Identification of Haploid Maize Kernel Using NIR Spectroscopy in Reflectance and Transmittance Modes: A Comparative Study.

The spectra measurements mode that suitable for haploid maize kernel identification was explored using MicroNIR-1700 series of miniature near infrared spectrometer by JDSU company. Based on Near Infrared Spectroscopy (NIRS) qualitative analysis techniques, we conducted a comparative study using reflectance and transmittance spectra to identify haploid maize kernels. Partial least squares-discriminant analysis (PLS-OLDA) was used to compress the pretreated spectral data, and then the identification models were built based on Support Vector Machine (SVM). The measured data were recorded in reflectance and transmittance modes and the recognition correct rates were calculated. For measurements taken in reflectance mode, the average recognition rate was less than 60% regardless of embryo side positions. In transmittance mode, however, the average recognition rate reached 93.2%. The experiment results show that diffuse reflection spectrum could only obtain corn grain surface information, so embryo side positions severely affect haploid maize kernel identification effect when reflectance measurements mode have been employed, but they have far less impact on transmittance mode. The near infrared diffuse transmittance spectra analyzes non-uniform samples can achieve the analysis of optical path depth information accumulation, all information of the sample interior can be obtained, so transmittance spectra could identify haploid maize effectively and be desensitized to kernel positions. NIRS qualitative analysis techniques with features of rapid, nondestructive could identify the haploid and Micro-NIR spectrometer scan fast and cost less, which have utility for automatically selecting haploid maize kernels from hybrid kernels.

[Study of Modeling Samples Selection Method Based on Near Infrared Spectrum].

For more wheat varieties classification problem, we use near infrared spectrumto do qualitative analysis. Increasing the size of modeling sample could increase information of the model, however, at the same time, it also makes information redundancy so that modeling time and storage space will increase, thus, we need to decrease the size of modeling sample though selecting them. Some information must be lost and the effects of the model must be worse if we select samples blindly. We put forward the k nearest neighbor-density sample selection based on the traditional selection methods. Experiments use the near infrared diffuse reflection spectrum of wheat seed from lots of days. First, we use preprocessing and feature extraction to deal with the wheat original spectrum, then select modeling sample by three methods that are random sampling, k nearest neighbor and k nearest neighbor-density, finally, we establish the models of BPR(Biomimetic Pattern Recognition) and BPRI(Biomimetic Pattern Recognition Improved). The experimental results show that in the model of BPR we get the best results using the selection method of k nearest neighbor-density, especially it also decreases the size of modeling sample deeply, and in the model of BPRI the results using the selection method of k nearest neighbor-density are much better than random sampling and a little better than k nearest neighbor, but in the meanwhile the size of modeling sample using the selection method of k nearest neighbor-density are much smaller than k nearest neighbor. The experimental results prove that the sample selection method of k nearest neighbor-density can not only greatly reduce the modeling sample size, and ensure the quality of the model, it has obvious effect on varieties classification problem of wheat.

[Maize Hybrid Seed Purity Identification Based on Near Infrared Reflectance (NIR) and Transmittance (NIT) Spectra].

This article explore the feasibility of using Near Infrared Reflectance (NIR) and Transmittance (NIT) Spectroscopy (908.1-1677.2 nm wavelength range) to identify maize hybrid purity, and compare the performance of NIR and NIT spectroscopy. Principle Component Analysis (PCA) and Orthogonal Linear Discriminant Analysis (OLDA) were used to reduce the dimension of spectra which have been pretreated by first derivative and vector normalization. The hybrid purity identification model of Nonghua101 and Jingyu16 were built by SVM. Models based on NIR spectra obtained correct identification rate as 100% and 90% for Nonghua101 and Jingyu16 respectively. But NIR spectra were greatly influenced by the placement of seeds, and there existed significant difference between NIR spectra of embryo and non-embryo side. Models based on NIT spectroscopy yielded correct identification rate as 98% both for Nonghua101 and Jingyu16. NIT spectra of embryo and non-embryo side were highly similar. The results indicate that it is feasible to identify maize hybrid purity based on NIR and NIT spectroscopy, and NIT spectroscopy is more suitable to analyze single seed kernel than NIR spectroscopy.

Discrimination of Maize Haploid Seeds from Hybrid Seeds Using Vis Spectroscopy and Support Vector Machine Method.

