A method for coordinating near-infrared spectral data with temperature and humidity

Abstract

The present invention relates to the field of infrared spectroscopy. In order to solve the problem that the interval difference of temperature and humidity is nonlinearly correlated with the mean of spectral data under the joint action of temperature and humidity, the conventional linear correction method cannot correct the sample spectral data under the joint action of temperature and humidity. The present invention provides a method for coordinating near-infrared spectral data with temperature and humidity, the core idea of ​​which is: using a near-infrared spectrometer to collect spectral data of samples with different temperatures and humidity, and calculating the mean of the spectral data of samples with different temperatures and humidity, and then calculating the temperature and humidity spectral correction coefficient in combination with the sample temperature and humidity values ​​and the corresponding spectral mean, and calculating the temperature and humidity spectral correction formula in combination with the coefficient, and finally correcting the sample spectral data to be corrected in combination with the temperature and humidity spectral correction formula to complete the correction of the spectral data. The method for coordinating near-infrared spectral data with temperature and humidity proposed by the present invention is mainly used to correct portable infrared spectral data.

Description

Technical Field

[0001] The invention relates to the field of near infrared spectroscopy, and in particular to a method for coordinating near infrared spectroscopy data with temperature and humidity. Background Art

[0002] Miniaturized near-infrared spectrometers are small and inexpensive, but due to the limitations of their own hardware conditions, the stability of their spectral data is weak and the accuracy is low. Especially in the actual application of miniaturized near-infrared detection, the spectral data acquisition process requires extremely precise, irregular acquisition process, uneven contact between equipment and samples during acquisition, changes in the acquisition environment and other factors can easily cause spectral data deviation, thereby affecting the prediction ability of the spectral model and reducing the detection accuracy of sample detection. Therefore, solving the problem of unstable spectral data and low accuracy in miniaturized near-infrared detection technology has become an urgent problem to be solved.

[0003] In actual testing, temperature and humidity often act on the test samples at the same time. In the correction of individual temperature or humidity spectral data, the mean of the spectral data is often linearly correlated with the temperature or humidity interval difference, and it can be corrected by using conventional linear correction methods. However, under the combined effect of temperature and humidity, the temperature and humidity interval difference is nonlinearly correlated with the mean of the spectral data, and conventional linear correction methods cannot correct the spectral data of the wine lees sample under the combined effect of temperature and humidity. For example, in the detection of wine lees samples, under different ambient temperature or humidity conditions, the temperature and humidity of the wine lees sample itself are also easily affected by the environment and change, resulting in the problem of poor consistency of the wine lees sample, which in turn causes the spectral data collected by the miniaturized near-infrared spectrometer to have a large deviation, further affecting the prediction ability of the miniaturized near-infrared spectrometer for wine lees samples. Summary of the invention

[0004] The technical problems to be solved by the present invention are:

[0005] The invention provides a method for coordinating near-infrared spectral data with temperature and humidity to solve the problem of unstable spectral data and low precision when temperature and humidity act on the sample to be tested at the same time in near-infrared detection technology.

[0006] The technical solution adopted by the present invention to solve the above problems is as follows:

[0007] A method for coordinating near-infrared spectral data using temperature and humidity, comprising the following steps:

[0008] Step 1: Use a near-infrared spectrometer to collect spectral data of samples at different temperatures and humidities; use it as baseline correction data;

[0009] Step 2: Calculate the average light intensity value of the sample spectrum under different temperatures and humidities collected in step 1;

[0010] Step 3: Calculate the temperature and humidity spectrum correction coefficient by combining the sample temperature and humidity values ​​and the corresponding light intensity average value;

[0011] Step 4: Calculate the temperature and humidity spectrum correction formula according to the temperature and humidity spectrum correction coefficient;

[0012] Step 5: Use the temperature and humidity spectrum correction formula to correct the light intensity value of the wavelength point of the sample to be corrected.

[0013] Further, the step 1 is as follows: the temperature and humidity values ​​of the highest point, the lowest point, and the mean point of the temperature and humidity of the sample are used as correction reference points, and the highest temperature point of the sample is C max , the lowest temperature point is C min , the mean temperature point is C mean =(C max +C min ) / 2, the highest humidity point of the sample is RH max , the lowest humidity point is RH min , the mean humidity point is RH mean =RH max +RH min ) / 2; respectively collect (C max , RH max )、(C max , RH mean )、(C max , RH min )、(C mean , RH max )、(C mean , RH mean )、(C mean , RH min )、(C min , RH max )、(C min , RH mean )、(C min , RH min ) The spectral data under nine reference conditions are used as the spectral correction reference data.

