A farmland pest real-time monitoring and early warning method and device based on acoustic detection

Abstract

The application provides a farmland pest real-time monitoring and early warning method and device based on acoustic detection, sound signals in farmland are collected by using a microphone array, the sound signals are denoised by using an audio preprocessing unit, and a feature vector of the sound signals is extracted and sent to a local data processing unit; the local data processing unit compares the feature vector of the sound signals with a standard data set by using a classification algorithm, obtains information about a pest type and a frog type corresponding to the sound signals, and samples the sound signals according to the information about the pest type and the frog type, determines the number of pests and the number of frogs, and sends the number of pests and the number of frogs to a central data processing unit; the central data processing unit sets an early warning threshold according to the number of frogs and the number of pests, and reminds a user to remove pests when the early warning threshold is exceeded. The application combines sound source detection and big data analysis, realizes accurate monitoring and early warning of farmland pests, has low cost, can effectively reduce the use of pesticides, ensures product safety, and is beneficial to the protection of the ecological environment.

Description

Technical Field

[0001] This invention relates to the field of pest and disease control technology, specifically to a method and device for real-time monitoring and early warning of agricultural pests based on acoustic detection. Background Technology

[0002] Pests and diseases are a global agricultural problem that has long caused enormous losses to agriculture and remains a challenge in agricultural production. They not only directly affect crop yield and quality but can also lead to fluctuations in agricultural product prices, threatening food security.

[0003] Existing pest control methods mostly include biological control, physical control, and chemical control. Biological control reduces pest populations by introducing natural enemies, but it is costly, unaffordable for small-scale farmers, and its slow effect makes it unsuitable for rapid control. Physical control uses physical or mechanical methods to control pests, such as light repellent, hot water soaking, high-temperature treatment, and netting. Although physical control methods do not involve chemicals and are therefore more environmentally friendly, their efficiency in controlling large-scale pests is very low, making them only suitable as temporary supplementary control measures. Chemical control kills pests in farmland by spraying pesticides, is highly efficient, and easy to implement, making it the mainstream method for pest control. Currently, chemical control methods include general spraying, precision spraying, spot spraying, and strip spraying. General spraying causes significant environmental pollution and leads to the rapid development of pesticide resistance in pests, while also wasting resources. Precision spraying, compared to other methods, is the most economical in terms of pesticide use and has the least environmental pollution. To achieve precise spraying, the mainstream identification method currently combines technologies such as computer vision, artificial intelligence, and machine learning algorithms. However, due to the involvement of expensive hardware, the cost of this technology is high, especially for small and medium-sized farmers. Even if the subsequent returns outweigh the investment, they may not have sufficient funds to deploy this technology upfront. Summary of the Invention

[0004] Purpose of the invention: The purpose of this invention is to provide a low-cost, high-recognition-rate method and device for real-time monitoring and early warning of farmland pests based on acoustic detection.

[0005] Technical solution: A method for real-time monitoring and early warning of farmland pests based on acoustic detection, comprising the following steps:

[0006] (1) Collect sound signals in the target area;

[0007] (2) Denoise the sound signal and extract the feature vector of the sound signal;

[0008] (3) The feature vector of the sound signal is compared with the standard dataset using a classification algorithm to obtain the insect species information and frog species information corresponding to the sound signal; the sound signal is sampled based on the insect species information and frog species information to determine the insect quantity information and frog quantity information.

[0009] (4) Set an early warning threshold based on the number of frogs and insects, and issue an alarm when the threshold is exceeded.

[0010] Specifically, step (2) includes:

[0011] (21) Discretize the sound signal A(t) to obtain A(n), and calculate the signal frequency based on A(n):

[0012]

[0013] In the formula: A(t) represents the sound pressure level at each time point, n in A(n) is the index of the sampling point, the total number of sampling points is N, and f k f is the signal frequency. s The sampling frequency is set when acquiring sound signals, and k is the index of the sampling point in the frequency domain;

[0014] (22) A filter is used to denoise the audio signal. The frequency response formula of the filter is:

[0015]

[0016] In the formula: H(f) k f is the frequency response of the filter. k f is the signal frequency. c Here, n is the cutoff frequency, and n is the order of the filter.

