CN117204826A

CN117204826A - Pulse condition acquisition method of multipoint pulse condition sensor

Info

Publication number: CN117204826A
Application number: CN202310882962.3A
Authority: CN
Inventors: 杨志强; 孙承瑞; 许斌; 金军; 李凯
Original assignee: Shenyang Tianren Heyi Technology Co ltd
Current assignee: Shenyang Tianren Heyi Technology Co ltd
Priority date: 2023-07-19
Filing date: 2023-07-19
Publication date: 2023-12-12

Abstract

A pulse condition acquisition method of a multipoint pulse condition sensor is technically characterized by comprising the following steps: step K1, obtaining an original pulse graph curve of each acquisition point through an array sensor; step K2, determining the direction of each pulse graph curve; step K3, determining the fluctuation period Ti of each pulse graph curve: step K4, determining the duration Twin of the sliding window; step K5, periodically determining the base line of each pulse graph curve, and synchronizing Ti of each forward/reverse pulse graph curve into a unified pulse condition period Tk; step K6, outputting a pulse graph curve after baseline correction: taking a window starting point Ts-pulse condition period Tk-1 as a real-time output period, wherein the output interval is 1 time/s, and finally obtaining a corrected pulse graph curve set of each acquisition point. And calibrating each pulse graph curve by acquiring a reference line and a pulse condition period so as to keep synchronization during data storage and reproduction.

Description

Pulse condition acquisition method of multipoint pulse condition sensor

Technical Field

The invention relates to a use method of a sensor, in particular to a pulse condition acquisition method of a multipoint pulse condition sensor, which is mainly applicable to baseline correction and time domain correction of the multipoint sensor.

Background

The pulse signals can reflect rich physiological and pathological information, and the traditional Chinese medicine pulse diagnosis mainly applies 'floating', 'middle', 'sinking' pressing forces to 'cun', 'guan', 'chi' three areas to collect the pulse signals by pressing the cun-kou surface of the arm with fingers, and has the characteristics of rapidness, no wound and the like. Along with the development of biomedical detection technology, the development of pulse diagnosis instruments is becoming the focus of attention of all parties, and especially the wearable pulse diagnosis instrument has the advantage of monitoring pulse signals in real time, and becomes a hot spot in the field of pulse diagnosis instrument research in recent years. In the prior art, regarding the technical scheme of radial artery positioning and detection, the IPC classification number mainly related to the technical scheme is A61B5/02. The existing pulse condition acquisition and diagnosis system is mostly single-point acquisition and single-point reproduction, but those skilled in the art know that the pulse condition is actually diffused outwards by taking radial artery as a central line when beating, and the change trend of the amplitude from high to low is presented, so that the pulse is detected as far as possible in a large range to obtain the optimal acquisition point position during acquisition, but the optimal acquisition position is hardly obtained by a single-point remote acquisition mode, so that the situation of misdiagnosis is probably caused in the remote reproduction mode. In order to realize the acquisition and reproduction of pulse conditions, the following technical schemes are disclosed in the prior art. The utility model patent application of publication No. CN107898445A discloses a wearable intelligent pulse diagnosis instrument, which comprises a self-adaptive wrist strap and a measuring head group, wherein the measuring head group consists of four measuring heads capable of independently moving and a longitudinal linear motion driving unit, and can drive the tips of the measuring heads to automatically find the positions of a size, a closing position, a ruler and a lower position along X, Y, Z, automatically apply pressing force and stably monitor and record pulse signals under different pressure stimuli. However, this solution has the following drawbacks: according to the scheme, the general position of the radial artery is determined only according to the amplitude of the pressing veins of the finger, because visualization cannot be realized, and because of the physiological structural difference of the radial artery (blood vessel) among different patients, the optimal pulse taking position is difficult to accurately position. In addition, the utility model patent application of publication No. CN113440113A, CN114224308A also discloses a similar approach to locating the radial artery by acquiring a pulse amplitude signal after compression. The utility model application of publication number CN112842292a also discloses a scheme for detecting cerebral arteries by means of a scheme of combining an air bag type wristband with an array sensor. The utility model patent application of publication number CN114569087b also discloses obtaining radial artery fluctuation information through a combination of non-penetrating visible light, infrared light cameras and array sensor broadcast mill. The utility model patent application of publication No. CN114041758A discloses a method for non-contact positioning of radial artery based on deep learning, but obviously, the scheme is characterized in that a large amount of data is marked manually, and only the 'radial artery region' can be generally positioned only through video data, but the main vessel position of the radial artery is not precisely positioned, so that the training result cannot be used for interaction of a remote pulse diagnosis medical care terminal (reproduction terminal). The utility model patent of publication number CN217090710U discloses a state recognition bracelet for the aged at home based on pulse detection, which is used for detecting infrared light transmitted by an infrared light emitting lamp penetrating through wrist pulse in real time through infrared emitting lamps and photodiodes respectively arranged in an upper ring body and a lower ring body. This solution, although different from single-sided, infrared detection, determines heart rate by detecting only fluctuations in pulse, not for locating radial artery position.

In the prior art, regarding the technical scheme of radial artery signal reproduction, the IPC classification number mainly related to the technical scheme is A61B5/02: the invention patent applications of publication No. CN108742546A, publication No. CN112690764A, publication No. CN112690764A, publication No. CN113367669A and publication No. CN114864630A publication No. CN115721271A all adopt three single-point pulse condition sensors in the disclosed technical scheme, and the array sensor data of the acquisition terminal cannot be completely reproduced in a point-to-point mode due to the fact that the number of the 'touch feedback points' is only three. Pulse width, pulse length, and pulse rhythm (interval and contour of pulse) cannot be measured accurately. In addition, in the technical scheme disclosed in the patent application of publication number CN115721271a, a mode of combining a fingerstall type flexible electrode array and an electric stimulation element is adopted in the reproduction module, but obviously, compared with the physical reproduction of directly simulating vascular fluctuation, the electric stimulation reproduction effect is poorer, and the technical teaching of point-to-point reproduction of the acquisition terminal and the reproduction terminal is not disclosed. The pulse taking method of the existing pulse diagnosis instrument mostly adopts a cunkou pulse taking method, and the pulse signal of the pulse condition is converted into an electronic signal, so that informatization and intellectualization of pulse diagnosis are realized. The invention patent application of publication number CN106137147b discloses a device and a method for acquiring pulse condition data based on human-computer interaction, which mainly comprises the steps of judging whether a pulse signal acquired by a sensor array element at the most middle position is the strongest pulse signal, if so, indicating a user to apply corresponding pressure and respectively acquire pulse position waveforms corresponding to the pressure, and if not, indicating the user to move the sensor array element at the most middle position to the position of the strongest pulse signal according to the strongest pulse signal until the sensor array element at the most middle position is moved to the position of the strongest pulse signal; the strongest pulse signal is converted into pulse data and displayed to the user. The method mainly comprises the steps of obtaining the optimal pulse taking value of the 'off' part, wherein the feeling of fingers in the pulse diagnosis of a clinician is higher than that of a sensor, and the pulse taking pressures of the three fingers of the cun off ruler are different, so that the technical scheme is low in accuracy. The invention patent application of publication number CN105534490A discloses a finger pressure type pulse-taking instrument and a control method, wherein the method comprises the following steps: step S1, acquiring floating and sinking voltage output signals of three parts of a finger cun guan ruler of a person to be acquired by adopting a pulse diagnosis instrument with a plurality of finger sleeve sensors; step S2, the pulse-taking instrument transmits the floating-middle sinking voltage output signal to a computer; step S3, the computer converts the floating and sinking voltage output signal into a floating and sinking pulse pressure value, compares the floating and sinking pulse pressure value with a preset floating and sinking pulse pressure range, and carries out corresponding operation on the floating and sinking pulse pressure value according to the comparison result; and S4, the computer stores a comparison result and an operation result obtained after the operation of the floating and sinking pulse pressure value. The technical proposal focuses on the three floating, sinking and taking pulse, but the optimal pulse taking pressure of each part is not clear.

