Standardized method for determining digestibility of whole-plant corn silage for near-infrared modeling

Abstract

The present invention relates to the technical field of infrared spectral data dimensionality reduction, and proposes a standardized determination method for whole-plant corn silage digestibility for near-infrared modeling, including: extracting silage samples, determining the suspected absorption peak points on the near-infrared spectrum curve of the silage samples; determining the absorbance and wavelength of the suspected absorption peak points and characteristic absorption peaks of the target absorption peaks, determining the absorption peak possibility of the suspected absorption peak points of the target absorption peaks, and determining the absorption peaks; marking the target absorption peaks, determining the dimensionality reduction weight influence coefficient of the target absorption peaks; establishing a NIR model according to the dimensionality reduction weight influence coefficients of all absorption peaks of the near-infrared spectrum curves of all silage samples, and using the NIR model to achieve the standardized determination of the digestibility of whole-plant corn silage. The present invention aims to improve the accuracy of the results of the standardized determination of the digestibility of whole-plant corn silage.

Description

Technical Field

[0001] The invention relates to the technical field of infrared spectrum data dimensionality reduction, and in particular to a standardized determination method for whole-plant corn silage digestibility for near-infrared modeling. Background Art

[0002] NIR near infrared spectroscopy is a fast, non-destructive analytical technique that can quickly determine the chemical composition of a sample by analyzing its optical properties in the near infrared spectral region. Near infrared modeling can achieve standardized determination of the digestibility of whole corn silage. Specifically, samples are extracted during the whole corn silage process, and the samples are tested using near infrared spectroscopy. A NIR model for predicting digestibility is established based on the test results. Based on the NIR model, rapid and standardized determination of the digestibility of whole corn silage samples can be achieved. Digestibility refers to the proportion of organic matter or nutrients that an animal can effectively digest and absorb.

[0003] When using near-infrared spectroscopy to detect samples and determine the main components of silage corn, it will be affected by the heterogeneity of the whole-plant corn silage in the silage barrel. That is to say, different parts of the corn stems, leaves, ears, etc. cannot be completely mixed evenly at different positions in the silage barrel. Therefore, the fermentation conditions at different positions in the silage barrel may be different, resulting in different near-infrared spectral test results obtained from samples collected at different positions, affecting the accuracy of the established NIR model, and further leading to inaccurate standardized measurement results of the digestibility of whole-plant corn silage. Summary of the invention

[0004] The present invention provides a method for standardizing the digestibility of whole-plant corn silage for near infrared modeling, so as to solve the problem that the mixing of different parts of corn makes the established NIR model inaccurate, resulting in inaccurate standardized measurement results of the digestibility of whole-plant corn silage. The technical scheme adopted is as follows:

[0005] One embodiment of the present invention provides a standardized method for determining whole-plant corn silage digestibility for near infrared modeling, the method comprising the following steps:

[0006] Extract a preset number of silage samples, obtain a near-infrared spectrum curve of each silage sample, and determine the suspected absorption peak point on the near-infrared spectrum curve, where each point on the near-infrared spectrum curve corresponds to an absorbance and a wavelength;

[0007] Determine the absorbance and wavelength of the characteristic absorption peak of each main chemical component of the whole corn silage, record any suspected absorption peak point as the target suspected absorption peak point, and determine the absorption peak possibility of the target suspected absorption peak point based on the difference between the absorbance of different suspected absorption peak points in the vicinity of the target suspected absorption peak point and the absorbance of the characteristic absorption peak, and the difference between the wavelength of the target suspected absorption peak point and two adjacent suspected absorption peak points, and determine the absorption peak based on the absorption peak possibility of the target suspected absorption peak point;

[0008] Any absorption peak is recorded as the target absorption peak, and the dimensionality reduction weight influence coefficient of the target absorption peak is determined according to the difference between the wavelength of the characteristic absorption peak of the main chemical component corresponding to the absorption peak possibility and the wavelength of the target absorption peak, and the difference between the wavelengths of the absorption peaks of the near-infrared spectrum curves of all silage samples corresponding to the same real characteristic absorption peak as the target absorption peak;

[0009] A NIR model was established based on the dimensionality reduction weight influence coefficients of all absorption peaks in the near-infrared spectral curves of all silage samples, and the NIR model was used to achieve standardized determination of whole-plant corn silage digestibility.

[0010] Furthermore, the specific method of determining the suspected absorption peak point on the near-infrared spectrum curve includes:

[0011] Obtain the minimum period value of the near-infrared spectrum curve of the silage sample, take half of the minimum period value as the sampling time interval, start sampling from the minimum value of the horizontal coordinate of the near-infrared spectrum curve, and obtain the sampling point;

[0012] All sampling points whose slopes are within a range of greater than or equal to -t and less than or equal to t are marked as suspected absorption peak points, where t represents a preset range threshold.

