Open Access
Add to BibSonomy
Abstract
Imaging and computational processing fusion technologies have expanded the wavelength range that can be visualized. However, it is still challenging to realize a system that can image a wide range of wavelengths, including non-visible regions, in a single system. Here, we propose a broadband imaging system based on femtosecond-laser-driven sequential light source arrays. The light source arrays allow us to form ultra-broadband illumination light depending on the excitation target and irradiated pulse energy. We demonstrated X-ray and visible imaging under atmospheric pressure by using a water film as an excitation target. Furthermore, by applying a compressive sensing algorithm, the imaging time was reduced while maintaining the number of pixels in the reconstructed image.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Imaging technologies provide a variety of visual information corresponding to the wavelength of light to be acquired. Generally, the wavelengths that an imaging system can visualize are limited by the intrinsic operating wavelength of devices employed in the system, such as an image sensor and an illumination light source. By contrast, the fusion of imaging and computational processing has enabled systems that can visualize multiple wavelength regions. Multispectral imaging techniques, which allow simultaneous acquisition of information in the spectral domain and the spatial domain, have been proposed using a scattering medium [1] and a method combining a color-coded aperture and diffractive optical elements [2] to acquire object light and reconstruct an image through computational processing. Single-pixel imaging techniques [3] which reconstruct an image using object light encoded with spatial light modulators (SLMs) and acquired by a single-pixel detector have extended the detectable range to non-visible wavebands. Rotating or translating disks having holes serving as intensity masking patterns have been used to capturing the visible region [4] and the X-ray region [5]. Simultaneous imaging of visible and shortwave infrared wavebands was achieved by using a digital mirror device and detectors corresponding to the respective wavelengths [6]. Interestingly, the simultaneous detection of these different wavelength regions realized imaging that combined the advantages of each wavelength region, such as using infrared illumination to produce an image in the visible region [7]. Furthermore, SLMs using a metamaterial have been used to demonstrate imaging in the terahertz region [8]. The main reason why these techniques can visualize these wavebands is that single-pixel detectors have high sensitivity in a wide range of wavelengths. However, the objects that can be imaged are still limited by the wavelength range supported by the light source and the SLMs.
In this paper, we propose an imaging system using femtosecond-laser-driven sequential light source arrays for visualizing broadband light information. Light source arrays based on spatial control of the focus position of a femtosecond laser enables broadband light sources including the non-visible range and time sequential formation of illumination patterns. Therefore, the system is free from the limitation of the the operating wavelengths of the light source and SLM and can visualize objects with broadband light information. In this research, we selected X-rays, which have a wide range of applications due to their high transmission, and the visible region from among the many wavelengths that femtosecond laser-driven light sources can produce. We constructed an imaging system and evaluated the femtosecond-laser-driven light source in the X-ray and visible region. Finally, some examples of X-ray and visible imaging using our system were demonstrated.
2. Principles
2.1 Imaging based on laser-driven sequential light source arrays
The proposed system acquires the object light information encoded by an illumination pattern formed by 2D scanned femtosecond-laser-driven light sources with a single pixel detector, and reconstructs an image based on computer processing using the detected values. The imaging technique, generally called single-pixel imaging, uses a light source, an SLM, and a single-pixel detector, as shown in Fig. 1(a). This technique captures images by using the detector to acquire the object light encoded by the SLM. The image is reconstructed through multiple measurements using different encoding patterns. The intensity obtained from M measurements is represented by a vector $\textbf {y}$ given by
where $\textbf {x} = \{x_1, x_2, \ldots, x_\textrm {N}\}$ is an N-elements object, and C is an M $\times$ N encoding matrix. The $\textrm {m}_{\textrm {th}}$ measurement $\textbf {y}_{\textrm {m}}$ is calculated as $\textbf {y}_{\textrm {m}} = \textbf {c}_{\textrm {m}} \cdot \textbf {x}$, which is the inner product of $\textbf {x}$ and an encoding pattern displayed on the SLM based on the row components $\textbf {c}_\textrm {m} = \{ \textbf {c}_{\textrm {m1}}, \textbf {c}_{\textrm {m2}}, \ldots, \textbf {c}_{\textrm {mN}} \}$ of $\textbf {C}$. If $\textbf {C}$ is a regular matrix, object x can be reconstructed by the inverse calculation $\textbf {x} = \textbf {C}^{-1}\textbf {y}$. Many different patterns have been used as $\textbf {C}$, such as the identity matrix (raster scan) and Hadamard matrices [9].
