CN113261954B - Multi-element sensing and high-data-precision noninvasive blood glucose detector and detection method - Google Patents

Abstract

The invention relates to a noninvasive blood glucose detection method, in particular to a noninvasive blood glucose detection method with multivariate perception and high data accuracy, which can perform limit, classification recognition, detection position determination and fingertip extrusion degree judgment on a finger before performing noninvasive blood glucose detection and perform the most stable spectrum detection on the finger under the condition of reducing all interference factors to the minimum.

Description

Multi-element sensing and high-data-precision noninvasive blood glucose detector and detection method

Technical Field

The invention relates to improvement of a noninvasive blood glucose detection technology, in particular to a noninvasive blood glucose detector with multivariate perception and high data accuracy and a detection method.

Background

Blood sugar detection is a key link in diabetes treatment, but blood is required to be taken in traditional detection, and the blood sugar is difficult to be monitored daily by a patient due to the wound of blood taking. This has also been a problem that has plagued the medical community for many years.

The near infrared spectrum blood sugar detection technology is expected to realize noninvasive detection all the time, but the reasons for preventing the noninvasive detection by the near infrared spectrum blood sugar detection technology are mainly as follows:

1 weak signal

More than 90 percent of the human blood is water, the proportion of the blood is only 7 to 8 percent, and the content of blood sugar in the blood is very low; moreover, water absorbs near infrared light seriously, which causes serious interference to noninvasive detection; in addition, the core of the near infrared spectral band is the frequency doubling and combined frequency absorption of molecules, the absorption peaks are wide and are overlapped seriously, and the absorbance magnitude has magnitude difference compared with the fundamental frequency of the mid infrared. In view of the above, if the change information of the blood glucose component with a weak content is to be detected accurately, a higher requirement is put forward on the performance of the spectrum acquisition system.

2 background interference

Human tissues such as skin, muscle, bone and the like all belong to strong near-infrared absorbers, human spectra carry a large amount of interference information related to the tissues, and effective information which can be used for analysis is easily submerged in a strong background. Therefore, the tissue background interference problem is one of the important reasons for influencing the accuracy of non-invasive blood glucose detection.

3 difference of individuals

Tissue characteristics such as blood, skin and muscle are greatly different among different individuals, even tissue background components of different parts of the same individual are different, so that the acquired spectrum background noise is complicated, and the difficulty of extracting blood component information from a human spectrum is further increased.

4 change in blood flow volume

The human body belongs to a complex living body, the physiological phenomena of heart pulsation, blood circulation and the like can cause the periodic fluctuation of blood flow volume, the time-varying characteristic of the blood flow volume can cause the change of absorbance in the near infrared spectrum of the human body, and obviously influences the measurement result, which is mainly represented as the instability of the spectrum time domain.

5 the wavelength range of the photoelectric sensor is too wide

The wavelength receiving range of the photoelectric sensor which can be used for detecting the blood sugar is too wide, the spectral information of the LED with the specific wavelength cannot be accurately received, the wavelength receiving range needs to be cut off through the optical filter, so that the wavelength receiving range is narrowed, and the accuracy is influenced due to insufficient light passing rate after the optical filter is used.

6LED power is too low, light transmission is insufficient

The power of LEDs with specific wavelengths which can be used for detecting blood sugar in the market is too small and is concentrated at 1-3mw, so that the passing rate of the light irradiating the fingers is insufficient.

Just like the above six reasons, signal is weak, background interference, individual difference, blood flow volume change, photoelectric sensor wavelength range is too wide, LED power is too small, and light transmission is not enough, so how to extract effective information from strong background spectrum is the key problem that need to be solved for near infrared spectrum noninvasive blood sugar detection.

The current Chinese patent CN108593593A discloses a serial double infrared spectrum noninvasive blood sugar measuring device, which comprises a broadband infrared light source, a measuring hole, a double-filter switcher, an infrared photoelectric sensor, a signal acquisition and processing circuit and the like; the emission spectrum energy of the broadband infrared light source is distributed in the near infrared (800 nm-1100 nm) or short wave infrared (1000-1800 nm) spectrum range; the broadband infrared light source can adopt a halogen lamp light source with stable emission power, and can also adopt other types of infrared light sources, such as a thermal emission infrared light source; the double-filter switcher is formed by combining two band-pass infrared filters with different central wavelengths, the central wavelengths of the two filters are designed according to the spectral absorption characteristics of blood sugar, can be distributed in the near infrared (800 nm-1100 nm) or short wave infrared (1000-1800 nm) range, and has a quick switching function; only one filter light beam is allowed to pass through each time of switching; the infrared photoelectric sensor has the spectral sensitivity of near infrared (800 nm-1100 nm) or short wave infrared (1000-1800 nm); the infrared photoelectric sensor is positioned behind the double-optical-filter switcher and can convert the spectral energy information into corresponding voltage signals.