[0013] Further, the absorption peak possibility of the target absorption peak suspected point is determined according to the difference between the absorbance of different absorption peak suspected points in the vicinity of the target absorption peak suspected point and the absorbance of the characteristic absorption peak, and the difference between the wavelength of the target absorption peak suspected point and two adjacent absorption peak suspected points, including the specific method of:

[0014] Recording a first preset number of absorption peak suspected points adjacent to the target absorption peak suspected point as adjacent absorption peak suspected points of the target absorption peak suspected point;

[0015] The average value of the absorbance of the target absorption peak suspected point and all adjacent absorption peak suspected points is recorded as the average absorbance of the target absorption peak suspected point; the absolute value of the difference between the absorbance of the adjacent absorption peak suspected points of the target absorption peak suspected point and the average absorbance of the target absorption peak suspected point is recorded as the absorbance difference of the adjacent absorption peak suspected points; the average value of the absorbance difference between the target absorption peak suspected point and all adjacent absorption peak suspected points is recorded as the change speed of the target absorption peak suspected point;

[0016] The minimum value of the change speed of all the target absorption peak suspected points is used as the change speed correction value of the target absorption peak suspected points;

[0017] The absorption peak possibility of the target absorption peak suspected point is determined according to the difference between the wavelengths of the target absorption peak suspected point and two adjacent absorption peak suspected points and the change speed correction value of the target absorption peak suspected point.

[0018] Further, the absorption peak possibility of the target absorption peak suspected point is determined according to the difference between the wavelength of the target absorption peak suspected point and two adjacent absorption peak suspected points, and the change speed correction value of the target absorption peak suspected point, including the specific method of:

[0019] The sum of the absolute values ​​of the wavelength differences between the target absorption peak suspected point and two adjacent absorption peak suspected points is recorded as the wavelength difference of the target absorption peak suspected point;

[0020] The normalized value of the ratio of the wavelength difference of the target absorption peak suspected point to the change speed correction value is recorded as the absorption peak possibility of the target absorption peak suspected point.

[0021] Further, the specific method for determining the absorption peak is:

[0022] When the absorption peak possibility of the target absorption peak suspected point is greater than or equal to the absorption peak judgment threshold, the target absorption peak suspected point is recorded as the absorption peak.

[0023] Further, the dimensionality reduction weight influence coefficient of the target absorption peak is determined based on the difference between the wavelength of the characteristic absorption peak of the main chemical component corresponding to the absorption peak possibility and the wavelength of the target absorption peak, and the difference between the wavelengths of the absorption peaks of the near-infrared spectrum curves of all silage samples corresponding to the same real characteristic absorption peak as the target absorption peak. The specific method for obtaining the coefficient is as follows:

[0024] The characteristic absorption peak of the main chemical component corresponding to the correction value of the absorption peak change rate corresponding to the absorption peak possibility is recorded as the real characteristic absorption peak of the target absorption peak;

[0025] According to the difference between the real characteristic absorption peak of the target absorption peak and the wavelength of the target absorption peak, and the difference between the wavelengths of the absorption peaks of the near-infrared spectral curves of all silage samples corresponding to the same real characteristic absorption peak as the target absorption peak, the dimensionality reduction weight influence coefficient of the target absorption peak is determined.

[0026] Further, the dimensionality reduction weight influence coefficient of the target absorption peak is determined according to the difference between the real characteristic absorption peak of the target absorption peak and the wavelength of the target absorption peak, and the difference between the wavelengths of the absorption peaks of the near-infrared spectrum curves of all silage samples corresponding to the same real characteristic absorption peak as the target absorption peak, including the specific method of:

[0027] The absolute value of the difference between the wavelength of the target absorption peak and the wavelength of the true characteristic absorption peak is recorded as the silage component wavelength difference of the target absorption peak;

[0028] The dimensionality reduction weight influence coefficient of the target absorption peak is determined based on the difference between the wavelengths of the absorption peaks of the near-infrared spectral curves of all silage samples corresponding to the same real characteristic absorption peak as the target absorption peak, and the wavelength difference of the silage component of the target absorption peak.