[0014] Furthermore, the implementation method of step 2 is as follows: the spectral data under each reference state in step 1 is collected multiple times, and the average calculation is performed. The averaged data is the actual spectral data of the sample, and then the number of wavelength points included in the sample spectrum under the nine reference states is calculated according to the selected wavelength range and the resolution of the near-infrared spectrometer, and the light intensity value (P1, P2..., P n ), the light intensity value of the sample is averaged to obtain the light intensity mean value P of the spectrum mean =(P1+P2+……P n) / n, the average light intensity values ​​under the nine reference conditions are (P max，max , P max，mean , P max，min , P mean，max , P mean，mean , P mean，min , P min，max , P min，mean , P min，min ).

[0015] Furthermore, the specific method for calculating the temperature and humidity spectrum correction coefficient in step 3 is:

[0016] Keep the humidity at the mean point and obtain the maximum temperature point (C max , RH mean ), mean temperature point (C mean , RH mean )、Minimum temperature point (C min , RH mean ) corresponds to the mean light intensity value (P max，mean , P mean，mean , P min，mean ), with P mean，mean As the center, calculate P max，mean , P min，mean The average of the two light intensity values ​​and P mean，mean The deviation of the center point is:

[0017]

[0018] Among them A max With A min P max，mean , P min，mean The average value of the light intensity at two wavelengths is P mean，mean Deviation value of the center point;

[0019] Keep the temperature at the mean point and obtain the maximum humidity point (C mean , RH max ), mean humidity point (C mean , RH mean )、Minimum humidity point (C mean , RH min ) corresponds to the mean light intensity value (P mean，max , P mean，mean , P mean，min ), with P mean，mean As the center point, mean，max , P mean，min The deviation values ​​of the two light intensity values ​​are calculated by the average value, and the deviation values ​​are:

[0020]

[0021] Among them B max With B min P mean，max , P mean，min The average value of the light intensity at two wavelengths is P mean，mean Deviation value of the center point;

[0022] P mean，mean As the center point, obtain all extreme points and P under the combined effect of temperature and humidity mean，mean Deviation of light intensity value:

[0023]

[0024] Where D max，max , D max，min , D min，max With D min，min P max，max , P max，min , P min，max , P min，min The average value of the light intensity at four wavelengths and P mean，mean Deviation value of the center point.

[0025] Calculate the temperature and humidity spectrum correction coefficients α1, α2, α3, and α4 based on the deviation values ​​obtained above:

[0026]

[0027] Wherein, C is the actual temperature value of the sample to be corrected, and RH is the actual humidity value of the sample to be corrected.

[0028] Further, the method for calculating the temperature and humidity spectrum correction formula described in step 4 is:

[0029] When C mean <C≤C max , RH mean <RH≤RH max When the temperature and humidity extreme point (C max , RH max ) and the corresponding wavelength point light intensity value P max，max The calculation spectrum correction formula M1 is as follows:

[0030]

[0031] When C min ≤C <C mean , RH min ≤RH <RH mean When the temperature and humidity extreme point (C min , RH min) and the corresponding wavelength point light intensity value P min，min The calculation spectrum correction formula M2 is as follows:

[0032]

[0033] When C mean <C≤C max , RH min ≤RH <RH mean When the temperature and humidity extreme point (C max , RH min ) and the corresponding wavelength point light intensity value P max，min The calculation spectrum correction formula M3 is as follows:

[0034]

[0035] When C min ≤C <C mean , RH mean <RH≤RH max When the temperature and humidity extreme point (C min , RH max ) and the corresponding wavelength point light intensity value P min，max The calculation spectrum correction formula M4 is as follows:

[0036]

[0037] Among them, C is the actual temperature value of the sample to be corrected, RH is the actual humidity value of the sample to be corrected, T1, T2, T3, and T4 are the average correction values ​​of the light intensity values ​​of the spectrum to be corrected under different temperature and humidity conditions.

