A method for detecting active ingredients of green pesticides

Through empirical modal decomposition algorithm and adaptive noise intensity adjustment, the low signal accuracy caused by baseline drift in green pesticide detection is solved, and more efficient detection accuracy of green pesticide active ingredients is achieved.

CN118425388BActive Publication Date: 2024-12-06RIZHAO XIANGWO INTELLIGENT AGRICULTURE DEVELOPMENT CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410563362.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-05-08
Publication Date
2024-12-06
Estimated Expiration
2044-05-08

AI Technical Summary

Technical Problem

Traditional pesticide active ingredient detection methods have the problem of low signal data accuracy when detecting green pesticides, especially noise interference caused by baseline drift, which affects detection accuracy.

Method used

The empirical modal decomposition (EMD) algorithm is used to perform preliminary decomposition of the digital signal, obtain the initial IMF component, and iteratively reduce the noise intensity, adaptively determine the optimal noise intensity to be introduced, remove baseline drift, and improve the accuracy of signal data.

Benefits of technology

It improves the accuracy of detection of active ingredients of green pesticides, overcomes noise interference, and enhances the robustness of signal data and the reliability of detection results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118425388B_ABST
    Figure CN118425388B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of data processing technology applicable to specific functions, and specifically to a method for detecting active ingredients of green pesticides, obtaining a digital signal of an active ingredient in a green pesticide to be detected, and further obtaining each initial IMF component of the active ingredient; obtaining the noise intensity to be introduced at each iteration according to the fluctuation characteristics of each initial IMF component, and determining the reasonable degree of the noise intensity to be introduced at each iteration; obtaining the optimal noise intensity to be introduced by combining the reasonable degree of the noise intensity to be introduced at each iteration with the same reference IMF component after introducing two adjacent iterations, and then obtaining a digital signal after removing the baseline drift, and judging whether the active ingredient of the green pesticide meets the standard. The present invention obtains a more accurate digital signal of the active ingredient by improving the determination method of the parameter noise intensity to be introduced, thereby improving the accuracy of the detection of the active ingredient of the green pesticide.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of data processing applicable to specific functions, and in particular to a method for detecting effective components of green pesticides. Background Art

[0002] Compared with traditional pesticides, green pesticides are more environmentally friendly and have less impact on the ecosystem. The active ingredients of green pesticides usually come from natural plant extracts, microorganisms, etc., and have lower ecotoxicity and environmental damage. In the current agricultural field, judging whether the active ingredients of green pesticides meet the standards is a very important part of the green pesticide testing process.

[0003] Traditional methods for detecting active ingredients in pesticides face some limitations when applied to green pesticides. On the one hand, the complexity and diversity of green pesticides increase the demand for efficient, accurate and reliable detection methods; on the other hand, some technologies in traditional methods for detecting active ingredients in pesticides cannot fully adapt to the special properties of green pesticides, such as natural extracts and microorganisms. Therefore, the characteristics of active ingredient signal data are analyzed by high performance liquid chromatography to achieve the detection of active ingredients in green pesticides.

[0004] In the process of collecting signal data of active ingredients in green pesticides using high performance liquid chromatography, the detector may experience baseline drift in the signal data due to reasons such as instrument drift, volumetric purity problems or aging of the chromatographic column. The baseline drift problem will affect the accuracy of the signal data. Accurate detection of the active ingredients in green pesticides cannot be achieved based on the inaccurate active ingredient signal data. Summary of the invention

[0005] In order to solve the technical problem that the accuracy of the collected active ingredient signal data is low, resulting in poor accuracy in detecting the active ingredients of green pesticides, the purpose of the present invention is to provide a method for detecting the active ingredients of green pesticides. The technical solution adopted is as follows:

[0006] An embodiment of the present invention provides a method for detecting active ingredients of green pesticides, the method comprising the following steps:

[0007] The digital signal of the active ingredient in the green pesticide to be detected is obtained, and the digital signal is preliminarily decomposed to obtain each initial IMF component of the active ingredient; wherein the horizontal axis of the digital signal is time, and the vertical axis is absorbance;

[0008] According to the fluctuation characteristics of each initial IMF component, the effective component characteristic value of each initial IMF component is determined, and then the fluctuation degree of the difference between any two effective component characteristic values ​​is analyzed to obtain the initial noise intensity to be introduced; the initial noise intensity to be introduced is iteratively reduced in sequence until the noise intensity to be introduced is not greater than 0, and the noise intensity to be introduced at each iteration is obtained;

[0009] A noise component signal is randomly generated according to the noise intensity to be introduced at each iteration; based on each initial IMF component, the noise component signal at each iteration and each reference IMF component after the noise component signal at each iteration is introduced, the decomposition effect of the decomposition result after the noise component signal at each iteration is introduced is analyzed to determine the reasonableness of the noise intensity to be introduced at each iteration;

[0010] According to the reasonableness of the noise intensity to be introduced in each iteration and any reference IMF component after the noise component signal in two adjacent iterations is introduced, the optimal noise intensity to be introduced is determined, and then the digital signal after removing the baseline drift corresponding to the effective component is determined;

[0011] According to the digital signal corresponding to the active ingredient after removing the baseline drift, it is judged whether the active ingredient of the green pesticide to be tested meets the standard.

[0012] Furthermore, the step of obtaining each reference IMF component includes:

[0013] For any iteration, determine the noise component signal of the noise intensity to be introduced at the iteration, introduce the noise component signal into the digital signal of the effective component, and obtain the digital signal of the noise component signal introduced at the iteration;

[0014] Performing collective empirical mode decomposition on the digital signal of the noise component signal introduced at the iteration to obtain various IMF components, and using the various IMF components as various reference IMF components after the noise component signal at the iteration is introduced.

[0015] Furthermore, according to the fluctuation characteristics of each initial IMF component, the effective component characteristic value of each initial IMF component is determined, including:

[0016] For any initial IMF component, obtain a frequency spectrum of the initial IMF component and determine each peak point in the frequency spectrum;

[0017] Select the peak point with the largest ordinate from each peak point in the spectrum diagram of the initial IMF component as the target peak point, calculate the abscissa difference between the two valley points closest to the target peak point, and determine it as the frequency component of the initial IMF component;

[0018] Obtain each extreme point in the initial IMF component, calculate the average value of the absorbance difference of all adjacent extreme points in the initial IMF component, and determine the fluctuation degree of the distribution curve of the initial IMF component;

[0019] The number of extreme value points of the initial IMF component is determined as the fluctuation frequency of the distribution curve of the initial IMF component;

[0020] The effective component characteristic value of the initial IMF component is determined by combining the frequency component of the initial IMF component, the fluctuation degree of the distribution curve and the fluctuation frequency of the distribution curve.

