Open Access
Add to BibSonomy
Abstract
An optical vortex is used to enhance the ranging accuracy of an underwater pulsed laser ranging system. An experiment is conducted whereby an underwater object is illuminated by a pulsed Gaussian beam, and both the object-reflected and scattered light are passed through a diffractive spiral phase plate prior to being imaged at the receiver. An optical vortex is formed from the spatially coherent non-scattered component of the return, providing an effective way to discriminate the desired objected reflected light from the spatially incoherent scatter. Experimental results show that the optical vortex permits a spatially coherent ballistic target return to be more easily discriminated from spatially incoherent forward scattered light up to eight attenuation lengths. The results suggest new optical sensing techniques for underwater imaging or lidar.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Light detection and ranging (lidar) is a method for obtaining high resolution 3-dimensional images in maritime environments [1–4]. In turbid media, or over long distances, backscattering from the interrogating beam can reduce image contrast. Receiver gating can be used to effectively discriminate against the this source of clutter. However, receiver gating is less effective against forward scattered light, which can reduce both spatial accuracy (due to beam spread) and range accuracy (caused by the temporal spread of photon arrival times among scattered and non-scattered light) [5]. As an alternative to short-pulse/range-gated measurement systems, the hybrid lidar-radar technique in which the lidar carrier is intensity modulated by a radar sub-carrier [6], has demonstrated the ability to improve both spatial resolution [7] and range resolution [8] in turbid media. Unfortunately, this technique requires custom lasers which are large and expensive.
A new method was recently proposed which uses an optical vortex to discriminate against scattered light in a turbid underwater environment [9]. For an underwater object illuminated with a Gaussian beam, the light collected at a receiver is composed of both non-scattered object-reflected light as well as backscattered and forward scattered clutter. By passing the return through a spiral phase plate the spatially coherent, non-scattered, object-reflected light forms a vortex. The spatially incoherent scatter does not. In this way, the spiral phase plate and optical vortex function as an analyzer to spatially separate and discriminate the target signature from undesired clutter. The method is conceptually similar to that of the optical vortex coronagraph, which exploits differences in spatial coherence between celestial bodies in order to detect faint objects obscured by the presence of brighter ones [10–12].
In the previous underwater study [9] no range information could be obtained due to the use of a continuous wave source and time-independent optical receiver. In this study, we combine the ranging capabilities of a pulsed laser source and time sensitive receiver with the spatial discrimination techniques of the optical vortex. It will be shown that the addition of the optical vortex provides effective discrimination against forward scattered light out to eight attenuation lengths, thus improving range accuracy.
2. Experimental set-up
2.1. Set-up
The experiment is shown in Fig. 1. The source is a frequency doubled, diode seeded, and fiber amplified pulsed laser with a wavelength of 532 nm, a pulse width of 1 ns, and a repetition rate of 20 kHz. The laser pulse enters a 1 m × 1 m × 3.56 m test tank, where it illuminates a white painted PVC target placed at a range of z = 2 m. The transmitter/receiver separation distance is s = 6.5 cm.
At the receiver, an optical vortex is formed by placing a diffractive spiral phase plate (order m = 16) 12 cm in front of a 150 mm focusing lens, which was used to fill the slit aperture of a streak tube. Upon passing through the slit aperture of the streak tube, a cross section of the optical vortex is imaged onto the streak tube and recorded by the CCD camera. The spatial information is retained in the horizontal axis of the CCD image, while the temporal evolution of the pulse as it travels through the medium is retained along the vertical axis of the CCD image, via the streak tube’s sweep. Experiments were repeated without the diffractive spiral phase plate (i.e., using order m = 0) in order to provide a baseline measurement comparable to standard detection methods.
A small portion of the laser pulse is sampled by a photodiode prior to entering the tank and is used to trigger a pulse generator which resets the streak tube. The streak tube can be reset at a maximum rate of 2 kHz, meaning 1-in-10 laser pulses are viewed. The laser pulse travels through an optical delay tube, accumulating 67 ns of delay, before entering the tank. This optical delay matches the electrical delay of the system to assure that the same pulse that triggers the streak unit is also imaged by the camera. The streak time was set to 20 ns in order to sufficiently capture the transmit pulse’s round trip path to and from the object, and the trigger delay was adjusted so the target return falls in the middle of the streak sweep. The CCD camera integration time is set at 10 seconds, over which the camera averages 20,000 pulses to form an image. Twenty images are then averaged to further minimize noise.
When turbid water experiments were performed, the tank was initially filled with tap water and filtered. Equate antacid was used to alter the turbidity of the water, as antacid particles closely mimic particle scattering in the ocean [13]. An AC-9 transmissometer (WET Labs, Philomath, OR) was used to measure the total attenuation coefficient, c m−1, in-situ. The distance to the target was quantified in terms of attenuation length, which is the product of the attenuation coefficient, c m−1, and physical range, z [m]. A sump pump was used for the duration of the experiment to keep the antacid in suspension. Since we did not create any significant temperature gradients within the tank, it is assumed that scattering is the dominant process, not turbulence due to water mixing.
