Target object frequency and amplitude detection system and method

Abstract

The invention discloses a target object frequency and amplitude detection system and method, and belongs to the technical field of traditional Chinese medicine diagnosis, and the method specifically comprises the steps: obtaining basic parameters and positioning reference information required by synchronous detection; based on the information, a human pulse condition-oriented tracking-deviation compensation strategy and a vibration signal-oriented vibration filtering-coordinate anchoring strategy are adopted, and a detection area is marked to generate a positioning datum point; after it is confirmed that the sensor is stable, multi-channel self-adaptive synchronous acquisition is executed with the positioning datum point as reference, and an original signal data set is generated; performing layered preprocessing on an original signal data set, obtaining a purified signal set through a traceability noise reduction model, and extracting frequency characteristic values and amplitude characteristic values of human pulse conditions and vibration signals through multi-algorithm game fusion; and finally, dynamic threshold verification is performed on the characteristic value, a detection report is generated, and a data ledger is formed. According to the method, the signal acquisition stability and the feature extraction accuracy are improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of traditional Chinese medicine diagnosis, in particular to a target object frequency and amplitude detection system and method. BACKGROUND

[0002] In the field of traditional Chinese medicine diagnosis, pulse-taking is an important part of four diagnostic methods. Doctors press the radial artery of the subject with their fingers to determine the physical condition and illness according to the frequency, amplitude, and morphology of the pulse. However, traditional pulse-taking relies on the subjective experience of doctors, and different doctors have different standards for judging pulse characteristics, resulting in poor consistency and difficulty in quantifying the detection results, which cannot provide objective data support for traditional Chinese medicine diagnosis. Existing traditional Chinese medicine four diagnostic methods collection and identification instruments attempt to collect pulse signals through sensors, but there are the following problems in the detection process: the alignment of the sensor and the pulse collection point relies on manual experience, which is prone to deviation, resulting in distorted collected signals; the collection process is easily affected by power frequency interference, electromyographic interference, and environmental noise, making it difficult to accurately extract the true characteristics of the pulse signal; most detection methods only focus on a single parameter of frequency or amplitude, and no standardized feature judgment criteria are established, so the practicality of the detection results is limited; the pulse signal intensity and fluctuation characteristics of different subjects differ, and existing methods lack adaptive adjustment capability, which cannot guarantee the accuracy of the detection results of different individuals. Therefore, there is an urgent need for a frequency and amplitude detection method that can accurately locate the target object, especially the pulse signal, resist interference, detect multiple parameters cooperatively, and have adaptive capability to solve the drawbacks of traditional detection methods. SUMMARY

[0003] To overcome the shortcomings of the prior art, the present application provides a target object frequency and amplitude detection system and method, which obtains the basic parameters and positioning reference information required for synchronous detection; based on the above information, a tracking-deviation compensation strategy for human pulse and a vibration filtering-coordinate anchoring strategy for vibration signals are adopted to generate positioning reference points for the detection area marker; after confirming the stability of the sensor, multi-channel adaptive synchronous collection is performed with the positioning reference points as the reference to generate the original signal data set; the original signal data set is preprocessed in layers, and the purified signal set is obtained through a traceable noise reduction model, and then the frequency characteristic value and amplitude characteristic value of the human pulse and vibration signal are extracted through multi-algorithm game fusion; finally, dynamic threshold verification is performed on the characteristic values to generate a detection report and form a data account; this method improves the signal collection stability and feature extraction accuracy.

[0004] To achieve the above purpose, the present application provides the following technical solutions:

[0005] A target object frequency and amplitude detection method, comprising:

[0006] obtaining the basic parameters and positioning reference information for human pulse-vibration synchronous detection;

[0007] Based on the basic parameters and the positioning reference information, a dynamic positioning strategy is adopted to perform a marking operation on the synchronous detection area to generate a positioning reference point; the dynamic positioning strategy includes a tracking-deviation compensation dynamic positioning strategy for the human pulse signal and a vibration filtering-coordinate anchoring dynamic positioning strategy for the vibration signal;

[0008] After confirming that the sensor is in a stable collection state, a multi-channel adaptive sensing and synchronous data collection is performed with the positioning reference point as a collection position reference to generate a human pulse-vibration synchronous original signal dataset;

[0009] A hierarchical preprocessing is performed on the human pulse-vibration synchronous original signal dataset, a purified human pulse-vibration synchronous signal set is generated through a traceability denoising model, and a multi-algorithm game fusion feature extraction is performed according to the purified human pulse-vibration synchronous signal set to generate human pulse signal frequency feature values, human pulse signal amplitude feature values, vibration signal frequency feature values, and vibration signal amplitude feature values, respectively;

[0010] Dynamic threshold verification is performed on the frequency feature values and the amplitude feature values of the human pulse signal and the vibration signal, respectively, a human pulse-vibration synchronous detection report is generated according to the verification results, and a synchronous detection data account is formed at the same time.

[0011] Specifically, the basic parameters include a synchronous detection target object type, a synchronous detection area physical feature, and an accuracy calibration parameter of an infrared positioning module; the positioning reference information includes a preset marking coordinate of a human pulse detection area, a preset marking coordinate of a vibration synchronous sensing area, and a common allowed deviation range of human pulse signal and vibration signal detection; the synchronous detection target object type includes a human pulse signal and a vibration signal; the synchronous detection area physical feature includes a physiological feature of a human arm detection part and a vibration conduction path feature;

[0012] When the basic parameters and the positioning reference information for human pulse-vibration synchronous detection are acquired, a preset parameter input interface is used to receive a synchronous detection target object type selection instruction input by a user, and a corresponding type of synchronous detection parameter template is called based on the synchronous detection target object type selection instruction; the synchronous detection parameter template includes a basic parameter framework and a positioning reference information framework for detection of a corresponding type of target object;

[0013] A sensor deployed in the detection area is controlled to scan the synchronous detection area and automatically identify the physical feature of the synchronous detection area, wherein for a human pulse detection area, physiological features including skin thickness and blood vessel distribution density are identified; for a vibration detection area, path features including the material and hardness of the vibration conduction medium are identified to generate physical feature data;

[0014] The synchronous detection parameter template and the user input synchronous detection target type selection instruction are taken as initial input for obtaining basic parameters, and the initial input is matched and fused with the generated physical feature data to fill in specific values for the parameter items in the synchronous detection parameter template, determine the specific values of the preset marker coordinates and the accurate values of the common allowable deviation range in the positioning reference information, and finally generate the basic parameters and the positioning reference information.

[0015] Specifically, the marker operation is performed on the synchronous detection area by adopting the dynamic positioning strategy to generate the positioning reference point, including:

[0016] For the target area corresponding to the human pulse signal in the synchronous detection, the accuracy calibration parameter of the infrared positioning module in the obtained basic parameters and the preset marker coordinates and the common allowable deviation range of the human pulse detection area in the positioning reference information are called, the target area corresponding to the human pulse signal in the synchronous detection is tracked in real time to generate the real-time position coordinates of the human pulse detection area, the real-time position coordinates of the human pulse detection area obtained are compared with the preset marker coordinates of the human pulse detection area in the positioning reference information, the three-dimensional position deviation value in the three-dimensional orthogonal coordinate system is calculated, and the space module of the three-dimensional position deviation value is calculated; the relationship between the calculated space module and the common allowable deviation range is judged, if the space module is less than or equal to the common allowable deviation range, the corresponding real-time position coordinates are directly determined as the first temporary marker point of the human pulse detection; if the space module is greater than the common allowable deviation range, the accuracy calibration parameter of the infrared positioning module is called to compensate and correct the three-dimensional position deviation value to obtain the compensated human pulse detection position coordinates, and the compensated human pulse detection position coordinates are determined as the first temporary marker point of the human pulse detection; the accuracy calibration parameter of the infrared positioning module includes a temperature drift compensation coefficient and an optical distortion correction matrix.

[0017] Specifically, the marker operation is performed on the synchronous detection area by adopting the dynamic positioning strategy to generate the positioning reference point, including:

[0018] For the sensing area corresponding to the vibration signal caused by the arm micro-motion in the synchronous detection, the preset marker coordinates of the vibration synchronous sensing area in the positioning reference information are called, the vibration sensor is started to collect the signal of the sensing area corresponding to the vibration signal caused by the arm micro-motion in the synchronous detection to generate the original vibration signal; the original vibration signal is filtered to obtain the filtered vibration signal, and the filtered vibration signal is analyzed in amplitude to determine the stable position interval of the vibration sensing area and calculate the geometric center coordinates of the stable position interval; the obtained geometric center coordinates are taken as the anchoring reference, and the preset marker coordinates of the vibration synchronous sensing area in the positioning reference information are mapped to the geometric center to generate the anchored vibration detection position coordinates, which are determined as the second temporary marker point of the vibration detection.