[0021] Furthermore, the fluctuation degree of the difference between the characteristic values ​​of any two effective components is analyzed to obtain the initial noise intensity to be introduced, including:

[0022] According to the effective component eigenvalues ​​of each initial IMF component, the difference value of any two effective component eigenvalues ​​is calculated and recorded as the characteristic difference value, and each characteristic difference value is obtained to form a characteristic difference value sequence; the variance of the characteristic difference value sequence is calculated, and the variance of the characteristic difference value sequence is used as the initial noise intensity to be introduced.

[0023] Further, according to each initial IMF component, the noise component signal at each iteration, and each reference IMF component after the noise component signal at each iteration is introduced, the decomposition effect of the decomposition result after the noise component signal at each iteration is introduced is analyzed to determine the reasonableness of the noise intensity to be introduced at each iteration, including:

[0024] For the noise component signal at any iteration, according to the absorbance difference between each initial IMF component and the noise component signal at this iteration, the degree to which the noise component signal participates in signal decomposition is analyzed, and the first reasonable factor of the noise intensity to be introduced at this iteration is determined;

[0025] Determine the effective component characteristic values ​​of each reference IMF component after the noise component signal at this iteration is introduced; analyze the independent distribution characteristics of the reference IMF components according to the characteristic difference values ​​of any two reference IMF components after the noise component signal at this iteration is introduced, and determine the second reasonable factor of the noise intensity to be introduced at this iteration;

[0026] According to the time of each extreme point in each reference IMF component after the noise component signal at this iteration is introduced, the degree of distinction between the high-frequency and low-frequency components of the reference IMF component is analyzed to determine the third reasonable factor of the noise intensity to be introduced at this iteration;

[0027] The rationality of the intensity of the noise to be introduced at the iteration is determined by combining the first rationality factor, the second rationality factor and the third rationality factor of the intensity of the noise to be introduced at the iteration.

[0028] Furthermore, the calculation formula of the first reasonable factor of the noise intensity to be introduced in this iteration is:

[0029] Where R is the first reasonable factor of the noise intensity to be introduced at this iteration, u is the number of initial IMF components, e is the serial number of the initial IMF component, N is the number of data points in the initial IMF component or noise component signal, y is the serial number of the data point in the initial IMF component or noise component signal, T y o is the absorbance of the yth data point in the noise component signal at this iteration, T y e is the absorbance of the yth data point in the eth initial IMF component.

[0030] Furthermore, according to the time of each extreme point in each reference IMF component after the noise component signal at this iteration is introduced, the degree of distinction between the high and low frequency components of the reference IMF component is analyzed, and a third reasonable factor of the intensity of the noise to be introduced at this iteration is determined, including:

[0031] For any reference IMF component after the noise component signal of this iteration is introduced, the time difference between all two adjacent extreme value points in the reference IMF component is determined, recorded as the time difference value, and the average value of each time difference value is calculated as the frequency degree of the reference IMF component;

[0032] According to the frequency levels of each reference IMF component after the noise component signal is introduced in this iteration, the difference between the frequency levels of two adjacent reference IMF components is calculated and recorded as the frequency level difference value, and the average value of all frequency level difference values ​​is used as the third reasonable factor of the noise intensity to be introduced in this iteration.

[0033] Further, according to the reasonableness of the noise intensity to be introduced in each iteration and any reference IMF component after the noise component signal in two adjacent iterations is introduced, the optimal noise intensity to be introduced is determined, including:

[0034] According to any reference IMF component after the noise component signal of the current iteration and the previous iteration is introduced, the component stability after the noise component signal of the current iteration is introduced is calculated;

[0035] Determine the introduction evaluation value of the noise intensity to be introduced at the current iteration by combining the rationality of the noise intensity to be introduced at the current iteration and the component stability after the noise component signal at the current iteration is introduced;

[0036] The noise intensity to be introduced corresponding to the maximum introduction evaluation value is taken as the optimal noise intensity to be introduced.

[0037] Furthermore, the calculation formula for the component stability after the noise component signal at the kth iteration is introduced is:

[0038] Where W k is the component stability after the noise component signal is introduced at the kth iteration, c is the sequence number of the data point in the reference IMF component, x is the number of data points in the reference IMF component, and s dc k is the absorbance of the cth data point in the dth reference IMF component after the noise component signal at the kth iteration is introduced, s dc k-1 is the absorbance of the cth data point in the dth reference IMF component after the noise component signal at the k-1th iteration is introduced, and || is the absolute value function.

[0039] Furthermore, judging whether the active ingredient of the green pesticide to be tested meets the standard according to the digital signal after removing the baseline drift corresponding to the active ingredient includes:

[0040] De-noising the digital signal corresponding to the effective component after removing the baseline drift to obtain a de-noised digital signal; in the de-noised digital signal, selecting a band in a preset nanometer range as a detection wavelength band;

[0041] Calculate the absorption peak area of ​​the detection wavelength and the total absorption peak area of ​​all wavelength bands in the digital signal, and use the ratio of the absorption peak area to the total absorption peak area as the effective component detection index;

[0042] When the active ingredient detection index is greater than the preset empirical threshold, it is determined that the active ingredient of the green pesticide to be detected meets the standard; otherwise, it is determined that the active ingredient of the green pesticide to be detected does not meet the standard.

[0043] Furthermore, according to the characteristic difference values ​​of any two reference IMF components after the noise component signal at the iteration is introduced, the independent distribution characteristics of the reference IMF components are analyzed to determine the second reasonable factor of the noise intensity to be introduced at the iteration, including:

[0044] Determine the characteristic difference value between any two reference IMF components after the noise component signal at this iteration is introduced, obtain all the corresponding characteristic difference values ​​after the noise component signal at this iteration is introduced, and calculate the cumulative sum of all the characteristic difference values ​​as the second reasonable factor of the noise intensity to be introduced at this iteration.

[0045] Further, determining the digital signal corresponding to the effective component after removing the baseline drift includes:

[0046] According to the last reference IMF component corresponding to the optimal noise intensity to be introduced and the last reference IMF components corresponding to each noise intensity to be introduced before the optimal noise intensity to be introduced, the average value of all the last reference IMF components is calculated as the trend item signal; the digital signal of the effective component is subtracted from the trend item signal to obtain the digital signal corresponding to the effective component after removing the baseline drift.

