Optical Light Manipulation and Imaging Through Scattering Media

Citation

Xu, Jian (2021) Optical Light Manipulation and Imaging Through Scattering Media. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/4hkq-dz43. https://resolver.caltech.edu/CaltechTHESIS:09092020-162015646

Abstract

Typical optical systems are designed to be implemented in free space or clean media. However, the presence of optical scattering media scrambles light waves and becomes a problem in light field control, optical imaging, and sensing.

To address the problem caused by optical scattering media, we discuss two types of solutions in this thesis. One type of solution is active control, where active modulators are used to modulate the light wave to compensate the wave distortion caused by optical scattering. The other type of solution is computational optics, where physical and mathematical models are built to computationally reconstruct the information from the measured distorted wavefront.

In the part of active control, we first demonstrate coherent light focusing through scattering media by transmission matrix inversion. The transmission matrix inversion approach can realize coherent light control through scattering media with higher fidelity compared to conventional transmission matrix approaches. Then, by combining the pre-designed scattering metasurface with wavefront shaping, we demonstrate a beam steering system with large angular and high angular resolution. Next, we present optical-channel-based intensity streaming (OCIS), which uses only intensity information of light fields to realize light control through scattering media. This solution can be used to control spatially incoherent light propagating through scattering media. In the part of computational optics, we first demonstrate the idea of interferometric speckle visibility spectroscopy (ISVS) to measure the information cerebral blood flow. In ISVS, a camera records the speckle frames of diffused light from the human subject interferometrically, and the speckle statistics is used to calculate the speckle decorrelation time and consequently the blood flow index. Then, we compare the two methods of decorrelation time measurements - temporal sampling methods and spatial ensemble methods - and derive unified mathematical expressions for them in terms of measurement accuracy. Based on current technology of camera sensors and single detectors, our results indicate that spatial ensemble methods can have higher decorrelation time measurement accuracy compared to commonly used temporal sampling methods.

Thesis Files