Open Access
Add to BibSonomy
Abstract
Since its inception, digital holography has been mainly carried out using visible light but soon various other regions of the electromagnetic spectrum has been fruitfully explored. The Short Wavelength Infrared (SWIR) region, despite its interesting peculiarities and appealing applications in the imaging field, has not yet been fully investigated in Digital Holography (DH). Here we present a compact and robust SWIR DH setup and demonstrate the peculiar and remarkable advantages offered by SWIR radiation for macroscopic object investigation. In particular, we focus our attention to a specific application, previously demonstrated using Long Wavelength IR DH, that is vision through smoke and flames and demonstrate the great potential of SWIR DH in the field of fire rescue.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Digital Holography is an interferometric imaging technique [1] which, thanks to its flexibility, has various and significant applications in many different areas, such as non-destructive testing, 3D imaging, microscopy, etc. [2–5]. Digital holography is usually realized using visible radiation; in this region of the electromagnetic spectrum, however, the stability requirements necessary to preserve the interferometric pattern, containing the information about the investigated scene, are quite stringent. Furthermore, the field of view of digital holography is directly proportional to the wavelength and inversely proportional to the pixel pitch [1] and, considered the typical modern visible detectors (CCD, CMOS), it is rather modest at these wavelengths. Moving toward longer wavelengths the stability requirements become less stringent and a larger field of view is available. Indeed, various interesting applications have been demonstrated in the IR range [6]. The Near IR region ($0.75-1.4 ~\mu m$) has been explored mainly using lasers at $980 ~\mu m$ and $1064 ~\mu m$ [7,8]. The Mid IR region ($3-8 ~\mu m$) has been explored using Quantum Cascade Lasers (QCLs) as coherent sources and thermal cameras [9]. In this case, by exploiting the broad wavelength tunability of QCLs, it is possible to acquire holograms at different wavelengths and extract phase images not subjected to phase wrapping at synthetic wavelengths ranging from hundreds of micrometres to several millimetres. In the Long wavelength IR range ($8-15 ~\mu m$), DH has been demonstrated in 2003 obtained using CO2 laser at $10.6 ~\mu m$ as coherent source and a pyroelectric camera for the hologram recording [10]. At this wavelength various important applications have been demonstrated [11–13], such as vision through smoke and flames [14] and remote monitoring of buildings oscillation modes [15]. Working with CO2 lasers, which are quite powerful and coherent lasers, makes it possible to easily record video holograms of large size dynamic scenes from remote distances. CO2 lasers are however quite bulky and this prevents a desirable efficient miniaturization of the system. The Far IR or THz range ($15-1000 ~\mu m$) has also been fruitfully investigated using THz QCLs and thermal cameras [16]. At these wavelengths various interesting applications exploiting the peculiarities of this radiation has been demonstrated. Indeed, various dielectric material, opaque in the visible range, becomes transparent in the THz range so that it has been demonstrated, for example, the possibility to realize images of scenes hidden behind some plastic materials. In this case, however, important limitations to the applicability of the technique arise from the low power of current sources and from the necessity to cool the sources at very low temperatures. Up to now, no attempt has been dedicated to the realization of DH in SWIR range [6], despite its significant applicative potential. The SWIR region refers to the portion of the electromagnetic spectrum adjacent to the NIR, between $1.4$ and $3 ~\mu m$. As there is no definitive standard of where each waveband separates, these wavelength ranges are described as approximations. Some classifications even think of SWIR as an extension of the NIR band. The SWIR radiation has various important characteristics: it is able to pass undisturbed through smoke, dust and fog. Very sensitive detectors, InGaAs focal plane arrays, with high resolution, small pixel pitch and high frame rate do exist at these wavelengths. Furthermore, very compact fiber sources with relatively high output power and extremely high coherence, such as telecom lasers at $1.55 ~\mu m$, are available. It becomes thus possible to easily shape and direct the beams by means of existing in fiber commercial systems. Finally, SWIR radiation is eye safe. The SWIR region has so far been little explored in DH probably due to the still high costs of sources and detectors but the increasing interest in the SWIR imaging and the peculiarities that SWIR radiation could represent, especially in large size sample investigation, an important incentive to explore this possibility.
