A method for bidirectional correction of near-infrared spectroscopy data for non-invasive blood glucose detection

Abstract

The present invention relates to spectral data correction technology, which discloses a method for bidirectional correction of non-invasive blood glucose near-infrared spectral data, effectively improving the accuracy and stability of spectral data, and thus improving the prediction ability of the non-invasive blood sample spectral model. The method first uses a miniaturized near-infrared spectrometer to non-invasively collect spectral data of blood samples at different temperatures, and calculates the mean of the spectral data of blood samples at different temperatures, and then calculates the temperature spectrum correction coefficient in combination with the temperature value of the non-invasively collected blood sample and the corresponding spectral mean, and calculates the temperature spectrum correction formula in combination with the temperature spectrum correction coefficient, and finally corrects the spectral data of the non-invasively collected blood sample to be tested in combination with the temperature spectrum correction formula to complete the data correction.

Description

Technical Field

[0001] The invention relates to a spectrum data correction technology, and in particular to a method for bidirectionally correcting near-infrared spectrum data for non-invasive blood sugar detection. Background Art

[0002] Diabetes seriously threatens people's life and health, but there is no clinical cure for diabetes. In clinical practice, frequent blood sugar concentration testing and drug control are often used to maintain the blood sugar concentration of diabetic patients at a normal level. Scholars at home and abroad have invested a lot of energy and resources in blood sugar concentration detection technology and have achieved certain research results. The current blood sugar detection methods include three categories: non-invasive, minimally invasive and invasive. Among them, invasive and minimally invasive detection have higher accuracy and can meet the requirements of clinical blood sugar detection. However, both invasive and minimally invasive detection will bring pain and discomfort to diabetic patients. Therefore, the demand for non-invasive blood sugar detection has emerged.

[0003] Among the many non-invasive blood glucose detection methods, the miniaturized near-infrared detection method has become the most widely used detection method due to its simple equipment, high efficiency and cost advantages. When using miniaturized near-infrared equipment for blood glucose detection, temperature is an important influencing factor that cannot be ignored. Temperature mainly changes the vibration spectrum of molecules by changing the intermolecular forces. In the spectrum of blood, temperature will affect the OH stretching vibration and its frequency harmonics: when the temperature increases, the OH frequency band moves to low wavelengths and becomes sharper. The researchers' quantitative research found that the near-infrared light absorption caused by a temperature change of 0.1°C is approximately equivalent to a change in glucose concentration of 1 mmol / L. The above shows that when the temperature changes, the spectral data collected by the near-infrared will change significantly. In addition, the temperature change of the measured part will also affect the optical property parameters of the tissue, further increasing the detection error. Therefore, temperature is an important factor that cannot be ignored in the near-infrared light detection of blood glucose. The near-infrared non-invasive blood glucose detection technology is easily affected by temperature, resulting in poor stability and low accuracy of the blood near-infrared spectrum data, which in turn affects the predictive ability of the spectral model and reduces the detection accuracy of the non-invasive blood glucose detection technology. Therefore, how to implement a spectral data correction method to correct blood spectral data under the influence of temperature has become an urgent problem to be solved. Summary of the invention

[0004] The technical problem to be solved by the present invention is to propose a method for bidirectionally correcting near-infrared spectral data of non-invasive blood glucose detection, which effectively improves the accuracy and stability of spectral data, and further improves the predictive ability of the spectral model of non-invasive blood sample collection.

[0005] The technical solution adopted by the present invention to solve the above technical problems is:

[0006] A method for bidirectionally correcting non-invasive blood sugar near-infrared spectroscopy data comprises the following steps:

[0007] S1. Non-invasive collection of spectral data of blood samples under different temperature conditions;

[0008] S2, calculating the mean of the spectral data of the blood samples under different temperature conditions;

[0009] S3, calculating the temperature spectrum correction coefficient by combining the temperature value of the non-invasively collected blood sample and the corresponding spectrum mean;

[0010] S4, calculating the temperature spectrum correction formula in combination with the temperature spectrum correction coefficient;

[0011] S5. Use the temperature spectrum correction formula to correct the spectrum data of the non-invasively collected blood sample to be tested.

