Postharvest Biology and Technology 67 (2012) 118–123
Contents lists available at SciVerse ScienceDirect
Postharvest Biology and Technology journal homepage: www.elsevier.com/locate/postharvbio
Temperature effect on apple biospeckle activity evaluated with different indices Andrzej Kurenda ∗ , Anna Adamiak, Artur Zdunek Department of Microstructure and Mechanics of Biomaterials, Institute of Agrophysics PAS Do´swiadczalna 4, 20-290 Lublin 27, Poland
a r t i c l e
i n f o
Article history: Received 7 December 2010 Accepted 31 December 2011 Keywords: Biospeckle Apple Temperature
a b s t r a c t The physical explanation of the biospeckle phenomenon is well known; however, there is lack of biologically related interpretation, which limits a possible application of the method as a new, nondestructive technique for evaluation of fruit quality. Physically, biospeckles are the result of scattering of coherent light on moving particles inside living tissue. Almost all biological processes are temperature-dependent; therefore, the aim of this study was to investigate a temperature effect on biospeckle activity in apples and determine the extent to which this activity results from biological processes in the tissues of the fruit. Apples of ‘Idared’ cultivar were cooled in a storage room from high room temperature 29 ◦ C down to 2 ◦ C, which resulted in apple surface temperatures in the range 29–4 ◦ C. To evaluate the biospeckle activity, three methods of image analysis were used: correlation coefficient, speckle contrast, and the moment of inertia. The results showed that biospeckle activity measured by each method decreases with temperature decreases. However, the correlation coefficient was found as the most robust indicator of biospeckle activity and Q10 factor indicates a mostly biological basis for this phenomenon. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Fruits and vegetables are biologically variable and highly perishable (Johnston et al., 2002; Bobelyn et al., 2010). Hence, there is a need to evaluate their quality at different stages pre- and post-harvest in order to provide best quality product to consumers. Recently, a few interesting optical techniques and devices have been developed for nondestructive evaluation of fruits and vegetables: Vis/NIR spectrophotometry (Zude-Sasse et al., 2002; Zude et al., 2011; Mireei et al., 2010; Rutkowski et al., 2008), time-resolved reflectance spectroscopy (Zerbini et al., 2003), hyperspectral backscattering imaging (Qin and Lu, 2006; Peng and Lu, 2008) or laser-induced light backscattering (Qing et al., 2006; Baranyai and Zude, 2009). Nicolaï et al. (2007) have reviewed most of the above techniques, using the term NIR spectroscopy, and have shown their feasibility and areas where more research is still needed. Biospeckle is another optical technique, although less known, that has been developing for the last fifteen years for evaluation of properties of biological materials (Xu et al., 1995; Zhao et al., 1997). The physics of biospeckle is well developed. Laser speckle is an interference pattern of backscattered light observed, for example by a CCD camera, at a some distance from the illuminated object. If the sample does not show activity, the speckle pattern is stable in time. However, in the case of biological samples, the speckle
∗ Corresponding author. Tel.: +48 81 743 8558; fax: +48 81 744 5067. E-mail address:
[email protected] (A. Kurenda). 0925-5214/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.postharvbio.2011.12.017
pattern consists of two components: static, from nonmoving elements of an object, and variable in time, from moving particles within a material. This variable-in-time speckle pattern is characteristic for biological tissues and has been called a biospeckle (Zhao et al., 1997; Xu et al., 1995). So far, attempts to apply the biospeckle method in a biological study have included measurements of blood flow in blood vessels (Briers and Fercher, 1982), viability of seeds (Braga et al., 2003; Sendra et al., 2005), activity of parasites in living tissues (Pomarico et al., 2004; Braga et al., 2005), analysis of maturation and bruising of fruits and vegetables (Xu et al., 1995; Pajuelo et al., 2003; Rabelo et al., 2005), and, monitoring of apple shelf life (Zdunek et al., 2007, 2008). In many publications on the biological applications, one can find a statement that biospeckle activity corresponds to activity of biological samples; however, exactly what kind of activity is not specified anywhere or is only suspected. Braga et al. (2009) has summarized that processes related to