The technical scheme is that a broadband infrared light source is used for distributing spectral energy in a near infrared (800 nm-1100 nm) or short wave infrared (1000-1800 nm) spectral range, and an infrared photoelectric sensor has the spectral sensitivity of the near infrared (800 nm-1100 nm) or short wave infrared (1000-1800 nm); the infrared photoelectric sensor is located behind the double-optical-filter switcher, and after the invalid wavelength is filtered, the wavelength carrying the blood sugar information is analyzed in a mode of receiving the effective wavelength, although the effective wavelength carries the blood sugar information, the wavelength of the specific effective blood sugar information is concentrated where and still can not be determined, conversion is usually carried out in a mode of adopting an algorithm, and the specific blood sugar value is judged by combining big data.

At present, the scheme is a mainstream scheme for detecting the blood sugar value, and has the problems that the adopted wavelength range is wide, the invalid information mixed in the wavelength is more, the algorithm and the big data are excessively depended on, the algorithm and the big data need to be input with the valid information, and when the invalid information is more, the accuracy is still low even if the algorithm and the big data are adopted for obtaining the result.

Some technical developers have also found the problem, so in the blood sugar detection by means of spectral measurement, designers also concentrate on narrowing the wavelength of the light source to realize that the detected wavelength is an effective wavelength, and the range of the wavelength is narrow, so as to filter out invalid optical information as much as possible.

For another example, the non-invasive blood glucose detecting method based on the spectrum sensor, which is applied by the applicant in chinese patent CN112022167a, includes the following steps: the method comprises the following steps: a spectrum sensor is designed at the fingertip position, and an LED is designed at the other side relative to the fingertip position; step two: a tunable filter of a Fabry-Perot interferometer is adapted in the spectrum sensor, and the optical receiving range of the tunable filter is adjusted to reach the nm level; step three: light rays emitted by a 1650nmLED penetrate through human tissues and are collected by a spectral sensor with the wavelength of 1350nm-1650 nm; step four: the light emitted by 1720nm LED is collected by 1550nm-1850nm spectrum sensor after passing through human tissue.

The design principle is that a light source with customized wavelength penetrates through the designated position of the end part of the finger, the wavelength after attenuation is detected after the wavelength passes through the end part of the finger by the spectrum sensor, the spectrum sensor adopts a mode of scanning near infrared light in a specific range, the scanning range of the spectrum sensor can be effectively realized by controlling the input specific voltage of the spectrum sensor, after the spectrum sensor scans the near infrared light, a matrix type light wavelength set can be obtained, effective values are selected for the wavelength set, the effective values and the original fixed value wavelength are converted, and therefore effective blood sugar values can be obtained.

The blood sugar value measured by the proposal in the actual use process is basically the same as the detection result of the invasive blood sugar value under the normal condition, but a problem is also found in the actual use process,

the main reason for finding that the measured indication value has a certain deviation when the same finger is sent into the noninvasive glucometer for detection is that the spectral detection mode of the noninvasive glucometer mainly aims at 2-5 mm of the fingertip part, the part has no skeleton blockage, so that when the spectrum sensor is used for collection, the wavelength containing effective blood sugar information can be collected as far as possible, but the part belongs to the forefront of the fingertip and has the influence of the nail length, so that when a user of the noninvasive glucometer measures, the part has some extruded states and some relaxed states, but the finger density can be changed after extrusion, and the finger force detected into the noninvasive glucometer is different, the extruded states are different, even when some finger parts are detected, the finger parts are not in an ideal light source, and the error is caused by one of the factors of the noninvasive glucometer for detection.

In summary, it can be seen that how to further reduce the error of the spectrum technique for implementing the non-invasive blood glucose detection is a technical problem to be solved.

Disclosure of Invention

The technical problem to be solved by the invention is to overcome the existing defects, and provide a noninvasive blood glucose detector with multivariate perception and high data accuracy and a detection method, which can perform the most stable spectrum detection on fingers under the condition of limiting, classifying and identifying, determining the detection position and judging the fingertip squeezing degree before performing noninvasive blood glucose detection on the fingers and reducing all interference factors to the lowest.