[0017] (23) After filtering, the denoised audio signal A′(t) is obtained. A fast Fourier transform is performed on the denoised audio signal A′(t) to convert the time-domain signal into a frequency-domain signal:

[0018]

[0019] In the formula: B(f) is the frequency domain signal, A′(t) is the denoised time domain signal, f is the frequency, t is the time, and j is the imaginary unit;

[0020] (24) Extract the main frequency components and amplitude information from the frequency domain signal B(f) to obtain the amplitude spectrum M(f), and form the feature vector C, as shown in the following formula:

[0021] M(f)=|B(f)|

[0022] C = [M(f1), M(f2), ..., M(f... n )]

[0023] In the formula: M(f) is the amplitude spectrum, f1, f2, ..., f n It is the main frequency component.

[0024] Preferably, the above filter is a fourth-order filter.

[0025] Specifically, classification algorithms include support vector machine algorithm, random forest algorithm, decision tree algorithm, K-nearest neighbor algorithm, artificial neural network algorithm, and Naive Bayes algorithm.

[0026] Preferably, in step (3), when the classification algorithm is the K-nearest neighbor algorithm, the classification algorithm is used to compare the feature vector of the sound signal with the standard dataset to obtain the insect species information and frog species information corresponding to the sound signal, including the following steps:

[0027] (311) Normalize each eigenvalue of the eigenvector C of the sound signal:

[0028]

[0029] In the formula: Let x be the j-th eigenvalue of the normalized eigenvector C. j This represents the j-th eigenvalue of the eigenvector C;

[0030] (312) Extract the feature vectors of various insect or frog samples from the standard dataset, and perform the normalization process in step (311) in the same way. Calculate the Euclidean distance between the feature vector C of the normalized sound signal and the feature vectors of various insect or frog samples in the standard dataset:

[0031]

[0032] In the formula: d i C is the Euclidean distance between C and the i-th feature vector in the standard dataset. j Let j be the j-th eigenvalue of the eigenvector in the standard dataset. Let be the j-th eigenvalue of the normalized eigenvector C, and D be the dimension of the eigenvector.

[0033] (313) Arrange the Euclidean distances between all feature vectors in the standard dataset and feature vector C in order;

[0034] (314) Given a value of k, select the k feature vectors with the smallest Euclidean distance to the feature vector C, and put the corresponding k sample indices into set I. neighbours ={i|b i In}, b i One of the k samples with the smallest Euclidean distance is selected, and the insect or frog species corresponding to the k samples are then placed into the set E = {ei |i∈I neighbours In}, e i The insect or frog species corresponding to the i-th sample;

[0035] (315) Based on the insect or frog species in set E, use majority voting to determine the insect or frog species corresponding to feature vector C, and obtain insect or frog species information:

[0036]

[0037] In the formula: As an indicator function, when c i When the expression is equal to c, the result is 1; otherwise, it is 0.

[0038] Specifically, in step (3), the sound signal sampling based on the insect species information and frog species information to determine the insect quantity information and frog quantity information includes the following steps:

[0039] (321) Based on the insect or frog species information, take the average time T for one call of the insect or frog of that type, and collect the sound signal A(t) at intervals T, for a total of M collections, to obtain the discrete amplitude s. i :

[0040] s i =A(t) i ), i∈[1,M]

[0041] In the formula: A(t) i (i) represents the sound signal acquired in the i-th acquisition.

[0042] (322) Transfer the sound signal s i Converted to digital signal s′ i :

[0043] s′ i =ADC(s) i )

[0044] (323) Based on the digital signal s′ i Calculate energy E k :

[0045]

[0046] (324) Set an energy threshold if energy E k If the energy value exceeds the energy threshold, it is determined as an event. The total number of events within M time periods T is calculated to obtain the corresponding insect or frog quantity information.

[0047] Specifically, step (4) includes:

[0048] When the number of frogs in the area is less than the set value, the warning threshold is set to 1 / 4 of the insect population carrying capacity K. When the warning threshold is exceeded, a warning message is sent to the user.

[0049] When the number of frogs in the area is greater than or equal to a set value, an early warning threshold is set based on the number of frogs. This early warning threshold is greater than 1 / 4K and less than 1 / 2K.

