Lensless ghost imaging through the strongly scattering medium
Yang Zhe1, Zhao Lianjie2, Zhao Xueliang1, Qin Wei3, Li Junlin1, †,
State Key Laboratory of Low Dimensional Quantum Physics and Department of Physics, Tsinghua University, Beijing 100084, China
Quantum Engineering Research Center, Beijing Institute of Aerospace Control Devices, Beijing 100854, China
School of Physics, Beijing Institute of Technology, Beijing 100081, China

 

† Corresponding author. E-mail: center@mail.tsinghua.edu.cn

Project supported by the National Natural Science Foundation of China (Grant Nos. 11175094 and 91221205), the National Basic Research Program of China (Grant No. 2015CB921002), and partially supported by the Basic Research Fund of Beijing Institute of Technology (Grant No. 20141842005).

Abstract
Abstract

Lensless ghost imaging has attracted much interest in recent years due to its profound physics and potential applications. In this paper we report studies of the robust properties of the lensless ghost imaging system with a pseudo-thermal light source in a strongly scattering medium. The effects of the positions of the strong medium on the ghost imaging are investigated. In the lensless ghost imaging system, a pseudo-thermal light is split into two correlated beams by a beam splitter. One beam goes to a charge-coupled detector camera, labeled as CCD2. The other beam goes to an object and then is collected in another charge-coupled detector camera, labeled as CCD1, which serves as a bucket detector. When the strong medium, a pane of ground glass disk, is placed between the object and CCD1, the bucket detector, the quality of ghost imaging is barely affected and a good image could still be obtained. The quality of the ghost imaging can also be maintained, even when the ground glass is rotating, which is the strongest scattering medium so far. However, when the strongly scattering medium is present in the optical path from the light source to CCD2 or the object, the lensless ghost imaging system hardly retrieves the image of the object. A theoretical analysis in terms of the second-order correlation function is also provided.

1. Introduction

Ghost imaging (GI), also known as correlated imaging, is a nonlocal imaging technology which retrieves an unknown object by measuring the second-order intensity correlation function between the test light field and the reference light field. The test light field goes to the object and then is collected by a bucket detector without spatial resolution, and the reference light field is detected by a spatial resolving detector without interacting with the object. The spatially point-to-point correlation of both entangled light source[1,2] and the thermal light[38] have been applied to achieve GI. GI also touches the fundamental principles of physics and aroused heated debates and discussions.[912] In addition, many different aspects of the GI have been studied, such as high-order GI,[1315] compressed sensing GI,[1620] differential ghost imaging (DGI),[21] normalized ghost imaging (NGI),[22] and so on.[2330] In recent years, spatial light modulator (SLM) and digital micro-mirror device (DMD) are being developed, and they have also been employed in GI, for instance in the computational ghost imaging (CGI).[31,32] Based on the CGI, three-dimensional image and full-color image of the object have been obtained.[33,34] Furthermore, the influence of atmospheric turbulence, turbid media, or scattering media on GI have also been investigated.[3542]

The lensless ghost imaging (LGI), where there is no lens in the optical paths, has drawn much attention in recent years.[4350] In x-ray or γ-ray imaging, LGI possesses many advantages since the corresponding optical components, e.g., lenses, are very scarce, and are vulnerable to radiation damage. The longitudinal coherence length of LGI was studied in Refs. [51]–[55]. Zhang et al. studied the LGI of a phase object.[56] LGIs by measuring reflected photons were reported in Refs. [57] and [58]. For practical applications of LGI, the effects of atmospheric turbulence, turbid media, or scattering media in the optical paths must be studied.[59] In 2009, Cheng analytically derived a formula for LGI, taking into account of atmospheric turbulence, and showed that the turbulence can result in the degradation of the imaging quality.[60] In 2011, Shih et al. proposed a turbulence-free GI with non-factorizable, point-to-point correlation of thermal light.[61] Later, Bina et al. presented a scheme to achieve the backscattering DGI through the turbid media.[62]

It is desirable to study the influences of the strongly scattering medium on LGI. Ground glass is usually used as a scattering medium in optical systems.[6365] Therefore, we use a pane of ground glass disk, which brings strong scatter and even attenuation, as the scattering medium in this work. We report the first (to the best of our knowledge) experimental investigation of the robustness of LGI in a scattering medium.