In order to solve the technical problems, the invention provides the following technical scheme: the utility model provides a many first perceptions and accurate noninvasive blood glucose detector of height data, includes casing, controller, power module, fin, switch, LED light source and spectral sensor, be equipped with the finger storehouse in the casing, power module is connected and all establishes in the casing with the controller electricity, the LED light source is established at the tip in finger storehouse and is located finger storehouse top surface, the fin is connected with spectral sensor, spectral sensor is located the tip in finger storehouse and is located the bottom surface in finger storehouse, LED light source and spectral sensor are located the same straight line, the tip in finger storehouse is equipped with the notch, the edge of notch is equipped with domatic dog, be equipped with a plurality of temperature sensor side by side on the domatic dog, temperature sensor is connected with the controller electricity, be equipped with the through-hole that is used for detecting the light to pass through on the domatic dog, still include fingertip pressure detection mechanism, fingertip pressure detection mechanism includes detection support, fingertip dog and pressure sensor, the detection support is established on the casing, the fingertip dog passes through pressure sensor and detects the leg joint; the pressure sensor is positioned at the notch and at the edge of the slope surface check block and is also electrically connected with the controller.

Preferably, the detection bracket is detachably connected with the shell, and the detection bracket is abutted to the controller.

Preferably, the pressure sensor also comprises a guide spring, the guide spring is nested on the pressure sensor, and the guide spring and the pressure sensor are inclined.

Preferably, a guide rail groove is formed in the side wall of the finger bin, a guide rail is arranged on the fingertip check block, and the fingertip check block is in sliding fit with the guide rail groove through the guide rail.

Preferably, a fingertip limiting groove with the width of 5 mm to 1 cm is arranged on the fingertip stop block.

Preferably, the top of the fingertip stop is arc-shaped, and the height of the notch is greater than that of the fingertip stop.

Preferably, the spectrum sensor further comprises a semiconductor refrigeration piece and a heat dissipation copper piece, the spectrum sensor is further connected with the heat dissipation copper piece, the heat dissipation piece is connected with the heat dissipation copper piece through the semiconductor refrigeration piece, the refrigeration end of the semiconductor refrigeration piece is attached to the heat dissipation copper piece, and the heating end of the semiconductor refrigeration piece is attached to the heat dissipation piece.

In order to solve the technical problems, the invention provides the following technical scheme: a noninvasive blood glucose detection method with multivariate perception and high data accuracy comprises the following steps:

turning on a power switch, and respectively electrifying an LED light source, a spectrum sensor, a temperature sensor and a pressure sensor by a power module through a controller;

inserting a finger into the finger bin and extending into the bottom of the finger bin, wherein the nail part of the fingertip part is positioned in a gap between the fingertip stop block and the notch, the front end of the fingertip is limited by the fingertip limiting groove and enables the fingertip to be horizontal, and the guide spring provides a damping effect and limits the forward displacement of the fingertip;

dividing the fingers into three categories, wherein the first category is the thumb, the second category is the index finger, the middle finger and the ring finger, and the third category is the little finger;

step four, when the pressure sensor receives the pressure signal and keeps stable, the fingertip is determined to reach the bottom of the finger bin and keep horizontal, and the pressure value received by the pressure sensor is a coefficient K; when the temperature received by the temperature sensor changes, the fingertip temperature T1 is judged, the unchanged temperature is the environment temperature T2, and the number N of the changed temperature sensors is determined;

after the LED light source emits light with specified wavelength, the light passes through the through hole and passes through a detection part at the front end of the fingertip, a tunable filter of a Fabry-Perot interferometer is adapted in the spectrum sensor, and the optical receiving range of the tunable filter is adjusted to reach the nm level for spectrum receiving;

and step six, judging whether the type of the finger belongs to the first type, the second type or the third type according to the changed number N of the temperature sensors, combining the coefficient K, the fingertip temperature T1 and the environment temperature T2, sending an ADC (analog to digital converter) value converted after the optical signal is received by the spectrum sensor to the controller together, and calling an algorithm in a built-in data storage module through the controller to calculate the blood sugar value.

Preferably, the LED light source emits light with specified wavelengths of 1500nm, 1525nm, 1550nm and 1575nm.

Preferably, when the power switch is turned on, the power module also powers on the semiconductor refrigeration piece through the controller, and the semiconductor refrigeration piece directly cools the spectrum sensor through the heat dissipation copper sheet, so that the spectrum sensor performs light sensing reception at a specified temperature.

