Electro-Optical Sensors for Atmospheric Turbulence Strength Characterization with Embedded Edge AI Processing of Scintillation Patterns (original) (raw)
Related papers
Analysis of deep learning architectures for turbulence mitigation in long-range imagery
2020
In long range imagery, the atmosphere along the line of sight can result in unwanted visual effects. Random variations in the refractive index of the air causes light to shift and distort. When captured by a camera, this randomly induced variation results in blurred and spatially distorted images. The removal of such effects is greatly desired. Many traditional methods are able to reduce the effects of turbulence within images, however they require complex optimisation procedures or have large computational complexity. The use of deep learning for image processing has now become commonplace, with neural networks being able to outperform traditional methods in many fields. This paper presents an evaluation of various deep learning architectures on the task of turbulence mitigation. The core disadvantage of deep learning is the dependence on a large quantity of relevant data. For the task of turbulence mitigation, real life data is difficult to obtain, as a clean undistorted image is ...
Remote Sensing, 2020
Spectral aerosol optical depth (AOD) estimation from satellite-measured top of atmosphere (TOA) reflectances is challenging because of the complicated TOA-AOD relationship and a nexus of land surface and atmospheric state variations. This task is usually undertaken using a physical model to provide a first estimate of the TOA reflectances which are then optimized by comparison with the satellite data. Recently developed deep neural network (DNN) models provide a powerful tool to represent the complicated relationship statistically. This study presents a methodology based on DNN to estimate AOD using Himawari-8 Advanced Himawari Imager (AHI) TOA observations. A year (2017) of AHI TOA observations over the Himawari-8 full disk collocated in space and time with Aerosol Robotic Network (AERONET) AOD data were used to derive a total of 14,154 training and validation samples. The TOA reflectance in all six AHI solar bands, three TOA reflectance ratios derived based on the dark-target assu...
Deep learning velocity signals allows to quantify turbulence intensity
2019
Turbulence, the ubiquitous and chaotic state of fluid motions, is characterized by strong and statistically non-trivial fluctuations of the velocity field, over a wide range of length- and time-scales, and it can be quantitatively described only in terms of statistical averages. Strong non-stationarities hinder the possibility to achieve statistical convergence, making it impossible to define the turbulence intensity and, in particular, its basic dimensionless estimator, the Reynolds number. Here we show that by employing Deep Neural Networks (DNN) we can accurately estimate the Reynolds number within 1515\%15 accuracy, from a statistical sample as small as two large-scale eddy-turnover times. In contrast, physics-based statistical estimators are limited by the rate of convergence of the central limit theorem, and provide, for the same statistical sample, an error at least 100100100 times larger. Our findings open up new perspectives in the possibility to quantitatively define and, theref...
SPIE Proceedings, 2004
During long-term experiments FGAN-FOM measured C n 2 values over land with identical scintillometers in two different climates, in moderate climate in mid-Europe and in arid climate. Since C n 2 usually changes as a function of time-of-day and of season its influence on electro-optical systems can only be expressed in a statistical way. The cumulative frequencies of occurrence were calculated for a time period of one month for different times of the day. The statistical analysis was applied to calculate the effects of atmospheric turbulence on sensor performance like turbulence MTF, the resolution limit due to turbulence and intensity fluctuation. The calculations were performed for a SWIR sensor (active imaging system) and for typical MWIR and LWIR warning sensors. Turbulence MTF were calculated for a slant path of 5 km from the ground to a height of 100 m for upward and downward looking cases. For horizontal paths at a height of 2 m and 30 m the resolution limits due to turbulence were compared with the corresponding diffraction-limited ones. Calculations of the normalized intensity fluctuations were carried out for two slant propagation paths (zenith angle β = 30° and 80°).
IEEE Transactions on Geoscience and Remote Sensing
In this study, a new deep learning method was developed to estimate the spatiotemporal properties of the hourly aerosol optical depth (AOD) because existing physical models are limited in their abilities to separate the reflectance between aerosols and the underlying surface over land, accurately and effectively. By incorporating geostationary ocean color imagery (GOCI), multispectral bands were applied to train data-driven models to estimate the high-spatiotemporal-resolution AOD over Northeast Asia. Physical model and traditional machine learning (ML) models (the random forest (RF) and support vector regression (SVR) models) were compared with the deep neural network (DNN) model to evaluate its accuracy, implementing hold-out validation and k-fold cross-validation approaches. In the statistical results of the hold-out validation, the DNN model showed the higher accuracy (root mean square error (RMSE) = 0.112, mean bias error (MBE) = 0.007, and correlation coefficient (R) = 0.863) relative to the traditional SVR (RMSE = 0.123, MBE = −0.010, and R = 0.833) and RF (RMSE = 0.125, MBE = 0.004, and R = 0.825) models. The DNN model also exhibited the best performance for most statistical metrics among the traditional SVR, RF, and selected physical models (except for the correlation coefficients and index of agreement) in the spatial and temporal cross-validation analyses. Although the DNN model was trained using the match-up dataset between the top of atmosphere (TOA) reflectance from GOCI multispectral bands and AErosol RObotic NETwork measurements, it showed high spatial and temporal generalization performance owing to its deeper and more complicated network structure. Hourly GOCI AOD data obtained using a deep learning approach with high accuracy are expected to be useful for the quantification of aerosol contents and monitoring of diurnal variations in the AOD.
