Laser-induced Fluorescence Spectroscopy for In-vivo Monitoring of Plant Activities

New spectroscopic method based on laser-induced fluorescence spectroscopy (LIFS) is proposed for the purpose of in vivo plant activity monitoring. To discuss LIFS application to plant monitoring, we developed three different LIF measurement systems which were used properly according to measurement applications. The basic configuration of the LIF system consisted of a laser, a spectroscopic device, a detector and a PC. Here, we report the systems’ details and some plant leaf monitoring results. 1) Tomato leaf growth monitoring by a mobile LIF spectrum monitoring system: LIF spectrum of tomato leaf was generally separated into two wavelength regions from 400 nm to 650 nm (blue-green LIF) and from 650 nm to 800nm (red & far-red LIF). It was suggested that a slope of the bluegreen intensity to the red & far-red one could be a parameter for evaluation of leaf growth process. 2) Vitality map of coffee tree leaf by a LIF imaging system: The system made possible to show an intensity distribution as a visualized LIF image with any wavelengths. The intensity was gradually decreased from the root part toward the top part of the leaf. Relation of the LIF intensity to plant vitality as effective light use inside the leaf was discussed. 3) Remote estimation of chlorophyll concentration of tree leaves by a LIF imaging lidar (light detection and ranging): It was found that the ratio of LIF intensity at 740nm to that at 685 nm had a linear correlation to chlorophyll concentration of leaf. Monthly variation in chlorophyll concentration of natural living tree leaves 65 m away from the system was remotely monitored. Through these experimental results, we discuss the effectiveness and adaptability of the LIFS idea and the usefulness of the developed systems for precise monitoring of agricultural-related plant activities. INTRODUCTION In vitro chemical treatments have been used to understand plant activities. This can offer very sophisticated physiological information, unfortunately it is not suitable for living materials, i.e. plant and agricultural products, because the treatment needs crash process. Also the system is generally designed for laboratory use. Otherwise in agricultural field, previous experience using information as size, shape, weight, color and so on judges growth stage of agricultural products and determines the harvest time. More objective criteria not depending on individual’s judgment are required. Plants use the sun light in any growth stage and after using the light energy for living activities, unused energy in that stage is reemitted outside as fluorescence. As Information and Technology for Sustainable Fruit and Vegetable Production FRUTIC 05, 12 – 16 September 2005, Montpel l ier France Precision Agriculture 700 fluorescence reflects the primary process of photosynthesis, information on plant growth status, health and disease, light use and so on can be obtained by detecting the fluorescence, which offers non-destructive in vivo plant monitoring. Idea of fluorescence use for remote plant monitoring was firstly indicated by Hemphill (1968) who used Fraunhofer Line Discrimination principle (FLD) (Kozyrev, 1956). Recent research of FLD plant monitoring is addressed to imaging (Saito et al., 2003) and aircraft monitoring (Moya et al., 2004). Stoll et al. (1999) proposed a space mission for screening vegetated area. The FLD is a passive monitoring technique. Active fluorescence, i.e. laser-induced fluorescence (LIF), monitoring technique has been also studied. Usage of laser source in place of the sun light induces fluorescence effectively. Laser’s single wavelength as an excitation source makes it easy to analyze fluorescence data. Measures et al., (1973) reported blue fluorescence detection from sugar maple tree at range of 542 feet by a LIF lidar (light detection and ranging). McFarlane (1980) applied LIF information to plant stress detection. Especially, UV laser-induced LIF has been widely studied in plant biology (Cerovic et al., 1999). In this paper, we describe our developed LIF measurement systems first, then show LIF spectra and LIF images of plant leaf, and finally through the experimental results, we discuss the effectiveness and adaptability of the laser-induced fluorescence spectroscopy and the usefulness of the developed systems for precise monitoring of plant activities. METHODS AND MATERIALS Mobile LIF Spectrum Monitoring System and Tomato Seedling A picture of a mobile LIF spectrum monitoring system is shown in Fig. 1. A LIF excitation source was a pulsed Nd:YAG laser with wavelength of 355 nm, pulse energy of 0.2 mJ, pulse width of 10 ns, beam diameter of 6 mm and repetition rate of 10 Hz. The laser beam was delivered by a transmitting fiber to a tomato leaf. LIF from the tomato leaf was collected and sent to a monochromator with a receiving fiber and the output from the monochromator was detected with an intensified CCD detector. Outlet of the