A method for in-situ quality grading of platycladus orientalis seedlings based on electrical impedance spectroscopy

Abstract

The application belongs to the technical field of seedling quality grading, and discloses a quality grading method for Platycladus orientalis seedlings in situ based on electrical impedance spectroscopy, which comprises the following steps: a Platycladus orientalis seedling in-situ measurement device is built for non-destructive measurement of root EIS in situ; an EIS-100 system is used to measure EIS parameters; based on a GSFS-FCM algorithm, input data are clustered, and the optimal cluster number K and the optimal cluster center coordinate matrix are output; according to the optimal cluster number K and the optimal cluster center coordinate matrix, the final membership degree is calculated and the classification is divided through the FCM algorithm, the quality grade to which the Platycladus orientalis seedling belongs is output, and the whole process from electrical impedance measurement to quality grading is completed. The above-mentioned grading method is adopted, the GSFS-FCM algorithm can automatically determine the classification number, automatically find the cluster center, reduce the iteration number of the algorithm, has high accuracy, and breaks away from the traditional grading method which relies on morphological parameters of the aboveground part.

Description

Technical Field

[0001] This invention relates to the field of seedling quality grading technology, and in particular to an in-situ grading method for Chinese arborvitae seedlings based on electrical impedance spectra. Background Technology

[0002] High-quality seedlings are the material foundation and prerequisite for the high-quality development of the Chinese arborvitae seedling industry. The quality of seedlings is directly related to the success or failure of Chinese arborvitae afforestation and greening, as well as operating costs and economic benefits. Scientific and reasonable seedling quality evaluation methods are an important means to ensure high-quality seedlings and have important research significance.

[0003] The demand for Chinese arborvitae seedlings is strong, and the industry's development continues to heat up. However, the seedling research and development cycle is long, inefficient, and costly. The seedling quality evaluation system is still incomplete. For a long time, the quality of Chinese arborvitae seedlings has been graded based on a few morphological indicators of the above-ground parts, such as seedling height and ground diameter. This not only makes it difficult to accurately reflect the true quality level of Chinese arborvitae seedlings, but also induces the production of Chinese arborvitae seedlings to unilaterally pursue the growth of seedling height and ground diameter, resulting in excessive fertilizer and water management. This leads to problems such as insufficient lignification of seedlings after transplanting, poor adaptability and stress resistance, low survival rate of afforestation, and poor performance in the later stages of afforestation, which seriously affects the quality of Chinese arborvitae afforestation and the healthy development of the Chinese arborvitae seedling industry.

[0004] To address the aforementioned issues, this invention proposes a method that combines in-situ non-destructive electrical impedance spectroscopy (EIS) measurement technology with an improved fuzzy C-means clustering algorithm to achieve accurate grading of Chinese arborvitae seedlings. Summary of the Invention

[0005] The purpose of this invention is to provide a method for in-situ quality grading of Chinese arborvitae seedlings based on electrical impedance spectra. By using an in-situ non-destructive electrical impedance spectra (EIS) measurement device to obtain key parameters, and then using an improved fuzzy C-means clustering algorithm to achieve accurate grading of Chinese arborvitae seedlings.

[0006] To achieve the above objectives, this invention provides an in-situ quality grading method for Platycladus orientalis seedlings based on electrical impedance spectra, implemented using an EIS-100 hardware system and an improved fuzzy C-means clustering algorithm, comprising the following steps: Step S1: Construct an in-situ measurement device for arborvitae seedlings to perform in-situ non-destructive EIS measurement of the electrical impedance spectrum of arborvitae seedling roots. Step S2: Based on the in-situ measurement device for Chinese arborvitae seedlings, the EIS parameters of the Chinese arborvitae root system are measured using the EIS-100 system. Step S3: The fuzzy C-means clustering algorithm GSFS-FCM, optimized based on the genetic sparrow family algorithm, is used to cluster the input data and output the optimal number of clusters K. And the optimal cluster center coordinate matrix; Step S4: Calculate the optimal cluster number K based on the GSFS-FCM algorithm. The optimal cluster center coordinate matrix is ​​used to perform the final membership degree calculation and category classification through the FCM algorithm, outputting the quality grade of each Chinese arborvitae seedling, thus completing the entire process from electrical impedance measurement to quality grading.

