of insect wings. The smart beehive built-in sensor module is connected to the data processing module of the smart beehive through a signal line.

[0102] Temperature and humidity sensor: installed inside the beehive, responsible for monitoring the temperature and humidity inside the beehive, and connected to the computer processing module through a signal line.

[0103] Weight sensor: Installed at the bottom of the beehive to monitor the weight change of the honeycomb. The weight sensor is connected to the computer processing module through a signal line.

[0104] Computer processing module: The core processor located in the beehive receives data from each sensor module and performs real-time analysis and processing to obtain recognition results.

[0105] Data storage and analysis module: used to store all collected vibration frequency data and recognition results, and perform behavior analysis and trend analysis. It can also include temperature, humidity and weight data storage and analysis.

[0106] Data processing module of the smart beehive: transmits processed data to the beekeeper's mobile terminal via wireless network, supporting remote management and real-time early warning.

[0107] Power module: Solar panel: installed on the top of the beehive to provide power to the hive, connected to the internal battery and power management system.

[0108] Battery and power management system: Stores the electricity generated by the solar panels and powers each module to ensure the continuous operation of the beehive.

[0109] Beehive early warning module: When harmful insects or abnormalities in the beehive are detected, an alarm notification will be sent through the beekeeper's mobile terminal.

[0110] The beehive shell protects all internal components and provides a suitable environment.

[0111] The sensor module is responsible for data collection, the vibration sensor detects harmful insects, the temperature and humidity sensor monitors the internal environment, and the weight sensor monitors the weight of the honey comb.

[0112] The computer processing module processes and analyzes the data, decides whether to trigger an early warning, and sends the data to the beekeeper's mobile terminal through the wireless communication module.

[0113] The power module provides stable power support for the entire system, and the solar panels ensure round-the-clock operation through the power management system.

[0114] The early warning system ensures that beekeepers can be notified in time when harmful insects invade or the environment is abnormal, so that they can take appropriate measures.

[0115] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0116] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A harmful insect detection system based on vibration frequency recognition, characterized in that: include: The smart beehive has a built-in sensor module, a smart beehive data processing module, a computer processing module, and a beehive early warning module; Wherein, the built-in sensor module of the smart beehive collects vibration frequency data in the smart beehive; The data processing module of the intelligent beehive compresses and transmits the vibration frequency data; The computer processing module extracts the vibration frequency features based on the vibration frequency data, and performs insect species identification based on the pre-trained SVM model and random forest model respectively; The beehive early warning module sends an alarm notification based on the identification result.

2. A harmful insect detection system based on vibration frequency recognition according to claim 1, characterized in that: The smart beehive built-in sensor module includes: a vibration sensor and a filter; Among them, the vibration sensor collects the vibration frequency data in the smart beehive, and the filter is used to process the influence of environmental noise on the vibration frequency data in the smart beehive collected by the vibration sensor. The expression of the filtering process is: y[n] = Filter(x[n]); Where x[n] is the original frequency-time domain signal, and y[n] is the filtered frequency-time domain signal.

3. A harmful insect detection system based on vibration frequency recognition according to claim 1, characterized in that: In the data processing module of the intelligent beehive, extracting the vibration frequency features based on the vibration frequency data includes: Where Y(k) is the frequency domain representation, y[n] is the filtered frequency time domain signal, and N is the number of samples.

4. The harmful insect detection system based on vibration frequency recognition according to claim 1, characterized in that: In the data processing module of the intelligent beehive, insect species identification is performed based on the pre-trained SVM model and random forest model, including Obtain annotated data sets and use them to train SVM models and random forest models respectively; The extracted frequency features are standardized to make the dimensions of different features consistent; For the standardized frequency features, the random forest model combines the prediction results of multiple decision trees and uses expert scoring to obtain the final classification result as the first detection result; For the standardized vibration frequency features, the SVM model calculates the probability of belonging to each harmful insect category, and takes the category with the highest probability as the second detection result; The first detection result and the second detection result are combined to output the final insect species recognition result.

5. A harmful insect detection system based on vibration frequency recognition according to claim 4, characterized in that: The annotated data set includes: the vibration frequency characteristics of various insects and their corresponding category labels.

6. The harmful insect detection system based on vibration frequency recognition according to claim 4, characterized in that: Training a random forest model using annotated datasets involves: Construct multiple decision trees, each trained on a random subset of the labeled dataset, so that the decision tree can make classification decisions based on the feature vector.

7. The harmful insect detection system based on vibration frequency recognition according to claim 4, characterized in that: Training the SVM model using annotated datasets involves: Divide the labeled data set into a training set and a test set; The test set is input into the SVM model, and the penalty parameter c and the parameter g of the kernel function are calculated respectively; Gradually expand the range of the penalty parameter c and the kernel function parameter g and reduce the step size, saving the penalty parameter c and the kernel function parameter g each time; Randomly select the SVM model corresponding to the penalty parameter c and kernel function parameter g saved in the test set from the test set and perform multiple tests to obtain the optimal penalty parameter c and kernel function parameter g; Save the optimal penalty parameter c, the kernel function parameter g and the corresponding SVM model.

8. The harmful insect detection system based on vibration frequency recognition according to claim 4, characterized in that: The first detection result and the second detection result are combined to output the final insect species identification result, including: Wherein, W represents the final insect species recognition result; F() represents the symbol of the function; W1 represents the first detection result of the output of the random forest model; W2 represents the first detection result of the output of the SVM model; β1 represents the error value of the random forest model; β2 represents the error value of the SVM model; ω1 represents the weight parameter of the first detection result of the output of the random forest model; ω2 represents the weight parameter of the first detection result of the output of the SVM model; θ represents the parameter.

9. The harmful insect detection system based on vibration frequency recognition according to claim 1, characterized in that: Also includes: The data storage and analysis module is used to store all collected frequency data and recognition results, and perform behavior analysis and trend analysis.

10. The harmful insect detection system based on vibration frequency recognition according to claim 1, characterized in that: Also includes: The regular system calibration module is used to perform regular self-inspection and calibration on the smart beehive built-in sensor module, the smart beehive data processing module and the computer processing module set in the smart beehive.

Images