[0029] Further, the dimensionality reduction weight influence coefficient of the target absorption peak is determined according to the difference between the wavelengths of the absorption peaks of the near-infrared spectrum curves of all silage samples corresponding to the same real characteristic absorption peak as the target absorption peak, and the wavelength difference of the silage component of the target absorption peak, including the specific method of:

[0030] The average value of the wavelengths of the absorption peaks of the near-infrared spectrum curves of all silage samples corresponding to the same real characteristic absorption peak as the target absorption peak is recorded as the average value of the wavelengths of the silage samples corresponding to the real characteristic absorption peak;

[0031] The absolute value of the difference between the wavelength of the target absorption peak and the mean wavelength of the silage sample corresponding to the real characteristic absorption peak of the target absorption peak is recorded as the wavelength deviation of the absorption peak;

[0032] The normalized value of the product of the wavelength difference of the silage component of the target absorption peak and the wavelength deviation is recorded as the first influence coefficient of the target absorption peak, and the difference between the number 1 and the first influence coefficient of the target absorption peak is recorded as the dimensionality reduction weight influence coefficient of the target absorption peak.

[0033] Furthermore, the NIR model is established based on the dimensionality reduction weight influence coefficients of all absorption peaks of the near infrared spectrum curves of all silage samples, including the specific method of:

[0034] The spectral data corresponding to all absorption peaks of the near-infrared spectral curves of all silage samples were processed using the PCA principal component analysis method to obtain the projection coefficient of the spectral data of each absorption peak on the principal component; the normalized value of the product of the dimensionality reduction weight influence coefficient of the absorption peak and the projection coefficient of the spectral data of the absorption peak on the principal component was recorded as the dimensionality reduction weight of the absorption peak;

[0035] The dimension reduction weight of the absorption peak is used as the projection coefficient of the spectral data of the absorption peak on the principal component, and the spectral data corresponding to all absorption peaks of the near-infrared spectral curves of all samples are processed using the PCA principal component analysis method to obtain the spectral data after dimension reduction;

[0036] The spectral data after dimension reduction were used as independent variables, and the digestibility of each silage sample was used as the dependent variable. The NIR model was established using the PLS partial least squares regression method.

[0037] Furthermore, the use of the NIR model to achieve standardized measurement of whole-plant corn silage digestibility includes the following specific methods:

[0038] For whole-plant corn silage that needs to be subjected to standardized digestibility determination, a silage sample to be tested of the whole-plant corn silage is collected, all absorption peaks of the near-infrared spectrum curve of the silage sample to be tested are obtained, and spectral data corresponding to all absorption peaks of the near-infrared spectrum curve of the silage sample to be tested are input into the NIR model to obtain the digestibility of the silage sample to be tested;

[0039] The digestibility of the silage sample to be tested is used as the digestibility of the whole corn silage that needs to be standardized for digestibility determination.

[0040] The beneficial effects of the present invention are:

[0041] Taking into account the different compositions of different parts of corn stems, leaves, ears, etc. contained in different silage samples, resulting in differences in near-infrared spectral curves, in order to establish a more accurate NIR model and obtain more accurate standardized measurement results of the digestibility of whole-plant corn silage, this application screens out the absorption peaks that affect the digestibility based on the near-infrared spectral curves of different samples, and determines the accurate information on the main chemical components corresponding to the whole-plant corn silage in the silage barrel. First, based on the characteristics of the slower the increase or decrease trend of the slopes of the suspected points of different absorption peaks in the vicinity of the real absorption peak on the near-infrared spectral curve of the silage sample, the absorption peak possibility of the suspected point of the target absorption peak is determined, and screening is performed based on the absorption peak possibility. All absorption peaks are identified; secondly, according to the characteristic that the wavelengths of the absorption peaks of the near-infrared spectral curves of all silage samples corresponding to the same true characteristic absorption peak are slightly different, the significance of the absorption peak in reflecting the main components in the whole-plant corn silage is evaluated, and the dimensionality reduction weight influence coefficient of the absorption peak is obtained; finally, according to the dimensionality reduction weight influence coefficient of all absorption peaks of the near-infrared spectral curves of all silage samples, a NIR model is established, and the NIR model is used to realize the standardized determination of the digestibility of the whole-plant corn silage, so as to solve the problem that the mixing of different parts of corn makes the established NIR model inaccurate, resulting in inaccurate standardized determination results of the digestibility of the whole-plant corn silage, and improve the accuracy of the standardized determination results of the digestibility of the whole-plant corn silage. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0043] Figure 1 A schematic diagram of a process flow of a standardized method for determining whole-plant corn silage digestibility for near infrared modeling provided by one embodiment of the present invention;

[0044] Figure 2 A flowchart for obtaining an absorption peak provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] See also Figure 1, which shows a flow chart of a standardized whole-plant corn silage digestibility determination method for near infrared modeling provided by an embodiment of the present invention, the method comprising the following steps:

[0047] Step S001, extract a preset number of silage samples, obtain the near-infrared spectrum curve of each silage sample, and determine the suspected absorption peak point on the near-infrared spectrum curve. Each point on the near-infrared spectrum curve corresponds to an absorbance and a wavelength. Extract a preset number of silage samples from the silage barrel of whole corn silage. The silage samples need to include different parts such as corn stems, leaves, and ears to ensure the representativeness and diversity of the silage samples. All silage samples are crushed and mixed separately to ensure the uniformity of the chemical composition of each silage sample. Select the transmission or reflection mode, place the silage sample in the sample chamber of the near-infrared spectrometer, use the near-infrared spectrometer to detect the silage sample, and obtain the near-infrared spectrum curve of the silage sample.

[0048] The near-infrared spectrum curve can reflect the light absorption or reflection information of silage samples at different wavelengths. The horizontal axis of the near-infrared spectrum curve is wavelength, ranging from 780nm to 2500nm, and the vertical axis is absorbance. The absorbance indicates the degree of light absorption, and the value range is a constant greater than or equal to 0 and less than or equal to 1.

[0049] The near-infrared spectral curves of silage samples were smoothed, denoised and spectrally normalized.

[0050] Preferably, in one embodiment of the present application, the near-infrared spectrometer is a Fourier transform near-infrared spectrometer; the near-infrared spectrum curve of the silage sample is smoothed and denoised using a Savitzy-Golay filter to remove random noise in the near-infrared spectrum curve; the near-infrared spectrum curve of the silage sample is spectrally normalized using an SNV standard normal variable transformation or an MSC multi-scattering correction to eliminate the baseline drift problem caused by the scattering effect and unify the scale of the spectral data; the number of extracted silage samples is 20, that is, the preset number is 20. In actual application, as other implementation methods, the implementer may select other near-infrared spectrometers, denoising methods, spectral normalization methods, and preset number values, and this application does not impose any special restrictions.

[0051] The near infrared spectrum curve of the silage sample is Fourier transformed to obtain the minimum period value of the near infrared spectrum curve of the silage sample. According to the Nyquist sampling theorem, half of the minimum period value is used as the sampling time interval, and sampling is performed from the minimum value of the horizontal coordinate of the near infrared spectrum curve, 780nm, to obtain the sampling point.

[0052] All sampling points whose slopes are within the range of -t and t are marked as suspected absorption peak points, where t represents a preset range threshold, and the value of t in this embodiment is 0.01.

[0053] At this point, all the suspected absorption peaks on the near-infrared spectrum curve are determined.

[0054] Step S002, determine the absorbance and wavelength of the characteristic absorption peak of each main chemical component of the whole corn silage, record any one of the absorption peak suspected points as the target absorption peak suspected point, and determine the absorption peak possibility of the target absorption peak suspected point according to the difference between the absorbance of different absorption peak suspected points in the vicinity of the target absorption peak suspected point and the absorbance of the characteristic absorption peak, and the difference between the wavelength of the target absorption peak suspected point and the two adjacent absorption peak suspected points, and determine the absorption peak according to the absorption peak possibility of the target absorption peak suspected point. The composition of different parts such as corn stems, leaves, and ears contained in different silage samples is different, so there are differences in the specific chemical components of different silage samples, and there are also differences in the near-infrared spectral curves of different silage samples. When the NIR model for predicting digestibility is established according to the near-infrared spectral curve, it will be affected by the specific chemical components of the silage sample, resulting in inaccurate NIR model, and then the standardized measurement result of the digestibility of whole corn silage is inaccurate. Therefore, it is necessary to screen out the absorption peak that affects the digestibility according to the near-infrared spectral curves of different samples, and determine the accurate main chemical component information corresponding to the whole corn silage in the silage barrel.

[0055] The main chemical components of whole corn silage include starch, protein, crude fat, water-soluble carbohydrates, acid detergent fiber, neutral detergent fiber, acid detergent lignin, crude ash, calcium, phosphorus and lactic acid. The characteristic absorption peak corresponding to each main chemical component is obtained. The ordinate corresponding to the characteristic absorption peak in the near-infrared spectrum curve is absorbance, and the abscissa is wavelength.

[0056] When the suspected absorption peak point is the real absorption peak on the near-infrared spectrum curve of the silage sample, the slopes of different suspected absorption peak points in the vicinity of the suspected absorption peak point should show a trend of slowly increasing or slowly decreasing.

[0057] Any absorption peak suspected point is recorded as a target absorption peak suspected point. This embodiment is described by taking any absorption peak suspected point as an example.