[0019] Specifically, the step of using a dynamic positioning strategy to perform a marking operation on the synchronous detection area and generate positioning reference points further includes:

[0020] The spatial coordinate system parameters of the synchronous detection area in the basic parameters are called, and the obtained human pulse detection first temporary marker point and vibration detection second temporary marker point are transformed to the same three-dimensional coordinate system. The spatial deviation value between the two points is calculated. If the spatial deviation value is greater than the preset first deviation threshold, coordinate association calibration is performed until the deviation value is less than or equal to the first deviation threshold. Finally, the calibrated human pulse detection first temporary marker point and the calibrated vibration detection second temporary marker point are jointly determined as the positioning reference point for synchronous detection.

[0021] The coordinate correlation calibration process includes:

[0022] Based on the coordinates of the first temporary marker point for human pulse detection after coordinate transformation and the second temporary marker point for vibration detection after coordinate transformation, historical coordinate data are collected to form a human pulse marker point coordinate sequence and a vibration marker point coordinate sequence.

[0023] The least squares method was used to fit the coordinate sequences of human pulse marker points and vibration marker points to construct a coordinate offset correction matrix;

[0024] Substitute the current coordinates of the second temporary vibration detection marker into the coordinate offset correction matrix to calculate the calibrated vibration detection marker coordinates.

[0025] Calculate the spatial deviation between the calibrated vibration detection marker coordinates and the human pulse marker coordinate sequence. If it is still greater than the first deviation threshold, return to the historical coordinate data acquisition process until the spatial deviation is less than or equal to the first deviation threshold, and obtain the calibrated first temporary marker for human pulse detection and the calibrated second temporary marker for vibration detection.

[0026] Specifically, the process of confirming that the sensor is in a stable data acquisition state includes:

[0027] Call the coordinates of the generated synchronous detection and positioning reference point to determine the pre-acquisition range. At the same time, call the sensor's pre-acquisition parameters in the basic parameters.

[0028] Based on the determined preset acquisition range and pre-acquisition parameters, the multi-channel sensors deployed in the detection area are controlled to perform pre-acquisition, generating a set of pre-acquisition signals including pulse pre-acquisition signals and vibration pre-acquisition signals, and each signal is accompanied by an acquisition timestamp and the corresponding sensor channel identifier.

[0029] From the pre-acquired signal set, the human pulse image pre-acquired signal subset and the vibration pre-acquired signal subset are extracted respectively, and the stability analysis is performed on the two types of signal subsets, and the standard deviation and the coefficient of variation of the pre-acquired signal subset are calculated;

[0030] If , and At the same time, , and , it is determined that the sensor is in a stable acquisition state, otherwise, the contact pressure of the control sensor is adjusted and the pre-acquisition process is repeated until all signals meet the stability determination threshold, and it is confirmed that the sensor is in a stable acquisition state, wherein and respectively represent the standard deviation and the coefficient of variation of the human pulse image pre-acquired signal subset, and respectively represent the standard deviation and the coefficient of variation of the vibration pre-acquired signal subset, and respectively represent the standard deviation threshold and the variation threshold of the human pulse image pre-acquired signal subset, and respectively represent the standard deviation threshold and the variation threshold of the vibration pre-acquired signal subset.

[0031] Specifically, the generation of the purified synchronous signal set through the traceability denoising model comprises:

[0032] The hierarchical pre-processed human pulse-vibration synchronous original signal data set is input into the pre-trained traceability denoising model; the traceability denoising model is constructed based on a deep learning network and includes an input layer, an attention mechanism intermediate layer and an output layer;

[0033] The input layer of the traceability denoising model performs time-frequency conversion on the input hierarchical pre-processed human pulse-vibration synchronous original signal, and performs noise marking through the attention mechanism intermediate layer, and finally performs denoising processing through the output layer, and outputs the purified human pulse-vibration synchronous signal set.

[0034] Specifically, the multi-algorithm game fusion feature extraction according to the purified human pulse-vibration synchronous signal set comprises:

[0035] The purified human pulse-vibration synchronous signal set is split into independent human pulse signal subset and vibration signal subset according to signal type;

[0036] The Fourier transform algorithm is used to perform frequency analysis on the human pulse signal subset and the vibration signal subset respectively, and the human pulse preliminary frequency characteristic value and the vibration preliminary frequency characteristic value are obtained;

[0037] The wavelet transform algorithm is used for time-frequency analysis on the human pulse signal subset and the vibration signal subset respectively, and human pulse preliminary amplitude characteristic values and vibration preliminary amplitude characteristic values are obtained.

[0038] The support vector machine algorithm is used for screening on the human pulse preliminary frequency characteristic values, the vibration preliminary frequency characteristic values, the human pulse preliminary amplitude characteristic values and the vibration preliminary amplitude characteristic values, and human pulse effective characteristic subsets and vibration effective characteristic subsets are generated.

[0039] The Nash equilibrium strategy in the game theory is used for fusion on the human pulse effective characteristic subsets and the vibration effective characteristic subsets, and final human pulse signal frequency characteristic values, human pulse signal amplitude characteristic values, vibration signal frequency characteristic values and vibration signal amplitude characteristic values are obtained.

[0040] A target frequency and amplitude detection system comprises an information acquisition module, a reference point generation module, a feature extraction module and a threshold verification module.

[0041] The information acquisition module is used for collecting and integrating basic parameters and positioning reference information of human pulse-vibration synchronous detection.

[0042] The reference point generation module adopts a differentiated dynamic positioning strategy for human pulse and vibration signals to perform a marking operation on a synchronous detection area, and generates a synchronous detection positioning reference point.

[0043] The feature extraction module is used for performing hierarchical preprocessing and noise reduction on an original signal data set, and then extracting frequency characteristic values and amplitude characteristic values of human pulse signals and vibration signals through multi-algorithm fusion.

[0044] The threshold verification module is used for performing dynamic threshold verification on output characteristic values, generating a detection report and a data account, and completing visual output and data archiving of the detection results.

[0045] Compared with the prior art, the present application has the following advantages:

[0046] 1. The present application proposes a target frequency and amplitude detection method, which adopts a whole-process design of accurate parameter positioning-dynamic tracking calibration-stable acquisition-depth preprocessing-game fusion extraction-dynamic verification, effectively solves the problems of large positioning deviation, many signal noise interferences and low feature extraction accuracy in traditional detection, and dynamically positions the pulse and vibration signals by tracking-deviation compensation and filtering-coordinate anchoring, combines a traceability noise reduction model and multi-algorithm game fusion, improves the positioning accuracy, signal purity and feature value reliability of synchronous detection, and ensures stable sensor acquisition and accurate data analysis.

[0047] 2.The application provides a target frequency and amplitude detection method, which can not only generate a standard detection report and complete data account through dynamic threshold verification, realize traceability of the detection process and visual presentation of the results, but also be widely applicable to medical health monitoring, device vibration collaborative analysis and the like, a standardized process design thereof reduces the complexity of detection operation, improves detection efficiency, and meanwhile, high-quality feature extraction and verification results can provide accurate data support for human health state evaluation, device operation fault prediction and the like. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 FIG. 1 is a schematic diagram of a target frequency and amplitude detection method according to the application;

[0049] Figure 2 FIG. 2 is a principle flowchart of a target frequency and amplitude detection method according to the application;

[0050] Figure 3 FIG. 3 is a system architecture diagram of a target frequency and amplitude detection system according to the application. DETAILED DESCRIPTION

[0051] Embodiment 1

[0052] Referring to Figure 1 and Figure 2 , the application provides an embodiment: a target frequency and amplitude detection method, which comprises S1-S5 and specifically comprises:

[0053] S1: acquiring basic parameters and positioning reference information of human pulse-vibration synchronous detection;

[0054] S2: based on the basic parameters and positioning reference information, performing a marking operation on a synchronous detection area by using a dynamic positioning strategy to generate a positioning reference point; the dynamic positioning strategy comprises a tracking-deviation compensation dynamic positioning strategy facing the human pulse signal and a vibration filtering-coordinate anchoring dynamic positioning strategy facing the vibration signal;

[0055] S3: after confirming that the sensor is in a stable acquisition state, taking the positioning reference point as an acquisition position reference to perform multi-channel adaptive sensing and synchronous data acquisition, and generating a human pulse-vibration synchronous original signal data set;

[0056] S4: performing hierarchical preprocessing on the human pulse-vibration synchronous original signal data set, generating a purified human pulse-vibration synchronous signal set through a traceability denoising model, and performing multi-algorithm game fusion feature extraction according to the purified human pulse-vibration synchronous signal set to respectively generate a human pulse signal frequency characteristic value, a human pulse signal amplitude characteristic value, a vibration signal frequency characteristic value and a vibration signal amplitude characteristic value;

[0057] S5: Perform dynamic threshold verification on the frequency characteristic value and the amplitude characteristic value of the human pulse signal and the vibration signal respectively, generate a human pulse-vibration synchronous detection report according to the verification result, and form a synchronous detection data account at the same time.

