System and method for monitoring slack thread insect pests based on intelligent sensor

Abstract

The invention provides a pine thread pest monitoring system and method based on an intelligent sensor. The method comprises the following steps: firstly, acquiring near infrared spectrum information of each pine tree sample through an intelligent spectrum sensor; further determining a characterization wave band of each pine tree sample; further determining the erosion state of each pine tree sample, and determining the identification quantity of each pine tree sample during pest grading through all erosion states and the characterization wave band of each pine tree sample; then, according to the similarity characteristic quantity of the near infrared spectrum image collected by each intelligent spectrum sensor and the humidity gradient at different sampling points, determining the recognition gain when performing pest qualitative treatment on each pine tree sample; and further determining the erosion trends of the pine nematodes at different sampling points in the pine nematode influence area, and sending all the erosion trends to a pine nematode pest monitoring center. By adopting the scheme of the application, the propagation trend of insect attack behaviors can be monitored at the initial stage of pine nematode infection, so that the accuracy of predicting the diffusion trend of pine nematode at the initial stage is improved.

Description

Technical Field

[0001] The present application relates to the field of near-infrared light detection technology, and more specifically, to a pine nematode pest monitoring system and method based on intelligent sensors. Background Art

[0002] Near-infrared light detection is an analytical technology based on near-infrared spectroscopy. Its working principle is to obtain the composition and property information of a substance by measuring the absorption or reflection characteristics of a substance to near-infrared light. In the existing technology, near-infrared light in the 700-2500 nanometer band can be used to analyze the near-infrared spectral characteristics of plants, which can effectively identify and quantitatively measure the presence and extent of pests. The impact of different pests on plants will cause changes in spectral reflectance characteristics, such as changes in moisture content, pigment distribution, and chemical composition. Combined with machine learning algorithms, a large amount of spectral data can be quickly processed to achieve real-time monitoring and early warning of pests. In addition, near-infrared spectroscopy can also help identify the potential impact of pests on the ecological environment. This non-destructive monitoring method has the advantages of being fast, efficient, and environmentally friendly.

[0003] Pine nematode pest monitoring uses machine learning and deep learning technology to establish a predictive model through a large amount of near-infrared spectral image data to analyze the spread trend and diffusion path of pine nematodes. However, the early symptoms of pine trees infected with pine nematodes are often not obvious. Traditional visual inspections make it difficult to detect pest behavior in the early stages, resulting in monitoring delays and the failure to promptly detect the invasion and spread of pine nematodes. Therefore, how to monitor the spread trend of pest behavior in the early stages of pine nematode infection to improve the accuracy of early predictions of pine nematode diffusion trends and avoid large-scale damage has become a difficult problem faced by the industry. Summary of the invention

[0004] The present application provides a pine nematode pest monitoring system and method based on intelligent sensors, which can monitor the spread trend of pest behavior in the early stage of pine nematode infection, so as to improve the accuracy of early prediction of pine nematode spread trend and prevent large-scale damage.

[0005] In a first aspect, the present application provides a pine nematode pest monitoring method based on an intelligent sensor, comprising the following steps: Setting up smart spectral sensors at different sampling points in the pine nematode-affected areas; Based on the pre-trained machine learning model, the near-infrared spectral information collected by each intelligent spectral sensor is extracted to obtain the characteristic bands of the pine samples at each sampling point under the invasion of pine nematodes; The erosion state of the pine tree samples at each sampling point under the invasion of pine nematodes is determined according to the band differences between all near-infrared spectral information and the distribution characteristic values ​​of the spectral curves of each near-infrared spectral information in the absorption valley. The identification value for pest classification of each pine tree sample is determined through all the erosion states and the characteristic bands of the pine tree samples at each sampling point under the invasion of pine nematodes. Automatically collect near-infrared spectral images of each pine tree sample by intelligent spectral sensors at each sampling point, and determine the recognition gain when qualitatively identifying the pests of each pine tree sample according to the similarity feature quantity of each near-infrared spectral image and the humidity gradient at different sampling points in the pine nematode-affected area; The erosion trend of pine nematodes at different sampling points in the pine nematode-affected area was determined by the recognition gain when each pine sample was subjected to pest qualitative characterization and the discrimination value when each pine sample was subjected to pest grading, and all erosion trends were sent to the pine nematode pest monitoring center.

[0006] In some embodiments, the near-infrared spectral information collected by each intelligent spectral sensor is subjected to band extraction based on a pre-trained machine learning model, and the characterization bands of the pine tree samples at each sampling point under pine nematode invasion are obtained, specifically including: The relevant features of each near-infrared spectrum information were selected to obtain the correlation coefficient of pine nematode invasion; Initialize the pre-trained machine learning model; The key characteristic bands of pests in each near-infrared spectral information are extracted through a pre-trained machine learning model; The characterization bands of the pine tree samples at each sampling point under pine nematode invasion are determined based on the pine nematode invasion correlation coefficient and the key characteristic bands of the pest in each near-infrared spectral information.

[0007] In some embodiments, determining the erosion state of the pine tree sample at each sampling point under the invasion of pine nematodes according to the band differences between all near-infrared spectral information and the distribution characteristic values ​​of the spectral curves of each near-infrared spectral information in the absorption valley specifically includes: Obtain the correlation coefficient of pine nematode invasion; Performing band difference analysis on all near-infrared spectral information by using the pine nematode invasion correlation coefficient to obtain the band differences between all near-infrared spectral information; Analyze the spectral curve of each near-infrared spectral information, and identify the distribution characteristic value of the spectral curve of each near-infrared spectral information in the absorption valley; Perform linear fitting on the distribution characteristic values ​​of all spectral curves in the absorption valley to obtain the distribution characteristic fitting curve of the absorption valley; The pine tree samples at each sampling point are evaluated for pest infestation through the band differences between the distribution characteristic fitting curve and all near-infrared spectral information, and the erosion status of the pine tree samples at each sampling point under the invasion of pine nematodes is obtained.

[0008] In some embodiments, determining the identification value for pest classification of each pine tree sample by characterizing the wavelength bands of all erosion states and pine tree samples at each sampling point under pine nematode invasion specifically includes: Cluster all erosion states to obtain multiple erosion state clusters; Determine the state characteristic value of each erosion state cluster; The identification value for pest grading of each pine tree sample is determined based on all state characteristic values ​​and the absorbance and light intensity corresponding to all wavelengths in the characterization band of the pine tree samples under pine nematode invasion at each sampling point.

[0009] In some embodiments, determining the recognition gain for pest qualitative analysis of each pine tree sample according to the similarity feature value of each near-infrared spectrum image and the humidity gradient at different sampling points in the pine nematode-affected area specifically includes: Perform brightness feature separation on each near-infrared spectrum image to obtain brightness feature values ​​of each near-infrared spectrum image; According to the humidity gradient at different sampling points in the pine nematode-affected area and the brightness characteristic values ​​of each near-infrared spectral image, the collinearity characteristic values ​​of each near-infrared spectral image corresponding to the pine tree sample were determined; The recognition gain when performing pest qualitative analysis on each pine tree sample is determined by the similarity feature quantity of each near-infrared spectrum image and the collinearity feature quantity of each pine tree sample.