[0012] Furthermore, in step S1, a wavelength-averaged miniaturized near-infrared spectrometer is used to non-invasively collect spectral data of blood samples under different temperature conditions. The method for collecting spectral data of blood samples under different temperature conditions is:

[0013] Set the highest ambient temperature point of non-invasive data collection in actual application to C max , the lowest temperature point is C min , the highest temperature point of the body surface temperature is S max , the lowest temperature point is S min ; Take the highest and lowest temperature points as the correction reference points, and collect the temperature (C max , S max )、(C max , S min )、(C min , S max )、(C min , S min )Near-infrared spectral data under four reference conditions.

[0014] Furthermore, when the collection temperature is (C max , S max )、(C max , S min )、(C min , S max )、(C min , S min ) when obtaining near-infrared spectral data under four reference conditions, three spectral data are collected for each blood sample under each reference condition, and the light intensity values ​​of the same wavelength points on the three spectral data are averaged and then one spectral data is output. The spectral data finally output is the actual spectral data of the corresponding blood sample under the reference condition, which is used as the near-infrared spectral correction reference data.

[0015] Furthermore, in step S2, the calculation of the mean of the spectral data of the blood samples under different temperature conditions specifically includes: setting the spectral data of the blood samples to G = (G1, G2..., G n ), where n is the number of wavelength points of the miniaturized near-infrared spectrometer, G n is the light intensity value at the nth wavelength point;

[0016] The spectral data of the blood sample is averaged to obtain the mean value G of the spectral data. mean for:

[0017] G mean =(G1+G2+……G n ) / n

[0018] For temperature (C max , S max )、(C max , S min )、(C min , S max )、(C min , S min ) The near-infrared spectral correction benchmark data under the four benchmark conditions are averaged to obtain the corresponding spectral mean values ​​G under the four benchmark conditions. max，max , G max，min , G min，max , G min，min .

[0019] Furthermore, in step S3, the temperature spectrum correction coefficient is calculated by combining the temperature value of the non-invasively collected blood sample and the corresponding spectrum mean, specifically including:

[0020] Taking the lowest body surface temperature as the reference point, the maximum offset A of the spectral data under the influence of ambient temperature is calculated as:

[0021] A=G max，min -G min，min

[0022] Taking the lowest ambient temperature as the reference point, the maximum offset B of the spectral data under the influence of body surface temperature is calculated as:

[0023] B=G min，max -G min，min

[0024] Combining the spectral data of the lowest ambient and body surface temperature with the spectral data of the highest ambient and body surface temperature, the maximum offset D of the spectral data under the combined effect of the two is calculated as:

[0025] D=G max，max -G min，min

[0026] Then the temperature spectrum correction coefficient α is calculated as:

[0027] C min <C<C max , S min <S<S max

[0028] Among them, C is the actual ambient temperature of the person being measured, and S is the actual body surface temperature of the person being measured.

[0029] Furthermore, in step S4, the temperature spectrum correction formula is calculated in combination with the temperature spectrum correction coefficient, specifically including:

[0030] Taking the lowest ambient temperature and the lowest body surface temperature as reference points, combined with the spectrum correction coefficient and the temperature extreme point, the temperature spectrum correction formula is calculated as follows:

[0031] C min <C<C max , S min <S<S max

[0032] Where T is the mean correction value of the spectral data.

[0033] Furthermore, in step S5, the spectral data of the non-invasively collected blood sample to be tested is corrected using the temperature spectrum correction formula, specifically including:

[0034] If the ambient temperature of the blood sample collected by the miniaturized near-infrared spectrometer is C h , the body surface temperature is S h , the spectral data is G h =(G1', G2'..., G n '), when C min <C h <C max , S min <S h <S max When the ambient temperature is C h , body surface temperature S h Substitute the spectrum correction formula to calculate the mean correction value T1 of the spectrum data under the temperature condition, and use this correction value to calculate the actual spectrum data G of the non-invasively collected blood sample to be tested. h The light intensity values ​​in each wavelength range are corrected one by one to obtain the corrected spectral data G h1 =(G1'+T1, G2'+T1..., G n '+T1).

[0035] Furthermore, step S5 further comprises: at an ambient temperature C h=C min And the body surface temperature S h ≠S min When the surface temperature spectrum data is combined with the single correction, the spectrum data mean correction value T2 is calculated as:

[0036] S min <S h ≤S max

[0037] The correction value is used to calculate the actual spectral data G of the blood sample to be tested. h The light intensity values ​​in each wavelength range are corrected one by one to obtain the corrected spectral data G h2 =(G1'+T2, G2'+T2..., G n '+T2).

