Ghost Imaging

Rachel Corzine Author

07/28/2021 Added

83 Plays

Description

Rachel Corzine's presentation on Ghost Imaging, completed under the mentorship of Matthias Fuchs as a part of UNL's Summer Research Program

Searchable Transcript

Search:

[00:00:01.410]Hi everyone,my name is Rachel Corzine
[00:00:03.470]and I conducted my research this summer in ghost imaging.
[00:00:08.860]Ghost imaging is a fairly new imaging method
[00:00:11.050]that was first carried out in the 1990s by Dr. Yanhua Shih
[00:00:14.350]and his coworkers at
[00:00:15.183]the University of Maryland, Baltimore County
[00:00:18.120]and their research was based in the theoretical work
[00:00:20.200]done by David Klyshko.
[00:00:21.870]Figure 1 is a reconstruction of the first ghost image
[00:00:24.240]that they took.
[00:00:25.100]You can see that they imaged the UNBC logo.
[00:00:28.210]So this leads us to the question of what exactly is
[00:00:30.680]ghost imaging.
[00:00:31.780]Conventional camera systems use light that is transmitted or
[00:00:34.870]backscattered by an object,
[00:00:36.415]to form an image of the object on a film
[00:00:38.880]or on a focal point detector array.
[00:00:41.050]Ghost imaging, by contrast, is an indirect imaging method
[00:00:44.130]that takes the image of an object using
[00:00:45.880]spatial intensity correlation measurements.
[00:00:48.520]It is called ghost imaging because
[00:00:49.970]unlike conventional photography,
[00:00:51.570]the light transmitted by the object
[00:00:53.180]never interacts with the camera.
[00:00:55.170]There can be two main types of ghost imaging,
[00:00:57.300]quantum ghost imaging and classical ghost imaging.
[00:01:00.080]Quantum ghost imaging is any ghost imaging methods that
[00:01:02.500]relies on quantum correlations,
[00:01:04.460]that is the correlations of measurements of
[00:01:06.330]individual photons,
[00:01:07.475]and classical ghost imaging is any ghost imaging method
[00:01:10.080]that relies on classical correlations,
[00:01:12.700]that is correlations not on the individual photon scale.
[00:01:17.810]The goal of our research in ghost imaging was to
[00:01:19.810]investigate and comprehend the different methods
[00:01:21.960]of ghost imaging and how they work,
[00:01:24.100]and to develop our own successful ghost imaging design.
[00:01:27.080]The long term goal of our research is to
[00:01:28.710]develop sufficiently robust ghost imaging design
[00:01:31.500]to be used as an imaging method in future research at UNL.
[00:01:34.906]Our research in each type of ghost imaging is to prepare for
[00:01:37.880]carrying out quantum ghost imaging,
[00:01:39.890]as it is the most difficult method of ghost imaging to do,
[00:01:43.070]and is expected to be the most advantageous
[00:01:45.020]ghost imaging method,
[00:01:45.937]having applications in both low light imaging
[00:01:48.207]and subwavelength resolution crystallography.
[00:01:52.200]Figure 2 is a simplified diagram
[00:01:53.980]explaining the design of quantum ghost imaging.
[00:01:56.490]To do quantum ghost imaging,
[00:01:58.800]a laser passes through a nonlinear optical crystal,
[00:02:01.640]producing a pair of quantum entangled photons
[00:02:03.750]through the process of spontaneous parametric
[00:02:05.560]down-conversion.
[00:02:07.070]The quantum entanglement of the photons gives the two beams
[00:02:09.840]of photons correlations in both their position and momentum
[00:02:13.160]for each individual photon.
[00:02:15.200]One beam of photons, the idler beam,
[00:02:17.200]does not interact with the object,
[00:02:18.850]hence its name, and is directed at a pixelated detector,
[00:02:21.815]which is a detector that can record spatial information.
[00:02:25.660]The other beam, the signal beam, is directed at the object
[00:02:28.480]and the transmitted light is recorded by a bucket detector.
[00:02:31.940]A buckets detector cannot record spatial information,
[00:02:34.272]it can only detect when a photon has hit the detector.
[00:02:37.281]Even though no spatial information of the object is able to