[0007] Preferably, in step S1, an in-situ measurement device for arborvitae seedlings is constructed for EIS in-situ non-destructive measurement of the root system of arborvitae seedlings. The specific process is as follows: Step S11: The in-situ measurement device for Chinese arborvitae seedlings includes a stainless steel needle electrode, a stainless steel base electrode, connecting wires, and an EIS-100 system. Step S12: First, insert the container and seedling, along with the soil, into the measuring stainless steel base electrode; then, insert the stainless steel needle electrode into the stem of the arborvitae seedling 2cm above the root-soil junction; the diameter of the stainless steel needle electrode is 0.3mm. Step S13: The measurement circuit includes a stem section, the contact surface between the stem and the soil, the root, the contact surface between the root and the soil, the soil, and the measuring base electrode; based on the connected circuit, EIS in-situ non-destructive measurement of the root system of the Chinese arborvitae seedling is realized.

[0008] Preferably, in step S2, based on the in-situ measurement device for Chinese arborvitae seedlings, the EIS parameters of the Chinese arborvitae seedling root system are measured using the EIS-100 system. The specific process is as follows: Step S21: Based on the in-situ measurement device for Chinese arborvitae seedlings, the EIS-100 system performs in-situ non-destructive measurement of the EIS parameters of the Chinese arborvitae seedling root system in the frequency range of 1.2kHz to 1MHz by a voltage excitation of 0.1V. Step S22, the principle of in-situ non-destructive EIS parameter measurement of the root system of Chinese arborvitae seedlings, is as follows: As the frequency of impedance measurement increases, the real and imaginary parts of the impedance exhibit a periodic variation pattern; the real part of the impedance represents the change in the root impedance of the Chinese arborvitae seedling, while the imaginary part represents the change in the capacitive reactance of the Chinese arborvitae seedling root system. At low frequencies, due to the large capacitive reactance of the cell wall, alternating current can only flow through the cell wall; as the frequency increases, the capacitive reactance of the cell membrane decreases, and the conductivity of the cell wall increases; the absolute values ​​of the real and imaginary parts of the roots of arborvitae seedlings with good quality are small, while the absolute values ​​of the data for the roots of arborvitae seedlings with poor quality are large. Step S23: Use Trifacta Wrangler software to clean the raw measurement data and calculate the derived parameters using LEVM 7 software; The derived parameters specifically include: series resistance R, distributed resistance R, relaxation time τ1, relaxation time distribution coefficient ψ1, distributed resistance R2, relaxation time τ2, relaxation time distribution coefficient ψ2, CPE relaxation time constant τ3, and CPE relaxation time distribution coefficient ψ3. Step S24: Based on the processed data, extract two EIS parameters and combine them with three morphological parameters as input data for the clustering algorithm to characterize the quality status of each seedling.

[0009] Preferably, in step S3, a fuzzy C-means clustering algorithm optimized based on the genetic sparrow family algorithm is used to cluster the input data and output the optimal number of clusters K. The specific process for finding the optimal cluster center coordinate matrix is ​​as follows: Step S31: Set the parameters of the GSFS-FCM algorithm, specifically including: maximum number of GSFS-FCM iterations, sparrow population size, sparrow family population size, crossover probability, number of elites, mutation probability, number of elite individuals, number of producers, number of sparrows that have detected danger, and enable minimum warning. Step S32: Construct a weight matrix for the parameters based on their correlation levels. , , , ..., ;in, The number of features, and have ; Step S33: Set the boundary of each solution space and determine the minimum and maximum values ​​of each solution space based on the constraints of the actual problem or prior knowledge. Step S34: Use minimum-maximum standardization, i.e. deviation standardization, to normalize the attribute variables and map the values ​​to the range [0,1]. Step S35: Perform iterative search based on the GSFS-FCM algorithm and output the optimal cluster number K. And the optimal cluster center coordinate matrix.

[0010] Preferably, in step S34, the normalization of the attribute variables is calculated as follows: (1); in, This represents the new value of variable x after normalization. The original value; To find the minimum value in the solution space; To find the maximum value in the solution space.