The paper is organized as follows. In Section 2, we present the experimental results and theory analysis of LGI system. Section 3 encompasses the study of the robust property of LGI when the scattering medium presents in different optical paths. The discussion and summary are given in Section 4.

2. LGI system

The experimental setup is depicted in Fig. 1. A laser beam working at λ = 780 nm illuminates a rotating ground glass with an angular frequency at 0.1 rotation per minute (RPM) to produce a pseudo-thermal light beam which is spatially incoherent. Then the pseudo-thermal light beam is divided into two beams by a beam splitter (BS). The test beam interacts with an object and is collected by CCD1 which simulates a bucket detector while the reference beam propagates freely to CCD2. d and d1 are the respective distances from the light source to the object and to CCD2. The distance from the object to the photosensitive array of CCD1 is d2. In our experiment we choose d = d1 = 30 cm and d2 = 2 cm. The object is a mask of letters THU depicted in Fig. 2(a). To study the robust property of the system, a strongly scattering medium (a pane of 2-mm thick ground glass disk) are placed in three different positions marked A, B, and C which are sketched respectively in Fig. 1. Position A is in the optical path from the object to CCD1, and position B or C is in the optical path from the light source to the object or CCD2. For simplicity, positions A and B are very close to the object and position C is very close to CCD2.

Fig. 1. Schematic diagram of the lensless ghost imaging system. BS: beam splitter. CCD1, CCD2: charge-coupled device camera (Mintron MTV-1881EX). The ground glass disk is placed in three different positions: (a) behind the object, (b) in front of the object, and (c) in front of CCD2. x1, x2 and y are the transverse coordinates on the test plane, the reference plane and the object plane, respectively.
Fig. 2. (a) The object which is a mask of letters THU. (b) The image captured by CCD1. (c) The ghost image of THU retrieved by calculating the second-order correlation function between CCD1 and CCD2 after averaging over N = 2 × 104 iterations.

In the classical interpretation of GI,[51] the normalized second-order correlation function between the two spatial points of test and reference beams plays a central role and reads

where I1(x1) and I2(x2) denote the intensities of light fields on the test plane and the reference plane at the transverse coordinates x1 and x2, respectively. If the pseudo-thermal light field obeys the Gaussian stochastic process, equation (1) could be rewritten as[46]

where E(x1) and E(x2) are the light fields on the test detector plane and the reference plane, respectively. By optical coherence theory, the last term of Eq. (2) |〈E*(x1)E(x2)〉|2 takes the form of

with G(1)(u1,u2)=I0δ(u1u2) being the first-order correlation function of the spatially incoherent pseudo-thermal light source, where u1 and u2 are the transverse coordinates on light source plane and I0 is constant. According to Huygens–Fresnel theory and under the paraxial approximation, the point spread function (PSF) of the test beam is

where t(y) is the transmission function of the object. The PSF of the reference beam is

Combining Eqs. (3)–(5) the second-order correlation function reads[54]

where sinc function is defined as sinc(x) = sin(x)/x and the coefficient πD/λd determines the spatial resolution of the LGI. Therefore, when the coefficient πD/λd is very large, sinc function in Eq. (6) could be approximated as the δ function and the spatial correlation term takes the form of a convolution product

which is the foundation of the LGI. According to Eqs. (6) and (7), the value of d2 has little influence on LGI as long as it is within the spatial longitudinal coherence length of lensless ghost imaging system with pseudo-thermal light.[52] Therefore, in the optical part from object to CCD1 the LGI is robust against the phase change of the light field.

In this experiment, the traditional ghost imaging algorithm, i.e., normalized iteration algorithm is employed to retrieve the image of the object in the LGI,

where

is the average over N-time measured iterations, S1 is the bucket signal of the test arm and is obtained by summing up intensities of all of the pixels of CCD1.