[0038] Furthermore, step S5 further includes: at the body surface temperature S h =S min And the ambient temperature C h ≠C min When the ambient temperature spectrum data is combined with the single correction, the spectrum data mean correction value T3 is calculated as:

[0039] C min <C h ≤C max

[0040] The correction value is used to calculate the actual spectral data G of the blood sample to be tested. h The light intensity values ​​in each wavelength range are corrected one by one to obtain the corrected spectral data G h3 =(G1'+T3, G2'+T3..., G n '+T3).

[0041] The beneficial effects of the present invention are:

[0042] This method not only performs bidirectional correction on the miniaturized near-infrared spectral data from the two influencing factors of ambient temperature and body surface temperature, but also calculates the temperature spectral correction coefficient through the spectral data of blood samples at different temperatures. The influence of the two influencing factors on the blood samples is characterized by the spectral correction coefficient, which effectively improves the accuracy and stability of the spectral data. After correcting the spectral data through this method, the predictive ability of the spectral model of non-invasively collected blood samples can be effectively improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a flow chart of a method for bidirectionally correcting non-invasive blood sugar near-infrared spectroscopy data in an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The present invention aims to propose a method for bidirectional correction of non-invasive blood glucose near-infrared spectral data, effectively improving the accuracy and stability of spectral data, and thus improving the prediction ability of the non-invasive blood sample spectral model. The method first uses a miniaturized near-infrared spectrometer to non-invasively collect spectral data of blood samples at different temperatures, and calculates the mean of the spectral data of blood samples at different temperatures, and then calculates the temperature spectrum correction coefficient in combination with the temperature value of the non-invasively collected blood sample and the corresponding spectral mean, and calculates the temperature spectrum correction formula in combination with the temperature spectrum correction coefficient, and finally corrects the spectral data of the non-invasively collected blood sample in combination with the temperature spectrum correction formula to complete the data correction.

[0045] Example:

[0046] like Figure 1 As shown, the method for bidirectionally correcting non-invasive blood glucose near-infrared spectroscopy data in this embodiment includes the following steps:

[0047] Step 101: non-invasively collect spectral data of blood samples under different temperature conditions:

[0048] In this embodiment, the use of a wavelength-averaged miniaturized near-infrared spectrometer to non-invasively collect blood sample spectral data under different temperature conditions can collect sample spectral data with the greatest uniformity, wherein the different temperature conditions are specifically different ambient temperatures and different body surface temperatures. According to the actual blood sample to be tested in this embodiment, a miniaturized near-infrared spectrometer with a wavelength range of 1300nm to 1850nm and a resolution of 10nm is selected for non-invasive collection of spectral data. It can be calculated that the wavelength points contained in each blood sample spectral data are K=1+(1850-1300) / 10=56, and the wavelength range corresponding to the 1st to 56th wavelength points is (1300nm, 1310nm, ... 1850nm). The spectral data collected from each blood sample is actually represented as a matrix set of light intensity values ​​at 56 wavelength points. During the collection process, three spectral data are collected for each blood sample, and the light intensity values ​​of 56 wavelength points on the three spectral data are averaged respectively. After averaging the three spectral data, one spectral data is output. This spectral data is the actual spectral data of the sample. The advantage of this is that it can effectively reduce collection errors and increase data reliability.

[0049] In addition, in the actual application of near-infrared non-invasive blood glucose analysis, for the same non-invasively collected blood sample spectral data, when the measured person's body surface temperature is fixed and the ambient temperature is different, the non-invasively collected blood spectral data will have obvious differences. At the same time, when the measured person's ambient temperature is fixed and the body surface temperature is different, the non-invasively collected blood spectral data will also have obvious differences. At the same time, because in actual applications, the measured person's ambient temperature and body surface temperature are always within a threshold range, if the highest ambient temperature point is set to C max , the lowest temperature point is C min Similarly, the highest temperature point of the body surface temperature is set as S max , the lowest temperature point is S min The highest and lowest temperature points are taken as the correction reference points, and the temperature is collected at (C max , S max )、(C max , S min )、(C min , S max )、(C min , S min ) Near-infrared spectral data under four reference conditions are obtained, and these spectral data are used as spectral correction reference data.

[0050] Step 102: Calculate the mean of the spectral data of the blood samples under different temperature conditions:

[0051] The mean of the spectral data of the blood sample can well reflect the range of the spectral data. That is, when calculating the spectral data, the spectral mean can be used to approximately replace the spectral data for calculation. The mean calculation of the spectral data is implemented by averaging the light intensity values ​​at 56 wavelength points in the spectral data.