[0058] The change speed of the target absorption peak suspected point is determined according to the difference between the absorbance of different absorption peak suspected points in the vicinity of the target absorption peak suspected point and the absorbance of the characteristic absorption peak.

[0059] Preferably, as an embodiment of the present application, the first preset number of absorption peak suspected points adjacent to the target absorption peak suspected point are recorded as adjacent absorption peak suspected points of the target absorption peak suspected point; the average of the absorbances of the target absorption peak suspected point and all adjacent absorption peak suspected points is recorded as the average absorbance of the target absorption peak suspected point; the absolute value of the difference between the absorbances of the adjacent absorption peak suspected points of the target absorption peak suspected point and the average absorbance of the target absorption peak suspected point is recorded as the absorbance difference of the adjacent absorption peak suspected points; the average of the absorbance differences between the target absorption peak suspected point and all adjacent absorption peak suspected points is recorded as the change rate of the target absorption peak suspected point.

[0060] Among them, the value of the first preset number in this embodiment is 100. It can be understood that the first preset number of absorption peak suspected points adjacent to the target absorption peak suspected point are the 50 absorption peak suspected points adjacent to the target absorption peak suspected point and the 50 absorption peak suspected points adjacent to the target absorption peak suspected point. When the difference between the absorbance of the target absorption peak suspected point and all adjacent absorption peak suspected points is smaller, the slope of different absorption peak suspected points in the vicinity of the target absorption peak suspected point shows a slower increase or decrease trend, the target absorption peak suspected point is more likely to be a real absorption peak on the near-infrared spectrum curve of the silage sample, and at this time, the change rate of the target absorption peak suspected point is slower.

[0061] It is understandable that there are multiple main chemical components of whole-plant corn silage, so there are also multiple characteristic absorption peaks of the main chemical components of whole-plant corn silage. For each characteristic absorption peak, the change rate of a target absorption peak suspected point can be calculated, and the minimum value of the change rate of all target absorption peak suspected points is used as the change rate correction value of the target absorption peak suspected point.

[0062] The changing speed of any suspected absorption peak point can be obtained in the same way.

[0063] When the difference between the wavelengths of the target absorption peak suspected point and the two adjacent absorption peak suspected points is greater, the target absorption peak suspected point presents more significant characteristics of a non-single noise absorption peak, and the probability that the target absorption peak suspected point corresponds to a characteristic absorption peak is greater.

[0064] The absorption peak possibility of the target absorption peak suspected point is determined according to the difference between the wavelengths of the target absorption peak suspected point and two adjacent absorption peak suspected points and the change speed correction value of the target absorption peak suspected point.

[0065] Preferably, as an embodiment of the present application, the sum of the absolute values ​​of the differences in wavelength between the target absorption peak suspected point and the two adjacent absorption peak suspected points is recorded as the wavelength difference of the target absorption peak suspected point; the normalized value of the ratio of the wavelength difference of the target absorption peak suspected point to the change rate correction value is recorded as the absorption peak possibility of the target absorption peak suspected point.

[0066] The greater the difference between the wavelength of the target absorption peak suspected point and the two adjacent absorption peak suspected points, and the smaller the change speed correction value of the target absorption peak suspected point, the greater the possibility of the absorption peak of the target absorption peak suspected point. At this time, the possibility that the target absorption peak suspected point is the real absorption peak on the near-infrared spectrum curve of the silage sample is greater.

[0067] It should be noted that this embodiment uses the Z-Score standard normalization method to calculate the normalized value. In actual application, the implementer may use other methods in the prior art such as the maximum and minimum value normalization method, the sigmoid function, etc. to calculate the normalized value, which is not limited here.

[0068] When the absorption peak possibility of the target absorption peak suspected point is greater than or equal to the absorption peak judgment threshold, the target absorption peak suspected point is recorded as the absorption peak.

[0069] In this embodiment, the value of the absorption peak judgment threshold is 0.9.

[0070] The same method can be used to obtain the absorption peak possibility of any absorption peak suspected point and screen out all absorption peaks.

[0071] At this point, all absorption peaks on the near-infrared spectrum curve are determined. The absorption peak acquisition flow chart is as follows: Figure 2 shown.

[0072] Step S003, record any absorption peak as the target absorption peak, and determine the dimensionality reduction weight influence coefficient of the target absorption peak based on the difference between the wavelength of the characteristic absorption peak of the main chemical component corresponding to the possibility of the absorption peak and the wavelength of the target absorption peak, and the difference between the wavelengths of the absorption peaks of the near-infrared spectral curves of all silage samples corresponding to the same real characteristic absorption peak as the target absorption peak.

