Construction method of oat hay crude protein detection model

Abstract

The invention discloses a construction method of an oat hay crude protein detection model, and relates to the technical field of agricultural product quality detection.The construction method comprises the following steps that samples are collected, and standard value calibration is conducted on the samples; performing spectrum scanning on the sample through a near-infrared spectrometer to obtain near-infrared spectrum data; preprocessing the near infrared spectrum data; constructing an oat hay crude protein detection model based on the preprocessed near infrared spectrum data; predicting the content of crude protein in the sample through the oat hay crude protein detection model, and verifying the oat hay crude protein detection model according to the standard value; high-flux and rapid detection of the oat hay crude protein is achieved, a single sample only needs 3 minutes, the efficiency is improved by 50 times or more compared with a traditional chemical method, meanwhile, the model is constructed based on multiple samples, the pertinence is high, stability is high, chemical reagents are not needed, operation is easy and convenient, cost is low, and the method is suitable for large-scale breeding and production application.

Description

Technical Field

[0001] This invention relates to the field of agricultural product quality testing technology, and in particular to a method for constructing a crude protein detection model for oat hay. Background Technology

[0002] Crude protein content in oat hay is a key indicator for evaluating its nutritional quality, directly affecting the nutritional value of livestock feed and animal production performance. Rapid and accurate determination of crude protein content is of great significance for oat variety selection, feed formulation optimization, and quality control.

[0003] In existing technologies, chemical analysis methods such as the Kjeldahl method are commonly used for determination. Firstly, the process is cumbersome, requiring multiple steps including sample digestion, distillation, and titration, demanding high levels of expertise from operators and making the results susceptible to inconsistencies due to operational errors. Secondly, the entire testing process is time-consuming, typically requiring several hours to complete for a single sample, failing to meet the demands of modern breeding and production for rapid, large-scale sample testing. Furthermore, chemical methods rely on harmful chemical reagents such as sulfuric acid and catalysts, posing safety risks, generating waste liquid, burdening the environment, and increasing testing costs. Finally, this method is destructive; samples cannot be reused after testing, hindering the preservation and subsequent analysis of valuable breeding materials.

[0004] In conclusion, although chemical methods have reference value under laboratory conditions, their low efficiency, high cost, and limited throughput make them unsuitable for large-scale, high-efficiency oat hay quality screening and monitoring scenarios.

[0005] Therefore, a method for constructing a crude protein detection model for oat hay is provided to solve the above problems. Summary of the Invention

[0006] The purpose of this invention is to provide a method for constructing a crude protein detection model for oat hay, which solves the problems of low detection efficiency of traditional chemical methods and poor adaptability of existing near-infrared models, and enables rapid batch detection of crude protein in oat hay.

[0007] To achieve the above objectives, the present invention provides a method for constructing a crude protein detection model for oat hay, comprising the following steps: S1: Collect samples and calibrate the samples according to standard values; S2: The sample is spectrally scanned using a near-infrared spectrometer to obtain near-infrared spectral data; S3: Preprocess near-infrared spectral data; S4: Construct a crude protein detection model for oat hay based on preprocessed near-infrared spectral data; S5: Predict the crude protein content in the sample using the oat hay crude protein detection model, and validate the oat hay crude protein detection model based on standard values.

[0008] Preferably, step S1 specifically includes the following steps: S11: Samples of oats from different planting areas and with different genetic characteristics were collected for two consecutive years. The genetic characteristics included early maturity, late maturity, drought resistance and tolerance to poor soil. The samples were collected during the milk stage and the above-ground parts of the oat plants were collected. S12: The collected samples are dried in a 60℃ oven to constant weight, pulverized by a high-speed pulverizer, and then passed through a 60-mesh sieve. S13: The crude protein content of each sample was determined by the Kjeldahl method. The determination was performed in three parallel trials, and the average value was taken. The average value was set as the standard value for the crude protein detection model of oat hay.

[0009] Preferably, in step S2, the near-infrared spectrometer is a TANGO-R near-infrared spectrometer, with a spectral wavelength range set to 1000 nm-2500 nm and a scanning resolution set to 8 cm⁻¹. -1 The number of scans was set to three, and the output near-infrared spectral data was the average of the three scan results.

[0010] Preferably, step S3 specifically includes the following steps: S31: Perform multi-source scattering correction on near-infrared spectral data; multi-source scattering correction results. Specifically set as follows: ; in, Represents near-infrared spectral data, Represents average spectral data. Indicates standard deviation; S32: Eliminate baseline drift in near-infrared spectral data through first-order derivative processing; S33: Eliminate scattering differences between different samples through standardization.

