2.1.

Preparation of Seed Samples

Capsicum annum seeds were obtained from the Institute for Microorganism, Kyungpook National University, Daegu, South Korea, to ensure high-seed viability and homogeneity. The seeds were germinated using petri dishes with tissue papers. The seeds were dried at room temperature (27°C) prior to the germination process. The OCT experiment was conducted under laboratory conditions ($27\pm 4°\mathrm{C}$; relative humidity, 80%), and the expected results were obtained within 8 consecutive days. The experiment was performed with 60 seed samples in triplicate ($3\times 20$), divided into three experimental groups. All the selected seeds had an equal initial fresh weight of 1.45 g. The first group of 20 seeds was treated with 1-hexadecene. The second group was treated with butandiol, and the remaining 20 seeds were germinated in sterile distilled water (SDW) medium. To ascertain the optimal chemical concentration, the corresponding two growth-promoting chemicals (1-hexadecene and butandiol) were evaluated initially. Seed samples with an equivalent fresh weight (1.45 g) were germinated for 15 consecutive days according to the gold standard method (fresh weight or length variation analysis) under various concentrations, such as 0.1, 1, 10, 100, 1000, and $2000\mu \mathrm{M}$, for 1-hexadecene and butandiol.

2.2.

Optical Coherence Tomography System Description

Figure 1 depicts the schematic diagram of the customized swept source OCT (SS-OCT) system used for the experiment. A high-speed 1310-nm swept source (Axsun Technology, USA) with a sweep bandwidth of 110 nm, sweeping rate of 50 kHz, and average output power of 20 mW was used as the swept laser source. The light beam from the laser source was passed into the input port of an 80:20 optical fiber coupler (Gooch & Housego, UK). A few parts (20%) of the incident beam were connected to the input port of a circulator. Simultaneously, the remaining 80% of the beam was connected to another circulator. Both of these corresponding circulators were connected to the reference and sample arms. The backscattered light from both the reference and sample arms was propagated through the output terminals of the circulators. This backscattered light of the two arms was interfered in the 50:50 optical fiber coupler (Gooch & Housego). The two output terminals of the fiber coupler were connected to a balanced photodetector (Thorlabs, USA). This obtained signal was digitized using a data acquisition card (Alazar Technologies Inc., Canada). A software-based data processing method was used for image construction. The maximum imaging depth and the imaging width of the customized SS-OCT system were 12 and 10 mm, respectively. The applied refractive index of plant cells, which affects the depth scale of 2-D OCT images, was 1.42. All of the seed specimens were scanned with a $3\mathrm{mm}\times 3\mathrm{mm}$ scanning range. The axial resolution of our system was $6.8\mu \mathrm{m}$ (in air), and the 2-D and 3-D OCT images were generated by scanning the laser beam through a scan lens, achieving transverse resolution of $10\mu \mathrm{m}$ in air and $7\mu \mathrm{m}$ in plant tissues.

Fig. 1

Schematic of the SS-OCT system configuration. AD, analog-digitizer; BD, balanced detector; C, collimator; CIR, circulator; FC, fiber coupler; GS, galvano-scanner; L, lens; M, mirror; PC, polarization controller; S, sample; SS, swept source laser (1310 nm).

3.1.

Fresh Weight Fluctuation Analysis According to the Gold Standard Method

For this experiment, we selected $0.1\text{-}\mu \mathrm{M}$ 1-hexadecene and $1\text{-}\mu \mathrm{M}$ butandiol, along with SDW, since the seed samples treated with $0.1\text{-}\mu \mathrm{M}$ 1-hexadecene and $1\text{-}\mu \mathrm{M}$ butandiol attained the maximum fresh weight.^34,35

Prior to OCT image acquisition, we performed gold standard methods to measure the increase in physical fresh weight and lengthening of the sprout of the chemically treated seeds for 15 consecutive days. Figure 2(a) depicts the physical fresh weights, while Fig. 2(b) outlines the monitored maximum length of the seed sprouts of the three chemically treated samples and the untreated sample. All the samples had an initial fresh weight of 1.45 g. The results revealed that the 1-hexadecene sample had the highest fresh weight of 2.15 g as well as the longest seed sprout.

Fig. 2

Physical fresh weight measurements of the three chemically treated seeds and the untreated seed for 15 consecutive days. (a) Fresh weight fluctuations of the chemically treated seeds on the basis of seed germination. (b) Photographs of the analyzed seed sprout length on the 15th day.

3.2.