Doubled haploid (DH) lines are routinely applied in the hybrid maize breeding programs of many institutes and companies for their advantages of complete homozygosity and short breeding cycle length. A key issue in this approach is an efficient screening system to identify haploid kernels from the hybrid kernels crossed with the inducer. At present, haploid kernel selection is carried out manually using the"red-crown" kernel trait (the haploid kernel has a non-pigmented embryo and pigmented endosperm) controlled by the R1-nj gene. Manual selection is time-consuming and unreliable. Furthermore, the color of the kernel embryo is concealed by the pericarp. Here, we establish a novel approach for identifying maize haploid kernels based on visible (Vis) spectroscopy and support vector machine (SVM) pattern recognition technology. The diffuse transmittance spectra of individual kernels (141 haploid kernels and 141 hybrid kernels from 9 genotypes) were collected using a portable UV-Vis spectrometer and integrating sphere. The raw spectral data were preprocessed using smoothing and vector normalization methods. The desired feature wavelengths were selected based on the results of the Kolmogorov-Smirnov test. The wavelengths with p values above 0. 05 were eliminated because the distributions of absorbance data in these wavelengths show no significant difference between haploid and hybrid kernels. Principal component analysis was then performed to reduce the number of variables. The SVM model was evaluated by 9-fold cross-validation. In each round, samples of one genotype were used as the testing set, while those of other genotypes were used as the training set. The mean rate of correct discrimination was 92.06%. This result demonstrates the feasibility of using Vis spectroscopy to identify haploid maize kernels. The method would help develop a rapid and accurate automated screening-system for haploid kernels.

[Detection of Adulteration in Milk Powder with Starch Near Infrared].

Three China trademarks of milk powder called Mengniu, Yili, Wandashan were taken as testing samples. Each of them mixed varied amount of starch in different gradient, which were consisted of 32 adulterated milk powder samples mixed with starch, was taken as standard samples for constructing predicted model. To those 32 samples, the reflecting spectrum characteristics in middle wave of near infrared spectrum with Near Infrared Spectrum Analyzer (Micro NIR 1700) produced by JDSU Ltd. USA were collected for five repeats in five different days. The time span was nearly two months. Firstly, we build the model used the reflecting spectrum characteristics of those samples with biomimetic pattern recognition (BPR) arithmetic to do the qualitative analysis. The analysis included the reliability of testing result and stability of the model. When we took ninety percent as the evaluation threshold of testing result of CAR (Correct Acceptance Rate) and CRR (Correct Rejection Rate), the lowest starch content of adulterate milk powder in all tested samples which the tested result were bigger than that abovementioned threshold was designated CAR threshold (CAR-T) and CRR threshold (CRR-T). CAR means the correct rate of accepting a sample which is belong to itself, CRR means correct rate of refusing to accept a sample which is not belong to itself. The results were shown that, when we constructed a model based on the near infrared spectrum data from each of three China trademark milk powders, respectively, if we constructed a model with infrared spectrum data tested in a same day, both the CAR-T and CRR-T of adulterate starch content of a sample can reach 0.1% in predicting the remainder infrared spectrum data tested within a same day. The three China trademarks of milk powder had the same result. In addition, when we ignored the trademarks, put the spectrum data of adulterate milk powder samples mixed with the same content of starch of three China trademarks milk powder together to construct a model, the CAR-T of mixed starch content of a sample may reach 0.1%, the CRR-T can reach 1%, if the model construction and predicting were performed with near infrared spectrum data tested in a same day. However, the CAR-T can just stably reach up to 5% and the CRR-T have the same result, if the model construction and predicting were crossly performed with mixed near infrared spectrum data tested in different days. Furthermore, the correct recognizing threshold mixed starch of a sample can stably reach up to 1% and the CAR-T can reach 5%, if the model construction was based on near infrared spectrum data combined the previous four days to predict the output of the another day. On the other hand, we also engaged quantitative analysis to the starch content in milk power with two kinds of arithmetic (PLSR, LS-SVR). In contrast with the testing outputs, the reliability of both the CAR-T and CRR-T in qualitative analysis was further validated.

[Feasibility study on an approach for identifying corn kernel varieties with seed coating agents via near infrared spectroscopy].

It is generally accepted that near infrared reflectance spectroscopy (NIRS) can be used to identify variety authenticity of bare maize seeds. In practical, maize seeds are covered with seed coating agents. Therefore it's of huge significance to investigate the feasibility of identifying coated maize seeds by NIRS. This study employed NIRS to quickly determine the variety of coated maize seeds. Influence of seed coating agent on NIR spectra was discussed. The NIR spectra of coated maize seeds were obtained using an innovative method to avoid the impact of the seed coating agent. Coated seeds were cut open, and the sections were scanned by the spectrometer, so as to acquire the information of the seed itself. Then, support vector machine (SVM), soft independent modeling of class analogy (SIMCA), and biomimetic pattern recognition (BPR) was employed to establish the identification model for four maize varieties, and yield 93%, 95.8%, 98% average correct rate respectively. BPR model showed better performance than SVM and SIMCA models. The robustness of identification model was tested by seeds harvested from four regions and model showed good performance.