[0058] Further, the specific steps of performing dynamic threshold verification include:

[0059] (1) Determine the core basis for matching the dynamic threshold template, that is, the synchronous detection target object type, and then call the dynamic threshold template corresponding to the current synchronous detection target object type from the threshold template library pre-stored in the system; the dynamic threshold template must include the frequency and amplitude threshold parameters of the human pulse and vibration signals, and the frequency and amplitude threshold parameters of the human pulse and vibration signals must be set in combination with industry standards and historical detection data, wherein the threshold range of the frequency of the human pulse signal template is set according to the statistical data of the pulse of healthy people, and the threshold range of the frequency of the vibration signal template is set according to the rated parameters of the vibration source;

[0060] (2) From the multi-algorithm game fusion feature extraction module, call the four types of characteristic values finally generated, which are human pulse signal frequency characteristic value, human pulse signal amplitude characteristic value, vibration signal frequency characteristic value, and vibration signal amplitude characteristic value, and record the background information at the time of generating each type of characteristic value, including the amount of historical data used in the fusion process and the optimal weight used in the calculation, then associate each type of characteristic value with the threshold parameters in the corresponding template, so that each characteristic value to be verified has a clear reference standard, avoiding parameter confusion in the verification process;

[0061] (3) Calculate the deviation rate of the characteristic value and the threshold range, including:

[0062] Determine the center value and half width of the threshold range, wherein the center value is the middle reference point of the threshold range, and the calculation method is to add the upper limit value and the lower limit value of the threshold range, and then divide by 2; the half width is half of the width of the threshold range, and the calculation method is to subtract the lower limit value from the upper limit value of the threshold range, and then divide by 2;

[0063] Calculate the actual deviation value and the deviation rate, wherein the actual deviation value is the absolute value of the difference between the characteristic value and the center value, and the absolute value is taken to avoid the situation that positive and negative deviations cancel each other out; the deviation rate is the percentage of the actual deviation value to the half width;

[0064] (4) Perform double-checking in combination with the two conditions of whether the deviation rate meets the allowed upper limit and whether the characteristic value is within the threshold range, to avoid incorrect judgment caused by only looking at one condition;

[0065] (5) The verification results of the four types of characteristic values are sorted into a dynamic threshold verification table. The dynamic threshold verification table includes characteristic value name, actual value, threshold range, center value, half-width, actual deviation value, deviation rate, deviation allowance upper limit, verification result, and abnormal reason. When sorting the table, it is classified according to signal type, including two categories of human pulse and vibration. Each category further includes detailed verification information of corresponding frequency and amplitude characteristic values.

[0066] Further, the specific steps of forming the synchronous detection data account include:

[0067] (1) Determine the account data entry and storage format, including:

[0068] Basic parameters: including synchronous detection target type, detection time, detection personnel number, detection equipment model, infrared positioning module accuracy calibration parameter, preset marker coordinates of human pulse detection area, preset marker coordinates of vibration synchronous sensing area, and allowed deviation range of the two types of signals;

[0069] Positioning data: including human pulse detection first temporary marker point coordinates, vibration detection second temporary marker point coordinates, positioning reference point coordinates after coordinate correlation calibration, and deviation value during calibration;

[0070] Acquisition data: including sensor stable state parameters, sampling frequency, channel number, acquisition duration, storage path of original signal data set, and pre-acquisition number. The sensor stable state parameters include standard deviation and coefficient of variation of human pulse signal, and standard deviation and coefficient of variation of vibration signal;

[0071] Preprocessing and noise reduction data: including hierarchical preprocessing parameters, traceable noise reduction model number, signal-to-noise ratio of purified signal, and noise reduction log;

[0072] Feature extraction and verification data: including four types of final characteristic values, fusion weight, dynamic threshold template number, deviation rate of each type of characteristic value, verification result, and abnormal reason;

[0073] Report: including detection report number, report storage path, and report conclusion;

[0074] All data is stored in a structured manner, with MySQL database table as the storage format. The table structure sets joint primary keys according to detection time and report number to ensure that each account record is unique and does not have duplicate records;

[0075] (2) Automatically retrieve the data required for the above entry, and enter it into the account table in chronological order. At the same time, automatically perform data verification;

[0076] If the data verification passes, the entry for this ledger is complete;

[0077] If the data verification fails, the data will be marked as abnormal and an alarm will be triggered to remind the operator to investigate the cause of the missing or incorrect data, such as an incorrect original signal storage path. After the operator corrects the data, it will be re-entered and verified.

[0078] (3) Generate ledger reports.

[0079] Furthermore, the synchronized detection data ledger is also used to provide feedback for optimizing the dynamic positioning strategy:

[0080] When the synchronous detection data ledger shows consecutive verification failure records caused by the same positioning reference point coordinate deviation, a positioning reference point recalibration command is triggered, and the positioning reference point recalibration control infrared positioning module re-executes the tracking-deviation compensation operation.

[0081] The basic parameters include the type of target object to be detected synchronously, the physical characteristics of the synchronous detection area, and the accuracy calibration parameters of the infrared positioning module; the positioning reference information includes the preset marker coordinates of the human pulse detection area, the preset marker coordinates of the vibration synchronous sensing area, and the common allowable deviation range for the detection of human pulse signals and vibration signals; the type of target object to be detected synchronously includes human pulse signals and vibration signals; the physical characteristics of the synchronous detection area include the physiological characteristics of the human arm detection site and the characteristics of the vibration transmission path;

[0082] When acquiring the basic parameters and positioning reference information required for human pulse-vibration synchronous detection, the system receives the user's input instruction for selecting the type of synchronous detection target through a preset parameter input interface, and calls the pre-stored synchronous detection parameter template of the corresponding type based on the synchronous detection target type selection instruction; the synchronous detection parameter template contains the basic parameter framework and positioning reference information framework required for the detection of the corresponding type of target.

[0083] The system controls the sensors deployed in the detection area to scan the synchronous detection area and automatically identify the physical characteristics of the synchronous detection area. Specifically, for the human pulse detection area, it identifies physiological characteristics including skin thickness and blood vessel distribution density; for the vibration detection area, it identifies path characteristics including the material and hardness of the vibration transmission medium and generates physical feature data.

[0084] The synchronous detection parameter template and the user-input synchronous detection target type selection command are used as the initial inputs to obtain the basic parameters. The initial inputs are matched and fused with the generated physical feature data. The parameter items in the synchronous detection parameter template are filled with specific values ​​to determine the specific values ​​of the preset marker coordinates and the precise values ​​of the common allowable deviation range in the positioning reference information. Finally, the basic parameters and positioning reference information are generated.

[0085] Furthermore, the initial input is matched and fused with the generated physical feature data, and specific values ​​are filled into the parameter items in the synchronous detection parameter template, including:

[0086] Construct bidirectional matching rules: On the one hand, extract the parameter items to be filled from the synchronous detection parameter template, such as the preset marker coordinate values ​​and the upper limit of the common allowable deviation range, as the matching targets; on the other hand, filter the data that is semantically related to the parameter items from the physical feature dataset, such as the preset marker coordinates of human pulse.

[0087] A layered fusion strategy is adopted: the first layer performs precise matching, directly filling the explicit values ​​in the physical feature data into the corresponding parameter items of the synchronous detection parameter template; the second layer performs computational matching, filling the range values ​​by taking the mean or weighted median value when the physical feature data is a range value; the third layer performs rule matching, determining the final value for parameter items that require comprehensive judgment by combining the industry standards of the target object type with the physical feature data.

[0088] The matched and fused values ​​are filled into the null parameter fields of the synchronous detection parameter template one by one to form a complete basic parameter table.