Fig. 1. (a) Traditional set up for imaging with single-pixel detector. (b) Concept of imaging with femtosecond-laser-driven sequential light source array.
Figure 1(b) shows the concept of the proposed system. In this method, imaging is performed by using light sources driven by a focused femtosecond laser, and the encoding pattern formed is represented by the spatial position of the light sources. The femtosecond-laser-driven light sources serve as illumination light having a broadband spectrum depending on the excitation target. The broadband light sources directly form the encoding patterns $\textbf {c}_\textrm {m}$ by changing their positions using spatial scanning of the laser. Hence, the encoded illumination patterns formed by femtosecond laser excitation, which we call a laser-driven sequential light source array, overcome the wavelength limitations in the encoding of object light due to the light sources and SLMs in a traditional setup.
2.2 Compressive light source arrays
In imaging technologies based on sequential scanning of sampling points, including our proposed light source arrays, the time for acquiring an image increases significantly as the number of sampling points increases. In this study, compressive sensing (CS) theory and optimization algorithms were used to reduce the imaging time. CS-based imaging has also been applied to solve time-consuming problems in atomic force microscopy [10,11]. In eq. (1), considering M < N, which means that the number of light sources is less than the desired number of pixels in an image, an example of a light source matrix $\textbf {C}$ compressed by removing some rows of the identity matrix is
2.3 Femtosecond-laser-driven light sources on water film
Focused irradiation of materials with near-infrared femtosecond lasers induces photon conversion to non-visible regions, such as X-rays and terahertz waves, in association with a white-light continuum covering the visible region due to their high peak intensity ($>10^{13}$ W/$\textrm {cm}^{2}$). Solid targets such as metals [14] and glasses [15] have been proposed as laser excitation targets for X-ray generation. A liquid target such as a thin water film can refresh the laser irradiation position constantly, and thus enable stable generation of bright X-ray emission with energies of 3–20 KeV (0.4–0.06 nm) under atmospheric pressure [16]. Furthermore, such liquid targets make it possible to control of X-ray spectral features and intensity enhancement by adding with solute chemical agents such as CsCl [17] and gold nanoparticles [18].
X-rays using liquid films are produced based on resonant absorption at the liquid/air interface, which depends on the laser absorption density $W_{\rm abs} \propto \dfrac {n_{\rm e}}{n_{\rm cr}}F_{\rm p}$, where ne and ncr are the free electron density and critical plasma density, respectively, and $\textrm {F}_{\textrm {p}}$ is the laser fluence [19]. A laser irradiating a liquid film produces conduction electrons $n_{\rm e} \propto \textrm {F}_{\textrm {p}}^{\textrm {n}}$ based on n-photon absorption and subsequent heating by the oscillating electric field of the laser light, which triggers a nonlinear increase in $n_{\rm e}$ at the irradiated area. As a result, $n_{\rm e}$ reaches $n_{\rm cr}$ and the surface of the liquid film temporarily becomes metal-like, which induces resonant absorption at the liquid/air interface and emission of bright X-rays. In this study, water films were employed as the excitation target. They are easy to handle under atmospheric pressure, and can produce sufficiently bright illumination sources in the visible and X-ray regions, as well as having scalability in terms of different non-visible light regions, such as terahertz waves [20,21].