Disclosure of Invention

The invention aims to provide a pulse condition acquisition method of a multipoint pulse condition sensor, which fundamentally solves the problems, and has the advantages of convenient use, accurate radial artery positioning, high pulse condition reproduction precision and the like. Better simulate clinical pulse taking and improve the accuracy of pulse taking of a pulse diagnosis instrument.

In order to achieve the above purpose, the present invention provides the following technical solutions: the pulse condition acquisition method of the multipoint pulse condition sensor is technically characterized by comprising the following steps of: step K1, obtaining an original pulse graph curve of each acquisition point through an array sensor; step K2, determining the direction of each pulse graph curve; step K3, determining the fluctuation period Ti of each pulse graph curve; step K4, determining the duration Twin of the sliding window; step K5, periodically determining the base line of each pulse graph curve, and synchronizing Ti of each forward/reverse pulse graph curve into a unified pulse condition period Tk: step K6, outputting a pulse graph curve after baseline correction: taking a window starting point Ts-pulse condition period Tk-1 as a real-time output period, wherein the output interval is 1 time/s, and finally obtaining a corrected pulse graph curve set of each acquisition point.

The invention has the beneficial effects that: according to the whole technical scheme, the acquisition and reproduction of the pulse condition can be realized in a mode of point-to-point matching of the flexible array sensor and the finger die assembly with the multipoint reproduction function. On the premise that the acquisition terminal locates the radial artery in advance through a visualization method, the reproduction terminal simulates a pulse searching method of a doctor in an actual pulse diagnosis scene in an active pressing mode, so that the condition of misdiagnosis is effectively avoided, and further, high-precision pulse position and pulse condition acquisition and reproduction are realized.

Structurally, the pulse searching process is different from that of the existing acquisition terminal, and the acquisition terminal directly positions the acquisition instruction on the radial artery in a visual pre-positioning mode. One side of the pulse taking frame component is propped against the root of the palm to be used as a positioning pulse taking origin, and the radial artery is positioned at the center of the pulse taking port to be used as secondary positioning through visual observation of real-time development, so that the problem that the XYZ space coordinates of the finger are required to be acquired in the remote control positioning in the existing pulse searching process is avoided, and the pulse searching accuracy is effectively improved. When the pulse is collected manually, the reproduction terminal can complete the pulse collection only by controlling the pressing force of the collection finger on the cun guan chi acupoint by a far end. When the pulse is automatically collected, the point position amplitude peak value is obtained by setting a certain amplitude value, for example 25g, on the collection finger component corresponding to the cun guan ruler. Meanwhile, through multi-point high-precision reproduction simulation, the problem that the existing acquisition terminal needs repeated positioning in the XZ direction for multiple times and the accurate positioning is still difficult for pressing in the Y direction is solved. And misdiagnosis caused by inaccurate positioning during single-point acquisition is avoided.

The visual pulse taking principle that hemoglobin absorbs infrared light more strongly than other tissues is utilized, near infrared light with specific wavelength is projected to the skin surface through an infrared lamp module at the bottom of the wrist, and a skin infrared image is acquired through a photosensitive element of an infrared camera. Infrared vascular imaging methods are clear for venous vascular imaging, but not for arterial blood imaging. Only the radial artery for pulse diagnosis between the flexor carpi radialis and the brachiocephalic muscle is identified, and the interference of most veins is shielded by utilizing a physical device in a rectangular area, and the outline of the radial artery is clearly displayed on a display terminal in real time after the acquired infrared original image is processed. And presetting a positioning program, and giving prompt information of positioning completion on a display terminal when the radial artery is positioned in a specific area range of the pulse taking port.

The wrist strap rectangular window mode is utilized to combine the images to collect pulse conditions, so that the range and difficulty of pulse collection are reduced; the circular positioning device is used for assisting in image calculation, and the pulse picking image is corrected, so that pulse picking can be conveniently carried out at all angles. The up-and-down motion can be realized independently by adopting the mechanical fingers of the gear rack structure and the up-and-down limiting assembly. And the pulse taking method of floating, middle and sinking of traditional Chinese medicine is realized through pressure value feedback of the array pressure sensors. Pulse condition period identification for a multi-point array pulse condition sensor; distinguishing forward and backward waveforms of pulse signals of the multi-point array pulse sensor, and performing a baseline drift removing algorithm of the forward and backward waveforms; a method for outputting pulse conditions in real time by adopting a sliding window mode; the multipoint sensor can collect all key finger sense information such as pulse position, pulse force, tension, length, frequency, rhythm, thickness, fluency, eight major factors, unique and the like.

The method comprises the steps of reproducing a mode that multiple array points of a terminal uniquely correspond to multiple voice coil motor assemblies, receiving amplitude-time waveform data from an acquisition terminal through a gateway board in a low-time delay mode, sending the amplitude-time waveform data to driving boards of the voice coil motor assemblies after processing the amplitude-time waveform data through an upper computer main board, outputting action currents to coils through the driving boards, driving a cover body to drive valve rods of a reciprocating hydraulic cylinder to act along a linear guide rail, driving hydraulic oil to rhythmically impact bionic skin of a finger die assembly, and finally projecting fluctuation of corresponding point positions of an array sensor of the acquisition terminal on the bionic skin of the finger die assembly in a uniquely corresponding mode in real time. Meanwhile, the pressure sensor at the bottom of the finger die seat is used for collecting pressing force data of doctors, so that the finger die seat is suitable for the difference of pulse taking force channels of different schools of traditional Chinese medicine, and the pulse diagnosis result difference is caused, and the compatibility is effectively improved. In addition, the array points in each row are arranged at the bottom of the bionic skin in a staggered mode on the premise of a certain number of array points as high precision reduction as possible. Specifically, the array points of the second row are disposed at spaced locations of the array points of the first row, the array points of the third row are disposed at spaced locations of the second row (which may coincide with the lateral coordinates of the first row), and so on. The invention mainly takes four groups of array points which are staggered as an example. By the arrangement mode, the coverage area of the array points is increased when the same number of array points are adopted.