[0011] Preferably, in step S35, an iterative search is performed based on the GSFS-FCM algorithm to output the optimal number of clusters K. The specific process for finding the optimal cluster center coordinate matrix is ​​as follows: Step S351: Initialize the initial sparrow positions using the Halton sequence to ensure that the initial population is evenly distributed in the solution space; A variable-length chromosome encoding scheme is used to initialize multiple sparrow families; each family represents a complete clustering scheme, the chromosome length is variable, corresponding to the number of clusters K, and the gene value on the chromosome, i.e. the sparrow position, represents the coordinates of the cluster center; Step S352: When the current iteration number is less than the maximum iteration number, the number of members in the initial sparrow family is used as the cluster number, and the initial sparrow position is used as the initial cluster center; run the FCM algorithm to perform clustering and perform fuzzy division on all cypress seedling samples; Step S353: Calculate the fitness function, clustering effectiveness evaluation index FPI, and NCE for each sparrow family; wherein, the clustering effectiveness evaluation index FPI and NCE are used to evaluate the clustering quality from the perspective of the clarity and entropy of the fuzzy partition. Step S354: Sort by fitness value, clustering effectiveness evaluation index FPI and NCE, and extract the top 5% of elite sparrow families to enter the next generation population; for the discoverer sparrow families ranked 5%~20%, update the member positions of the discoverers and perform local fine search; for the joiner sparrow families ranked after 20%, update the member positions of the joiners and perform global exploration. Step S355: In order to adapt to the changes in chromosome length during evolution, this invention designs a crossover operator, randomly selects two sparrow families, deletes a segment of gene from the chromosome of one sparrow family, and inserts this segment of gene into a designated position on the chromosome of the other sparrow family, generating two new families and increasing population diversity. Step S356: Select the top 10%-20% of individuals with better fitness as scouts, update the member positions of the scouts, and perform vigilance behavior; perform family member mutation operation on a small portion of the scout families, randomly change the positions of some members, and introduce random perturbation. Step S357: Merge elite individuals, updated discoverers, and new individuals (including those generated by crossover) and scouts to form a new generation of sparrow family population; if the fitness function of the new sparrow family individuals is better, update the sparrow family population; Step S358: After the loop ends, output the optimal sparrow family found during the entire evolution process; decode the optimal sparrow family to obtain the globally optimal cluster size K. , corresponding to the optimal number of levels and the optimal cluster center, and corresponding to the standard feature vectors of each level.

[0012] Preferably, in step S354, the member positions of the discoverer are updated, as follows: (2); in, This indicates the updated location information; Indicates the first The sparrow in the first Location information in the dimension , The number of features; Indicates the current iteration number; Represents the maximum number of iterations; It is a random number; Indicates the warning value; Indicates a safe value; It is a random number that follows a normal distribution; Represent a Matrix; The matrix contains all elements equal to 1.

[0013] Preferably, in step S354, the member position of the joiner is updated, as follows: (3); in, It is the optimal location of the current discoverer; Indicates the worst position at present; Represent a A matrix, where each element is randomly assigned the value 1 or -1, and ;when When, it indicates the first The individual participant is at a disadvantage; if , representing the Each scrounger will randomly find a location near the optimal location to perform a local search.

[0014] Preferably, in step S356, the positions of the scout members are updated, as follows: (4); in, It is the current global optimal position; β, as the step size control parameter, is a random number that follows a normal distribution with a mean of 0 and a variance of 1. It is a random number; This is the current fitness value of the individual sparrow; and These are the current best and worst fitness values ​​globally, respectively. It is the smallest constant to avoid zero in the denominator.

[0015] Therefore, this invention adopts the above-mentioned in-situ arborvitae seedling quality grading method based on electrical impedance spectrum, and uses in-situ non-destructive electrical impedance spectrum (EIS) measurement equipment to obtain key parameters, realizing in-situ non-destructive testing, and getting rid of the traditional grading method that relies on the morphological parameters of the aboveground parts; the GSFS-FCM algorithm can automatically determine the number of classifications, automatically find the cluster centers, reduce the number of algorithm iterations, and has high accuracy.

[0016] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0017] Figure 1 This is a flowchart of an in-situ quality grading method for Platycladus orientalis seedlings based on electrical impedance spectra according to the present invention; Figure 2 This is a diagram of the in-situ measurement device for Chinese arborvitae seedlings according to the present invention; wherein, (a) is the actual measurement scenario; (b) is the electrode connection; (c) is the bottom electrode; and (d) is the measurement principle. Figure 3 This is a flowchart illustrating the implementation of the GSFS-FCM algorithm of this invention. Detailed Implementation

