CHAPTER 1. GENERAL INTRODUCTION
1.1 INTRODUCTION
Near infrared (NIR) spectroscopy, a nondestructive sensing method based on specific
absorptions within a given range of wavelength corresponding to the constituents in the
sample (McClure, 2003), has been widely applied for the evaluation of internal quality
of agricultural products (Davey et al., 2009; Lebot et al., 2011). Since NIR spectra of a
mixture is the linear summation of individual spectra of the constituents in the mixture,
such a mixture spectra thus can be regarded as ‘blind sources’ (Hyvärinen et al., 2001)
as the proportion of constituents in the samples remains unknown. Many attempts have
been made in recent years to extract critical features from the spectra using multivariate
analysis (Blanco and Villarroya, 2002; Burns and Ciurczak, 2008), including multiple
linear regression (MLR) (Chang et al., 1998), principal component regression (PCR)
(Wold, 1987), and partial least squares regression (PLSR) (Wold et al., 2001). However,
these methods were not designed for resolving the ‘blind source’ problem and may not
correlate well with the properties of constituents in the mixture, consequently hindering
the applicability of the spectra for chemometric analysis of the constituents
(Al-Mbaideen and Benaissa, 2011; Chen and Wang, 2001; Kaneko et al., 2008).
A multiuse statistical approach originally used to implement ‘blind source separation’
in signal processing (Herault and Jutten, 1986; Vittoz and Arreguit, 1989), independent
component analysis (ICA) is capable of disassembling the mixture signals of Gaussian
distribution into non-Gaussian independent constituents with only a little loss of
information and does not require any information to be added to the source (Comon,
1994). In practice, multiple ICA algorithms have been developed, including JADE
algorithm (joint approximate diagonalization of eigenmatrices) (Cardoso and
Souloumiac, 1993; Cardoso, 1999) and FastICA algorithm (Hyvärinen and Oja, 1997;
Hyvärinen, 1999), making ICA a high-speed and reliable tool (Hyvärinen and Oja, 2000)
for analytical chemistry (Lathauwer et al., 2000; Al-Mbaideen and Benaissa, 2011),
biomedical signal processing, telecommunications, econometrics, audio processing, and
image processing (Hyvärinen et al., 2001).
Application of ICA for spectrum analysis has been demonstrated by Chen and Wang
(2001) to separate the pure spectra of various constituents from the NIR spectra of the
mixture and to build qualitative relationship between the estimated independent
components and the constituents. Such a capability also enabled complete explanation of the constituents’ properties for NIR qualitative analyses (Westad and Kermit, 2003).
In addition, ICA was used to obtain statistically independent and chemically
interpretable latent variables (LVs) in multivariate regression (Gustafsson, 2005). It was
also noted that the number of independent components extracted from the spectra of
mixtures is related to the performance of ICA (Westad, 2005). Moreover, ICA was
employed to identify the infrared spectrum of mixtures containing two pure materials
(Hahn and Yoon, 2006) as well as the constituents in commercial gasoline (Pasadakis
and Kardamakis, 2006; Kardamakis et al., 2007). Equally noteworthy is that the
calibration model built through MLR, after using ICA to extract independent
components of aqueous solutions, gave good predictability (Kaneko et al., 2008),
whereas NIR estimation of sucrose concentration (Chuang et al., 2010) and glucose
concentration (Al-Mbaideen and Benaissa, 2011) were enhanced by using ICA.
While application of ICA for spectral analysis appears promising, available literature
still focuses mainly on chemical samples or non-natural products, such as (1)
identification of constituents in the mixture, especially for chemical compounds (Chen
and Wang, 2001; Hahn and Yoon, 2006; Pasadakis and Kardamakis, 2006; Kardamakis
et al., 2007; Liu et al., 2008; Kaneko et al., 2008; Al-Mbaideen and Benaissa, 2011); (2)
a preprocessing method for improving predictability of calibration model (Zou and
Zhao, 2006); and (3) combination of ICA and other multivariate analysis methods, such
as PCA-ICA (Pasadakis and Kardamakis, 2006), ICA-MLR (Kaneko et al., 2008; Liu et
al., 2009), ICA-PLS (Liu et al., 2009), ICA-LS-SVM (Wu et al., 2008) and ICA-NNR
(Fang and Lin, 2008) to deal with linear or nonlinear problems. However, no literature
exists by using ICA with NIR spectroscopy as the sole tool to quantify internal quality
or constituents of biomaterials without any other assisted multivariate analysis methods.
The applicability of ICA for quantitative inspection of biomaterials thus should be
evaluated and studied. In this dissertation, ICA was first applied as the sole tool in
conducting NIR quantitative analyses of biomaterials, including wax jambu fruit (see
CHAPTER 2), medicinal plant Gentiana scabra Bunge (see CHAPTER 3 and 4), and
milled white rice (see CHAPTER 5), to evaluate the applicability of this method.
Influence due to various types of sample (sucrose solution, intact fruit, dry powder of
Gentiana scabra Bunge, and cargo rice) was also studied.