The classical image of THU is directly captured by CCD1 as depicted in Fig. 2(b), and the image is blurry due to the diffraction of the light field. As showed in Fig. 2(c), the image of THU is retrieved by calculating the second-order correlation function between the test and the reference beam according to Eq. (8), after averaging over N = 2 × 104 iterations. It can be seen from the Figs. 2(b) and 2(c) that the LGI with pseudo-thermal light is robust against the diffraction in the optical path from the object to CCD1 which is consistent with Eqs. (6) and (7).

3. Robust property of the LGI system in the scattering medium

In the following, we report the investigation of the robust property for the LGI system when the strongly scattering medium (a 2-cm thick pane of ground glass disk) is put in the optical path. Firstly, the strongly scattering medium is placed in position A. Then, CCD1 is totally sheltered by the medium. The picture directly captured by CCD1 is showed in Fig. 3(a), so the classical image of THU is thoroughly destroyed by the strongly scattering medium. However, the ghost image of THU can be clearly retrieved by the LGI system as illustrated in Fig. 3(b). Figures 2(c) and 3(b) make it clear that, in position A, the LGI system with a pseudo-thermal light source is robust against the strongly scattering medium. Moreover, even when the ground glass is rotating with different angular frequencies, 0.01 RPM, 0.1 RPM, and 1 RPM, respectively, the quality of the image for the LGI can also be maintained with total iteration number N = 2 × 104 as presented in Fig. 4.

Fig. 3. The strongly scattering medium present in position A: (a) image captured by CCD1 and (b) the ghost image of the THU with N = 2 × 104 iterations.

The robust property against stationary strongly scattering medium in position A of the LGI system is easy to be understood. According to Eq. (8), when the light field has propagated through the object, only the total intensity of CCD1 camera S1 is useful. Thus, even if the strongly scattering medium exists in position A and causes the diffuse transmission of the test light field, S1 only needs to be effectively multiplied by a coefficient factor β < 1. Then, equation (8) changes to

It is seen that β can be reduced, therefore, the LGI system with pseudo-thermal light is robust against the stationary strongly scattering medium in position A. Although CCD1 cannot collect all of the light scattered by the ground glass disk like the real bucket detector, part of the total light intensity is enough to retrieve the image of the object as shown in Fig. 3(b). If the ground glass rotates with a certain angular frequency, the β will change with time: β(t) = β + δ(t), where β is the attenuation term and δ(t) is the fluctuation due to the rotating of the medium. In Eq. (9), β(t) cannot be reduced directly because δ(t) introducing a new noise to the second-order correlation function. However, as long as the fluctuation of β(t) is small and cannot annihilate the fluctuation property of the bucket signal S1, the LGI system will still work as presented in Fig. 4. In other words, if the roughness of the surface for the strongly scattering medium (ground glass) is distributed evenly, LGI is also robust against the rotating strongly scattering medium in position A.

Fig. 4. The ghost image of THU with the ground glass in position A which rotates in different angular frequencies of (a) 0.01 RPM (b) 0.1 RPM, and (c) 1 RPM respectively.

Furthermore, when the strongly scattering medium is presented in other positions of this part of optical path from the object to CCD1, the only change is the value of β when the ground glass disk is stationary or the mean value of the β in the case that the ground glass disk is rotating. Hence, we can infer that not only position A but also the whole optical path from the object to CCD1 are robust against the strongly scattering medium for the LGI system with pseudo-thermal light.

When the strongly scattering medium is located in position B, no effective image could be retrieved from the LGI system as depicted in Fig. 5(a). The reason is that the PSF h(x,u) from the light source to the object cannot maintain the form of Eq. (4) due to the presence of the scattering medium. Therefore, the second-order correlation function between two points of the test and reference beam will not be a sinc or δ function and the spatially point-to-point correlation is not satisfied. Even through the mean value of S1 in the case of position B is nearly equal to that in the case of position A, the LGI cannot work anymore. When the ground glass disk is very close to the object, the ghost imaging system will retrieve t1(y) = t(y) × tglass(y), instead of t(y), where tglass(y) is the transmission function of the ground glass disk, leading to the original t(y) of the object is hard to obtain. Thus, the LGI system is not robust against the strongly scattering medium in position B. When the strongly scattering medium is located in position C, the result is similar to position B as shown in Fig. 5(b). The PSF h(x,u) from the light source to CCD2 will be radically changed by the strongly scattering medium and the spatially point-to-point second-order correlation will also be destroyed, yielding the collapse of the GI. We also could infer that when the strongly scattering medium is present in the other positions of optical paths from the light source to the object or CCD2, LGI will not work.