Consecutively Monitored Seed Morphological Changes Based Germination Rate

Figure 3 shows the SS-OCT images of seed morphology variations that occurred along with seed germination. Figures 3(a), 3(e), and 3(i) represent the 2-D OCT images of initial seed samples treated with SDW, butandiol ($1\mu \mathrm{M}$), and 1-hexadecene ($0.1\mu \mathrm{M}$), respectively. The results were obtained within 8 consecutive days. The speed of the morphological changes monitored using 2-D images was compared to evaluate the germination speed as well as to verify the growth-promoting chemical compound that led to the optimal increase in germination rate. Both testa and endosperm could be visualized in all three samples at the initial stage. Seed sprouting was observed with the progress of germination after 8 days. Morphological variations, such as the formation of cotyledons and embryonic regions, and the structural changes of the micropylar endosperm region were detected between the initial and final steps of germination. The samples treated with 1-hexadecene showed more rapid morphological changes compared with the samples treated with butandiol and SDW. Figure 3(c) shows a structural change of the micropylar endosperm in the SDW-treated sample on the sixth imaging day; this was observed on the third imaging day for the samples treated with butandiol and 1-hexadecene [Figs. 3(f) and 3(j)]. Similarly, Fig. 3(d) shows the initial stage of the embryonic form on the eighth imaging day [Fig. 3(d)], a stage that was identified on the sixth [Fig. 3(g)] and third [Fig. 3(j)] imaging days for the samples treated with butandiol and 1-hexadecene, respectively. However, the region of the micropylar endosperm in Fig. 3(j) was slightly disappeared, and the thickness of the seed coat was reduced because of rapid germination relative to that observed in Fig. 3(f). Therefore, the samples treated with 1-hexadecene showed a significant acceleration in the germination rate when compared with the samples treated with SDW and butandiol, which correlates with the gold standard results. Additionally, to gain a better understanding about the accelerated germination rate, the duration of the aforementioned embryonic formation in each corresponding seed sample was emphasized using a blue color dashed circular region to highlight the efficacy of 1-hexadecene, which is two times faster than butandiol and more than 2.5 times faster than SDW.

Fig. 3

Day-to-day morphological changes in the chemically treated seeds from the initial stage to the sprouting stage. (a)–(d) Seed germinated in SDW, (e)–(h) butandiol-treated seed, and (i)–(l) 1-hexadecene-treated seed. C, cotyledons; E, endosperm; EM, embryo; ME, micropylar endosperm; SC, storage cotyledons; T, testa (seed coat). The horizontal scale bars: $700\mu \mathrm{m}$ and vertical scale bars: $200\mu \mathrm{m}$.

3.3.

A-Scan Depth Profile Analysis Based Total Signal Intensity Fluctuation

We performed an A-scan depth profile analysis to identify the total signal intensity fluctuation of seed cross-sectional OCT images on the basis of germination speed. For a precise evaluation of intensity fluctuation, a program was coded using MATLAB (Mathworks, USA) to analyze the total pixel intensity of the image in depth direction. In the A-scan process, 2-D OCT image was loaded and the intensity signals belonging to the entire seed cross-section (approximately 300 A-scan lines) were applied. The algorithm detects the maximum intensity in individual A-scan line sequentially. Then all the peak positions in all intensity signals were rearranged while matching the peak intensity index in A-scans to flatten the image. Owing to the physical structure of the seed, the acquired cross-sectional images are nonflattened images containing maximum intensity index positions at different positions. Therefore, the index positions with high intensity should be rearranged and matched linearly to acquire a flattened image. To normalize the A-scan lines, the intensities of the A-scans were divided by the maximum intensity values. Finally, the intensities of all the rearranged, flattened, and normalized A-scan lines were summed up and averaged to obtain the required total intensity fluctuation in depth direction. The evaluated total intensity fluctuations of days 1, 3, 6, and 8 are shown in Figs. 4(a) and 4(d). There was not much intensity fluctuation identified between SDW (blue dotted curve), butandiol (black dashed curve), and 1-hexadecene (red solid curve) on the first day of the experiment, as shown in Fig. 4(a). However, due to the formation of cotyledons and embryonic regions of butandiol- and 1-hexadecene-treated samples, a reduction of the intensity than SDW-treated samples (blue dotted curve) was identified in the A-scans of butandiol (black dashed curve) and the least intensity was identified in 1-hexadecene (red solid curve), as shown in Fig. 4(b). Similar behavior to Fig. 4(b) was identified in the total signal intensity fluctuations on the remaining experimental days [Figs. 4(c) and 4(d)] due to the formation of internal regions such as a large embryonic region and storage cotyledons. As a result of the accelerated seed germination rate of 1-hexadecene and secondly butandiol, internal seed regions, such as embryo, micropylar endosperm, and storage cotyledons, are formed by clearing the internal solid structure of seed. Due to the less germination rate of SDW-treated samples, clearing rate of the solid seed structure and the formation rate of the aforementioned internal seed regions were less than the other growth-promoting chemicals, which leads to a change of refractive index of the internal seed structure. Therefore, owing to the remaining internal solid seed structure, SDW-treated samples have a higher scattering coefficient and perform a higher normalized total intensity compared with butandiol and 1-hexadecene. Hence, this analysis plays a vital role by confirming the higher germination speed through the detection of less total intensity of an A-scan depth profile. Although a reduction of total intensity due to the morphological changes could be identified in all three samples, the samples treated with 1-hexadecene showed the intensity reduction earlier than the other samples, which confirms the highest germination rate.