3. Experimental setup
Figure 2(a) shows the experimental setup. In this system, imaging is performed using a 2D light source array obtained by scanning the focal point of a femtosecond laser (Micra and Legend Elite Duo, Coherent) on a water film (see Visualization 1). The femtosecond laser has a center wavelength of 800 nm, a repetition frequency of 1 kHz, a maximum average power of 7 W, and a pulse duration of >34 fs. The laser pulses pass through a 2D galvanometer scanner (GM-2020, Canon) and are focused onto a water film by an F$\theta$ lens (S4LFT0089, Sill optics) with a focal length of 88.6 mm and a numerical aperture NA = 0.07. The galvanometer scanner has two mirrors each with a maximum beam deflection angle of 40 degree and a resolution of 3.3 microdegree, and scans the focal spot in the 2D plane to generate the light source arrays. As shown in Fig. 2(b), a water film with a thickness of less than 20 $\mathrm {\mu }$m is produced by colliding distilled water jetted from two nozzles connected to a liquid circulation pump (PMD-221B2M, Sanso) at 45 degrees, and is oriented perpendicular to the optical axis. The light source array in air shown in Fig. 2(c) emits brighter light at the same energy as shown in Fig. 2(d) when the excitation target is a water film. The light source generated on the water film has a spectrum that covers the entire visible region [21]. The available generation region of light sources depends on the size of the water film, which is 10 mm $\times$ 10 mm in this setup.
Fig. 2. (a) X-ray and visible imaging system based on femtosecond-laser-driven sequential light source arrays on a water film. (b) Water flow system. Sequential light source arrays generated by using (c) air and (d) a water film as an excitation target. These photographs were taken by a camera with an exposure time of 13 s.
The generated light source array illuminates an imaging target placed 2 mm behind it. The transmitted light at each illuminated position is selected to be either visible light or X-rays by a flippable mirror, and the respective signals are acquired by a photodiode (S2281-01, Hamamatsu) and a Geiger counter (SS315, Southern Scientific), respectively. The photodiode and the Geiger counter each act as a single-pixel detector. In this system, we call the imaging in each wavelength region the "visible mode" and the "X-ray mode", respectively. The Geiger counter was set at distance of 100 mm from the light source array, and the photodiode was placed via a reduction imaging system formed of 200 mm and 100 mm focal length lenses. Excitation light in the visible mode was removed by an IR cut filter (FESH0750, Thorlabs). The acquisition of spectra and X-ray intensity and the operation of the scanning position of the galvanometer scanner were performed by software written in C++ code on a Windows 10 computer. All experiments were performed under atmospheric pressure at a temperature of around 22 $^{\circ }$C.
4. Experimental results
4.1 Evaluation of light sources
We evaluated the spatial generation of femtosecond-laser-driven light sources in the visible and X-ray regions. Figure 3(a) shows the images of the femtosecond-laser-driven light sources on the water film and a plot of the line through the pixel with the highest intensity in the image. The laser light was radiated with a repetition rate of 1 kHz and a pulse energy of 4.3 $\mathrm {\mu }$J. These images were captured at different positions along with the optical axis using a CMOS image sensor (DMK33UX273, Imaging Source) in the visible mode. The sensor was set above the image plane of a light source serving as 0.0 mm. The position of the sensor was changed toward the laser propagation direction. The image of the light source was observed to expand its illumination area as the sensor moved from the 0.0 mm. At +2.0 mm, where the sample was to be placed, the illumination area was approximately 55 $\mathrm {\mu }$m from the full width at half maximum (FWHM) of the intensity plot. Figure 3(b) shows the images of light sources generated at 8 $\times$ 8 different spatial positions. The sources at each point were generated by beam scanning with a galvanometer scanner and then opening a shutter placed in the optical path. In the visible mode, we demonstrated that femtosecond-laser-driven light sources on a water film could be generated in a 2D array.
Fig. 3. (a) Femtosecond-laser-driven light sources on a water film taken at different distances along the optical direction. Each image shows a plot line through the pixel value with the highest intensity in the image. (b) 8 $\times$ 8 light source array generated by 2D beam scanning.