[0089] The step of using a dynamic positioning strategy to perform a marking operation on the synchronous detection area and generate positioning reference points includes:

[0090] For the target area corresponding to the human pulse signal in synchronous detection, the accuracy calibration parameters of the infrared positioning module in the acquired basic parameters, as well as the preset marker coordinates and common allowable deviation range of the human pulse detection area in the positioning reference information, are called. The target area corresponding to the human pulse signal in synchronous detection is tracked in real time to generate the real-time position coordinates of the human pulse detection area. The obtained real-time position coordinates of the human pulse detection area are compared with the preset marker coordinates of the human pulse detection area in the positioning reference information to calculate the three-dimensional position deviation value in the three-dimensional rectangular coordinate system, and the spatial modulus of the three-dimensional position deviation value is calculated. The relationship between the calculated spatial modulus and the common allowable deviation range is determined. If the spatial modulus is less than or equal to the common allowable deviation range, the corresponding real-time position coordinates are directly determined as the first temporary marker point for human pulse detection. If the spatial modulus is greater than the common allowable deviation range, the accuracy calibration parameters of the infrared positioning module are called to compensate and correct the three-dimensional position deviation value to obtain the compensated human pulse detection position coordinates, and this is determined as the first temporary marker point for human pulse detection. The accuracy calibration parameters of the infrared positioning module include the temperature drift compensation coefficient and the optical distortion correction matrix.

[0091] Furthermore, the calculation process for the three-dimensional position deviation value is as follows: the X-axis position deviation value is obtained by subtracting the X-axis value of the preset marker coordinates from the X-axis value of the real-time position coordinates of the human pulse detection area; similarly, the Y-axis position deviation value is obtained by subtracting the Y-axis value of the preset marker coordinates from the Y-axis value of the real-time position coordinates of the human pulse detection area, and the Z-axis position deviation value is obtained by subtracting the Z-axis value of the preset marker coordinates from the Z-axis value of the real-time position coordinates of the human pulse detection area. Together, they constitute the three-dimensional position deviation value.

[0092] Furthermore, the calculation process for the spatial modulus of the three-dimensional position deviation value is as follows: First, the position deviation values ​​of the X-axis, Y-axis, and Z-axis in the three-dimensional position deviation value are converted into non-negative values ​​respectively; then, the three non-negative values ​​are squared respectively to obtain three squared values; next, the three squared values ​​are added together to obtain a sum of squares; finally, the square root operation is performed on the sum of squares to obtain the spatial modulus of the three-dimensional position deviation value.

[0093] Furthermore, the temperature drift compensation calculation process is as follows: First, determine the difference between the current ambient temperature and the calibration temperature of the infrared positioning module, i.e., the temperature change; then multiply the temperature change by the temperature drift compensation coefficient of the corresponding axis to obtain the temperature drift correction amount of the axis; finally, subtract the temperature drift correction amount from the original position deviation value of the axis to obtain the position deviation value of the axis after temperature compensation, thereby obtaining the three-dimensional position deviation value after temperature compensation.

[0094] Furthermore, the optical distortion correction calculation process is as follows: the temperature-compensated three-dimensional position deviation value is taken as a three-dimensional vector and multiplied by the optical distortion correction matrix. Specifically, each element in the three-dimensional vector is multiplied by the element of the corresponding row of the optical distortion correction matrix, and the product results of each row are added together to obtain three new values. These three new values ​​are the three-dimensional position deviation values ​​after optical distortion correction.

[0095] Furthermore, the optical distortion correction matrix is ​​mainly obtained through the camera calibration process, which aims to determine the camera's internal and external parameters.

[0096] Furthermore, the calculation process for the compensated human pulse detection position coordinates is as follows: the X-axis value of the real-time position coordinates of the human pulse detection area is subtracted from or added to the deviation value of the axis after temperature drift compensation and optical distortion correction to obtain the X-axis compensated coordinate value; similarly, the Y-axis and Z-axis compensated coordinate values ​​are calculated, and the three together constitute the compensated human pulse detection position coordinates. The subtraction or addition is determined according to the direction of the deviation value after the axis correction.

[0097] For the sensing area corresponding to the vibration signal caused by the slight arm movement in the synchronous detection, the preset marker coordinates of the vibration synchronous sensing area in the positioning reference information are called. The vibration sensor is started to collect signals from the sensing area corresponding to the vibration signal caused by the slight arm movement in the synchronous detection, generating the original vibration signal. The original vibration signal is filtered to obtain the filtered vibration signal. Then, the amplitude of the filtered vibration signal is analyzed to determine the stable position interval of the vibration sensing area and calculate the geometric center coordinates of the stable position interval. The obtained geometric center coordinates are used as the anchoring reference. The preset marker coordinates of the vibration synchronous sensing area in the positioning reference information are then mapped to the geometric center to generate the anchored vibration detection position coordinates, which are determined as the second temporary marker point for vibration detection.

[0098] Furthermore, the original vibration signal is filtered using wavelet threshold filtering. However, wavelet threshold filtering is a prior art in this field and is not an inventive solution of this application, so it will not be described in detail here.

[0099] Furthermore, amplitude analysis is performed on the filtered vibration signal to determine the stable position range of the vibration-induced region, including:

[0100] (1) Extract the vibration amplitude at each sampling time in the filtered vibration signal and calculate the amplitude statistics parameters of the entire filtered vibration signal, including the average amplitude, amplitude standard deviation, maximum amplitude and minimum amplitude;

[0101] (2) Based on the amplitude statistics parameters, set a stable fluctuation threshold. Usually, the average amplitude plus or minus twice the amplitude standard deviation is used as the stable fluctuation range.

[0102] (3) Select all sampling times within the stable fluctuation range. The vibration state corresponding to these sampling times can be regarded as stable vibration.

[0103] (4) By analyzing the sensor sensing positions corresponding to the stable vibration moments, the spatial distribution of these sensing points is statistically analyzed: if more than 90% of the stable sensing points are concentrated in a circular area with a radius of 3 mm centered on the preset marker coordinates, then the circular area is determined as the stable position interval of the vibration sensing area; if the distribution is more scattered, the signal acquisition time needs to be extended again and the number of stable vibration samples increased until a concentrated stable position interval is determined.

[0104] It should be noted that there is a corresponding relationship between the location of the vibration sensing area and the vibration amplitude. The larger the amplitude, the closer the sensing point is to the vibration source, which is the core area of ​​the arm's micro-movement; the smaller the amplitude, the farther the sensing point is from the core area.

[0105] Furthermore, the process of calculating the coordinates of the geometric center of this interval includes:

[0106] (1) Determine the boundary coordinates of the stable position interval: For a circular stable position interval, measure the maximum diameter and minimum diameter of the interval. The maximum diameter is the distance between the two farthest sensing points in the interval, and the minimum diameter is the distance between the two farthest sensing points in the direction perpendicular to the maximum diameter. Take the midpoint of the maximum diameter and the minimum diameter and draw a line. The intersection of the two lines is the preliminary position of the geometric center.

[0107] (2) Collect the three-dimensional coordinates of all sensing points within the stable position interval, and calculate the mean values ​​of the three-dimensional coordinates on the X-axis, Y-axis and Z-axis. The mean value of the X-axis is the arithmetic mean of the X coordinates of all sensing points, the mean value of the Y-axis is the arithmetic mean of the Y coordinates of all sensing points, and the mean value of the Z-axis is the arithmetic mean of the Z coordinates of all sensing points. The mean value of the three axes is used as the final value of the geometric center coordinates.

[0108] Furthermore, the obtained geometric center coordinates are used as the anchoring reference, and the preset marker coordinates of the vibration synchronous sensing area in the positioning reference information are mapped to the geometric center to generate the anchored vibration detection position coordinates, which are then determined as the second temporary marker point for vibration detection, including:

[0109] (1) Analyze the spatial deviation between the preset mark coordinates and the geometric center coordinates, calculate the coordinate difference between the two on the X-axis, Y-axis and Z-axis, and obtain the deviation vector;

[0110] (2) Based on the characteristics of the stable position range, determine whether the deviation is due to the offset between the actual position of the sensing area and the preset position. For example, if the arm placement position is slightly deviated from the preset area, and the deviation is within the preset range, such as the deviation of each axis being less than 5 mm, then the preset mark coordinates are translated to the geometric center coordinates. The translation amount is equal to the negative value of the deviation vector, so that the translated coordinates coincide with the geometric center coordinates.