[0010] In some embodiments, determining the pine nematode erosion trend at different sampling points in the pine nematode-affected area through the recognition gain when each pine tree sample is subjected to pest qualitative evaluation and the discrimination value when each pine tree sample is subjected to pest classification specifically includes: Select a sampling point as the selected sampling point; The auxiliary scale for evaluating pest invasion at the selected sampling point was determined based on the recognition gain of pine tree samples from adjacent sampling points when conducting pest qualitative analysis; The trend characteristic value of pest invasion at the selected sampling point is determined by the identification value of pine tree samples at adjacent sampling points during pest classification; The erosion trend of pine nematodes at the selected sampling points in the pine nematode-affected area is determined based on the trend characteristic value of the pest invasion at the selected sampling points and the auxiliary scale of the pest invasion at the selected sampling points; Continue to determine the pine nematode encroachment trends at the remaining sampling sites in the pine nematode-affected area.

[0011] In some embodiments, the smart spectral sensor is a device that integrates the collection and analysis of spectral data of an object within different wavelength ranges.

[0012] In a second aspect, the present application provides a pine nematode pest monitoring system based on an intelligent sensor, comprising: A setting module for setting up intelligent spectral sensors at different sampling points in the pine nematode-affected area; A processing module is used to extract the near-infrared spectral information collected by each intelligent spectral sensor based on a pre-trained machine learning model to obtain the characterization bands of the pine tree samples at each sampling point under the invasion of pine nematodes; The processing module is further used to determine the erosion state of the pine tree samples at each sampling point under the invasion of pine nematodes according to the band differences between all near-infrared spectral information and the distribution characteristic values ​​of the spectral curves of each near-infrared spectral information in the absorption valley, and determine the identification value when performing pest classification on each pine tree sample through all the erosion states and the characteristic bands of the pine tree samples at each sampling point under the invasion of pine nematodes; The processing module is also used to automatically collect near-infrared spectral images of each pine tree sample through intelligent spectral sensors at each sampling point, and determine the recognition gain when performing pest qualitative analysis on each pine tree sample according to the similarity feature quantity of each near-infrared spectral image and the humidity gradient at different sampling points in the pine nematode-affected area; The execution module is used to determine the pine nematode erosion trend at different sampling points in the pine nematode affected area through the recognition gain when each pine tree sample is subjected to pest qualitative characterization and the identification amount when each pine tree sample is subjected to pest grading, and send all erosion trends to the pine nematode pest monitoring center.

[0013] In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory stores codes, and the processor is configured to obtain the codes and execute the above-mentioned smart sensor-based pine nematode pest monitoring method.

[0014] In a fourth aspect, the present application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned pine nematode pest monitoring method based on smart sensors.

[0015] The technical solution provided by the embodiments disclosed in this application has the following beneficial effects: In the pine nematode pest monitoring system and method based on smart sensors provided in the present application, firstly, smart spectral sensors are set at different sampling points in the pine nematode-affected area; based on the pre-trained machine learning model, band extraction is performed on the near-infrared spectral information collected by each smart spectral sensor to obtain the characterization band of the pine tree samples at each sampling point under the invasion of pine nematodes; according to the band differences between all the near-infrared spectral information and the distribution characteristic values ​​of the spectral curves of each near-infrared spectral information in the absorption valley, the erosion state of the pine tree samples at each sampling point under the invasion of pine nematodes is determined; and the erosion state of the pine tree samples at each sampling point under the invasion of pine nematodes is determined by all the erosion states and the pine tree samples at each sampling point under the invasion of pine nematodes. The characterization band under nematode invasion determines the identification value when grading the pests of each pine tree sample; the near-infrared spectral image of each pine tree sample is automatically collected by the intelligent spectral sensor at each sampling point, and the recognition gain when qualitatively identifying the pests of each pine tree sample is determined according to the similarity feature value of each near-infrared spectral image and the humidity gradient at different sampling points in the pine nematode-affected area; the erosion trend of pine nematodes at different sampling points in the pine nematode-affected area is determined by the recognition gain when qualitatively identifying the pests of each pine tree sample and the identification value when grading the pests of each pine tree sample, and all the erosion trends are sent to the pine nematode pest monitoring center.

[0016] It can be seen that in the present application, the erosion trend of pine nematodes at different sampling points in the pine nematode-affected area can be determined based on the recognition gain when each pine tree sample is subjected to pest qualitative identification and the identification amount when each pine tree sample is subjected to pest grading; wherein, firstly, the humidity gradient at different sampling points in the pine nematode-affected area and the brightness characteristic values ​​of each near-infrared spectral image are used to extract the collinear feature quantity used to describe the linear correlation characteristics between the brightness characteristic value of each near-infrared spectral image and the humidity gradient at the sampling point, and then the similarity feature quantity of each near-infrared spectral image and the collinear feature quantity of each pine tree sample are used to obtain the recognition gain when each pine tree sample is subjected to pest qualitative identification. Since the recognition gain can be used to perform pest qualitative identification on the pine tree samples, the recognition gain can be used to perform pest qualitative identification on the pine tree samples. In the early stage of pine nematode infection, the spread trend of the pest is often hidden, manifested as a slight change in the humidity gradient or a slight fluctuation in the spectral characteristic value. By identifying the gain, these slight changes in the environment and spectral characteristics can be integrated to more accurately capture the initial spread of the pest. Secondly, the spectral curve of the near-infrared spectral information of each pine sample is analyzed, and the distribution characteristic values ​​of the spectral curve of each near-infrared spectral information in the absorption valley are identified. Since the distribution characteristic values ​​represent the characteristic values ​​reflecting the slight changes in the chemical composition of the pine tissue in the early stage of pine nematode infection, these changes will be manifested in the absorption valley characteristics of the near-infrared spectrum. Through the distribution characteristic values ​​of the absorption valley, it is possible to sensitively These early changes can be captured sensitively, and then all near-infrared spectral information is analyzed for band differences to obtain the band differences between all near-infrared spectral information. The band differences can reflect the health status of pine trees when attacked by pests. The band differences between all near-infrared spectral information and the distribution characteristic values ​​of the spectral curve of each near-infrared spectral information in the absorption valley are further quantified to obtain the erosion state that describes each pine sample under the invasion of pine nematodes, which causes the change in the physiological state of the pine sample. Therefore, the identification value of each pine sample during pest classification is determined through all the erosion states and the characterization bands of each pine sample under the invasion of pine nematodes. By monitoring the identification values ​​of multiple pine samples, the spread trend of pest behavior can be analyzed. The spread path and speed of the pest are determined, and the accuracy of the early pest spread prediction model is enhanced. Then, the erosion trend of pine nematodes at different sampling points in the pine nematode-affected area is quantified by quantifying the recognition gain when the pest is qualitatively characterized for each pine sample and the identification amount when the pest is graded for each pine sample. Since the erosion trend represents the parameters of the dynamic spread trend of pine nematodes initially infected in pine samples at the sampling points in space, finally, all erosion trends are sent to the pine nematode pest monitoring center and pest trend monitoring is carried out. In summary, the scheme of the present application can monitor the spread trend of pest behavior in the early stage of pine nematode infection, thereby improving the accuracy of the initial prediction of the spread trend of pine nematodes and avoiding large-scale damage. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is an exemplary flow chart of a pine nematode pest monitoring method based on a smart sensor according to some embodiments of the present application; Figure 2 is a schematic diagram of a process for determining a characterization band in some embodiments of the present application; Figure 3 is a schematic diagram of a flow chart of determining the identification amount when grading pests in some embodiments of the present application; Figure 4 is a schematic diagram of the structure of a pine nematode pest monitoring system based on an intelligent sensor in some embodiments of the present application; Figure 5 It is a structural schematic diagram of a computer device for implementing a pine nematode pest monitoring method based on a smart sensor as shown in some embodiments of the present application. DETAILED DESCRIPTION