2. Theoretical background
2.1 Digital holography
Digital Holography (DH) [1] is the digital version of analog holography [17]. DH provides amplitude and phase information on the wavefront backscattered by an object irradiated with coherent radiation. The hologram $H$ is the intensity interference pattern resulting from the superposition between the radiation coming from the object (Object Beam, OB) and the reference radiation (Reference Beam, RB), the latter directly sent on a recording device.
Explicitly, we can write $H$ as follow
From the product $R(x_H,y_H )\cdot H(x_H,y_H)$, it can be retrieved real and imaginary contributions of the OB wavefront
The OB wavefront is numerically obtained making use of the Rayleigh-Sommerfeld diffraction integral [18]
From the obtained $O(x_R,y_R)$, the intensity $I(x_R,y_R)$ and the phase $\phi (x_R,y_R)$ of the object wavefront are calculated in the usual way
If $x_H$, $y_H$, $x_R$, $y_R$ are small compared with $\rho$, the Rayleigh-Sommerfeld integral can be simplified by adopting the Fresnel approximation (in the applications presented here we will always operate in this approximation)
In digital holography we are dealing with digitized holograms into MxN matrixes, thus Eq. (5) becomes the following discrete function
Amplitude and phase of the discrete object wavefront $O(m, n)$ are computed by means of an FFT algorithm.
2.2 SWIR digital holography and vision through smoke and flames
SWIR DH has various advantages both with respect to LWIR DH and to visible DH. First, according to the expressions for resolution and field of view (FOV) in DH [1], SWIR DH has a higher resolution than LWIR DH and a larger FOV than visible DH, for fixed values of the detector parameters. Considering a wavelength of $1.55~\mu m$, a 640x512 detector resolution and a $15~\mu m$ pixel pitch, the reconstruction resolution and the FOV, at $5~m$ distance, are about $1~mm$ x $0.8~mm$ and $0.5~m$ x $0.5~m$, respectively. These values fit very well with the desirable performances of an imaging system working as a torch within a distance of about 20 m. Secondly, the reduction in stability with respect to LWIR DH, due to the shorter wavelength, is compensated by the shorter exposure times allowed by typical SWIR detectors. Indeed, reducing the exposure time, as we will see in section 4, it becomes possible to freeze the fringe pattern and maintain a good hologram visibility even when the system is handheld. Furthermore, using a fiber coupled laser, it is possible to separate the source from to the optical system thus facilitating the realization of a portable device. Finally, thanks to the high sensitivity of SWIR cameras, it is possible to obtain large size sample images employing relatively low laser power. All these elements, as we demonstrate in what follows, contribute in realising a compact imaging device with unprecedented potential application in the field of vision through smoke and flames.
The capability of imaging a scene through smoke and flames is a highly desirable goal due to its potential application in the field of fire rescue. Vision in such impaired conditions would let rescuers move safely in a fire scenario and would help them in identifying/individuating human being or things hidden behind smoke and flames.
Current visible imaging systems cannot see either through smoke or through flames. Standard thermo-cameras can see through smoke (and are indeed used by firefighters for this purpose) but are blinded by flames. Indeed the image of the flames, focused on the focal plane array by the lens detector, covers and hides the scene of interest. In particular, the radiation emitted by the flames of a real fire scenario, may reach very high power density values which, in principle, can saturate any detector, depending on various parameters: detector sensitivity, detector spectral range, detector-flames distance, detector exposure time, flame temperature, presence of a lens in front of the detector. This last factor, in particular, always impair the vision of any scene behind a flame, even if it does not lead to the detector saturation, since the flame temperature is usually higher than the remaining part of the scene. On the contrary, as demonstrated in [14], DH in the Long Wave Infrared (LWIR) is able to see through both smoke and flames. Vision through smoke is mainly due to the well know capacity of IR radiation to penetrate almost undisturbed through smoke. Vision through flames is instead related to two different peculiarities of the technique. First of all, since the radiation emitted by the flame is not coherent with the laser radiation, it does not contribute to the generation of the interferometric pattern containing the information relative to the scene. Secondly, DH allows to numerically reconstruct a scene without the need to focus its image on the focal plane array of the detector. For this reason, the radiation emitted by the flame does not focus onto the detector but is instead distributed over the whole focal plane array and any saturation effect is reduced. The main limitation of LWIR DH is that CO2 lasers are quite bulky and it is difficult to imagine a miniaturization of the system in order to integrate it into the equipment of rescuers. As shown in what follows, working with fiber coupled laser in the SWIR region makes it possible to miniaturize the system. It is thus possible to hold it in hand like a flashlight, despite it is an interferometric imaging system, and see behind an otherwise impenetrable curtain of smoke and flames. This device would obviously represent a great help for firefighters and first responders to coordinate the rescue operations and work safely in hostile environments.