movement of the scattering centers in the tissue such as cytoplasmic streaming, organelle movement, cell growth and division during fruit maturation, and biochemical reactions are responsible for certain biospeckle activity. Brownian motions should be considered as the source of biospeckle activity, too (Zhao et al., 1997). In the case of living organisms, whose operation is based on enzymes, the rate of biological processes strongly depends on temperature (apart from a few temperature-compensated ones). An example of that temperature-dependent process is the cytoplasmic streaming in Nitella obtusa and Elodea canadensis cells (Shimmen and Yoshida, 1993; Vorob’ev et al., 2004) that is suspected to be a biospeckle source. Therefore, in this study, we decided to
A. Kurenda et al. / Postharvest Biology and Technology 67 (2012) 118–123
investigate the relation of the biospeckle activity with temperature. The main objective of this experiment is to determine the extent to which the activity coefficients, currently used in measuring, are most useful for practical application, and whether biospeckle activity corresponds to the physical or biological phenomena. We have chosen apples, which were stored at temperatures used in postharvest technology, as a model object. The rate of processes based on enzymatic reactions is determined by the temperature coefficient Q10 , which describes a process activity gradient with temperature rise of 10 ◦ C (Hegarty, 1973; Xiao, 2000). When the process is independent of the temperature, the value of Q10 is 1. When the process depends on temperature, Q10 > 1. The value of Q10 for diffusion is approximately 1, for a typical chemical reaction 2, and for some metabolic reactions, even 3 (Kotov et al., 2007). Brownian motion velocity depends on the temperature in a linear way and its Q10 ratio is about 1.1 (Jia et al., 2007). Since temperature in the range of 5–29 ◦ C is studied in this paper, the temperature compensation phenomenon should be taken into account. Temperature compensation is a phenomenon observed in many living organisms and maintains a constant living activity in lower temperature compared to the optimal temperature (Ruoff et al., 2007). It is assumed that the mechanism of temperature compensation involves conjunction of many temperature-dependent metabolic reactions in a temperature-independent system based on feedbacks (Rajan and Abbott, 2007).
2. Materials and methods The device for biospeckle measurements was similar to that which was previously used by Zdunek et al. (2008) (Fig. 1). The system consists of a low power He–Ne laser (1 mW, = 632.8 nm, LLR 811, Optel, Opole, Poland), with a microscope objective PZO Poland 10/0.24, 160/- (as a beam expander) to illuminate the sample, and a CCD camera (Monochrome FireWire Astronomy Camera DMK 21AF04.AS, The Imaging Source Europe GmbH, Bremen, Germany) with a 25 mm objective with a 20 mm extension ring, as a detector of scattered light. The distance from camera to object was 37 mm and laser to object was 180 mm. The illumination angle was ≈ 30◦ . Biospeckle activity was recorded as uncompressed movies (AVI, 8 bits, RGB24 codec) lasting 4 s with 15 frames per second (fps) rate. The image resolution was 320 × 240 pixels, which corresponded to an observation area of 7.5 mm2 . In the observations with constant temperatures, movies were recorded for 95 min with lag of 5 min. In the observations with decreasing temperature, 4 s movies were registered at the moments when the sample temperature decreased by 1 ◦ C. The image exposure time of the CCD camera was 1/250 s. Gain and brightness of the CCD camera was found experimentally so as to avoid overexposures of pixels on the image histogram. The image acquisition settings were kept unchanged during the experiment. The experiment was carried out in a temperature range that suits both, the conditions in the orchard as well as those in which fruit can be stored and sold. Measurements were planned in three temperature regimes: 24 ◦ C (room conditions), 2.5 ◦ C (cold storage room, normal atmosphere), and gradually free-falling temperature from 29 ◦ C (warmer than the usual room temperature) to 2.5 ◦ C (cold storage room, normal atmosphere). In the latter regime, typically the fruit did not reach the lower target temperature. Storage at constant temperatures 2.5 ◦ C and 24 ◦ C was aimed at checking the stability of the measuring system and checking possible existence of natural oscillations of biological activity. The lowest temperature achieved by the fruit was about 4–5 ◦ C. When the samples reached the lowest temperature, they were kept for another 24 h to check if the temperature still decreased. 35 apples (Malus domestica Borkh