Drawings

Fig. 1 is a schematic view of a terminal according to the present invention.

Fig. 2 is a schematic diagram of the reproduction terminal (pipeline not shown) of the present invention.

Fig. 3 is a schematic structural diagram I of the voice coil motor assembly of fig. 2.

Fig. 4 is a schematic structural diagram II of the voice coil motor assembly of fig. 2.

Fig. 4a is a schematic view of a part of the quick-connect structure in fig. 4.

Fig. 5 is a schematic view of the finger die assembly of fig. 2.

Fig. 6 is a schematic perspective view of the finger die holder in fig. 5.

Fig. 6a is a schematic perspective view of another finger die holder according to the present invention.

Fig. 6b is a schematic isometric side view of the biomimetic skin of fig. 6.

Fig. 7 is a schematic structural diagram along a direction a of fig. 6, which illustrates one array mode of the present invention.

Fig. 7a is a schematic cross-sectional view of fig. 7.

FIG. 7b is a schematic diagram showing the pulse replication results of one of the finger die assemblies of FIG. 7.

Fig. 7c is a schematic diagram of a wave form acquired by one of the acquisition finger assemblies.

Fig. 7d is a schematic diagram of the hydraulic principle of one of the finger die assemblies in fig. 7.

Fig. 8 is a schematic structural diagram of the use state of the acquisition terminal according to the present invention.

Fig. 9 is a schematic structural diagram of an acquisition terminal according to the present invention.

Fig. 10 is an exploded view of the acquisition terminal of the present invention (different strap configurations).

Fig. 11 is a schematic view of the finger assembly of fig. 10.

Fig. 12 is an infrared raw image of a visualized pulse taking time.

Fig. 12a is an infrared image after rotation correction.

Fig. 12b is an image of fig. 12a after the positioning cut and after the primary and secondary edge sharpening.

Fig. 12c is the binarized image of fig. 12 b.

Fig. 12d is an image of the bone extracted from fig. 12 c.

Fig. 12e is a schematic diagram of the longest path in fig. 12 d.

Fig. 12f is a schematic diagram of the position of the three adjacent pixels of the current pixel taken by the micro rectangle.

Fig. 12g is a schematic view of the radial artery region.

FIG. 12h is a schematic diagram of the result of the reduction of FIG. 12g to FIG. 12.

FIG. 13 is a flow chart of the pulse taking method of the present invention.

FIG. 14 is a pulse diagram of an array sensor.

Fig. 15 is a three-dimensional schematic diagram of time domain analysis and frequency domain analysis of the acquisition terminal of the present invention.

FIG. 16 is a schematic diagram of the overall architecture of the system of the present invention.

FIG. 17 is a flow chart of the multipoint pulse picking method according to the present invention.

FIG. 17a shows the raw pulse signals at each acquisition point without baseline calibration.

FIG. 17b is a plot of the positive and negative pulse and pulse derivative curves of an acquisition point in FIG. 17a at a detection interval.

Fig. 17c shows the forward pulse plot of the collection points of fig. 17b and the pulse plot derivative curve (unfiltered).

FIG. 17d shows the filtered derivative of the pulse pattern of FIG. 17c and its corresponding peak pulse pattern.

Fig. 17e is a graph of the pulse after removal of the baseline shift.

FIG. 17f shows the pulse signals at each of the acquisition points after baseline calibration.

Detailed Description

The following describes the details of the present invention in detail through specific embodiments with reference to fig. 1 to 17. In the overall design of the reproduction terminal 1, since a single pressure reproduction point cannot provide enough information to accurately reproduce the pressure range in the pulse condition. If a plurality of pressure reproduction points are combined together and arranged in a staggered way, the gully sense of each pressure reproduction point can be eliminated, so that pulse condition data with higher resolution and more accurate positioning can be obtained, and misdiagnosis is avoided. As shown in fig. 1-2, the reproduction terminal 1 adopts a drawer type mounting bracket 15 as an integral structural support, and is matched with an external streamline housing as a main body framework. Because the voice coil motor assembly 13 array is adopted, a drawer type structure is adopted for facilitating daily use, the corresponding voice coil motor assembly 13 is conveniently and rapidly positioned during detection, and rapid disassembly and assembly are completed. In order to realize visual communication inquiry between doctors and patients, common components in the field, such as a camera 111, a touch screen 112, a microphone 113, a power button 114, a loudspeaker 115 and the like, are also arranged on the panel assembly 11. In order to expand the use function, a data interface (for example, a USB interface of a reproduction terminal, not shown in the figure) is reserved at the back of the shell, so that the debugging system is convenient or data can be imported and exported, and a power interface and a main switch (not shown in the figure) are reserved. To ensure stable operation of the hardware device, a heat sink (not shown) is also reserved on the back.

For complete machine control, the electric control assembly 14 mainly adopts a gateway board 141 for remote communication, an AC-DC power adapter board 142 for complete machine power supply, and an upper computer main board 143 for loading a main control system. In order to match with the high-precision reproduction of pulse conditions, a finger die assembly 12 with multi-point position control is arranged below a touch screen 112, a plurality of voice coil motor assemblies 13 matched with the number of the reproduction points are arranged on a drawer type mounting bracket 15, and a hydraulic oil tank 16 serving as a liquid source is arranged on the drawer type mounting bracket 15 in order to match with the use of the voice coil motor assemblies 13. The hydraulic oil tank 16, the voice coil motor assemblies 13 and the finger die assemblies 12 are sequentially arranged according to the space height from top to bottom. Preferably, the number of layers on the drawer type mounting bracket 15 where the voice coil motor assembly 13 is located corresponds to the height of the connection end 125 on the finger die pad 121. As shown in fig. 7d, the hydraulic line system mainly includes a hydraulic tank 16, a plurality of solenoid valve blocks 161 (not shown), a reciprocating cylinder 135, a plurality of single-point repeating oil chambers 129, and corresponding connecting lines.

As shown in fig. 6, 6a, 6b, 7 and 7a, the bionic skin 123 made of flexible material (such as oil-resistant rubber) is sealed and fixed (such as adhesion) by the convex connection parts 1211 on four corners of the top end of the finger die seat 121 made of rigid material (such as aluminum alloy), air is discharged through the air outlet 1231 in the fixing process, so that the bionic skin 123 is completely attached to the top of the finger die seat 121, the single-point reproduction oil cavity 129 in the bionic skin 123 corresponds to the convex (not marked in the figure) at the tail end of the runner 127 at the top of the finger die seat 121 one by one, and the outlet end of each runner 127 can be just covered and sealed by the single-point reproduction oil cavity 129, so that the hydraulic oil is prevented from leaking outwards.