[0073] The characteristic absorption peak corresponding to the correction value of the change rate of the absorption peak is recorded as the real characteristic absorption peak of the absorption peak.

[0074] According to the difference between the wavelength of the absorption peak and the real characteristic absorption peak, the wavelength difference of the silage component of the absorption peak is determined.

[0075] Preferably, as an embodiment of the present application, the absolute value of the difference between the wavelength of the absorption peak and the wavelength of the true characteristic absorption peak is recorded as the silage component wavelength difference of the absorption peak.

[0076] The smaller the difference between the wavelength of the absorption peak and the true characteristic absorption peak, the closer the wavelength of the absorption peak is to the wavelength of the true characteristic absorption peak, and the more the absorption peak can reflect the characteristics of the main components in the whole corn silage. When the NIR model is determined based on the near-infrared spectrum curve of the silage sample, more attention should be paid to the information on the main components of silage provided by the absorption peak. At this time, the smaller the wavelength difference of the silage component of the absorption peak is.

[0077] Any absorption peak is recorded as a target absorption peak. This embodiment is described by taking any absorption peak as an example.

[0078] The dimensionality reduction weight influence coefficient of the target absorption peak is determined based on the difference between the wavelengths of the absorption peaks of the near-infrared spectral curves of all silage samples corresponding to the same real characteristic absorption peak as the target absorption peak, and the wavelength difference of the silage component of the target absorption peak.

[0079] Preferably, as an embodiment of the present application, the average value of the wavelength of the absorption peak of the near-infrared spectrum curves of all silage samples corresponding to the same real characteristic absorption peak as the target absorption peak is recorded as the average wavelength of the silage samples corresponding to the real characteristic absorption peak. The absolute value of the difference between the wavelength of the target absorption peak and the average wavelength of the silage samples corresponding to the real characteristic absorption peak of the target absorption peak is recorded as the wavelength deviation of the absorption peak. The normalized value of the product of the wavelength difference of the silage component of the target absorption peak and the wavelength deviation is recorded as the first influence coefficient of the target absorption peak, and the difference between the number 1 and the first influence coefficient of the target absorption peak is recorded as the dimensionality reduction weight influence coefficient of the target absorption peak.

[0080] It should be noted that this embodiment uses the Z-Score standard normalization method to calculate the normalized value. In actual application, the implementer may use other methods in the prior art such as the maximum and minimum value normalization method, the sigmoid function, etc. to calculate the normalized value, which is not limited here.

[0081] It can be understood that, for each absorption peak, the characteristic absorption peak corresponding to the correction value of the change rate of the absorption peak may correspond to the characteristic absorption peak of any one of the main chemical components of the whole-plant corn silage, and the main chemical components of the whole-plant corn silage include starch, protein, crude fat, water-soluble carbohydrates, acid detergent fiber, neutral detergent fiber, acid detergent lignin, crude ash, calcium, phosphorus and lactic acid. When the characteristic absorption peaks corresponding to the correction values ​​of the change rates of two absorption peaks are the main chemical components of the same whole-plant corn silage, the two absorption peaks are recorded as corresponding to the same real characteristic absorption peak. For example: when the characteristic absorption peaks corresponding to the correction values ​​of the change rates of two absorption peaks are both characteristic absorption peaks of acid detergent fiber, it is considered that the two absorption peaks are recorded as corresponding to the same real characteristic absorption peak. When the difference between the wavelengths of the absorption peaks of the near-infrared spectral curves of all silage samples corresponding to the same real characteristic absorption peak is smaller and the wavelength difference of the silage component of the absorption peak is smaller, the dimensionality reduction weight influence coefficient of the absorption peak is larger, and the absorption peak can better reflect the characteristics of the main components in the whole corn silage. When the NIR model is determined according to the near-infrared spectral curve of the silage sample in the future, more attention should be paid to the information on the main components of silage provided by the absorption peak.

[0082] The dimensionality reduction weight influence coefficient of any absorption peak can be obtained in the same way.

[0083] At this point, the dimensionality reduction weight influence coefficients of all absorption peaks of the near-infrared spectral curves of all silage samples are obtained.

[0084] Step S004: Establish a NIR model based on the dimensionality reduction weight influence coefficients of all absorption peaks of the near infrared spectral curves of all silage samples, and use the NIR model to achieve standardized determination of whole-plant corn silage digestibility.

[0085] The spectral data corresponding to all absorption peaks of the near-infrared spectral curves of all silage samples were processed using the PCA principal component analysis method to obtain the projection coefficient of the spectral data of each absorption peak on the principal component.

