KEYWORDS: Machine learning, Computer vision technology, Computer graphics, Education and training, Data modeling, Field emission displays, Deep learning, Visualization, RGB color model, Performance modeling, Principal component analysis, Image classification
Synthetically-generated imagery holds the promise of being a panacea for the challenges of real world datasets. Yet it continues to be frequently observed that deep learning model performance is not as good when trained with synthetic data versus real measured imagery. In this study we present analyses and illustration of the use of several statistical metrics, measures, and visualization tools based on the distance and similarity between real and synthetic data empirical distributions in the latent feature embedding space, which provide a quantitative understanding of the relevant image-domain distribution discrepancy issues hampering the generation of performant simulated datasets. We also demonstrate the practical applications of these tools and techniques in a novel study comparing latent space embedding vector distributions of real, pristine synthetic, and synthetic modified by physics-based degradation models. The results may assist deep learning practitioners and synthetic imagery modelers with evaluating latent space embedding distributional dissimilarity and improving model performance when using simulation tools to generate synthetic imagery training data.
Modeling and simulation of the full electro-optical/infrared observation chain remains an incompletely solved problem, and approximations are made at many stages. Including support for current advances in sensor and imaging-system technology with greater spatial, spectral, and temporal resolution only increases the challenge. In this paper we will present results of a US Navy effort to develop an integrated tool that provides enhanced 3D physics-based EO/IR observation chain modeling support for complex dynamic scenes at hyperspectral radiometric fidelity levels, to support research and development in multiple areas of importance for EO and IR imaging systems. A new prototype software system integrates the US Navy TrueView EO/IR/hyperspectral scene simulation and signature modeling tool with the mature US Army Integrated Performance Model (IPM).
We present some preliminary results and discussion of our ongoing effort to develop a prototype volumetric atmospheric optical refraction simulator which uses 3D nonlinear ray-tracing and state-of-art physics-based rendering techniques. The tool will allow simulation of optical curved-ray propagation through nonlinear refractivity gradient profiles in volumetric atmospheric participating media, and the generation of radiometrically accurate images of the resulting atmospheric refraction phenomena, including inferior and superior mirages, over-the-horizon viewing conditions, looming and sinking, towering and stooping of distant objects. The ability to accurately model and predict atmospheric optical refraction conditions and phenomena is important in both defense and commercial applications. Our nonlinear refractive ray-trace method is currently CPU-parallelized and is well-suited for GPU compute implementation.
We present a novel method for efficient and physically-accurate modeling & simulation of anisoplanatic imaging through the atmosphere; in particular we present a new space-variant volumetric image blur algorithm. The method is based on the use of physical atmospheric meteorology models, such as vertical turbulence profiles and aerosol/molecular profiles which can be in general fully spatially-varying in 3 dimensions and also evolving in time. The space-variant modeling method relies on the metadata provided by 3D computer graphics modeling and rendering systems to decompose the image into a set of slices which can be treated in an independent but physically consistent manner to achieve simulated image blur effects which are more accurate and realistic than the homogeneous and stationary blurring methods which are commonly used today. We also present a simple illustrative example of the application of our algorithm, and show its results and performance are in agreement with the expected relative trends and behavior of the prescribed turbulence profile physical model used to define the initial spatially-varying environmental scenario conditions. We present the details of an efficient Fourier-transform-domain formulation of the SV volumetric blur algorithm and detailed algorithm pseudocode description of the method implementation and clarification of some nonobvious technical details.
The Navy is actively developing diverse optical application areas, including high-energy laser weapons and free- space optical communications, which depend on an accurate and timely knowledge of the state of the atmospheric channel. The Optical Channel Characterization in Maritime Atmospheres (OCCIMA) project is a comprehensive program to coalesce and extend the current capability to characterize the maritime atmosphere for all optical and infrared wavelengths. The program goal is the development of a unified and validated analysis toolbox. The foundational design for this program coordinates the development of sensors, measurement protocols, analytical models, and basic physics necessary to fulfill this goal.
The characterization of atmospheric effects on a propagated laser beam is important to applications ranging from free-space optical communications to high-energy laser systems for ship defense. These applications are frequently developed for a dynamic propagation environment in which either one or both ends of the optical link are moving. The instruments are often constrained by size, weight, and power limitations due to the platforms on which they will be installed. The dynamic nature of the optical link induces several difficulties in link-path instrumentation: turbulence statistics on a continuously changing path are hard to interpret, and the optical instruments must be designed to maintain a high-quality link between beacon and receiver. We will review some of the scintillometer designs and we examine the associated data produced by these different instruments.