Through unified control of the upper computer main board 143, the voice coil motor assembly 13 periodically moves in a large stroke (not the stroke of the repeated pulse condition), so that air in the pipeline moves to the electromagnetic valve group 161 at the highest position of the pipeline under the action of buoyancy, at the moment, the electromagnetic valve group 161 is opened, the air is returned to the hydraulic oil tank 16, meanwhile, hydraulic oil is supplemented to the corresponding pipeline, and finally, periodic automatic oil supplementing and air exhausting are realized. The breather valve 163 is a one-way valve, and is used for realizing exchange of gas and hydraulic oil, maintaining pressure balance in the hydraulic oil tank 16, avoiding overhigh air pressure in the hydraulic oil tank 16, and preventing external impurities from entering. When the level sensor 162 detects that the liquid level is too low, a host signal is sent upward to manually replenish the hydraulic oil in the hydraulic oil tank 16.

As shown in fig. 3-4 and fig. 4a, the voice coil motor assembly 13 comprises a base 139, a permanent magnet 133, a yoke (not shown in the drawings), a coil 131, a cover 137, a linear potentiometer 138 and the like which are common in the prior art. Specifically, a pair of C-shaped concave portions arranged at intervals are arranged on one side of the cover 137 corresponding to the reciprocating hydraulic cylinder 135 as a quick assembly structure 1371, the rear part of the main body of the reciprocating hydraulic cylinder 135 is fixed on the base 139 through bolts during installation, then the valve rod 1351 is clamped into the quick assembly structure 1371 in a state of not contacting the inner wall of the quick assembly structure 1371, and the valve plate 1352 is limited in a quick assembly gap 1372 of the quick assembly structure 1371. Because the valve rod 1351 is not in direct contact with the quick-assembly structure 1371, an action space is reserved for installation or machining errors (form and position tolerance deviation generated by installation or manufacturing precision), so that the valve rod 1351 can only bear axial force when in motion, and further, the problem that the valve rod 1351 bears extra radial force to cause cylinder-stringing and liquid leakage is avoided, and the operation stability of equipment is improved.

The present invention also improves the drive module of the voice coil motor assembly 13. The linear guide 134 ensures form and position tolerances of the components; the drive board 132 reduces remote interference of high-frequency AD analog signals, and the CAN bus is independently calibrated and controlled, so that the maintenance is convenient; the reciprocating cylinder 135 achieves precise, high frequency motion, converting linear motion into vascular pulsations, and the combination of multiple points simulates a complete pulse condition. Specifically, the drive plate 132 integrates position control, current control, and PID control algorithms to achieve more precise motion control. In order to improve the precision, the position control can provide higher precision, and the pulse amplitude output position can be more accurate and repeatable. In addition, the current control can also provide higher control precision, and the influence of factors such as a sensor, system noise and the like on the performance of a control system is eliminated; in order to increase the response speed, a fast, smooth linear transition can be achieved. In addition, the position control and the current control can monitor the mechanical load in real time, and simultaneously, the current is quickly adjusted to maintain the pulse graph curve coincidence rate; in order to improve stability, linear motors can handle load variations better and maintain stability because they can perform full position and current control. Ensuring the movement pressure and amplitude under the pressing of the pulse taking finger force; in order to reduce the energy consumption, linear motors generally have a higher efficiency, since they can handle dynamic loads better. In addition, position and current control may reduce power consumption of the motor during start-up and steady state, reducing motor operating temperature.

As shown in FIG. 5, the finger die assembly 12 is a three bar-shaped structure positioned within the base 128 corresponding to the respective cun, guan and chi acupoints. One perspective structure of the finger die assembly 12 is shown in fig. 6, and the finger die assembly 12 includes a pressure sensor 124 disposed in a base 128, a strip finger die seat 121 disposed on the pressure sensor 124, a plurality of runners 127 communicating from the side wall of the finger die seat 121 to the top, connection ends 125 disposed on the side wall of the finger die seat 121 at the input end of the runners 127, array points 126 disposed at the output end of the runners 127, and bionic skin 123 (such as silica gel) disposed on the top of the finger die seat 121 and matched with the output end of the runners 127, wherein each connection end 125 is fixed with a quick-twist elbow 122. The array points on the top of the finger die holder 121 are arranged in one-to-one correspondence with the electrode points of the flexible array sensor 2142 of the acquisition terminal 2.

Meanwhile, in order to simulate the wrist size as much as possible, the size design of the bionic skin 123 is limited; in order to avoid signal deviation caused by the fact that the finger die holder 121 is too high to press the pressure sensor 124 downwards, the height of the finger die holder 121 is limited; to ensure that the flow channels 127 have a consistent trend within finger die block 121, the flow channels 127 have a consistent tube diameter and do not interfere with each other, and the direction of entry from connecting end 125 is consistent, thereby defining the installation direction of quick twist elbow 122. Because of the minimum size constraints of the quick-twist elbow 122, it should also be ensured that it does not interfere with installation. If the quick-twist elbows 122 are vertically arranged, in order to avoid interference, it is impossible to install all the quick-twist elbows 122 in an upward or downward manner at the same time, so that different impulsive forces are generated on the bionic skin 123 under the same hydraulic conditions, and the recurrence accuracy is reduced. Thus, the liquid inlet end of the quick twist 122 is disposed entirely laterally. In order to meet the uniformity of the load on the pressure sensor 124, the weight of each finger die holder 121 should also be ensured. On the premise that all of the above conditions are satisfied, it may be necessary to deform the structure of the finger die holder 121 to some extent. In another finger die holder configuration, as shown in fig. 6a, the mounting location of the connecting end 125 is protruded (i.e., widened) outward by lowering and widening the finger die holder 121, thereby avoiding interference between the snap bends 122 of the middle and side finger die assemblies 12. The middle finger rest 121 is designed to be reduced in height in order to ensure that the middle finger rest 121 is comparable in weight to the side finger rest 121. Through the modified design, the wide-narrow staggered installation mode can be realized, the quick-screwing elbow 122 is installed in a limited space at high density, the product volume is greatly reduced, and the pipeline arrangement is optimized. The weight center of gravity is adjusted through volume calculation and drilling, so that the weight of the two repeated finger die holders 121 is unified, the influence on the pressure sensor 124 at the lower part of the finger die is avoided, and the accurate feedback of the floating and sinking finger force is ensured.