[00:02:40.640]be collected by the bucket detector,
[00:02:42.520]and no light transmitted by the object
[00:02:43.883]reaches the pixilated detector,
[00:02:46.190]a coincidence circuit between
[00:02:47.530]the pixelated and bucket detector
[00:02:49.330]allows the image to be formed.
[00:02:51.500]The coincidence circuit works by having that when
[00:02:53.950]the bucket detector receives the photon and the signal arm,
[00:02:56.870]then the pixelated detector will accord the spatial
[00:02:59.270]information of the correlated photons in the idler beam.
[00:03:02.700]Summing over all the photons detected,
[00:03:04.770]a shadow image of the object is formed.
[00:03:07.870]The work we accomplished in ghost imaging included
[00:03:10.070]preparing different aspects of the experiment to be ready to
[00:03:12.237]do quantum ghost imaging,
[00:03:14.170]as the actual full set up to do quantum ghost imaging
[00:03:16.490]is more detailed than figure 2 shows.
[00:03:18.870]To produce the quantum entangled photons
[00:03:20.910]an 800 nanometer laser passes through
[00:03:23.510]a non-linear frequency doubling crystal which yields
[00:03:26.577]400 and 800 nanometer light that are not spatially aligned.
[00:03:30.780]We then focused the 400 and 800 nanometer light into
[00:03:33.790]a third harmonic generation crystal,
[00:03:35.940]which yields 266, 400 and 800 nanometer light.
[00:03:39.820]Directing the 266 nanometer light into
[00:03:42.560]the spontaneous parametric down-conversion crystal
[00:03:45.138]would give 400 and 800 nanometer lights,
[00:03:48.110]but they're now quantum entangled and allow us to do
[00:03:50.660]two color quantum ghost imaging.
[00:03:52.920]We also experimented with finding the most efficient
[00:03:54.939]equipment to use as our bucket detectors specifically
[00:03:58.740]looking into using a photo multiplier tool.
[00:04:03.050]Comparing classical ghost imaging to quantum ghost imaging,
[00:04:06.060]the entangled quantum pair generation source is replaced by
[00:04:09.440]a classically correlated source that is divided into two by
[00:04:12.430]a beamsplitter as shown in figure three.
[00:04:15.220]The classical source can be anything that generates
[00:04:17.330]a changing random pattern in the intensity of a light source
[00:04:20.430]such as passing a coherent light beam through a
[00:04:22.440]rotating ground piece of glass.
[00:04:24.270]One copy of the beam is detected by a camera in the
[00:04:26.620]idler beam.
[00:04:27.640]The other beam eliminates the object in the signal arm,
[00:04:30.250]where the beamsplitter-object distance
[00:04:31.860]is equal to the beamsplitter-camera distance,
[00:04:34.260]so that each detector receives the same pattern
[00:04:36.380]at exactly the same time.
[00:04:38.330]A photodiode acting as a bucket detector detects the light
[00:04:41.210]transmitted by the object, where the photodiode position
[00:04:44.160]relative to the object is unimportant so long as
[00:04:46.890]the full beam is detected.
[00:04:49.940]One type of classical ghost imaging we did
[00:04:51.960]was computational ghost imaging,
[00:04:53.880]which we carried out using two different methods.
[00:04:56.300]The first method of computational ghost imaging we did
[00:04:58.680]is shown in figure 4.
[00:05:00.210]We created a spatially structured electric field using a
[00:05:02.810]spatial light modulator or SLM,
[00:05:05.200]that created a time-evolving photo-mask.
[00:05:08.020]SLM have an array of digital micromirrors
[00:05:10.470]that can reflect light towards an observer
[00:05:12.320]or deflect light away.
[00:05:14.200]The array creates a light pattern of black or white.
[00:05:17.120]A camera is not needed to record the light beam since the
[00:05:19.410]pattern generated is known.
[00:05:21.090]Therefore the light beam does not need to be split
[00:05:23.010]by a beamsplitter.
[00:05:24.280]A coherent light beam eliminates the SLM,
[00:05:26.265]producing a spatially modulated light beam,
[00:05:28.518]and the light beam is then directed at the object
[00:05:31.100]where the backscattered light is recorded by a photodiode.