Deep Neural Networks for Aerosol Optical Depth Retrieval
Atmosphere, 2022
Aerosol Optical Depth (AOD) is a measure of the extinction of solar radiation by aerosols in the atmosphere. Understanding the variations of global AOD is necessary for precisely determining the role of aerosols. Arctic warming is partially caused by aerosols transported from vast distances, including those released during biomass burning events (BBEs). However, measuring AODs is challenging, typically requiring active LIDAR systems or passive sun photometers. Both are limited to cloud-free conditions; sun photometers provide only point measurements, thus requiring more spatial coverage. A more viable method to obtain accurate AOD may be found through machine learning. This study uses DNNs to estimate Svalbard’s AODs using a minimal set of meteorological parameters (temperature, air mass, water vapor, wind speed, latitude, longitude, and time of year). The mean absolute error (MAE) between predicted and true data was 0.00401 for the entire set and 0.0079 for the validation set. It w...
Experimental Machine Learning Approach for Optical Turbulence and FSO Outage Performance Modeling
Electronics
A laser beam propagating in the free space suffers numerous degradation effects. In the context of free space optical communications (FSOCs), this results in reduced availability of the link. This study provides a comprehensive comparison between six machine learning (ML) regression algorithms for modeling the refractive index structure parameter (Cn2). A single neural network (ANN), a random forest (RF), a decision tree (DT), a gradient boosting regressor (GBR), a k-nearest neighbors (KNN) and a deep neural network (DNN) model are applied to estimate Cn2 from experimentally measured macroscopic meteorological parameters obtained from several devices installed at the Naval Postgraduate School (NPS) campus over a period of 11 months. The data set was divided into four quarters and the performance of each algorithm in every quarter was determined based on the R2 and the RMSE metric. The corresponding RMSE were 0.091 for ANN, 0.064 for RF, 0.075 for GBR, 0.073 for KNN, 0.083 for DT and...
IET Optoelectronics, 2022
One of the main barriers of free space optical (FSO) communication systems is atmospheric turbulence. Various processing techniques at the transmitter, receiver, and transceiver sides are available for addressing this issue; however, they have either high complexity or low performance. Considering this problem, in this study, deep learning (DL) is deployed at the transmitter, receiver, and transceiver sides of an FSO system for constellation shaping, detection, and joint constellation-shaping detection, respectively. Furthermore, the proposed DL-based structures are deployed in an FSO-multi-input multi-output (MIMO) system. As the first investigation over DL for the FSO-MIMO system, different combining schemes including the maximum ratio combiner, equal gain combiner, and the selection combiner are considered. Considering a wide range of atmospheric turbulence, from the weak to the strong regime, the performance of the proposed structures are compared with that of the maximum likelihood (ML) detection. To the best of the authors' knowledge, the main contributions and novelties of this work include considering transmitter learning in the FSO system, designing low complexity DL structures for FSO system applications, and providing complexity analysis for the proposed DL algorithms. The results indicate that the proposed DL-based FSO systems achieve the optimum performance with lower complexity compared with the state-of-the-art conventional FSO systems. For instance, the proposed DL-based detector is almost 2, 3, and 7.5 times faster than the ML detector for modulation orders of 16, 64, and 256, respectively.
Deep Learning Enabled Performance Monitoring of Free Space Optical Communication System
The 9th Optoelectronics Global Conference, 2024
Free space optics is a form of optical communication that uses free space instead of optical fibers to transmit signals. As a result, the optical signal is vulnerable to free-space channel characteristics, such as atmospheric turbulence, fog, rain, smoke, etc., that degrade its quality. Hence, optical performance monitoring seeks to assess the amount of distortion caused by these impairments from the received signals and predicting the parameters associated with the channel condition aids in the construction of adaptive and reliable optical links. In this paper, we investigate the performance of FSO communication system under mild to strong turbulent weather conditions. The optical intensity fluctuation at the receiver due to varying channel conditions in the form of scintillation index and the jitter variance, which dictates the turbulence and pointing errors, is first modeled using a statistical approach. Later, we exploit convolutional neural network (CNN) to predict these parameters. Overall 25 channel scenarios corresponding to various channel conditions are successfully estimated by CNN with a normalized mean square error < -26 dB.
Restoration of turbulence profile from scintillation indices
Monthly Notices of the Royal Astronomical Society, 2003
An algorithm that permits one to measure atmospheric turbulence by statistical analysis of light flux fluctuations in four concentric-ring apertures is described in detail. It consists of computing the scintillation indices for each aperture and pairwise aperture combination and in fitting the set of measured indices to a model with a small number of turbulent layers. The performance of this method is analysed by means of simulations and using the real data from a multi-aperture scintillation sensor. It is shown that a turbulence profile with a vertical resolution of h/h ∼ 0.5 can be reconstructed and that the errors of the measured intensities of turbulent layers are typically around 10 per cent of the integrated intensity. The integral parameters such as the seeing and the isoplanatic angle are measured with few per cent accuracy.