The invention has the beneficial effects that: compared with the prior art, the design of the slope stop block can not only block the end part of the finger to a certain extent and remind the finger to reach the designated position in a touch manner, but also can detect the temperature of the environment and the finger, and can judge the width of the finger tip according to the number of the changed temperature sensors so as to judge the type of the finger inserted;

the end part of the finger bin is provided with a notch, the position of the notch is provided with a fingertip pressure detection mechanism, a guide spring of the fingertip pressure detection mechanism supports a fingertip stop block and keeps inclining, the guide rail is matched with the guide rail groove to limit the movement, so that the fingertip can touch the fingertip stop block to play a damping effect, and meanwhile, the pressure sensor can further detect the extrusion variable of the end part of the finger;

the notch height of pointing the storehouse tip is higher than the fingertip stop dog height, and the top of fingertip stop dog can be the arc, and the clearance that the centre formed can supply the nail to pass, can not be because of the light leak or the excessive extrusion problem of tissue that the nail warp and lead to when guaranteeing to detect.

For sponge foaming material in the soft material layer that the fingertip stop dog adopted, outer lining soft surface fabric is equipped with the fingertip spacing groove simultaneously and can prevent the fingertip skew to fingertip levelness fine setting, avoids the fingertip to excessively extrude, if the excessive extrusion appears, then pressure sensor's variable can intervene and regard as the reference volume that blood glucose detected.

The cooperation semiconductor refrigeration piece is to the cooling of heat dissipation copper sheet to guarantee that spectral sensor's constancy of temperature can not lead to detecting data drift because of temperature variable, dispel the heat fast through fin and stainless steel heat dissipation etching net, and combine radiator fan to dispel the heat in to the casing, can guarantee that the instrument also can guarantee the temperature stability of detection ring border when using the longer time.

Drawings

FIG. 1 is a blood vessel distribution map of a palm;

FIG. 2 is a schematic view of a fingertip portion of a finger;

FIG. 3 is a schematic view of the detector of the present invention;

FIG. 4 is a first internal schematic view of the detector of the present invention;

FIG. 5 is an enlarged partial view of the first embodiment of the present invention;

FIG. 6 is a second internal schematic view of the detector of the present invention;

FIG. 7 is an enlarged partial view of the second embodiment of the present invention;

FIG. 8 is an exploded view of the detector of the present invention;

fig. 9 is a block diagram of the controller of the present invention.

Reference numerals: 1. a housing; 2. a controller; 3. a power supply module; 4. a heat sink; 5. a power switch; 6. an LED light source; 7. a spectral sensor; 8. a finger compartment; 9. a notch; 10. a slope surface stop block; 11. a temperature sensor; 12. a through hole; 13. a fingertip pressure detection mechanism; 14. detecting the bracket; 15. a fingertip stop block; 16. a pressure sensor; 17. a guide spring; 18. a guide rail groove; 19. a guide rail; 20. a fingertip limiting groove; 21. an indicator light; 22. heat dissipation holes; 23. an upper cover; 24. a retaining ring; 25. fixing a bracket; 26. a fixing plate; 27. stainless steel heat dissipation etching net; 28. a rubber seal ring; 29. a heat radiation fan; 30. a radiator fan fixing support; 31. a heat dissipation copper sheet; 32. semiconductor refrigeration piece.

Detailed Description

The preferred embodiments of the present invention will be described in conjunction with the accompanying drawings, and it will be understood that they are described herein for the purpose of illustration and explanation and not limitation.

As shown in fig. 1, blood vessels of the palm are distributed, and it can be known from observation of the graph that finger tips are distributed with rich capillary vessel networks, which can effectively reflect blood sugar content in the human body, and the position has obvious blood volume change characteristics; the muscle and skeletal tissue of the finger is relatively thin, so background interference information has relatively little influence; in addition, the front end of the finger is convenient to measure, and the examinee has no psychological burden, so that a stable high signal-to-noise ratio spectrum signal can be obtained; in the prior art, the size and the shape of a finger chamber of all noninvasive glucose meters are fixed, the mounting positions of a spectrum sensor and an LED light source in the finger chamber are fixed, although the specification of the equipment can explain in advance which finger needs to be used for detection, generally an index finger and a middle finger, as shown in fig. 2, after light penetrates through a detection part of a fingertip, the light is collected by the spectrum sensor, if the finger size, the direction and the extrusion degree are as shown in fig. 2 in each detection, the detection part is determined and kept uniform, which is the most ideal state, but the use habits of each person in the actual use process are different and are easy to ignore, meanwhile, the noninvasive glucose meter of each person is not necessarily kept horizontal in the use process, the finger does not necessarily keep horizontal when inserted into the bottom end of the finger chamber, and the finger tip, the nail length and the insertion force of the end part of the finger cannot be determined, so that more variables exist before measurement, and the variables are all the influence of noninvasive glucose detection.