[0018] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0019] like Figure 1 As shown, an in-situ quality grading method for Platycladus orientalis seedlings based on electrical impedance spectra is implemented using an EIS-100 hardware system and an improved fuzzy C-means clustering algorithm, and includes the following steps: Step S1: Construct an in-situ measurement device for arborvitae seedlings to perform in-situ non-destructive measurement of electrical impedance spectra (EIS) of arborvitae seedling roots. Step S2: Based on the in-situ measurement device for Chinese arborvitae seedlings, the EIS parameters of the seedling root system are measured using the EIS-100 system; Step S3: The fuzzy C-means clustering algorithm (GSFS-FCM) optimized by the genetic sparrow family search algorithm is used to cluster the input data and output the optimal number of clusters K. And the optimal cluster center coordinate matrix; Step S4: Calculate the optimal cluster number K based on the GSFS-FCM algorithm. The optimal cluster center coordinate matrix is ​​used to perform the final membership degree calculation and category classification through the FCM algorithm, outputting the quality grade of each Chinese arborvitae seedling, thus completing the entire process from electrical impedance measurement to quality grading.

[0020] Example 1 Step S1: Construct an in-situ measurement device for the Chinese arborvitae seedlings, such as... Figure 2 As shown, this is an example of EIS in-situ non-destructive measurement of the root system of Chinese arborvitae seedlings.

[0021] Step S11: The in-situ measurement device for Chinese arborvitae seedlings includes a stainless steel needle electrode, a stainless steel base electrode, connecting wires, and an EIS-100 system.

[0022] Step S12: First, insert the container and seedling, along with the soil, into the measuring stainless steel base electrode; then, insert the stainless steel needle electrode (0.3mm in diameter) into the stem of the cypress seedling 2cm above the root-soil junction.

[0023] Step S13: The measurement circuit includes a stem section, the contact surface between the stem and the soil, the root, the contact surface between the root and the soil, the soil, and the measuring base electrode; based on the connected circuit, EIS in-situ non-destructive measurement of the root system of the Chinese arborvitae seedling is realized.

[0024] Step S2: Based on the in-situ measurement device for Chinese arborvitae seedlings, the EIS parameters of the seedling root system are measured using the EIS-100 system.

[0025] Step S21: Based on the in-situ measurement device for Chinese arborvitae seedlings, the EIS-100 system performs in-situ non-destructive measurement of the EIS parameters of the Chinese arborvitae seedling root system in the frequency range of 1.2kHz to 1MHz by a voltage excitation of 0.1V.

[0026] Step S22, the principle of in-situ non-destructive EIS parameter measurement of the root system of Chinese arborvitae seedlings, is as follows: Frequency is one of the most important factors affecting impedance characteristics. As the frequency of impedance measurement increases, the real and imaginary parts of the impedance exhibit a periodic variation. The real part of the impedance represents the change in the root impedance of the arborvitae seedlings, while the imaginary part represents the change in the capacitive reactance of the arborvitae seedlings' roots.

[0027] At low frequencies, due to the high capacitive reactance of the cell wall, alternating current can only flow through the cell wall. As the frequency increases, the capacitive reactance of the cell membrane begins to decrease, and the conductivity of the cell wall gradually increases. With increasing frequency, the absolute values ​​of the real and imaginary parts of the root system of good-quality arborvitae seedlings become smaller, while the absolute values ​​of the data for poor-quality arborvitae seedling root systems become larger.

[0028] These changes are due to polarization and relaxation phenomena, as well as the energy dissipation patterns at different interfaces (such as cell membranes, root / soil interfaces) and chambers (such as apoplasts, symplasts, soil pores, and soil particles) when alternating current passes through the sample.

[0029] Step S23: Use Trifacta Wrangler software to clean the raw measurement data and calculate the derived parameters using LEVM 7 software, specifically including: series resistance R, distributed resistance R, relaxation time τ1, relaxation time distribution coefficient ψ1, distributed resistance R2, relaxation time τ2, relaxation time distribution coefficient ψ2, CPE relaxation time constant τ3, and CPE relaxation time distribution coefficient ψ3.

[0030] Step S24: Based on the processed data, extract two representative EIS parameters and combine them with three morphological parameters as input data for the clustering algorithm to characterize the quality status of each seedling.

[0031] Step S3: The fuzzy C-means clustering algorithm (GSFS-FCM algorithm) optimized based on the genetic sparrow family algorithm is used to cluster the input data and output the optimal number of clusters K. And the optimal cluster center coordinate matrix.