transmitting fiber and inlet of the receiving fiber was set in one united holder. The holder was gently pressed on the leaf through a soft rubber to prevent incoming ambient light. Tomato seedlings were cultivated in a greenhouse of Research Institute of KAGOME Co., Ltd (Nishinasuno-machi, Nasu-gun, Tochigi, Japan). LIF Imaging System and Coffee Tree A picture of a LIF imaging system is shown in Fig. 2. A sample leaf set in a dark box was irradiated by a violet laser diode with wavelength of 398 nm, average (CW) power of 30 mWmax. The beam diameter was expanded by a beam expander to cover the whole area of the leaf. A liquid crystal tunable filter was used for selection of a desired wavelength from broad LIF spectrum of the leaf. The wavelength selection from 420 nm to 750 nm could be easily and rapidly done by voltage tuning. This tuning method was ideal for image analysis, because it had no mechanical moving which sometimes made difficulties for precise positioning of pixels . An image intensified CCD camera (1024 x 1024 pixels) detected LIF and made it visible as an image with the selected wavelength. Sample was a coffee tree leaf. Coffee trees were cultivated at UCC UESHIMA COFFEE Co., Ltd. Hawaii Farm (Kona, Hawaii, U.S.A.) under natural weather conditions. Information and Technology for Sustainable Fruit and Vegetable Production FRUTIC 05, 12 – 16 September 2005, Montpel l ier France Precision Agriculture 701 LIF Imaging Lidar System and Natural Tree A picture of a LIF imaging lidar system (Saito et al., 2002) and experimental configuration are shown in Fig. 3. The system was developed to investigate the feasibility of long-range plant monitoring. A pulsed Nd:YAG laser beam coming from a laboratory building irradiated target trees. Laser was with wavelength of 532 nm, pulse energy of 10 mJ, pulse width of 6 ns, beam diameter of 6 mm and repetition rate of 10 Hz. The beam was magnified by a negative lens so that the beam could cover the whole area of the target tree. LIF from tree leaves was collected by a 42 mm diameter camera lens, and in front of the lens three band-pass filters were inserted alternately to obtain chlorophyll fluorescence (685 nm and 740 nm) and scattered light (532 nm) from the leaves. They were detected by a gated image intensified CCD camera (510 x 492 pixels). The gate with 40 ns was opened at 413 ns after the laser was fired. This synchronized delay setting with only a short opening period allowed the system to obtain weak fluorescence signal only from the tree and to reduce ambient light. The target trees were poplar (Populus nigra var. italica), ginkgo (Ginkgo biliba Linn) and hiba (Thujopsis dolabrata var. Hondae Makino) growing under natural conditions in our campus, which was located about 65 m away from the system. RESULTS AND DISCUSSION Growth Stage Monitoring of Tomato Leaf Using the mobile LIF spectral monitoring system, LIF spectrum of the tomato leaf was measured. The spectrum shown in Fig. 4 was separated into two wavelength regions; blue-green fluorescence which was shorter than 650 nm and red & far-red fluorescence which was from 650 nm to 800nm with two peaks at 685 nm and 740 nm. The same spectral shapes also obtained in many plant leaves (Saito et al, 1998), so the spectrum shape was a common feature of plant leaf. Figure 4 shows a slope between the blue-green intensity (peak) to the red & far-red one. The leaf at the middle part of the seedling had larger gradient, and at the bottom part the slope took the opposite gradient. It was suggested that the top part leaves were under growing, the middle part ones were mature, the lower part ones were older, and the bottom part ones were withered or dead. Cerovic et al. (1999) reported that main origins of the blue-green fluorescence were ferulic acid derivatives, other phenylpropanoids, and NAD(P)H, and that of the red & far-red fluorescence was chlorophyll a. The slope reflected the balance of those pigments which changed according to the growth stage. Direction of the slope (positive or negative) together with the entire LIF shape can be a good parameter for estimation of the leaf growth process related to those pigments. Vitality Map of Coffee Tree Leaf LIF images with different wavelengths of a coffee leaf are shown in Fig. 5. The wavelengths of 460 nm, 685 nm and 740 nm corresponded to the featured one of LIF spectrum (see Fig. 4). Each image at each wavelength contains information on pigments mentioned above. A distribution pattern in the LIF intensity image was clearly seen even in one leaf. This means that living status was different depending on the leaf area. As a common feature, the intensity in every image gradually decreased from the root part toward the top part of the leaf. Two possible considerations on this were considered; 1)if fluorescence is a process of dissipation of absorbed light energy which should be used for photosynthesis activity, the leaf top where emitted fluorescence was very low used the Information and Technology for Sustaina