[0011] Preferably, step S4 specifically includes the following steps: S41: A crude protein detection model for oat hay was constructed using partial least squares regression algorithm, and a near-infrared spectral matrix was established. and sample component content matrix Linear correlation analysis was performed between near-infrared spectral data and sample component content; S42: By maximizing the near-infrared spectral matrix and sample component content matrix The principal components of the crude protein detection model of oat hay were extracted, and the number of principal components was adjusted in real time. The number of principal components was set to a range of 5-8. S43: Pair the preprocessed near-infrared spectral data with the standard values ​​of the oat hay crude protein detection model, and divide the pairing results into a training set and a validation set, with the ratio of the training set to the validation set set being 7:3.

[0012] Preferably, in step S41, the near-infrared spectral matrix and sample component content matrix Specifically set as follows: ; ; in, The score matrix represents the near-infrared spectrum. This represents the transpose of the near-infrared spectral loading matrix. The residual matrix representing the near-infrared spectrum. The score matrix represents the content of sample components. This represents the transpose of the sample component content loading matrix. The residual matrix represents the content of sample components.

[0013] Preferably, step S5 specifically includes the following steps: S51: The near-infrared spectral data in the validation set are processed using the oat hay crude protein detection model, and the corresponding predicted values ​​are output. S52: Compare the predicted values ​​with the standard values ​​of the oat hay crude protein detection model to obtain the prediction accuracy of the oat hay crude protein detection model. The comparison includes the correlation coefficient. and relative analysis error Correlation coefficient and relative analysis error Specifically set as follows: ; ; in, Indicates the first The true standard value of crude protein content for each sample Indicates the first The model-predicted value of crude protein content for each sample. This represents the average of the true standard values ​​of crude protein content across all samples in the validation set. The standard deviation of the validation set is represented by the standard deviation of the validation set. This represents the root mean square error of the validation set; S53: Record the detection time for each sample, and analyze the high-throughput detection effect of the oat hay crude protein detection model based on the detection time and prediction accuracy.

[0014] Therefore, the present invention employs the above-mentioned method for constructing a crude protein detection model for oat hay, which has the following beneficial effects: (1) This scheme has significant advantages in high throughput, short detection time for a single sample, and can achieve batch detection of hundreds of oat germplasms, which is suitable for the large-scale screening needs in the breeding stage and can greatly improve detection efficiency. (2) This scheme has strong adaptability. Based on a large number of oat germplasms with different genetic backgrounds, it constructs a crude protein detection model for oat hay. Compared with the general model, it is more targeted to oat hay and has higher detection stability. (3) This solution is low in cost and easy to operate. It does not require chemical reagents, and the detection process is lossless, which reduces the detection cost. The operation does not require professional chemical analysis skills, which makes it easy for production and breeding units to promote and apply it.

[0015] The method of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0016] Fig. 1 This is a flowchart illustrating a method for constructing a crude protein detection model for oat hay according to the present invention. Fig. 2 This is a schematic diagram illustrating the correlation between the predicted and actual crude protein content values ​​of this invention. Detailed Implementation

[0017] The method of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0018] Unless otherwise defined, the methodological or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0019] The terms "comprising" or "including" as used in this invention mean that the element preceding the term encompasses the element listed after the term, and do not exclude the possibility of encompassing other elements. Terms such as "inner," "outer," "upper," and "lower" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. When the absolute position of the described object changes, the relative positional relationship may also change accordingly. In this invention, unless otherwise explicitly specified and limited, the term "attached" and similar terms should be interpreted broadly. For example, it can refer to a fixed connection, a detachable connection, or an integral part; it can refer to a direct connection or an indirect connection through an intermediate medium; it can refer to the internal communication of two elements or the interaction relationship between two elements. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0020] Example like Figs. 1-2 As shown, this invention provides a method for constructing a crude protein detection model for oat hay, comprising the following steps: S1: Collect samples and calibrate the samples according to standard values; Step S1 specifically includes the following steps: S11: Samples of oats from different planting areas and with different genetic characteristics were collected for two consecutive years. The genetic characteristics included early maturity, late maturity, drought resistance and tolerance to poor soil, to ensure the representativeness of the samples. The samples were collected at the milk stage and the above-ground parts of the oat plants were collected. In this embodiment, the collection period was two consecutive years, the planting areas were Zhangjiakou and Ningxia, and the number of samples collected was 200.