[0111] (3) Verify the coordinates after translation, i.e. the vibration detection position coordinates after anchoring, and calculate the average distance between the coordinates and all sensing points in the stable position interval. If the average distance is less than the radius of the stable position interval, such as less than 3 mm, it means that the anchoring coordinates are in the center of the stable interval and meet the requirements. If the average distance exceeds the radius, the geometric center coordinate calculation process needs to be checked again, and the mapping is repeated after eliminating the calculation error.

[0112] (4) After verification, the coordinates of the anchored vibration detection position are determined as the second temporary marker point for vibration detection. The coordinate value, generation time, and radius of the stable position interval of the second temporary marker point for vibration detection are recorded and stored in the database.

[0113] The spatial coordinate system parameters of the synchronous detection area in the basic parameters are called, and the obtained human pulse detection first temporary marker point and vibration detection second temporary marker point are transformed to the same three-dimensional coordinate system. The spatial deviation value between the two points is calculated. If the spatial deviation value is greater than the preset first deviation threshold, coordinate association calibration is performed until the deviation value is less than or equal to the first deviation threshold. Finally, the calibrated human pulse detection first temporary marker point and vibration detection second temporary marker point are jointly determined as the positioning reference point for synchronous detection.

[0114] The coordinate correlation calibration process includes:

[0115] A1: Based on the coordinates of the first temporary marker point for human pulse detection after coordinate transformation and the second temporary marker point for vibration detection after coordinate transformation, historical coordinate data are collected to form a human pulse marker point coordinate sequence and a vibration marker point coordinate sequence.

[0116] A2: The coordinate sequence of human pulse marker points and the coordinate sequence of vibration marker points are fitted using the least squares method to construct a coordinate offset correction matrix. The least squares method is a prior art in this field and is not an inventive solution of this application, so it will not be described in detail here.

[0117] A3: Substitute the current coordinates of the second temporary vibration detection marker into the coordinate offset correction matrix to calculate the calibrated coordinates of the vibration detection marker;

[0118] Furthermore, the specific steps in A3 include:

[0119] (1) Obtain the second temporary marker point for vibration detection and the coordinate offset correction matrix;

[0120] (2) Perform matrix multiplication on the second temporary marker point of vibration detection and the coordinate offset correction matrix to obtain the corrected coordinate column vector, including the median value of the corrected X-axis coordinate, the median value of the corrected Y-axis coordinate, and the median value of the corrected Z-axis coordinate.

[0121] The median value of the corrected X-axis coordinate is equal to the first element of the first row of the coordinate offset correction matrix multiplied by the X-axis coordinate of the second temporary marker point of the vibration detection, plus the second element of the first row of the coordinate offset correction matrix multiplied by the Y-axis coordinate of the second temporary marker point of the vibration detection, plus the third element of the first row of the coordinate offset correction matrix multiplied by the Z-axis coordinate of the second temporary marker point of the vibration detection. The sum of the three products is the median value of the X-axis.

[0122] The median value of the corrected Y-axis coordinate is equal to the first element of the second row of the coordinate offset correction matrix multiplied by the X-axis coordinate of the second temporary marker point of the vibration detection, plus the second element of the second row of the coordinate offset correction matrix multiplied by the Y-axis coordinate of the second temporary marker point of the vibration detection, plus the third element of the second row of the coordinate offset correction matrix multiplied by the Z-axis coordinate of the second temporary marker point of the vibration detection. The sum of the three products is the median value of the Y-axis.

[0123] The median value of the corrected Z-axis coordinate is equal to the first element of the third row of the coordinate offset correction matrix multiplied by the X-axis coordinate of the second temporary marker point of the vibration detection, plus the second element of the third row of the coordinate offset correction matrix multiplied by the Y-axis coordinate of the second temporary marker point of the vibration detection, plus the third element of the third row of the coordinate offset correction matrix multiplied by the Z-axis coordinate of the second temporary marker point of the vibration detection. The sum of the three products is the median value of the Z-axis.

[0124] (3) Round the intermediate values ​​of the three axis coordinates obtained by matrix multiplication to obtain the final corrected X-axis, Y-axis and Z-axis coordinates, thereby obtaining the calibrated vibration detection marker coordinates;

[0125] (4) Store the calibrated vibration detection marker coordinates in the vibration detection coordinate library of the system database according to the preset format, and the storage path is associated with the original coordinates of the second temporary marker of vibration detection.

[0126] A4: Calculate the spatial deviation between the coordinates of the calibrated vibration detection marker and the coordinate sequence of the human pulse marker. If it is still greater than the first deviation threshold, return to the historical coordinate data acquisition process until the spatial deviation is less than or equal to the first deviation threshold, and obtain the calibrated first temporary marker for human pulse detection and the calibrated second temporary marker for vibration detection.

[0127] The process of confirming that the sensor is in a stable data acquisition state includes:

[0128] B1: Call the coordinates of the generated synchronous detection and positioning reference point to determine the pre-acquisition range. At the same time, call the sensor's pre-acquisition parameters in the basic parameters.

[0129] B2: Based on the determined preset acquisition range and pre-acquisition parameters, control the multi-channel sensors deployed in the detection area to perform pre-acquisition, generate a set of pre-acquisition signals including pulse pre-acquisition signals and vibration pre-acquisition signals, and each signal is accompanied by an acquisition timestamp and the corresponding sensor channel identifier.

[0130] B3: Extract human pulse pre-acquisition signal subsets and vibration pre-acquisition signal subsets from the pre-acquisition signal set, perform stability analysis on the two types of signal subsets respectively, and calculate the standard deviation and coefficient of variation of the pre-acquisition signal subsets;

[0131] like ,and ,at the same time, ,and If the signal meets the stability threshold, the sensor is determined to be in a stable acquisition state. Otherwise, the contact pressure of the control sensor is adjusted and the pre-acquisition process is repeated until all signals meet the stability threshold, confirming that the sensor is in a stable acquisition state. and These represent the standard deviation and coefficient of variation of the pre-acquired subset of human pulse signals, respectively. and These represent the standard deviation and coefficient of variation of the pre-acquired vibration signal subset, respectively. and These represent the standard deviation threshold and variation threshold of the pre-acquired subset of human pulse signals, respectively. and These represent the standard deviation threshold and variation threshold of the vibration pre-acquisition signal subset, respectively.

[0132] The generation of the purified synchronization signal set through the source denoising model includes:

[0133] C1: Input the hierarchically preprocessed human pulse-vibration synchronous original signal dataset into the pre-trained source denoising model; the source denoising model is built based on a deep learning network and includes an input layer, an attention mechanism intermediate layer and an output layer. The deep learning network is existing technology in this field and is not an inventive solution of this application, so it will not be described in detail here.

[0134] C2: The input layer of the source denoising model performs time-frequency conversion on the hierarchical preprocessed human pulse-vibration synchronization original signal, performs noise labeling through the attention mechanism intermediate layer, and finally performs noise reduction processing through the output layer to output the purified human pulse-vibration synchronization signal set.

[0135] The step of performing multi-algorithm game-theoretic fusion feature extraction based on the purified human pulse-vibration synchronization signal set includes:

[0136] D1: The purified human pulse-vibration synchronization signal set is split into independent human pulse signal subsets and vibration signal subsets according to signal type;

[0137] D2: The Fourier transform algorithm is used to perform frequency analysis on the subset of human pulse signal and the subset of vibration signal respectively to obtain the preliminary frequency characteristic values ​​of human pulse and vibration. The Fourier transform algorithm is the prior art in this field and is not an inventive solution of this application, so it will not be described in detail here.

[0138] D3: Wavelet transform algorithm is used to perform time-frequency analysis on the human pulse signal subset and vibration signal subset respectively to obtain the preliminary amplitude characteristic value of human pulse and the preliminary amplitude characteristic value of vibration. The wavelet transform algorithm is the prior art in this field and is not an inventive solution of this application, so it will not be described in detail here.

[0139] D4: The support vector machine algorithm is used to filter the preliminary frequency feature value of human pulse, the preliminary frequency feature value of vibration, the preliminary amplitude feature value of human pulse, and the preliminary amplitude feature value of vibration to generate a subset of effective features of human pulse and a subset of effective features of vibration. The support vector machine algorithm is the prior art in this field and is not an inventive solution of this application, so it will not be described in detail here.

[0140] D5: Using the Nash equilibrium strategy in game theory, the effective feature subsets of human pulse and vibration are fused to obtain the final frequency feature values ​​of human pulse signal, amplitude feature values ​​of human pulse signal, frequency feature values ​​of vibration signal, and amplitude feature values ​​of vibration signal.

