Assessment of biological leaf tissue using biospeckle laser imaging technique

2.1.1. IM of THSP.

The biospeckle data was numerically evaluated using the biospeckle index (the IM). In each leaf sample, the region of interest (ROI) was selected and analysed by means of the IM index; this was followed by the creation of a THSP and then a co-occurrence matrix (COM). Figure 1 shows an example of the formation of THSP and COM matrices from raw speckle frames of a biospeckle video sequence.

Zoom In Zoom Out Reset image size

The IM index can then be computed as [3, 20, 21]

$$\begin{align} \newcommand{\e}{{\rm e}} \displaystyle {\rm IM}=\sum\limits_{i,j}{{\rm CO}{{{\rm M}}_{i,j}}{{\left| i-j \right|}^{2}},}\nonumber \end{align} \tag{ 1 }$$

where COM is related to the THSP, and i and j represent the dimensions of the COM matrix [21].

The activity of the biological sample was evaluated using the methods of Fujii, generalised difference (GD) [5], and the temporal LASCA [18]. These processing methods are based on the intensity variations in the biospeckle sequence pattern. The measurements are defined as functions of a sequence of n integer values $s=\left[ {{x}_{0}},{{x}_{1}},...,{{x}_{n-1}} \right]$, as illustrated in figure 2.

Zoom In Zoom Out Reset image size

The Fujii function (F) for a sequence ${{s}_{p}}=\left[ {{x}_{0}},{{x}_{1}},...,{{x}_{n-1}} \right]$ can be defined as follows [5]:

$$\begin{align} \newcommand{\e}{{\rm e}} \displaystyle F\left(s \right)=\sum\limits_{i=0}^{n-1}{\frac{\left| {{x}_{i}}-{{x}_{i+1}} \right|}{{{x}_{i}}+{{x}_{i+1}}}}.\nonumber \end{align} \tag{ 2 }$$

The GD can be expressed by

$$\begin{align} \newcommand{\e}{{\rm e}} \displaystyle G\left(s \right)=\sum\limits_{i}{\sum\limits_{j}{\left| {{x}_{i}}-{{x}_{i+j}} \right|}},\nonumber \end{align} \tag{ 3 }$$

where i and j are indices spanning all possible numbers in the sequence ${{s}_{p}}=\left[ {{x}_{0}},{{x}_{1}},...,{{x}_{n-1}} \right]$.

2.2.1. LASCA using temporal correlation (tLASCA).

Temporal speckle contrast is a method in which averaging in the spatial domain is performed [18, 22]. For each frame, the contrast value ${{C}_{t}}$ of pixel (i, j) of a particular frame can be computed by using

$$\begin{align} \newcommand{\e}{{\rm e}} \displaystyle {{C}_{t}}\left(i,j \right)=\sum\limits_{r=i-1}^{i+1}{\sum\limits_{k=j-1}^{\,j+1}{\frac{{{\delta }_{i,j,t}}}{\left\langle {{I}_{i,j,t}} \right\rangle }}}.\nonumber \end{align} \tag{ 4 }$$

Here, ${{\delta }_{i,j,t}}$ is the standard deviation of all pixels at (i, j), and $\left\langle {{I}_{i,j,t}} \right\rangle$ is the mean intensity of all pixels at (i, j) of n frames along the temporal dimension. This method is typically used to distinguish the contrast between slow- and fast-moving speckles, which correspond to the mean and standard deviation, respectively.

Two different plants leaves, the P. rubra leaf and E. aureum leaf, were selected for the biospeckle study. Five leaves with almost similar shape and size from each plant were taken. The torn leaves were then used for the biospeckle laser imaging.

Two basic experimental optical setups were employed for the biospeckle assessment of the leaf tissue. Figure 3 shows the first experimental configuration, which is normally used to assess the biospeckle in Fraunhoffer's plane. This geometry allows for quantitative evaluation of the leaf tissue. Figure 4 shows the second configuration in which the phenomenon is followed in the image plane; this is also used for the quantitative evaluation of the specimen. In figure 5, we demonstrate a third backscattering configuration that allows for the study of the spatial distribution of activity across the leaf tissue.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

A He–Ne laser of 2 mW operating at 632.8 nm was used to illuminate the leaf surface. The laser beam was spatially filtered and expanded to cover the ROI of the specimen. A CCD camera (Basler, Germany), with a resolution of 1294  ×  964 pixels (pixel size of 3.75 square pixels) and having a frame rate of 32 fps, was used to collect the biospeckle data. The camera exposure time was set to 8.9 ms.