[0047] The present invention has the following beneficial effects:

[0048] The present invention provides a method for detecting the effective components of green pesticides, which mainly performs high-performance liquid chromatography detection on green pesticide samples, and collects signal data in the detection process through a detector. In order to overcome the problem of baseline drift in the collected signal data, the trend item in the digital signal is removed to improve the accuracy of the collected signal data, overcome the adverse effects of noise data, and further improve the accuracy of green pesticide effective component detection; in order to improve the accuracy of trend item signal extraction, compared with the existing preset noise intensity to be introduced, the initial noise intensity to be introduced adaptively determined based on the component signal fluctuation characteristics of this embodiment is more robust and more suitable for green pesticide effective component detection scenarios; different sizes of noise intensity to be introduced are obtained by iterative method, and the decomposition effect of the decomposition result after the noise component signal of different sizes of noise intensity is introduced is analyzed, and the reasonable degree of the noise intensity to be introduced at each iteration is quantified, and the decomposition effect of the decomposition result after the noise component signal is introduced is analyzed, which is conducive to improving the numerical accuracy of the reasonable degree; combined with any reference IMF component after the noise component signal is introduced during two adjacent iterations, the stability during iteration can be analyzed, and the reliability of the optimal noise intensity to be introduced is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0050] Figure 1 This is a flow chart of a method for detecting active ingredients of green pesticides according to an embodiment of the present invention;

[0051] Figure 2 An example diagram of a digital signal of an effective component in an embodiment of the present invention;

[0052] Figure 3 This is an example diagram of a digital signal after removing the baseline drift corresponding to the effective component in an embodiment of the present invention. DETAILED DESCRIPTION

[0053] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects of the technical solutions proposed by the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0054] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0055] The scenario targeted by the embodiment of the present invention is: the methods for detecting green pesticides mainly include chemical analysis and biological analysis. The chemical analysis method determines whether the green component standard is met by monitoring the active ingredients in the green pesticides. This embodiment mainly uses the chemical analysis method of high performance liquid chromatography to detect green pesticides. In the process of using a UV detector to monitor green pesticide samples, in order to overcome the baseline offset problem of the signal caused by instrument drift, volume purity problems or aging of the chromatographic column, it is necessary to obtain more accurate active component signal data.

[0056] In order to detect the active ingredients of green pesticides and overcome the limitations of traditional methods, this embodiment provides a method for detecting the active ingredients of green pesticides, such as Figure 1 As shown, the following steps are included:

[0057] S1, obtaining the digital signal of the active ingredient in the green pesticide to be detected, performing preliminary decomposition on the digital signal, and obtaining each initial IMF component of the active ingredient.

[0058] The first step is to obtain the digital signal of the active ingredient in the green pesticide to be detected.

[0059] First, collect samples of green pesticides to be tested, and perform appropriate pretreatment according to the sample type, such as extraction, purification, concentration, etc.; second, prepare standard samples of green pesticides to be tested, prepare the standard samples into working solutions of appropriate concentrations, and add them to the pretreated samples; then, inject the samples into the HPLC (High Performance Liquid Chromatography) system, and different injection techniques can be used, such as automatic injector or manual injection; then, use gradient elution and set the mobile phase gradient, that is, change the ratio of the mobile phase composition over time, which helps to separate the different components in the sample; third, select appropriate chromatographic columns, mobile phases, flow rates, detection wavelengths and other conditions, and use ultraviolet (ultraviolet absorption The detector measures the absorbance in the ultraviolet spectrum, and the commonly used wavelengths include 200-400 nanometers, to obtain the analog signal of the active ingredient; finally, it is converted into digital form by an analog-to-digital converter (ADC), and then the digital signal is recorded by data acquisition software. The digital signal of the active ingredient in the sample has the horizontal axis of time and the vertical axis of absorbance.

[0060] Among them, the example of the digital signal of the effective component is as follows Figure 2 As shown; in addition, nicotine pesticides can be used as test samples in the process of detecting the active ingredients in green pesticides using high performance liquid chromatography.

[0061] The second step is to preliminarily decompose the digital signal to obtain the initial IMF components of the effective components.

[0062] It should be noted that in the process of detecting the active ingredients of pesticides using high performance liquid chromatography, it is necessary to effectively process the digital signal data of the active ingredients collected by the ultraviolet detector. In order to achieve accurate identification of digital signals, it is necessary to effectively remove the noise components in the digital signals, which may come from instruments, solvents or other environmental factors. Before removing the noise, the digital signal data needs to be preprocessed to overcome the baseline offset caused by instrument drift, solvent purity problems or aging of the chromatographic column, avoid affecting the signal denoising processing results, and improve the accuracy of the digital signal data. In the process of removing the trend item component caused by the baseline offset in the digital signal, the digital signal of the active ingredient is first preliminarily decomposed to improve the data support for the noise component signal of the subsequent adaptive determination of the noise intensity.

[0063] In this embodiment, the EMD algorithm (Empirical Mode Decomposition) can be used to perform preliminary decomposition of the digital signal to obtain the various IMF components (Intrinsic Mode Functions) of the digital signal, which are recorded as initial IMF components to obtain the initial IMF components of the effective components. Of course, the implementer can also use other methods to achieve the preliminary decomposition of the digital signal. The implementation process of the EMD algorithm is a prior art and is not within the scope of protection of the present invention, so it will not be elaborated in detail this time.

[0064] So far, this embodiment obtains various initial IMF components by preliminarily decomposing the obtained digital signal corresponding to the green pesticide to be detected.

[0065] S2, determining the intensity of the noise to be introduced in each iteration according to the fluctuation characteristics of each initial IMF component of the active ingredient in the green pesticide to be detected.

[0066] It should be noted that the introduction of random noise into the digital signal of the effective component can avoid the modal problem in the signal decomposition process to a certain extent, so that the components of the different components of the digital signal of the effective component contained in the decomposition result can be more distinguished, which can be achieved through the EEMD algorithm (Ensemble Empirical Mode Decomposition). The intensity of the noise to be introduced in the existing EEMD algorithm is inconsistent with the actual signal fluctuation. The noise intensity is also the noise amplitude. In order to improve the robustness of the setting of the noise intensity to be introduced, the initial noise intensity to be introduced is determined based on the fluctuation characteristics of each initial IMF component obtained by the preliminary decomposition, and then the noise intensity to be introduced at each iteration is determined. The specific implementation steps may include:

[0067] The first step is to determine the effective component characteristic value of each initial IMF component according to the fluctuation characteristics of each initial IMF component.

