Capturing Images with Limited Light Using Ghost Imaging

Table of Contents

Introduction
An Inherently Uncertain Technology
A Dynamic Field
     Ditch the Camera, Bring More Computers
     Get it All at Once
     Catching Ghosts in the Dark
Looking “Forward,” in a Slightly Different Direction

Introduction

How many photons are required to create an image? Where do these photons originate? Presently the understanding of capturing an image definitely provides non-supernatural answers to these two questions. When light shines onto an object, the sum of all the photons is reflected into some kind of detector, thus creating an image. If the image is captured digitally as done nowadays, then there may be some amplification of the image, which is then mapped to pixels to generate a two-dimensional representation of the object. Higher pixels and lower noise will provide a better resolution of the resulting image.

Current work in quantum imaging and a method called “ghost imaging” provides basically different answers. In the weird world of quantum mechanics, an image of an object can be formed with light that has never interacted with that object; and the minimum number of photons required may very well be closer to “1” than actually expected.

Ghost imaging enables a high-resolution camera to form an object’s image which is actually not viewed by the camera itself. It employs two sensors: one that looks at the object and another that looks at a light source. Both sensors face different directions and differ considerably: the sensor facing the light source (away from the object) is a high-resolution sensor such as a camera, whereas the sensor facing the object is a single-pixel bucket sensor such as a photodiode.

first ghost images of an opaque object

Now Yanhua Shih of the University of Maryland, Baltimore, and colleagues Ron Meyers and Keith Deaconat the US Army Research Laboratory, also in Maryland, took the first ghost images of an opaque object – a toy soldier.

Later, a computer program compares and integrates the patterns received from the light and object. This creates a “ghost image” - a black-and-white or color picture of the object being photographed. The earliest ghost images were silhouettes, but present setups can produce more detailed objects.

An Inherently Uncertain Technology

By nature, there are some uncertainties in quantum mechanics. Similarly, ghost imaging is not an exception to this. Researchers are not completely sure that a given ghost image is created, or best explained, using classical physics or quantum theory. One possibility for the ghost image is due to quantum entanglement, where separate particles can become “entangled” when measured, thereby losing their individual qualities and becoming parts of a more complicated probability function describing both particles together. Einstein called this “spooky action at a distance.” Another possibility is that enough photons could be similar to each other, and since they are “close enough”, the comparison works well and classical physics still has a strong influence. Either of the two theories or both may be the reason behind ghost imaging.

Ghost imaging

An image based on single-pixel camera data where the single-pixel camera data using random spatial patterns - the number of patterns is 25% of the total number of pixels (a). and an image derived from the same data but under the assumption of same data but filtered under the assumption of the pixel-to-pixel intensity (b). Source: An introduction to ghost imaging: quantum and classical (2016)

A Dynamic Field

There is no need to feel bad if you have not heard much about ghost imaging. The concept is new and is just beginning to be understood, and also the images are still very primitive. The military has been interested in this field since the 00’s, but tactical (or even practical) applications still have a long way to go.

Still, the potential is enormous. Scientists visualize soldiers using a quantum ghost imager to look through battlefield smoke and see around corners. The fact that ghost imaging can use almost any light source from a laser, fluorescent bulb, or even the sun, expands its applications to a large extent. Satellites could capture earth imagery that couldn’t be seen directly due to smoke, clouds, or fog which is beyond the capability of traditional imaging. Therefore, these are only the initial days of this technology, although it has received a lot of attention in 2017 in the form of new research, searching for ways to enhance the technique:

Ditch the Camera, Bring More Computers

In computational ghost imaging (CGI), you don't even need a camera. Instead, many known random patterns are projected on the object to be imaged and a lens captures the photons reflected by the object. The light intensities are measured by the bucket detector. By calculating the correlations between the known random patterns and the measured light intensities, an image of the object is then created. CGI can create object images even in noisy environments. As random patterns are used to image objects, the reconstructed images are contaminated by noise. Improved correlation calculation methods have been formulated to enhance the quality of CGI images.

Although CGI can attain two- or three-dimensional images with a single or a few bucket detectors, noise caused by the reconstruction of images from random patterns reduces the quality of the reconstructed images. In this 2017 study, “Computational ghost imaging using deep learning,” researchers enhanced the quality of CGI images using a deep neural network to automatically learn the features of noise-contaminated CGI images.

Get it All at Once

Ghost imaging usually needs long acquisition times, making it less efficient for dynamic scenes. To deal with this problem, researchers published Single-shot thermal ghost imaging using wavelength-division multiplexing, which puts forth a single-shot thermal ghost imaging scheme through wavelength-division multiplexing method. They produced thousands of patterns at the same time by modulating a broadband light source with a wavelength dependent diffuser. These patterns have the spatial information of the scene and then the correlated measurements are coupled into a spectrometer for reconstructing the final image.

Catching Ghosts in the Dark

Traditional imaging at decreased light level needs hundreds of detected photons per pixel to reduce the Poisson noise for accurate reflectivity inference. Researchers suggest a high-efficiency photon-limited imaging method called First-Photon Ghost Imaging, which recovers the image from the first-photon detection by utilizing the framework of ghost imaging and the physics of low-flux measurements. The experimental results demonstrated that it could retrieve an image by only 0.1 photon detection per pixel, which is three orders lower than the traditional imaging method. This method could find applications in everything from biological microscopy to remote sensing.

In this presentation, from the WITS Structured Light Laboratory, researchers showed how you could even build a “simple” ghost imaging system to teach graduate students.

Looking “Forward,” in a Slightly Different Direction

The first ghost imaging experiment was demonstrated in 1995 and after that the techniques have moved rapidly. The increasing ability to create and manipulate quantum states of light has enabled the development of new technologies that utilize the apparently strange properties of quantum states as information and now imaging. Quantum imaging offers the new possibility to overcome the limits of current optics, which are ruled by classical physics. Instead of just lenses, polarization, SNR, and wavelengths, quantum resources like sub-poissonian statistics and entanglement can be used to carry out ghost imaging, sub Rayleigh imaging, and sub-shot noise imaging for actual practical applications.

Teledyne DALSA

This information has been sourced, reviewed and adapted from materials provided by Teledyne DALSA.

For more information on this source, please visit Teledyne DALSA.

Tell Us What You Think

Do you have a review, update or anything you would like to add to this article?

Leave your feedback
Submit