The array point 126 of the reproduction terminal 1, the voice coil motor assembly 13 and the flexible array sensor 2142 of the acquisition terminal 2 adopt a one-to-one correspondence manner to complete the acquisition and reproduction process of the pulse condition. Specifically, each array spot 126 is driven by a completely separate voice coil motor assembly 13, and each voice coil motor assembly 13 communicates with a relatively separate hydraulic reservoir 16. The illustration in fig. 2 is only a schematic for describing the structure of the reproduction terminal 1 and is not a practical number of components. After the flexible array sensor 2142 of one of the collection finger assemblies 214 acquires the waveform signal as shown in fig. 7c, the waveform signal is transmitted to the gateway board 141 through a wireless communication module (not shown) of a terminal device (such as a smart phone), the data is decoded and then sent to the driving board 132 of the corresponding voice coil motor assembly 13 by the host computer main board 143, then the driving board 132 outputs driving current to the coil 131, the coil 131 drives the cover 137 to act along the linear guide rail 134 under the action of the driving current, and then drives the valve rod 1351 of the reciprocating hydraulic cylinder 135 to act, so that hydraulic oil is pumped into the finger module 12 with different forces when the valve rod 1351 extends or contracts, and further fluctuation and fluctuation are generated on the corresponding array point 126 of the bionic skin 123. The other array points 126, after synchronized action, produce amplitude fluctuations as shown in fig. 7 b. Fig. 7b shows an application schematic of dense array points 126 in an ideal state, in which, in view of cost and overall weight, when the number of array points 126 is small, the array points 126 still conform to the law that the array points 126 of a certain row fluctuate within a range from zero to a maximum amplitude, and the array points 126 of other rows fluctuate within a range from zero to less than the maximum amplitude, and the maximum amplitude decreases from the middle row to the upper and lower sides.

One embodiment of the invention is shown in fig. 8. The acquisition terminal 2 comprises a pulse condition acquisition part 21 and a wearable part 22 which are connected in a sliding and clamping manner. The wearable part 22 is changed into a double-sided tightening structure based on the existing adjustable wrist strap 223, and the pulse taking frame component 225 and the infrared lamp module 224 are respectively arranged at the upper end and the lower end of the adjustable wrist strap 223 in a relative mode, and the infrared lamp module 224 is arranged in the slidable lamp box 221. The radial artery side of the pulse taking frame component 225 is provided with a pulse taking opening 2251, and the pulse taking frame component 225 is in sliding fit with the collecting seat 212 of the pulse condition collecting part 21 through a pair of linear slide ways 2252 arranged along the radial artery direction. The side of the pulse taking frame component 225 near the palm is also provided with an arc-shaped indent 2254 for matching with the root of the palm. In another embodiment of the present invention shown in fig. 9-11, the wearable portion 22 adopts a single-sided tightening structure, and the collecting seat 212 and the pulse taking frame assembly 225 are clamped on the self-locking protrusions 2253 through the self-locking members 216. The self-locking member 216 may be configured in the prior art, such as a commercially available push spring buckle. The self-locking member 216 is switched between the locking and unlocking operating states by the cooperation of a spring and a resilient catch (not shown). So that the pulse condition collecting part 21 can be quickly clamped on the pulse taking frame component 225 or quickly removed from the pulse taking frame component 225.

The pulse condition collecting part 21 comprises a collecting seat 212 serving as a main body supporting framework, a protective shell 211, a main control board 219 serving as a main control part, a (lithium) battery 215, three sets of collecting finger assemblies 214 used for collecting pulse condition signals and arranged side by side, an infrared camera 218 used for being matched with an infrared lamp module 224, a self-locking piece 216 used for being matched with a self-locking protrusion 2253, a rack 217 used for being matched with a gear 2141 of the collecting finger assembly 214, and a pair of limiting modules 213 used for limiting the upper and lower stroke end points of the collecting finger assembly 214 and provided with a pulling piece 2131. The main control board 219 is provided with a main control key 2191, a light guide column 2192 and a USB interface (not labeled in the figure) for charging or transmitting data.

The light guide posts 2192 may be used as status indication lamps, or may be multi-color single indication lamps for displaying the current pressure status, data transmission status, battery power, etc. For example, when the indicator lights blink at high frequencies, it indicates that data is being transmitted; when the low frequency flicker is generated, the error states of acquisition data abnormality, network connection abnormality, bluetooth connection abnormality and the like are indicated; when the red indicator lamp is normally on, the electric quantity is insufficient; when the green indicator lamp is normally on, the state is good. The indication state is reflected in the mobile phone APP at the same time.

The collection finger assembly 214 includes a motor assembly 2145, a gear 2141 disposed at an output end thereof, a motor bracket 2144 for mating with an adjacent collection finger assembly 214, a flexible array sensor 2142 disposed at the bottom, and a silicone finger pattern 2143. Wherein a silicone finger die 2143 overlies the flexible array sensor 2142.

When in use, the pulse condition acquisition part 21 is fixed on the wearable part 22 after the radial artery is positioned by a visual positioning method. (see the following application method for details), racks 217 corresponding to the number of the collection finger assemblies 214 are limited in the collection seat 212, and when the motor assembly 2145 drives the gear 2141 to rotate, the collection finger assemblies 214 can act along the vertical direction of the racks 217 as a whole, so that the bottom silica gel finger model 2143 is abutted against the wrist of the patient, and pulse condition signals with different intensities are fed back to the flexible array sensor 2142 according to the pressing force. When the doctor of the reproduction terminal 1 presses down the bionic skin 123, the pressure sensor 124 at the bottom of the finger die holder 121 receives the pressure signal and sends the pressure signal to the main control board 219 through the gateway board 141, and the main control board 219 drives the motor component 2145 to drive the collection finger component 214 to generate a motion corresponding to the strength of the reproduction terminal 1, so as to simulate the face-to-face pulse feeling situation with high precision.

The pulse taking technique is mainly used for automatically taking pulses of the finger collecting assembly 214 by a step-by-step pressurizing method on the cun-guan-chi pulse position until the maximum amplitude range available in the array sensor is obtained. As shown in fig. 13, the specific pulse taking method includes the following steps:

step S1, positioning and covering radial artery: positioning the radial artery beating range by visual sense or finger touch, placing the collection finger assembly 214 of three cunguan ulna above the radial artery, so that each flexible array sensor 2142 covers the radial artery beating range (4*6 array sensors are staggered in this embodiment as an example);

step S2, obtaining the optimal pulse taking pressure: step S21, generating a pulse potential diagram: the amplitude signal detected by the flexible array sensor 2142 is combined through time domain analysis and frequency domain analysis to generate a two-dimensional (time-pulse width) (fig. 14) and a three-dimensional vector diagram (time-pulse width-frequency) (fig. 15), and the flexible array sensor 2142 measures pulse potential diagram changes of three cun-guan-chi parts at 25-250 g; the three parts of the cunguan ruler are used for taking the pulse according to the gradually increasing pressure value so as to simulate the pulse taking technique which is always pressed by a clinician; step S22, determining the optimal pulse taking pressure: five continuous pulse pressure values with highest pulse potential graph values and no baseline drift are obtained from each part of the cunguan ruler; the method comprises the following steps: gradually increasing the pressure value in a multi-period equivalent incremental mode until a maximum pulse amplitude range is obtained in a certain period; the pressure amplification is 15-35 g, the period time is 3-8 s, and the measurement period is more than two;