Figure 4(a) shows the X-ray intensity versus the irradiated pulse energy. 20 measurements were made for a single pulse energy, and each measurement value and its average were connected with a straight line. The X-ray intensities at ${\rm z_1}$, ${\rm z_2}$ and ${\rm z_3}$ were acquired at different positions of the water film in the axial direction relative to the focal point. These positions were found by changing the axial position of water film surface so that the maximum X-ray intensity was obtained when pulses with energy of 172 $\mathrm {\mu }$J, 344 $\mathrm {\mu }$J and 516 $\mathrm {\mu }$J were incident on the water film. The water film positions at ${\rm z_2}$ and ${\rm z_3}$ were shifted by 200 $\mathrm {\mu }$m and 300 $\mathrm {\mu }$m in the opposite direction of the laser propagation with respect to ${\rm z_1}$. The changes in the peak values of ${\rm z_1}$, ${\rm z_2}$ and ${\rm z_3}$ indicate that the X-ray intensity increased as the incident pulse energy increased. On the other hand, changes in pulse energy decreased the X-ray intensity at each water film position. This is considered to be due to the shift of the focal point caused by the self-focusing induced by the optical Kerr effect, resulting in a misalignment between the water film and the focal point. Figure 4(b) shows the spatial X-ray intensity distribution of 8 $\times$ 8 light sources generated by the galvanometer scanner. The X-ray intensity was detected by a Geiger counter. The light sources at each position were generated by irradiation with a pulse energy of 320 $\mathrm {\mu }$J. As a result, we observed that the system could generate X-ray array sources as well as in the visible mode, and we demonstrated that they could serve as illumination light in the respective wavebands.
Fig. 4. (a) X-ray intensity versus the energy of irradiated pulses. ${\rm z_1}$, ${\rm z_2}$ and ${\rm z_3}$ were acquired at the location of the water film where the X-ray intensity was maximum when the incident pulse energies were 172 $\mathrm {\mu }$J, 344 $\mathrm {\mu }$J and 516 $\mathrm {\mu }$J. (b) 8 $\times$ 8 spatial X-ray intensity distribution.
4.2 Visible and X-ray imaging
Figure 5 shows the reconstructed images acquired in the visible and X-ray modes, and the captured samples. The light sources were generated on an array of 32 $\times$ 32 points in an area of 10 $\times$ 10 mm using raster scanning to illuminate the sample. The sample in visible mode was a stainless steel plate with a thickness of 500 $\mathrm {\mu }$m having a hole formed in the shape of a letter "V". The laser light was radiated with a pulse energy of 212 $\mathrm {\mu }$J. The results shows that the system was able to visualize the shape of the sample by illumination using the visible region. The sample in X-ray mode was the same stainless steel plate having an "X" shaped hole placed behind a sheet of paper. The irradiated pulse energy was 267 $\mathrm {\mu }$J. Spatially generated X-ray illumination effectively visualized the shape of the letter behind the paper. Thus, we achieved visible and X-ray imaging in a single system by using femtosecond-laser-driven sequential light source arrays on a water film as the illumination light.
Fig. 5. Imaging target samples and reconstructed images in visible and X-ray modes with 32 $\times$ 32 pixels using light source arrays formed by raster scanning on a water film.
4.3 X-ray images with compressive light source arrays
Figure 6 shows the reconstructed images acquired in the X-ray mode using a compressed light source array, and the light source generation pattern. The coordinates of the white dots in the generation pattern image indicate the position of the light source. A generation rate of 100 % means that 1024 light sources are generated, which is the same as the number of pixels in the reconstructed image. The image target sample was a rectangular stainless steel plate with a thickness of 500 $\mathrm {\mu }$m, illuminated by a light source generated with an irradiated pulse energy of 344 $\mathrm {\mu }$J.
Fig. 6. Reconstructed X-ray images with illumination patterns using light source arrays constructed with 1024, 819, 614 and 204 points, respectively.
The results show that the image was well maintained even when the number of light sources was compressed by a factor of five. We demonstrated that image reconstruction with fewer light sources than the number of image pixels was possible by integrating a compressed sensing algorithm into the system, and that this was effective in reducing the measurement time.