[0052] In this embodiment, the spectrum data of the blood sample is set to G=(G1, G2..., G 56 ), where G1 is the light intensity at a wavelength of 1300 nm, G2 is the light intensity at a wavelength of 1310 nm, and so on. The spectral data of the blood sample is averaged to obtain the mean value G of the spectral data. mean for:

[0053] G mean =(G1+G2+……G 56 ) / 56

[0054] After the averaging operation, the original spectral data containing 56 light intensity value points are replaced by 1 light intensity value point for calculation, which greatly improves the calculation efficiency.

[0055] For temperature (C max , Smax )、(C max , S min )、(C min , S max )、(C min , S min ) The near-infrared spectral correction benchmark data under the four benchmark conditions are averaged to obtain the spectral mean values ​​(G max，max , G max，min , G min，max , G min，min ).

[0056] Step 103, calculate the temperature spectrum correction coefficient by combining the temperature value of the non-invasively collected blood sample and the corresponding spectrum mean value:

[0057] In this step, a two-segment balanced difference method is used for calculation, and the two-segment balanced data correspond to two temperature reference point data, namely, two segments of spectrum data corresponding to the maximum temperature point and the minimum temperature point.

[0058] In this embodiment, in the actual near-infrared non-invasive blood glucose analysis application, the sum of the spectral data offsets caused by a single change in ambient temperature or body surface temperature is not equal to the offset caused by the combined effect of ambient temperature and body surface temperature. For example, when the body surface temperature is constant, the ambient temperature of the person being measured is offset by C from the ambient temperature of the lowest reference point. off degrees, the spectral data offset is G off1 When the ambient temperature is constant, the actual body temperature of the person being measured deviates from the lowest reference point body temperature by S off degrees, the spectral data offset is G off2 When the measured person satisfies the ambient temperature and the lowest reference point ambient temperature offset by C off degrees and the actual body surface temperature deviates from the lowest reference point body surface temperature by S off When the spectral data is not offset by (G off1 +G off2 ), it is necessary to calculate the temperature spectrum correction coefficient to correspond the relationship between the spectral data offset caused by a single temperature change and the spectral data offset under the combined effect of the two.

[0059] The specific calculation method of the spectral correction coefficient is:

[0060] Taking the lowest body surface temperature as the reference point, the maximum offset A of the spectral data under the influence of ambient temperature is calculated as:

[0061] A=G max，min -G min，min

[0062] Taking the lowest ambient temperature as the reference point, the maximum offset B of the spectral data under the influence of body surface temperature is calculated as:

[0063] B=G min，max -G min，min

[0064] Combining the spectral data of the lowest ambient and body surface temperature with the spectral data of the highest ambient and body surface temperature, the maximum offset D of the spectral data under the combined effect of the two is calculated as:

[0065] D=G max，max -G min，min

[0066] The deviation values ​​calculated by fixing the ambient temperature or body surface temperature at the lowest reference point respectively reflect the influence of a single factor (ambient temperature or body surface temperature) on the spectral data of non-invasively collected blood samples. It is necessary to convert the single factor into the influence of the two together, that is, it is necessary to calculate the corresponding temperature spectrum correction coefficient for the numerical conversion between the two. From the above, the temperature spectrum correction coefficient α is calculated as:

[0067] C min <C<C max , S min <S<S max

[0068] Among them, C is the actual ambient temperature of the person being measured, and S is the actual body surface temperature of the person being measured.

[0069] Step 104, calculate the temperature spectrum correction formula in combination with the temperature spectrum correction coefficient:

[0070] In the correction of the spectral data of the ambient temperature or body surface temperature alone, the mean of the spectral data is often linearly correlated with the temperature interval difference, and it can be corrected by the conventional linear correction method. However, under the combined effect of ambient temperature and body surface temperature, the temperature interval difference is nonlinearly correlated with the mean of the spectral data, and the conventional linear correction method cannot correct the spectral data of blood samples under the combined effect of the two. The temperature spectrum correction coefficient can well reflect the influence of temperature on blood samples. The temperature spectrum correction formula can be obtained by jointly calculating the temperature correction coefficient and the temperature extreme value data. This formula comprehensively considers the combined effect of ambient temperature and body surface temperature, and can well correct the spectral data of blood samples under the combined effect.