2.2. Processing and analysis
Since the streak tube aperture is a narrow slit, only a cross-section of the incident vortex beam is imaged on the CCD. It will be instructive to consider the intensity of the light falling in three distinct regions of this cross-section: the optical vortex (Ivortex), the core inside the vortex (Icore), and the area outside the vortex (Ioutside). Figure 2(a) illustrates the optical return from a mirrored target in clean water. Since this return remains highly spatially coherent, a clear vortex cross-section is observed, as shown in Fig. 2(b). In this case,
where IT,coherent represents the target reflected light which remains spatially coherent. Since the reflection from the mirror does not significantly alter the spatial coherence of the interrogating beam, nearly all of the received light is converted into the optical vortex by the spiral phase plate. Note that even if the target location is unknown (along the vertical time axis), knowledge of the width of each region (i.e. the columns corresponding to each region along the horizontal axis) is assumed to be known a priori either by calibration or through knowledge of the spiral phase plate order which determines the vortex size.
Fig. 2 A conceptual sketch of target returns (right column) with a representative streak camera image (right) for (a–b) a mirror target in clean water, (c–d) a PVC target in clean water, and (e–f) a PVC target in turbid water.
Figure 2(c) illustrates the optical return from a painted PVC target in clean water. The corresponding streak image in Fig. 2(d) shows that a vortex is still formed with the portion of target-reflected rays that are spatially coherent at the input of the phase plate. Target reflected rays arriving at larger angles are spatially incoherent, and fall outside the vortex. Mathematically,
A small amount of spatially incoherent target light may be observed in the core, however in practice it was observed to be nearly an order of magnitude smaller than the coherent light in the vortex, and is considered negligible.As turbidity increases (Fig. 2(e)–2(f)), scattered light is distributed both temporally (vertically) and spatially (horizontally) across the CCD image. Since the streak tube is gated close to the target return, backscattered light is effectively gated out. Light arriving near the expected return time of the target is attributed to forward scatter. Since the forward scattered light is spatially incoherent relative to the ballistic target-reflected light, it is spatially (horizontally) distributed across the vortex. Furthermore, since the forward scatter has traveled a longer distance than the target-reflected photons, it is also temporally (vertically) delayed relative to the ballistic target return. However, since the relative time differences are small, it can be difficult to temporally discriminate the target from the clutter. This results in the pulse broadening which reduces range accuracy. In this scenario, the light in and around the target return is described as,
where Ifwd is the forward scattered component. A key observation (similar to that made in [10]), is that the vortex core acts as a ‘window’ to collect only the forward scattered component. The spatial separation of this clutter term can be exploited to correct the portions of the image which contain target information but are obscured by forward scatter.To perform this correction, Icore = Ifwd is estimated for each time bin (i.e. row) by averaging the columns corresponding to the vortex core, as these columns contain only scattered light. This average is then subtracted from from the remaining pixels in that row (Ivortex − Icore and Ioutside − Icore). The procedure is repeated for each row in the image. A spatial cross-correlation is then performed using the corrected image and the clean water image as a reference. Target range is determined by the location of maximum intensity in the cross-correlation. Range error is determined by comparing the range given by the cross-correlation in turbid water relative to those given by the cross-correlation in clean water.
3. Results
To establish a baseline, we first compared the ranging performance of the receiver with and without the spiral phase plate. Figure 3(a) shows the normalized intensity taken along the time (vertical) axis for increasing turbidities without the spiral phase plate. As turbidity increases, the width of the signature is both delayed and broadened in time. This is a common observation in pulsed ranging systems in turbid media. Figure 3(b) shows range error increases with increasing turbidity, due to the increased collection of temporally delayed forward scattered light.
Fig. 3 (a) The normalized intensity along the time axis of the streak camera, and (b) the relative error of the measured target position versus the actual target position (as measured in clean water), with no optical vortex (i.e., m = 0). The relative error increases with attenuation length due to forward scattering.
Next, the spiral phase plate (order m = 16) is added, and the scattering correction was applied to the resultant images. Figure 4 shows the raw streak camera images with the optical vortex (left column), the images after the scatter correction was applied (center column), and the cross-correlations of the scatter-corrected images (right column) for three increasing turbidities (cz = 0.7, 4 and 8). The ‘tilt’ observed in the images are due to slight misalignments of the streak tube receiver, and a non-uniform intensity distribution across the vortex due to receiver alignment and field of view. However, this will not degrade the accuracy of the system because the clean water target return, which also contains the effects of slight misalignment, is used as the template for cross correlation.