[0050] Preferably, step (4) further includes: calculating the insect density in the target area at different time periods based on the insect quantity information, statistically determining the time period with the highest insect density, and sending a warning message to the user.

[0051] The present invention also provides a real-time monitoring and early warning device for farmland pests based on acoustic detection, characterized in that it includes several monitoring units and a central data processing unit. The monitoring unit includes a microphone array, an audio preprocessing unit, and a local data processing unit. The microphone array includes several symmetrically distributed microphones. The microphone array, the audio preprocessing unit, and the local data processing unit are connected in sequence. The local data processing unit is connected to the central data processing unit.

[0052] Microphone array: Used to collect sound signals in the target area and transmit them to the audio preprocessing unit;

[0053] Audio preprocessing unit: used to denoise the sound signal, extract the feature vector of the sound signal, and transmit it to the local data processing unit;

[0054] Local data processing unit: Used to compare the feature vector of the sound signal with the standard dataset using a classification algorithm to obtain the insect species information and frog species information corresponding to the sound signal; sample the sound signal based on the insect species information and frog species information to determine the insect quantity information and frog quantity information, and transmit it to the central data processing unit;

[0055] Central data processing unit: Used to set early warning thresholds based on frog and insect count information, and issue an alarm when the thresholds are exceeded.

[0056] Specifically, the microphone arrays can be arranged in the following ways: uniform grid distribution: each microphone is equidistantly placed in a rectangular grid; boundary reinforcement distribution: the number of microphones is increased at the boundary of the target area and the number of microphones is reduced at the center; and ring distribution: the microphones are arranged in a circular or multiple concentric circles around the center of the target area.

[0057] Beneficial effects: Compared with the prior art, the significant effects of the present invention are:

[0058] (1) By combining acoustic detection and big data analysis technology, this invention realizes the real-time collection and analysis of information on the types and quantities of insects and frogs in farmland, achieving automatic and efficient information collection.

[0059] (2) By using the collected information on the types and numbers of insects and frogs, it is possible to determine in real time whether the number of pests has reached the warning threshold, thereby sending pest control reminders to farmers in a timely and effective manner. At the same time, it achieves accurate judgment in terms of time and area, improves the accuracy and efficiency of pest control, reduces the excessive use of pesticides, helps protect the ecological environment, and reduces pesticide residues in crops.

[0060] (3) The equipment used is inexpensive and has a cost advantage, making it suitable for small and medium-sized farmers. At the same time, the equipment is easy to set up and a single monitoring unit can be used independently offline, making it suitable for various scenarios. Attached Figure Description

[0061] Figure 1 This is a flowchart of the method in Embodiment 1 of the present invention.

[0062] Figure 2 This is a structural diagram of the device in Embodiment 2 of the present invention. Detailed Implementation

[0063] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments.

[0064] Example 1

[0065] In this embodiment, the standard datasets used are three public datasets: Macaulay Library, GBIF, and AmphibiaWeb, which contain a large amount of sound information from insects and frogs.

[0066] Please see Figure 1 As shown in the figure, this embodiment provides a method for real-time monitoring and early warning of farmland pests based on acoustic detection, including the following steps:

[0067] (1) Collect the sound signal A(t) in the target area, where A(t) represents the sound pressure level at each time point.

[0068] (2) The sound signal A(t) is denoised to remove background noise such as wind and rain, and the feature vector of the sound signal is extracted, including information such as frequency, loudness and duration.

[0069] Discretize the sound signal A(t) to obtain A(n), and calculate the signal frequency based on A(n):

[0070]

[0071] In the formula: A(t) represents the sound pressure level at each time point, n in A(n) is the index of the sampling point, the total number of sampling points is N, and f k f is the signal frequency. sThe sampling frequency is a value set when acquiring the signal, and k is the index of the sampling point in the frequency domain.