Fig. 5. The ghost imaging of THU with N = 2 × 104 iterations when the ground glass disk (a) in position B and (b) in position C.
4. Summary

In summary, we have studied the robustness of the LGI system with a pseudo-thermal light source using a pane of ground glass as the strongly scattering medium. When such a medium is located at position A, the LGI is robust against both the stationary and rotating strongly scattering medium. In the case that the strongly scattering medium is stationary, the total intensity of the test beam S1 only needs to be multiplied by an attenuation factor of β < 1 which could be reduced in the normalized second-order correlation function. When the strongly scattering medium is rotating, the attenuation factor β will vary with time, leading to β (t) = β + δ(t). However, as long as the roughness of the surface of the strongly scattering medium is distributed evenly, which means that the fluctuation of δ(t) is small, the fluctuation of the bucket signal will not be destroyed and the image quality of the LGI system could be maintained. When the strongly scattering medium is present in position B or C, the image of the object cannot be retrieved by the LGI system because the spatially point-to-point correlation function between the test and the reference beams has been destroyed by the strongly scattering medium.

In addition to the three positions reported here, we have also investigated other positions experimentally, finding that the results are similar to these three cases. These show that in the LGI system with a pseudo-thermal light source, the optical path from the object to CCD1 is of good robustness against the strongly scattering medium while the paths from the light source to the object and to CCD2 are not robust against the strongly scattering medium. This work can help us deepen the understanding of the LGI and will be of great benefit to the application of the ghost imaging.