[0038] Furthermore, when C x ≠C mean And RH x ≠RH mean When the temperature of the sample to be corrected collected by the near infrared spectrometer is C x , humidity is RH x , the light intensity value at the wavelength point is P x =(P1 ' , P2 ' ..., P n '), the specific method for correcting the light intensity value of the sample wavelength point to be corrected described in step 5:

[0039] When C mean <C x ≤C max , RH mean <RH x ≤RH max When the temperature C x, humidity RH x Substitute the spectrum correction formula M1 to calculate the mean wavelength point light intensity value correction value T1 under the temperature and humidity conditions, and use this correction value to correct the actual wavelength point light intensity value P of the sample. x The light intensity values ​​at each wavelength range are corrected one by one to obtain the corrected wavelength point light intensity value P x1 =(P1′+T1, P2′+T1…,P n ′+T1);

[0040] When C min ≤C x <C mean , RH min ≤RH x <RH mean When the temperature C x , humidity RH x Substitute the spectrum correction formula M2 to calculate the mean wavelength point light intensity value correction value T2 under the temperature and humidity conditions, and use this correction value to correct the actual wavelength point light intensity value P of the sample. x The light intensity values ​​at each wavelength range are corrected one by one to obtain the corrected wavelength point light intensity value P x2 =(P1′+T2, P2′+T2…,P n ′+T2);

[0041] When C mean <C x ≤C max , RH min ≤RH x <RH mean When the temperature C x , humidity RH x Substitute the spectrum correction formula M3 to calculate the mean correction value T3 of the wavelength point light intensity value under the temperature and humidity conditions, and use this correction value to correct the actual wavelength point light intensity value P of the sample. x The light intensity values ​​at each wavelength range are corrected one by one to obtain the corrected wavelength point light intensity value P x3 =(P1′+T3, P2′+T3……,P n ′+T3);

[0042] When C min ≤C x <C mean , RH mean <RH x ≤RH max When the temperature C x , humidity RH x Substitute the spectrum correction formula M4 to calculate the mean correction value T4 of the wavelength point light intensity value under the temperature and humidity conditions, and use this correction value to correct the actual wavelength point light intensity value P of the sample.x The light intensity values ​​at each wavelength range are corrected one by one to obtain the corrected wavelength point light intensity value P x4 =(P1′+T4, P2′+T4……,P n ′+T4).

[0043] Furthermore, when C x =C mean or RH x =RH mean When the temperature of the sample to be corrected collected by the near infrared spectrometer is C x , humidity is RH x , the light intensity value at each wavelength point is P x =(P1′, P2′, P n '), at this time, the wavelength point light intensity value of the sample to be corrected only needs to be corrected for temperature or humidity. The specific method for correcting the wavelength point light intensity value of the sample to be corrected in step 5 is:

[0044] At a temperature of C x =C mean And humidity RH x ≠RH mean When the humidity RH of the sample to be corrected is x RH mean <RH x ≤RH max , RH min ≤RH x <RH mean When the wavelength point light intensity mean correction values ​​T5 and T6 are:

[0045]

[0046] This correction value is used to calculate the light intensity value P at the actual wavelength point of the sample to be tested. x The light intensity values ​​at each wavelength range are corrected one by one to obtain the corrected wavelength point light intensity value P x5 =(P1′+T5, P2′+T5……,P n ′+T5) and P x6 =(P1′+T6, P2′+T6……,P n ′+T6);

[0047] At humidity RH x =RH mean And the temperature is C x ≠C mean When the temperature of the sample to be corrected is C, a single correction is performed in combination with the light intensity value at the temperature wavelength point. x Belong to C mean <Cx ≤C max , C min ≤C x <C mean When the wavelength point light intensity mean correction values ​​T7 and T8 are calculated, they are:

[0048]

[0049] Use this correction value to correct the actual wavelength point light intensity value P of the sample x The light intensity values ​​at each wavelength range are corrected one by one to obtain the corrected wavelength point light intensity value P x7 =(P1′+T7, P2′+T7……,P n ′+T7) and P x8 =(P1′+T8, P2′+T8……,P n ′+T8).

