Open Access
Add to BibSonomy
Abstract
An 808 nm single-band Mie scattering Scheimpflug lidar system is developed in Dalian, Northern China, for real-time, large-area atmospheric aerosol/particle remote sensing. Atmospheric measurement has been performed in urban area during a typical haze weather condition, and time-range distribution of atmospheric backscattering signal is recorded from March 18th to 22nd, 2017, by employing the Scheimpflug lidar system. Atmospheric extinction coefficient is then retrieved according to the Klett-inversion algorithm, while the boundary value is obtained by the slope-method in the far end where the atmosphere is homogeneous in a subinterval region. The correlation between the extinction coefficients retrieved from the Scheimpflug lidar technique and the PM10/PM2.5 concentrations measured by a conventional air pollution monitoring station is also studied. The good agreement between the measurement results, i.e., a correlation coefficient of 0.85, successfully demonstrates the feasibility and great potential of the Scheimpflug lidar technique for atmospheric studies and applications.
© 2017 Optical Society of America
1. Introduction
As a powerful active remote sensing technique, lidar has been employed for atmospheric aerosol/particle monitoring since decades [1]. Nowadays, lidar systems are mainly based on the photon time-of-flight principle [2,3]. By transmitting nanosecond-scale laser pulses into atmosphere, range-resolved backscattering light from aerosols and molecules can be first collected by a large-aperture telescope, and then captured by sensitive detectors such as photomultiplier tubes (PMT), etc [4]. Atmospheric parameters, e.g., backscattering and extinction coefficients, can be retrieved from the backscattering echoes by employing Klett-Fernald inversion algorithm or their variances [5–7]. Despite the rapid development of pulsed laser sources (e.g., diode laser pumped Nd:YAG lasers) and photon-detection electronics, the construction of a pulsed lidar system is still of high-cost and complicated in particular when multi-wavelength and Raman channels are integrated [8,9].
Extensive effort has been devoted to develop robust and relatively low-cost Mie-scattering lidar systems for various application scenarios, e.g., micro-pulse-lidar [10,11], LED-based lidar [12] and diode-laser-based ceilometers [13,14], etc. Recently, the Scheimpflug lidar technique, based on the Scheimpflug principle, has been demonstrated for atmospheric gas and aerosol sensing [15–18]. The Scheimpflug principle reveals that the laser beam transmitted into atmosphere can be in-focused on a tilted image sensor while employing a large-aperture telescope, as long as the image sensor plane, the lens plane of the receiving telescope and the laser beam plane intersect with each other. As shown in Fig. 1, when transmitting a continuous-wave laser beam into atmosphere, the backscattering light originating from different distances is sharply imaged on different pixels of the image sensor. Thus, range-resolved atmospheric backscattering signal can be retrieved by recording the backscattering image and deducing pixel-distance relationship according to geometrical optics. The Scheimpflug lidar technique significantly reduces system complexity and cost by utilizing high-power continuous-wave laser diodes as light sources and highly integrated CMOS/CCD sensors as detectors.
Fig. 1 Principles and simplified system diagrams of the Scheimpflug lidar and the conventional pulsed lidar techniques.
In previous work, the Scheimpflug lidar technique has been successfully applied for qualitative studies of time-range transportation of atmospheric aerosols. In spite of the great potential of utilizing the Scheimpflug lidar technique for atmospheric remote sensing, no quantitative analysis about the backscattering/extinction coefficient, the height of the boundary layer, etc., has yet been performed. In this work, we demonstrate atmospheric extinction coefficient measurement in urban area by employing the Scheimpflug lidar technique developed recently by researchers in Dalian University of Technology (DLUT), China. Klett- inversion algorithm is employed for extinction coefficient retrieval in haze weather conditions. Extinction coefficients retrieved by the Scheimpflug lidar technique are further validated by comparing the experimental results with the PM2.5/PM10 concentrations measured by a conventional air pollution monitoring station.
2. Instrumentation and methods
2.1. Single-band Mie-scattering Scheimpflug lidar system in DLUT
The Scheimpflug lidar principle has been thoroughly introduced in [16], and thus would not be described here. We recently develop a Scheimpflug lidar system in DLUT, Dalian. The schematic of the Scheimpflug lidar system as well as the photograph are shown in Fig. 2. The system configurations are different in some aspects compared to those employed in our previous work. We hereby describe the Scheimpflug lidar system of DLUT as following.