[0068] In this embodiment, by combining the different fluctuation characteristics of the component signals, the distribution characteristics of the component signals are quantified, and the partial data characteristics of the active ingredients in the green pesticide to be detected corresponding to the component signals are determined, so as to provide data support for the subsequent determination of the initial noise intensity to be introduced. The calculation process of the active ingredients of each initial IMF decomposition is consistent. For the purpose of description, any one of the initial IMF components is taken as an example to determine the characteristic value of the active ingredient.

[0069] The first sub-step is to obtain the spectrum of the initial IMF component and determine each peak point in the spectrum.

[0070] In this embodiment, the initial IMF component is converted into the frequency domain using FFT (Fast Fourier Transform) to obtain the spectrum of the initial IMF component; and the peak point detection algorithm is used to obtain each peak point in the initial IMF component. The implementation process of FFT and the peak point detection algorithm is prior art and is not within the scope of protection of the present invention, and will not be elaborated here.

[0071] In the second sub-step, the peak point with the largest ordinate is selected as the target peak point from each peak point in the spectrum diagram of the initial IMF component, and the abscissa difference between the two valley points closest to the target peak point is calculated to determine the frequency component of the initial IMF component.

[0072] In this embodiment, a representative maximum peak point, that is, a peak point with the largest amplitude, is selected from all peak points and recorded as a target peak point. The horizontal axis in the spectrum graph is the frequency and the vertical axis is the amplitude. Two valley points adjacent to the target peak point are selected and marked as f a and f b , the two valley points adjacent to the target peak point are the two valley points closest to the target peak point; the bandwidth of the target peak point is characterized by the horizontal coordinate difference between the two valley points, the size of the bandwidth can characterize the frequency component of the initial IMF component, and the horizontal coordinate difference between the two valley points can be the absolute value of the frequency difference between the two valley points.

[0073] The third sub-step is to obtain each extreme point in the initial IMF component, calculate the average value of the absorbance difference of all adjacent extreme points in the initial IMF component, and determine the fluctuation degree of the distribution curve of the initial IMF component.

[0074] In this embodiment, the curve signal fluctuation characteristics of the initial IMF component can be analyzed through extreme points. The greater the difference between two adjacent extreme points, the greater the fluctuation degree of the distribution curve, which indicates that the fluctuation degree of the signal distribution curve in the initial IMF component is greater.

[0075] In the fourth sub-step, the number of extreme value points of the initial IMF component is determined as the fluctuation frequency of the distribution curve of the initial IMF component.

[0076] In this embodiment, the more extreme value points there are in the signal distribution curve in the initial IMF component, the higher the fluctuation frequency of the initial IMF component appears, and the higher the fluctuation frequency of the distribution curve will be.

[0077] The fifth sub-step is to determine the effective component characteristic value of the initial IMF component by combining the frequency component of the initial IMF component, the fluctuation degree of the distribution curve and the fluctuation frequency of the distribution curve.

[0078] In this embodiment, different characteristic expressions of the initial IMF components are combined by multiplication to obtain a characteristic value, namely, an effective component characteristic value, and each initial IMF component has its corresponding effective component characteristic value.

[0079] As an example, the calculation formula for the effective component eigenvalue of the e-th initial IMF component can be:

[0080] Where Y e is the effective component eigenvalue of the e-th initial IMF component, n is the total logarithm of adjacent extreme value points in the e-th initial IMF component, m is the logarithmic sequence number of adjacent extreme value points in the e-th initial IMF component, h i is the absorbance of the ith extreme point in the eth initial IMF component, h i+1 is the absorbance of the i+1th extreme point in the eth initial IMF component, |h i -h i+1 | m is the absorbance difference of the mth pair of adjacent extreme points in the eth initial IMF component, G e is the number of extreme points of the e-th initial IMF component, f a is a valley point adjacent to the target peak point in the e-th initial IMF component, f b is another valley point adjacent to the target peak point in the e-th initial IMF component, || is the absolute value function; |f a -f b | represents the frequency component of the e-th initial IMF component, represents the fluctuation degree of the distribution curve of the e-th initial IMF component, G e It is also the fluctuation frequency of the distribution curve of the e-th initial IMF component, and || is the absolute value function.

[0081] The second step is to analyze the fluctuation degree of the difference between the characteristic values ​​of any two effective components to obtain the initial noise intensity to be introduced.

[0082] In this embodiment, according to the effective component characteristic values ​​of each initial IMF component, the difference value of any two effective component characteristic values ​​is calculated, recorded as the characteristic difference value, and each characteristic difference value is obtained to form a characteristic difference value sequence; the variance of the characteristic difference value sequence is calculated, and the variance of the characteristic difference value sequence is used as the initial noise intensity to be introduced. Among them, the value range of the initial noise level to be introduced is 0 to positive infinity. When the initial noise level to be introduced is 0, it means that all characteristic difference values ​​are equal. At this time, the characteristic distribution of the digital signal is stable, and there is a better decomposition result.

[0083] It should be noted that the smaller the variance of the characteristic difference value sequence, the worse the stability of the characteristic difference value distribution, the smaller the fluctuation of the characteristic difference value distribution, and the smaller the noise intensity in the digital signal of the effective component. The variance of the characteristic difference value sequence is correlated with the noise intensity to be introduced, so the variance of the characteristic difference value sequence can be directly used as the initial noise intensity to be introduced, so as to facilitate the subsequent noise cancellation in the digital signal and obtain a component signal with higher accuracy.

[0084] In the third step, the initial noise intensity to be introduced is iteratively reduced in sequence until the noise intensity to be introduced is no greater than 0, and the noise intensity to be introduced in each iteration is obtained.

[0085] In this embodiment, after obtaining the initial noise intensity to be introduced, the initial noise intensity to be introduced is iteratively decreased in sequence, and the noise intensity to be introduced at each iteration can be obtained until the noise intensity to be introduced after decreasing is not greater than 0, that is, less than or equal to 0. The iteration step size can be set to 2, and the implementer can set the size of the iteration step size according to the actual situation, and no specific limitation is made.

[0086] It should be noted that the noise intensity is obtained based on the frequency overlap that may occur during the decomposition process of the EMD. However, this noise intensity is not the optimal parameter for EEMD decomposition. In order to facilitate the subsequent analysis of the decomposition effects corresponding to different sizes of noise intensities to be introduced, the noise intensity to be introduced at each iteration is determined by iterative reduction.

[0087] So far, this embodiment has determined the intensity of the noise to be introduced in each iteration based on the fluctuation characteristics of the component signals initially decomposed.