[Near infrared spectroscopy analysis method of maize hybrid seed purity discrimination].

Near infrared spectroscopy analysis method of discrimination of maize hybrid seed purity was studied with the sample of Nong Hua 101 (NH101) from different origins and years. Spectral acquisition time lasted for 10 months. Using Fourier transform (FT) near infrared spectroscopy instruments, including 23 days in different seasons (divided into five time periods), a total of 920 near infrared diffuse reflectance spectra of single corn grain of those samples were collected. Moving window average, first derivative and vector normalization were used to pretreat all original spectra, principal component analysis (PCA) and linear discriminant analysis (LDA) were applied to reduce data dimensionality, and the discrimination model was established based on biomimetic pattern recognition (BPR) method. Spectral distortion was calibrated by spectra pretreatment, which makes characteristics spatial distribution range of sample spectra set contract. The relative distance between hybrid and female parent increased by nearly 70-fold, and the discrimination model achieved the identification of hybrid and female parent seeds. Through the choice of representative samples, the model's response capacity to the changes in spectral acquisition time, place and environment, etc. was improved. Besides, the model's response capacity to the changes in time and site of seed production was also improved, and the robustness of the model was enhanced. The average correct acceptance rate (CAR) of the test set reached more than 95% while the average correct rejection rate (CRR) of the test set also reached 85%.

[Study on method of maize hybrid purity identification based on hyperspectral image technology].

The feasibility of employing hyperspectral image technology to identify maize hybrid purity was studied by analyzing the spectral information of maize hyperspectral image. The hyperspectral images of hybrid and female parent of maize variety NH101 in the range of 871-1 699 nm including 308 wavelengths were collected by hyperspectral imaging system. We extracted average spectral information of interested region on maize seed and built identification models of hybrid and female parent of maize variety NH101 based on processed spectral data. The influences of different sample laying modes (seed embryo facing the light source, seed embryo backward light source, and seed put in different locations on sample stage) and experimental environments on the performance of identification models were discussed. Spectral collected under different sample laying modes and experimental environments were used to test the robustness of identification models. The average correct acceptance rates and average correct rejection rates are more than 90%. The feature spectral bands (1 195-1 246 nm) with which the differences between hybrid and female parent are the largest were extracted by a wavelength selection method based on standard deviations, called Qs. The performance of identification models built based on spectral data in feature spectral bands reached the same level of models built based on spectral data in the full range of 925-1 597 nm. The results demonstrated the feasibility of using hyperspectral image technology as an objective and rapid method for the identification of maize hybrid purity.

[Study on spectral measurement methods in identification of maize variety authenticity based on near infrared spectra of single kernels].

In order to establish the better maize variety identification models based on single kernel samples, the near-infrared spectral measurement methods were studied by comparing the direction of the maize seed's embryo, diffuse reflectance and transmission mode, devices of holding the sample according to their impacts on the performance of variety identification models. Partial least squares-discriminant analysis (PLSDA) was used to compress the pretreated spectral data into 9 variables, and then the identification models were built based on biomimetic pattern recognition (BPR). The results show that with the maize grain's embryo facing the light source the models can be made perform better than with embryo backing toward the light source, diffuse reflectance mode is better than transmission mode, and small sample pool performs better than the small aperture. The measurement method of acquiring the diffuse reflectance near infrared spectra of maize by small pool with the seed embryo facing the light source can make models have the best performance. The average correct identification rate of the models is 94.6%, and the average correct rejection rates for the varieties not belonging to the models reached 96.5%.

[The utility research on NIR diffuse reflectance and transmittance measurements mode in authenticity identification of maize population samples].

In the present paper, the spectral measurements of maize population samples were researched so as to identify their authenticity. Diffuse reflectance and transmittance measure modes were used to collect spectral data of 8 maize varieties. DPLS-DA was used to compress pretreated data. The recognition models of eight maize varieties were built based on biomimetic pattern recognition (BPR). The average correct recognition rate and average correct rejection rate of identification models built by two modes were calculated. The average recognition rate and rejection rate of diffuse reflectance method reached 94.50% and 96.71%, and those of transmittance method reached 98.5% and 98.00%, respectively. Both of them met the requirements of maize preliminary screening, and the recognition rate and rejection rate of transmittance method are higher than diffuse reflectance method by 4% and 1.3% respectively.