[0050] Furthermore, when the temperature C x =C mean And humidity RH x =RH mean When the temperature of the sample to be corrected collected by the near infrared spectrometer is C x , humidity is RH x , the light intensity value at each wavelength is P x =(P1′, P2′, P n ′), the light intensity value of each wavelength point at this time is taken as the most preferred spectral data, and no spectral data correction is performed.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] The method for collaboratively correcting the near-infrared wavelength point light intensity value by temperature and humidity described in the present invention not only corrects the near-infrared spectral data from the two dimensions of temperature and humidity, but also calculates the temperature and humidity spectral correction coefficient through the spectral data of samples with different temperatures and humidity, and characterizes the influence of the combined effect of temperature and humidity on the sample through the spectral correction coefficient, which can effectively improve the accuracy and stability of the spectral data. After correcting the spectral data through this method, the problem of poor consistency of spectral samples under different environmental conditions can be effectively solved, thereby improving the prediction ability of the sample spectral model. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a schematic diagram of the steps of a method for collaboratively correcting near-infrared spectral data using temperature and humidity according to the present invention. DETAILED DESCRIPTION

[0054] like Figure 1As shown, the method for collaboratively correcting near-infrared spectral data by temperature and humidity described in the present invention first uses a miniaturized near-infrared spectrometer to collect spectral data of wine lees samples with different temperatures and humidity, and calculates the mean of the spectral data of wine lees samples with different temperatures and humidity, then calculates the temperature and humidity spectrum correction coefficient in combination with the temperature and humidity values ​​of the wine lees samples and the corresponding spectral mean, and calculates the temperature and humidity spectrum correction formula in combination with the coefficient, and finally corrects the spectral data of the wine lees sample to be tested in combination with the temperature and humidity spectrum correction formula to complete the correction of the spectral data.

[0055] Example:

[0056] This embodiment specifically introduces the method of coordinating near-infrared spectral data with temperature and humidity according to the present invention in combination with the detection of wine lees samples.

[0057] In this embodiment, a wavelength-averaged miniaturized near-infrared spectrometer is used to collect spectral data of the wine lees sample. The wavelength range of the collected spectral data is 1300nm to 1800nm, and the resolution is 10nm. The spectral data of each wine lees sample is actually composed of light intensity values ​​at 51 wavelength points. During the collection process, multiple spectral data are collected for each wine lees sample, and the multiple spectral data are averaged. The averaged data is the actual spectral data of the sample.

[0058] If the highest temperature of the lees is set to C max , the lowest temperature point is C min , we can calculate its mean temperature point as C mean =(C max +C min ) / 2, and similarly, the highest humidity point of the lees is set to RH max , the lowest humidity point is RH min , the mean humidity point can be calculated as RH mean =(RH max +RH min ) / 2. The highest, lowest and average temperature and humidity values ​​are taken as the correction reference points, and the temperature and humidity are collected at (C max , RH max )、(C max , RH mean )、(C max , RH min )、(C mean , RH max )、(C mean , RH mean )、(C mean , RH min )、(C min , RH max )、(C min , RH mean )、(Cmin , RH min ) Near-infrared spectral data under nine reference conditions are obtained, and these spectral data are used as spectral correction reference data.

[0059] In this embodiment, the spectrum data of the wine lees sample is set to P = (P1, P2..., P 51 ), where P1 is the light intensity at a wavelength of 1300 nm, P2 is the light intensity at a wavelength of 1310 nm, and so on. The spectral data of the wine lees sample is averaged to obtain the mean value P of the spectral data. mean For: P mean =(P1+P2+……P 51 ) / 51. For temperature and humidity (C max , RH max )、(C max , RH mean )、(C max , RH min )、(C mean , RH max )、(C mean , RH mean )、(C mean , RH min )、(C min , RH max )、(C min , RH mean )、(C min , RH min ) The near-infrared spectrum correction benchmark data under nine benchmark conditions were averaged to obtain the mean spectral intensity values ​​(P max，max , P max，mean , P max，min , P mean，max , P mean，mean , P mean，min , P min，max , P min，mean , P min，min ).

[0060] The temperature and humidity spectrum correction coefficient is calculated using the three-segment balanced difference method. The three-segment balanced data correspond to the three-segment temperature reference point data and the three-segment humidity reference point data, namely, the three-segment spectrum data corresponding to the maximum temperature point, the mean temperature point, and the minimum temperature point, and the three-segment spectrum data corresponding to the maximum humidity point, the mean humidity point, and the minimum humidity point. The specific calculation method is:

[0061] When calculating the temperature and humidity spectrum correction coefficient, keep the humidity at the mean point, and obtain the mean of the three spectral data corresponding to the maximum temperature point, the mean temperature point, and the minimum temperature point, that is, (C max , RHmean )、(C mean , RH mean )、(C min , RH mean ) corresponds to the spectral mean (P max，mean , P mean，mean , P min，mean ), with P mean，mean As the center point, max，mean , P min，mean The deviation values ​​of the two spectral data are calculated by the mean, and the deviation values ​​are:

[0062]

[0063] Among them A max With A min P max，mean , P min，mean The mean of the two spectral data is P mean，mean Deviation value of the center point.