[0088] S3, determining a reasonable degree of the intensity of the noise to be introduced at each iteration according to each initial IMF component, the noise component signal at each iteration, and each reference IMF component after the noise component signal at each iteration is introduced.

[0089] It should be noted that the decomposition results of the EEMD algorithm for different noise intensities to be introduced are different. In order to obtain effective component signal data with higher quality and accuracy and less noise influence, the reasonable degree of the noise intensity to be introduced at each iteration is evaluated based on the decomposition effect of different noise intensities to be introduced. The calculation implementation process of the reasonable degree of the noise intensity to be introduced at each iteration is the same. Taking the determination process of the reasonable degree of the noise intensity to be introduced at any iteration as an example, the specific implementation steps may include:

[0090] In the first step, according to the absorbance difference between each initial IMF component and the noise component signal at this iteration, the degree of participation of the noise component signal in signal decomposition is analyzed, and the first reasonable factor of the noise intensity to be introduced at this iteration is determined.

[0091] It should be noted that the first reasonable factor can characterize the degree to which the noise component signal participates in the signal decomposition. The larger the first reasonable factor is, the greater the difference between the noise component signal at this iteration and the initial IMF components is, which can ensure that the noise component introduced in the decomposition process is also decomposed accordingly; if the first reasonable factor is smaller, the fewer noise components participate in the decomposition process, and the frequency overlap in the decomposition process cannot be effectively solved.

[0092] In this embodiment, the random noise to be introduced can be decomposed together with the components of the effective component digital signal. If the difference between the noise component signal of the noise intensity to be introduced and the initial IMF components of the preliminary decomposition is small, modal problems are likely to occur in the decomposition process of the signal introduced with noise. Therefore, the greater the difference between the noise component signal to be introduced and the initial IMF components, the more helpful it is to avoid the problem of modal aliasing in the decomposition process.

[0093] First, the noise component signal at this iteration is obtained. The noise intensity to be introduced at this iteration is known. A set of random numbers is generated according to the noise intensity to be introduced. The data property of the random number is absorbance. The absorbance whose standard deviation meets the noise intensity is selected as the noise data to be introduced. The noise data generation process is a prior art and is not within the scope of protection of the present invention. It will not be elaborated in detail here. Secondly, a curve fitting is performed on the generated set of noise data to obtain the noise component signal of the noise intensity to be introduced at this iteration, which is referred to as the noise component signal at this iteration. It is worth noting that the length of the noise component signal at each iteration is consistent with the length of the initial IMF component signal.

[0094] Finally, according to the absorbance difference between each initial IMF component and the noise component signal at this iteration, the first reasonable factor of the noise intensity to be introduced at this iteration is calculated. The first reasonable factor can characterize the degree to which the noise component signal of the noise intensity to be introduced at this iteration participates in the signal decomposition.

[0095] As an example, the calculation formula of the first reasonable factor of the noise intensity to be introduced in this iteration may be:

[0096] Where R is the first reasonable factor of the noise intensity to be introduced at this iteration, u is the number of initial IMF components, e is the serial number of the initial IMF component, N is the number of data points in the initial IMF component or noise component signal, y is the serial number of the data point in the initial IMF component or noise component signal, T y o is the absorbance of the yth data point in the noise component signal at this iteration, T y eis the absorbance of the yth data point in the eth initial IMF component.

[0097] In the calculation formula of the first rational factor, It can characterize the mean square error between the noise component signal at this iteration and the e-th initial IMF component. The smaller the mean square error, the higher the similarity between the noise component signal and the e-th initial IMF component, and the worse the decomposition effect of the noise component signal at this iteration on the e-th initial IMF component. It can represent the average value of the mean square error between the noise component signal at this iteration and all the initial IMF components. The larger the average value of the mean square error, the more reasonable the noise intensity to be introduced at this iteration is, and the more helpful it is to reduce the possibility of modal problems in the decomposition process.

[0098] The second step is to determine the effective component characteristic values ​​of each reference IMF component after the noise component signal of this iteration is introduced; based on the characteristic difference values ​​of any two reference IMF components after the noise component signal of this iteration is introduced, the independent distribution characteristics of the reference IMF components are analyzed to determine the second reasonable factor of the noise intensity to be introduced at this iteration.

[0099] The first sub-step is to obtain each reference IMF component after the noise component signal of the iteration is introduced.

[0100] In this embodiment, the noise component signal at this iteration is introduced into the digital signal of the effective component to obtain the digital signal of the noise component signal introduced at this iteration; the digital signal of the noise component signal introduced at this iteration is subjected to collective empirical mode decomposition to obtain various IMF components, and the various IMF components are used as various reference IMF components after the noise component signal at this iteration is introduced. The implementation process of collective empirical mode decomposition is a prior art, which is not within the protection scope of the present invention and will not be elaborated in detail this time.

[0101] The second sub-step is to determine the effective component characteristic value of each reference IMF component.

[0102] In this embodiment, the effective component value of the reference IMF component can be obtained by referring to the calculation process of the effective component value of the initial IMF component, and the calculation process of the effective component characteristic value will not be repeated here.

[0103] The third sub-step is to determine a second reasonable factor of the noise intensity to be introduced in this iteration.

[0104] It should be noted that the second reasonable factor can characterize the independent distribution characteristics of the reference IMF components. The more obvious the independent distribution characteristics are, the better the decomposition effect corresponding to the noise intensity to be introduced at this iteration. The independent distribution characteristics of all reference IMF components can be quantified by analyzing the difference in the effective component characteristic values ​​between any two reference IMF components after the noise component signal at this iteration is introduced, that is, the characteristic difference value.

[0105] In this embodiment, the characteristic difference value between any two reference IMF components after the noise component signal at the iteration is introduced is determined, and all characteristic difference values ​​corresponding to the noise component signal at the iteration are obtained. The independent distribution characteristics of all reference IMF components are measured by calculating the cumulative sum of all characteristic difference values, and the second reasonable factor is determined. The larger the second reasonable factor is, the more obvious the independent distribution characteristics of the reference IMF components after the noise component signal at the iteration is introduced, the smaller the possibility that the effective component characteristics between some reference IMF components are similar, the better the decomposition effect is, and the more reasonable the intensity of the noise to be introduced at the iteration is.

[0106] The third step is to analyze the degree of distinction between high-frequency and low-frequency components of the reference IMF components according to the time of each extreme point in each reference IMF component after the noise component signal of this iteration is introduced, and determine the third reasonable factor of the noise intensity to be introduced at this iteration.