1.1.1 WAX JAMBU
Wax jambu (Syzygium samarangense Merrill & Perry), an endemic fruit in Taiwan
and parts of southeast Asia has very unique surface and texture that are easily bruised or
damaged, hence requiring wax jambu to be handled delicately from harvest to shipping
and distribution. To date, several researches aimed to develop a non-invasive and rapid
detection method for the analysis of internal quality of wax jambu (You, 2002; Lin,
2002; Chung et al., 2004). For further applications of ICA with NIR spectroscopy in the
inspection of fruits, wax jambu is suitable to serve as an example for discussion. In the
present study, ICA was integrated with NIR spectral analysis to quantify the sugar
content in intact wax jambu. The results of wax jambu were also compared with those
of sucrose solutions –– mixtures of sucrose and de-ionized water. Spectral pretreatments
and linear regression were then used to build spectral calibration models of sugar
content. The analysis results of ICA were also compared with those of PLSR to assess
the abilities in predicting sugar content in wax jambu.
1.1.2 GENTIANA SCABRA BUNGE
Medicinal plants have always been considered an important and reliable source of
pharmacy, since they are rich in many bioactive components. The international trade
market for medicinal plant products continues to expand and covers food, beverages,
drugs, cosmetics, and skin care products. Gentiana scabra Bunge, a perennial
herbaceous plant, is mainly grown in temperate regions such as Taiwan, China, Japan,
South Korea, and Russia. Dried root and rootstock of Gentiana scabra Bunge are
commonly used as pharmaceutical raw materials, since they are rich in many
secoiridoid glycosides such as gentiopicroside, swertiamarin and sweroside (Kakuda et
al., 2001). In particular, gentiopicroside has been shown to protect liver, inhibit liver
dysfunction, and promote gastric acid secretion in addition to its antimicrobial and
anti-inflammatory effects, making it a popular ingredient in Chinese herbal medicine
and health products (Kim et al., 2009).
In early days, Gentiana scabra Bunge was mainly collected from the wild. As the
demand for Gentiana scabra Bunge increases, the wild resources diminish gradually,
thus restoration of Gentiana scabra Bunge became an important issue (Zhang et al.,
2010). Studies in recent years used tissue culture technology to cultivate of Gentiana
scabra Bunge (Cai et al., 2009), by domesticating the tissue culture of Gentiana scabra
Bunge, then transplanting it to the greenhouse for cultivation. In order to monitor the
change of Gentiana scabra Bunge during the growth process, it is necessary to measure
the bioactive components of Gentiana scabra Bunge. However, the commonly used
methods such as micellar electrokinetic capillary chromatography (MECC) (Glatz et al.,
2000), high performance liquid chromatography (HPLC) (Szücs et al., 2002; Kikuchi et
al., 2005; Carnat et al., 2005; Kušar et al., 2010; Hayta et al., 2011a; Hayta et al.,
2011b), liquid chromatography-mass spectrometry (LC-MS) (Aberham et al., 2007;
Aberham et al., 2011), and ultra-performance liquid chromatography (UPLC)
(Nastasijević et al., 2012) are all time-consuming and energy-intense, hence cannot be
applicable for daily quality inspection of Gentiana scabra Bunge during cultivation.
NIR spectroscopy has been widely used in dispensation, such as herbal component
analysis of Chinese herbal plants Angelicae gigantis Radix (Woo et al., 2005), Rhubarb
(Zhang and Tang, 2005), licorice (Glycyrrhizia uralensis Fisch,) (Wang et al., 2007),
Panax Species (Chen et al., 2011), and Lonicera japonica (Wu et al., 2012), as well as
the content detection of active pharmaceutical ingredients (APIs) in tablets (Paris et al.,
2006; Jamrógiewicz, 2012; Porfire et al., 2012). However, it has not been employed to
qualitatively monitor the growth of Gentiana scabra Bunge. In recent years, ICA has
been used in medicinal tests (Fang and Lin, 2008; Wang et al., 2009; Shao et al., 2009).
Considering there hasn’t been any study applying NIR spectroscopy in inspection on
internal components of Gentiana scabra Bunge currently, it is the intent of this study to
apply ICA, which could analyze various components simultaneously, in NIR
spectroscopy analysis on gentiopicroside and swertiamarin to discuss qualitative and
quantitative relationships of the two bioactive components. Efforts were also made to
build spectral calibration models with high predictability in order to evaluate the
potentiality of NIR for quality inspection on Gentiana scabra Bunge.
1.1.3 RICE
Rice is one of the most important and popular food crops in the world, and freshness
of rice depends on the storage conditions such as storage time, storage temperature, and
relative humidity. Among them, the storage time of rice has an enormous effect on its
appearance, flavor, and quality of the nutrients (Zhou et al., 2002). Previous studies
demonstrated that most lipids in rice hydrolyze into free fatty acids and cause the acidity
of rice to increase with prolonged storage (Takano, 1989; Hu, 2011; Chen et al., 2011).
Therefore, the determination of rice freshness is one of the main goals in site
examination. There is a strong need to develop a non-invasive, rapid detection method
for the analysis of freshness. Therefore, the objective of the current study was to inspect
rice freshness in terms of qualitative and quantitative approaches using NIR
spectroscopy. Rice freshness was expressed by both pH value and fat acidity. The pH
values were determined by bromothymol blue - methyl red (BTB-MR) method (Hsu
and Song, 1988) and fat acidity by AACC International method 02-02.02 (AACC
International, 2000). By means of a calibration curve, a relationship between pH and fat
acidity was established. ICA was subsequently integrated with NIR spectral analysis to
quantify the pH in rice. Linear regression was then used to build spectral calibration
models of pH value.