[0064] Similarly, keep the temperature at the mean point and obtain the mean of the three spectral data corresponding to the maximum humidity point, the mean humidity point, and the minimum humidity point, that is, (C mean , RH max )、(C mean , RH mean )、(C mean , RH min ) corresponds to the spectral mean (P mean，max , P mean，mean , P mean，min ), also with P mean，mean As the center point, mean，max , P mean，min The deviation values ​​of the two spectral data are calculated by the mean, and the deviation values ​​are:

[0065]

[0066] Among them B max With B min P mean，max , P mean，min The mean of the two spectral data is P mean，mean Deviation value of the center point.

[0067] Since the combined effects of temperature and humidity are considered when performing spectrum correction on the lees sample to be tested, it is necessary to obtain the extreme point under the combined effects of temperature and humidity, that is, (C max , RH max )、(C max , RH min )、(C min , RHmax )、(C min , RH min ) corresponds to the spectral mean (P max，max , P max，min , P min，max , P min，min ), also with P mean，mean As the center point, the deviation values ​​of the mean values ​​of the four spectral data are calculated, and the deviation values ​​are:

[0068]

[0069] Where D max，max , D max，min , D min，max With D min，min P max，max , P max，min , P min，max , P min，min The average value of four spectral data is P mean，mean Deviation value of the center point.

[0070] The deviation value A calculated by fixing the temperature or humidity at the mean point max , A min , B max , B min They respectively reflect the influence of a single factor (temperature or humidity) on the spectral data of the lees sample. It is necessary to convert the single factor into the influence under the combined action of temperature and humidity, that is, it is necessary to calculate the corresponding temperature and humidity spectrum correction coefficient for the numerical conversion between the two. From the above, the temperature spectrum correction coefficient (α1, α2, α3, α4) is calculated as:

[0071]

[0072] Wherein, C is the actual temperature value of the lees sample to be corrected, and RH is the actual humidity value of the lees sample to be corrected.

[0073] Take the mean temperature and humidity point (C mean , RH mean ) corresponds to the mean value of the spectral data P mean，mean As the central reference point, the temperature and humidity spectrum correction formula is calculated by combining the spectrum correction coefficient and the temperature and humidity extreme points.

[0074] When C mean <C≤C max , RH mean <RH≤RH max When the temperature and humidity extreme point (C max , RH max ) and its corresponding spectral data mean P max，max The calculation spectrum correction formula M1 is as follows:

[0075]

[0076] Similarly, when C min ≤C <C mean , RH min ≤RH <RH mean When the temperature and humidity extreme point (C min , RH min ) and its corresponding spectral data mean P min，min The calculation spectrum correction formula M2 is as follows:

[0077]

[0078] Similarly, when C mean <C≤C max , RH min ≤RH <RH mean When the temperature and humidity extreme point (C max , RH min ) and its corresponding spectral data mean P max，min The calculation spectrum correction formula M3 is as follows:

[0079]

[0080] Similarly, when C min ≤C <C mean , RH mean <RH≤RH max When the temperature and humidity extreme point (C min , RH max ) and its corresponding spectral data mean P min，max The calculation spectrum correction formula M4 is as follows:

[0081]

[0082] Among them, C is the actual temperature value of the sample to be corrected, RH is the actual humidity value of the sample to be corrected, and T1, T2, T3, and T4 are the average correction values ​​of the light intensity values ​​under different temperature and humidity conditions of the spectrum to be corrected.