The applicant found in the development that the apparatus of the non-invasive glucometer has the following three main problems in use:

firstly, after the finger is inserted into the finger bin, the finger tip shapes and widths of the index finger, the middle finger and the ring finger are similar, but the widths and lengths of the little finger and the thumb are greatly different, after the finger is inserted into the finger bin, the extrusion degree of the detection part is influenced because the extrusion degree of the detection part influences the wavelength change and the light intensity attenuation of the LED light source after penetrating through the detection part, the finger tip shapes and the widths of the index finger, the middle finger and the ring finger are different;

secondly, the finger tip nails of each finger are different in length according to personal habits, when the fingers are inserted into the finger chamber, the finger tip nails abut against the front end of the finger chamber to form deformation to cause the light leakage phenomenon, and the detection part moves backwards or forwards, so that the ideal position of the detection part is deviated from the position of the test light source, and the deformation of the nails can cause the change of the tissue density of the actual detection part;

thirdly, when each finger is inserted into the finger bin, the fingertip is not always kept horizontal, the fingertip part can incline to a certain degree in the insertion process, and the inclined test part and the tissue density can deviate from the preset value;

the technology of the noninvasive blood glucose meter has certain resistance in the popularization and promotion processes, and part of the reason is that equipment for researching and developing noninvasive blood glucose based on near infrared spectroscopy in the market is designed based on a specific wavelength range or a specific wavelength LED and a traditional photoelectric sensor, and because the receiving wavelength range of the photoelectric sensor is wide, if optical data with high precision is to be obtained, the wavelength needs to be cut off by an optical filter, but the optical signal precision is influenced by the processing precision of the optical filter, so that the optical signal precision is insufficient.

On the other hand, due to the fact that individuals have differences, users have different habits and the structural design is solidified, although calibration is needed before use, the calibration is adapted to the characteristics of the individuals, and the problems can exist during each independent detection.

The invention relates to a noninvasive blood glucose detector with multivariate perception and high data accuracy, which is specifically assembled as shown in figure 8, wherein a power switch 5 is arranged on an upper cover 23, the power switch 5 is electrically connected with a controller 2, the upper cover 23 is detachably connected with a shell 1 through a retaining ring 24, a heat dissipation hole 22 is arranged on the side wall of the shell 1, a heat dissipation fan 29 is arranged on the side wall of the shell 1 through a heat dissipation fan fixing support 30 and is positioned beside the heat dissipation hole 22, a power module 3 is arranged with the controller 2 through a fixing plate 26, the power module 3 is positioned at the bottom of the fixing plate 26, the controller 2 is positioned above the fixing plate 26, and after the power module 3, the fixing plate 26 and the controller 2 are assembled into a control module, the control module is arranged on the shell 1 through a fixing support 25 and has adjustable height; the indicator light 21 is installed on the controller 2 and exposed on the surface of the shell 1 to indicate whether the power switch 5 is opened or not, the spectrum sensor 7 is electrically connected with the controller 2 after being clamped with the heat dissipation copper sheet 31, the finger bin 8 is clamped with the shell 1 through the rubber sealing ring 28, the semiconductor refrigeration sheet 32 is attached to the heat dissipation copper sheet 31 through one surface of the heat conduction silica gel, the other surface of the semiconductor refrigeration sheet is attached to the heat dissipation sheet 4, the shell 1 is further provided with the stainless steel heat dissipation etching net 27, the position of the stainless steel heat dissipation etching net 27 is attached to the heat dissipation sheet 4, the stainless steel heat dissipation etching net 27 is light and thin in thickness of 0.2-0.3 mm and convenient to install, and is more attractive compared with a plastic heat dissipation net; meanwhile, compared with a plastic material, the metal material has better heat conductivity and can improve the heat dissipation efficiency; due to temperature deviation, for the spectrum sensor, a detection value can be caused to drift, so that the heat dissipation of the spectrum sensor is extremely severe, after the power switch 5 is turned on, the semiconductor chilling plate 32 is connected with electricity, the semiconductor chilling plate 32 can be used for cooling while heating after being electrified, the heating surface is attached to the radiating fin 4, the cooling surface is attached to the radiating copper sheet 31, the radiating copper sheet 31 is in contact with the spectrum sensor 7 to realize heat conduction, so that the environment temperature is controllable when the spectrum sensor 7 detects, air flow is formed in the shell 1 through the radiating fan 29, the stainless steel radiating etching net 27 is combined to carry out rapid heat dissipation, and the situation that the detection indication value of the spectrum sensor 7 drifts due to the environment temperature can be avoided under the condition that the working time of the equipment is long.

