Puncture operation tissue recognition system based on bioelectrical impedance

Abstract

The invention belongs to the field of biomedical engineering, and particularly relates to a puncture operation tissue recognition system based on bioelectrical impedance, which comprises a dual-electrode impedance needle used for collecting electric signal data in real time; the feedback display equipment is internally integrated with a trained random forest model and is used for processing the electric signal data according to the trained random forest model to obtain a classification result; the multi-frequency bioelectrical impedance analyzer is used for displaying the real-time acquired waveform and displaying the classification result of the random forest model at the same time; the dual-electrode impedance needle comprises a needle tube structure and an electrode structure located at the head of the needle tube structure. According to the invention, electric signals of different tissues are effectively obtained, and measurement errors and signal interference are reduced.

Description

Technical Field

[0001] The invention belongs to the field of biomedical engineering, and in particular relates to a puncture surgery tissue identification system based on bioelectrical impedance. Background Art

[0002] Puncture surgery is one of the important means of treating diseases and protecting human health. It is used to introduce catheters into the patient's body, obtain tissue samples or perform treatments. The surgeon's lack of experience, wrist fatigue and hand tremors after long hours of work, changes in the position of organs during surgery, and complex structures of the surgical site can cause the surgical puncture to accidentally injure other organs, nerves or blood vessels. During puncture surgery, doctors need to accurately locate and identify the target tissue, and traditional ultrasound or CT guidance techniques may be limited.

[0003] In recent years, bioimpedance technology has developed rapidly. Bioimpedance measurement technology has been widely used in basic science and clinical science. By extracting the modulus and phase angle of biological tissue signals, the different conduction and impedance characteristics of current due to the structures, components and biological health conditions of skin, muscle, fat, bone, tumor, etc. are used to clearly evaluate the tissue state. Specifically for the present invention, by measuring the impedance value of biological tissue in the puncture area, the clinical state of the organism can be analyzed from the measurement results, and the physiological state of the biological tissue can also be inferred. This technology has the characteristics of being cheap, safe, non-toxic and harmless, and easy to operate.

[0004] In the prior art, CN 114271912 A discloses a medical puncture surgery navigation system and method, which uses CT scanning and QR code technology to provide navigation in puncture surgery, and displays information such as the depth and length of the puncture needle through a display. The technology relies on CT scanning and image navigation, and there are radiation risks, poor real-time performance, and insufficient tissue recognition accuracy. Summary of the invention

[0005] In order to solve the above problems and improve the accuracy and safety of puncture surgery, the present invention provides a puncture surgery tissue identification system based on bioelectrical impedance, comprising:

[0006] Dual-electrode impedance needle for real-time acquisition of electrical signal data;

[0007] A feedback display device, which has a trained random forest model integrated therein, is used to process the electrical signal data according to the trained random forest model to obtain a classification result;

[0008] Multi-frequency bioelectrical impedance analyzer, used to display real-time acquisition waveforms and classification results of random forest models;

[0009] Among them, the dual-electrode impedance needle includes a needle tube structure and an electrode structure located at the head of the needle tube structure; the needle tube structure includes an outer needle tube 1, an insulating tube layer 2, and an inner needle core 3 from the inside to the outside; the electrode structure includes an excitation electrode layer 6, an insulating layer 5, and a measuring electrode layer 4 from the inside to the outside, the excitation electrode layer 6 is connected to an excitation electrode external cable 8, and the measuring electrode layer 4 is connected to a measuring electrode external cable 7; wherein the outer needle tube 1 is connected to the measuring electrode layer 4, and the inner needle core 3 is connected to the excitation electrode layer 6.

[0010] Beneficial effects of the present invention:

[0011] The present invention develops a threshold-based impedance mutation detection algorithm, that is, a statistical threshold method is used to process data and monitor the mutation of signals in real time, which can quickly identify sudden changes in tissue types during the puncture process and provide timely feedback on the change information.