[0072] Using analog signal processing techniques, the sound signal is optimized using amplifiers and filters. Taking insect sound signals as an example, by analyzing insect sound data in the Macaulay Library, it was found that effective insect sound signals are concentrated in the frequency band of 2kHz to 8kHz. A filter is used to filter out waves outside the mid-frequency band of the sound signal. In this embodiment, considering efficiency, performance, cost, and implementation complexity, a fourth-order filter is selected, with the low cutoff frequency set to 2kHz and the high cutoff frequency set to 8kHz to preserve the sound characteristics of insects while filtering out low-frequency and high-frequency environmental noise. The frequency response formula of the filter is:

[0073]

[0074] In the formula: H(f) k f is the frequency response of the filter. k f is the signal frequency. c Here, n is the cutoff frequency, and n is the order of the filter.

[0075] After filtering, the denoised audio signal A′(t) is obtained. A Fast Fourier Transform (FFT) is then performed on the denoised audio signal A′(t) to convert the time-domain signal into a frequency-domain signal.

[0076]

[0077] In the formula: B(f) is the frequency domain signal, A′(t) is the denoised time domain signal, f is the frequency, t is the time, and j is the imaginary unit;

[0078] Extract the main frequency components and amplitude information from the frequency domain signal B(f), take the modulus of the amplitude to obtain the amplitude spectrum M(f), and form the feature vector C, as shown in the following formula:

[0079] M(f)=|B(f)|

[0080] C = [M(f1), M(f2), ..., M(f... n )]

[0081] In the formula: M(f) is the amplitude spectrum, f1, f2, ..., f n It is the main frequency component.

[0082] (3) The feature vector of the sound signal is compared with the standard dataset using a classification algorithm to obtain the insect species information and frog species information corresponding to the sound signal; the sound signal is sampled based on the insect species information and frog species information to determine the insect quantity information and frog quantity information.

[0083] In the task of processing the classification of sound signal data, common classification algorithms include the support vector machine algorithm (SVM), random forest algorithm, decision tree algorithm, K-nearest neighbor algorithm (KNN), artificial neural network algorithm, naive Bayes algorithm, etc. In this embodiment, considering the combined use of the dataset and test samples, the K-nearest neighbor algorithm is simple to implement, has no parameter model, is suitable for small datasets, and can be dynamically updated, which is in line with the usage scenario of this method. The following describes the actual application of the K-nearest neighbor algorithm in this method.

[0084] Perform normalization processing on each eigenvalue of the feature vector C of the sound signal:

[0085]

[0086] In the formula: is the j-th eigenvalue of the feature vector C after normalization processing, and x j represents the j-th eigenvalue of the feature vector C.

[0087] To distinguish the types of insects and frogs, extract the feature vectors of various insect or frog samples in the standard dataset, perform the same normalization processing, and calculate the Euclidean distance between the feature vector C of the sound signal after normalization processing and the feature vectors of various insect or frog samples in the standard dataset:

[0088]

[0089] In the formula: d i is the Euclidean distance between C and the i-th feature vector in the standard dataset, C j is the j-th eigenvalue of the feature vector in the standard dataset, is the j-th eigenvalue of the feature vector C after normalization processing, and D is the dimension of the feature vector.

[0090] Arrange the Euclidean distances between all feature vectors in the standard dataset and the feature vector C in order;

[0091] Given the k value, which is generally determined according to the size of the target area and environmental conditions, select the k feature vectors with the smallest Euclidean distance from the feature vector C, and put the corresponding k sample indices into the set I neighbours ={i|b i}, and then put the corresponding insect types or frog types of the k samples into the set E = {e i |i∈I neighbours}; b i is one of the k data samples with the smallest Euclidean distance, and e i is the corresponding insect type or frog type of the i-th data sample.

[0092] In this embodiment, the k value in the K-nearest neighbor algorithm is determined as follows: the samples in the standard dataset are divided into ten groups, one group is tested each time, and the remaining nine groups are used as samples. The steps are the same as described above. The accuracy obtained by the voting method is compared, and finally the average value of the ten groups is taken. The k value with the highest average accuracy is used as the k value in actual application.

[0093] Based on the insect or frog species in set E, the majority voting method is used to obtain the insect or frog species with the most votes as the output result, thereby determining the insect or frog species corresponding to feature vector C and obtaining insect or frog species information:

[0094]

[0095] In the formula: As an indicator function, when c i When the expression is equal to c, the result is 1; otherwise, it is 0.