The noninvasive blood glucose detecting instrument with multi-element sensing and accurate height data has the internal structure shown in figures 4 to 7, an LED light source 6 is arranged on a shell 1 and is electrically connected with a controller 2, the LED light source 6 is arranged at the top surface position of the end part of a finger bin 8, the LED light source 6 and a spectrum sensor 7 are positioned on the same straight line, the end part of the finger bin 8 is provided with a notch 9, the edge of the notch 9 is provided with a slope stop block 10, the slope stop block 10 is provided with a plurality of temperature sensors 11 side by side, the temperature sensors 11 are electrically connected with the controller 2, the slope stop block 10 is provided with a through hole 12 for detecting the passing of light, the noninvasive blood glucose detecting instrument also comprises a fingertip pressure detecting mechanism 13, the fingertip pressure detecting mechanism 13 comprises a detecting bracket 14, a fingertip stop block 15 and a pressure sensor 16, the detecting bracket 14 is arranged on the shell 1, and the fingertip stop block 15 is connected with the detecting bracket 14 through the pressure sensor 16; the pressure sensor 16 is positioned at the notch 9 and positioned at the edge of the slope stopper 10, and the pressure sensor 16 is also electrically connected with the controller 2; the detection bracket 14 is detachably connected with the shell 1, and the detection bracket 14 is also abutted against the controller 2; the design can ensure that the position of the detection bracket 14 is stable and does not deviate; the test device further comprises a guide spring 17, wherein the guide spring 17 is nested on the pressure sensor 16, the guide spring 17 is designed according to the size and parameters of a GB/T1973.3-2005 small cylindrical helical compression spring, the middle diameter of the spring is 0.8mm, the test load is 2.16N, and the guide spring 17 and the pressure sensor 16 are in an inclined shape; the specific angle can be 30 degrees, a guide rail groove 18 is formed in the side wall of the finger bin 8 close to the movement direction pressed by the finger tip of a person, a guide rail 19 is arranged on the finger tip stop dog 15, and the finger tip stop dog 15 is in sliding fit with the guide rail groove 18 through the guide rail 19; the soft material layer of the fingertip stop block 15 is internally made of sponge foaming materials, the outer layer of the soft material layer is lined with soft fabrics, and the soft material layer is matched with guide rails and springs, so that fingers are not pressed by components when extending in, a certain damping effect is achieved, the position of the fingertips is limited, and a fingertip limiting groove 20 with the width of 5 mm to 1 cm is formed in the fingertip stop block 15; when the fingertip is inserted, the fingertip can be automatically kept horizontal and centered according to the touch habit, the top of the fingertip stop block 15 is arc-shaped, the height of the notch 9 is larger than that of the fingertip stop block 15, and the nail can exceed the fingertip stop block 15 and cannot be extruded by the end part of the finger chamber 8.

The method for detecting the finger specifically includes the following steps, as shown in fig. 3 to 9:

step one, turning on a power switch 5, and respectively electrifying an LED light source 6, a spectrum sensor 7, a temperature sensor 11 and a pressure sensor 16 by a power module 3 through a controller 2;

inserting a finger into the finger bin 8 and extending into the bottom of the finger bin 8, wherein the nail part of the fingertip part can be positioned in a gap between the fingertip stop block 15 and the notch 9, so that the length of the nail can not interfere with the detection part of the fingertip, the position of the front end of the fingertip is limited by the fingertip limiting groove 20, the fingertip is horizontal, and the guide spring 17 provides a damping effect and limits the forward displacement of the fingertip; the unique material design of the fingertip stop block 15 can ensure that when the end part of a finger has a blocking effect, excessive extrusion cannot be generated, the movement direction is limited by the direction of the guide rail 19 and the direction of the guide spring 17, the finger basically keeps stable and horizontal, and deflection cannot occur;

dividing the fingers into three categories, wherein the first category is the thumb, the second category is the index finger, the middle finger and the ring finger, and the third category is the little finger;

step four, when the pressure sensor 16 receives the pressure signal and keeps stable, the fingertip is determined to reach the bottom of the finger chamber 8 and keep horizontal, and the pressure value received by the pressure sensor 16 is a coefficient K; when the temperature received by the temperature sensor 11 changes, the fingertip temperature T1 is determined, the unchanged temperature is the ambient temperature T2, and the number N of changes of the temperature sensor 11 occurs;

after the LED light source 6 emits light with specified wavelength, the light passes through the through hole 12 and passes through a detection part at the front end of a fingertip, a tunable filter of a Fabry-Perot interferometer is adapted in the spectrum sensor 7, and the optical receiving range of the tunable filter is adjusted to reach the nm level for spectrum receiving;

and step six, judging whether the type of the finger belongs to a first type, a second type or a third type according to the changed number N of the temperature sensors 11, combining the coefficient K, the fingertip temperature T1 and the ambient temperature T2, sending the ADC value converted after the optical signal is received by the spectrum sensor 7 to the controller 2 together, and calling an algorithm in a built-in data storage module through the controller 2 to calculate the blood sugar value.

