Aquatic environment monitoring using a drone-based fluorosensor
A drone-based system for monitoring of laser-induced fluorescence from the aquatic environment was constructed. Fixed-range remote-sensing demonstration measurements were performed, and field recordings of natural river water fluorescence, oil-slicks as well as dye-marked natural water volumes were taken at drone flying heights of about 10 m. Our fluorosensor, weighing only 1.5 kg, and carried by a commercial drone, illustrates how airborne remote sensing based on fluorescence can be made cost-effective and readily applicable, while presently only in ambient low-light-level conditions.
Oil spills have caused worldwide attention because of major impact on marine ecosystems. Offshore drilling platform and ship accidents constitute major sources of oil spills . Oil is a hydrocarbon-rich polymer mixture including cycloparaffins, alkanes, and aromatic hydrocarbons, and has very complex physical and chemical characteristics. In the first days of an oil spill, oil floats on the surface of the water. Then, under the actions of physical transport, dissolution, emulsification, oxidation, and degradation, its components change. The rapid and reliable detection and discrimination of oils are of crucial importance for marine oil-spill control.
Remote-sensing technologies have contributed much to the development of oil-spill detection [2, 3]. Several available measurement methods, for example IR/UV sensor- and radar systems have been proven capable of detecting oil spills when operated from airplanes. Also, space-born equipment, such as the MODIS and MERIS satellites, were introduced and enabled oil-spill detection, also with thickness measurement capability .
Laser-induced-fluorescence (LIF) provides a further powerful technique for environmental monitoring. Water bodies can be studied using the characteristic fluorescence signatures of, e.g., algae and oil spills. Some early work is presented and reviewed in [5, 6, 7, 8], with subsequent developments exemplified in [9, 10, 11, 12, 13, 14]. Remote-sensing instruments using LIF techniques mostly employ pulsed laser sources with wavelengths ranging between 308 and 355 nm for excitation, and such instruments are also useful for vegetation monitoring (see, e.g., ) and also have applications in the cultural heritage sector (see, e.g., ). The laser-induced fluorescence approach, utilizing specific spectral signatures, provides the possibility to identify different kinds of released oil and also to studying the influence of weathering, etc. [17, 18, 19, 20].
The available detection methods for oils are mainly designed for airborne- or satellite-borne applications. Passive imaging cameras are heavily dependent on the sunlight conditions. Reported laser-based airborne systems for fluorescence monitoring are complex and demanding. Even if an individual airborne lidar system can be applied over considerable ranges , it is bulky and not easy to operate. This is much related to the fact, that pulsed lasers are customarily employed, leading to heavy and costly systems, even if recently lighter-weight solutions adaptable for unmanned aerial vehicle (UAV) operation as analyzed in  now seem feasible.
Remote sensing of oil pollution using fluorescence has almost exclusively been pursued with pulsed laser systems, for which background radiation can conveniently be suppressed using a gated and intensified detector. With the development of CW blue and UV lasers based on semiconductor materials , it became possible to create simplified fluorosensors, first for laboratory studies  and then also for remote-sensing fluorescence applications . Oil pollution discrimination by an inelastic hyperspectral Scheimpflug lidar system based on CW lasers was demonstrated in the laboratory  using a system which was similar to one previously used for range-resolved monitoring of algae and zooplankton in water . We here report on the construction of a fully functional, compact and low-cost laser-induced fluorescence system, which is operated from a readily available commercial drone. We have recently described a drone-based, hyperspectral CW lidar system with range-resolution capability and demonstrated fluorescence height profiles of trees . Drones carrying lidar systems have been used extensively, also as commercial undertakings, as discussed in , but as it seems only for elastic backscattering terrain and city profiling. Since range-resolution is not required in surface monitoring, e.g., of oils on water, the optics can be adjusted for a suitable fixed flying altitude, and a fluorescence monitoring system can be made simpler and of even lighter weight.
Below, initial laboratory test measurements with the new fluorosensor, its deployment on a bridge crossing a river, as well as airborne monitoring of natural waters and contaminants using the drone are described.
2 Instrument description
3.1 Fixed-range test measurements
The process method of the recorded data is illustrated in Fig. 3a. C (F500) is the oil fluorescence at 500 nm. A (F480) is the fluorescence intensity at 480 nm and B is the free-standing water Raman signal. The increasing ratio between F500 and F480 is observed in Fig. 3b, where the 10 ml and 20 ml data points were measured in a separate test. The ratio gets saturated with more oil added. In the evaluation of the oil fluorescence, shown in Fig. 3c, the fluorescence intensity at 500 nm is used directly, since the DOM contribution becomes very small for thicker layers. To get the absolute (free-standing) water Raman peak intensity from our measured results, first, we set the intensity values at the characteristic water Raman band (470–490 nm) to zero. We then used the method of interpolation in the curve fitted to the non-Raman parts of the composed spectrum, also considering the filter fluorescence, for achieving a background spectrum. Then, the resulting background spectra were subtracted from the measured spectra to get the absolute Raman peak. The result is shown in Fig. 3d. To compensate for different exposure time to get everything on the same scale, the spectral intensities are normalized to the different exposure times used as mentioned at the beginning of Sect. 2.1. We note, that even for relatively thick oil films the oil fluorescence still increases and there is a remaining water Raman signal, reflecting the fact that oil has a comparatively low absorption at the 412 nm wavelength. The curves have some irregularities, most likely due to possible instability in the experimental set-up reaching over three building stories.
