Qualitative Analysis of Maize Haploid Kernels Based on Calibration Transfer by Near-Infrared Spectroscopy

Jia Li,Xiao-Ning Yu,Wen-Zhang Ge,Dong An

DOI: https://doi.org/10.1080/00032719.2018.1459656

2018-05-09

Analytical Letters

Abstract:Haploid breeding is one of the most important modern crop selection technologies. Near-infrared spectroscopy (NIRS) has been used to identify haploids rapidly and to non-destructively accelerate the selection process. However, the change of the external environment weakens the performance of the model, as the training and the test spectra may be collected separately from different environments. Thus, a novel calibration transfer method is proposed to calibrate the model in order to reduce the impact of the environment. The near-infrared spectra of 400 maize kernels of two varieties were collected from 9000 to 4000 cm−1. Principal component analysis was performed to construct a feature space and extract features. In the constructed feature space, the calibration transfer method was used to calibrate test sets. Finally, support vector machine was employed to establish a haploid identification model. The results show that when the spectra of the test set and the training set were collected in the same environment, the corrected acceptance of the model was above 90%. While the spectra of the test set and the training set were collected from different environments, the corrected acceptance was 77.87%. However, when the model used the calibration transfer method, the corrected acceptance increased by 12.46%. Moreover, compared with direct standardization, this calibration transfer method achieved better results without detailed sample chemical information and many standards. The results demonstrate that the calibration transfer method based on NIRS was effective for identifying maize haploid kernels in variable environments.

chemistry, analytical