[0096] Based on the insect or frog species information, the average time T for one call of that insect or frog is taken, and sound signals A(t) are collected at intervals T, for a total of M collections, to obtain the discrete amplitude s. i :

[0097] s i =A(t) i ), i∈[1,M]

[0098] In the formula: A(t) i ) represents the sound signal acquired in the i-th acquisition.

[0099] The sound signal s i Converted to digital signal s′ i :

[0100] s′ i =ADC(s) i )

[0101] According to the digital signal s′ i Calculate energy E k :

[0102]

[0103] Set energy threshold E thresh If energy E k Greater than the energy threshold E thresh If a number is found to be an event, calculate the total number of events within M time periods T. We can assume that the total number of events represents the number of insects or frogs in the target area. Repeat the above steps to obtain the information on the number of insects or frogs in the target area.

[0104] (4) Based on the information on the number of frogs and insects, and combined with the carrying capacity of the insect population, set an early warning threshold. When the early warning threshold is exceeded, an alarm will be triggered to remind the user to carry out pest control operations.

[0105] Frog populations are an important indicator of pest infestation levels. Since frogs' predatory behavior has a natural regulatory effect on insect populations, the impact of frogs should be considered when assessing the severity of pest infestations.

[0106] When the number of frogs in the area is less than a set value, such as when the number of frogs is 0, its impact is not considered. In this case, the carrying capacity of the insect population (K value) is considered, that is, the maximum number of insect populations that can be maintained under the support of existing resources and environmental conditions in a specific environment. When the population number reaches K / 2, the insect growth rate is the fastest and difficult to control. Therefore, it is necessary to control the population before it reaches K / 2. In this embodiment, the warning threshold is set to K / 4. When the warning threshold is exceeded, a warning message is sent to the user to remind the user to take timely pest control measures.

[0107] When the number of frogs in the area is greater than or equal to the set value, such as exceeding 10% of the number of insects, an early warning threshold is set based on the number of frogs. This early warning threshold is greater than 1 / 4K and less than 1 / 2K. If it reaches 10% of the number of insects, the early warning threshold is adjusted to K / 3.

[0108] This embodiment also provides an additional early warning method, which calculates the insect density in the target area at different time periods based on the insect quantity information, plots the curve of insect density changing with time, statistically identifies the time period with the highest insect density, sends an early warning message to the user, and suggests that the user take pest control measures during the corresponding time period.

[0109] Example 2

[0110] Please refer to Figure 2 As shown, this embodiment provides a real-time monitoring and early warning device for farmland pests based on acoustic detection. The device includes several monitoring units and a central data processing unit. Each monitoring unit includes a microphone array, an audio preprocessing unit, and a local data processing unit. The monitoring units are set up in the target area, i.e., farmland. The microphone array, audio preprocessing unit, and local data processing unit are connected sequentially via wired / wireless means. All local data processing units are connected to the central data processing unit via wired / wireless means.

[0111] Microphone array: Used to collect sound signals in the target area and transmit them to the audio preprocessing unit.

[0112] In this embodiment, a planar array structure is used to design the microphone array. A microphone array contains multiple microphones and is symmetrically distributed on the array plane to ensure effective processing of sound sources from all directions. The construction of a planar array is much simpler and less expensive than that of a spherical array, but its performance is sufficient to meet the needs of detecting insect and frog sounds in farmland environments.

[0113] In this embodiment, the microphone array is arranged in the following ways: uniform grid distribution: each microphone is equidistantly placed in a rectangular grid. This distribution helps the system to uniformly cover the entire farmland and simplifies the complexity of sound source localization; boundary reinforcement distribution: the number of microphones is increased at the boundary of the target area and the number of microphones is reduced at the center. Since the boundary area is the primary area for insects to invade the farmland, this distribution can better capture the sound of insects entering the farmland from the outside; ring distribution: the microphones are arranged in a circular or multiple concentric circles around the center of the target area. This distribution is suitable for farmland areas that are close to a circle.

[0114] Audio preprocessing unit: used to denoise the audio signal, extract the feature vector of the audio signal, and transmit it to the local data processing unit.

[0115] Local data processing unit: Used to compare the feature vector of the sound signal with the standard dataset using a classification algorithm to obtain the insect species information and frog species information corresponding to the sound signal; sample the sound signal according to the insect species information and frog species information to determine the insect quantity information and frog quantity information, and transmit it to the central data processing unit.