Step S3, each flexible array sensor 2142 takes pulse in a synchronous or asynchronous mode under the optimal pulse taking pressure value, the pulse taking time is 20-40S, and the optimal pulse taking diagram takes the highest value of five continuous pulse potential diagrams, no baseline drift and no deviation as the reference; synchronous pulse taking simulation cunguan ruler three parts of total floating, middle sinking pulse taking and total pressing of the three parts under the optimal pulse taking pressure value; asynchronous pulse taking simulation cunguan ruler three parts are totally floating, middle and sinking to take pulses and three parts are pressed singly under the optimal pulse taking pressure value;

step S4, generating an optimal pulse potential diagram: recording the trend change of the pulse width of the whole optimal pulse potential diagram in the step S2 under different pressing pressures to obtain a three-dimensional pulse potential diagram, and recording three optimal pulse potential diagrams of the cunguan ruler in the step S3.

The change curves of pulse amplitudes of different pressure values are recorded, so as to determine pulse positions, pulse forces and pulse potentials in pulse condition elements, three optimal pulse potential diagrams are recorded respectively, and pulse lengths, pulse widths, fluency and tension are analyzed from two-dimensional and three-dimensional angles. In summary, the embodiment includes three parts of the cunguan ruler for synchronously taking pulse, so as to distinguish pulse position, pulse force and pulse potential, and further extract an optimal pulse potential diagram of each part under the optimal pulse taking pressure, so that the rest factors of pulse conditions are analyzed and measured, the acquisition method is closer to clinic, the operation time is shorter, the patient can tolerate the pulse condition, and the pulse taking ruler is simple and portable and is suitable for clinic.

In the prior art, for a single-point sensor, only the fluctuation period (hereinafter referred to as Ti) of a pulse profile curve of a single curve is required to be used as a unified pulse profile period (hereinafter referred to as Tk). As shown in FIG. 17a, the original pulse signals acquired by the multi-point array sensor always obtain two pulse curves of forward and reverse directions because the sensor corresponding to the acquisition point cannot completely cover the artery and has compression on surrounding tissues, so that the base lines of the sensors are different. In addition, the base line of each pulse graph curve has drift phenomenon, so that it is difficult to synchronously output pulse signals, the pulse start points (hereinafter referred to as T0) of each curve have differences, the amplitude of the acquisition point positioned at the edge is smaller, and it is difficult to accurately judge Ti of each curve. Therefore, the pulse cycle and the base line of the pulse graph curve obtained by each acquisition point must be corrected, and finally, the pulse data of each acquisition point is stored and reproduced in the form of amplitude gradient. The forward pulse graph curve and its derivative curve of one of the acquisition points is shown in fig. 17 b. The original signal has baseline drift, the baseline drift of a single sensor is corrected, and Tk is the time difference between two adjacent T0.

The correction method comprises the following steps:

step K1, obtaining an original pulse graph curve of each acquisition point through an array sensor;

step K2, determining each pulse graph curve direction (forward/reverse pulse graph curve): according to the average heartbeat frequency (about 1.0 s-1.5 s/time) in the resting state, the initial Twin is set to be 4.2s (unit: second, three fluctuation periods are taken as a basis, and the acquisition duration is 3.5 x 1.2=4.2);

the derivative (hereinafter, ∂) of each map curve was obtained, and the absolute value of each of the forward peak time (hereinafter, th) and the reverse Th of the derivative of any of the map curves was determined when twin=4.2, and the direction in which the absolute value was large was defined as the map direction. For example, for any pulse graph curve there are n derivative positive Th, m derivative negative Th within Twin: s is S _{Positive direction} =|∑∂ _peak1+ ∂ _peak2+....+ ∂ _peakn |S _{Negative =} |∑∂ _peak1+ ∂ _peak2+....+ ∂ _peakm I (I); when S is _{Positive direction} >S _{Negative pole} The pulse pattern is positive and negative.

Step K3, determining Ti of each pulse graph curve: step K31, obtaining Th of the forward/reverse pulse graph curve: for the forward pulse graph curve, as shown in fig. 17d, correcting the baseline drift of the pulse graph curve (curve 1) of the acquisition point to obtain a derivative curve (curve 2), filtering the derivative curve to obtain a derivative curve (curve under fig. 17 d) shown in fig. 17d, and obtaining Th of the forward pulse graph curve; for the reverse pulse graph, as shown in fig. 17b, a time (hereinafter, td) corresponding to the minimum value of the derivative (hereinafter, ∂ min) is obtained, and Td decreases most rapidly on the pulse graph; looking forward for Th of the reverse pulse graph curve according to Td as a reference point for baseline drift (hereinafter Tb); step K32, obtaining T0 of the forward/reverse pulse graph curve: for the forward pulse graph, as shown in fig. 17e, the time difference of the adjacent T0 is obtained to be Ti of the forward pulse graph; for the reverse pulse graph, as shown in fig. 17b, the time difference of acquiring adjacent Th is Ti of the reverse pulse graph; step K33, removing baseline drift of each forward/reverse pulse graph curve: for the forward pulse graph curve, as shown in 17e, connecting all Th to obtain a base line of the forward pulse graph curve; for the reverse pulse graph, as shown in fig. 17b, connecting all Th to obtain the baseline of the reverse pulse graph, and subtracting the baseline from the original reverse pulse graph to remove the baseline drift of the reverse pulse graph.

Step K4, determining a sliding window duration (or a data capturing duration, hereinafter referred to as Twin): as shown in fig. 17b, taking Th of pulse derivative as pulse characteristic point, the time difference between adjacent Th is about Tk from the starting point of sliding window; for a forward pulse condition, the time difference of adjacent T0 needs to be identified; for the forward pulse plot, then Twin is positive = t+ (Th-T0) +t, and Twin is positive <3T; for reverse pulse, then Twin inverse = t+ (Th-reverse Tm) +t, and Twin inverse <3T; in order to avoid possible unable acquisition, an additional 0.5T safety interval is provided, i.e. twin=3.5t

Step K5, periodically determining the base line of each pulse graph curve, and synchronizing Ti of each forward/reverse pulse graph curve into Tk: and counting the period Ti of each acquisition point pulse graph curve, counting the Ti ratio of each pulse graph curve, and taking the statistical value N of a certain Ti as Tk when the statistical value N is more than N/2 (N is the number of sensors).