Fig. 2 (a) System schematic of a Scheimpflug lidar system, (b) Photograph of the Scheimpflug lidar system developed in DLUT. The small (alignment) telescope with a white-black camera, which is not illustrated in figure (a), is installed for assisting system alignment and laser beam observation. LD: laser diode, TEC: thermoelectric cooling. The Scheimpflug lidar system is located on the second floor of a university building.
An 808-nm laser diode with approximate 4-W output power and 3-nm full width at half maximum (FWHM) is employed as the light source. The laser diode chip is a wide-stripe with an emission area of approximately 230 μm (slow axis) × 1 μm (fast axis). The laser beam FWHM divergence is about 6°║ (slow axis) × 8°┴ (fast axis). Such a small divergence is rather uncommon for commercial multimode laser diodes operating at other wavelengths. The laser diode is housed by a homemade aluminum mount with large-capacity thermoelectric cooling to stabilize the case temperature, thus ensuring an emitting wavelength centered at 808 nm. The accuracy of the temperature stabilization is better than ± 0.1°, which is sufficient for relatively broadband (3 nm) lidar applications in the present work as the temperature coefficient of wavelength for semiconductor laser diodes is small, e.g., 0.06 nm/K. The infrared laser beam is collimated by a refractor telescope (F6) with the eyepiece removed. Due to the small divergence of the laser beam, the geometrical transmission of the laser beam when passing through the F6 refractor is above 90%.
An 808-nm interference filter (3 nm FWHM) and a high-pass color filter (RG780) are employed to suppress the background noise. The backscattering light of the laser beam is collected by a Newtonian telescope with approximately 806 mm separation to the refractor telescope. Both the refractor telescope and the Newtonian telescope are mounted on an aluminum bar, which is carried by an equatorial mount (Skywatcher, Eq. (8). A 2D CMOS camera is mounted with 45° titled to the optical axis of the receiving telescope. A third telescope with a white-black camera, which is not illustrated in Fig. 2(a), is also mounted on the aluminum bar for assisting the alignment of the Scheimpflug lidar system. The laser beam reflected from a hard-target object/building can be clearly observed even during daytime by employing the white-black camera and a long pass filter (RG780).
A trigger signal generated by the camera is fed to a 2-bit Johnson ring counter, and the output is utilized to switch the laser diode on and off alternatively through the driving current. Accordingly, the CMOS camera records the images of the laser beam and the background light. The background signal can then be subtracted from the recorded raw images dynamically. The preprocessed laser beam image is vertically (perpendicular to the propagation direction of the laser beam) binned in a LabVIEW-based program developed by our research group. The binning process leads to a single pixel-intensity curve. After system calibration with a hard target at a known distance, the pixel-distance relationship can be calculated and the distance-intensity curve, i.e., the range-resolved lidar signal is retrieved.
Detailed specifications of the instruments as well as their approximate cost are presented in Table 1. The total hardware cost for the 808-nm single-band Scheimpflug lidar system is about $4800 (excluding the equatorial mount), which is less than tenth of the hardware cost of conventional pulsed lidar systems.
Table 1. System configurations of the Scheimpflug lidar system developed in DLUT, as well as approximate cost of each instrument.
2.2. Solving the Scheimpflug lidar equation
The atmospheric backscattering echo measured by the Scheimpflug lidar is given by [16]:
Here is the distance to the lidar system, K is the system constant, and are the backscattering and extinction coefficients at the wavelength, respectively. In this work, experimental measurements are performed for near ground atmosphere in urban area. Besides, light scattering by aerosols dominates over the molecular scattering in a typical haze weather condition, thus the single-component Klett inversion algorithm can be utilized [4]. As the Scheimpflug lidar equation is analogous to the pulsed lidar equation except the term [16], the Klett solution for the Scheimpflug lidar equation can be deducedHere is the calibration (boundary) distance while is the corresponding extinction coefficient. In order to retrieve, the boundary value has to be determined first. In conventional pulsed lidar technique, a common solution is to find a region in which the assumption of homogeneous atmosphere is reasonable, i.e., the slope-method [19]. In this work, the slope-method is employed for boundary value determination in a subinterval range where the atmosphere is homogeneous [20].3. Atmospheric measurements
Atmospheric measurements are performed in urban area of Dalian from March 18th 18:53 to March 22nd 18:23, 2017 (in total four days). The Scheimpflug lidar system is located on the second floor of a university building and the laser beam is about 6-m height above ground at the operation site. Thus, the risk of eye exposure of pedestrians/drivers is very low during the measurement. Besides, the elevation angle of the lidar system is set to approximately 5° to avoid tall buildings within the field of view. During this period, the concentrations of atmospheric pollutants vary significantly, e.g., PM 2.5 concentration varies between 10 µg/m3 to 150 µg/m3. Thus, the measurement campaign covers a wide range of atmospheric conditions in terms of optical extinction. The pixel-distance relationship is calibrated by measuring the backscattering echo from a hard target with known distance, specifically a tall building located at 961.5 m away from the lidar system in this work. Atmospheric backscattering signals are recorded with an exposure time of 20 milliseconds. The electrical gain of the CMOS camera is adjusted manually to avoid sensor saturation in a sunny weather condition when the lidar system nearly straight looks into the sun (e.g., at around 15:00-16:00 o’clock).