[0018] In order to better understand the technical solution of the present application, the technical solution of the present application will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0019] refer to Figure 1 , which is an exemplary flow chart of a pine nematode pest monitoring method based on a smart sensor according to some embodiments of the present application. The pine nematode pest monitoring method 100 based on a smart sensor mainly includes the following steps: In step 101, intelligent spectral sensors are set at different sampling points in the area affected by pine nematodes.

[0020] In specific implementation, the pine nematode-affected area can be divided into multiple sub-areas, and then the pine trees in each sub-area are randomly sampled, and the pine trees obtained after random sampling are used as a sampling point, and intelligent spectral sensors are set at each sampling point.

[0021] It should be noted that the intelligent spectral sensor described in this application is a device that integrates the collection and analysis of spectral data of objects in different wavelength ranges. It combines traditional spectral sensors and intelligent algorithms, and can automatically identify and judge the characteristics, status or changes of target objects through analysis of spectral data.

[0022] In step 102, band extraction is performed on the near-infrared spectral information collected by each intelligent spectral sensor based on a pre-trained machine learning model to obtain the characterization bands of the pine tree samples at each sampling point under the invasion of pine nematodes.

[0023] In some embodiments, reference Figure 2As shown, this figure is a schematic diagram of the process of determining the characterization band in some embodiments of the present application. In this embodiment, the near-infrared spectral information collected by each intelligent spectral sensor is extracted based on the pre-trained machine learning model, and the characterization band of the pine tree sample at each sampling point under the invasion of pine nematodes can be obtained by the following steps: First, in step 1021, relevant features are selected for each near infrared spectrum information to obtain a correlation coefficient of pine nematode invasion; Secondly, in step 1022, the pre-trained machine learning model is initialized; Then, in step 1023, the key characteristic bands of the pests in each near infrared spectrum information are extracted through the pre-trained machine learning model; Finally, in step 1024, the characterization bands of the pine tree samples at each sampling point under pine nematode invasion are determined based on the pine nematode invasion correlation coefficient and the key characteristic bands of the pest in each near-infrared spectrum information.

[0024] It should be noted that in the present application, the near-infrared spectral information of the pine samples is collected by the intelligent spectral sensors at each sampling point, wherein the pine samples are root segments of the pine roots, and the near-infrared spectral information is information consisting of multiple wavelengths, multiple absorbances and multiple spectral intensities, wherein the absorbance indicates the degree of absorption of light of a specific wavelength, and the spectral intensity indicates the response intensity to light of a specific wavelength, and at the same time, each wavelength corresponds to an absorbance and a spectral intensity.

[0025] In specific implementation, relevant feature selection is performed on each near-infrared spectral information to obtain multiple pine nematode invasion correlation coefficients, which can be achieved in the following manner, namely: first, near-infrared spectral information of historical pine tree samples invaded by pine nematodes is collected and obtained, and the obtained infrared spectral information is used as historical infrared spectral information, and the Pearson correlation coefficient is calculated for the absorbance corresponding to each wavelength in the historical infrared spectral information and the absorbance corresponding to the same wavelength in each near-infrared spectral information, so as to obtain the Pearson correlation coefficient of absorbance between the historical infrared spectral information and each near-infrared spectral information, and further hierarchical clustering is performed on all the Pearson correlation coefficients, and the cluster center of the largest cluster obtained after hierarchical clustering is used as the pine nematode invasion correlation coefficient. In other embodiments, other methods can also be used for implementation, which are not limited here.

[0026] It should be noted that the pine nematode invasion correlation coefficient described in the present application represents the parameter value between the wavelength absorbance related to the pine nematode invasion status in the near-infrared spectral information. The most relevant features can be screened out through the pine nematode invasion correlation coefficient, thereby constructing a more accurate pine nematode invasion prediction model and reducing redundant information.

[0027] In addition, it should be noted that the machine learning model initialized for pre-training in the present application can collect near-infrared spectral data of historical pine tree samples, including healthy samples and pine nematode invasion samples. Then, all the near-infrared spectral data are used to train the machine learning model (for example, classification models such as support vector machines, random forests and convolutional neural networks, which are not limited here), and the model obtained after the training is used as the pre-trained machine learning model. The machine learning model can evaluate which bands contribute the most to the classification results based on the importance of features such as absorbance and spectral intensity, that is, the output result of the machine learning model is a band, where the band represents the range of wavelengths, which can be divided into an upper wavelength bound and a lower wavelength bound, and the range composed of the upper wavelength bound and the lower wavelength bound is used as a band. In other embodiments, other methods can also be used to achieve this, which will not be repeated here.

[0028] In specific implementation, the key characteristic bands of pests in each near-infrared spectral information can be extracted through a pre-trained machine learning model in the following manner, namely: using a pre-trained machine learning model to extract the band that contributes the most to the classification result of each near-infrared spectral information, and using the obtained band as the key characteristic band of pests in the corresponding near-infrared spectral information. In other embodiments, other methods can also be used for implementation, which are not limited here.

[0029] It should be noted that each key characteristic band in this application corresponds to a pine tree sample at a sampling point.

[0030] In specific implementation, the characterization bands of the pine samples at each sampling point under pine nematode invasion are determined according to the pine nematode invasion correlation coefficient and the key characteristic bands of the pests in each near-infrared spectral information, which can be implemented in the following manner, namely: a key characteristic band of the pests in the near-infrared spectral information is selected as the selected key characteristic band, the upper wavelength limit and the lower wavelength limit of the selected key characteristic band are multiplied by the pine nematode invasion correlation coefficient, and the multiplied values ​​are respectively used as the upper wavelength limit and the lower wavelength limit of the invasion-related wavelength of the pine nematode, and the range composed of the upper wavelength limit and the lower wavelength limit of the invasion-related wavelength is further used as the characterization band of the pine samples at the sampling points corresponding to the selected key characteristic band under pine nematode invasion, and the characterization bands of the pine samples at the sampling points corresponding to the remaining key characteristic bands are continued to be determined. In other embodiments, other methods can also be used for implementation, which will not be repeated here.