[0032] The GSFS-FCM algorithm, through the fusion of genetic algorithms and sparrow algorithms, combines the powerful global search capability of genetic algorithms with the fast local search speed of sparrow algorithms. By using variable-length sparrow families, it solves the problems of the FCM clustering algorithm, which is unable to determine the number of clusters and is sensitive to the initial cluster centers. Furthermore, it optimizes the details of the sparrow algorithm through various strategies, such as... Figure 3 As shown.

[0033] Step S31: Set the parameters of the GSFS-FCM algorithm, including: maximum number of GSFS-FCM iterations, sparrow population size, sparrow family population size, crossover probability, number of elites, mutation probability, number of elite individuals, number of producers, number of sparrows that have detected danger, and enable minimum warning.

[0034] Step S32: Construct a weight matrix for the parameters based on their correlation levels. , , , ..., ;in, The number of features, and have .

[0035] Step S33: Set the boundaries of each solution space. Usually, the minimum and maximum values ​​of each solution space are determined based on the constraints of the actual problem or prior knowledge.

[0036] Step S34: Use min-maximum standardization, i.e., deviation standardization, to normalize the attribute variables, mapping the values ​​to the range [0,1]. The new variable values ​​are shown below: (1); in, This represents the new value of variable x after normalization. The original value; To find the minimum value in the solution space; To find the maximum value in the solution space.

[0037] Step S35: Perform iterative search based on the GSFS-FCM algorithm and output the optimal cluster number K. The specific process for finding the optimal cluster center coordinate matrix is ​​as follows: Step S351: Initialize the initial sparrow positions using the Halton sequence to ensure that the initial population is evenly distributed in the solution space, avoiding the initial positions of the population individuals being concentrated near local extreme points and prematurely falling into local optima; A variable-length chromosome encoding scheme is used to initialize multiple sparrow families. Each family represents a complete clustering scheme, with a variable chromosome length (corresponding to the number of clusters K), and the gene values ​​(sparrow positions) on the chromosome represent the coordinates of the cluster centers.

[0038] Step S352: When the current iteration number is less than the maximum iteration number G, the number of members in the initial sparrow family is used as the cluster number, and the initial sparrow position is used as the initial cluster center; run the FCM algorithm to perform clustering and perform fuzzy division on all cypress seedling samples.

[0039] Step S353: Calculate the fitness function, clustering effectiveness evaluation index FPI, and NCE for each sparrow family; wherein, the clustering effectiveness evaluation index FPI and NCE are used to evaluate the clustering quality from the perspective of the clarity and entropy of the fuzzy partition.

[0040] Step S354: Sort by fitness value, cluster effectiveness evaluation index FPI and NCE, and extract the top 5% of elite sparrow families to enter the next generation population; for the discoverer sparrow families ranked 5% to 20%, update the member positions of the discoverers according to the specific formula (2) and perform local fine search; for the joiner sparrow families ranked after 20%, update the member positions of the joiners according to the formula (3) and perform global exploration.

[0041] (2); in, Indicates the first The sparrow in the first Location information in the dimension , The number of features; Indicates the current iteration number; Represents the maximum number of iterations; It is a random number; Indicates the warning value; Indicates a safe value; It is a random number that follows a normal distribution; Represent a Matrix; The matrix contains all elements equal to 1.

[0042] (3); in, It is the optimal location of the current discoverer; Indicates the worst position at present; Represent a A matrix, where each element is randomly assigned the value 1 or -1, and ;when When, it indicates the first The individual participant is at a disadvantage; if , representing the Each scrounger will randomly find a location near the optimal location to perform a local search.

[0043] Step S355: In order to adapt to the changes in chromosome length during evolution, this invention designs a crossover operator, randomly selects two sparrow families, deletes a segment of gene from the chromosome of one sparrow family, and inserts this segment of gene into a designated position on the chromosome of the other sparrow family, generating two new families and increasing population diversity.

[0044] Step S356: Select the top 10%-20% of individuals with better fitness as scouts, update the member positions of the scouts according to formula (4), and perform vigilance behavior; perform family member mutation operation on a small portion of the scout families of 5%-2%, randomly change the positions of some members, and introduce random perturbation.

[0045] (4); in, It is the current global optimal position; β, as the step size control parameter, is a random number that follows a normal distribution with a mean of 0 and a variance of 1. It is a random number; This is the current fitness value of the individual sparrow; and These are the current best and worst fitness values ​​globally, respectively. It is the smallest constant to avoid zero in the denominator.

[0046] when This indicates that the sparrows are on the fringes of their population and are extremely vulnerable to predators. This indicates that the sparrow in this location is in the best and safest position in the population; This indicates that sparrows in the middle of the population are aware of the danger and need to move closer to other sparrows to minimize their risk of being preyed upon.