3. Experimental setup
In order to demonstrate DH in the SWIR range we have realized the holographic system, shown in Fig. 1, composed of few essential elements: a SWIR laser source, a SWIR InGaAs camera and optical elements suitable for the used wavelength. In particular, the laser source is a single frequency mode fiber laser, CNI FL-1550-SF, working at $1.55~\mu m$, with a maximum power of $1~W$ in CW mode operation and a laser linewidth lower than $2~MHz$, which guarantees an imbalance between the two arms of the interferometric system up to more than $50~m$. The camera, NIT Widy SenS 640 V-ST, has 640x512 pixels with pixel pitch of $15~\mu m$ and can work at a maximum frame rate of $230~Hz$ and with a minimum exposure time of $10~\mu s$. The laser radiation is split into two different beams by means of a polarization maintaining fiber coupler with a split ratio of 99:1. The higher power beam, called object beam, is reshaped by means of an appropriate positive lens, of $6~cm$ focal length and $2.54~cm$ diameter (optical element (c) in Fig. 1), in order to have a desired divergence of the beam and is then sent toward the scene to be investigated; the lower power beam, constitutes the reference beam, is sent toward a non-polarizing beam splitter cube ($1~cm$ side, optical element (b) in Fig. 1), with reflection-transmission ratio 10:90, in order to recombine, on the detector surface, with the radiation scattered from the irradiated sample; the beam splitter cube is tilted at the correct angle in order to obtain narrow and correctly sampled interferometric fringes. A positive lens with $7~cm$ of focal length and $1.5~cm$ diameter (optical element (a) in Fig. 1) is inserted in front of the beam splitter cube in order to create an image of the sample at the appropriate position to work in a Fourier configuration [1]; in such configuration the distance between the cube center and the sample image and the distance between the cube center and the reference beam fiber output are the same. The minimalism of such an optical system makes it robust and portable. The optical system (laser source excluded) is indeed mounted on a metallic base and occupies a volume of $30~cm$ x $7~cm$ x $10~cm$ and has a top aluminium surface equipped with a special handle to hold it and direct it towards the investigated scene (Fig. 1, bottom).
Fig. 1. Schematic representation of the SWIR holographic experimental setup (up) and its current implementation (bottom). In the scheme: (a) positive lens, $7~cm$ focal length and $1.5~cm$ diameter; (b) non-polarizing beam splitter cube $1~cm$ side and reflection-transmission ratio 10:90; (c) positive lens, $6~cm$ focal length and $2.54~cm$ diameter.
4. Results
The described system has been used to carry out three different measurements. The numerical reconstruction of the amplitude image of the sample was obtained using standard reconstruction algorithms of DH. For the first test, a small bronzed resin statue ($30~cm$ high) reproducing the famous Perseo statue by Benvenuto Cellini was used as a target. The statue was positioned at a distance of $2.5~m$ from the detector and the hologram was registered using a laser power of $900~mW$ and an exposure time of $100~\mu s$. In Fig. 2 it is shown, from left to right, a visible image of the sample, its SWIR hologram and the amplitude reconstructed image. The hologram amplitude reconstruction, in this case, is obtained applying a third order zero padding to the hologram matrix.