[0083] Combined with the temperature and humidity spectrum correction formula, the spectrum data of the distiller's grains sample to be tested is corrected to complete the data correction:

[0084] In this embodiment, if the temperature of the wine lees sample to be tested collected by the miniaturized near-infrared spectrometer is C x , humidity is RH x , the spectral data is P x =(P1′, P2′, P 51 ′), when C mean <Cx ≤C max , RH mean <RH x ≤RH max When the temperature is C x , humidity is RH x Substitute the spectrum correction formula M1 to calculate the mean correction value T1 of the spectrum data under the temperature and humidity conditions, and use this correction value to correct the actual spectrum data P of the wine lees sample. x The light intensity values ​​in each wavelength range are corrected one by one to obtain the corrected spectral data P x1 =(P1′+T1, P2′+T1…,P 51 ′+T1);

[0085] Similarly, when C min ≤C x <C mean , RH min ≤RH x <RH mean When the temperature is C x , humidity is RH x Substitute the spectrum correction formula M2 to calculate the mean correction value T2 of the spectrum data under the temperature and humidity conditions, and use this correction value to correct the actual spectrum data P of the wine lees sample. x The light intensity values ​​in each wavelength range are corrected one by one to obtain the corrected spectral data P x2 =(P1′+T2, P2′+T2…,P 51 ′+T2);

[0086] Similarly, when C mean <C x ≤C max , RH min ≤RH x <RH mean When the temperature is C x , humidity is RH x Substitute the spectrum correction formula M3 to calculate the mean correction value T3 of the spectrum data under the temperature and humidity conditions, and use this correction value to correct the actual spectrum data P of the wine lees sample. x The light intensity values ​​in each wavelength range are corrected one by one to obtain the corrected spectral data

[0087] P x3 =(P1′+T3, P2′+T3……,P 51 ′+T3);

[0088] Similarly, when C min ≤C x <C mean , RH mean <RHx ≤RH max When the temperature is C x , humidity is RH x Substitute the spectrum correction formula M4 to calculate the mean correction value T4 of the spectrum data under the temperature and humidity conditions, and use this correction value to correct the actual spectrum data P of the wine lees sample. x The light intensity values ​​in each wavelength range are corrected one by one to obtain the corrected spectral data

[0089] P x4 =(P1′+T4, P2′+T4……,P 51 ′+T4);

[0090] In addition to the above four cases, when there is C x =C mean or RH x =RH mean In this case, the spectral data of the wine lees sample to be corrected only needs a single correction of temperature or humidity, and there is no need to calculate the spectral correction coefficient, and the conventional linear correction method can be used to correct it.

[0091] At a temperature of C x =C mean And humidity RH x ≠RH mean When the humidity spectrum data is combined with the single correction, the spectrum data mean correction values ​​T5 and T6 are calculated as follows:

[0092]

[0093] The correction value is used to calculate the actual spectral data P of the wine lees sample to be tested. x The light intensity values ​​in each wavelength range are corrected one by one to obtain the corrected spectral data P x5 =(P1′+T5, P2′+T5……,P 51 ′+T5) and P x6 =(P1′+T6, P2′+T6……,P 51 ′+T6);

[0094] Similarly, when the humidity is RH x =RH mean And the temperature is C x ≠C mean When the temperature spectrum data is combined with the single correction, the spectrum data mean correction values ​​T7 and T8 are calculated as:

[0095]

[0096] The correction value is used to calculate the actual spectral data P of the wine lees sample to be tested. xThe light intensity values ​​in each wavelength range are corrected one by one to obtain the corrected spectral data P x7 =(P1′+T7, P2′+T7……,P 51 ′+T7) and P x8 =(P1′+T8, P2′+T8……,P 51 ′+T8);

[0097] At temperature C x =C mean And humidity RH x =RH mean In this case, there is no need to correct the spectral data of the wine lees sample to be corrected under this condition.