[0047] Step S357: Merge elite individuals, updated discoverers, and new individuals (including those generated by crossover) and scouts to form a new generation of sparrow family population; if the fitness function of the new sparrow family individuals is better, update the sparrow family population.

[0048] Step S358: After the loop ends, output the optimal sparrow family found during the entire evolution process; decode the optimal sparrow family to obtain the globally optimal cluster size K. (Optimal number of levels) and optimal cluster centers (standard feature vectors of each level).

[0049] Step S4: Calculate the optimal clustering number K for all the arborvitae seedlings to be graded using the feature data of all seedlings. The optimal cluster center coordinate matrix is ​​used to perform the final membership degree calculation and category classification through the FCM algorithm, outputting the quality grade of each Chinese arborvitae seedling, thus completing the entire process from electrical impedance measurement to quality grading.

[0050] Example 2 In this embodiment, two-year-old arborvitae seedlings were selected and cultivated in brown soil (pH=8.0) in the same area with an organic matter content of 3%. The seedlings were randomly dug up and transported back to the laboratory for morphological and EIS parameter measurements.

[0051] First, when measuring the EIS parameters of the Chinese arborvitae seedlings, the container and the seedling with soil were inserted into the base of the measuring stainless steel electrode. Then, the stainless steel needle electrode was inserted into the stem at four angles (0°, 90°, 180°, and 270°) perpendicular to the stem (without penetrating it). The electrical impedance spectrum was measured four times, and the average value of the four measurements was taken to obtain the raw data of the seedling EIS parameter values, as shown in Table 1.

[0052] Table 1. Raw data of EIS parameter values ​​for Chinese arborvitae seedlings

[0053] Then, Trifacta Wrangler was used for data cleaning, and the derived parameters were calculated using LEVM 7 software. Two parameters and three morphological parameters of EIS were extracted as input data, including the phase angle at 4 kHz and the tangent value at 400 kHz, as well as stem diameter, plant height and number of branches.

[0054] Secondly, hierarchical clustering was performed using the GSFS-FCM algorithm to obtain the optimal number of clusters K. The specific process for finding the optimal cluster center coordinate matrix is ​​as follows: (1) Set the parameters of the GSFS-FCM algorithm as shown in Table 2.

[0055] Table 2 Parameter settings for the GSFS-FCM algorithm

[0056] (2) Construct a weight matrix for the parameters based on their correlation levels. , , , , .

[0057] (3) Set the boundary of each dimension of the solution space, denoted as: (V1 min V1 max ); (V2 min V2 max ); (V3 min V3 max ); (V4 min V4 max ); (V5 min V5 max ).

[0058] (4) Use minimum-maximum standardization, i.e. deviation standardization method, to normalize the attribute variables and map the values ​​to the range [0,1].

[0059] Finally, the quality grading results of the Chinese arborvitae seedlings were obtained, as shown in Table 3.

[0060] Table 3. Quality grading results of Platycladus orientalis seedlings based on GSFS-FCM clustering algorithm

[0061] Therefore, this invention adopts the above-mentioned in-situ arborvitae seedling quality grading method based on electrical impedance spectrum, and uses in-situ non-destructive electrical impedance spectrum (EIS) measurement equipment to obtain key parameters, realizing in-situ non-destructive testing, and getting rid of the traditional grading method that relies on the morphological parameters of the aboveground parts; the GSFS-FCM algorithm can automatically determine the number of classifications, automatically find the cluster centers, reduce the number of algorithm iterations, and has high accuracy.

[0062] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for in-situ quality grading of Platycladus orientalis seedlings based on electrical impedance spectra, characterized in that, Implemented using the EIS-100 hardware system and an improved fuzzy C-means clustering algorithm, the process includes the following steps: Step S1: Construct an in-situ measurement device for arborvitae seedlings to perform in-situ non-destructive EIS measurement of the electrical impedance spectrum of arborvitae seedling roots. Step S2: Based on the in-situ measurement device for Chinese arborvitae seedlings, the EIS parameters of the Chinese arborvitae root system are measured using the EIS-100 system. Step S3: The fuzzy C-means clustering algorithm GSFS-FCM, optimized based on the genetic sparrow family algorithm, is used to cluster the input data and output the optimal number of clusters K. And the optimal cluster center coordinate matrix; Step S4: Calculate the optimal cluster number K based on the GSFS-FCM algorithm. The optimal cluster center coordinate matrix is ​​used to perform the final membership degree calculation and category classification through the FCM algorithm, outputting the quality grade of each Chinese arborvitae seedling, thus completing the entire process from electrical impedance measurement to quality grading.