[Study of feature extraction methods for maize's near infrared spectra in biomimetic pattern recognition].

Near infrared spectrum is an important step in near infrared spectrum qualitative analysis, which influences the qualitative analysis results directly. Diffuse transmittance measurements mode was used to collect spectral data of eight maize varieties. PCA, ICA, PLS-DA and wavelet transformation were used to extract features of pretreated data. Finally, we used the test set data to test the recognition models of eight maize varieties which were built based on biomimetic pattern recognition (BPR). We draw a conclusion that PLS-DA can make models get higher average correct recognition rate than PCA, ICA and Wavelet transformation.

[Analysis method of the fingerprint of LiuWei DiHuang pills by near infrared spectroscopy].

The purpose of the present paper is to establish a method of the fingerprint of LiuWei DiHuang pills by the near infrared spectroscopy. First, the authors established the fingerprint of LiuWei DiHuang pills by high performance capillary electrophoresis (HPCE), and calculated its similarity. At the same time, the authors scanned its near infrared spectrogram. Then the authors established the mathematical model between the similarity of fingerprint and the near infrared spectrogram. Through the optimization of the model, the correlation(r), calibration standard deviation and the average relative error of the modeling set were 0.9046, 0.058 and 6.12%, respectively. It proved that the linearity between calculated and forecast of fingerprint was clear. The results showed that the method of the fingerprint of LiuWei DiHuang pills by the near infrared spectroscopy was feasible.

[Progress in the study of impact of scattering on stability of quantitative analysis model using near infrared spectroscopy technology and correction methods].

Near infrared spectroscopy technology for inner quality detection and analysis is a rapid and accurate technique. The factors affecting the stability of quantitative analysis models based on NIRS especially the influence of scattering, were investigated. The methods of determination of the parameters of optical properties of sample (scattering coefficient and absorption coefficient) and the interaction between light and sample were summarized. The progress in pretreatment methods to improve the model's stability, such as transformation of standard normal variate (SNV) and multiplication scatter correction (MSC); as well as the multivariate calibration model techniques, such as partial least square (PLS) and primary component analysis regression (PCA), were also described. The model with scatter correction pretreatments and different regression methods were compared to establish a better regression equation by near Infrared reflectance spectroscopy (NIRS). Although pretreatment methods could improve the robustness of model for influence of scattering, new techniques should be employed to eliminate the error of predication under scattering condition when samples are complicated in physical states and are in the large range of concentration.

[The rapid analysis of fatty acids in vegetable oils by near infrared spectrum].

In this research, The functional components of vegetable oils were analyzed by near infrared (NIR) spectral technology. The optimum conditions of mathematics model of four components (C16 : 0, C18 : 0, C18 : 1, C18 : 2) were studied, including the sample set selection, chemical value analysis, the detection methods and condition. Chemical value was analyzed by HPLC. 52 samples were selected, 41 for modeling set and 11 for testing set. All samples were placed in 5mm thick sample pools and swept by near infrared (NIR) with discrimination factor 8 cm(-1) without any other disposal. Using PLS methods sated model. Data were processed by first derivative method and centering method. 5 000-9 000 cm(-1) spectral region was analyzed. Correlating index (r), RMSECV and RMSEP were chose as evaluation index. The result demonstrated that the correlation between the reference value of the modeling sample set and the near infrared predictive value were r(C16 : 0) = 0.891, r(C18 : 0) = 0.837, r(C18 : 1) = 0.982, r(C18 : 2) = 0.971, respectively. And the correlation between the reference value of the testing sample set and the near infrared predictive value were 0.921, 0.891, 0.946 and 0.949, respectively. It proved that the near infrared predictive value was linear with chemical value and the mathematical model established for components of vegetable oils was feasible. For validation, 8 unknown samples were selected to be analysis by infrared (NIR). The result demonstrated that error between predict value and chemical value was less than 10%. That was to say infrared (NIR) had a good veracity in analysis components of vegetable oil. Because infrared (NIR) spectral technology is convenient, rapid than HPLC in oil components analysis, moreover, infrared (NIR) can analyze many components at the same time. It must have great application prospect in vegetable oil components analysis.

[Quality analysis of Chinese bacon with near infrared spectroscopy].