In atmospheric lidar techniques, backscattering signals are frequently averaged for thousands of times and then smoothed/filtered by moving average or low-pass filter such as Savitzky-Golay algorithm, etc [21,22]. In this work, to improve the signal-to-noise ratio (SNR), a single lidar curve is retrieved from the statistic median value of 1000 acquisitions, with a total measurement time of approximate 45 seconds. Apart from atmospheric background noise, the Scheimpflug lidar signals also suffer from readout noise, dark current noise and fixed pattern noise (FPN) of the CMOS camera. The FPN describes the dark current non-uniformity and photo response non-uniformity of the image sensor, and is highly relevant to the exposure time, electrical gain and incident illumination. During daytime, sunlight background noise dominates while the camera noise dominates in nighttime. Due to the readout noise and FPN, the SNR of the lidar signals during nighttime is not improved compared to that in daytime, although the sunlight background is significantly lower. In order to improve the SNR, a low-pass finite impulse response (FIR) filter is employed to process the lidar signals with a bandwidth of five pixels. The SNR of the lidar signals can be improved by a factor of 2-3. The noise level of the lidar signal is about 2 counts after applying the FIR filter. The SNR at the maximum intensities of recorded lidar curves generally varies from 100 to 700 when the atmosphere changes from excellent to moderately polluted condition, with a measurement time less than 1 minute. It should be emphasized here that the SNR of the Scheimpflug lidar technique does not decrease with the square of the distance, which is the case for the conventional pulsed lidar technique.
The lidar system is located in DLUT campus, and the laser beam is transmitted into the west-southwest (WSW) direction. There are several national monitoring stations in Dalian that can report the concentrations of atmospheric pollutants (e.g., PM10/PM2.5, O3, NO2, SO2) as well as other atmospheric parameters (e.g., humidity, wind speed, pressure) every hour. The PM10/PM2.5 concentration is measured by instruments based on β-ray method with a detection sensitivity of 2-5 µg/m3. As shown in Fig. 3, the closest national monitoring station is approximately 2.5 km away from the lidar system, i.e., Dalian Qixianling station. Weather parameters as well as the PM10/PM2.5 concentrations reported by this station are recorded as references for the lidar measurement analysis and comparison studies between the experimental results measured by the Scheimpflug lidar system and the national monitoring station.
Fig. 3 Experimental site map, the Qixianling national monitoring station is about 2.5 km away from the lidar system.
4. Results and discussion
The time-range map of the pre-processed atmospheric backscattering signals is shown in Fig. 4. Besides, atmospheric backscattering signals with significantly higher intensities can be clearly observed during March 19th 09:00 to March 20th 15:00 o’clock, and March 22nd 07:00 to March 22nd 15:00 o’clock. In fact, during the period of the measurement campaign, the whole northern China suffers from severe haze. According to the PM2.5/PM10 AQI (air quality index) values reported by the Qixianling national monitoring station, atmosphere in Dalian city is moderately polluted and light polluted on March 20th and 22nd, respectively. In the afternoon on March 20th, atmospheric pollutants are blew away by strong wind (5-7 m/s) from north. As can be seen from Fig. 4, the backscattering intensity gradually decreases during this period. On March 21st, the northwest wind (5-9 m/s) further cleans the atmosphere. However, local emissions such as automobile exhaust and construction dust can still be observed occasionally from the backscattering time-range map. Lidar signals measured in several typical hours are presented in Fig. 5. In this work, the maximum measurement distance of the Scheimpflug lidar system is defined by the distance where the SNR is about 10, corresponding to a backscattering intensity of around 20 counts. In a moderately polluted weather condition, e.g., PM2.5 concentration of 70 µg/m3 (PM2.5 AQI≈160), the maximum measurement distance is around 4 km. As can be seen from Fig. 5, in good/excellent weather conditions (PM2.5/PM10 AQI<100), the SNRs of the lidar signals are beyond 10 even at around 10 km. However, the range-resolution for the far range decreases significantly attributed to the Scheimpflug principle. Under this circumstance, range-resolution is the main consideration for further data analysis, instead of the SNR.