[0021] S12: The collected samples are dried in a 60℃ oven to constant weight, then pulverized by a high-speed pulverizer and passed through a 60-mesh sieve to avoid particle size differences affecting spectral detection; S13: The crude protein content of each sample was determined by the Kjeldahl method, which conforms to the GB / T6432-2018 standard. The average value was taken as the average value of three parallel determinations and set as the standard value of the crude protein detection model of oat hay.

[0022] S2: The sample is spectrally scanned using a near-infrared spectrometer to obtain near-infrared spectral data; In step S2, a TANGO-R near-infrared spectrometer was used, with the spectral wavelength range set to 1000 nm-2500 nm. This range represents the concentrated spectral region of protein characteristics, and the scanning resolution was set to 8 cm⁻¹. -1 The number of scans was set to three, and the output near-infrared spectral data was the average of the three scan results, eliminating random errors.

[0023] S3: Raw spectra are susceptible to baseline drift, instrument noise, and sample scattering. Therefore, near-infrared spectral data are preprocessed to remove abnormal spectral data and ensure the reliability of the spectral data. Step S3 specifically includes the following steps: S31: Perform multi-source scattering correction on near-infrared spectral data; multi-source scattering correction results. Specifically set as follows: ; in, Represents near-infrared spectral data, Represents average spectral data. Indicates standard deviation; S32: Eliminate baseline drift in near-infrared spectral data through first-order derivative processing; S33: Eliminate scattering differences between different samples through standardization.

[0024] S4: Construct a crude protein detection model for oat hay based on preprocessed near-infrared spectral data; Step S4 specifically includes the following steps: S41: A crude protein detection model for oat hay was constructed using partial least squares regression algorithm, and a near-infrared spectral matrix was established. and sample component content matrix Linear correlation analysis was performed between near-infrared spectral data and sample component content; In step S41, the near-infrared spectral matrix and sample component content matrix Specifically set as follows: ; ; in, The score matrix represents the near-infrared spectrum. This represents the transpose of the near-infrared spectral loading matrix. The residual matrix representing the near-infrared spectrum. The score matrix represents the content of sample components. This represents the transpose of the sample component content loading matrix. The residual matrix represents the content of sample components.

[0025] S42: By maximizing the near-infrared spectral matrix and sample component content matrix The principal factors of the crude protein detection model of oat hay were extracted, and the number of principal factors was adjusted in real time to avoid overfitting or underfitting of the model. The number of principal factors was set to 5-8, and in this embodiment, the number of principal factors was set to 6. S43: The pretreated oat hay powder is loaded into a special sample cup, and the pretreated near-infrared spectral data is paired with the standard value of the oat hay crude protein detection model. The pairing results are divided into a training set and a validation set. The ratio of the training set to the validation set is set to 7:3. In this embodiment, the training set contains 140 samples and the validation set contains 60 samples.

[0026] S5: Predict the crude protein content in the sample using the oat hay crude protein detection model, and validate the oat hay crude protein detection model based on standard values.

[0027] Step S5 specifically includes the following steps: S51: The near-infrared spectral data in the validation set are processed using the oat hay crude protein detection model, and the corresponding predicted values ​​are output. S52: Compare the predicted values ​​with the standard values ​​of the oat hay crude protein detection model to obtain the prediction accuracy of the oat hay crude protein detection model. The comparison includes the correlation coefficient. and relative analysis error Correlation coefficient and relative analysis error Specifically set as follows: ; ; in, Indicates the first The true standard value of crude protein content for each sample Indicates the first The model-predicted value of crude protein content for each sample. This represents the average of the true standard values ​​of crude protein content across all samples in the validation set. The standard deviation of the validation set is represented by the standard deviation of the validation set. This represents the root mean square error of the validation set; Correlation coefficient in this embodiment The relative analysis error is 0.91. The value is 3.27, which meets the testing requirements.

[0028] S53: Record the detection time for each sample, and analyze the high-throughput detection effect of the oat hay crude protein detection model based on the detection time and prediction accuracy.

[0029] The results showed that the crude protein detection model of oat hay achieved an accuracy of 88%, and the detection time for a single sample was only 3 minutes, which is more than 50 times more efficient than the traditional chemical method, meeting the needs of high-throughput detection.

[0030] Therefore, the present invention adopts the above-mentioned method for constructing a crude protein detection model for oat hay. By constructing a crude protein detection model for oat hay based on near-infrared spectroscopy and partial least squares regression, the present invention achieves rapid, non-destructive, and high-throughput detection of crude protein content in oats, which greatly improves detection efficiency and reduces costs.