[0012] Traditional single-electrode designs often fail to provide accurate tissue electrical property data, leading to inaccurate tissue identification. The present invention uses dual-electrode impedance needles with positive inside and negative outside to more effectively obtain electrical signals from different tissues and reduce measurement errors and signal interference. The dual-electrode design improves the reliability and accuracy of the data, and the difference in different impedance values ​​sampled by the positive inside and negative outside structure is more stable. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 It is a schematic diagram of the architecture of the puncture surgery tissue identification system based on bioelectrical impedance of the present invention;

[0014] Figure 2 This is a schematic diagram of a dual-electrode impedance needle of the present invention;

[0015] Figure 3 The training process flow chart of the random forest model of the present invention;

[0016] Figure 4 Schematic diagram of the random forest model of the present invention;

[0017] Among them, 1-outer needle tube, 2-insulating tube layer, 3-inner needle core, 4-measuring electrode layer, 5-insulating layer, 6-excitation electrode layer, 7-measuring electrode external cable, 8-excitation electrode external cable. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] The present invention provides a puncture surgery tissue recognition system based on bioelectrical impedance, such as Figure 1 , Figure 2 As shown, including:

[0020] Dual-electrode impedance needle for real-time acquisition of electrical signal data;

[0021] A feedback display device, which has a trained random forest model integrated therein, is used to process the electrical signal data according to the trained random forest model to obtain a classification result;

[0022] Multi-frequency bioelectrical impedance analyzer, used to display real-time acquisition waveforms and classification results of random forest models;

[0023] Among them, the dual-electrode impedance needle includes a needle tube structure and an electrode structure located at the head of the needle tube structure; the needle tube structure includes an outer needle tube 1, an insulating tube layer 2, and an inner needle core 3 from the inside to the outside; the electrode structure includes an excitation electrode layer 6, an insulating layer 5, and a measuring electrode layer 4 from the inside to the outside, the excitation electrode layer 6 is connected to an excitation electrode external cable 8, and the measuring electrode layer 4 is connected to a measuring electrode external cable 7; wherein the outer needle tube 1 is connected to the measuring electrode layer 4, and the inner needle core 3 is connected to the excitation electrode layer 6.

[0024] Specifically, since the measurement position of the multi-frequency bioelectrical impedance analyzer is small and cannot directly and continuously measure the electrical signals of biological tissues, a dual-electrode impedance needle is externally connected to the multi-frequency bioelectrical impedance analyzer to obtain the electrical signal data of different tissues. Subsequently, these data will be analyzed by the external multi-frequency bioelectrical impedance analyzer.

[0025] Specifically, the present invention designs a dual-electrode impedance needle to achieve the purpose of visualizing the puncture position through bioelectrical impedance measurement technology, thereby reducing the dependence of the operation on the experience of clinicians and improving the success rate of puncture. Figure 2 As shown. The needle tube structure gradually tapers from the head to the tail, wherein the outer needle tube 1 is the external structure of the needle tube structure, which provides the necessary support and conductive path for the puncture operation; the outer needle tube 1 is connected to the measuring electrode layer 4 and participates in the measurement process of bioelectrical impedance. The inner needle core 3 is located inside the outer needle tube 1, which provides additional structural support and conductive path, and is also used for puncture. The inner needle core 3 is connected to the excitation electrode 6, which is used to stimulate tissue response in bioelectrical impedance measurement. The insulating tube layer 2 is located between the outer needle tube 1 and the inner needle core 3 to ensure the isolation of different circuit parts, ensure the accuracy of signal transmission, prevent current leakage, and protect the internal circuit.

[0026] The electrode structure is arranged at the head of the needle tube structure, wherein the excitation electrode layer 6 is partially placed inside the inner needle core 3. The excitation electrode layer 6 is responsible for evenly distributing the electrical signal in the tissue to be tested, stimulating the change of bioelectrical impedance, and providing the necessary electrical signal for measurement. The measuring electrode layer 4 is located outside the excitation electrode layer 6, and is used to collect and process the test signal stimulated by the excitation electrode, accurately measure the electrical impedance of the tissue, and provide data support for puncture. The insulating layer 5 is located between the excitation electrode layer 6 and the measuring electrode layer 4 to ensure accurate transmission of the signal and prevent current leakage.