Fig. 4 Time-range map of atmospheric backscattering intensity. The “dark lines” observed at around 15:00 o’clock on March 21st are saturated lidar signals because of not in time adjustment on the electrical gain of the CMOS camera. The total measurement time of a single lidar curve is about 45 seconds.
Fig. 5 Atmospheric lidar signals recorded at several typical hours, and the solid-black curves are the corresponding linear fittings in a subinterval region, according to the slope-method. The corresponding boundary values at 95% confidence level for each lidar curve are 0.42 ± 0.004 km−1, 0.08 ± 0.01 km−1, 0.42 ± 0.02 km−1, 0.27 ± 0.02 km−1, 0.07 ± 0.01 km−1, 0.17 ± 0.01 km−1, respectively.
In order to retrieve the extinction coefficient from backscattering signals, firstly, the boundary value for each individual lidar signal has to be determined. As can be seen from Fig. 4 and Fig. 5, the measurement distance of the Scheimpflug lidar system varies significantly during the measurement campaign. Thus, the subinterval range for solving the boundary values has also to be selected according to the SNR of individual lidar signal. To precisely retrieve the boundary extinction coefficient and optimize the maximum measurement distance, we set two criterions for boundary region selection. First, a threshold intensity (20 counts), corresponding to a SNR of 10, is set to define the minimum signal intensity for analysis. Second, as the range resolution decreases with increasing measurement distance, which could deteriorate the measurement accuracy, we set the maximum distance for the extinction coefficient retrieval to be 7 km.
Based on the above criterions, the subinterval homogenous region for each lidar curve can be found out by searching a region with minimized linear fitting residual in the far end of the log-scale lidar curve. The linear fittings for typical lidar curves as well as the corresponding boundary values are presented in Fig. 5. The ranges of the subinterval region, containing 4-pixel data points, vary from 200 m to 700 m. The uncertainties of the boundary values are about 1%-14%, implying that the assumption of homogeneous atmosphere is reasonable in the corresponding subinterval region. With the boundary value of the extinction coefficient as input, the extinction coefficient distribution for each lidar curve can be retrieved according to the Klett-method, i.e., Eq. (2). Figure 6 shows the retrieved profiles of the extinction coefficients in situations with high and low atmospheric extinction. As has been known in conventional pulsed lidar technique, the uncertainty of the extinction coefficient is highly relevant to the error of the boundary value, and it decreases as the decreasing of the measurement distance for the backward-integration Klett-method. The time-range map of the extinction coefficient is shown in Fig. 7. As can be seen, moderately polluted atmosphere can be readily identified in the period mentioned above.
Fig. 6 Retrieved extinction coefficients and the corresponding upper- and lower-limits calculated according to the uncertainties of the boundary values shown in Fig. 5.
To validate extinction coefficients retrieved from the lidar measurements, the correlation between the measured extinction coefficients and the PM10/PM2.5 concentrations recorded by the monitoring station has been performed in this work. As the lidar system measured the extinction coefficient distribution in a very large region, we thus calculate the statistic median value of the extinction coefficients from the closest measurement distance to the maximum measurement distance for each lidar measurement. The median value is a good representation of atmospheric extinction coefficient as large extinction coefficients due to suddenly appeared local emissions can be eliminated. As shown in Fig. 8, the trends of the one-hour averaged extinction coefficients obtained from the lidar system and the PM10/PM2.5 concentrations recorded by the monitoring station generally agree well with each other.
Fig. 8 Trends of atmospheric extinction coefficient and PM10/PM2.5 concentrations from March 18th to 22nd, 2017. The extinction coefficient is 10-time averaged.