For the qualitative evaluation of the leaf tissue, a set of 500 speckle images (500  ×  500 pixels) were acquired at every 0.08 s. However, a raw speckle data containing 256 images (256  ×  256 pixels) was used for the quantitative evaluation of the leaf tissue.

After harvest from the plants, leaves present different degrees of activity as a result of several phenomena, including evaporation, exchanges across cell membranes, attacks of microorganisms, micro flow, etc. This activity can be effectively obtained using the biospeckle optical technique.

The biospeckle pattern on a leaf tissue varies in the space and time dimensions. The temporal information can be extracted by generating a 2D map called the THSP, where the vertical direction represents the time evolution of the phenomenon. For a specimen showing high activity, the corresponding biospeckle pattern quickly evolves with time, and the THSP image resembles a pattern of a general speckle. However, for samples displaying low activity, the THSP pattern becomes similar to parallel vertical bars.

The evolution of activity can be obtained through the generation of a matrix called the COM, and the subsequent computation of the numerical index IM. In the COM matrix, activity is obtained through the points spread around its principle diagonal. For a sample showing high activity, the points on the COM matrix become more spread.

Figure 6 shows a bright field image of an E. aureum leaf. The leaves were submitted for biospeckle illumination as shown in figure 3. For a quantitative evaluation of the leaf, an ROI was selected from the tissue, and raw speckle frames were collected from the given ROI (as shown in figure 6). From figure 6, one can see the time evolution of a single pixel intensity as well as the mean intensity variation on the given ROI.

Zoom In Zoom Out Reset image size

Figures 7 and 8 respectively present the THSP image and corresponding COM matrix for E. aureum and P. rubra leaves. As can be seen, the THSPs and COM matrices for the leaves are distinctive and can be easily differentiated.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 1 shows the computed values of the IM index for the two biological leaves. As can be seen, the IM values for both leaves are statistically significant and can be discriminated. The IM index can therefore be employed as a biospeckle numerical index for the biological signature to a particular leaf species.

From figure 9, one can see the statistical plot of the IM index of the two leaves. The values are significant and statistically different for both leaf samples.

Zoom In Zoom Out Reset image size

Three infected (due to microbiological organism attacks) E. aureum leaves were collected from the plant and submitted for biospeckle imaging.

Figure 10 shows the spatial distribution of biospeckle activity of E. aureum leaf tissue generated using the methods mentioned in section 2.2. By following the activity maps, one can easily identify the infected regions within the leaf. The regions infected by microbiological organisms in different proportions have been highlighted within the circles, and show a lower biospeckle activity (presented in blue colour).

Zoom In Zoom Out Reset image size

Figure 11 presents a time evolution of the biospeckle intensity computed within the healthy and infected regions (as indicated by dotted yellow squares in figure 10). In comparison to the healthy region, intensity fluctuation within the infected zone is almost constant, which causes a lower biospeckle activity as can be observed using the activity maps in figure 10.

Zoom In Zoom Out Reset image size

Using the tLASCA, Fujii, and GD images, one can easily observe the infected zones within the tissue. However, in the image generated using the tLASCA algorithm, it is not visible to identify the activity of the midvein. This activity has clearly been identified in the Fujii and GD activity maps (see figures 10(c) and (d)). The activity of the vein corresponds to the internal microflow of fluid through the vein.

Five E. aureum leaves were collected from the plant. Leaves exhibiting an initial degree of infection caused by microorganism attack were selected and used for laser biospeckle imaging.

Figure 12 presents the biospeckle activity distribution in an E. aureum leaf tissue showing an initial degree of infection. The maps were generated using the methods mentioned in section 2.2. As can be observed, the regions containing initial infection were clearly identified, reflecting a lower biospeckle activity. The infected region is highlighted by circles, and is hardly visible in the original leaf (figure 12(a)).

Zoom In Zoom Out Reset image size

One can easily differentiate the infected region within the tissue in all activity images computed using the different methods. Moreover, using the Fujii and GD maps, it is easy to observe the higher activity of the midvein caused by internal fluid flow. Conversely, in the tLASCA image this activity is hardly visible.

With the passage of time, microbes continue to consume nutrients from the plant. This results in the infected region drying rapidly, which then causes a lower biospeckle activity.