[0027] The excitation electrode external cable 8 and the measurement electrode external cable 7 are connected to the excitation electrode layer and the measurement electrode layer respectively, and are used to transmit signals to external equipment for further processing and analysis to achieve circuit integrity.

[0028] Specifically, Figure 3 As shown in Figure 1, the training process of the random forest model includes the following steps:

[0029] S1. The electrical signal datasets of five types of biological tissues, namely skin, muscle, fat, bone, and tumor, are collected through a dual-electrode impedance needle. Each data in the electrical signal dataset of each type of biological tissue includes frequency, resistance, reactance, and phase.

[0030] Specifically, step 1 includes:

[0031] S11. Maintain an environment with a temperature of 20-25°C and a humidity of 40-60% rh.

[0032] S12. In the frequency range of 1kHz to 100MHz, sample p frequency points through a multi-frequency bioelectrical impedance analyzer, collect the frequency, resistance, reactance and phase of the same skin sample at each frequency point, and form an electrical signal data set Y of the skin tissue pf ;

[0033] S13. In the frequency range of 1kHz to 100MHz, sample p frequency points through a multi-frequency bioelectrical impedance analyzer, and collect the frequency, resistance, reactance and phase of the same muscle sample at each frequency point; form an electrical signal data set Y of muscle tissue jr ;

[0034] S14. In the frequency range of 1kHz to 100MHz, sample p frequency points through a multi-frequency bioelectrical impedance analyzer, and collect the frequency, resistance, reactance and phase of the same fat sample at each frequency point; form an electrical signal data set Y of the fat tissue zf ;

[0035] S15. In the frequency range of 1kHz to 100MHz, sample p frequency points through a multi-frequency bioelectrical impedance analyzer, and collect the frequency, resistance, reactance and phase of the same bone sample at each frequency point; form an electrical signal data set Y of bone tissue gg ;

[0036] S16. In the frequency range of 1kHz to 100MHz, sample p frequency points by a multi-frequency bioelectrical impedance analyzer, and collect the frequency, resistance, reactance and phase of the same tumor sample at each frequency point; form an electrical signal data set Y of the tumor tissue zl .

[0037] In the embodiment of the present invention, Indicates the i=1, 2, ..., pth data in the electrical signal data set of biological tissue type L, Represent the frequency, resistance, reactance, and phase of the i-th data respectively; therefore, the electrical signal data set of biological tissue type L is expressed as

[0038]

[0039] Electrical signal dataset is a matrix of size p×q, q is the number of feature categories, and in the present invention, q=4. The biological tissue types L of the present invention include five categories: skin, muscle, fat, bone, and tumor. When L=pf, it means that the biological tissue type is skin, when L=jr, it means that the biological tissue type is muscle, when L=zf, it means that the biological tissue type is fat, when L=gg, it means that the biological tissue type is bone, and when L=zl, it means that the biological tissue type is tumor.

[0040] S2. Based on the multi-frequency bioelectrical impedance analyzer, the statistical threshold method is used to preprocess the electrical signal data set of each type of biological tissue to obtain the corresponding preprocessed data set.

[0041] In step S2, a statistical threshold method is used to preprocess the electrical signal data set of biological tissue type L to obtain a corresponding preprocessed data set, including:

[0042] S21. For the electrical signal dataset Y L Eliminate invalid data and obtain valid data set Y LY .