The scatter plots of PM10/PM2.5 concentrations and extinction coefficients are shown in Fig. 9. Extinction coefficients are calculated from the one-hour mathematic average of extinction coefficients presented in Fig. 8. In spite that the monitoring station is a bit away from the lidar system, a correlation coefficient of 0.85 for both PM10 and PM2.5 is still achieved, implying a good agreement between the particle concentrations and extinction coefficients. By employing a PM10/PM2.5 monitoring station or instrument that is closer to the lidar system as a reference, an even higher correlation coefficient can be expected [23,24].
Fig. 9 Scatter plot of the PM2.5 and PM10 concentrations and extinction coefficients (), and the solid curves are the corresponding linear fittings. The correlation coefficient between the PM10/PM2.5 concentrations and extinction coefficients is about 0.85.
5. Conclusion
In this work, it has been illustrated an 808 nm single-band Mie scattering Scheimpflug lidar system in DLUT developed for real-time, large area atmospheric aerosol/particle remote sensing. Our measurements are performed in urban area during typical haze weather conditions with variations in the atmospheric pollution conditions from March 18th to 22nd, 2017. The maximum measurement distance of the present Scheimpflug lidar system is about 4 km (SNR≈10) in moderately polluted weather condition, while it is limited mainly by the range-resolution in good/excellent weather condition. The Klett-inversion algorithm is employed for atmospheric extinction coefficient retrieval. The boundary values of the extinction coefficient is calculated by the slope-method in the far end. The subinterval region for boundary value calculation of each lidar profile is determined taking into account the constraints of both the SNR and the range-resolution. The trends of the extinction coefficients retrieved from the Scheimpflug lidar system are in good agreement with the PM10/PM2.5 concentrations measured by a conventional air pollution monitoring station, and a correlation coefficient of 0.85 is achieved. The good agreement successfully demonstrates the feasibility of the Scheimpflug lidar technique for atmospheric studies. This work paves the way of the Scheimpflug lidar technique for atmospheric remote sensing applications in the near future.
Funding
National Key Research and Development Program (2016YFC0200600); Fundamental Research Funds for the Central Universities (DUT15RC(3)107); Natural Science Foundation of Liaoning Province, China (201602163).
Acknowledgments
The authors greatly appreciate the strong support of Prof. Qingxu Yu and Dr. Xiaona Wang in the development of the Scheimpflug lidar system. The authors thank the reviewers for their detailed reviews that improved the manuscript.
References and links
1. J. Lenoble, L. Remer, and D. Tanre, Aerosol Remote Sensing (Springer, 2013).
2. E. J. Welton, J. R. Campbell, J. D. Spinhirne, and V. S. Scott, “Global monitoring of clouds and aerosols using a network of micro-pulse lidar systems,” in Proc. SPIE 4153 Lidar Remote Sensing for Industry and Environment Monitoring (2001). [CrossRef]
3. G. Pappalardo, A. Amodeo, A. Apituley, A. Comeron, V. Freudenthaler, H. Linne, A. Ansmann, J. Bosenberg, G. D’Amico, I. Mattis, L. Mona, U. Wandinger, V. Amiridis, L. Alados-Arboledas, D. Nicolae, and M. Wiegner, “EARLINET: towards an advanced sustainable European aerosol lidar network,” Atmos. Meas. Tech. 7(8), 2389–2409 (2014). [CrossRef]
4. V. A. Kovalev and W. E. Eichinger, Elastic Lidar: Theory, Practice, and Analysis Methods (John Wiley & Sons, 2004).
5. F. G. Fernald, “Analysis of atmospheric lidar observations: Some Comments,” Appl. Opt. 23(5), 652–653 (1984). [CrossRef] [PubMed]
6. J. D. Klett, “Lidar inversion with variable backscatter/extinction ratios,” Appl. Opt. 24(11), 1638–1643 (1985). [CrossRef] [PubMed]
7. V. A. Kovalev, “Lidar measurement of the vertical aerosol extinction profiles with range-dependent backscatter-to-extinction ratios,” Appl. Opt. 32(30), 6053–6065 (1993). [CrossRef] [PubMed]
8. R. Engelmann, T. Kanitz, H. Baars, B. Heese, D. Althausen, A. Skupin, U. Wandinger, M. Komppula, I. S. Stachlewska, V. Amiridis, E. Marinou, and I. Mattis, H. L. andA. Ansmann, “EARLINET Raman lidar PollyXT: the neXT generation,” Atmos. Meas. Tech. Discuss. 8, 7737–7780 (2015). [CrossRef]
9. S. Wu, X. Song, B. Liu, G. Dai, J. Liu, K. Zhang, S. Qin, D. Hua, F. Gao, and L. Liu, “Mobile multi-wavelength polarization Raman lidar for water vapor, cloud and aerosol measurement,” Opt. Express 23(26), 33870–33892 (2015). [CrossRef] [PubMed]
10. J. D. Spinhirne, “Micro pulse lidar,” IEEE Trans. Geosci. Remote Sens. 31(1), 48–55 (1993). [CrossRef]
11. C. J. Flynn, A. Mendoza, Y. Zheng, and S. Mathur, “Novel polarization-sensitive micropulse lidar measurement technique,” Opt. Express 15(6), 2785–2790 (2007). [CrossRef] [PubMed]
12. M. Koyama and T. Shiina, “Development of LED mini-lidar,” in Quantum Electronics Conference & Lasers and Electro-Optics, (2011).