In the pulse graph with period not Tk: for the reverse curves, finding any other reverse pulse graph curve with the period Tk, setting Ti of the two reverse pulse graph curves as Tk, and overlapping the wave crest points of the two reverse pulse graph curves; for the forward curve, T0 is set to be the same as the forward curve of period Tk. The pulse graph after baseline drift removal is shown in fig. 17 f;

Step K6, outputting a pulse graph curve after baseline correction: in order to ensure that all curves can be processed correctly, a window starting point Ts-pulse condition period Tk-1 (unit: second) is taken as a real-time output period, the output interval is 1 time/s, and finally, a pulse graph curve set of each corrected acquisition point shown in fig. 17f is obtained.

In step S1, the pulse condition collection unit 21 of the collection terminal 2 is connected to a display terminal, such as a smart phone, a tablet, or a display module integrated on the pulse condition collection unit 21, through bluetooth. For cost reduction, the pulse condition acquisition part 21 is not integrated with a display module in the embodiment shown in the drawings of the invention. Wearing the wearable part 22 on a wrist, placing the pulse taking opening 2251 close to the radial artery, placing the cun guan ulna pulse in the pulse taking opening 2251 with the palm facing upwards, placing the infrared lamp module 224 at the bottom of the wrist, and turning on the infrared lamp module 224 by using the mobile phone APP;

step S2, the other hand holds the adjusting collecting seat 22, irradiates the pulse taking opening 2251 with the bottom infrared camera 218, and observes the development position of the radial artery in real time through the display terminal;

step S3, according to the developed image on the display terminal, the arc-shaped indent 2254 of the pulse taking frame assembly 225 rotates along the palm root portion under the condition of being close to the palm, so as to adjust the position of the pulse taking port 2251, and the mobile phone end or the collection seat 22 prompts that the pulse taking port 2251 is located at the correct position in a voice, indicator light or message notification mode, and the pulse taking port 2251 is clamped on the pulse taking frame assembly 225 through the self-locking cooperation of the self-locking protrusion 2253 and the self-locking piece 216, or the pulse taking port 21 is clamped on the pulse taking frame assembly 225 in a sliding manner through the linear slide 2252, so that pulse taking can be started.

The radial artery detection method in the step S2 comprises the following steps:

step P1, obtaining an original image: the infrared camera 218 irradiates wrist skin in the vein access 2251 to acquire an infrared original image graph 12 in real time (marks are only used as marks and illustrations, and corresponding mark contents do not exist during image acquisition);

step P2, correcting the image: step P21, positioning the radial side and the ulnar side: the specific positioning points (two circular points in this embodiment) for correcting the original image orientation are preset on the ruler side of the pulse taking frame component 225, so that the positioning accuracy and the positioning speed are improved. Other shapes can be adopted, or the number of positioning points can be increased or decreased, so long as the manner of rapidly positioning the position of the pulse taking port 2251 can be realized. Locating the specific locating points through an edge detection algorithm such as a Canny algorithm and a Hough transformation algorithm; step P22, positioning the pulse taking port 2251: continuously utilizing an edge detection algorithm to identify and obtain a rectangular area; step P23, rotation correction: and (3) obtaining a calibrated horizontal line according to the rectangular position and the specific locating point, determining the rotation angle of the image, and cutting the rectangular region obtained in the step (P22) after rotation correction. For example, setting conditions in the program, rotating the image until the center line of the specific anchor point is rotated to the horizontal, and calculating the position of the vein fetch 2251 according to the length of the anchor point line;

Step P3, sharpening the boundary of the rectangular region ROI: performing edge sharpening twice by using a Laplace filter to respectively obtain an upper part and a lower part of fig. 12b (upper part) and 12b (lower part), wherein a Laplace operator (formula 1) adopts a four-adjacent-domain convolution kernel (formula 2);

(1)/(1)>(2)

Step P4, marking arterial blood vessels: step P41, binarization: after (lower) binarization processing of fig. 12b, fig. 12c is obtained; step P42, bone extraction: extracting bones from the graph 12c by using a Zhang-Suen' sAlgorithm algorithm to obtain a graph 12D, and obtaining a plurality of connected areas A-D; step P43, locating the longest path: finding the longest path L (i) for each connected region in fig. 12d using DFS depth-first algorithm; step P44, locating radial artery: comparing Li with La according to a preset length threshold La <2.73cm, pruning the area with the longest path smaller than the threshold in the graph 12d (marked area in the graph 12 d), and judging the area A in the graph 12d to be radial artery. Normally, the length of the radial artery is 2.73 cm-4.10 cm, and although the length of the radial artery is different according to the conditions of different people, for insurance, a length threshold la=20mm is taken to judge whether the radial artery is a radial artery vessel;

step P5, marking radial artery: the path of fig. 12d is used to generate a plurality of tiny rectangles, the number of which is n, and the tiny rectangles take three adjacent pixels of the current pixel as shown in fig. 12 f. Inputting each rectangle to the Grabcut algorithm on fig. 12d to obtain a selected region S (i) to obtain S (1) us (2) us @ us (n) =s to mark S to obtain fig. 12g;

Step P6, image restoration: fig. 12g is restored to fig. 12 and output to the display terminal.

As shown in fig. 16, the pulse condition acquisition and reproduction system includes a reproduction terminal 1, an acquisition terminal 2, and a server terminal, wherein a data transmission channel can be directly established between the acquisition terminal 2 and the reproduction terminal 1 through a C/S architecture, the acquisition terminal 2 (patient end) is used as a server end, pulse condition data is sent to the reproduction terminal 1 (client end) as a client end in real time, and the reproduction terminal 1 reproduces the pulse condition data in a one-to-one correspondence manner with the flexible array sensor 2142 of the acquisition terminal 2 through a plurality of reproduction points capable of converting waveform signals into fluctuation, so as to complete pulse condition reproduction.

In order to establish a data set of the intelligent pulse condition acquisition system, data samples of pulse conditions of different ages need to be acquired. At this time, a data transmission channel is established by taking the server terminal as a relay, and repeated pulse condition data with specific duration is intercepted and uploaded to the server in real time. The pulse condition signals acquired by the acquisition terminal 2 are acquired by the flexible array sensor 2142, and the data acquired by each acquisition point are two-dimensional data formed by time-pulse amplitude or three-dimensional data formed by time-pulse amplitude-acquisition frequency. The acquisition terminal 2 side sends pulse condition data to intelligent equipment through a Bluetooth module, and then the intelligent equipment such as a smart phone or a tablet is sent to a server terminal. The pulse condition data and/or the diagnosis result of the server terminal are downloaded to the reproduction terminal 1. As shown in fig. 15, on the intelligent device of the reproduction terminal 1, two-dimensional data formed by time-pulse amplitude or three-dimensional data formed by time-pulse amplitude-acquisition frequency is restored to a more visual graphic result, thereby realizing real-time signal acquisition of pulse condition or directly acquiring pulse diagnosis result without diagnosis of doctor.

