Method for Identifying Maize Haploid Seeds by Applying Diffuse Transmission Near-Infrared Spectroscopy

Abstract

The identification for haploid seeds is an important process in maize haploid breeding. Thanks to the diffuse transmission (DT) technology of near-infrared (NIR) spectroscopy, maize haploid seeds can be selected automatically using NIR spectrum features. However, the NIR spectra of maize seeds contain a large number of redundant features and noise that will degrade the identification performance. We resolved this problem by designing a low dimension and uniform space of seed spectrum features to improve the collected spectra. The zero-phase component analysis (ZCA) method was utilized to uniform the feature space and the partial least squares regression (PLSR) was employed to design the low dimension space. Then, by using the classifier of back propagation neural network (BPNN), a high qualitative identification method was developed for selecting maize haploid seeds. The study results demonstrate that the average accuracy of the proposed method is outstanding (96.16%) with a minor standard deviation (SD) compared with other methods. Therefore, our proposed method is potentially useful for automatically identifying maize haploid seeds.

A data transfer method for improving seed identification of maize (Zea mays) haploid breeding based on genetic similarity

“…In the maize haploid breeding technology, the haploid production rate has been improved with the invention of the inducer line method (Chen et al, ). When the NIR analysis is employed to lessen the human identification work in this technology (Li & Li, ; Lin et al, ), plenty of various inducer lines cause the difficulty in data collection for training models to identify different haploids. For example, when experts want to breed seeds in a certain place, they usually prepare several alternative male parents to cross with a common female parent to deal with the practical situations, such as the needs for induction rate, pollen quality, high‐temperature resistance and florescence.…”

“…In each group, the data from a different male parent were employed as D S data, and the progeny data of haploid and hybrid was used as D T data. The spectra data of seeds were acquired by a cross‐interval way under the stable instrument (Figure ) and environment after calibration (Lin et al, ; Yu et al, ). The collection lasted for 10 days, and 100 haploid seed and 100 hybrid seed were collected each day.…”

“…For better identification, the data needs to be processed to remove noises and redundant features in the spectrum. The data preprocessing of zero component analysis (ZCA) can make the spectrum data more independent than the standard normalization (SN) method (Kessy, Lewin, & Strimmer, ; Lin et al, ). Similarly, the partial least squares regression (PLSR) (Höskuldsson, ) is a common method that can extract features X from the spectrum.…”

“…The neural network (NN) has been widely used in agricultural technologies (Samuel & Ismail, ; Shi, Zhao, Zhang, & Ren, ; Silva et al, ; Takahashi et al, ; Tetyana, Cassinao, Joerg, & Bernd, ). It can make a better identification of seeds with its non‐linear ability (Lin et al, ). However, after data transfer, the number of output positive samples appears to be no longer equal to the number of output negative samples.…”

“…The NIR has the advantages of being non‐destructive (Jiang et al, ; Pasquini, ; Rhiel, Cohen, Murhammer, & Arnold, ; Sandora et al, ), real‐time (Alves‐Rausch, Bienert, Grimm, & Bergmaier, ; Blanco & Villarroya, ; Sandnes et al, ; Sato, Zahlner, Berghofer, Lošák, & Vollmann, ) and of less cost (Sandora et al, ) than other methods, like nuclear magnetic resonance (NMR) (Chen & Song, ), weighting (Smelser et at., ) and image (De La Fuente et al, ). Through the diffuse transmittance (DT) mode, the spectra of non‐uniform organism seeds can be effectively collected with less disturbing effects from other side of seed (Qin et al, , ), and this improvement shows a high performance of seed identification (Li & Li, ; Lin, Yu, Li, & Qin, ; Yu et al, ). However, due to a great variety of seed parents, the identification models need to be established independently for different parents' progeny.…”