[0031] In step 103, the erosion state of the pine samples at each sampling point under the invasion of pine nematodes is determined based on the band differences between all near-infrared spectral information and the distribution characteristic values ​​of the spectral curves of each near-infrared spectral information in the absorption valley, and the identification value for pest grading of each pine sample is determined through all the erosion states and the characterization bands of the pine samples at each sampling point under the invasion of pine nematodes.

[0032] In some embodiments, the following steps can be used to determine the erosion state of the pine tree samples at each sampling point under the invasion of pine nematodes based on the band differences between all near-infrared spectral information and the distribution characteristic values ​​of the spectral curves of each near-infrared spectral information in the absorption valley: Obtain the correlation coefficient of pine nematode invasion; Performing band difference analysis on all near-infrared spectral information by using the pine nematode invasion correlation coefficient to obtain the band differences between all near-infrared spectral information; Analyze the spectral curve of each near-infrared spectral information, and identify the distribution characteristic value of the spectral curve of each near-infrared spectral information in the absorption valley; Perform linear fitting on the distribution characteristic values ​​of all spectral curves in the absorption valley to obtain the distribution characteristic fitting curve of the absorption valley; The pine tree samples at each sampling point are evaluated for pest infestation through the band differences between the distribution characteristic fitting curve and all near-infrared spectral information, and the erosion status of the pine tree samples at each sampling point under the invasion of pine nematodes is obtained.

[0033] In specific implementation, all near-infrared spectral information is subjected to band difference analysis through the pine nematode invasion correlation coefficient, and the band difference between all near-infrared spectral information can be obtained by the following method, namely: the standard deviation of the absorbance in all near-infrared spectral information at the same wavelength is calculated, so as to obtain the standard deviation of the absorbance at different wavelengths, and then all the obtained standard deviations are multiplied by the pine nematode invasion correlation coefficient, and the maximum value and the minimum value are further selected from all the multiplied values, so as to subtract the minimum value from the maximum value, and the subtracted value is used as the band difference between all near-infrared spectral information. In other embodiments, other methods can also be used for implementation, which are not limited here.

[0034] It should be noted that the band difference described in the present application represents the difference in absorbance of the near-infrared spectrum at a specific wavelength, and the band difference can reflect the health status of the pine trees when attacked by insect pests.

[0035] In specific implementation, the spectral curve of each near-infrared spectral information is analyzed, and the distribution characteristic value of the spectral curve of each near-infrared spectral information in the absorption valley is identified. The following method can be used to achieve this, namely: select a near-infrared spectral information as the selected near-infrared spectral information, draw a spectral curve for the selected near-infrared spectral information with wavelength as the horizontal axis and absorbance as the vertical axis, and then use a smoothing algorithm (such as a Savitzky-Golay filter, which is not limited here) to smooth the spectral curve. The smoothing process can reduce noise and make the spectral curve smoother, thereby calculating the first-order derivative of the spectral curve, and finding points where the derivative is zero and the second-order derivative is greater than zero. These points are the local minima of the spectral curve, indicating the position that may correspond to the absorption valley. Further, a smoothing window is used to scan along the spectral curve, and the absorbance of each point is compared with that of its neighboring points. The local minimum point (i.e., the minimum value) is identified as the absorption valley, thereby obtaining multiple absorption valleys, and an absorption valley is selected as the selected absorption valley. First, the absorbance threshold is calculated. The absorbance threshold is a certain proportion of the valley bottom absorbance value of the selected absorption valley. The following formula is used for determination, namely: absorbance threshold value = minimum absorbance value at the selected absorption valley + 0.5 × (background absorbance on both sides of the selected absorption valley - minimum absorbance value at the selected absorption valley), and then starting from the bottom position of the selected absorption valley, the wavelength is scanned step by step to the left (reducing the wavelength) to find the first position where the absorbance rises to exceed the absorbance threshold, and the wavelength at this position is used as the left wavelength. Similarly, starting from the bottom position of the selected absorption valley, the wavelength is scanned step by step to the right (increasing the wavelength value) to find the first position where the absorbance rises to or exceeds the absorbance threshold, and the wavelength at this position is used as the right wavelength, and then the right wavelength is subtracted from the left wavelength, and the subtraction result is used as the absorption valley width of the selected absorption valley, and the absorption valley widths of the remaining absorption valleys are continued to be determined, and the sum of all the absorption valley widths is used as the distribution characteristic value of the spectral curve of the selected near-infrared spectral information in the absorption valley, and the distribution characteristic values ​​of the spectral curve of the remaining near-infrared spectral information in the absorption valley are continued to be determined. In other embodiments, other methods can also be used to achieve this, which will not be repeated here.

[0036] It should be noted that the smoothing window described in this application can select an initial window size of 3 to 5 times the spectral resolution. For example, if the spectral resolution is 1 nm, the initial window can be set to 3-5 nm.

[0037] In addition, it should be noted that the distribution characteristic values ​​described in the present application represent characteristic values ​​reflecting slight changes in the chemical composition of pine tissues in the early stages of pine nematode infection. These changes will be manifested in the absorption valley characteristics of the near-infrared spectrum. Through the distribution characteristic values ​​of the absorption valley, these early changes can be sensitively captured, and each distribution characteristic value corresponds to a near-infrared spectrum information.

[0038] In specific implementation, linear fitting is performed on the distribution characteristic values ​​of all spectral curves in the absorption valley to obtain the distribution characteristic fitting curve of the absorption valley, which can be achieved in the following manner, namely: the distribution characteristic values ​​of all spectral curves in the absorption valley are sorted in ascending order, and then the least squares support vector machine algorithm is used to fit the sequence obtained after sorting, and the fitted curve is used as the distribution characteristic fitting curve of the absorption valley. In other embodiments, other fitting algorithms can also be used, which is not limited here.

[0039] It should be noted that in the present application, insect infestation will cause changes in the physiological state of the pine samples, and this change will be reflected in the spectral intensity. In addition, in the near-infrared spectrum, the spectral curve of the pine samples will show multiple absorption valleys, especially in the bands related to water content, chlorophyll and other chemical components. When the plants are infested by insect pests, the depth and width of the absorption valleys will change, resulting in changes in the spectral intensity. These changes can be used to evaluate the impact of insect pests.