[0071] In this embodiment, the temperature spectrum correction formula is calculated by taking the lowest ambient temperature and the lowest body surface temperature as reference points and combining the spectrum correction coefficient and the temperature extreme point as follows:

[0072]

[0073] Where T is the mean correction value of the spectral data.

[0074] Step 105: Correct the spectrum data of the non-invasively collected blood sample using the temperature spectrum correction formula:

[0075] The miniaturized near-infrared spectroscopy device adds a temperature sensor to the spectral sensor to detect the current ambient temperature and the body surface temperature of the person being measured. After obtaining the temperature value, the value is substituted into the temperature spectrum correction formula to obtain the spectral data deviation, and then the current blood sample spectral data is corrected based on the deviation, thus completing the entire correction process and improving the accuracy and stability of the spectral data.

[0076] In this embodiment, if the ambient temperature of the blood sample to be tested collected by the miniaturized near-infrared spectrometer is C h , the body surface temperature is S h , the spectral data is G h =(G1', G2'..., G 56 '), when C min <C h <C max , S min <S h <S max When the ambient temperature is C h , body surface temperature S h Substitute the spectrum correction formula to calculate the mean correction value T1 of the spectrum data under the temperature condition, and use this correction value to calculate the actual spectrum data G of the non-invasively collected blood sample to be tested. h The light intensity values ​​in each wavelength range are corrected one by one to obtain the corrected spectral data G h1 =(G1'+T1, G2'+T1..., G 56 '+T1);

[0077] In addition to the above situation, when there is C h =C min or S h =S min In this case, the spectral data of the blood sample to be tested only needs to be corrected for the ambient temperature or the body surface temperature, and there is no need to calculate the spectral correction coefficient, and the conventional linear correction method can be used to correct it.

[0078] At ambient temperature C h =C min And the body surface temperature S h ≠S min When the surface temperature spectrum data is combined with the single correction, the spectrum data mean correction value T2 is calculated as:

[0079] S min<S h ≤S max

[0080] The correction value is used to calculate the actual spectral data G of the blood sample to be tested. h The light intensity values ​​in each wavelength range are corrected one by one to obtain the corrected spectral data G h2 =(G1'+T2, G2'+T2..., G 56 '+T2);

[0081] At the body surface temperature S h =S min And the ambient temperature C h ≠C min When the ambient temperature spectrum data is combined with the single correction, the spectrum data mean correction value T3 is calculated as:

[0082] C min <C h ≤C max

[0083] The correction value is used to calculate the actual spectral data G of the blood sample to be tested. h The light intensity values ​​in each wavelength range are corrected one by one to obtain the corrected spectral data G h3 =(G1'+T3, G2'+T3..., G 56 '+T3);

[0084] At ambient temperature C h =C min And the body surface temperature S h =S min In this case, there is no need to correct the spectral data of the non-invasively collected blood sample to be tested under this condition.

[0085] Finally, it should be noted that the above embodiments are only preferred implementations and are not intended to limit the present invention. It should be pointed out that for those skilled in the art, several modifications, equivalent replacements, improvements, etc. can be made without departing from the scope of the present invention and the scope of protection of the claims, and all of these should be included in the protection scope of the present invention.

