Determination of Seed Soundness in Conifers Cryptomeria japonica and Chamaecyparis obtusa Using Narrow-Multiband Spectral Imaging in the Short-Wavelength Infrared Range

Abstract

Regeneration of planted forests of Cryptomeria japonica (sugi) and Chamaecyparis obtuse (hinoki) is the pressing importance to the forest administration in Japan. Low seed germination rate of these species, however, has hampered low-cost production of their seedlings for reforestation. The primary cause of the low germinability has been attributed to highly frequent formation of anatomically unsound seeds, which are indistinguishable from sound germinable seeds by visible observation and other common criteria such as size and weight. To establish a method for sound seed selection in these species, hyperspectral imaging technique was used to identify a wavelength range where reflectance spectra differ clearly between sound and unsound seeds. In sound seeds of both species, reflectance in a narrow waveband centered at 1,730 nm, corresponding to a lipid absorption band in the short-wavelength infrared (SWIR) range, was greatly depressed relative to that in adjacent wavebands on either side. Such depression was absent or less prominent in unsound seeds. Based on these observations, a reflectance index SQI, abbreviated for seed quality index, was formulated using reflectance at three narrow SWIR wavebands so that it represents the extent of the depression. SQI calculated from seed area-averaged reflectance spectra and spatial distribution patterns of pixelwise SQI within each seed area were both proven as reliable criteria for sound seed selection. Enrichment of sound seeds was accompanied by an increase in germination rate of the seed lot. Thus, the methods described are readily applicable toward low-cost seedling production in combination with single seed sowing technology.

Hyperspectral imaging for seed quality and safety inspection: a review

“…Such depression was obvious in sound seeds and absent or less prominent in unsound seeds. Average spectra based and pixel-wise spectra based SQI showed the feasibility to select sound seeds [31]. Mo et al used a hyperspectral imaging system with various ranges of spectra induced by blue, green, red and RGB LED (400–500 nm for blue LED, 500–600 nm for green LED, 600–700 nm for red LED and 400–700 nm for RGB LED) to predict the germination quality of cucumber seeds.…”

“…The analysis of spectral features is easier than that of image features, and results have proved the efficiency of models based on spectral features. Given this background, most of the references focus on only the spectral features [12, 13, 22, 30, 31, 33, 57, 61, 67].…”

“…For discriminant models and regression models, PLS was the widely used chemometric method for data analyses of hyperspectral images. PLS had the characteristic of dealing with large number of data rapidly and efficiently, and it worked well for both discriminant (PLS-DA) [14, 27, 30, 32] and regression analysis (PLSR) [31, 68, 72]. Except for PLS, neural network (BPNN), LDA, QDA, SVM, LS-SVM, PCA were also widely used chemometric methods.…”

Determination of the viability of retinispora (Hinoki cypress) seeds using shortwave infrared hyperspectral imaging spectroscopy

“…It should be noted that the C-H group is characteristic of fatty acids, which represent the main difference in chemical constituents between the viable and nonviable seeds. Indeed, it was previously reported 35 that lipids were the key constituents in determining both germinability and the spectral depression between Hinoki and Sugi seeds (conifers). Moreover, other studies have also confirmed that lipids are the dominant reserve compounds for many conifers, 52 which agrees with the findings of this study.…”

“…32 As such, the purpose of optimal wavelength selection involves extraction of the wavebands that are composed of essential information while eliminating the unwanted wavebands from the spectral data. 33,34 Among the recent applications of shortwave infrared HSI (SWIR-HSI) in the context of tree seeds, one study 35 reported the application of narrow multiband spectral imaging to conifer seeds to establish seed quality index parameters for seed soundness in Sugi and Hinoki. The authors reported that for a seed to be tested for its germinability, the viability at the time of sowing must be evaluated in a nondestructive manner.…”