Claims

1. A method for coordinating near-infrared spectral data with temperature and humidity, characterized in that: The following steps are involved: Step 1: Use a near-infrared spectrometer to collect spectral data of samples at different temperatures and humidities, and use it as benchmark correction data; Step 2: Calculate the average light intensity value of the sample spectrum under different temperatures and humidities collected in step 1; Step 3: Calculate the temperature and humidity spectrum correction coefficient by combining the sample temperature and humidity values ​​and the corresponding light intensity average value; Step 4: Calculate the temperature and humidity spectrum correction formula according to the temperature and humidity spectrum correction coefficient; Step 5: Correct the light intensity value of the wavelength point of the sample to be corrected by combining the temperature and humidity spectrum correction formula; The specific step 1 is as follows: the temperature and humidity values ​​of the highest point, the lowest point, and the mean point of the temperature and humidity of the sample are used as correction reference points, and the highest temperature point of the sample is C max , the lowest temperature point is C min , the mean temperature point is C mean =(C max +C min ) / 2, the highest humidity point of the sample is RH max , the lowest humidity point is RH min , the mean humidity point is RH mean =(RH max +RH min ) / 2; respectively collect (C max , RH max )、(C max , RH mean )、(C max , RH min )、(C mean , RH max )、(C mean , RH mean )、(C mean , RH min )、(C min , RH max )、(C min , RH mean )、(C min , RH min ) Spectral data under nine reference conditions as reference data for spectrum correction; The implementation method of step 2 is as follows: the spectrum data of each state under the nine reference states in step 1 is collected multiple times, and the average calculation is performed, and the averaged data is used as the actual spectrum data of the sample, and then the number of wavelength points included in the sample spectrum under the nine reference states is calculated according to the selected wavelength range and the resolution of the near-infrared spectrometer, and the light intensity value (P1, P2..., P n ), the light intensity value of the sample is averaged to obtain the light intensity mean value P of the spectrum mean =(P1+P2+……P n ) / n, the average light intensity values ​​under the nine reference conditions in step 1 are (P max，max , P max，mean , P max，min , P mean，max , P mean，mean , P mean，min , P min，max , P min，mean , P min，min ); The specific method for calculating the temperature and humidity spectrum correction coefficient in step 3 is: keep the humidity at the mean point, obtain the maximum temperature point (C max , RH mean ), mean temperature point (C mean , RH mean )、Minimum temperature point (C min , RH mean ) corresponds to the mean light intensity value (P max，mean , P mean，mean , P min，mean ), with P mean，mean As the center, calculate P max，mean , P min，mean The average of the two light intensity values ​​and P mean，mean The deviation of the center point is: Among them A max With A min P max，mean , P min，mean The average value of the light intensity at two wavelengths is P mean，mean Deviation value of the center point; Keep the temperature at the mean point and obtain the maximum humidity point (C mean , RH max ), mean humidity point (C mean , RH mean )、Minimum humidity point (C mean , RH min ) corresponds to the mean light intensity value (P mean，max , P mean，mean , P mean，min ), with P mean，mean As the center point, mean，max , P mean，min The deviation values ​​of the two light intensity values ​​are calculated by the average value, and the deviation values ​​are: Among them B max With B min P mean，max , P mean，min The average value of the light intensity at two wavelengths is P mean，mean Deviation value of the center point; P mean，mean As the center point, obtain all extreme points and P under the combined effect of temperature and humidity mean，mean Deviation of light intensity value: Where D max，max , D max，min , D min，max With D min，min P max，max , P max，min , P min，max , P min，min The average value of the light intensity at four wavelengths and P mean，mean Deviation value of the center point; Calculate the temperature and humidity spectrum correction coefficients α1, α2, α3, and α4 based on the deviation values ​​obtained above: Wherein, C is the actual temperature value of the sample to be corrected, and RH is the actual humidity value of the sample to be corrected; Method for calculating the temperature and humidity spectrum correction formula described in step 4: When C mean <C≤C max , RH mean <RH≤RH max When the temperature and humidity extreme point (C max , RH max ) and the corresponding wavelength point light intensity value P max，max The calculation spectrum correction formula M1 is as follows: When C min ≤C <C mean , RH min ≤RH <RH mean When the temperature and humidity extreme point (C min , RH min ) and the corresponding wavelength point light intensity value mean P min，min The calculation spectrum correction formula M2 is as follows: When C mean <C≤C max , RH min ≤RH <RH mean When the temperature and humidity extreme point (C max , RH min ) and the corresponding wavelength point light intensity value mean P max，min The calculation spectrum correction formula M3 is as follows: When C min ≤C <C mean , RH mean <RH≤RH max When the temperature and humidity extreme point (C min , RH max ) and the corresponding wavelength point light intensity value P min，max The calculation spectrum correction formula M4 is as follows: Among them, C is the actual temperature value of the sample to be corrected, RH is the actual humidity value of the sample to be corrected, T1, T2, T3, and T4 are the average correction values ​​of the light intensity values ​​of the spectrum to be corrected under different temperature and humidity conditions.