[0040] In a specific implementation, the pine tree samples at each sampling point are evaluated for pest infestation through the band difference between the distribution characteristic fitting curve and all near-infrared spectral information, and the erosion state of the pine tree samples at each sampling point under the invasion of pine nematodes can be obtained by the following method, namely: a sampling point is selected as the selected sampling point, and the spectral intensities with the same number of absorption valleys as the near-infrared spectral information of the pine tree sample at the selected sampling point are selected from the near-infrared spectral information of the pine tree sample at the selected sampling point in order from large to small spectral intensities, and then all spectral intensities are summed, and the summed value is used as the global spectral intensity, and the derivative of the distribution characteristic fitting curve is further obtained to obtain multiple derivatives, wherein each derivative corresponds to A distribution characteristic value is selected, and the distribution characteristic value corresponding to the near-infrared spectrum information of the pine sample at the selected sampling point is used as the selected distribution characteristic value. First, the derivative corresponding to the selected distribution characteristic value is taken as a negative exponential function with the natural logarithm e as the base, and then the value obtained after calculating the negative exponential function with the natural logarithm e as the base is multiplied by the selected distribution characteristic value, and then the multiplied value is added to the global spectral intensity, and then the added value is divided by the band difference between all near-infrared spectrum information, and the divided value is used as the erosion state of the pine sample at the selected sampling point under the invasion of pine nematodes, and the erosion state of the pine samples at the remaining sampling points under the invasion of pine nematodes is continued to be determined. In other embodiments, other methods can also be used to achieve this, which is not limited here.

[0041] It should be noted that the erosion status described in the present application represents an indicator of the growth status of pine tree samples attacked by pine nematodes. By evaluating the erosion status of multiple trees in an area, it can help predict the spread of pests.

[0042] In some embodiments, reference Figure 3As shown, this figure is a schematic diagram of the process of determining the identification amount when the pest is graded in some embodiments of the present application. In this embodiment, the identification amount when each pine sample is graded for pests is determined by characterizing the bands of all erosion states and pine samples at each sampling point under pine nematode invasion, which can be achieved by the following steps: First, in step 1031, all erosion states are clustered to obtain multiple erosion state clusters; Next, in step 1032, the state characteristic values ​​of each erosion state cluster are determined; Then, in step 1033, the identification value for pest classification of each pine tree sample is determined based on all state characteristic values ​​and the absorbance and light intensity corresponding to all wavelengths in the characterization band of the pine tree sample under pine nematode invasion at each sampling point.

[0043] In the specific implementation, all erosion states are clustered to obtain multiple erosion state clusters, which can be achieved in the following manner: first, the erosion level can be divided into 4 levels, namely, healthy (no erosion), mild erosion, moderate erosion and severe erosion; secondly, the number of 4 clusters is specified, and K-means clustering is used to cluster all erosion states to obtain multiple clusters, and the obtained clusters are all used as erosion state clusters. In other embodiments, other methods can also be used to achieve this, which will not be repeated here.

[0044] In specific implementation, the state characteristic value of each erosion state cluster can be determined in the following manner, namely: the erosion state of each erosion state cluster is divided by all erosion states, and the value obtained by the division is used as the state weight of the corresponding erosion state cluster, and then the obtained state weight is multiplied by the value of the cluster center of each erosion state cluster, and the multiplied value is used as the state characteristic value of the corresponding erosion state cluster. In other embodiments, other methods can also be used for implementation, which are not limited here.

[0045] It should be noted that the state characteristic value described in the present application represents the parameter value for identifying the health state (healthy, lightly eroded, moderately eroded, severely eroded) of the pine tree sample. In addition, each pine tree sample corresponds to a state characteristic value.

[0046] In specific implementation, the identification amount for pest grading of each pine sample is determined according to all state characteristic values ​​and the absorbance and light intensity corresponding to all wavelengths in the characterization band of the pine sample at each sampling point under pine nematode invasion. The method can be adopted as follows, namely: a sampling point is selected as the selected sampling point, the absorbance and light intensity corresponding to the same wavelength in the characterization band of the pine sample at the selected sampling point under pine nematode invasion are divided, and then all the divided values ​​are summed, and the summed value is further multiplied by the state characteristic value corresponding to the pine sample at the selected sampling point, and the multiplied value is used as the identification amount for pest grading of the pine sample at the selected sampling point, and the identification amount for pest grading of the pine samples at the remaining sampling points is continued to be determined. In other embodiments, other methods can also be used for implementation, which will not be repeated here.

[0047] It should be noted that the identification value described in the present application represents a parameter for evaluating the degree of pine nematode infection when grading pine tree samples. By monitoring the identification values ​​of multiple pine tree samples, the spread trend of pest behavior can be analyzed to clarify the spread path and speed of the pest, while enhancing the accuracy of the early pest spread prediction model.

[0048] In step 104, the near-infrared spectral image of each pine tree sample is automatically collected by the intelligent spectral sensor at each sampling point, and the recognition gain for pest qualitative analysis of each pine tree sample is determined based on the similarity feature of each near-infrared spectral image and the humidity gradient at different sampling points in the pine nematode-affected area.

[0049] In the specific implementation, the near-infrared spectral images of the pine tree samples at each sampling point are automatically collected through the intelligent spectral sensor.

[0050] In some embodiments, the recognition gain for qualitatively identifying the pests of each pine tree sample is determined based on the similarity feature of each near-infrared spectral image and the humidity gradient at different sampling points in the pine nematode-affected area, and the following steps can be used: Perform brightness feature separation on each near-infrared spectrum image to obtain brightness feature values ​​of each near-infrared spectrum image; According to the humidity gradient at different sampling points in the pine nematode-affected area and the brightness characteristic values ​​of each near-infrared spectral image, the collinearity characteristic values ​​of each near-infrared spectral image corresponding to the pine tree sample were determined; The recognition gain when performing pest qualitative analysis on each pine tree sample is determined by the similarity feature quantity of each near-infrared spectrum image and the collinearity feature quantity of each pine tree sample.

[0051] In specific implementation, brightness feature separation is performed on each near-infrared spectrum image, and the brightness feature value of each near-infrared spectrum image can be obtained by the following method, namely: select a near-infrared spectrum image as the selected near-infrared spectrum image, first, convert the selected near-infrared spectrum image into a grayscale image, and use the grayscale value in the grayscale image as the brightness value, and secondly, count the brightness values ​​of all pixels in the grayscale image into a histogram, and the brightness value can be divided into 256 intervals (from 0 to 255), and then calculate the number of pixels in each interval to form a brightness distribution histogram, and then use a global threshold (such as Otsu algorithm, which is not limited here) to determine the segmentation point according to the brightness distribution histogram, divide the pixels into low, medium and high brightness areas, calculate the average brightness and brightness standard deviation of each brightness area, divide the average brightness of each brightness area by the brightness standard deviation of each brightness area, and then sum all the values ​​obtained by the division, and use the summed value as the brightness feature value of the selected near-infrared spectrum image, and continue to determine the brightness feature values ​​of the remaining near-infrared spectrum images. In other embodiments, other methods can also be used to implement, which will not be repeated here.

[0052] It should be noted that the brightness characteristic value of the near-infrared spectral image in the present application represents the statistical characteristics obtained after quantifying the brightness of each pixel in the infrared spectral image. The brightness characteristic value can be associated with the health status of the tree and potential pest behavior through the brightness changes of the pine tree samples under the infrared spectrum.

