SPIE Membership Get updates from SPIE Newsroom
  • Newsroom Home
  • Astronomy
  • Biomedical Optics & Medical Imaging
  • Defense & Security
  • Electronic Imaging & Signal Processing
  • Illumination & Displays
  • Lasers & Sources
  • Micro/Nano Lithography
  • Nanotechnology
  • Optical Design & Engineering
  • Optoelectronics & Communications
  • Remote Sensing
  • Sensing & Measurement
  • Solar & Alternative Energy
  • Sign up for Newsroom E-Alerts
  • Information for:
SPIE Photonics West 2019 | Call for Papers

2018 SPIE Optics + Photonics | Register Today



Print PageEmail PageView PDF

Remote Sensing

Using polarization to separate overlapping spectra of elastic reflectance and fluorescence

The chlorophyll fluorescence spectrum of algae in seawater can be extracted from the overlapping elastic-scattering spectrum by exploiting the polarization of the scattered signal.
20 April 2006, SPIE Newsroom. DOI: 10.1117/2.1200603.0166

Techniques for determining coastal-zone chlorophyll-a concentrations (Chl) using remote-sensing spectra are frequently based on the analysis of features in the near-infrared1 (NIR). This is because of the deficiencies the simple blue-green ratio approach, which exploits the fact that the ratio of blue and green irradiance at the sea surface shows a good correlation with chlorophyll concentration in open oceans. It fails when the strong scattering of mineral particles and the absorption of colored dissolved organic matter(CDOM) are present. However, depending on the water composition, chlorophyll fluorescence and the reflectance peak in the NIR band may overlap.2 The accuracy of [Chl] retrieval therefore depends on the ability to distinguish the contributions of fluorescence in the NIR reflectance from those due to elastic scattering (interactions with matter that don't change the light's frequency). Both contributions can vary significantly: particularly in coastal waters, where mineral particles and CDOM can substantially affect scattering and absorption.3

To date, there have been no good experimental techniques for separating elastic scattering and fluorescence when the two spectra overlap. Approaches using a series of filters in conjunction with spectrometers4 are cumbersome and incomplete, and therefore prone to inaccuracy. A baseline subtraction method5 has also been used to extract the chlorophyll fluorescence from the ocean reflectance, but the interplay of water and chlorophyll absorption can result in apparent spectral peak features near the fluorescence peaks (around 685nm), leading this method to overestimate the fluorescence signal.

We have developed and applied a new technique that uses polarization information to separate overlapping spectra of elastic reflectance and algae fluorescence in seawater.5 We recently showed that it is possible, using the technique, to separate the chlorophyll fluorescence signal that peaks at 685nm from elastic scattering in the water, thus leaving radiance.6

The basic set-up to implement the approach in a laboratory setting is shown in Figure 1. White light is guided by fibers and collimated to illuminate water that contains algae. The signal returned from the water, R, is passed through a polarizer, collected, and delivered by a fiber bundle into a spectrometer.

Figure 1. Shown is the experimental set-up. L: lens. FP: fiber probe. P: polarizer. C: sample. i1, i2: incident, detection angles. ∑: scattering angle.

R consists of two components: elastic scattering Rs, and chlorophyll fluorescence Fl:


Except in the forward (0°) and backward (180°) directions, elastic scattering is always at least partially polarized, even if the incident light is unpolarized. Fluorescence is generally unpolarized. The degree of polarization reaches a maximum when the angle between the incident illumination and detector axes, i1 + i2, is 90°, which is the optimum situation for the technique. The detected reflectance signal will change from maximum to minimum as the polarizer is rotated:


where // and ∑ represent the parallel and perpendicular polarized components of the scattered light, with respect to the scattering plane formed by the incident and reflected light path. Note that R = Rmin + Rmax, and RS = R// + R. Subtracting (3) from (2) gives an expression that eliminates the fluorescence Fl:


Generally, it can be shown that strong correlations exist between RS and RD, allowing an estimate of RS based on the fluorescence-free RD. A linear regression is used to fit R with RD outside the fluorescence region, where R is, in principal, identical with RS. The resulting relationship,RS = A RD + B, is then extended across the fluorescence region to estimate the scattering component there. The fluorescence Fl can then be obtained by subtracting the estimated scattering signal RS from the total reflectance R: Fl = R - RS

This technique has been successfully applied in the laboratory to a wide variety of algae species. Figure 2 shows the spectrum estimate for Tetraselmis striata fluorescence. The retrieved spectrum is a good match for the laser-induced fluorescence spectrum of the same species, also shown in Figure 2. Similar good matches were achieved for a wide variety of water conditions, including appreciable surface roughness (as long as sufficient averaging is performed), and high concentrations of mineral particles.7

Figure 2. Shown are the detected and estimated spectra for algae Tetraselmis striata. λ: wavelength. R, RS: signal, estimated elastic scattering spectra. Fl, Fllaser: estimated, laser-induced fluorescence spectra.

Figure 3 shows spectra and fluorescence estimates based on measurements in Shinnecock Bay, Long Island, NY, in June 2004, where the chlorophyll concentration was about 8∫g/l. The slight negative values are due to imperfect curve-fitting stemming from small non-linearities between the two orthogonally-polarized components outside the fluorescence band. However, as can be seen, the shape and band of the fluorescence peak conform to laboratory measurements.

In conclusion, we have developed a technique that separates the chlorophyll fluorescence spectrum of algae in seawater and the overlapping elastic-scattering reflectance spectrum by exploiting the polarization of the scattered light. The excellent agreements between fluorescence spectrum shapes from laboratory application of the technique with laser-induced fluorescence of the same species confirm the technique's validity. Laboratory and in-situ measurements under various water conditions show that this technique is effective for extracting chlorophyll fluorescence in the 685nm region from reflectance in the case of algae dominated by chlorophyll pigments.

Figure 3. Shown are the detected and estimated spectra (in arbitrary units) for measurements in Shinnecock Bay. Rmax, Rmin: maximum, minimum signal spectra under polarizer rotation.

Samir Ahmed, Jing Zhou, Alexander Gilerson, Fred Moshary, Barry Gross
Optical Remote Sensing Lab, Department of Electrical Engineering, The City College of City University of New York
New York, NY
Samir Ahmed is the Herbert G. Kaiser Professor of Electrical Engineering at the City College of the City University of New York and Director of the Optical Remote Sensing Lab, which is partially funded by contracts under the NOAA-CREST and NASA COSI centers.
Dr. Jing Zhou works as a research associate at Optical Remote Sensing Lab at the City College of the City University of New York. Her research interests mainly focus on the spectral sensing technique of water constituents, especially for coastal water.

1. G. Dall'Olmo, A. A. Gitelson, Effect of bio-optical parameter variability on the remote estimation of chlorophyll-a concentration in turbid productive waters: experimental results,
Appl. Opt.,
Vol: 44, no. 3, pp. 412-422, 2005.
2. D. Pozdnyakov, A. Lyaskovsky, H. Grassl, L. Pettersson, Numerical modeling of trans spectral processes in natural waters: implications for remote sensing,
Int. J. Rem. Sens.,
Vol: 23, no. 8, pp. 1581-1607, 2002.