Claims

1. A method for bidirectionally correcting non-invasive near-infrared spectroscopy data for blood glucose detection, characterized in that: The following steps are involved: S1. Non-invasive collection of spectral data of blood samples under different temperature conditions; S2, calculating the mean of the spectral data of the blood samples under different temperature conditions; S3, calculating the temperature spectrum correction coefficient by combining the temperature value of the non-invasively collected blood sample and the corresponding spectrum mean; S4, calculating the temperature spectrum correction formula in combination with the temperature spectrum correction coefficient; S5, using a temperature spectrum correction formula to correct the spectrum data of the non-invasively collected blood sample to be tested; In step S1, a wavelength-averaged miniaturized near-infrared spectrometer is used to non-invasively collect spectral data of blood samples under different temperature conditions. The method for collecting spectral data of blood samples under different temperature conditions is: Set the highest ambient temperature point of non-invasive data collection in actual application to C max , the lowest temperature point is C min , the highest temperature point of the body surface temperature is S max , the lowest temperature point is S min ; Take the highest and lowest temperature points as the correction reference points, and collect the temperature (C max , S max )、(C max , S min )、(C min , S max )、(C min , S min )Near infrared spectral data under four reference conditions; In step S2, the calculation of the mean of the spectral data of the blood samples under different temperature conditions specifically includes: setting the spectral data of the blood samples to G = (G1, G2..., G n ), where n is the number of wavelength points of the miniaturized near-infrared spectrometer, G n is the light intensity value at the nth wavelength point; The spectral data of the blood sample is averaged to obtain the mean value G of the spectral data. mean for: G mean =(G1+G2+……G n ) / n For temperature (C max , S max )、(C max , S min )、(C min , S max )、(C min , S min ) The near-infrared spectral correction benchmark data under the four benchmark conditions are averaged to obtain the corresponding spectral mean values ​​G under the four benchmark conditions. max，max , G max，min , G min，max , G min，min ; In step S3, the temperature spectrum correction coefficient is calculated by combining the temperature value of the non-invasively collected blood sample and the corresponding spectrum mean, specifically including: Taking the lowest body surface temperature as the reference point, the maximum offset A of the spectral data under the influence of ambient temperature is calculated as: A=G max，min -G min，min Taking the lowest ambient temperature as the reference point, the maximum offset B of the spectral data under the influence of body surface temperature is calculated as: B=G min，max -G min，min Combining the spectral data of the lowest ambient and body surface temperature with the spectral data of the highest ambient and body surface temperature, the maximum offset D of the spectral data under the combined effect of the two is calculated as: D=G max，max -G min，min Then the temperature spectrum correction coefficient α is calculated as: Wherein, C is the actual ambient temperature of the person being measured, and S is the actual body surface temperature of the person being measured; In step S4, the temperature spectrum correction formula is calculated in combination with the temperature spectrum correction coefficient, specifically including: Taking the lowest ambient temperature and the lowest body surface temperature as reference points, combined with the spectrum correction coefficient and the temperature extreme point, the temperature spectrum correction formula is calculated as follows: Where T is the mean correction value of the spectral data.

2. A method for bidirectionally correcting non-invasive blood glucose near-infrared spectroscopy data as claimed in claim 1, characterized in that: When the collection temperature is (C max , S max )、(C max , S min )、(C min , S max )、(C min , S min ) when obtaining near-infrared spectral data under four reference conditions, three spectral data are collected for each blood sample under each reference condition, and the light intensity values ​​of the same wavelength points on the three spectral data are averaged and then one spectral data is output. The spectral data finally output is the actual spectral data of the corresponding blood sample under the reference condition, which is used as the near-infrared spectral correction reference data.

3. A method for bidirectionally correcting non-invasive blood glucose near-infrared spectroscopy data as claimed in claim 1, characterized in that: In step S5, the temperature spectrum correction formula is used to correct the spectrum data of the non-invasively collected blood sample to be tested, specifically including: If the ambient temperature of the blood sample collected by the miniaturized near-infrared spectrometer is C h , the body surface temperature is S h , the spectral data is G h =(G1', G2'..., G n '), when C min <C h <C max , S min h max When the ambient temperature is C h , body surface temperature S h Substitute the spectrum correction formula to calculate the mean correction value T1 of the spectrum data under the temperature condition, and use this correction value to calculate the actual spectrum data G of the non-invasively collected blood sample to be tested. h The light intensity values ​​in each wavelength range are corrected one by one to obtain the corrected spectral data G h1 =(G1'+T1, G2'+T1..., G n '+T1).​​ 4. A method for bidirectionally correcting non-invasive blood glucose near-infrared spectroscopy data as claimed in claim 3, characterized in that: Step S5 also includes: at an ambient temperature C h =C min And the body surface temperature S h ≠S min When the surface temperature spectrum data is combined with the single correction, the spectrum data mean correction value T2 is calculated as: The correction value is used to calculate the actual spectral data G of the blood sample to be tested. h The light intensity values ​​in each wavelength range are corrected one by one to obtain the corrected spectral data G h2 =(G1'+T2, G2'+T2..., G n '+T2).

5. A method for bidirectionally correcting non-invasive near-infrared spectroscopy data for blood sugar detection as claimed in claim 3, characterized in that: Step S5 also includes: at the body surface temperature S h =S min And the ambient temperature C h ≠C min When the ambient temperature spectrum data is combined with the single correction, the spectrum data mean correction value T3 is calculated as: The correction value is used to calculate the actual spectral data G of the blood sample to be tested. h The light intensity values ​​in each wavelength range are corrected one by one to obtain the corrected spectral data G h3 =(G1'+T3, G2'+T3..., G n '+T3).

Images