The feasibility of fast and correctly detecting the quality of Chinese bacon by NIR was studied. The acid value (AV) can reflect the quality of Chinese bacon during processing and storage which is prescribed in the Chinese national standard methods definitely. The fat is abundant in Chinese bacon, so the AV index is important for the quality of Bacon. Samples were scanned on the Bruker FTNIR reflected spectra instrument after being ground. The preprocess method of Additional Scattered Correction was used for the mathematic model of AV and moisture content of Chinese Bacon by PLS. The correlation ratio and the RMSCV of AV and moisture content of the prediction set were 0. 98, 0. 25, 0. 90 and 0. 02 respectively. The results showed that NIR spectroscopy analysis technology can be used for fast detecting AV and moisture content of Chinese Bacon.

[Study of temperature correction in the PLS near-infrared quantitative analysis of the wheat].

The influence of sample temperature on the wheat PLS near infrared (NIR) quantitative analysis was studied with 45 wheat seeds. The original spectra of the wheat seeds were measured at different temperatures. The temperature corrections were subtracted from the original spectra of the samples by using the PLS algorithm. The corrected spectra were used to build the mathematical model of the near-infrared quantitative analysis of the wheat. The results show that the influence of sample temperature on the mathematical model can be eliminated effectively when the temperature correction was performed in the calculation, improving the stability and prediction of this mathematical model.

[The effects of noise on NIR analysis and related mathematic pretreatments and models].

The feasibility of using a relatively high noise NIR spectrometer for analysis was examined by using maize powder samples. The results showed that with four-time averaged NIR spectrum data without more pretreatments, PLS mathematic models and CAU-NIR software, the relative high noise scan NIR spectrometer could be used to get satisfied prediction results compared with other low noise NIR spectrometers. The prediction coefficient could reach 98% and the CV (variation coefficient) was 6.2%. It was proved that when the S/N of NIR spectrometer was lower than 10(5), it still could be used for quantity analysis with the help of some mathematic pretreatments and models.

[Study on NPLS model with three-dimensional hyperspectrum for assessing per-mu-yield of winter wheat].

The present study piled a cube array with the winter wheat's cap reflection hyperspectra sampled from 35 experimental districts during 8 growth periods, and then used a multiway partial least squares (NPLS) algorithm to establish a model to assess the per-mu-yield of winter wheat. The model was used to predict the per-mu-yields of other 13 experimental districts. The correlation coefficient of predicted and actual values was 0.9366, and the mean relative error was 4.44%. These results showed a good prediction of the model. The study has guiding meaning to actual yield assessment.

[The PLS calibration model optimization and determination of rice protein content by near-infrared reflectance spectroscopy].

A hundred and ninety one representative brown rice samples from the Chinese Rice Genebank and the DH population derived from the cross of japonica upland rice IRAT109 with paddy rice Yuefu were selected for this study. Their protein content range was 5.90%-14.50%. Near-infrared diffusive spectroscopy (NIDRS) and partial least square (PLS) were used to determine protein content with different wavelength ranges and data preprocessing methods for regression and information extraction. The object function [R/(1+RMSECV)] of quantitative model was defined, and the samples of calibration and validation tests were classified by projective distribution of PLS loadings. These methods were applied to the optimization of the calibration model. It is demonstrated that the calibration model developed by the spectral data pretreatment of the first derivative + standard vector normalization with the same spectral region (5 000-9 000 cm(-1) resulted in the best determination of protein content in brown rice when the maximum values of the object function were reached. The maximum values of the object functions of calibration and validation sets were 0.701 and 0.687, respectively. Projective distributions of PLS loadings were used to validate the models, and the result was the same as that of validating model by object function [R/(1+RMSECV)].

[Prediction of IVDMD with near infrared reflectance spectroscopy (NIRS) in maize stalk].

The in vitro dry matter digestion (IVDMD) in maize stalk was analyzed with 161 samples selected from 600 samples of different eco-environments, hybrids and inbred lines, development stages, and various parts of the plants in two years. The technique of near infrared reflectance spectroscopy (NIRS) and partial least square regression (PLS) were used to establish the models by comparing several preprocessing procedures and wavelength ranges. The optimal models could be obtained in the range of 6 101. 7-5 773. 8 cm(-1) and 4 601. 3-4 246. 5 cm(-1) by the spectral data preprocessing of the Max-Min normalization. The model is suitable for measuring various sample IVDMD. The determination coefficients of the modes were 0.907 3 and 0.906 6 for cross and external validation, respectively. The root mean square error of prediction was 2.08%, and the coefficient of correlation(r) was 0.956 between NIRS predicted and actual IVDMD in these materials. The results showed that NIRS is a simple effective means for measuring IVDMD in maize stalk. The results are of great value of application in screening and evaluating quality constituents of silage maize.