Reference
1Pittman T BShih Y HStrekalov D VSergienko A V 1995 Phys. Rev. A 52 R3429
2Strekalov D VSergienko A VKlyshko D NShih Y H 1995 Phys. Rev. Lett. 74 3600
3Bennink R SBentley S JBoyd R W 2002 Phys. Rev. Lett. 89 113601
4Bennink R SBentley S JBoyd R WHowell J C 2004 Phys. Rev. Lett. 92 033601
5Gatti ABrambilla EBache MLugiato L A 2004 Phys. Rev. Lett. 93 093602
6Gatti ABrambilla EBache MLugiato L A 2004 Phys. Rev. A 70 013802
7Ferri FMagatti DGatti ABache MBrambilla ELugiato L A 2005 Phys. Rev. Lett. 94 183602
8Valencia AScarcelli GD’ngelo MShih Y H 2005 Phys. Rev. Lett. 94 063601
9Scarcelli GBerardi VShih Y H 2006 Phys. Rev. Lett. 96 063602
10Shih Y H 2012 Quantum Inform. Process. 11 995
11Shapiro J HBoyd R W 2012 Quantum Inform. Process. 11 1003
12Shapiro J HBoyd R W 2012 Quantum Inform. Process. 11 949
13Li HShi JChen ZZeng G 2012 J. Opt. Soc. Am. A 29 2256
14Liu QChen X HLuo K HWu WWu L A 2009 Phys. Rev. A 79 053844
15Chan K W CO’ullivan M NBoyd R W 2009 Opt. Lett. 34 3343
16Katz OBromberg YSilberberg Y 2009 Appl. Phys. Lett. 95 131110
17Katkovnik VAstola J 2012 J. Opt. Soc. Am. A 29 1556
18Zerom PChan K W CHowell J CBoyd R W 2011 Phys. Rev. A 84 061804
19Li L ZYao X RLiu X FYu W KZhai G J 2014 Acta Phys. Sin. 63 224201 (in Chinese)
20Zhao S MZhuang P 2014 Chin. Phys. B 23 054203
21Ferri FMagatti DLugiato L AGatti A 2010 Phys. Rev. Lett. 104 253603
22Sun BWelsh S SEdgar M PShapiro J HPadgett M J 2012 Opt. Express 20 16892
23Wang K GCao D Z 2004 Phys. Rev. A 70 041801
24Xiong JCao D ZHuang FLi H GSun X JWang K G 2005 Phys. Rev. Lett. 94 173601
25Zhang DZhai Y HWu L AChen X H 2005 Opt. Lett. 30 2354
26Zhai Y HChen X HZhang DWu L A 2005 Phys. Rev. A 72 043805
27Cai YWang F 2007 Opt. Lett. 32 205
28Cai YLin QKorotkova O 2009 Opt. Express 17 2453
29Liu X FYao X RLi M FYu W KChen X HSun Z BWu L AZhai G J 2013 Acta Phys. Sin. 62 184205 (in Chinese)
30Chen L XLei J JRomero J 2014 Light Sci. Appl. 3 e153
31Shapiro J H 2008 Phys. Rev. A 78 061802
32Bromberg YKatz OSilberberg Y 2009 Phys. Rev. A 79 053840
33Sun BEdgar M PBowman RVittert L EWelsh SBowman APadgett M J 2013 Science 340 844
34Welsh S SEdgar M PBowman RJonathan PSun BPadgett M J 2013 Opt. Express 21 23068
35Zhang PGong WShen XHan S 2010 Phys. Rev. A 82 033817
36Gong WHan S 2011 Opt. Lett. 36 394
37Hardy N DShapiro J H 2011 Phys. Rev. A 84 063824
38Dixon P BHowland G AChan K W Cet al. 2011 Phys. Rev. A 83 051803
39Chan K W CSimon D SSergienko A Vet al. 2011 Phys. Rev. A 84 043807
40Shi DFan CZhang PZhang JShen HQiao CWang Y 2012 Opt. Express 20 27992
41Li HShi JZhu YZeng G 2013 Appl. Phys. Lett. 103 051901
42Zhang YShi JLi HZeng G 2014 Chin. Opt. Lett. 12 011102
43Cao D ZXiong JWang K G 2005 Phys. Rev. A 71 013801
44Scarcelli GBerardi VShih Y H 2006 Appl. Phys. Lett. 88 061106
45Basano LOttonello P 2006 Appl. Phys. Lett. 89 091109
46Cheng J 2008 Phys. Rev. A 78 043823
47Chen X HLiu QLuo K HWu L A 2009 Opt. Lett. 34 695
48Cheng JHan S 2004 Phys. Rev. Lett. 92 093903
49Zhang M HWei QShen XLiu Y FLiu H LCheng JHan S 2007 Phys. Rev. A 75 021803
50Cai YWang F 2007 Opt. Lett. 32 205
51Ferri FMagatti DSala V GGatti A 2008 Appl. Phys. Lett. 92 261109
52Liu HLand Han S 2008 Opt. Lett. 33 824
53Gong WZhang PShen XHan S 2009 Appl. Phys. Lett. 95 1110
54Gong WHan S 2010 J. Opt. Soc. Am. B 27 675
55Lu GTian JLin H L 2015 Chin. Phys. Lett. 32 014202
56Zhang D JTang QWu T Fet al. 2014 Appl. Phys. Lett. 104 121113
57Meyers RDeacon K SShih Y H 2008 Phys. Rev. A 77 041801
58Bisht N SSharma E KKandpal H C 2010 Opt. Lasers Eng. 48 671
59Yao Y PWan R GZhang S WZhang T Y 2013 Optik-International Journal for Light and Electron Optics 124 6973
60Cheng J 2009 Opt. Express 17 7916
61Meyers R EDeacon K SShih Y H 2011 Appl. Phys. Lett. 98 111115
62Bina MMagatti DMolteni MGatti ALugiato L AFerri F 2013 Phys. Rev. Lett. 110 083901
63Toida MKondo MIchimura TInaba H 1991 Appl. Phys. B 52 391
64Leith E NDilworth DChen CChen HChen YLopez JSun P C 1991 Opt. Lett. 16 1820
65Indebetouw GKlysubun P 2000 Opt. Lett. 25 212