Assisting target recognition through strong turbulence with the help of neural networks.

Imaging and target recognition through strong turbulence is regarded as one of the most challenging problems in modern turbulence research. As the aggregated turbulence distortion inevitably degrades remote targets and makes them less recognizable, both adaptive optics approaches and image correction methods will become less effective in retrieving correct attributes of the target. Meanwhile, machine learning (ML)-based algorithms have been proposed and studied using both hardware and software approaches to alleviate turbulence effects. In this work, we propose a straightforward approach that treats images with turbulence distortion as a data augmentation in the training set, and investigate the effectiveness of the ML-assisted recognition outcomes under different turbulence strengths. Retrospectively, we also apply the recognition outcomes to evaluate the turbulence strength through regression techniques. As a result, our study helps to build a deep connection between turbulence distortion and imaging effects through a standard perceptron neural network (NN), where mutual inference between turbulence levels and target recognition rates can be achieved.

Quadrant Fourier transform and its application in decoding OAM signals.

We present a new, to the best of our knowledge, concept of using quadrant Fourier transforms (QFTs) formed by microlens arrays (MLAs) to decode complex optical signals based on the optical intensity collected per quadrant area after the MLAs. From a computational optics viewpoint, we show the most promising use of the QFT in low-cost and passive decoding of laser signals carrying optical angular momenta (OAM) that are prevalent in research frontiers of optical communications, computation, and imaging. There are numerous ways of creating, adding, and combining OAM states in optical waves, while decoding or demultiplexing approaches often turn out to be complicated or expensive. The simple OAM decoder formed by a pair of identical MLAs, which are concatenated in the focal plane and transversely offset by half-pitch length, can accomplish the imaging task with four pixels per cell. By sorting the gradient curls of the optical wave into local quadrant cells, the decoder analyzes the intensity reallocation that is proportional to the gradients and computes the gradient curls accordingly. The low-cost, compactness, and simplicity of the proposed OAM sensor will further promote OAM-based applications, as well as many other applications that exploit the spatial complexity of optical signals.

Light field camera study of near-ground turbulence anisotropy and observation of small outer-scales.

Understanding turbulence effects on laser beam propagation is critical to the emerging design, study, and test of many long-range free space optical (FSO) communication and directed energy systems. Conventional studies make the prevalent assumption of isotropic turbulence, while more recent results suggest anisotropic turbulence for atmospheric channels within a few meters elevation above the ground. As countless FSO systems have been and continue to be deployed in such channels, analysis of anisotropic modelings has become one of the fastest growing areas in FSO research. This in turn motivates new tools that can distinguish anisotropic characteristics to improve both modeling accuracy and physical interpretations. Wavefront sensors such as Shack-Hartmann sensors, interferometers, and plenoptic sensors have been devised and used in experiments; however, they all require rigid alignments that lack resilience against temperature gradient buildup and beam wander. We find that by using a light field camera (LFC) that extracts perturbation of individual light rays, the wave structure function of turbulence can be retrieved with high reliability. Furthermore, we find through experiments that the outer scales of near-ground turbulence tend to be a magnitude smaller than conventional theoretical assumptions, agreeing with new findings by others but being absent in current theoretical modelings. As a result, we believe that the LFC is an ideal candidate in the frontier of turbulence research; it is both commercially available and easy to adapt to turbulence experiments.

Measuring the turbulence profile in the lower atmospheric boundary layer.

Optical turbulence can have a severe effect on the propagation of laser beams through the atmosphere. In free space optics and directed energy applications, these laser beams quite often propagate along a slant or vertical path. In these cases, the refractive index structure function parameter cannot be assumed constant, since it varies with height. How it varies with height, especially in the first few meters above the ground, is not well behaved. Turbulence height profiles have been measured since the 1970s, mainly for astronomical observations. These profiles are usually measured for the atmospheric boundary layer (the layer of air from the ground up to approx. 1 km during day and 100 m during night) and some kilometers above it. We have measured the temperature fluctuations in the first few meters above ground level using a system containing eight resistance thermometer devices, mounted in a row at different spacings. Measurements were made flying this system under a tethered balloon or mounted on a telescoping mast. The temperature structure function parameter, CT2, can be estimated from the temperature fluctuations measured by the 28 different probe pairs and the unique distances between the two probes. Finally, Cn2 is estimated from this temperature structure function parameter and compared to values predicted by a turbulence profile model.

Using turbulence scintillation to assist object ranging from a single camera viewpoint.

Image distortions caused by atmospheric turbulence are often treated as unwanted noise or errors in many image processing studies. Our study, however, shows that in certain scenarios the turbulence distortion can be very helpful in enhancing image processing results. This paper describes a novel approach that uses the scintillation traits recorded on a video clip to perform object ranging with reasonable accuracy from a single camera viewpoint. Conventionally, a single camera would be confused by the perspective viewing problem, where a large object far away looks the same as a small object close by. When the atmospheric turbulence phenomenon is considered, the edge or texture pixels of an object tend to scintillate and vary more with increased distance. This turbulence induced signature can be quantitatively analyzed to achieve object ranging with reasonable accuracy. Despite the inevitable fact that turbulence will cause random blurring and deformation of imaging results, it also offers convenient solutions to some remote sensing and machine vision problems, which would otherwise be difficult.

Phase and amplitude beam shaping with two deformable mirrors implementing input plane and Fourier plane phase modifications.

We find that ideas in optical image encryption can be very useful for adaptive optics in achieving simultaneous phase and amplitude shaping of a laser beam. An adaptive optics system with simultaneous phase and amplitude shaping ability is very desirable for atmospheric turbulence compensation. Atmospheric turbulence-induced beam distortions can jeopardize the effectiveness of optical power delivery for directed-energy systems and optical information delivery for free-space optical communication systems. In this paper, a prototype adaptive optics system is proposed based on a famous image encryption structure. The major change is to replace the two random phase plates at the input plane and Fourier plane of the encryption system, respectively, with two deformable mirrors that perform on-demand phase modulations. A Gaussian beam is used as an input to replace the conventional image input. We show through theory, simulation, and experiments that the slightly modified image encryption system can be used to achieve arbitrary phase and amplitude beam shaping within the limits of stroke range and influence function of the deformable mirrors. In application, the proposed technique can be used to perform mode conversion between optical beams, generate structured light signals for imaging and scanning, and compensate atmospheric turbulence-induced phase and amplitude beam distortions.

Multi-aperture laser transmissometer system for long-path aerosol extinction rate measurement.

We present the theory, design, simulation, and experimental evaluations of a new laser transmissometer system for aerosol extinction rate measurement over long paths. The transmitter emits an ON/OFF modulated Gaussian beam that does not require strict collimation. The receiver uses multiple point detectors to sample the sub-aperture irradiance of the arriving beam. The sparse detector arrangement makes our transmissometer system immune to turbulence-induced beam distortion and beam wander caused by the atmospheric channel. Turbulence effects often cause spatial discrepancies in beam propagation and lead to miscalculation of true power loss when using the conventional approach of measuring the total beam power directly with a large-aperture optical concentrator. Our transmissometer system, on the other hand, combines the readouts from distributed detectors to rule out turbulence-induced temporal power fluctuations. As a result, we show through both simulation and field experiments that our transmissometer system works accurately with turbulence strength Cn2 up to 10-12 m-2/3 over a typical 1-km atmospheric channel. In application, our turbulence- and weather-resistant laser transmissometer system has significant advantages for the measurement and study of aerosol concentration, absorption, and scattering properties, which are crucial for directed energy systems, ground-level free-space optical communication systems, environmental monitoring, and weather forecasting.