In a second test we irradiated a scene (Fig. 3) composed of various objects of different materials (two bottles of different plastic material, a wood cube, a ball, etc.) and investigated the possibility to do a tracking shot on the scene while holding the system by hand; to this purpose we registered a video hologram with a frame rate of $50~Hz$, an exposure time of $10~\mu s$ and a laser power of $900~mW$ (see Visualization 1); thanks to the very low exposure time it was thus possible to irradiate the various portion of the scene with no particular attention to the stability. The portability of the system is allowed, with some attentions, working with exposure times lower than $100~\mu s$ and becomes easy with the minimum exposure time of $10~\mu s$. This test made it possible also to verify the good visibility of different materials. In this context, another plastic bottle (HDPE, High-density polyethylene), opaque in the visible range, was holographically investigated. This bottle, as expected, was partially transparent to SWIR radiation revealing its content and allowing to see the liquid level inside the bottle. In Fig. 4 it is shown a visible image of the bottle and the holographic reconstruction where the liquid level is clearly visible.
The last test was conducted in outdoor condition in order to demonstrate vision through smoke and flames by means of SWIR DH in a realistic fire scenario. To reach such goal, we mounted a tunnel using $2~m$ long galvanized aluminium corrugated walls; the entrance of the tunnel had approximately an $90~cm$ diameter. The tunnel base was covered with a layer of pine cones, a layer of hay and a layer of small caliber wood up to a height of about $35~cm$. An aluminium plate ($24~cm$ x $29~cm$) was positioned $50~cm$ after the end of the tunnel. This carpet of flammable material was then doused in alcohol and set on fire. At $2.5~m$ from the entrance of the tunnel we had positioned three different imaging systems: a visible camera (Jenoptik), a thermal camera (Miricle Thermoteknix 307k) and the compact SWIR DH system. We acquired a SWIR digital video-hologram after the ignition at a frame rate of $50~Hz$ with an exposure time of $100~\mu s$. As it is possible to see in Fig. 5 and in the attached video comparing visible and holographic images (see Visualization 2), the visible image is covered by the flame image while the thermal image results completely blinded by the flame radiation which saturated the micro-bolometric elements as the flames spread along the entire base of the tunnel; on the contrary, the holographic amplitude image of the plate remained clearly visible throughout the experiment. In order to improve the quality of the reconstructed images, we divided the video into subsets of five frames and each subset was replaced by a single frame obtained using the Z Project function (option Standard Deviation) of the ImageJ software. Using this option, the software takes, for each pixel, the corresponding value in each frame of the subset and calculates the standard deviation of those values. Then the original subset is replaced by an image of the calculated standard deviation values. This procedure allowed to increase the contrast between the target and the surrounding empty space, especially when the target signal was rapidly changing due to the turbulence of the flame.
5. Conclusions
We present a compact portable SWIR DH system which can be used to retrieve real-time coherent images of human size scenes and, in particular, is perfectly suitable for imaging in smoky and flames invaded environment.
Fig. 2. Statue of Perseo by Benvenuto Cellini. From left: visible image, hologram and amplitude reconstruction.
Fig. 3. Scene with objects of different materials; in the upper inset four holographic amplitude reconstructions of selected items.
Fig. 4. Plastic bottle (HDPE, High-density polyethylene) partially filled: visible and holographic amplitude reconstruction. In the amplitude image the liquid level can be observed.
Fig. 5. Images before and during the fire. Top raw, images without fire. From left: visible, thermal and holographic images respectively. Bottom raw, images during the fire. From left: visible, thermal and holographic images respectively.
Indeed, the portability of the SWIR system combined with the ability of SWIR radiation to penetrate smoke, fog and dust, and with the peculiarity of DH to see through flames, make it possible to realize a sort of torch which could be used, by firefighters and rescuers to see in a real fire scenario like in a burning apartment or tunnel.
Advancements in hardware miniaturization and computational algorithms and the integration of machine learning and artificial intelligence could enhance the efficiency and applicability of digital holography in challenging environments, in particular in the reconstruction and recognition of objects obscured by smoke and flames. The applications of digital holography in firefighting, industrial monitoring, and combustion research are just a few examples of the potential impact this technology could have.
Acknowledgments
The authors thank Mr. Antonio Cubattoli and Mr. Stefano Pianigiani for their contribution to the logistic organization of the fire scenario.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. U. Schnars, C. Falldorf, J. Watson, and W. Jüptner, Digital Holography and Wavefront Sensing, (Springer-Verlag, 2015).