[00:05:34.130]The photodiode signal varies between -1 and +1,
[00:05:38.070]indicating the strength of correlation between
[00:05:39.910]the random pattern generated and the shape of the object.
[00:05:42.890]A positive photodiode signal indicates a strong correlation
[00:05:45.840]between the random pattern and the shape of the object
[00:05:48.510]and a negative photodiode signal then indicates
[00:05:50.730]a weak correlation between the random pattern
[00:05:53.020]and the shape of the object,
[00:05:54.420]where the random pattern is again known because
[00:05:56.440]the pattern is programmed by the SLM.
[00:05:58.860]To produce the final image,
[00:06:00.040]one sums up all the random patterns generated
[00:06:02.422]multiplied by the average detector signal
[00:06:04.830]subtracted from the detector signal
[00:06:06.930]for that individual random pattern.
[00:06:09.410]The second computational ghost imaging method we did
[00:06:11.890]involves using simply a computer monitor
[00:06:14.010]and a camera as shown in figure five.
[00:06:16.460]In this method of computational ghost imaging,
[00:06:18.840]the light source is the computer monitor itself,
[00:06:21.280]where the random pattern is programmed to appear
[00:06:23.160]directly on the computer screen.
[00:06:24.956]The pattern that we generated was
[00:06:27.530]an alternating pattern of sines and cosines.
[00:06:30.540]The object is placed between the monitor
[00:06:32.190]and a camera directed at the screen.
[00:06:33.959]The camera functions as our bucket detector,
[00:06:36.500]and also allowed us to compare the ghost image result to a
[00:06:39.240]conventional image taken by the camera.
[00:06:41.490]I will be showing the results of this method of
[00:06:43.360]computational ghost imaging.
[00:06:46.500]Image 1 is the image taken by the camera
[00:06:48.710]of a UNL hand-cut paper logo
[00:06:51.160]and image two is the ghost image taken of the logo.
[00:06:53.910]Both original images were black and white
[00:06:56.050]and were recolored.
[00:06:57.036]Note looking at the ghost image that there's clearly an
[00:06:59.530]alternating intensity sinusoidal type pattern,
[00:07:02.730]and this is an artifact of the methodology that we used
[00:07:05.320]to do computational ghost imaging.
[00:07:08.380]Image 3 is the image of a Knox College hand-cut paper logo
[00:07:11.530]taken with the camera
[00:07:12.780]and image four is a ghost image taken of the logo.
[00:07:15.770]Again, the images were in black and white and recolored.
[00:07:18.340]Note that again, we see the same sinusodial type pattern.
[00:07:21.770]And just for fun, the last ghost image results we have
[00:07:24.700]are images we took of a paper ghost.
[00:07:26.870]Image 5 is the image taken with the camera
[00:07:29.040]and image 6 is the ghost image.
[00:07:32.550]I thoroughly went through and understood
[00:07:34.100]computational ghost imaging,
[00:07:35.716]both as a proof of concept
[00:07:37.720]and as preparation of materials to carry out
[00:07:39.860]quantum ghost imaging,
[00:07:40.852]which is more involved but also more advantageous.
[00:07:44.380]I explored and dealt with nonlinear, optical crystals
[00:07:46.850]and single photon counting sources and the challenges
[00:07:49.540]that they present in the modern laboratory.
[00:07:51.700]The research I did in ghost imaging this summer
[00:07:53.590]has brought us closer to being able to
[00:07:55.120]successfully carry out quantum ghost imaging.
[00:07:57.612]I would like to acknowledge the effort of Tom Reardon,
[00:08:00.340]Alexander Y-A Saw and Hunter Parker.
[00:08:02.420]This material is based on work supported by the
[00:08:04.670]National Science Foundation.
[00:08:06.350]These are the references that I used
[00:08:07.790]and thank you for listening.

The screen size you are trying to search captions on is too small!

You can always jump over to MediaHub and check it out there.

Comments

0 Comments

Ghost Imaging

Description

Searchable Transcript

Comments icon comment

Related Channels

Comments