[0053] In addition, it should be noted that in the present application, pine nematode infection can cause abnormal tissue metabolism of pine trees, resulting in damage to the water transport system of the trees, thereby causing changes in the temperature and infrared spectral characteristics of leaves or trunks. In infrared spectral images, this change often manifests itself as a decrease in brightness or an uneven distribution. By analyzing the brightness characteristic values, these initial abnormalities can be detected.

[0054] In specific implementation, the collinearity characteristic quantity of each near-infrared spectral image corresponding to the pine tree sample is determined according to the humidity gradient at different sampling points in the pine nematode affected area and the brightness characteristic value of each near-infrared spectral image. The method can be implemented as follows: first, the humidity value is measured at different sampling points in the pine nematode affected area, and the humidity gradient of each sampling point is recorded, and then the mean of all humidity gradients and the mean of all brightness characteristic values ​​are calculated, and the obtained values ​​are used as the humidity gradient equilibrium value and the brightness characteristic value equilibrium value respectively, and the Pearson correlation coefficient is further calculated for the humidity gradient equilibrium value, the brightness characteristic value equilibrium value, the humidity gradient at each sampling point, and the brightness characteristic value of the near-infrared spectral image of the pine tree sample at each sampling point, and The obtained values ​​are all used as the humidity correlation coefficients of the near-infrared spectrum images of the pine tree samples at each sampling point. Secondly, a decision tree model is constructed according to the brightness characteristic values ​​of the near-infrared spectrum images of all pine tree samples and the humidity gradients at the corresponding sampling points, wherein the decision tree model outputs the degree of influence of the brightness characteristic value of each near-infrared spectrum image on the humidity gradient (feature importance score). Then, the humidity correlation coefficient of each near-infrared spectrum image is multiplied by the degree of influence of the brightness characteristic value of each near-infrared spectrum image on the humidity gradient, and the multiplied value is used as the collinear feature quantity of the pine tree sample corresponding to each near-infrared spectrum image. In other embodiments, other methods can also be used for implementation, which are not limited here.

[0055] It should be noted that the collinear feature quantity in the present application represents the feature quantity of the linear correlation between the brightness feature value of the near-infrared spectral image of the pine sample and the humidity gradient at the sampling point. The smaller the collinear feature quantity, the weaker the resistance of the pine tree to pests, and the pests will spread rapidly, and vice versa.

[0056] In specific implementation, the recognition gain of each pine tree sample for pest qualitative identification is determined by the similarity feature of each near-infrared spectral image and the collinearity feature of each pine tree sample, which can be implemented in the following manner, namely: first, the collinearity feature of all pine tree samples is summed, and the summed value is used as the global collinearity feature, and then a near-infrared spectral image is selected as the selected near-infrared spectral image, and then 1 plus the influence of the brightness feature value of the selected near-infrared spectral image on the humidity gradient (feature importance score) is used, and the sum is used as the influence correction factor, and further, the collinearity feature of the pine tree sample corresponding to the selected near-infrared spectral image is divided by the global collinearity feature, and the value obtained by the division is multiplied by the similarity feature of the selected near-infrared spectral image, and then the multiplied value is divided by the influence correction factor, and the divided value is used as the recognition gain of the selected near-infrared spectral image corresponding to the pine tree sample for pest qualitative identification, and the recognition gain of the remaining near-infrared spectral images corresponding to the pine tree samples for pest qualitative identification is continuously determined. In other embodiments, other methods can also be used for implementation, which will not be repeated here.

[0057] It should be noted that the recognition gain described in the present application represents a measure of the sensitivity of judging the pest spread when conducting qualitative analysis of pine samples. In the early stages of pine nematode infection, the spread of pests is often more concealed, manifested as slight changes in humidity gradients or slight fluctuations in spectral characteristic values. The role of the recognition gain is to integrate these slight changes in the environment and spectral characteristics, improve the measure of sensitivity for judging the spread of pests, and thus more accurately capture the initial signs of pest spread.

[0058] In step 105, the erosion trend of pine nematodes at different sampling points in the pine nematode-affected area is determined by the recognition gain of each pine tree sample when performing pest qualitative analysis and the discrimination value of each pine tree sample when performing pest classification, and all erosion trends are sent to the pine nematode pest monitoring center.

[0059] In some embodiments, the following steps can be used to determine the pine nematode erosion trend at different sampling points in the pine nematode-affected area by using the recognition gain when each pine tree sample is subjected to pest qualitative analysis and the discrimination value when each pine tree sample is subjected to pest classification: Select a sampling point as the selected sampling point; The auxiliary scale for evaluating pest invasion at the selected sampling point was determined based on the recognition gain of pine tree samples from adjacent sampling points when conducting pest qualitative analysis; The trend characteristic value of pest invasion at the selected sampling point is determined by the identification value of pine tree samples at adjacent sampling points during pest classification; The erosion trend of pine nematodes at the selected sampling points in the pine nematode-affected area is determined based on the trend characteristic value of the pest invasion at the selected sampling points and the auxiliary scale of the pest invasion at the selected sampling points; Continue to determine the pine nematode encroachment trends at the remaining sampling sites in the pine nematode-affected area.

[0060] It should be noted that the number of adjacent sampling points of the selected sampling point in the present application can be determined by the following method: taking the square root of the total number of all sampling points and rounding down, and using the value obtained after rounding down as the number of local sampling points, and then selecting the sampling points that are closest to the selected sampling point and have the same number as the local sampling points, and all the selected sampling points are used as adjacent sampling points.

[0061] In a specific implementation, the auxiliary scale for evaluating the invasion of pests at the selected sampling point can be determined according to the recognition gain of the pine tree samples at the adjacent sampling points of the selected sampling point when performing pest qualitative analysis. The following method can be used, namely: the maximum recognition gain and the minimum recognition gain are selected from the recognition gains of the pine tree samples at the adjacent sampling points of the selected sampling point when performing pest qualitative analysis, the maximum recognition gain is subtracted from the minimum recognition gain, and the subtraction value is used as the auxiliary scale for evaluating the invasion of pests at the selected sampling point. In other embodiments, other methods can also be used for implementation, which will not be repeated here.

[0062] It should be noted that the auxiliary scale in the present application represents an indicator used to assist in evaluating the pest invasion intensity at selected sampling points in the pine nematode infection area. By analyzing the recognition gain values ​​near the selected sampling points, it can help determine the local change characteristics of pest transmission.

[0063] In specific implementation, determining the trend characteristic value of pest invasion at the selected sampling point by using the identification amount of pine tree samples at adjacent sampling points at the selected sampling point during pest grading can be implemented in the following manner, namely: selecting the largest identification amount and the smallest identification amount from the trend characteristic value of pest invasion at the selected sampling point determined by using the identification amount of pine tree samples at adjacent sampling points at the selected sampling point during pest grading, subtracting the smallest identification amount from the largest identification amount, and using the subtracted value as the trend characteristic value of pest invasion at the selected sampling point. Other methods may also be used in other embodiments, which are not limited here.