The LED light source 6 emits light with specified wavelengths of 1500nm, 1525nm, 1550nm and 1575nm.

When the power switch 5 is turned on, the power module 3 is powered on by the semiconductor refrigerating sheet 32 through the controller 2, and the semiconductor refrigerating sheet 32 directly cools the spectrum sensor 7 through the heat dissipation copper sheet 31, so that the spectrum sensor 7 performs light sensing and receiving at a specified temperature.

The algorithm used by the invention usually contains some errors according to the spectrum information measured by the spectrum sensor, such as stray light, human tissue influence and the like, so that the measured data has certain noise and the calculation accuracy of blood sugar is influenced, therefore, the collected original spectrum data needs to be preprocessed before modeling, wherein the preprocessing comprises the steps that the types of fingers belong to a first type, a second type or a third type, the errors are reduced by combining a coefficient K, a fingertip temperature T1, an environment temperature T2 and the like, and effective information in the data is extracted, and the preprocessing comprises the steps of pairing so as to improve the calculation accuracy of a blood sugar model. Modeling the multi-modal spectral data using MATLAB software according to support vector machine theory.

Two groups of near infrared spectrum data with different wavelengths are used as independent variable matrixes of the model to be input, blood glucose values adopted by the household glucometer are used as dependent variables of the model, and a sample training set and a test set are divided. In the early stage, 8 groups of data are measured as modeling data by testing diabetic patients. Sorting the screened spectral data according to the blood sugar values measured by a household blood sugar meter, and dividing the training set and the test set according to the ratio of 3: 1 to ensure that the selected samples cover all the blood sugar values, namely 6 groups of sample data are used as the training set, and 2 groups of sample data are used as the test set. The sample data is then normalized. The spectral data measured by two groups of different wavelengths are used as independent variables, the blood sugar values are used as dependent variables, normalization is respectively carried out, probability distribution is in the same range, the influence on a modeling result due to too large data distribution range and inconsistent order of magnitude is reduced, and training efficiency is improved. The optimal kernel function, penalty factor coefficient (c) and parameter coefficient (g) of the kernel function are required to be selected for SVM parameter setting, and due to the difference of models and data, the optimal parameters cannot be obtained before modeling, so that the optimal values of c and g are obtained by adopting a cross validation method in the modeling process for training and predicting. The experimental training set has 6 groups of data, and the spectral data and the blood glucose true value of 6 groups of samples are used for training to obtain a model between the spectral data and the blood glucose true value. Substituting the 2 groups of data in the test set into the model for calculation to obtain the calculated value of the blood sugar in the 2 groups of data.

For the established correction model, four indexes of a correlation coefficient R, a correction set Root Mean Square Error (RMSEC), a test set Root Mean Square Error (RMSEP) and a relative error E are adopted to evaluate the model, wherein the correlation coefficient reflects the similarity degree of a predicted value and a theoretical value, and the root mean square error and the relative error reflect the model precision.

When two kinds of specific wavelength spectrum data are used for modeling, the correlation coefficient of a training set is 97.29%, the root mean square error is 0.3558mmol/L, the correlation coefficient of a testing set is 96.3%, the root mean square error is 0.3804mmol/L, the maximum relative error is 13.68%, and the average relative error is 0.069%.

Since the present invention is not limited to the above embodiments, those skilled in the art can make modifications and variations to the above embodiments without departing from the scope of the present invention.

Claims (7)