[0116] The local data processing unit only needs to process the data of the microphone array it is located in, which has low computational requirements. It can meet the needs by running offline using low-cost hardware. This design reduces the dependence on real-time network connection and ensures stable operation in remote areas with poor network coverage.

[0117] Central data processing unit: Based on frog and insect population information, combined with the insect population carrying capacity, it sets early warning thresholds. When the thresholds are exceeded, it can send messages to users in real time to remind them to carry out pest control operations.

[0118] The central data processing unit collects and integrates the processing results of each local data processing unit, and performs global analysis and decision support. The algorithm designed in this invention requires less computing resources, has a low workload, and has low hardware requirements, which can effectively save costs. At the same time, the central data processing unit can also periodically upload data to the cloud for backup and further analysis.

Claims

1. A method for real-time monitoring and early warning of farmland pests based on acoustic detection, characterized in that, Includes the following steps: (1) Acquire sound signals in the target area; (2) Denoise the sound signal and extract the feature vector of the sound signal; Includes the following sub-steps: (21) Discretize the sound signal A(t) to obtain A(n), and calculate the signal frequency based on A(n): In the formula: A(t) represents the sound pressure level at each time point, and n in A(n) is the index of the sampling point, with a total of N sampling points. For signal frequency, The sampling frequency is set when acquiring sound signals, and k is the index of the sampling point in the frequency domain; (22) The frequency response formula of the filter is as follows: (The filter is used to denoise the sound signal.) In the formula: For the frequency response of the filter, For signal frequency, The cutoff frequency, The order of the filter; (23) After filtering, the denoised audio signal A′(t) is obtained. A fast Fourier transform is performed on the denoised audio signal A′(t) to convert the time domain signal into a frequency domain signal: In the formula: f is the frequency domain signal, A′(t) is the denoised time domain signal, f is the frequency, t is the time, and j is the imaginary unit; (24) From frequency domain signals Extract the main frequency components and amplitude information to obtain the amplitude spectrum. , forming feature vectors The formula is as follows: In the formula: For amplitude spectrum, , ,…, The main frequency components; (3) The feature vector of the sound signal is compared with the standard dataset using a classification algorithm to obtain the insect species information and frog species information corresponding to the sound signal; the sound signal is sampled based on the insect species information and frog species information to determine the insect quantity information and frog quantity information. (4) Set an early warning threshold based on the number of frogs and insects, and issue an alarm when the threshold is exceeded.

2. The method for real-time monitoring and early warning of farmland pests according to claim 1, characterized in that: The filter is a fourth-order filter.

3. The method for real-time monitoring and early warning of farmland pests according to claim 1, characterized in that: The classification algorithms include support vector machine algorithm, random forest algorithm, decision tree algorithm, K-nearest neighbor algorithm, artificial neural network algorithm, and Naive Bayes algorithm.

4. The method for real-time monitoring and early warning of farmland pests according to claim 1, characterized in that: In step (3), when the classification algorithm is the K-nearest neighbor algorithm, the feature vector of the sound signal is compared with the standard dataset to obtain the insect species information and frog species information corresponding to the sound signal, including the following steps: (311) Feature vector of sound signal Each feature value is normalized: In the formula: The feature vector after normalization The 1 eigenvalue, Representative eigenvector The One eigenvalue; (312) Extract the feature vectors of various insect or frog samples from the standard dataset, and perform the normalization process in step (311) to calculate the feature vector of the normalized sound signal. Euclidean distance to the feature vectors of various insect or frog samples in the standard dataset: In the formula: yes Compared with the standard dataset, the first Euclidean distance between eigenvectors The first feature vector in the standard dataset 1 eigenvalue, The feature vector after normalization The 1 eigenvalue, The dimension of the feature vector; (313) Combine all feature vectors in the standard dataset with the feature vectors. The Euclidean distances between them are arranged in order; (314) Given a value of k, select the eigenvector. The k feature vectors with the smallest Euclidean distance between each other are placed into a set containing the corresponding k sample indices. middle, The k samples with the smallest Euclidean distance are selected, and the insect or frog species corresponding to the k samples are then added to the set. middle, For the first The insect or frog species corresponding to each sample; (315) According to the set The feature vectors are determined using majority voting for the insect or frog species in the data. The corresponding insect or frog species information is obtained: In the formula: For indicator functions, when When the result is 1, the result is 0 otherwise.