2. T. Kreis, “Application of Digital Holography for Nondestructive Testing and Metrology: A Review,” IEEE Trans. Ind. Inf. 12(1), 240–247 (2016). [CrossRef]
3. E. Stoykova, F. YaraŞ, and H. Kang, “Visible reconstruction by a circular holographic display from digital holograms recorded under infrared illumination,” Opt. Lett. 37(15), 3120–3122 (2012). [CrossRef]
4. P. Marquet, B. Rappaz, P. J. Magistretti, E. Cuche, Y. Emery, T. Colomb, and C. Depeursinge, “Digital holographic microscopy: a noninvasive contrast imaging technique allowing quantitative visualization of living cells with subwavelength axial accuracy,” Opt. Lett. 30(5), 468–470 (2005). [CrossRef]
5. M. K. Kim, “Principles and techniques of digital holographic microscopy,” J. Photonics Energy 1(1), 018005 (2010). [CrossRef]
6. M. Georges, Y. Zhao, and J.-F. Vandenrijt, “Holography in the invisible. From the thermal infrared to the terahertz waves: outstanding applications and fundamental limits,” Light: Advanced Manufacturing 2(2), 1 (2022). [CrossRef]
7. D. Grujić, D. Vasiljević, D. Pantelić, L. Tomić, Z. Stamenković, and B. Jelenković, “Infrared camera on a butterfly’s wing,” Opt. Express 26(11), 14143–14158 (2018). [CrossRef]
8. S. A. Owens, M. F. Spencer, D. E. Thornton, and G. P. Perram, “Pulsed laser source digital holography efficiency measurements,” Appl. Opt. 61(16), 4823–4832 (2022). [CrossRef]
9. M. Ravaro, M. Locatelli, E. Pugliese, I. Di Leo, M. Siciliani de Cumis, F. D’Amato, P. Poggi, L. Consolino, R. Meucci, P. Ferraro, and P. De Natale, “Mid-infrared digital holography and holographic interferometry with a tunable quantum cascade laser,” Opt. Lett. 39(16), 4843–4846 (2014). [CrossRef]
10. E. Allaria, S. Brugioni, S. De Nicola, P. Ferraro, S. Grilli, and R. Meucci, “Digital holography at 10.6μm,” Opt. Commun. 215(4-6), 257–262 (2003). [CrossRef]
11. M. Georges, “Long-Wave Infrared Digital Holography,” in New Techniques in Digital Holography (John Wiley & Sons, Inc., USA, 2015), PP. 219–254.
12. J.-F. Vandenrijt, C. Thizy, L. Martin, F. Beaumont, J. Garcia, C. Fabron, É. Prieto, T. Maciaszek, and M. Georges, “Digital holographic interferometry in the long-wave infrared and temporal phase unwrapping for measuring large deformations and rigid body motions of segmented space detector in cryogenic test,” Opt. Eng. 55(12), 121723 (2016). [CrossRef]
13. M. Georges, “Speckle interferometry in the long-wave infrared for combining holography and thermography in a single sensor: applications to nondestructive testing: The FANTOM Project,” Proc. SPIE 9525, 95251L (2015). [CrossRef]
14. M. Locatelli, E. Pugliese, M. Paturzo, V. Bianco, A. Finizio, A. Pelagotti, P. Poggi, L. Miccio, R. Meucci, and P. Ferraro, “Imaging live humans through smoke and flames using far-infrared digital holography,” Opt. Express 21(5), 5379–5390 (2013). [CrossRef]
15. P. Poggi, M. Locatelli, E. Pugliese, D. Delle Donne, G. Lacanna, R. Meucci, and M. Ripepe, “Remote monitoring of building oscillation modes by means of real-time Mid Infrared Digital Holography,” Sci. Rep. 6(1), 23688 (2016). [CrossRef]
16. M. Locatelli, M. Ravaro, S. Bartalini, L. Consolino, M. S. Vitiello, R. Cicchi, F. Pavone, and P. De Natale, “Real-time terahertz digital holography with a quantum cascade laser,” Sci. Rep. 5(1), 13566 (2015). [CrossRef]
17. D. Gabor, “A new microscopic principle,” Nature 161(4098), 777–778 (1948). [CrossRef]
18. M. Born and E. Wolf, Principles of optics, Electromagnetic theory of propagation, interference and diffraction of light, (Cambridge University Press UK, 1999).