[0043] Specifically, step S21 performs the following operations on the electrical signal data set Y: L Invalid data exclusion processing is performed, including:

[0044] S211. Calculate the electrical signal data set Y LThe electrical impedance value of each data in is processed by the quartile method to obtain the lower quartile Q1, median Q2 and upper quartile Q3; traverse the electrical signal data set Y L For each piece of data in the impedance value, filter out the value less than Q1-1.5 (Q3-Q1) or greater than Q3+1.5 (Q3-Q1), and delete the corresponding complete data;

[0045] S212. According to all the data retained in step S211, the mean and standard deviation of the three characteristics of resistance, reactance and phase are calculated respectively, and the standard deviation judgment process is performed on each piece of data;

[0046] Specifically, in step S212, according to all the data retained in step S211, the resistance mean, reactance mean, phase mean, resistance standard deviation, reactance standard deviation, and phase standard deviation are calculated; in step S212, any data is subjected to standard deviation judgment processing, including

[0047] Calculate a first difference between the resistance and reactance mean, a second difference between the resistance and the phase mean, a third difference between the reactance and the resistance mean, a fourth difference between the reactance and the phase mean, a fifth difference between the phase and the resistance mean, and a sixth difference between the phase and the reactance mean;

[0048] If one or more of the six comparison rules are met, the data is deleted, otherwise the data is retained; the six comparison rules include the following six conditions:

[0049] Condition 1: The first difference is greater than 3 times the reactance standard deviation,

[0050] Condition 2: The second difference is greater than 3 times the phase standard deviation,

[0051] Condition 3: The third difference is greater than 3 times the resistance standard deviation,

[0052] Condition 4: The fourth difference is greater than 3 times the phase standard deviation,

[0053] Condition 5: The fifth difference is greater than 3 times the resistance standard deviation,

[0054] Condition 6: The sixth difference is greater than 3 times the standard deviation of the reactance.

[0055] S213. Calculate the mean and standard deviation of each piece of data retained in step S212, calculate the coefficient of variation based on the mean and standard deviation, and delete the piece of data if the coefficient of variation is less than 15%. Wherein, coefficient of variation = (standard deviation / mean) × 100%.

[0056] In step S213, for each piece of data, a mean value and a standard deviation are calculated according to the resistance, reactance and phase contained therein, and then the coefficient of variation corresponding to the piece of data is calculated according to the mean value and the standard deviation.

[0057] Specifically, in step S21, after the exclusion process, several pieces of data that do not meet the standards are deleted. To keep the number of data sets unchanged, the same number of new data entries are added according to the number of deleted entries for processing.

[0058] S22. For the valid data set Y LY Perform Gaussian filtering to obtain the filtered effective data set Y LL .

[0059] Specifically, Gaussian filtering is used to smooth each piece of data to effectively remove baseline drift and physiological noise. The Gaussian function is used as the smoothing window function, where σ controls the width of the filter. The effective data set Y is convolved LY Convolve with the Gaussian kernel G to obtain the filtered effective data set Y LL , Y LL =Y LY *G, this operation can suppress high-frequency fluctuations and retain important low-frequency information, thereby reducing the impact of noise caused by physiological activities such as breathing and heartbeat, and ensuring the stability and reliability of the signal in subsequent analysis.

[0060] Filter valid data set Y LL The i-th data can be expressed as

[0061]

[0062] Among them, * represents the convolution operation, is the Gaussian filter kernel function, and σ is the standard deviation. Respectively represent the filtered effective data set Y LL The resistance, reactance and phase of the i-th data obtained after Gaussian filtering.

[0063] S23. Filter the effective data set Y LL Perform standardization to obtain the preprocessed data set Y LB .

[0064] Specifically, calculate the standardized data set Y LB The mean and standard deviation of each type of feature indicator are calculated, and then each feature is Z-score standardized using the mean and standard deviation to eliminate the impact of inconsistent feature dimensions. Finally, the data set is updated and the processing results are verified to ensure the effectiveness of standardization.

[0065] Specifically, the mean and standard deviation of each type of characteristic index are calculated as follows:

[0066]

[0067] in, They represent the mean values ​​of three characteristic indicators: resistance, reactance and phase. σ φ They represent the standard deviations of the three characteristic indicators, resistance, reactance and phase, respectively, and p represents the number of sampling frequency points.

[0068] The normalization process can be expressed as

[0069]

[0070] in, Represents the preprocessed data set Y LB The resistance, reactance and phase of the i-th data in .