[0107] It should be noted that by analyzing the time of each extreme point in the reference IMF component, the frequency degree of the reference IMF component can be quantified. Based on the frequency degrees of different reference IMF components, the degree of distinction between high-frequency components and low-frequency components in all reference IMF components is analyzed. The decomposition effect after the noise component signal is introduced in this iteration is quantified again to determine the third reasonable factor. The specific implementation steps may include:

[0108] In the first sub-step, for any reference IMF component after the noise component signal is introduced during this iteration, the time difference between all two adjacent extreme points in the reference IMF component is determined, recorded as the time difference value, and the average value of each time difference value is calculated as the frequency degree of the reference IMF component.

[0109] In this embodiment, the difference between the time of two adjacent extreme value points can be the absolute value of the difference between the time, and the absolute value of the difference between the time is the time interval. The larger the time interval, the more likely the corresponding component signal is a high-frequency component. By quantifying the frequency levels of different reference IMF components, it is helpful to subsequently analyze the difference between the frequency levels of two adjacent reference IMF components. The larger the difference between the frequency levels, the more helpful it is to distinguish the high-frequency component and the low-frequency component after the noise component signal is introduced during the iteration.

[0110] In the second sub-step, based on the frequency levels of each reference IMF component after the noise component signal is introduced in this iteration, the difference between the frequency levels of two adjacent reference IMF components is calculated and recorded as the frequency level difference value, and the average value of all frequency level difference values ​​is used as the third reasonable factor of the noise intensity to be introduced in this iteration.

[0111] In this embodiment, after the noise component signal is introduced during this iteration, the frequency difference between adjacent reference IMF components is large, the high-frequency components and the low-frequency components in the reference IMF components can be better distinguished, the decomposition effect is better, the rationality of the noise intensity to be introduced during this iteration is better, and the third rational factor of the noise intensity to be introduced during this iteration is higher.

[0112] The fourth step is to determine the rationality of the intensity of the noise to be introduced at this iteration by combining the first rational factor, the second rational factor and the third rational factor of the intensity of the noise to be introduced at this iteration.

[0113] In this embodiment, the first reasonable factor, the second reasonable factor and the third reasonable factor of the noise intensity to be introduced in this iteration can be combined by multiplication, and the product of the three reasonable factors is used as the reasonable degree of the noise intensity to be introduced in this iteration.

[0114] It should be noted that analyzing the decomposition effect of the noise component signal introduced in this iteration from three different perspectives and quantifying the reasonable degree of the noise intensity to be introduced in this iteration can effectively improve the numerical accuracy and reliability of the reasonable degree of the noise intensity to be introduced.

[0115] As an example, the calculation formula for the reasonable degree of the noise intensity to be introduced during this iteration may be:

[0116] Where E is the reasonable degree of the noise intensity to be introduced at this iteration, R is the first reasonable factor of the noise intensity to be introduced at this iteration, M is the number of feature difference values ​​at this iteration, p is the sequence number of the feature difference value at this iteration, Δw p is the pth characteristic difference value corresponding to the noise component signal at this iteration, is the second reasonable factor of the noise intensity to be introduced at this iteration, v is the number of reference IMF components after the noise component signal is introduced at this iteration, d is the reference IMF component sequence number after the noise component signal is introduced at this iteration, t d is the frequency level of the dth reference IMF component after the noise component signal of this iteration is introduced, t d+1 is the frequency level of the d+1th reference IMF component after the noise component signal of this iteration is introduced, |t d -td+1 | is the frequency difference value after the noise component signal is introduced in this iteration, is the third reasonable factor of the noise intensity to be introduced in this iteration, and || is the absolute value function.

[0117] At this point, this embodiment quantifies the rationality of the noise intensity to be introduced at each iteration by analyzing the decomposition effect of the decomposition result after introducing the noise component signal at each iteration. The rationality here refers to the rationality of introducing the noise component signal corresponding to the noise intensity into the digital signal. The higher the signal decomposition effect, the higher the rationality of the corresponding introduced noise intensity, and the higher the accuracy of the subsequent extraction of the trend signal of the baseline offset.

[0118] S4, obtaining a digital signal after removing the baseline drift corresponding to the effective component according to the reasonableness of the intensity of the noise to be introduced in each iteration and any reference IMF component after the noise component signals in two adjacent iterations are introduced.

[0119] It should be noted that in the process of ensemble empirical mode decomposition, the last IMF component usually corresponds to the trend term of the signal, that is, the signal component corresponding to the baseline drift. When there is a baseline drift in the signal, the accuracy of the decomposed component signal is poor, and other components are easily misjudged as trend terms, resulting in inaccurate extraction of trend terms, which affects the accuracy of the decomposition. It also makes it difficult to distinguish between high-frequency components and low-frequency components in the signal, reducing the accuracy of the signal decomposition result. This embodiment analyzes the influence of each noise intensity to be introduced on the extraction accuracy of the trend term signal of the baseline drift in an iterative manner, so as to obtain the optimal noise intensity to be introduced and overcome the influence of baseline drift on the accuracy of the effective component signal data.

[0120] In the first step, the optimal noise intensity to be introduced is determined according to the reasonableness of the noise intensity to be introduced in each iteration and any reference IMF component after the noise component signal in two adjacent iterations is introduced.

[0121] In this embodiment, when determining the optimal noise intensity to be introduced, it is necessary not only to consider the decomposition effect of different noise intensities to be introduced for signal decomposition, but also to consider the stability of the reference IMF component signal after the noise component signal of the current iteration is introduced compared to the stability of the reference IMF component signal when the noise component signal of the previous iteration is introduced. The more stable the component signal is, the closer the noise intensity to be introduced in the current iteration is to the convergence condition. The step of determining the optimal noise intensity to be introduced may include:

[0122] In the first sub-step, according to any reference IMF component after the noise component signal of the current iteration and the previous iteration is introduced, the component stability after the noise component signal of the current iteration is introduced is calculated.

[0123] By calculating the difference between any two groups of first IFM components after two adjacent introductions of noise components, the component stability after the noise component signal at the current iteration is evaluated.

[0124] As an example, the calculation formula for the component stability after the noise component signal at the current iteration is introduced can be:

[0125] Where W k is the component stability after the noise component signal is introduced at the kth iteration, c is the sequence number of the data point in the reference IMF component, x is the number of data points in the reference IMF component, and s dc k is the absorbance of the cth data point in the dth reference IMF component after the noise component signal at the kth iteration is introduced, s dc k-1 is the absorbance of the cth data point in the dth reference IMF component after the noise component signal at the k-1th iteration is introduced, and || is the absolute value function.