2. The method for in-situ quality grading of Platycladus orientalis seedlings based on electrical impedance spectra according to claim 1, characterized in that, In step S1, an in-situ measurement device for arborvitae seedlings is constructed for EIS in-situ non-destructive measurement of the root system of arborvitae seedlings. The specific process is as follows: Step S11: The in-situ measurement device for Chinese arborvitae seedlings includes a stainless steel needle electrode, a stainless steel base electrode, connecting wires, and an EIS-100 system. Step S12: First, insert the container and seedling, along with the soil, into the measuring stainless steel base electrode; then, insert the stainless steel needle electrode into the stem of the arborvitae seedling 2cm above the root-soil junction; the diameter of the stainless steel needle electrode is 0.3mm. Step S13: The measurement circuit includes a stem section, the contact surface between the stem and the soil, the root, the contact surface between the root and the soil, the soil, and the measuring base electrode; based on the connected circuit, EIS in-situ non-destructive measurement of the root system of the Chinese arborvitae seedling is realized.

3. The method for in-situ quality grading of Platycladus orientalis seedlings based on electrical impedance spectra according to claim 1, characterized in that, In step S2, based on the in-situ measurement device for Chinese arborvitae seedlings, the EIS parameters of the Chinese arborvitae seedling root system are measured using the EIS-100 system. The specific process is as follows: Step S21: Based on the in-situ measurement device for Chinese arborvitae seedlings, the EIS-100 system performs in-situ non-destructive measurement of the EIS parameters of the Chinese arborvitae seedling root system in the frequency range of 1.2kHz to 1MHz by a voltage excitation of 0.1V. Step S22, the principle of in-situ non-destructive EIS parameter measurement of the root system of Chinese arborvitae seedlings, is as follows: As the frequency of impedance measurement increases, the real and imaginary parts of the impedance exhibit a periodic variation pattern; the real part of the impedance represents the change in the root impedance of the Chinese arborvitae seedling, while the imaginary part represents the change in the capacitive reactance of the Chinese arborvitae seedling root system. At low frequencies, due to the large capacitive reactance of the cell wall, alternating current can only flow through the cell wall; as the frequency increases, the capacitive reactance of the cell membrane decreases, and the conductivity of the cell wall increases; the absolute values ​​of the real and imaginary parts of the roots of arborvitae seedlings with good quality are small, while the absolute values ​​of the data for the roots of arborvitae seedlings with poor quality are large. Step S23: Use Trifacta Wrangler software to clean the raw measurement data and calculate the derived parameters using LEVM7 software; The derived parameters specifically include: series resistance R, distributed resistance R, relaxation time τ1, relaxation time distribution coefficient ψ1, distributed resistance R2, relaxation time τ2, relaxation time distribution coefficient ψ2, CPE relaxation time constant τ3, and CPE relaxation time distribution coefficient ψ3. Step S24: Based on the processed data, extract two EIS parameters and combine them with three morphological parameters as input data for the clustering algorithm to characterize the quality status of each seedling.

4. The method for in-situ quality grading of Platycladus orientalis seedlings based on electrical impedance spectra according to claim 1, characterized in that, In step S3, a fuzzy C-means clustering algorithm optimized based on the genetic sparrow family algorithm is used to cluster the input data and output the optimal number of clusters K. The specific process for finding the optimal cluster center coordinate matrix is ​​as follows: Step S31: Set the parameters of the GSFS-FCM algorithm, specifically including: maximum number of GSFS-FCM iterations, sparrow population size, sparrow family population size, crossover probability, number of elites, mutation probability, number of elite individuals, number of producers, number of sparrows that have detected danger, and enable minimum warning. Step S32: Construct a weight matrix for the parameters based on their correlation levels. , , , ..., ;in, The number of features, and have ; Step S33: Set the boundary of each solution space and determine the minimum and maximum values ​​of each solution space based on the constraints of the actual problem or prior knowledge. Step S34: Use minimum-maximum standardization, i.e. deviation standardization, to normalize the attribute variables and map the values ​​to the range [0,1]. Step S35: Perform iterative search based on the GSFS-FCM algorithm and output the optimal cluster number K. And the optimal cluster center coordinate matrix.