[0141] Furthermore, the specific steps of D5 include:

[0142] (1) Clarify the preconditions for constructing the game participants and payoff function. First, build game frameworks for the effective feature subsets of human pulse and vibration respectively. For the effective feature subset of human pulse, determine the human pulse frequency feature and human pulse amplitude feature as game participants. These two participants need to work together to generate the final feature value that can accurately reflect the actual state of human pulse. For the effective feature subset of vibration, set the vibration frequency feature and vibration amplitude feature as game participants. The two need to cooperate to output the final feature value that can accurately reflect the essential attributes of vibration signal.

[0143] (2) Retrieve the reference values ​​of human pulse signal and vibration signal from the basic parameters. The human pulse reference value includes the average value of pulse frequency and the average value of pulse amplitude of healthy people. The vibration reference value includes the rated frequency and rated amplitude of the vibration source.

[0144] (3) Define the profit function, i.e. the objective function. Its core purpose is to minimize the deviation between the feature values ​​and the reference benchmark values, so as to ensure that the fused feature values ​​are closer to the true state:

[0145] In the fusion of human pulse characteristics, the actual measured values ​​and corresponding reference values ​​of human pulse frequency characteristics, as well as the actual measured values ​​and corresponding reference values ​​of human pulse amplitude characteristics, are first determined. Then, a benefit function is defined for the participants in the human pulse frequency characteristic fusion. This benefit function is represented by the negative value of the absolute value of the difference between the actual measured value and the reference value. The negative sign is used to reflect that the smaller the deviation between the actual measured value and the reference value, the higher the benefit of the participant. Similarly, a similar benefit function is defined for the participants in the human pulse amplitude characteristic fusion, which is the negative value of the absolute value of the difference between the actual measured value and the reference value. Finally, the overall joint benefit function is calculated, which is the sum of the benefit functions of the two participants. That is, the negative value of the absolute value of the difference between the actual measured value and the reference value of the human pulse frequency characteristic is added to the negative value of the absolute value of the difference between the actual measured value and the reference value of the human pulse amplitude characteristic. The ultimate goal is to maximize the result of this joint benefit function, which is essentially to minimize the total deviation of the frequency characteristic value and the amplitude characteristic value from their respective reference values.

[0146] For vibration feature fusion, the approach is similar to that of human pulse feature fusion: first, clarify the actual measured value and corresponding reference value of vibration frequency feature, and the actual measured value and corresponding reference value of vibration amplitude feature; define a benefit function for the vibration frequency feature participant, which is the negative value of the absolute value of the difference between the actual measured value of vibration frequency and the reference value; define a benefit function for the vibration amplitude feature participant, which is the negative value of the absolute value of the difference between the actual measured value of vibration amplitude and the reference value; the overall joint benefit function is the sum of the benefit functions of these two participants, and the goal is to maximize the result of this joint benefit function, that is, to minimize the total deviation of vibration frequency feature value and amplitude feature value from their respective reference values;

[0147] (4) Determine the feature weight variables and corresponding constraints. Since the final feature value is a weighted fusion of frequency and amplitude features, weight variables need to be introduced, including:

[0148] Based on the characteristics of human pulse, the weights of frequency features and amplitude features in the fusion process are set. According to the basic logic of weight allocation, the sum of these two weights must be equal to 1. This is to ensure that the final feature value obtained after fusion is within a reasonable range. At the same time, neither of these weights can be negative to avoid meaningless negative features affecting the final result.

[0149] For vibration characteristics, the weights of frequency characteristics and amplitude characteristics are also set. These two weights also need to satisfy the constraint that the sum is 1 and both are non-negative. The specific values ​​of these weight variables will directly affect the accuracy of the final characteristic values. Therefore, the optimal weight values ​​need to be solved by the Nash equilibrium method.

[0150] (5) Construct a Nash equilibrium solution model. The core principle of Nash equilibrium is that, given the opponent's strategy, each participant's chosen strategy is optimal. In other words, adjusting the weight of a single feature alone cannot improve the overall payoff. Specifically, this includes:

[0151] When constructing a fusion model of human pulse features, we first assume that the weights of amplitude features are fixed. Then, we need to find the optimal weights of frequency features so that the overall joint benefit function is maximized under the constraint that the sum of the two weights is 1. We can represent the weight of frequency features by subtracting the weight of amplitude features from 1, and then substitute it into the joint benefit function. It is important to note that the values ​​of amplitude features should be converted into equivalent values ​​of the same dimension as frequency features to avoid interference with the calculation results due to the different units of measurement. For example, the millivolt value of amplitude can be converted into a corresponding value related to pulse frequency. Then, we need to find the specific values ​​of amplitude feature weights to maximize the result of the joint benefit function. Similarly, we fix the weights of frequency features again and look for the optimal weights of amplitude features. Finally, we find a set of weights such that the weights of frequency features are the optimal choice when the weights of amplitude features are known, and the weights of amplitude features are also the optimal choice when the weights of frequency features are known. This set of weights satisfies the Nash equilibrium condition.

[0152] When constructing the Nash equilibrium model for vibration feature fusion, a similar logic to that used for human pulse feature fusion is adopted. The weight of the vibration frequency feature is represented by 1 minus the weight of the vibration amplitude feature, and substituted into the joint benefit function of the vibration features. The value of the vibration amplitude feature is also converted into an equivalent value of the same dimension as the vibration frequency. For example, the voltage value of the amplitude is converted into a corresponding value associated with the vibration frequency. Then, the optimal weight combination of the vibration frequency feature and the amplitude feature that satisfies the Nash equilibrium condition is solved.

[0153] (6) Entering the stage of solving for the optimal weights, i.e., the Nash equilibrium point, in order to accurately find the optimal weights, it is necessary to introduce historical feature data to assist in the calculation. Specifically, this involves retrieving the effective feature data of human pulse obtained from N sets of similar synchronous detections in the past. These data include the actual values ​​of pulse frequency and pulse amplitude of each set of detections; at the same time, the corresponding N sets of effective vibration feature data are retrieved, including the actual values ​​of vibration frequency and vibration amplitude of each set. Then, for each set of data, the joint benefit value under different weight combinations is calculated, and a table of correspondence between benefit and weight is constructed.

[0154] To find the optimal weights for human pulse characteristics, we start from 0 and end at 1, following the step size... The process involves progressively selecting the weight values ​​for amplitude features. For each amplitude feature weight selected, the corresponding frequency feature weight is obtained according to the rule that the sum of the two weights equals 1. Then, the average value of the combined benefit of N sets of human pulse feature data under this weight combination is calculated. The amplitude feature weight value that maximizes the average value is found; the corresponding frequency feature weight is 1 minus this amplitude feature weight value. After finding this weight combination, it is validated, for example, by increasing the frequency feature weight. Reduced weight of amplitude feature Calculate the average of the combined returns of the N sets of data at this point, and see if it has decreased compared to the previous average; then increase the weight of the magnitude feature. Frequency feature weights reduced Similarly, calculate the average of the combined returns and observe whether it decreases; if both adjusted averages decrease, it indicates that this weighted combination is the Nash equilibrium point of the fusion of human pulse characteristics.

[0155] To determine the optimal weights for vibration characteristics, a method similar to that used for human pulse characteristics is employed, ranging from 0 to 1 according to step size. Select the weights of vibration amplitude features to obtain the corresponding vibration frequency feature weights. Calculate the average value of the joint income of N sets of vibration feature data under each set of weights. Find the amplitude feature weights and corresponding frequency feature weights that maximize the average value. Then, confirm whether this set of weights is the Nash equilibrium point of vibration feature fusion by adjusting the weights.

[0156] (7) After obtaining the optimal weights, calculate the final eigenvalues ​​based on the optimal weights, including:

[0157] Regarding the characteristics of human pulse, the optimal weights of the frequency and amplitude characteristics of the human pulse obtained from the solution are substituted into the fusion calculation. When calculating the final frequency characteristic value of the human pulse signal, the optimal weight of the frequency characteristic is multiplied by the actual measured value of the frequency characteristic, and then the optimal weight of the amplitude characteristic is multiplied by the equivalent frequency value after the amplitude characteristic is converted. When calculating the final amplitude characteristic value of the human pulse signal, the optimal weight of the amplitude characteristic is multiplied by the actual measured value of the amplitude characteristic, and then the optimal weight of the frequency characteristic is multiplied by the equivalent amplitude value after the frequency characteristic is converted. After the calculation is completed, the deviation rate between the final frequency characteristic value and the corresponding reference value is calculated. The deviation rate is calculated by dividing the absolute value of the difference between the final frequency characteristic value and the reference value by the reference value, and then multiplying by 100% to obtain the percentage. Similarly, the deviation rate between the final amplitude characteristic value and the corresponding reference value is calculated. If both deviation rates are less than or equal to the preset second deviation threshold, the two final characteristic values ​​are retained. If either deviation rate is greater than the preset second deviation threshold, the step of solving for the optimal weight is returned, and the amount of historical data is increased to resolve for the optimal weight until the calculated deviation rate meets the requirement of being less than or equal to the preset second deviation threshold.