[0031] Finally, it should be noted that the above embodiments are only used to illustrate the method of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the method of the present invention, and these modifications or equivalent substitutions should not cause the modified method to deviate from the spirit and scope of the method of the present invention.

Claims

1. A method for constructing a crude protein detection model for oat hay, characterized in that, Includes the following steps: S1: Collect samples and calibrate the samples according to standard values; S2: The sample is spectrally scanned using a near-infrared spectrometer to obtain near-infrared spectral data; S3: Preprocess near-infrared spectral data; S4: Construct a crude protein detection model for oat hay based on preprocessed near-infrared spectral data; S5: Predict the crude protein content in the sample using the oat hay crude protein detection model, and validate the oat hay crude protein detection model based on standard values.

2. The method for constructing a crude protein detection model for oat hay according to claim 1, characterized in that, Step S1 specifically includes the following steps: S11: Samples of oats from different planting areas and with different genetic characteristics were collected for two consecutive years. The genetic characteristics included early maturity, late maturity, drought resistance and tolerance to poor soil. The samples were collected during the milk stage and the above-ground parts of the oat plants were collected. S12: The collected samples are dried in a 60℃ oven to constant weight, pulverized by a high-speed pulverizer, and then passed through a 60-mesh sieve. S13: The crude protein content of each sample was determined by the Kjeldahl method. The determination was performed in three parallel trials, and the average value was taken. The average value was set as the standard value for the crude protein detection model of oat hay.

3. The method for constructing a crude protein detection model for oat hay according to claim 1, characterized in that, In step S2, a TANGO-R near-infrared spectrometer is used, with the spectral wavelength range set to 1000 nm-2500 nm and the scanning resolution set to 8 cm⁻¹. -1 The number of scans was set to three, and the output near-infrared spectral data was the average of the three scan results.

4. The method for constructing a crude protein detection model for oat hay according to claim 2, characterized in that, Step S3 specifically includes the following steps: S31: Perform multi-source scattering correction on near-infrared spectral data; multi-source scattering correction results. Specifically set as follows: ； in, Represents near-infrared spectral data, Represents average spectral data. Indicates standard deviation; S32: Eliminate baseline drift in near-infrared spectral data through first-order derivative processing; S33: Eliminate scattering differences between different samples through standardization.

5. The method for constructing a crude protein detection model for oat hay according to claim 4, characterized in that, Step S4 specifically includes the following steps: S41: A crude protein detection model for oat hay was constructed using partial least squares regression algorithm, and a near-infrared spectral matrix was established. and sample component content matrix Linear correlation analysis was performed between near-infrared spectral data and sample component content; S42: By maximizing the near-infrared spectral matrix and sample component content matrix The principal components of the crude protein detection model of oat hay were extracted, and the number of principal components was adjusted in real time. The number of principal components was set to a range of 5-8. S43: Pair the preprocessed near-infrared spectral data with the standard values ​​of the oat hay crude protein detection model, and divide the pairing results into a training set and a validation set, with the ratio of the training set to the validation set set being 7:

3.

6. The method for constructing a crude protein detection model for oat hay according to claim 5, characterized in that, In step S41, the near-infrared spectral matrix and sample component content matrix Specifically set as follows: ； ； in, The score matrix represents the near-infrared spectrum. This represents the transpose of the near-infrared spectral loading matrix. The residual matrix representing the near-infrared spectrum. The score matrix represents the content of sample components. This represents the transpose of the sample component content loading matrix. The residual matrix represents the content of sample components.

7. The method for constructing a crude protein detection model for oat hay according to claim 5, characterized in that, Step S5 specifically includes the following steps: S51: The near-infrared spectral data in the validation set are processed using the oat hay crude protein detection model, and the corresponding predicted values ​​are output. S52: Compare the predicted values ​​with the standard values ​​of the oat hay crude protein detection model to obtain the prediction accuracy of the oat hay crude protein detection model. The comparison includes the correlation coefficient. and relative analysis error Correlation coefficient and relative analysis error Specifically set as follows: ； ； in, Indicates the first The true standard value of crude protein content for each sample Indicates the first The model-predicted value of crude protein content for each sample. This represents the average of the true standard values ​​of crude protein content across all samples in the validation set. The standard deviation of the validation set is represented by the standard deviation of the validation set. This represents the root mean square error of the validation set; S53: Record the detection time for each sample, and analyze the high-throughput detection effect of the oat hay crude protein detection model based on the detection time and prediction accuracy.

Images