The authors have recently developed an optical transmissometer device used for estimation of the visibility and atmospheric extinction coefficient along a horizontal or slant terrestrial path of ranges from 500m out to 6 km. This is a bistatic device using a modulated LED beacon transmitter and an 8” (200mm) primary receiver lens with a silicon (Si) photodetector. We discuss how this device can be used to simultaneously obtain an estimate of the atmospheric turbulence characteristics along the same propagation path, using the optical intensity scintillation effect, without requiring any hardware modifications to the existing device. Device principles of operation are presented, followed by the results of a preliminary proof-of-concept field test which yielded encouraging results showing validity of the basic system design but indicating that additional engineering work is required to resolve some implementation details, and further field testing needed to verify and validate the system.
The selection of the “optimal” operating wavelength for Free Space Optical (FSO) systems has been a subject of some ongoing controversy over the past several decades. Practical FSO systems have been found to suffer severe performance degradation in adverse atmospheric visibility conditions (high extinction/low-transmission) such as fog, haze, and other atmospheric aerosols (smoke, dust). Claims have been made that certain wavelengths offer generally superior performance and reduced attenuation for FSO system operation. We will revisit the problem of optical propagation through atmospheric particulates, and will show that the specific details of the selected aerosol size distribution function (SDF), which specifies the aerosol number density distribution by radius, and the corresponding wavelength-dependent complex refractive indices can significantly influence the total extinction/transmission behavior of various wavelengths and hence the choice of “optimal” wavelength. We will use a variety of realistic atmospheric SDFs to highlight the sensitivity of the “optimal” wavelength to the SDF composition details. A primary result will be a comparison illustrating extinction performance at selected wavelengths across the spectrum of visible to LWIR for a variety of realistic and clearly-defined atmospheric scenarios: urban, desert, maritime, with fogs, hazes, smoke, and dust.
Current transmissometer designs can be physically bulky, electronically complex, and susceptible to background light; ultimately limiting performance. We describe a novel transmissometer design based upon a modulated LED source and an AC-coupled receiver to improve upon the aforementioned shortcomings. The design aims to reduce both complexity and SWAP through the use of a high frequency modulation technique, while ultimately improving SNR and measurement range over a variety of atmospheric conditions. The instrument is a dynamic atmosphere and range transmissometer (DART). First we discuss the theory associated with our technique; particularly addressing how the effects of atmospheric turbulence are handled. Next, we describe the radiometry and calibration procedures for the transmitter and the receiver. We describe the instrument hardware and how the DART was built and tested in the laboratory. Finally, we discuss the field experiment to test the DART against a commercial unit over a 700m coastal path in San Diego. The processed data are compared with concurrent measurements from the Optec LPV-3 commercial transmissometer. Transmission data from the DART tracks the commercial instrument very well over varying atmospheric conditions.
Obtaining accurate, precise and timely information about the local atmospheric turbulence and extinction conditions and aerosol/particulate content remains a difficult problem with incomplete solutions. It has important applications in areas such as optical and IR free-space communications, imaging systems performance, and the propagation of directed energy. The capability to utilize passive imaging data to extract parameters characterizing atmospheric turbulence and aerosol/particulate conditions would represent a valuable addition to the current piecemeal toolset for atmospheric sensing. Our research investigates an application of fundamental results from optical turbulence theory and aerosol extinction theory combined with recent advances in image-quality-metrics (IQM) and image-quality-assessment (IQA) methods. We have developed an algorithm which extracts important parameters used for characterizing atmospheric turbulence and extinction along the propagation channel, such as the refractive-index structure parameter C2n , the Fried atmospheric coherence width r0 , and the atmospheric extinction coefficient βext , from passive image data. We will analyze the algorithm performance using simulations based on modeling with turbulence modulation transfer functions. An experimental field campaign was organized and data were collected from passive imaging through turbulence of Siemens star resolution targets over several short littoral paths in Point Loma, San Diego, under conditions various turbulence intensities. We present initial results of the algorithm’s effectiveness using this field data and compare against measurements taken concurrently with other standard atmospheric characterization equipment. We also discuss some of the challenges encountered with the algorithm, tasks currently in progress, and approaches planned for improving the performance in the near future.