[0158] For vibration characteristics, the optimal weights of the obtained vibration frequency and amplitude characteristics are substituted into the fusion calculation. When calculating the final frequency characteristic value of the vibration signal, the optimal weight of the vibration frequency characteristic is multiplied by the actual measured value of the vibration frequency, and the optimal weight of the vibration amplitude characteristic is multiplied by the equivalent frequency value after vibration amplitude conversion. When calculating the final amplitude characteristic value of the vibration signal, the optimal weight of the vibration amplitude characteristic is multiplied by the actual measured value of the vibration amplitude, and the optimal weight of the vibration frequency characteristic is multiplied by the equivalent amplitude value after vibration frequency conversion. Then, the deviation rate between the final vibration frequency characteristic value and the reference value, and the deviation rate between the final vibration amplitude characteristic value and the reference value are calculated. If both deviation rates are less than or equal to the preset third deviation threshold, the two final characteristic values ​​are retained. If either deviation rate is greater than the preset third deviation threshold, the process returns to the step of solving for the optimal weights, and the historical data volume is increased before resolving until the deviation rate meets the requirements.

[0159] (8) Compare the calculated final frequency characteristic value of the human pulse signal, the final amplitude characteristic value of the human pulse signal, the final frequency characteristic value of the vibration signal, and the final amplitude characteristic value of the vibration signal with the reasonable range of the characteristic values ​​preset in the basic parameters. If all the final characteristic values ​​are within the corresponding preset reasonable range, the feature fusion process is confirmed to be complete, and the four final characteristic values ​​are output. If any final characteristic value exceeds the corresponding preset reasonable range, the reason for this situation is analyzed, and the construction process of the Nash equilibrium model is returned. The weight coefficients related to the deviation in the payoff function are adjusted, and the entire feature fusion process is restarted until all the final characteristic values ​​meet the corresponding reasonable range requirements.

[0160] Example 2:

[0161] Please see Figure 3 Another embodiment of the present invention provides: a target object frequency and amplitude detection system, comprising:

[0162] Information acquisition module, benchmark point generation module, feature extraction module, threshold verification module;

[0163] The information acquisition module is used to collect and integrate the basic parameters and positioning reference information required for human pulse-vibration synchronous detection, providing an initial basis for subsequent dynamic positioning and signal acquisition.

[0164] The reference point generation module employs a differentiated dynamic positioning strategy based on human pulse and vibration signals to perform marking operations on the synchronous detection area, generating accurate synchronous detection positioning reference points to ensure the stability and synchronization of sensor acquisition positions.

[0165] The feature extraction module is used to perform hierarchical preprocessing and noise reduction on the original signal dataset, and then extract the frequency and amplitude feature values ​​of human pulse signals and vibration signals through multi-algorithm fusion.

[0166] The threshold verification module is used to perform dynamic threshold verification on the output feature values, generate detection reports and data ledgers, and complete the visualization output and data archiving of detection results.

[0167] The information acquisition module includes: an instruction receiving unit, a template calling unit, a feature recognition unit, and a benchmark generation unit;

[0168] The instruction receiving unit is used to receive the synchronous detection target type selection instruction input by the user and temporarily store the instruction in the system cache;

[0169] The template calling unit is used to automatically call the pre-stored corresponding type synchronous detection parameter template based on the user's input synchronous detection target type selection instruction, and associate the synchronous detection parameter template with the user instruction to form an initial parameter framework;

[0170] The feature recognition unit is used to control the infrared scanning sensor and vibration pre-acquisition sensor deployed in the detection area to start scanning, automatically identify the physical features of the synchronous detection area, and generate structured physical feature data;

[0171] The baseline generation unit is used to match and fuse the initial parameter framework and synchronous detection target type selection instructions output by the template calling unit with the generated physical feature data, fill in the parameter items in the synchronous detection parameter template with specific values, determine the specific content of the positioning baseline information, and finally generate complete basic parameters and positioning baseline information.

[0172] The benchmark point generation module includes: a dynamic positioning unit and a coordinate association calibration unit;

[0173] A dynamic positioning unit is used to generate a first temporary marker point for human pulse detection and a second temporary marker point for vibration detection.

[0174] The coordinate association calibration unit is used to unify the first temporary marker point for human pulse detection and the second temporary marker point for vibration detection into the same coordinate system, eliminate coordinate deviation, and generate the final positioning reference point.

[0175] The feature extraction module includes: a hierarchical preprocessing unit, a source denoising unit, and a feature extraction unit;

[0176] Hierarchical preprocessing units are used to progressively eliminate invalid data, baseline drift, and dimensional differences in the original signal dataset;

[0177] The source denoising unit is used to remove inherent noise and interference noise from the signal through a pre-trained source denoising model, and output a purified human pulse-vibration synchronization signal set.

[0178] The feature extraction unit is used to split the purified human pulse-vibration synchronization signal set, and output the final feature value through Fourier transform, wavelet transform, support vector machine screening and Nash equilibrium fusion.

[0179] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments under the guidance of the present invention without departing from the spirit and scope of the present invention. All of these variations are within the protection scope of the present invention.

Claims

1. A method for detecting the frequency and amplitude of a target object, characterized in that, include: Acquire the basic parameters and positioning reference information for synchronous detection of human pulse and vibration; Based on basic parameters and positioning reference information, a dynamic positioning strategy is used to perform a marking operation on the synchronous detection area to generate positioning reference points; the dynamic positioning strategy includes a tracking-deviation compensation dynamic positioning strategy for human pulse signals and a vibration filtering-coordinate anchoring dynamic positioning strategy for vibration signals. After confirming that the sensor is in a stable acquisition state, using the positioning reference point as the acquisition position reference, multi-channel adaptive sensing and synchronous data acquisition are performed to generate a human pulse-vibration synchronous raw signal dataset. Hierarchical preprocessing is performed on the original human pulse-vibration synchronization signal dataset. A purified human pulse-vibration synchronization signal set is generated through a source denoising model. Based on the purified human pulse-vibration synchronization signal set, multi-algorithm game fusion feature extraction is performed to generate human pulse signal frequency feature value, human pulse signal amplitude feature value, vibration signal frequency feature value, and vibration signal amplitude feature value, respectively. Dynamic threshold verification is performed on the frequency and amplitude characteristic values ​​of human pulse signals and vibration signals respectively. Based on the verification results, a human pulse-vibration synchronous detection report is generated, and a synchronous detection data ledger is formed.

2. The target object frequency and amplitude detection method as described in claim 1, characterized in that, The basic parameters include the type of target object to be detected synchronously, the physical characteristics of the synchronous detection area, and the accuracy calibration parameters of the infrared positioning module; the positioning reference information includes the preset marker coordinates of the human pulse detection area, the preset marker coordinates of the vibration synchronous sensing area, and the common allowable deviation range for the detection of human pulse signals and vibration signals; the type of target object to be detected synchronously includes human pulse signals and vibration signals; the physical characteristics of the synchronous detection area include the physiological characteristics of the human arm detection site and the characteristics of the vibration transmission path; When acquiring the basic parameters and positioning reference information for human pulse-vibration synchronous detection, the system receives the user's input instruction for selecting the type of synchronous detection target through a preset parameter input interface, and calls the pre-stored synchronous detection parameter template of the corresponding type based on the instruction for selecting the type of synchronous detection target. The synchronous detection parameter template includes a basic parameter framework and a positioning reference information framework for the detection of corresponding types of target objects. The system controls the sensors deployed in the detection area to scan the synchronous detection area and automatically identify the physical characteristics of the synchronous detection area. Specifically, for the human pulse detection area, it identifies physiological characteristics including skin thickness and blood vessel distribution density; for the vibration detection area, it identifies path characteristics including the material and hardness of the vibration transmission medium and generates physical feature data. The synchronous detection parameter template and the user-input synchronous detection target type selection command are used as the initial inputs to obtain the basic parameters. The initial inputs are matched and fused with the generated physical feature data. The parameter items in the synchronous detection parameter template are filled with specific values ​​to determine the specific values ​​of the preset marker coordinates and the precise values ​​of the common allowable deviation range in the positioning reference information. Finally, the basic parameters and positioning reference information are generated.