3.2 Bridge-based measurements on river water
Figure 4a is the schematic graph of the bridge-based arrangements, and Fig. 4b is a photograph of the laser beam striking the river surface. Figure 4c shows the spectra recorded by the fluorosensor. The natural river deep water fluorescence shows a strong DOM signal from the polluted water, as well as a prominent chlorophyll peak at around 680 nm due to the abundant microscopic algae in the eutrophic water. The water Raman peak is observed at 480 nm, as induced by the 412 nm laser, and provides a convenient calibration, since the probed water volume is the same. This allows the observation, that the river water has about six times stronger DOM signal than Guangzhou tap water, the spectrum of which is included in Fig. 2b. When the floating container with oil is brought under the fluorosensor laser beam, the broad-band fluorescence increases about five times and Raman and algal signals disappear as expected. It should be noted, that in the open maritime environment, with a much lower DOM signal, the contrast to an oil spill is clearly much enhanced, and can be further increased by selecting an optimized, considerably lower, excitation wavelength, where the oil has a much higher absorption.
3.3 Drone-based measurements of river water
We have described a compact and light-weight fluorosensor, which provides high-quality fluorescence spectra of the superficial aquatic layer featuring spectral signatures from oils, dissolved organic matter (DOM), algae, etc., and allowing internal calibration using the water Raman signal [27, 28]. The full system weight is only 1.5 kg, achievable using a high-power CW semiconductor laser in conjunction with a compact digital spectrometer. The system was integrated with a compact drone carrier. To obtain range-resolved spectra of the upper few meters of the water overflown by the drone, a CW lidar system based on the Scheimpflug principle, as described for vegetation monitoring in our recent paper , could be adapted to aquatic applications, where many challenges regarding, e.g., the aquatic fauna are present . However, for studies of oil on water, and for hydrological studies of dye dispersion , the compact system described in the present paper provides a good alternative to conventional pulsed-laser fluorosensors and CW-laser Scheimpflug systems, and exhibits certain advantages in terms of performance and simplicity. With a CW semiconductor laser operating just below 400 nm, an eye-safe system would be achievable. The present version of the fluorosensor system could only be used at night time, which is a clear drawback compared to pulsed lidar systems with range gating, allowing daytime use, while again still functioning better in low ambient light level conditions. Using sufficiently high CW powers and combining with a suitable modulation scheme for background subtraction, daylight operation should be feasible, considerably extending the present capability. Clearly, limited range and flying duration constitute limitations for large-area surveillance. However, our study shows that the cost of airborne fluorescence mapping can be dramatically reduced, making the drone-based remote LIF technique a viable diagnostic alternative, especially in studies of limited areas.
The authors gratefully acknowledge the continuing support from Professors Sailing He and Guofu Zhou and the assistance by Xun Wang. This work was supported by the National Science Foundation of China (61705069) and the Chinese Ministry of Science and Technology through the National Key Research and Development Program of China (2018YFC1407503) and the Special Funds Program for the Cultivation of Guangdong College Students’ Scientific and Technological Innovation Climbing Project (pdjha0127).
- 3.M. Fingas, C.E. Brown, Chap. 6. Oil spill remote sensing: a review, in Spill science and technology, ed. by M. Fingas (Gulf Professional Publishing, Boston, 1990)Google Scholar
- 5.J.F. Fantasia, H.C. Ingrao, The development of an experimental airborne laser remote sensing system for the detection and classification of oil spill. In: Ann Arbor, Michigan. In: Proc. 9th International Conference on Remote Sensing of Environment, p. 1711 (1974)Google Scholar
- 10.M. Sasano, K. Hitomi, H. Yamanouchi, S. Yamagishi, Remote sensing of oil spills with a helicopter-based fluorescence imaging lidar. Trans. Vis. Soc. Jpn. 25, 421 (2005)Google Scholar
- 11.M. Lennon, N. Thomas, V. Mariette, S. Babichenko, Operational quantitative mapping of oil pollutions at sea by joint use of a hyperspectral imager and a fluorescence lidar system on-board a fixed-wing aircraft. Oceans 842, 848 (2005)Google Scholar
- 12.V. Drozdowska, Seasonal and spatial variability of surface seawater fluorescence properties in the Baltic and Nordic Seas: results of lidar experiments. Oceanologia 49, 59 (2007)Google Scholar
- 14.C.E. Brown, Chap. 7 Laser fluorosensors, in Oil spill science and technology, ed. by M. Fingas (Gulf Professional Publishing, Boston, 2011)Google Scholar
- 16.V. Raimondi, G. Cecchi, D. Lognoli, L. Palombi, R. Grönlund, A. Johansson, S. Svanberg, K. Barup, J. Hällström, The fluorescence LIDAR technique for the remote sensing of photoautotrophic biodeteriogens on outdoor cultural heritage: a decade of in situ experiments. Int. Biodeterior. Biodegrad. 63, 823 (2010)CrossRefGoogle Scholar
- 17.A.B. Utkin, A. Lavrov, R. Vilar, Evaluation of oil spills by laser induced fluorescence spectra. Proc. SPIE 7994, 48 (2010)Google Scholar
- 18.W. Feng, R. Wang, P. Sun, Z. Gao, L. Chen, Study on ultraviolet laser induced fluorescence spectra of several typical petroleum contaminants. Spectrosc. Spectr. Anal. 31, 1168 (2011)Google Scholar
- 27.H. Edner, J. Johansson, S. Svanberg, E. Wallinder, G. Cecchi, L. Pantani, Fluorescence lidar monitoring of the Arno river. EARSEL Adv. Rem. Sens. 1, 42 (1992)Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.