[0071] S3. Perform feature extraction on the preprocessed data set of each type of biological tissue to obtain a corresponding feature data set.

[0072] Specifically, step S3 includes:

[0073] S31. Calculate the electrical impedance of each piece of data based on resistance and reactance;

[0074] S32. For the electrical impedance corresponding to the current frequency point, the absolute value of the difference between the electrical impedance corresponding to the previous frequency point is calculated. If the absolute value of the difference is greater than a set threshold, the data corresponding to the current frequency point is marked as mutation data.

[0075] S4. The characteristic data set of each type of biological tissue is processed separately by principal component analysis to obtain the corresponding principal component characteristic parameter set.

[0076] Specifically, step S4 uses principal component analysis to analyze the characteristic data set Y of biological tissue type L. LT Processing includes:

[0077] S41. Based on feature dataset Y LT , calculate the mean and standard deviation of the three characteristic indicators of resistance, reactance and phase respectively.

[0078] Specifically, the average value of the characteristic index of resistance and standard deviation It can be expressed as

[0079]

[0080] Represents the feature dataset Y LTThe resistance of the i-th data in . Similarly, the mean and standard deviation of the other two characteristic indicators are calculated in this way.

[0081] S42. According to the mean and standard deviation, the feature data set Y LT Each characteristic indicator of each data is standardized.

[0082] Specifically, the feature dataset Y LT The resistance of the i-th data is standardized to obtain Expressed as

[0083]

[0084] Y LN Represents the standardized feature dataset, that is

[0085]

[0086] Represents the feature dataset Y LT The standardized result of the i-th data in .

[0087] S43. Calculate the covariance matrix Cov(Y according to the standardized feature data set LN ), calculate the eigenvalues ​​and eigenvectors of the covariance matrix, expressed as

[0088]

[0089] S44. Orthogonalize each eigenvector to obtain an orthogonal eigenvector;

[0090] S45. Arrange all eigenvalues ​​in descending order, and calculate the cumulative contribution rate of each eigenvalue in turn; obtain all eigenvalues ​​with a cumulative contribution rate greater than m, and extract the corresponding orthogonal eigenvectors to construct a principal component matrix;

[0091] Specifically, all eigenvalues ​​are arranged in descending order to obtain λ1>λ2>…>λ p ,λ i represents the i-th eigenvalue after sorting;

[0092] Calculate the cumulative contribution rate θ i , used to measure the proportion of data variance explained by the first i principal components

[0093]

[0094] The cumulative contribution rate indicates how much effective information of the bioelectrical impedance signal is retained after dimensionality reduction.

[0095] Select eigenvalues ​​with cumulative contribution rates greater than m, and construct the principal component matrix P with the orthogonal eigenvectors corresponding to these eigenvalues

[0096]

[0097] At this point, the principal component matrix P is an a×q matrix, which represents the standardized feature data set Y LN Projected into the new a-dimensional space, the main tissue identification features are retained. a is the number of eigenvalues ​​with a cumulative contribution rate of more than m.

[0098] S46. Calculate the principal component feature parameter set based on the principal component matrix and the standardized feature data set.

[0099] The principal component feature parameter set can be expressed as a matrix Z, which is defined as follows:

[0100] Z=Y LN ·P T

[0101] Among them, P is a matrix composed of the eigenvectors of the first a principal components, with a shape of a×q; Z is the principal component characteristic parameter set, with a shape of p×a, expressed as:

[0102] Each z i is the reduced-dimensional feature vector at the i-th frequency point, with dimension a, representing the main features at this frequency point.

[0103] S5. Construct a training data set based on the principal component feature parameter sets of five types of biological tissues; and train a random forest model based on the training data set.

