REMOTE MONITORING OF VEGETATION BY SPECTRAL MEASUREMENTS
AND MULTI-COLOUR FLUORESCENCE IMAGING
M. Andersson, H. Edner. J. Johansson, P. Ragnarson, S. Svanberg and E. Wallinder
Department of Physics, Lund Institute of Technology
P.O. Box 118, S-221 00 Lund, Sweden
ABSTRACT:
A brief review of remote monitoring of laser-induced fluorescence from vegetation is given together with
examples of results from measurements during field campaigns within the European LASFLEUR project.
A mobile fluorescence lidar system was used in the Swedish activities. Measurements were performed in a
spectrally resolving mode yielding a full fluorescence spectrum in one selected point, and in an imaging
mode, where spatially resolved data are captured in four selected spectral bands simultaneously. As a
transmitter a Nd:YAG laser system operating at 20 Hz was used, either frequency tripled with an output at
355 nm, or, preferentially, Raman shifted to obtain an output wavelength at 397 nm. A 40 cm diameter
telescope was used to collect fluorescence photons, that were analysed either in an optical multichannel
analyser system or in a four-colour imaging system, both equipped with intensified CCD detectors. This
system has been used in field campaigns during the last few years in Italy, Germany and France,
monitoring several species, such as beech, spruce and maize. Detailed studies on spruce and maize have
been made to identify the optimum excitation wavelength. From the results obtained, and also taking into
consideration eye-safety regulations, 397 nm was chosen as the optimum wavelength. The daily cycle of
these species was studied for samples under parallel plant physiological control. Maize grown under
different conditions with regard to nutrition and water supply was also investigated. Fluorescence imaging
was performed with a maximum distance of 100 m. Detailed studies were made on maize plants at a
distance of 40 m. Four fluorescence images were recorded at selected wavelengths, including the two
chlorophyll peaks at 685 and 735 nm. The simultaneous recording of fluorescence images makes the
system less vulnerable to wind movements of the vegetation. Sequential single-colour fluorescence
imaging, yielding sharper images was also evaluated.
KEY WORDS: multi-colour imaging, laser-induced fluorescence, vegetation, remote sensing, spectral
analysis 1
1. INTRODUCTION
Several areas in Europe exhibit damage to the vegetation due to environmental pollution of air, soil and
water. It is of considerable interest to be able to perform early detection and mapping of the damaged
vegetation. One possibility to map large areas in a short time is the use of reflectance spectroscopy
performed with satellite-borne multi-spectral imagery [1]. Active remote sensing using a transmitter can
provide additional information [2]. On excitation with laser light in the UV or blue region, vegetation
exhibits characteristic fluorescence from chlorophyll. The chlorophyll gives rise to two red fluorescence
peaks, one at 685 nm and one at 735 nm. The relative intensity of these peaks is a measure of the
physiological status of the plant [3]. In addition, a broadband fluorescence in the 450-600 nm region is
also obtained [4]. The molecules giving rise to this fluorescence are not yet fully identified, but several
fluorophores, carotenoids [5], riboflavin [5], cinnamic acids [ 6 ], coumarine [ 6 ] and NADPH [7,8] have
been suggested. Also, contributions from the surface wax layer can be seen in the blue region. The
possibility to use this blue fluorescence to decide on the plant status is investigated by several groups [9].
Swedish remote fluorescence work started in 1978. Early experience is described in [10].
Three years ago the European LASFLEUR project was started. The aim of the collaborative
project is to develop an airborne fluorosensor that can detect early damage on vegetation. Within the
LASFLEUR project several field campaigns have been performed. The participation from the Swedish
side has been in Pian di Novello (I) [11], Oberpfaffenhofen (D) [12], Karlsruhe (D) [13] and Avignon (F).
Our experience from this work is presented below, with special focus on the Avignon campaign,
September 1993.