[0086] The normalized value of the product of the dimensionality reduction weight influence coefficient of the absorption peak and the projection coefficient of the spectral data of the absorption peak on the principal component is recorded as the dimensionality reduction weight of the absorption peak. The dimensionality reduction weight of the absorption peak is used as the projection coefficient of the spectral data of the absorption peak on the principal component. The spectral data corresponding to all absorption peaks of the near-infrared spectral curves of all samples are processed using the PCA principal component analysis method to obtain the spectral data after dimension reduction.

[0087] The digestibility of the silage sample is determined based on the silage sample. The method for determining the digestibility of the silage sample based on the silage sample is a well-known technology and will not be described in detail.

[0088] The spectral data after dimension reduction were used as independent variables, and the digestibility of each silage sample was used as the dependent variable. The NIR model was established using the PLS partial least squares regression method.

[0089] For whole-plant corn silage that requires standardized digestibility measurement, collect silage samples to be tested from the whole-plant corn silage, use a near-infrared spectrometer to test the silage samples to obtain a near-infrared spectrum curve of the silage samples to be tested, obtain all absorption peaks of the near-infrared spectrum curve of the silage samples to be tested, input spectral data corresponding to all absorption peaks of the near-infrared spectrum curve of the silage samples to be tested into the NIR model, and the digestibility of the silage samples to be tested can be obtained, and the digestibility of the silage samples to be tested is used as the digestibility of the whole-plant corn silage that requires standardized digestibility measurement.

[0090] At this point, the NIR model for the standardized determination of the digestibility of whole-plant corn silage has been obtained, and the standardized determination of the digestibility of whole-plant corn silage can be achieved using the NIR model.

[0091] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A standardized method for determining the digestibility of whole-plant corn silage for near infrared modeling, characterized in that: The method comprises the following steps: Extract a preset number of silage samples, obtain a near-infrared spectrum curve of each silage sample, and determine the suspected absorption peak point on the near-infrared spectrum curve, where each point on the near-infrared spectrum curve corresponds to an absorbance and a wavelength; Determine the absorbance and wavelength of the characteristic absorption peak of each main chemical component of the whole corn silage, record any suspected absorption peak point as the target suspected absorption peak point, and determine the absorption peak possibility of the target suspected absorption peak point based on the difference between the absorbance of different suspected absorption peak points in the vicinity of the target suspected absorption peak point and the absorbance of the characteristic absorption peak, and the difference between the wavelength of the target suspected absorption peak point and two adjacent suspected absorption peak points, and determine the absorption peak based on the absorption peak possibility of the target suspected absorption peak point; Record any absorption peak as the target absorption peak; record the sum of the absolute values ​​of the difference in wavelength between the target absorption peak suspected point and the two adjacent absorption peak suspected points as the wavelength difference of the target absorption peak suspected point; record the characteristic absorption peak of the main chemical component corresponding to the absorption peak change speed correction value corresponding to the absorption peak possibility as the real characteristic absorption peak of the target absorption peak; record the absolute value of the difference between the wavelength of the target absorption peak and the wavelength of the real characteristic absorption peak as the silage component wavelength difference of the target absorption peak; record the mean value of the wavelength of the absorption peak of the near-infrared spectral curves of all silage samples corresponding to the same real characteristic absorption peak as the target absorption peak as the mean value of the wavelength of the silage samples corresponding to the real characteristic absorption peak; record the absolute value of the difference between the wavelength of the target absorption peak and the mean value of the wavelength of the silage samples corresponding to the real characteristic absorption peak of the target absorption peak as the wavelength deviation of the absorption peak; determine the dimensionality reduction weight influence coefficient of the target absorption peak according to the silage component wavelength difference and the wavelength deviation of the target absorption peak; A NIR model was established based on the dimensionality reduction weight influence coefficients of all absorption peaks in the near-infrared spectral curves of all silage samples, and the NIR model was used to achieve standardized determination of whole-plant corn silage digestibility.

2. The method for standardizing whole-plant corn silage digestibility for near infrared modeling according to claim 1, characterized in that: The specific method of determining the suspected absorption peak point on the near-infrared spectrum curve includes: Obtain the minimum period value of the near-infrared spectrum curve of the silage sample, take half of the minimum period value as the sampling time interval, start sampling from the minimum value of the horizontal coordinate of the near-infrared spectrum curve, and obtain the sampling point; All sampling points whose slopes are within a range of greater than or equal to -t and less than or equal to t are marked as suspected absorption peak points, where t represents a preset range threshold.