119
var. ‘Idared’) purchased at the local market were used for the experiment. The apples were not selected according to size or maturity stage, and no other treatments were performed. The fruit were conditioned for 24 h at the starting temperature for each temperature regime prior to measurements. The regime with gradual reduction of the temperature was performed on 15 apples, whereas the regimes with constant temperature involved 10 randomly selected apples. After preconditioning, a standard resistance temperature sensor, Pt 1000, was mounted approximately 1 mm under the apple skin on the opposite side of the fruit than the point of illumination. The temperature of the apples was measured with accuracy of 0.3 ◦ C by a digital ohmmeter using a temperature calibration chart. The temperature in the cold storage room was also monitored. 2.1. Indices of biospeckle activity Three indices characterizing the biospeckle activity were calculated from the biospeckle movies: (1) Cross-correlation coefficient Ck , where k = 0, 1, 2, . . . and = 1/15 s was calculated as the correlation coefficient of the data matrix of the first frame (k = 0) with the data matrices of the following frames (at k) from the analyzed biospeckle movie. In this study, C4 was analyzed only as the correlation coefficient between the first frame k = 0 and the frame at k = 4 s, because after this time, this parameter stabilized at a value corresponding to the biospeckle activity, as in Zdunek et al., 2008. Correlation coefficient Ck was determined by means of “corrcoef” function available in the Matlab® R2010a (The MathWorks, Inc., Massachusetts, U.S.A.) (2) Speckle contrast (SC) along time defined as the ratio of the standard deviation t of the intensity I to the temporal mean value of the intensities in each pixel I on the speckle pattern was t also used for describing biospeckle fluctuations. SC =
t I
(1) t
The moment of inertia (IM) which is based on the creation of the secondary image called the Time of the History Speckle Pattern (THSP). This display is formed by comparing, side by side, chosen rows extracted from registered frames. In our case, the line number 120 (the middle row in frame) was chosen for THSP formation. From THSP image, co-occurrence matrix COM is created:
COM = Nij
(2)
where the entries are the N number of occurrences of a certain intensity value i, that is followed by an intensity value j in the THSP. As a result of the normalization process of the co-occurrence matrix (Eq. (2)) a new, modified matrix is obtained, in which the sum of the components in each row equals 1. Mij =
Nij
(3)
Nij
j
Finally, the IM is calculated according to Eq. (4) (Arizaga et al., 1999): IM =
M ij (i − j)
2
(4)
ij
In summary, an increase in biospeckle activity corresponds to an increase in IM and a decrease in C4 and SC. C4 and SC can change in the range from 0 to 1.
120
A. Kurenda et al. / Postharvest Biology and Technology 67 (2012) 118–123
Fig. 1. Scheme of experimental setup for biospeckle activity measurements.
Table 1 Mean values of biospeckle activity indices with their percentage error expressed as standard deviation/mean value at constant temperatures of storage. Parameter
C4 in 4 ◦ C
C4 in 24 ◦ C
IM in 4 ◦ C
IM in 24 ◦ C
SC in 4 ◦ C
SC in 24 ◦ C
Mean value SD/mean (%)
0.741 0.41
0.484 1.80
96629 10.71
214700 7.07
0.538 3.15
0.425 2.23
C4 – correlation coefficient; IM – inertia moment; SC – speckle contrast; SD – standard deviation.
The temperature coefficient, Q10 , was calculated according to the equation: Q10 =
K 10/(T1 −T2 ) 1
K2
(5)
where K1 and K2 are rate constants of a biological process at temperatures T1 and T2 , respectively. Q10 was determined for all indices of biospeckle activity (Xiao, 2000). 2.2. Statistical analysis The experiment was performed on 10 or 15 apples that were treated as replicates. Mean values with standard deviations and other statistics were calculated with STASTICA® 8.0 (StatSoft, Inc. 1984–2008). One-way ANOVA was used for checking the temperature effect on the biospeckle activity indices. Significance of the effect was determined at p < 0.05. Post hoc analysis was performed using corrected variance ω2 . 3. Results Table 1 shows the mean values of biospeckle activity indices with their percentage error expressed as a ratio of standard deviation to the mean value at constant temperatures 4 ◦ C and 24 ◦ C. The most stable parameter determining the biospeckle activity was C4 , which was particularly evident at 4 ◦ C (SD/mean < 0.5% at 4 ◦ C and