[0104] like Figure 4 As shown in the figure, random forest is an ensemble learning method that improves the accuracy of classification or regression by constructing multiple decision trees. The basic idea is to generate multiple training subsets by sampling with replacement, thereby training multiple weak classifiers. Each decision tree, as a basic unit, is built based on different random samples and features. When making predictions, a single decision tree starts from the root node and divides the samples to be classified into different child nodes layer by layer according to the rules until it reaches the leaf node. The corresponding category on the leaf node is the prediction result. By combining the prediction results of each tree, random forest can effectively reduce overfitting and enhance the robustness of the model.

[0105] In the present invention, unless otherwise clearly stipulated and limited, the terms such as "installation", "setting", "connection", "fixation" and "rotation" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral one; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium; it can be the internal connection of two elements or the interaction relationship between two elements. Unless otherwise clearly defined, ordinary technicians in this field can understand the specific meanings of the above terms in the present invention according to the specific circumstances.

[0106] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A puncture surgery tissue identification system based on bioelectrical impedance, characterized in that: include: Dual-electrode impedance needle for real-time acquisition of electrical signal data; A feedback display device, which has a trained random forest model integrated therein, is used to process the electrical signal data according to the trained random forest model to obtain a classification result; Multi-frequency bioelectrical impedance analyzer, used to display real-time acquisition waveforms and classification results of random forest models; Among them, the dual-electrode impedance needle includes a needle tube structure and an electrode structure located at the head of the needle tube structure; the needle tube structure includes an outer needle tube, an insulating tube layer, and an inner needle core from the inside to the outside; the electrode structure includes an excitation electrode layer, an insulating layer, and a measuring electrode layer from the inside to the outside; the excitation electrode layer is connected to an excitation electrode external cable, and the measuring electrode layer is connected to a measuring electrode external cable; wherein the outer needle tube is connected to the measuring electrode layer, and the inner needle core is connected to the excitation electrode layer.

2. According to claim 1, a puncture surgery tissue identification system based on bioelectrical impedance is characterized in that: The training process of the random forest model consists of the following steps: S1. The electrical signal datasets of five types of biological tissues, namely skin, muscle, fat, bone, and tumor, are collected through a dual-electrode impedance needle. Each data item in the electrical signal dataset of each type of biological tissue includes frequency, resistance, reactance, and phase. S2. Preprocessing the electrical signal data set of each type of biological tissue by using a statistical threshold method to obtain a corresponding preprocessed data set; S3. extracting features from the preprocessed data set of each type of biological tissue to obtain a corresponding feature data set; S4. processing the characteristic data set of each type of biological tissue respectively by principal component analysis to obtain the corresponding principal component characteristic parameter set; S5. Construct a training data set based on the principal component feature parameter sets of five types of biological tissues; and train a random forest model based on the training data set.

3. The bioimpedance-based puncture surgery tissue identification system according to claim 2, characterized in that: Step 1 specifically includes: S11. Maintain an environment with a temperature of 20-25°C and a humidity of 40-60% rh; S12. In the frequency range of 1kHz to 100MHz, a multi-frequency bioelectrical impedance analyzer is used to sample p frequency points, and the frequency, resistance, reactance and phase of the skin sample are collected at each frequency point to form an electrical signal data set Y of the skin tissue. pf ; S13. In the frequency range of 1kHz to 100MHz, sample p frequency points through a multi-frequency bioelectrical impedance analyzer, and collect the frequency, resistance, reactance and phase of the muscle sample at each frequency point; form an electrical signal data set Y of the muscle tissue jr ; S14. In the frequency range of 1kHz to 100MHz, sample p frequency points through a multi-frequency bioelectrical impedance analyzer, and collect the frequency, resistance, reactance and phase of the fat sample at each frequency point; form an electrical signal data set Y of the fat tissue zf ; S15. In the frequency range of 1kHz to 100MHz, sample p frequency points through a multi-frequency bioelectrical impedance analyzer, and collect the frequency, resistance, reactance and phase of the bone sample at each frequency point; form an electrical signal data set Y of the bone tissue gg ; S16. In the frequency range of 1kHz to 100MHz, sample p frequency points by a multi-frequency bioelectrical impedance analyzer, and collect the frequency, resistance, reactance and phase of the tumor sample at each frequency point; form an electrical signal data set Y of the tumor tissue zl .