2. The method for coordinating near-infrared spectral data with temperature and humidity according to claim 1, characterized in that: When C x ≠C mean And RH x ≠RH mean When the temperature of the sample to be corrected collected by the near infrared spectrometer is C x , humidity is RH x , the light intensity value at the wavelength point is P x =(P1', P2'..., P n '), the specific method for correcting the light intensity value of the sample wavelength point to be corrected described in step 5: When C mean <C x ≤C max , RH mean <RH x ≤RH max When the temperature C x , humidity RH x Substitute the spectrum correction formula M1 to calculate the mean wavelength point light intensity value correction value T1 under the temperature and humidity conditions, and use this correction value to correct the actual wavelength point light intensity value P of the sample. x The light intensity values ​​at each wavelength range are corrected one by one to obtain the corrected wavelength point light intensity value P x1 =(P1′+T1, P2′+T1…,P n ′+T1); When C min ≤C x <C mean , RH min ≤RH x <RH mean When the temperature C x , humidity RH x Substitute the spectrum correction formula M2 to calculate the mean wavelength point light intensity value correction value T2 under the temperature and humidity conditions, and use this correction value to correct the actual wavelength point light intensity value P of the sample. x The light intensity values ​​at each wavelength range are corrected one by one to obtain the corrected wavelength point light intensity value P x2 =(P1′+T2, P2′+T2…,P n ′+T2); When C mean <C x ≤C max , RH min ≤RH x <RH mean When the temperature C x , humidity RH x Substitute the spectrum correction formula M3 to calculate the mean correction value T3 of the wavelength point light intensity value under the temperature and humidity conditions, and use this correction value to correct the actual wavelength point light intensity value P of the sample. x The light intensity values ​​at each wavelength range are corrected one by one to obtain the corrected wavelength point light intensity value P x3 =(P1′+T3, P2′+T3…,P n ′+T3); When C min ≤C x <C mean , RH mean <RH x ≤RH max When the temperature C x , humidity RH x Substitute the spectrum correction formula M4 to calculate the mean correction value T4 of the wavelength point light intensity value under the temperature and humidity conditions, and use this correction value to correct the actual wavelength point light intensity value P of the sample. x The light intensity values ​​at each wavelength range are corrected one by one to obtain the corrected wavelength point light intensity value P x4 =(P1′+T4, P2′+T4……,P n ′+T4).

3. The method for coordinating near-infrared spectral data with temperature and humidity according to claim 1, characterized in that: When C x =C mean or RH x =RH mean When the temperature of the sample to be corrected collected by the near infrared spectrometer is C x , humidity is RH x , the light intensity value at each wavelength point is P x =(P1', P2'..., P n '), at this time, the wavelength point light intensity value of the sample to be corrected only needs to be corrected for temperature or humidity. The specific method for correcting the wavelength point light intensity value of the sample to be corrected in step 5 is: At a temperature of C x =C mean And humidity RH x ≠RH mean When the humidity RH of the sample to be corrected is x RH mean <RH x ≤RH max , RH min ≤RH x <RH mean When the wavelength point light intensity mean correction values ​​T5 and T6 are: This correction value is used to calculate the light intensity value P at the actual wavelength point of the sample to be tested. x The light intensity values ​​at each wavelength range are corrected one by one to obtain the corrected wavelength point light intensity value P x5 =(P1'+T5, P2'+T5..., P n '+T5) and P x6 =(P1'+T6,P2'+T6...,P n ′+T6); At humidity RH x =RH mean And the temperature is C x ≠C mean When the temperature of the sample to be corrected is C, a single correction is performed in combination with the light intensity value at the temperature wavelength point. x Belong to C mean <C x ≤C max , C min ≤C x <C mean When the wavelength point light intensity mean correction values ​​T7 and T8 are calculated, they are: Use this correction value to correct the actual wavelength point light intensity value P of the sample x The light intensity values ​​at each wavelength range are corrected one by one to obtain the corrected wavelength point light intensity value P x7 =(P1'+T7, P2'+T7..., P n '+T7) and P x8 =(P1'+T8,P2'+T8...,P n ′+T8).

4. The method for coordinating near-infrared spectral data with temperature and humidity according to claim 1, characterized in that: When the temperature C x =C mean And humidity RH x =RH mean When the temperature of the sample to be corrected collected by the near infrared spectrometer is C x , humidity is RH x , the light intensity value at each wavelength is P x =(P1', P2'..., P n '), the light intensity value of each wavelength point at this time is taken as the most preferred spectral data and no correction is made.

Images