3. The method for standardizing whole-plant corn silage digestibility for near infrared modeling according to claim 1, characterized in that: The method of determining the absorption peak possibility of the target absorption peak suspected point according to the difference between the absorbance of different absorption peak suspected points in the vicinity of the target absorption peak suspected point and the absorbance of the characteristic absorption peak, and the difference between the wavelength of the target absorption peak suspected point and two adjacent absorption peak suspected points, includes the following specific methods: Recording a first preset number of absorption peak suspected points adjacent to the target absorption peak suspected point as adjacent absorption peak suspected points of the target absorption peak suspected point; The average value of the absorbance of the target absorption peak suspected point and all adjacent absorption peak suspected points is recorded as the average absorbance of the target absorption peak suspected point; the absolute value of the difference between the absorbance of the adjacent absorption peak suspected points of the target absorption peak suspected point and the average absorbance of the target absorption peak suspected point is recorded as the absorbance difference of the adjacent absorption peak suspected points; the average value of the absorbance difference between the target absorption peak suspected point and all adjacent absorption peak suspected points is recorded as the change speed of the target absorption peak suspected point; The minimum value of the change speed of all the target absorption peak suspected points is used as the change speed correction value of the target absorption peak suspected points; The absorption peak possibility of the target absorption peak suspected point is determined according to the difference between the wavelengths of the target absorption peak suspected point and two adjacent absorption peak suspected points and the change speed correction value of the target absorption peak suspected point.

4. The method for standardizing whole-plant corn silage digestibility for near infrared modeling according to claim 3, characterized in that: The method of determining the absorption peak possibility of the target absorption peak suspected point according to the difference between the wavelength of the target absorption peak suspected point and two adjacent absorption peak suspected points and the change speed correction value of the target absorption peak suspected point includes the following specific methods: The normalized value of the ratio of the wavelength difference of the target absorption peak suspected point to the change speed correction value is recorded as the absorption peak possibility of the target absorption peak suspected point.

5. The method for standardizing whole-plant corn silage digestibility for near infrared modeling according to claim 1, characterized in that: The specific method for determining the absorption peak is: When the absorption peak possibility of the target absorption peak suspected point is greater than or equal to the absorption peak judgment threshold, the target absorption peak suspected point is recorded as the absorption peak.

6. The method for standardizing whole-plant corn silage digestibility for near infrared modeling according to claim 1, characterized in that: The specific method of determining the dimensionality reduction weight influence coefficient of the target absorption peak according to the silage component wavelength difference and wavelength deviation of the target absorption peak includes: The normalized value of the product of the wavelength difference of the silage component of the target absorption peak and the wavelength deviation is recorded as the first influence coefficient of the target absorption peak, and the difference between the number 1 and the first influence coefficient of the target absorption peak is recorded as the dimensionality reduction weight influence coefficient of the target absorption peak.

7. The method for standardizing whole-plant corn silage digestibility for near infrared modeling according to claim 1, characterized in that: The NIR model is established according to the dimensionality reduction weight influence coefficients of all absorption peaks of the near infrared spectrum curves of all silage samples, and the specific method includes: The spectral data corresponding to all absorption peaks of the near-infrared spectral curves of all silage samples were processed using the PCA principal component analysis method to obtain the projection coefficient of the spectral data of each absorption peak on the principal component; the normalized value of the product of the dimensionality reduction weight influence coefficient of the absorption peak and the projection coefficient of the spectral data of the absorption peak on the principal component was recorded as the dimensionality reduction weight of the absorption peak; The dimension reduction weight of the absorption peak is used as the projection coefficient of the spectral data of the absorption peak on the principal component, and the spectral data corresponding to all absorption peaks of the near-infrared spectral curves of all samples are processed using the PCA principal component analysis method to obtain the spectral data after dimension reduction; The spectral data after dimension reduction were used as independent variables, and the digestibility of each silage sample was used as the dependent variable. The NIR model was established using the PLS partial least squares regression method.

8. The method for standardizing whole-plant corn silage digestibility for near infrared modeling according to claim 1, characterized in that: The method of using the NIR model to achieve the standardized determination of whole-plant corn silage digestibility includes: For whole-plant corn silage that needs to be subjected to standardized digestibility determination, a silage sample to be tested of the whole-plant corn silage is collected, all absorption peaks of the near-infrared spectrum curve of the silage sample to be tested are obtained, and spectral data corresponding to all absorption peaks of the near-infrared spectrum curve of the silage sample to be tested are input into the NIR model to obtain the digestibility of the silage sample to be tested; The digestibility of the silage sample to be tested is used as the digestibility of the whole corn silage that needs to be standardized for digestibility determination.

Images