[0126] In the calculation formula of component stability, The difference between the d-th reference IMF component after the introduction of the noise component signal at the k-th iteration and the d-th reference IMF component after the introduction of the noise component signal at the k-1-th iteration can be characterized. The smaller the difference, the closer the current iteration is to stability, and the higher the introduction evaluation value of the noise intensity to be introduced at the current iteration; the d-th reference IMF component can be the second-to-last reference IMF component after the introduction of the noise component signal at each iteration. The reason is that because the last reference IMF component in the iteration is a trend item signal, if the second-to-last reference IMF component tends to be stable during the iteration process, it means that the iteration process tends to the convergence condition.

[0127] In the second sub-step, based on the rationality of the noise intensity to be introduced at the current iteration and the component stability after the noise component signal is introduced at the current iteration, the introduction evaluation value of the noise intensity to be introduced at the current iteration is determined, and the noise intensity to be introduced corresponding to the maximum introduction evaluation value is taken as the optimal noise intensity to be introduced.

[0128] In this embodiment, the product of the rationality and the component stability is calculated by multiplication, and the product of the two is used as the introduction evaluation value of the noise intensity to be introduced at the current iteration. The higher the rationality and the component stability, the higher the introduction evaluation value, and the higher the possibility that the corresponding noise intensity to be introduced is the optimal noise intensity to be introduced. Therefore, the noise intensity to be introduced corresponding to the maximum introduction evaluation value can be used as the optimal noise intensity to be introduced.

[0129] The second step is to determine the digital signal corresponding to the effective ingredient after removing the baseline drift.

[0130] It should be noted that, in the process of ensemble empirical mode decomposition, the last component signal is usually considered to be a residual signal, that is, a trend term signal representing the reference offset of the signal.

[0131] In this embodiment, in order to eliminate the interference introduced by different noise intensities and improve the accuracy of trend item signal extraction, the average value of all the last reference IMF components is calculated as the trend item signal according to the last reference IMF component corresponding to the best noise intensity to be introduced and the last reference IMF component corresponding to each noise intensity to be introduced before the best noise intensity to be introduced. For example, when the best noise intensity to be introduced is the 20th iteration, the signal values ​​of the same time in the last reference IMF component of the noise intensity to be introduced during the first 20 iterations are averaged, and the average values ​​at different times constitute a curve signal as the trend item signal; in order to overcome the influence of baseline offset on the detection of active ingredients of green pesticides, the digital signal of the active ingredient is subtracted from the trend item signal to obtain the digital signal corresponding to the active ingredient after removing the baseline drift. Among them, the example figure of the digital signal corresponding to the active ingredient after removing the baseline drift is as follows: Figure 3 shown.

[0132] At this point, this embodiment analyzes the decomposition effect and decomposition stability of the noise intensity to be introduced in each iteration, quantifies the introduction of evaluation values, and improves the extraction accuracy of the trend item of baseline drift caused by the introduced noise component, which is beneficial to accurately remove the trend item signal of baseline offset in the digital signal of the effective ingredient, improves the accuracy of the digital signal of the effective ingredient collected during the green pesticide detection process, and helps to further improve the accuracy of green pesticide effective ingredient detection.

[0133] S5, judging whether the active ingredient of the green pesticide to be tested meets the standard according to the digital signal corresponding to the active ingredient after removing the baseline drift.

[0134] In this embodiment, first, the digital signal corresponding to the effective ingredient after removing the baseline drift is denoised to obtain the denoised digital signal; then, in the denoised digital signal, a band in the preset nanometer range is selected as the detection wavelength band; the absorption peak area of ​​the detection wavelength and the total absorption peak area of ​​all wavelength bands in the digital signal are calculated, and the ratio of the absorption peak area to the total absorption peak area is used as the effective ingredient detection index; when the effective ingredient detection index is greater than the preset empirical threshold, it is determined that the effective ingredient of the green pesticide to be detected meets the standard, otherwise, it is determined that the effective ingredient of the green pesticide to be detected does not meet the standard. The implementation process of signal denoising is a prior art, which is not within the protection scope of the present invention and will not be elaborated in detail here.

[0135] It should be noted that, taking nicotine insecticides among green pesticides as an example, the preset nanometer range of the detection wavelength is 260 to 272 nanometers, within which the nicotine content range can be detected; the preset empirical threshold can be set to 0.7, and the implementer can set the size of the empirical threshold according to the specific actual situation.

[0136] This concludes the present embodiment.

[0137] An embodiment of the present invention provides a method for detecting the active ingredient of green pesticides. The method decomposes the active ingredient signal data in the data acquisition stage, improves the method for determining the intensity of the noise to be introduced by the parameter in the EEMD algorithm, and is conducive to overcoming the influence of baseline offset on the active ingredient signal data, and obtaining more accurate active ingredient signal data, which helps to further improve the accuracy of green pesticide active ingredient detection.