We consider an optical beam propagated through the atmosphere and incident on an object causing a temperature rise. In clear air, the physical characteristics of the optical beam transmitted to the object surface are influenced primarily by the effect of atmospheric turbulence, which can be significant near the ground or ocean surface. We use a statistical model to quantify the expected power transfer through turbulent atmosphere and provide guidance toward the threshold of thermal blooming for the considered scenarios. The bulk thermal characteristics of the materials considered are used in a thermal diffusion model to determine the net temperature rise at the object surface due to the incident optical beam. These results of the study are presented in graphical form and are of particular interest to operators of high power laser systems operating over large distances through the atmosphere. Numerical examples include a CO 2 laser (λ=10.6 μm ) with: aperture size of 5 cm, varied pulse duration, and propagation distance of 0.5 km incident on 0.1-mm copper, 10-mm polyimide, 1-mm water, and 10-mm glass/resin composite targets. To assess the effect of near ground/ocean laser propagation, we compare turbulent (of varying degrees) and nonturbulent atmosphere.
The refractive index structure parameter C 2/n(z) as a function of vertical height z, is a key parameter describing the turbulent intensity of the atmosphere, and is important for modeling and predicting beam propagation behavior. Over the past several decades many vertical C 2/n models have been developed, many based on empirical data from field test campaigns involving difficult in situ measurements by radiosondes, or remote-sensing using scidar/lidar/radar techniques. Each model has its own set of limitations and caveats. We have developed an improved C 2/n parametric model for the maritime environment, which uses the Navy Surface-Layer Optical Turbulence model for the low-altitude surface boundary layer, and merges with a generalized Hufnagel-Valley model for the middle- and upper-altitudes (up to 24 km elevation). It takes inputs of local bulk meteorological measurements and forms an estimate of C 2/n based on Monin-Obukhov similarity theory. We present phase-screen wave-optics propagation simulations comparing our improved model with previous models, in terms of turbulence metrics such as Fried's atmospheric coherence width r0, the scintillation index, and the Strehl ratio for both the weak and strong turbulence regimes, for vertical and slant paths, and for various characteristic regimes of the ratio w0=r0, where w0 is the Gaussian beam waist radius.
Using an 850-nanometer-wavelength free-space optical (FSO)communications system of our own design, we acquired field data for the transmitted and received signals in fog at Point Loma, CA for a range of optical depths within the multiple-scattering regime. Statistical estimators for the atmospheric channel transfer function and the related coherency function were computed directly from the experimental data. We interpret the resulting channel transfer function estimates in terms of the physics of the atmospheric propagation channel and fog aerosol particle distributions. We investigate the behavior of the estimators using both real field-test data and simulated propagation data. We compare the field-data channel transfer function estimates against the outputs from a computationally-intensive radiative-transfer theory model-based approach, which we also developed previously for the FSO multiple-scattering atmospheric channel. Our results show that the data-driven channel transfer function estimates are in close agreement with the radiative transfer modeling, and provide comparable receiver signal detection performance improvements while being significantly less time and computationally-intensive.
We investigate the effects of beam wander on an uncorrected laser system. The goal is to enable an accurate
assessment of irradiance at a receiver or target for a shipboard laser system, and in this paper we show that a
maritime surface layer turbulence model is important for an accurate vertical profile. The approach is to provide
an appropriate and flexible hybrid between high fidelity surface layer similarity theory model and a parametric
regression-based model.
The performance of terrestrial free-space optical communications systems is severely impaired by atmospheric aerosol
particle distributions where the particle size is on the order of the operating wavelength. For optical and near-infrared
wavelengths, fog droplets cause multiple-scattering and absorption effects which rapidly degrade received symbol
detection performance as the optical depth parameter increases (visibility decreases). Using a custom free-space optical
communications system we measured field data in fog within the optical multiple-scattering regime. We investigate the
behavior of the estimated channel transfer function using both real field-test data and simulated propagation data based
on field-test conditions. We then compare the channel transfer function estimates against the predictions computed using
a radiative-transfer theory model-based approach which we also developed previously for the free-space optical
atmospheric channel.
We present some preliminary results from our recent free-space optical communications field test experiments in
the foggy littoral environment along the coast of Point Loma, San Diego, conducted between October 2009 and
June 2010. Our custom-built 850nm lasercomm system uses on-off keyed non-return-to-zero intensity-modulation
and direct-detection to transmit pseudo-random bit sequences (PN-11 codes) at 250 Mbps over a 300m horizontal
atmospheric path. We investigate improvements offered by using the latest Advanced Navy Aerosol Model to
calculate the aerosol size-distribution function, a fundamental input to the radiative transfer code which we use
to generate an estimate of the channel frequency response/impulse-response function. The estimated channel
response function is used to design an equalization filter to correct signal distortion due to multiple-scattering
effects and additive noise. We compare the performance of the Advanced Navy Aerosol Model against the
more simplistic log-normal, Gaussian, and Mooradian Pt. Loma distributions. In this conference proceeding
manuscript we are presenting only preliminary findings of our work in progress. Additional analysis, verification,
and study is required before any final results can be posited from these preliminary findings.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.