4. The bioimpedance-based puncture surgery tissue identification system according to claim 2, characterized in that: In step S2, a statistical threshold method is used to preprocess the electrical signal data set of any type of biological tissue to obtain a corresponding preprocessed data set, including: S21. Eliminate invalid data from the electrical signal data set to obtain a valid data set; S22. Performing Gaussian filtering on the valid data set to obtain a filtered valid data set; S23. Standardize the filtered valid data set to obtain a preprocessed data set.

5. The bioimpedance-based puncture surgery tissue identification system according to claim 4, characterized in that: Step S21 performs invalid data exclusion processing on the electrical signal data set, including: S211. Calculate the electrical impedance value of each data in the electrical signal data set, and use the quartile method to process all the electrical impedance values ​​to obtain the lower quartile Q1, the median Q2, and the upper quartile Q3; traverse the electrical impedance value of each data in the electrical signal data set, filter out the value less than Q1-1.5 (Q3-Q1) or greater than Q3+1.5 (Q3-Q1), and delete the corresponding complete data; S212. According to all the data retained in step S211, the mean and standard deviation of the three characteristics of resistance, reactance and phase are calculated respectively, and the standard deviation judgment process is performed on each piece of data; S213. Calculate the mean and standard deviation of each piece of data retained in step S212, calculate the coefficient of variation based on the mean and standard deviation, and if the coefficient of variation is less than 15%, delete the piece of data.

6. The bioimpedance-based puncture surgery tissue identification system according to claim 5, characterized in that: In step S212, the standard deviation judgment process is performed on any piece of data, including: Calculate a first difference between the resistance and reactance mean, a second difference between the resistance and the phase mean, a third difference between the reactance and the resistance mean, a fourth difference between the reactance and the phase mean, a fifth difference between the phase and the resistance mean, and a sixth difference between the phase and the reactance mean; If one or more of the six comparison rules are met, the data is deleted, otherwise the data is retained; the six comparison rules include the following six conditions: Condition 1: The first difference is greater than 3 times the reactance standard deviation, Condition 2: The second difference is greater than 3 times the phase standard deviation, Condition 3: The third difference is greater than 3 times the resistance standard deviation, Condition 4: The fourth difference is greater than 3 times the phase standard deviation, Condition 5: The fifth difference is greater than 3 times the resistance standard deviation, Condition 6: The sixth difference is greater than 3 times the standard deviation of the reactance.

7. The bioimpedance-based puncture surgery tissue identification system according to claim 2, characterized in that: Feature extraction of preprocessed datasets of any type of biological tissue includes: S31. Calculate the electrical impedance of each piece of data based on resistance and reactance; S32. For the electrical impedance corresponding to the current frequency point, the absolute value of the difference between the electrical impedance corresponding to the previous frequency point is calculated. If the absolute value of the difference is greater than a set threshold, the data corresponding to the current frequency point is marked as mutation data.

8. The bioimpedance-based puncture surgery tissue identification system according to claim 2, characterized in that: Step S4 processes the characteristic data set of any type of biological tissue by principal component analysis, including: S41. Based on the characteristic data set, the mean and standard deviation of the three characteristic indicators of resistance, reactance and phase are calculated respectively; S42. standardize each feature indicator of each piece of data in the feature data set according to the mean and standard deviation; S43. Calculate the covariance matrix based on the standardized feature data set, and calculate the eigenvalues ​​and eigenvectors of the covariance matrix; S44. Orthogonalize each eigenvector to obtain an orthogonal eigenvector; S45. Arrange all eigenvalues ​​in descending order, and calculate the cumulative contribution rate of each eigenvalue in turn; obtain all eigenvalues ​​with a cumulative contribution rate greater than m, and extract the corresponding orthogonal eigenvectors to construct a principal component matrix; S46. Calculate the principal component feature parameter set based on the principal component matrix and the standardized feature data set.

Images