13. E. Gregorio, F. Rocadenbosch, J. Tiana-Alsina, A. Comeron, R. Sanz, and J. R. Rosell-Polo, “Parameter design of a biaxial lidar ceilometer,” J. Appl. Remote Sens. 6(1), 063546 (2012). [CrossRef]
14. J. D. V. Hey, A Novel Lidar Ceilometer (Springer, 2015).
15. L. Mei and M. Brydegaard, “Continuous-wave differential absorption lidar,” Laser Photonics Rev. 9(6), 629–636 (2015). [CrossRef]
16. L. Mei and M. Brydegaard, “Atmospheric aerosol monitoring by an elastic Scheimpflug lidar system,” Opt. Express 23(24), A1613–A1628 (2015). [CrossRef] [PubMed]
17. P. Guan and L. Mei, “Range-resolved atmospheric aerosol monitoring by a Scheimpflug lidar system based on a continuous-wave laser diode,” in Asia Communications and Photonics Conferene (ACP, 2016), paper AS1F.5.
18. L. Mei and M. Brydegaard, “Development of a Scheimpflug lidar system for atmospheric aerosol monitoring,” EPJ Web of Conferences (ILRC 27), 119, 27005, (2016). [CrossRef]
19. C. Böckmann, U. Wandinger, A. Ansmann, J. Bösenberg, V. Amiridis, A. Boselli, A. Delaval, F. De Tomasi, M. Frioud, I. V. Grigorov, A. Hågård, M. Horvat, M. Iarlori, L. Komguem, S. Kreipl, G. Larchevêque, V. Matthias, A. Papayannis, G. Pappalardo, F. Rocadenbosch, J. A. Rodrigues, J. Schneider, V. Shcherbakov, and M. Wiegner, “Aerosol lidar intercomparison in the framework of the EARLINET project. 2. Aerosol backscatter algorithms,” Appl. Opt. 43(4), 977–989 (2004). [CrossRef] [PubMed]
20. J. D. Klett, “Stable analytical inversion solution for processing lidar returns,” Appl. Opt. 20(2), 211–220 (1981). [CrossRef] [PubMed]
21. E. Eloranta, “High spectral resolution lidar measurements of atmospheric extinction: progress and challenges,” in IEEE Aerospace Conference (2014), pp. 1-6.
22. W. J. Marais, R. E. Holz, Y. H. Hu, R. E. Kuehn, E. E. Eloranta, and R. M. Willett, “Approach to simultaneously denoise and invert backscatter and extinction from photon-limited atmospheric lidar observations,” Appl. Opt. 55(29), 8316–8334 (2016). [CrossRef] [PubMed]
23. T. Y. He, S. Stanic, F. Gao, K. Bergant, D. Veberic, X. Q. Song, and A. Dolzan, “Tracking of urban aerosols using combined LIDAR-based remote sensing and ground-based measurements,” Atmos. Meas. Tech. 5(5), 891–900 (2012). [CrossRef]
24. S. Parracino, M. Richetta, M. Gelfusa, A. Malizia, C. Bellecci, L. De Leo, C. Perrimezzi, A. Fin, M. Forin, F. Giappicucci, M. Grion, G. Marchese, and P. Gaudio, “Real-time vehicle emissions monitoring using a compact LiDAR system and conventional instruments: first results of an experimental campaign in a suburban area in southern Italy,” Opt. Eng. 55(10), 103107 (2016). [CrossRef]