Fig. 4 Post processing of streak camera images in increasingly turbid water. (a,d,g) Raw streak camera images, (b,e,h) scatter corrected images, and (c,f,i) cross-correlation between scatter corrected images and the clean water image.
The range to the target is calculated from the cross-correlation of the scatter-corrected images with the clean water reference image. The normalized intensity taken along the time axis, through the point of maximum intensity of the cross-correlation image, and is shown in Fig. 5(a). The range errors relative to clean water are plotted in Fig. 5(b). Compared to the m = 0 baseline, the range measurements obtained from the m = 16 optical vortex with scatter correction more accurately match the true target position up to 8 attenuation lengths. These results illustrate how the optical vortex can be used to spatially discriminate target light from incoherent forward scatter clutter.
Fig. 5 (a) The normalized intensity taken along the time axis, through the point of maximum intensity of the cross-correlation image, and (b) the relative error of the measured target position versus the actual target position (as measured in clean water), with optical vortex (m = 16) and scattering correction. Using the optical vortex and performing the scattering correction effectively reduces the range error caused by forward scatter. The range error when no optical vortex is used (m = 0) is shown for reference.
4. Conclusion and future work
The use of an optical vortex as a spatial coherence analyzer has been experimentally demonstrated in the optical detection of objects in a turbid underwater medium. This is achieved by exploiting the spatial relationships of coherent and incoherent light returning from the underwater scene after light from the scene is passed through a spiral phase plate. Time-resolved images of the spatial intensity pattern were obtained using a pulsed laser and a temporally sensitive streak camera. Improved range accuracy is achieved through simple image processing by exploiting the spatial dependence of coherent target and incoherent scattered light. Improved range accuracy is observed over eight attenuation lengths relative to a traditional detection scheme that does not leverage the spatial dependence of the optical vortex. Future work will investigate the effect of varying the object range and the order of the phase plate, as well as examine how this method may also be used to help discriminate against backscattered light.
Funding
Office of Naval Research (ONR) Code 32 (N00014-17-WX01846) and Code 31 (N00014-16-1-2779).
References and links
1. G. Kattawar, “Time of flight lidar measurements as an ocean probe,” Appl. Opt. 11(3), 662–666 (1972) [CrossRef] [PubMed]
2. G. C. Guenther, “Airborne Laser Hydrography,” U.S. Department of Commerce, National Oceanic and Atmospheric Administration, National Ocean Service, Charting and Geodetic Services (1985)
3. J. Jaffe, “Underwater Optical Imaging: The Past, the Present, and the Prospects,” IEEE J. Oceanic Eng. 40(3), 683–700, (2015) [CrossRef]
4. F. R. Dalgleish, F. M. Caimi, W. B. Britton, and C. F. Andren, “Improved LLS imaging performance in scattering-dominant waters,” Proc. SPIE 7317, 73170E (2009) [CrossRef]
5. E. A. McLean, H. R. Burris, and M. P. Strand, “Short-pulse range-gated optical imaging in turbid water,” Appl. Opt. 34(21), 4343–4351 (1995) [CrossRef] [PubMed]
6. L. Mullen and V. Contarino, “Hybrid LIDAR-radar: Seeing through the scatter,” IEEE Microwave Magazine 1(3), 42–48 (2000) [CrossRef]
7. B. Cochenour, S. O’Connor, and L. Mullen, “Suppression of forward-scattered light using high-frequency intensity modulation,” Opt. Eng. 53(5), 051406 (2014) [CrossRef]
8. J. K. Nash, R. W. Lee, and L. J. Mullen, “Tailoring of RF coded optical pulses for underwater 3D imaging,” OCEANS 2015 - MTS/IEEE Washington, Washington, DC, 1–8, (2015)
9. B. Cochenour, L. Rodgers, A. Laux, L. Mullen, K. Morgan, K. Miller, and E. Johnson, “The detection of objects in a turbid underwater medium using orbital angular momentum (OAM),” Proc. SPIE 10186, 1018603 (2017) [CrossRef]
10. D. Palacios, D. Rozas, and G. A. Swartzlander, “Observed Scattering into a Dark Optical Vortex Core,” Phys. Rev. Lett. 88(10), 103902 (2002) [CrossRef] [PubMed]
11. G. Foo, D. Palacios, and G. Swartzlander, “Optical vortex coronagraph,” Opt. Lett. 30(24), 3308–3310 (2005) [CrossRef]
12. G. Swartzlander Jr, E. Ford, R. Abdul-Malik, L. Close, M. Peters, D. Palacios, and D. Wilson, “Astronomical demonstration of an optical vortex coronagraph,” Opt. Express 16(14), 10200–10207 (2008) [CrossRef] [PubMed]
13. T. J. Petzold, “Volume Scattering Functions for Selected Ocean Waters,” Scripps Inst. of Oceanography, Visibility Laboratory, San Diego, CA (1972) [CrossRef]