1. The utility model provides a many first perceptions and accurate noninvasive blood glucose detector of high data, includes casing (1), controller (2), power module (3), fin (4), switch (5), LED light source (6) and spectral sensor (7), be equipped with finger storehouse (8) in casing (1), power module (3) are connected and all establish in casing (1) with controller (2) electricity, establish at the tip of finger storehouse (8) and be located finger storehouse (8) top surface LED light source (6), fin (4) are connected with spectral sensor (7), spectral sensor (7) are located the tip of finger storehouse (8) and are located the bottom surface of finger storehouse (8), LED light source (6) are located same straight line with spectral sensor (7), its characterized in that: the fingertip pressure detection device comprises a fingertip pressure detection mechanism (13), wherein the fingertip pressure detection mechanism (13) comprises a detection support (14), a fingertip stop block (15) and a pressure sensor (16), the detection support (14) is arranged on a shell (1), and the fingertip stop block (15) is connected with the detection support (14) through the pressure sensor (16); the pressure sensor (16) is positioned at the notch (9) and at the edge of the slope stop block (10), and the pressure sensor (16) is also electrically connected with the controller (2); the pressure sensor is characterized by further comprising a guide spring (17), wherein the guide spring (17) is nested on the pressure sensor (16), and the guide spring (17) and the pressure sensor (16) are inclined; a guide rail groove (18) is formed in the side wall of the finger bin (8), a guide rail (19) is arranged on the fingertip stop block (15), and the fingertip stop block (15) is in sliding fit with the guide rail groove (18) through the guide rail (19); and a fingertip limiting groove (20) with the width of 5 mm to 1 cm is arranged on the fingertip stop block (15).

2. The multivariate perception and high data accuracy noninvasive blood glucose monitor of claim 1, wherein: the detection bracket (14) is detachably connected with the shell (1), and the detection bracket (14) is also abutted against the controller (2).

3. The multivariate perception and high data accuracy noninvasive blood glucose monitor of claim 1, wherein: the top of the fingertip stop block (15) is arc-shaped, and the height of the notch (9) is larger than that of the fingertip stop block (15).

4. The multivariate perception and high data accuracy noninvasive blood glucose monitor of claim 1, wherein: still include semiconductor refrigeration piece (32) and heat dissipation copper sheet (31), spectral sensor (7) still are connected with heat dissipation copper sheet (31), fin (4) are connected with heat dissipation copper sheet (31) through semiconductor refrigeration piece (32), the refrigeration end of semiconductor refrigeration piece (32) is pasted to heat dissipation copper sheet (31), the end of heating of semiconductor refrigeration piece (32) is pasted to fin (4).

5. A multivariate sensing and highly data-accurate noninvasive blood glucose measuring method using the noninvasive blood glucose measuring instrument according to claim 1, characterized in that: the method comprises the following steps: step one, a power switch (5) is turned on, and a power module (3) is used for respectively electrifying an LED light source (6), a spectrum sensor (7), a temperature sensor (11) and a pressure sensor (16) through a controller (2); inserting a finger into the finger bin (8) and extending into the bottom of the finger bin (8), wherein the nail part of the fingertip part is positioned in a gap between the fingertip stop block (15) and the notch (9), the front end of the fingertip is limited by a fingertip limiting groove (20) to enable the fingertip to be horizontal, and a guide spring (17) provides a damping effect and limits the forward displacement of the fingertip; dividing the fingers into three categories, wherein the first category is the thumb, the second category is the index finger, the middle finger and the ring finger, and the third category is the little finger; step four, when the pressure sensor (16) receives the pressure signal and keeps stable, the fingertip is determined to reach the bottom of the finger bin (8) and keep horizontal, and the pressure value received by the pressure sensor (16) is a coefficient K; when the temperature received by the temperature sensor (11) changes, the fingertip temperature T1 is judged, the unchanged temperature is the environment temperature T2, and the changed number N of the temperature sensor (11) is obtained; after the LED light source (6) emits light with specified wavelength, the light passes through the through hole (12) and passes through a detection part at the front end of a fingertip, a tunable filter of a Fabry-Perot interferometer is adapted in the spectrum sensor (7), and the optical receiving range of the tunable filter is adjusted to reach the nm level for spectrum receiving; and step six, judging whether the type of the finger belongs to a first type, a second type or a third type according to the changed number N of the temperature sensors (11), combining the coefficient K, the fingertip temperature T1 and the ambient temperature T2, sending an ADC (analog to digital converter) value converted after the spectral sensor (7) receives the optical signal to the controller (2), and calling an algorithm in a built-in data storage module through the controller (2) to calculate the blood sugar value.

6. The multivariate perception and high-data-accuracy noninvasive blood glucose detection method according to claim 5, characterized in that: the LED light source (6) emits light with specified wavelengths of 1500nm, 1525nm, 1550nm and 1575nm.

7. The multivariate perception and high-data-accuracy noninvasive blood glucose detection method according to claim 5, characterized in that: when the power switch (5) is turned on, the power module (3) is also powered on for the semiconductor refrigerating sheet (32) through the controller (2), and the semiconductor refrigerating sheet (32) directly cools the spectrum sensor (7) through the heat dissipation copper sheet (31), so that the spectrum sensor (7) performs light sensing receiving at a specified temperature.

Images