5. The method for real-time monitoring and early warning of farmland pests according to claim 1, characterized in that: In step (3), the process of sampling sound signals based on insect and frog species information to determine insect and frog quantity information includes the following steps: (321) Based on the insect or frog species information, take the average time T for one call of the insect or frog of that type, and collect the sound signal A(t) at intervals T, for a total of M collections, to obtain the discrete amplitude. : In the formula: For the first The audio signal was collected in this second instance; (322) Transmit the sound signal Convert to digital signal : (323) Based on digital signals Calculate energy : (324) Set an energy threshold; if the energy... If the energy value exceeds the energy threshold, it is determined as an event. The total number of events within M time periods T is calculated to obtain the corresponding insect or frog quantity information.

6. The method for real-time monitoring and early warning of farmland pests according to claim 1, characterized in that: Step (4) includes: When the number of frogs in the area is less than the set value, the warning threshold is set to 1 / 4 of the insect population carrying capacity K. When the warning threshold is exceeded, a warning message is sent to the user. When the number of frogs in the area is greater than or equal to a set value, an early warning threshold is set based on the number of frogs. This early warning threshold is greater than 1 / 4K and less than 1 / 2K.

7. The method for real-time monitoring and early warning of farmland pests according to claim 1, characterized in that: Step (4) further includes: calculating the insect density in the target area at different time periods based on the insect quantity information, statistically determining the time period with the highest insect density, and sending a warning message to the user.

8. A real-time monitoring and early warning device for farmland pests based on acoustic detection, characterized in that, It includes several monitoring units and a central data processing unit. The monitoring unit includes a microphone array, an audio preprocessing unit, and a local data processing unit. The microphone array includes several symmetrically distributed microphones. The microphone array, the audio preprocessing unit, and the local data processing unit are connected in sequence. The local data processing unit is connected to the central data processing unit. Microphone array: Used to collect sound signals in the target area and transmit them to the audio preprocessing unit; Audio preprocessing unit: used to denoise the audio signal, extract the feature vector of the audio signal, and transmit it to the local data processing unit; specifically including: Discretize the sound signal A(t) to obtain A(n), and calculate the signal frequency based on A(n): In the formula: A(t) represents the sound pressure level at each time point, and n in A(n) is the index of the sampling point, with a total of N sampling points. For signal frequency, The sampling frequency is set when acquiring sound signals, and k is the index of the sampling point in the frequency domain; The frequency response formula of a filter is as follows: In the formula: For the frequency response of the filter, For signal frequency, The cutoff frequency, The order of the filter; After filtering, the denoised audio signal A′(t) is obtained. A Fast Fourier Transform is then performed on the denoised audio signal A′(t) to convert the time-domain signal into a frequency-domain signal. In the formula: f is the frequency domain signal, A′(t) is the denoised time domain signal, f is the frequency, t is the time, and j is the imaginary unit; From frequency domain signals Extract the main frequency components and amplitude information to obtain the amplitude spectrum. , forming feature vectors The formula is as follows: In the formula: For amplitude spectrum, , ,…, The main frequency components; Local data processing unit: Used to compare the feature vector of the sound signal with the standard dataset using a classification algorithm to obtain the insect species information and frog species information corresponding to the sound signal; sample the sound signal based on the insect species information and frog species information to determine the insect quantity information and frog quantity information, and transmit it to the central data processing unit; Central data processing unit: Used to set early warning thresholds based on frog and insect count information, and issue an alarm when the thresholds are exceeded.

9. The real-time monitoring and early warning device for farmland pests according to claim 8, characterized in that: The microphone array can be arranged in the following ways: uniform grid distribution: each microphone is equidistantly placed in a rectangular grid; boundary reinforcement distribution: the number of microphones is increased at the boundary of the target area and the number of microphones is reduced at the center; ring distribution: the microphones are arranged in a circular or multiple concentric circles around the center of the target area.

Images