3. The target object frequency and amplitude detection method as described in claim 2, characterized in that, The step of using a dynamic positioning strategy to perform a marking operation on the synchronous detection area and generate positioning reference points includes: For the target area corresponding to the human pulse signal in synchronous detection, the accuracy calibration parameters of the infrared positioning module in the acquired basic parameters, as well as the preset marker coordinates and common allowable deviation range of the human pulse detection area in the positioning reference information, are called. The target area corresponding to the human pulse signal in synchronous detection is tracked in real time to generate the real-time position coordinates of the human pulse detection area. The obtained real-time position coordinates of the human pulse detection area are compared with the preset marker coordinates of the human pulse detection area in the positioning reference information to calculate the three-dimensional position deviation value in the three-dimensional rectangular coordinate system, and the spatial modulus of the three-dimensional position deviation value is calculated. The relationship between the calculated spatial modulus and the common allowable deviation range is determined. If the spatial modulus is less than or equal to the common allowable deviation range, the corresponding real-time position coordinates are directly determined as the first temporary marker point for human pulse detection. If the spatial modulus is greater than the common allowable deviation range, the accuracy calibration parameters of the infrared positioning module are called to compensate and correct the three-dimensional position deviation value to obtain the compensated human pulse detection position coordinates, and this is determined as the first temporary marker point for human pulse detection. The accuracy calibration parameters of the infrared positioning module include the temperature drift compensation coefficient and the optical distortion correction matrix.

4. The target object frequency and amplitude detection method as described in claim 3, characterized in that, The step of using a dynamic positioning strategy to perform a marking operation on the synchronous detection area and generate positioning reference points also includes: For the sensing area corresponding to the vibration signal caused by the slight movement of the arm during synchronous detection, the preset marker coordinates of the vibration synchronous sensing area in the positioning reference information are called. The vibration sensor is activated to collect signals from the sensing area corresponding to the vibration signal caused by the slight movement of the arm during synchronous detection, generating the original vibration signal. The original vibration signal is filtered to obtain the filtered vibration signal. The amplitude of the filtered vibration signal is then analyzed to determine the stable position range of the vibration sensing area, and the geometric center coordinates of the stable position range are calculated. The obtained geometric center coordinates are used as the anchoring reference, and the preset marker coordinates of the vibration synchronous sensing area in the positioning reference information are mapped to the geometric center to generate the anchored vibration detection position coordinates, which are then determined as the second temporary marker point for vibration detection.

5. The target object frequency and amplitude detection method as described in claim 4, characterized in that, The step of using a dynamic positioning strategy to perform a marking operation on the synchronous detection area and generate positioning reference points also includes: The spatial coordinate system parameters of the synchronous detection area in the basic parameters are called to transform the first temporary marker point of human pulse detection and the second temporary marker point of vibration detection to the same three-dimensional coordinate system. The spatial deviation value between the two points is calculated. If the spatial deviation value is greater than the preset first deviation threshold, coordinate association calibration is performed until the deviation value is less than or equal to the first deviation threshold. Finally, the calibrated first temporary marker point of human pulse detection and the calibrated second temporary marker point of vibration detection are jointly determined as the positioning reference point of synchronous detection. The coordinate correlation calibration process includes: Based on the coordinates of the first temporary marker point for human pulse detection after coordinate transformation and the second temporary marker point for vibration detection after coordinate transformation, historical coordinate data are collected to form a human pulse marker point coordinate sequence and a vibration marker point coordinate sequence. The least squares method was used to fit the coordinate sequences of human pulse marker points and vibration marker points to construct a coordinate offset correction matrix; Substitute the current coordinates of the second temporary vibration detection marker into the coordinate offset correction matrix to calculate the calibrated vibration detection marker coordinates. Calculate the spatial deviation between the calibrated vibration detection marker coordinates and the human pulse marker coordinate sequence. If it is still greater than the first deviation threshold, return to the historical coordinate data acquisition process until the spatial deviation is less than or equal to the first deviation threshold, and obtain the calibrated first temporary marker for human pulse detection and the calibrated second temporary marker for vibration detection.

6. The target object frequency and amplitude detection method as described in claim 5, characterized in that, The process of confirming that the sensor is in a stable data acquisition state includes: Call the coordinates of the generated synchronous detection and positioning reference point to determine the pre-acquisition range. At the same time, call the sensor's pre-acquisition parameters in the basic parameters. Based on the determined preset acquisition range and pre-acquisition parameters, the multi-channel sensors deployed in the detection area are controlled to perform pre-acquisition, generating a set of pre-acquisition signals including pulse pre-acquisition signals and vibration pre-acquisition signals, and each signal is accompanied by an acquisition timestamp and the corresponding sensor channel identifier. From the pre-acquisition signal set, extract the human pulse pre-acquisition signal subset and the vibration pre-acquisition signal subset respectively, perform stability analysis on the two types of signal subsets respectively, and calculate the standard deviation and coefficient of variation of the pre-acquisition signal subsets; like ,and ,at the same time, ,and If the signal meets the stability threshold, the sensor is determined to be in a stable acquisition state. Otherwise, the contact pressure of the control sensor is adjusted and the pre-acquisition process is repeated until all signals meet the stability threshold, confirming that the sensor is in a stable acquisition state. and These represent the standard deviation and coefficient of variation of the pre-acquired subset of human pulse signals, respectively. and These represent the standard deviation and coefficient of variation of the pre-acquired vibration signal subset, respectively. and These represent the standard deviation threshold and variation threshold of the pre-acquired subset of human pulse signals, respectively. and These represent the standard deviation threshold and variation threshold of the vibration pre-acquisition signal subset, respectively.

7. The target object frequency and amplitude detection method as described in claim 6, characterized in that, The generation of the purified synchronization signal set through the source denoising model includes: The hierarchically preprocessed human pulse-vibration synchronous original signal dataset is input into a pre-trained source tracing and denoising model; the source tracing and denoising model is built based on a deep learning network and includes an input layer, an attention mechanism intermediate layer, and an output layer; The source denoising model's input layer performs time-frequency conversion on the hierarchically preprocessed original human pulse-vibration synchronization signal, and performs noise labeling through the attention mechanism intermediate layer. Finally, the output layer performs noise reduction processing and outputs a purified human pulse-vibration synchronization signal set.

8. The target object frequency and amplitude detection method as described in claim 7, characterized in that, The step of performing multi-algorithm game-theoretic fusion feature extraction based on the purified human pulse-vibration synchronization signal set includes: The purified human pulse-vibration synchronization signal set was split into independent human pulse signal subsets and vibration signal subsets according to signal type; The Fourier transform algorithm was used to perform frequency analysis on the subsets of human pulse signals and vibration signals to obtain preliminary frequency characteristic values ​​of human pulse and vibration. Wavelet transform algorithm was used to perform time-frequency analysis on the human pulse signal subset and vibration signal subset to obtain the preliminary amplitude characteristic values ​​of human pulse and vibration. The support vector machine algorithm is used to filter the preliminary frequency feature value of human pulse, the preliminary frequency feature value of vibration, the preliminary amplitude feature value of human pulse, and the preliminary amplitude feature value of vibration, and to generate effective feature subsets of human pulse and vibration. The effective feature subsets of human pulse and vibration are fused using the Nash equilibrium strategy in game theory to obtain the final frequency feature values ​​of human pulse signal, amplitude feature values ​​of human pulse signal, frequency feature values ​​of vibration signal, and amplitude feature values ​​of vibration signal.

9. A target object frequency and amplitude detection system, used to implement the target object frequency and amplitude detection method according to any one of claims 1-8, characterized in that, include: Information acquisition module, benchmark point generation module, feature extraction module, threshold verification module; The information acquisition module is used to collect and integrate the basic parameters and positioning reference information of human pulse-vibration synchronous detection. The reference point generation module uses a differentiated dynamic positioning strategy based on human pulse and vibration signals to perform a marking operation on the synchronous detection area and generate synchronous detection positioning reference points. The feature extraction module is used to perform hierarchical preprocessing and noise reduction on the original signal dataset, and then extract the frequency and amplitude feature values ​​of human pulse signal and vibration signal through multi-algorithm fusion. The threshold verification module is used to perform dynamic threshold verification on the output feature values, generate detection reports and data ledgers, and complete the visualization output and data archiving of detection results.

Images