[0064] It should be noted that the trend characteristic value described in the present application represents a characteristic value used to evaluate the diffusion trend of the selected sampling points in the pine nematode infection area in the early stage of infection. By analyzing the diffusion trend of neighboring sampling points, the spatial propagation pattern of the infection behavior can be captured.

[0065] In addition, it should be noted that the erosion trend represents the parameter of the dynamic diffusion trend of pine nematodes in the space of the pine samples initially infected with pine nematodes at the sampling points. As a preferred embodiment, the erosion trend of pine nematodes at the selected sampling points in the pine nematode-affected area is determined according to the trend characteristic value of the pest invasion at the selected sampling points and the auxiliary scale of the pest invasion at the selected sampling points. This can be achieved in the following manner, namely: multiplying the trend characteristic value of the pest invasion at the selected sampling points by the auxiliary scale of the pest invasion at the selected sampling points, and using the multiplied value as the erosion trend of pine nematodes at the selected sampling points in the pine nematode-affected area. Other methods may also be used in other embodiments, which will not be described in detail here.

[0066] In specific implementation, the pine nematode pest monitoring center can monitor the pest trend of pine trees in all erosion trend pine nematode-affected areas. For example: first, the gradient of the trend scale at each sampling point is calculated, and the obtained gradient is used as the trend scale gradient value at each sampling point. Then, the maximum gradient ascent method is used to find the path with the most significant gradient change, and the obtained path is used as the potential direction path for the spread of pests in the pine nematode-affected area, so that pest control work can be carried out on the potential direction path for the spread of pests (for example: setting up isolation belts, applying pesticides, etc.), so as to curb the spread and spread of pests. Other methods can also be used in other embodiments, which will not be repeated here.

[0067] In addition, in another aspect of the present application, in some embodiments, the present application provides a pine nematode pest monitoring system based on an intelligent sensor, referring to Figure 4 , which is a schematic diagram of the structure of a pine wire pest monitoring system based on a smart sensor according to some embodiments of the present application. The pine wire pest monitoring system based on a smart sensor 400 includes: a setting module 401, a processing module 402 and an execution module 403, which are described as follows: Setting module 401, in the present application, detection module 401 is mainly used to set intelligent spectral sensors at different sampling points in the area affected by pine nematodes; Processing module 402, in the present application, is used to extract bands of near-infrared spectral information collected by each intelligent spectral sensor based on a pre-trained machine learning model, and obtain the characterization bands of the pine tree samples at each sampling point under the invasion of pine nematodes; It should be noted that the processing module 402 in the present application is also used to determine the erosion state of the pine tree samples at each sampling point under the invasion of pine nematodes according to the band differences between all near-infrared spectral information and the distribution characteristic values ​​of the spectral curves of each near-infrared spectral information in the absorption valley, and determine the identification value when performing pest classification on each pine tree sample through all the erosion states and the characterization bands of the pine tree samples at each sampling point under the invasion of pine nematodes; In addition, it should be noted that the processing module 402 in the present application is also used to automatically collect near-infrared spectral images of each pine tree sample through intelligent spectral sensors at each sampling point, and determine the recognition gain when qualitatively identifying the pests of each pine tree sample according to the similarity feature value of each near-infrared spectral image and the humidity gradient at different sampling points in the pine nematode-affected area; Execution module 403. In the present application, execution module 403 is mainly used to determine the erosion trend of pine nematodes at different sampling points in the pine nematode-affected area through the recognition gain when performing pest qualitative characterization on each pine tree sample and the identification amount when performing pest grading on each pine tree sample, and send all erosion trends to the pine nematode pest monitoring center.

[0068] In addition, the present application also provides a computer device, which includes a memory and a processor, the memory stores codes, and the processor is configured to obtain the codes and execute the above-mentioned pine nematode pest monitoring method based on smart sensors.

[0069] In some embodiments, reference Figure 5 , which is a schematic diagram of the structure of a computer device using a pine nematode pest monitoring method based on an intelligent sensor according to some embodiments of the present application. The pine nematode pest monitoring method based on an intelligent sensor in the above embodiment can be Figure 5 The computer device 500 shown in the figure is implemented, and the computer device 500 includes at least one processor 501, a communication bus 502, a memory 503 and at least one communication interface 504.

[0070] Processor 501 may be a general-purpose central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more processors for controlling the execution of the smart sensor-based pine nematode pest monitoring method of the present application.

[0071] The communication bus 502 may be used to transmit information between the above-mentioned components.

[0072] The memory 503 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, an optical disc storage (including a compressed optical disc, a laser disc, an optical disc, a digital versatile disc, a Blu-ray disc, etc.), a magnetic disk or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory 503 may exist independently and be connected to the processor 501 via the communication bus 502. The memory 503 may also be integrated with the processor 501.

[0073] The memory 503 is used to store the program code for executing the solution of the present application, and the execution is controlled by the processor 501. The processor 501 is used to execute the program code stored in the memory 503. The program code may include one or more software modules. The method described in the above method embodiment can be implemented by the processor 501 and one or more software modules in the program code in the memory 503.

[0074] The communication interface 504 uses any transceiver or other device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0075] In a specific implementation, as an embodiment, a computer device may include multiple processors, each of which may be a single-CPU processor or a multi-CPU processor. The processor here may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0076] The above-mentioned computer device may be a general-purpose computer device or a special-purpose computer device. In a specific implementation, the computer device may be a desktop computer, a portable computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device or an embedded device. The embodiment of the present application does not limit the type of computer device.

[0077] In addition, the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned pine nematode pest monitoring method based on the intelligent sensor is implemented.

[0078] In summary, in the pine nematode pest monitoring system and method based on intelligent sensors disclosed in the embodiments of the present application, first, intelligent spectral sensors are set at different sampling points in the pine nematode-affected area; based on the pre-trained machine learning model, band extraction is performed on the near-infrared spectral information collected by each intelligent spectral sensor to obtain the characterization band of the pine tree samples at each sampling point under the invasion of pine nematodes; the erosion state of the pine tree samples at each sampling point under the invasion of pine nematodes is determined according to the band differences between all near-infrared spectral information and the distribution characteristic values ​​of the spectral curves of each near-infrared spectral information in the absorption valley, and the erosion state of the pine tree samples at each sampling point under the invasion of pine nematodes is determined according to all the erosion states and the pine tree samples at each sampling point. The characterization band of the sample under the invasion of pine nematodes determines the identification amount when grading the pests of each pine tree sample; the near-infrared spectral image of each pine tree sample is automatically collected by the intelligent spectral sensor at each sampling point, and the recognition gain when qualitatively determining the pests of each pine tree sample is determined according to the similarity feature of each near-infrared spectral image and the humidity gradient at different sampling points in the pine nematode-affected area; the erosion trend of pine nematodes at different sampling points in the pine nematode-affected area is determined by the recognition gain when qualitatively determining the pests of each pine tree sample and the identification amount when grading the pests of each pine tree sample, and all the erosion trends are sent to the pine nematode pest monitoring center. The scheme of the present application can monitor the spread trend of pest behavior in the early stage of pine nematode infection, thereby improving the accuracy of early prediction of the spread trend of pine nematodes and preventing large-scale damage.