5. Conclusion
We have proposed a broadband imaging system based on the spatially selective generation of femtosecond-laser-driven light sources and demonstrated imaging in the visible and X-ray regions. The light source arrays in the visible and X-ray regions were formed sequentially by employing a water film as the excitation target and scanning the focal point with a galvanometer scanner. The X-ray intensity was sensitive to the position of the focal point along the axial direction with respect to the water film. In order to produce a system that is robust and more stable against misalignment of the focal point, automatic positioning of the water film [22] and design of the position and intensity of the focal point [23] are expected to be effective methods. Our system enabled imaging in the visible and X-ray regions under atmospheric pressure. In the current system, visible and X-ray were switched by a flippable mirror. In the future, we will construct an optical system that can use each wavelength region simultaneously. Simultaneous measurement of visible and X-rays have possibilities to realize image fusion, such that the degraded image including spectral information of the object behind the scattering media can be recovered by the shape information acquired by the X-rays. The available area of generating light sources was 10 $\times$ 10 mm in this setup. If the water film size can be increased by adjusting the nozzle angle of the water jet, the area can be brought closer to 30 $\times$ 30 mm which is the available scanning range of the focal points. Applying a CS technique to the light source generation and image reconstruction reduced the number of light sources required for image acquisition to 20 %, and as a result, the imaging time for a 32 $\times$ 32 pixel X-ray image was reduced from 1024 s to 204 s. We believe that this system has the potential to realize ultra-broadband imaging from in the region from X-rays to terahertz under atmospheric pressure with a single system by providing an optical system that separates object light into specific wavelengths and detectors for each waveband.
Funding
the Collaborative Research Projects of Laboratory for Materials and Structures, Institute of Innovative Research (Tokyo Institute of Technology); the Cooperative Research Program of “Network Joint Research Center for Materials and Devices“, Nanotechnology Platform (Hokkaido University); the National Science and Technology Council (NSTC) of Taiwan (107-2112-M-001-014-MY3, 110-2112-M-001-054); Council for Science, Technology and Innovation (CSTI), Cross-ministerial Strategic Innovation Promotion Program (SIP), “Photonics and Quantum Technology for Society 5.0” (Funding agency: QST); Japan Society for the Promotion of Science (JP21H05584, JP21K17777).
Acknowledgments
K. H. acknowledges the Japan Science and Technology Agency (JST) PRESTO (Precursory Research for Embryonic Science and Technology) Program (SAKIGAKE, Innovative use of light and materials/life) for its support on the original project on X-ray/THz wave simultaneous emission, "Ultrawide band light conversion by controlling structures of microdroplets and ultrashort laser pulse (2009-2013)" and for the laser facilities for the current project.
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. S. K. Sahoo, D. Tang, and C. Dang, “Single-shot multispectral imaging with a monochromatic camera,” Optica 4(10), 1209–1213 (2017). [CrossRef]
2. H. Arguello, S. Pinilla, Y. Peng, H. Ikoma, J. Bacca, and G. Wetzstein, “Shift-variant color-coded diffractive spectral imaging system,” Optica 8(11), 1424–1434 (2021). [CrossRef]
3. G. M. Gibson, S. D. Johnson, and M. J. Padgett, “Single-pixel imaging 12 years on: a review,” Opt. Express 28(19), 28190–28208 (2020). [CrossRef]
4. Y. Hayasaki and R. Sato, “Single-pixel camera with hole-array disk,” Opt. Rev. 27(2), 252–257 (2020). [CrossRef]
5. Y. He, A. Zhang, M. Li, Y. Huang, B. Quan, D. Li, L. Wu, and L. Chen, “High-resolution sub-sampling incoherent x-ray imaging with a single-pixel detector,” APL Photonics 5(5), 056102 (2020). [CrossRef]
6. N. Radwell, K. J. Mitchell, G. M. Gibson, M. P. Edgar, R. Bowman, and M. J. Padgett, “Single-pixel infrared and visible microscope,” Optica 1(5), 285–289 (2014). [CrossRef]
7. R. S. Aspden, N. R. Gemmell, P. A. Morris, D. S. Tasca, L. Mertens, M. G. Tanner, R. A. Kirkwood, A. Ruggeri, A. Tosi, R. W. Boyd, G. S. Buller, R. H. Hadfield, and M. J. Padgett, “Photon-sparse microscopy: visible light imaging using infrared illumination,” Optica 2(12), 1049–1052 (2015). [CrossRef]
8. C. M. Watts, D. Shrekenhamer, J. Montoya, G. Lipworth, J. Hunt, T. Sleasman, S. Krishna, D. R. Smith, and W. J. Padilla, “Terahertz compressive imaging with metamaterial spatial light modulators,” Nat. Photonics 8(8), 605–609 (2014). [CrossRef]
9. N. Sloane and M. Harwit, “Masks for hadamard transform optics, and weighing designs,” Appl. Opt. 15(1), 107–114 (1976). [CrossRef]
10. S. B. Andersson and L. Y. Pao, “Non-raster sampling in atomic force microscopy: A compressed sensing approach,” in 2012 American Control Conference (ACC), (2012), pp. 2485–2490.