Fig. 4

Consecutive detection of total intensity fluctuation analyzed using A-scan depth profiles. (a)–(d) The total intensity fluctuations of treated samples quantified on the first, third, sixth, and eighth experimental days.

3.4.

Thickness Quantification Based Germination Rate Analysis

According to the obtained morphological analysis, a gradual expansion of the distance between endosperm layer and cotyledons layer was identified along with the progress of germination. To gain a better understanding and a rigorous confirmation about the efficacy of the applied growth-promoting chemicals, we quantified the thickness between endosperm layer and cotyledons layer. The selection of the exact location to measure the thickness between endosperm and cotyledons layers was a challenging task due to the unflattened biological nature of the seed morphology. Therefore, to increase the accuracy of the measurements, we developed a program-based cropped window with 50 intensity signals (A-scan lines) to select the region of interest. The cropped window selects the intensity signals of the region of interest and detects the maximum intensity positions to linearly index the positions and average, which flattens the entire unflattened region of interest. Owing to the detection capability of the cropped window, distinguishable peak information representing endosperm and cotyledons layers was clearly identified. Thus, for each single sample, thickness was measured by considering the distance between averaged peak information. To increase the accuracy of the peak detection method and to verify the exact location, we obtained multiple 2-D OCT images (six images) from a single seed and averaged the measurements. Then we analyzed the thickness for all 20 seed specimens per each seed category and averaged to obtain precise thickness information.

The table shown in Fig. 5 shows the statistical values such as averaged thickness, standard deviations errors, and minimum and maximum thickness fluctuation range of three seed categories, which were obtained on the third, sixth, and eighth experimental days. The bar graph shown in Fig. 5 clearly verifies the thickness increase between the aforementioned two seed internal layers in all three seed categories, and the maximum thickness was identified in 1-hexadecene-treated samples, which confirms the accelerated germination rate compared with the other two categories. As a consequence, the results can be implemented as a thickness threshold range to obtain a rapid confirmation about the germination state of a seed by only measuring the thickness difference between the endosperm layer and cotyledons layer, which will be beneficial to select the most appropriate growth-promoting chemical for plant seeds. Therefore, the efficacy of the proposed method can be enhanced by analyzing the internal physical state of the seed at an initial stage using the performed precise quantification method.

Fig. 5

The thickness quantification between endosperm layer and cotyledons layer (the error bars indicate $\pm \text{standard deviation}$). The horizontal scale bars: $700\mu \mathrm{m}$ and vertical scale bars: $200\mu \mathrm{m}$.

3.5.

Three-Dimensional Optical Coherence Tomography Image Analysis

To evaluate the proposed method more rigorously and to emphasize the influence of each chemical component for the fluctuation of 3-D morphological structures with respect to the germination, we concluded the monitoring process by obtaining 3-D OCT images after 12 days of the chemical treatments. Figure 6 shows the 3-D top view, enface view of the seed middle region, and orthosliced cross-sectional view of the seeds along with the monitored morphological changes. The obtained enface images revealed the morphological changes that occurred owing to the growth-promoting chemicals. In particular, additional structural layers and morphological changes were observed in the samples treated with 1-hexadecene, such as in the embryo, storage cotyledons, radicle, and micropylar endosperm. Moreover, an obvious long sprout was identified in the sample treated with 1-hexadecene, confirming the suitability of 1-hexadecene as the optimal growth-promoting chemical to enhance the germination speed.

Fig. 6

3-D OCT images of seed morphology variations that occurred owing to germination. (a)–(c) 3-D, enface, and cross-sectional OCT images of an SDW-treated seed. (d)–(f) 3-D, enface, and cross-sectional OCT images of a seed treated with butandiol. (g)–(i) 3-D, enface, and cross-sectional OCT images of a seed treated with 1-hexadecene. C, cotyledons; E, endosperm; EM, embryo; H, hypocotyl; ME, micropylar endosperm; R, radicle; SC, storage cotyledons; SP, sprout; T, Testa (seed coat). The horizontal scale bars: $700\mu \mathrm{m}$ and vertical scale bars: $200\mu \mathrm{m}$.