[0079] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0080] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A pine nematode pest monitoring method based on intelligent sensors, characterized in that: The steps include: Setting up smart spectral sensors at different sampling points in the pine nematode-affected areas; Based on the pre-trained machine learning model, the near-infrared spectral information collected by each intelligent spectral sensor is extracted to obtain the characteristic bands of the pine samples at each sampling point under the invasion of pine nematodes; The erosion state of the pine tree samples at each sampling point under the invasion of pine nematodes is determined according to the band differences between all near-infrared spectral information and the distribution characteristic values ​​of the spectral curves of each near-infrared spectral information in the absorption valley. The identification value for pest classification of each pine tree sample is determined through all the erosion states and the characteristic bands of the pine tree samples at each sampling point under the invasion of pine nematodes. Automatically collect near-infrared spectral images of each pine tree sample by intelligent spectral sensors at each sampling point, and determine the recognition gain when qualitatively identifying the pests of each pine tree sample according to the similarity feature quantity of each near-infrared spectral image and the humidity gradient at different sampling points in the pine nematode-affected area; The erosion trend of pine nematodes at different sampling points in the pine nematode-affected area was determined by the recognition gain when each pine sample was subjected to pest qualitative characterization and the discrimination value when each pine sample was subjected to pest grading, and all erosion trends were sent to the pine nematode pest monitoring center.

2. The method according to claim 1, characterized in that Based on the pre-trained machine learning model, the near-infrared spectral information collected by each intelligent spectral sensor is extracted to obtain the characterization bands of the pine samples at each sampling point under the invasion of pine nematodes, including: The relevant features of each near-infrared spectrum information were selected to obtain the correlation coefficient of pine nematode invasion; Initialize the pre-trained machine learning model; The key characteristic bands of pests in each near-infrared spectral information are extracted through a pre-trained machine learning model; The characterization bands of the pine tree samples at each sampling point under pine nematode invasion are determined based on the pine nematode invasion correlation coefficient and the key characteristic bands of the pest in each near-infrared spectral information.

3. The method according to claim 1, characterized in that According to the band differences between all near-infrared spectral information and the distribution characteristic values ​​of the spectral curves of each near-infrared spectral information in the absorption valley, the erosion state of the pine tree samples at each sampling point under the invasion of pine nematodes is determined, including: Obtain the correlation coefficient of pine nematode invasion; Performing band difference analysis on all near-infrared spectral information by using the pine nematode invasion correlation coefficient to obtain the band differences between all near-infrared spectral information; Analyze the spectral curve of each near-infrared spectral information, and identify the distribution characteristic value of the spectral curve of each near-infrared spectral information in the absorption valley; Perform linear fitting on the distribution characteristic values ​​of all spectral curves in the absorption valley to obtain the distribution characteristic fitting curve of the absorption valley; The pine tree samples at each sampling point are evaluated for pest infestation through the band differences between the distribution characteristic fitting curve and all near-infrared spectral information, and the erosion status of the pine tree samples at each sampling point under the invasion of pine nematodes is obtained.

4. The method according to claim 1, characterized in that The identification quantity for pest classification of each pine tree sample is determined by characterizing the bands of all erosion states and pine tree samples at each sampling point under pine nematode invasion, including: Cluster all erosion states to obtain multiple erosion state clusters; Determine the state characteristic value of each erosion state cluster; The identification value for pest grading of each pine tree sample is determined based on all state characteristic values ​​and the absorbance and light intensity corresponding to all wavelengths in the characterization band of the pine tree sample at each sampling point under pine nematode invasion.

5. The method according to claim 1, characterized in that Determining the recognition gain for qualitatively identifying the pests of each pine tree sample according to the similarity feature quantity of each near infrared spectrum image and the humidity gradient at different sampling points in the pine nematode-affected area specifically includes: Perform brightness feature separation on each near-infrared spectrum image to obtain brightness feature values ​​of each near-infrared spectrum image; According to the humidity gradient at different sampling points in the pine nematode-affected area and the brightness characteristic values ​​of each near-infrared spectral image, the collinearity characteristic values ​​of each near-infrared spectral image corresponding to the pine tree samples were determined; The recognition gain when performing pest qualitative analysis on each pine tree sample is determined by the similarity feature quantity of each near-infrared spectrum image and the collinearity feature quantity of each pine tree sample.

6. The method according to claim 1, characterized in that The pine nematode erosion trends at different sampling points in the pine nematode-affected area were determined by the recognition gain of each pine tree sample when the pest was qualitatively identified and the discrimination value of each pine tree sample when the pest was graded. Specifically, the following are the trends: Select a sampling point as the selected sampling point; The auxiliary scale for evaluating pest invasion at the selected sampling point was determined based on the recognition gain of pine tree samples from adjacent sampling points when conducting pest qualitative analysis; The trend characteristic value of pest invasion at the selected sampling point is determined by the identification value of pine tree samples at the adjacent sampling points during pest classification; Determine the erosion trend of pine nematodes at the selected sampling points in the pine nematode-affected area based on the trend characteristic value of the pest invasion at the selected sampling points and the auxiliary scale of the pest invasion at the selected sampling points; Continue to determine the pine nematode encroachment trends at the remaining sampling sites in the pine nematode-affected area.

7. The method according to claim 1, wherein the intelligent spectral sensor is a device that integrates the collection and analysis of spectral data of an object within different wavelength ranges.

8. A pine nematode pest monitoring system based on intelligent sensors, characterized in that: include: A setting module for setting up intelligent spectral sensors at different sampling points in the pine nematode-affected area; A processing module is used to extract the near-infrared spectral information collected by each intelligent spectral sensor based on a pre-trained machine learning model to obtain the characterization bands of the pine tree samples at each sampling point under the invasion of pine nematodes; The processing module is further used to determine the erosion state of the pine tree samples at each sampling point under the invasion of pine nematodes according to the band differences between all near-infrared spectral information and the distribution characteristic values ​​of the spectral curves of each near-infrared spectral information in the absorption valley, and determine the identification value when performing pest classification on each pine tree sample through all the erosion states and the characteristic bands of the pine tree samples at each sampling point under the invasion of pine nematodes; The processing module is also used to automatically collect near-infrared spectral images of each pine tree sample through intelligent spectral sensors at each sampling point, and determine the recognition gain when performing pest qualitative analysis on each pine tree sample according to the similarity feature quantity of each near-infrared spectral image and the humidity gradient at different sampling points in the pine nematode-affected area; The execution module is used to determine the pine nematode erosion trend at different sampling points in the pine nematode affected area through the recognition gain when each pine tree sample is subjected to pest qualitative characterization and the identification amount when each pine tree sample is subjected to pest grading, and send all erosion trends to the pine nematode pest monitoring center.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the pine nematode pest monitoring method based on the smart sensor according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the pine nematode pest monitoring method based on a smart sensor as described in any one of claims 1 to 7 is implemented.

Images