11. Z. Wang, J. Zhuang, Z. Gao, and X. Liao, “A fast scanning ion conductance microscopy imaging method using compressive sensing and low-discrepancy sequences,” Rev. Sci. Instrum. 89(11), 113709 (2018). [CrossRef]
12. R. Tibshirani, “Regression shrinkage and selection via the lasso,” J. Royal Stat. Soc. Ser. B 58(1), 267–288 (1996). [CrossRef]
13. S. Boyd, N. Parikh, E. Chu, B. Peleato, and J. Eckstein, “Distributed optimization and statistical learning via the alternating direction method of multipliers,” Found. Trends Mach. Learn. 3(1), 1–122 (2010). [CrossRef]
14. M. Yoshida, Y. Fujimoto, Y. Hironaka, K. Nakamura, K. Kondo, M. Ohtani, and H. Tsunemi, “Generation of picosecond hard x rays by tera watt laser focusing on a copper target,” Appl. Phys. Lett. 73(17), 2393–2395 (1998). [CrossRef]
15. K. Hatanaka and H. Fukumura, “Femtosecond laser-induced x-ray pulse emission from transparent materials including glasses,” New Glass 19, 61–66 (2008).
16. K. Hatanaka and H. Fukumura, “X-ray emission from cscl aqueous solutions when irradiated by intense femtosecond laser pulses and its application to time-resolved xafs measurement of i- in aqueous solution,” X-Ray Spectrom. 41(4), 195–200 (2012). [CrossRef]
17. K. Hatanaka, T. Miura, and H. Fukumura, “Ultrafast x-ray pulse generation by focusing femtosecond infrared laser pulses onto aqueous solutions of alkali metal chloride,” Appl. Phys. Lett. 80(21), 3925–3927 (2002). [CrossRef]
18. F. C. P. Masim, M. Porta, W. H. Hsu, M. T. Nguyen, T. Yonezawa, A. Balčytis, S. Juodkazis, and K. Hatanaka, “Au nanoplasma as efficient hard x-ray emission source,” ACS Photonics 3(11), 2184–2190 (2016). [CrossRef]
19. W. Hsu, F. C. P. Masim, A. Balčytis, S. Juodkazis, and K. Hatanaka, “Dynamic position shifts of x-ray emission froma water film induced by a pair of time-delayed femtosecond laser pulses,” Opt. Express 25(20), 24109–24118 (2017). [CrossRef]
20. H. Huang, T. Nagashima, W. Hsu, S. Juodkazis, and K. Hatanaka, “Dual thz wave and x-ray generation from a water film under femtosecond laser excitation,” Nanomaterials 8(7), 523 (2018). [CrossRef]
21. H. Huang, S. Juodkazis, E. G. Gamaly, T. Nagashima, T. Yonezawa, and K. Hatanaka, “Spatio-temporal control of thz emission,” Commun. Phys. 5(1), 134 (2022). [CrossRef]
22. W. Hsu, F. C. P. Masim, M. Porta, M. T. Nguyen, T. Yonezawa, A. Balčytis, X. Wang, L. Rosa, S. Juodkazis, and K. Hatanaka, “Femtosecond laser-induced hard x-ray generation in air from a solution flow of au nano-sphere suspension using an automatic positioning system,” Opt. Express 24(18), 19994–20001 (2016). [CrossRef]
23. H. Zhang, S. Hasegawa, H. Toyoda, and Y. Hayasaki, “Femtosecond laser-induced hard x-ray generation in air from a solution flow of au nano-sphere suspension using an automatic positioning system,” Opt. Lasers Eng. 151, 106884 (2022). [CrossRef]
