The ability to see through walls is a crucial need for special operations and security forces. Our previous research has demonstrated that centimeter wave (CMW) imaging system operated at around 5 GHz WiFi signal offers a low power solution with good range and penetration capabilities. However, the accuracy of the existing system in scene reconstruction was limited due to computational complexity. In this work, we aim to leverage deep learning (DL) based algorithms to design a scene reconstruction approach with significantly improved accuracy. We utilize the high-fidelity electromagnetic (EM) simulation tool, SABR (Shooting and Bouncing Ray), for RF (Radio Frequency) simulations across different scenes and sensor setups. The backbone of our approach is an encoder-decoder neural network. To accommodate the sparse distribution of transmitter and receiver locations in 3D space, we recognize that the transformer with position encoding is more suitable to be used as our building blocks, as opposed to convolution blocks whose receptive field is the neighborhood grid. Additionally, recognizing the sparse nature of point clouds, our decoder integrates sparse tensors and convolutions via the Minkowski Engine. This innovation in model design not only makes it memory-efficient but also supports higher resolution reconstructions and the utilization of deeper learning architectures. We notice that WiFi 3D scene reconstruction using DL technique is a relatively unexplored problem, and we demonstrate that we are able to reconstruct the scene with resolution close to Rayleigh limit. Our approach has great potential to allow scene reconstruction behind obstacles on low SWaP hardware. It has wide applications such as battlefield, security and surveillance to detect and locate threats, search and rescue mission for trapped or injured under rubble, debris, or even medical fields for remote diagnosis and/or treatment, etc.
Deep learning based vision models have had significant success recently. However, one of the biggest challenges is to apply the established models to new targets and environments where no samples exist for training. To solve this zero-shot learning problem, we formulated a heterogeneous learning domain adaptation scenario where labeled data from other domains are available for the new target and for some other unrelated classes. We developed an innovative zero-shot domain adaptation (ZSDA) method by implementing end-to-end adversarial training with class-aware alignment and latent space feature transformation. We demonstrated the performance improvements in several applications in comparison with traditional unsupervised domain adaptation (UDA) approach including threat detection in X-ray security screening imagery.
Formaldehyde is a trace species that plays a key role in atmospheric chemistry. It is an important indicator of nonmethane
volatile organic compound emissions. Also, it is a key reactive intermediate formed during the photochemical
oxidation in the troposphere. Because the lifetime of formaldehyde in the atmosphere is fairly short (several hours), its
presence signals hydrocarbon emission areas. The importance of measuring formaldehyde concentrations has been
recognized by the National Academy's Decadal Survey and two of NASA's forthcoming missions the GEO-CAPE and
GACM target its measurement. There are several techniques some of which are highly sensitive (detection limit ~ 50
parts-per-trillion) for in-situ measurement of formaldehyde and many reported atmospheric measurements. However
there appear to be no reported standoff lidar techniques for range resolved measurements of atmospheric formaldehyde
profiles. In this paper, we describe a formaldehyde lidar profiler based on differential laser induced fluorescence
technique. The UV absorption band in the 352 - 357nm is well suited for laser excitation with frequency tripled
Neodymium lasers and measuring the strong fluorescence in the 390 - 500nm region. Preliminary nighttime
measurements of formaldehyde were demonstrated with a lidar using a commercial Nd:YAG laser (354.7 nm) with a
rather large linewidth (~.02 nm). The measured sensitivity was ~1 ppb at 1 km with 100 meters range resolution even
with this non-optimized system. In this paper we describe our approach for increasing the sensitivity by many orders
and for daytime operation by improving the laser parameters (power and linewidth) and optimizing the receiver.
Water vapor is the most important atmospheric greenhouse gas, but its variability and distribution, particularly the
vertical profile, are not well known due to a lack of reliable long-term observations in the upper troposphere and
stratosphere. Additional design and testing is necessary to extend Water Vapor Sensor System (WVSS) sensitivity to
water vapor from a threshold of 100 ppmv to 2.8 ppmv to support operational and climate applications. Laser
photoacoustic spectroscopy (LPAS) technique can extend the sensitivity to this level without extending absorption
chamber path or using expensive laser emitting at stronger absorption line. A laser photoacoustic spectroscopy sensor
based on inexpensive telecommunication style packaged, fiber-coupled near IR distributed feedback (DFB) laser diodes
was developed to quantify concentrations of water vapor (H2O), CO2, and methane in ambient air. The LPAS sensor
assembled in a compact package was designed for airborne, real-time measurements of atmospheric components. A
resonant photoacoustic cell is used to increase the photoacoustic signal, electrical modulation is applied to replace
mechanical chopper, and wavelength modulation spectroscopy is used to minimize the interfering background signal
from window absorption in the sample cell. The minimum detection sensitivities (1σ) of 5 ppm at 1.39 μm (5 mW) for
water vapor, 6 ppm at 1.6 μm (15 mW) for CO2, and 3 ppm at 1.6 μm (15 mW) for methane, are reported.
US EPA's Clean Air Act lists 187 hazardous air pollutants (HAP) or airborne toxics that are considered especially
harmful to health, and hence the measurement of their concentration is of great importance. Numerous sensor systems
have been reported for measuring these toxic gases and vapors. However, most of these sensors are specific to a single
gas or able to measure only a few of them. Thus a sensor capable of measuring many of the toxic gases simultaneously is
desirable. Laser photoacoustic spectroscopy (LPAS) sensors have the potential for true broadband measurement when
used in conjunction with one or more widely tunable laser sources. An LPAS gas analyzer equipped with a continuous
wave, room temperature IR Quantum Cascade Laser tunable over the wavelength range of 9.4 μm to 9.7 μm was used
for continuous real-time measurements of multiple gases/chemical components. An external cavity grating tuner was
used to generate several (75) narrow line output wavelengths to conduct photoacoustic absorption measurements of gas
mixtures. We have measured various HAPs such as Benzene, Formaldehyde, and Acetaldehyde in the presence of
atmospheric interferents water vapor, and carbon dioxide. Using the preliminary spectral pattern recognition algorithm,
we have shown our ability to measure all these chemical compounds simultaneously in under 3 minutes. Sensitivity
levels of a few part-per-billion (ppb) were achieved with several of the measured compounds with the preliminary
laboratory system.
Raman lidar techniques have been used in remote sensing to measure the aerosol optical extinction in the lower atmosphere, as well as water vapor, temperature and ozone profiles. Knowledge of aerosol optical properties assumes special importance in the wake of studies strongly correlating airborne particulate matter with adverse health effects. Optical extinction depends upon the concentration, composition, and size distribution of the particulate matter. Optical extinction from lidar returns provide information on particle size and density. The influence of relative humidity upon the growth and size of aerosols, particularly the sulfate aerosols along the northeast US region, has been investigated using a Raman lidar during several field measurement campaigns. A particle size distribution model is being developed and verified based on the experimental results. Optical extinction measurements from lidar in the NARSTO-NE-OPS program in Philadelphia PA, during summer of 1999 and 2001, have been analyzed and compared with other measurements such as PM sampling and particle size measurements.
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.