Specifically, the server terminal includes: the communication module is used for receiving the pulse condition data uploaded by the acquisition terminal 2 or/and sending a diagnosis result to the reproduction terminal 1; the database module is used for storing two-dimensional or three-dimensional pulse condition data uploaded by the acquisition terminal 2; the deep learning model module is used for loading a pulse condition diagnosis model and returning pulse condition data uploaded by the acquisition terminal 2 to a diagnosis result; the pulse condition data labeling module is used for monitoring and training the pulse diagnosis model; and the model training module is used for processing the marked data into a pulse condition diagnosis model.

Reference numerals illustrate: 1 a reproduction terminal, 11 a panel assembly, 111 a camera, 112 a touch screen, 113 a microphone, 114 a power button, 115 a loudspeaker, 12 a die assembly, 121 a die holder, 1211 a connecting part, 122 a quick-screwing elbow, 123 a bionic skin, 1231 an exhaust port, 124 a pressure sensor, 125 a connecting end, 126 an array point, 127 a runner, 128 a base, 129 a single-point reproduction oil cavity, 13 a voice coil motor assembly, 131 a coil, 132 a driving plate, 133 a permanent magnet, 134 a linear guide rail, 135 a reciprocating hydraulic cylinder, 1351 a valve rod, 1352 a valve plate, 136 a three-way quick-screwing joint, 137 a cover, 1371 a quick-mounting structure, 1372 a quick-mounting gap, 138 a linear potentiometer, 139 a base, 14 an electric control assembly, 141 a gateway plate, 142 an AC-DC power supply plate, 143 an upper computer main board, 15 a drawer type mounting bracket, 16 a hydraulic oil tank, 161 electromagnetic valve set, 162 a liquid level sensor and 163 a breathing valve; 2 acquisition terminals, 21 pulse condition acquisition parts, 211 protection shells, 212 acquisition seats, 213 limit modules, 2131 poking sheets, 214 acquisition finger assemblies, 2141 gears, 2142 flexible array sensors, 2143 silica gel finger molds, 2144 motor brackets, 2145 motor assemblies, 215 batteries, 216 self-locking parts, 217 racks, 218 infrared cameras, 219 main control boards, 2191 main control keys, 2192 light guide columns, 22 wearable parts, 221 slidable lamp boxes, 223 adjustable wrist bands, 224 infrared lamp modules, 225 pulse taking frame assemblies, 2251 pulse taking ports, 2252 linear slide ways, 2253 self-locking bulges and 2254 arc concave parts.

Claims

1. A pulse condition acquisition method of a multipoint pulse condition sensor is characterized by comprising the following steps:

step K2, determining the direction of each pulse graph curve;

step K3, determining the fluctuation period Ti of each pulse graph curve;

step K4, determining the duration Twin of the sliding window;

step K5, periodically determining the base line of each pulse graph curve, and synchronizing Ti of each forward/reverse pulse graph curve into a unified pulse condition period Tk;

step K6, outputting a pulse graph curve after baseline correction: taking a window starting point Ts-pulse condition period Tk-1 as a real-time output period, wherein the output interval is 1 time/s, and finally obtaining a corrected pulse graph curve set of each acquisition point.

2. The pulse condition acquisition method of a multipoint pulse condition sensor according to claim 1, wherein the step K2 comprises the steps of: setting initial Twin to 3.0-6.0 s according to the average heartbeat frequency in the resting state; the derivative of each pulse graph curve is calculated, and for all forward Th and reverse Th of the derivative of any pulse graph curve, the absolute value of the pulse graph curve is judged when Tain=4.2, and the direction with the large absolute value is taken as the pulse graph direction.

3. The pulse condition acquisition method of a multipoint pulse condition sensor according to claim 1 or 2, wherein the step K3 comprises the steps of:

Step K31, obtaining the peak moment of the forward/reverse pulse graph curve:

correcting the baseline drift of the pulse graph curve of the acquisition point to obtain a derivative curve for the forward pulse graph curve, filtering the derivative curve, and obtaining Th of the forward pulse graph curve;

for the reverse pulse graph curve, acquiring a time Td corresponding to the minimum derivative value ∂ min, and searching the peak time Th of the reverse pulse graph curve forward according to the time Td of the minimum derivative value as a reference point Tb of baseline drift;

step K32, obtaining the pulse condition starting point T0 of the forward/reverse pulse chart curve:

for the forward pulse graph curve, obtaining the time difference of the starting point T0 of the adjacent pulse condition to obtain the fluctuation period Ti of the forward pulse graph curve;

for the reverse pulse graph curve, obtaining the time difference of adjacent peak time Th to obtain the fluctuation period Ti of the reverse pulse graph curve;

step K33, removing baseline drift of each forward/reverse pulse graph curve:

for the forward pulse graph curve, connecting all peak moments Th to obtain a base line of the forward pulse graph curve;

and (3) connecting all peak time Th for the reverse pulse graph curve to obtain a base line of the reverse pulse graph curve, and subtracting the base line from the original reverse pulse graph curve to remove the base line drift of the reverse pulse graph.

4. A pulse condition acquisition method of a multipoint pulse condition sensor according to claim 3, wherein step K4 comprises the steps of:

taking the peak time Th of the pulse condition derivative as a pulse condition characteristic point, and for the forward pulse condition, identifying the time difference of the adjacent pulse condition starting point T0;

for the forward pulse graph, the sliding window duration Twin is positive=t+ (Th-T0) +t, and the sliding window duration Twin is positive < 3T;

for the reverse pulse chart, the sliding window duration Twin is inverse=t+ (Th-reverse Tm) +t, and the sliding window duration Twin is inverse < 3T;

in order to avoid the situation that the acquisition cannot be performed, an insurance interval of 0.5T, namely twin=3.5t, is additionally arranged.

5. The pulse condition acquisition method of a multipoint pulse condition sensor according to claim 4, wherein the step K5 comprises the steps of:

counting the pulse condition period Ti of each acquisition point pulse graph curve, counting the pulse condition period Ti duty ratio of each pulse graph curve, and taking the pulse condition period Ti as a unified pulse condition period Tk when the counted value of the pulse condition period Ti of a certain curve exceeds 50% of the number of the acquisition points;

in the pulse graph curve with the period not being the unified pulse condition period Tk:

for the reverse curves, finding any other reverse pulse graph curve with the period Tk, setting Ti of the two reverse pulse graph curves as Tk, and overlapping the wave crest points of the two reverse pulse graph curves;

For the forward curve, setting the pulse condition starting point T0 to be the same as the forward curve with the pulse condition period Tk, and removing the baseline drift.