5. The method for in-situ quality grading of Platycladus orientalis seedlings based on electrical impedance spectra according to claim 4, characterized in that, In step S34, the normalization of attribute variables is calculated as follows: （1）； in, This represents the new value of variable x after normalization. The original value; To find the minimum value in the solution space; To find the maximum value in the solution space.

6. The method for in-situ quality grading of Platycladus orientalis seedlings based on electrical impedance spectra according to claim 4, characterized in that, In step S35, an iterative search is performed based on the GSFS-FCM algorithm to output the optimal number of clusters K. The specific process for finding the optimal cluster center coordinate matrix is ​​as follows: Step S351: Initialize the initial sparrow positions using the Halton sequence to ensure that the initial population is evenly distributed in the solution space; A variable-length chromosome encoding scheme is used to initialize multiple sparrow families; each family represents a complete clustering scheme, the chromosome length is variable, corresponding to the number of clusters K, and the gene value on the chromosome, i.e. the sparrow position, represents the coordinates of the cluster center; Step S352: When the current iteration number is less than the maximum iteration number, the number of members in the initial sparrow family is used as the cluster number, and the initial sparrow position is used as the initial cluster center; run the FCM algorithm to perform clustering and perform fuzzy division on all cypress seedling samples; Step S353: Calculate the fitness function, clustering effectiveness evaluation index FPI, and NCE for each sparrow family; wherein, the clustering effectiveness evaluation index FPI and NCE are used to evaluate the clustering quality from the perspective of the clarity and entropy of the fuzzy partition. Step S354: Sort by fitness value, clustering effectiveness evaluation index FPI and NCE, and extract the top 5% of elite sparrow families to enter the next generation population; for the discoverer sparrow families ranked 5%~20%, update the member positions of the discoverers and perform local fine search; for the joiner sparrow families ranked after 20%, update the member positions of the joiners and perform global exploration. Step S355: In order to adapt to the changes in chromosome length during evolution, this invention designs a crossover operator, randomly selects two sparrow families, deletes a segment of gene from the chromosome of one sparrow family, and inserts this segment of gene into a designated position on the chromosome of the other sparrow family, generating two new families and increasing population diversity. Step S356: Select the top 10%-20% of individuals with better fitness as scouts, update the member positions of the scouts, and perform vigilance behavior; perform family member mutation operation on a small portion of the scout families, randomly change the positions of some members, and introduce random perturbation. Step S357: Merge elite individuals, updated discoverers, and new individuals (including those generated by crossover) and scouts to form a new generation of sparrow family population; if the fitness function of the new sparrow family individuals is better, update the sparrow family population; Step S358: After the loop ends, output the optimal sparrow family found during the entire evolution process; decode the optimal sparrow family to obtain the globally optimal cluster size K. , corresponding to the optimal number of levels and the optimal cluster center, and corresponding to the standard feature vectors of each level.

7. The method for in-situ quality grading of Platycladus orientalis seedlings based on electrical impedance spectra according to claim 6, characterized in that, In step S354, the member positions of the discoverer are updated, as follows: （2）； in, This indicates the updated location information; Indicates the first The sparrow in the first Location information in the dimension , The number of features; Indicates the current iteration number; Represents the maximum number of iterations; It is a random number; Indicates the warning value; Indicates a safe value; It is a random number that follows a normal distribution; Represent a Matrix; The matrix contains all elements equal to 1.

8. The method for in-situ quality grading of Platycladus orientalis seedlings based on electrical impedance spectra according to claim 7, characterized in that, In step S354, the member position of the joiner is updated, as follows: （3）； in, It is the optimal location of the current discoverer; Indicates the worst position at present; Represent a A matrix, where each element is randomly assigned the value 1 or -1, and , For transpose operation; when When, it indicates the first The individual participant is at a disadvantage; if , representing the Each scrounger will randomly find a location near the optimal location to perform a local search.

9. A method for in-situ quality grading of Platycladus orientalis seedlings based on electrical impedance spectra according to claim 8, characterized in that, In step S356, the positions of the scout members are updated, as follows: （4）； in, It is the current global optimal position; β, as the step size control parameter, is a random number that follows a normal distribution with a mean of 0 and a variance of 1. It is a random number; This is the current fitness value of the individual sparrow; and These are the current best and worst fitness values ​​globally, respectively. It is the smallest constant to avoid zero in the denominator.

Images