[0138] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for detecting active ingredients of green pesticides, characterized in that: The following steps are involved: The digital signal of the active ingredient in the green pesticide to be detected is obtained, and the digital signal is preliminarily decomposed to obtain each initial IMF component of the active ingredient; wherein the horizontal axis of the digital signal is time, and the vertical axis is absorbance; According to the fluctuation characteristics of each initial IMF component, the effective component characteristic value of each initial IMF component is determined, and then the fluctuation degree of the difference between any two effective component characteristic values ​​is analyzed to obtain the initial noise intensity to be introduced; the initial noise intensity to be introduced is iteratively reduced in sequence until the noise intensity to be introduced is not greater than 0, and the noise intensity to be introduced at each iteration is obtained; A noise component signal is randomly generated according to the noise intensity to be introduced at each iteration; based on each initial IMF component, the noise component signal at each iteration and each reference IMF component after the noise component signal at each iteration is introduced, the decomposition effect of the decomposition result after the noise component signal at each iteration is introduced is analyzed to determine the reasonableness of the noise intensity to be introduced at each iteration; According to the reasonableness of the noise intensity to be introduced in each iteration and any reference IMF component after the noise component signal in two adjacent iterations is introduced, the optimal noise intensity to be introduced is determined, and then the digital signal after removing the baseline drift corresponding to the effective component is determined; According to the digital signal corresponding to the active ingredient after removing the baseline drift, it is judged whether the active ingredient of the green pesticide to be tested meets the standard; According to the fluctuation characteristics of each initial IMF component, the effective component characteristic value of each initial IMF component is determined, including: For any initial IMF component, obtain a frequency spectrum of the initial IMF component and determine each peak point in the frequency spectrum; Select the peak point with the largest ordinate from each peak point in the spectrum diagram of the initial IMF component as the target peak point, calculate the abscissa difference between the two valley points closest to the target peak point, and determine it as the frequency component of the initial IMF component; Obtain each extreme point in the initial IMF component, calculate the average value of the absorbance difference of all adjacent extreme points in the initial IMF component, and determine the fluctuation degree of the distribution curve of the initial IMF component; The number of extreme value points of the initial IMF component is determined as the fluctuation frequency of the distribution curve of the initial IMF component; The effective component characteristic value of the initial IMF component is determined by combining the frequency component of the initial IMF component, the fluctuation degree of the distribution curve and the fluctuation frequency of the distribution curve; Analyze the fluctuation degree of the difference between the characteristic values ​​of any two effective components to obtain the initial noise intensity to be introduced, including: According to the effective component characteristic values ​​of each initial IMF component, the difference value of any two effective component characteristic values ​​is calculated, recorded as the characteristic difference value, and each characteristic difference value is obtained to form a characteristic difference value sequence; the variance of the characteristic difference value sequence is calculated, and the variance of the characteristic difference value sequence is used as the initial noise intensity to be introduced; According to each initial IMF component, the noise component signal at each iteration, and each reference IMF component after the noise component signal at each iteration is introduced, the decomposition effect of the decomposition result after the noise component signal at each iteration is introduced is analyzed to determine the reasonableness of the noise intensity to be introduced at each iteration, including: For the noise component signal at any iteration, according to the absorbance difference between each initial IMF component and the noise component signal at this iteration, the degree to which the noise component signal participates in signal decomposition is analyzed, and the first reasonable factor of the noise intensity to be introduced at this iteration is determined; Determine the effective component characteristic values ​​of each reference IMF component after the noise component signal at this iteration is introduced; analyze the independent distribution characteristics of the reference IMF components according to the characteristic difference values ​​of any two reference IMF components after the noise component signal at this iteration is introduced, and determine the second reasonable factor of the noise intensity to be introduced at this iteration; According to the time of each extreme point in each reference IMF component after the noise component signal at this iteration is introduced, the degree of distinction between the high-frequency and low-frequency components of the reference IMF component is analyzed to determine the third reasonable factor of the noise intensity to be introduced at this iteration; The rationality of the intensity of the noise to be introduced at the iteration is determined by combining the first rationality factor, the second rationality factor and the third rationality factor of the intensity of the noise to be introduced at the iteration.

2. A green pesticide active ingredient detection method according to claim 1, characterized in that: The steps of obtaining each reference IMF component include: For any iteration, determine the noise component signal of the noise intensity to be introduced at the iteration, introduce the noise component signal into the digital signal of the effective component, and obtain the digital signal of the noise component signal introduced at the iteration; Performing collective empirical mode decomposition on the digital signal of the noise component signal introduced at the iteration to obtain various IMF components, and using the various IMF components as various reference IMF components after the noise component signal at the iteration is introduced.

3. A green pesticide active ingredient detection method according to claim 1, characterized in that: The calculation formula of the first reasonable factor of the noise intensity to be introduced in this iteration is: Where R is the first reasonable factor of the noise intensity to be introduced at this iteration, u is the number of initial IMF components, e is the serial number of the initial IMF component, N is the number of data points in the initial IMF component or noise component signal, y is the serial number of the data point in the initial IMF component or noise component signal, T y o is the absorbance of the yth data point in the noise component signal at this iteration, T y e is the absorbance of the yth data point in the eth initial IMF component.

4. A green pesticide active ingredient detection method according to claim 1, characterized in that: According to the time of each extreme point in each reference IMF component after the noise component signal at this iteration is introduced, the degree of distinction between high-frequency and low-frequency components of the reference IMF component is analyzed to determine the third reasonable factor of the intensity of the noise to be introduced at this iteration, including: For any reference IMF component after the noise component signal of this iteration is introduced, the time difference between all two adjacent extreme value points in the reference IMF component is determined, recorded as the time difference value, and the average value of each time difference value is calculated as the frequency degree of the reference IMF component; According to the frequency levels of each reference IMF component after the noise component signal is introduced in this iteration, the difference between the frequency levels of two adjacent reference IMF components is calculated and recorded as the frequency level difference value, and the average value of all frequency level difference values ​​is used as the third reasonable factor of the noise intensity to be introduced in this iteration.

5. A green pesticide active ingredient detection method according to claim 1, characterized in that: According to the reasonableness of the noise intensity to be introduced in each iteration and any reference IMF component after the noise component signal in two adjacent iterations is introduced, the optimal noise intensity to be introduced is determined, including: According to any reference IMF component after the noise component signal of the current iteration and the previous iteration is introduced, the component stability after the noise component signal of the current iteration is introduced is calculated; Determine the introduction evaluation value of the noise intensity to be introduced at the current iteration by combining the rationality of the noise intensity to be introduced at the current iteration and the component stability after the noise component signal at the current iteration is introduced; The noise intensity to be introduced corresponding to the maximum introduction evaluation value is taken as the optimal noise intensity to be introduced.

6. A method for detecting active ingredients of green pesticides according to claim 5, characterized in that: The calculation formula for the component stability after the noise component signal at the kth iteration is introduced is: Where W k is the component stability after the noise component signal is introduced at the kth iteration, c is the sequence number of the data point in the reference IMF component, x is the number of data points in the reference IMF component, and s dc k is the absorbance of the cth data point in the dth reference IMF component after the noise component signal at the kth iteration is introduced, s dc k-1 is the absorbance of the cth data point in the dth reference IMF component after the noise component signal at the k-1th iteration is introduced, and || is the absolute value function.

7. A method for detecting active ingredients of green pesticides according to claim 5, characterized in that: According to the digital signal corresponding to the active ingredient after removing the baseline drift, it is judged whether the active ingredient of the green pesticide to be tested meets the standard, including: De-noising the digital signal corresponding to the effective component after removing the baseline drift to obtain a de-noised digital signal; in the de-noised digital signal, selecting a band in a preset nanometer range as a detection wavelength band; Calculate the absorption peak area of ​​the detection wavelength and the total absorption peak area of ​​all wavelength bands in the digital signal, and use the ratio of the absorption peak area to the total absorption peak area as the effective component detection index; When the active ingredient detection index is greater than the preset empirical threshold, it is determined that the active ingredient of the green pesticide to be detected meets the standard; otherwise, it is determined that the active ingredient of the green pesticide to be detected does not meet the standard.