Retrieval of reflectance spectra as well as of other level 2 products from hyperspectral remotely sensed data demands an accurate analysis of the attenuation and scattering effects due to aerosols and gases distributed in the atmospheric path. Starting from radiometrically corrected data, target reflectance spectra were obtained by solving the radiative transfer equation using a rather simple physical model, which also takes into account the effects of molecular (Rayleigh) and aerosol (Mie) scattering. We have also investigated the problem of how the environment surrounding the observed target may affect the radiance reaching the imaging sensor. By comparing simulated with measured data improvements in visible and near infrared spectral range with respect to the usually adopted inversion techniques are shown. Some examples of atmospheric correction on data acquired by the MIVIS imaging spectrometer, are presented and discussed.

The Target Acquisition Weapons Software (TAWS) is a strike warfare tactical decision aid. TAWS was originally developed by the U.S. Air Force, but has been significantly upgraded and adapted to meet Army, Navy, and Coast Guard applications. Because of the new modules and enhancements, TAWS needs to be validated for accuracy. This presentation discusses the remote sensing methods used to validate TAWS operation over the marine environment. Zero-range target-background radiance contrasts as viewed along a slant path through the atmosphere are inferred using two different methods. The techniques involve making measurements at varying distances from a target with a calibrated imaging system. The data are extracted from the calibrated images and plotted with an exponential least squares curve fit. Zero-range target-background contrast, transmittance, and detection range can be derived from the resulting equation. A second method for determining the zero-range target-background contrast involves analysis of a single image and then correcting the data for the atmospheric influence using the MODTRAN code. The results, advantages and possible limitations of these techniques are discussed. Also, discussed is the possibility for future improvements to TAWS by replacing some burdensome model calculations with direct inputs from remote sensing and pre-calculated online atmospheric products.

In this paper we present MATISSE 1.1 a new background scene generator, whose goal is to compute spectral or integrated radiance images of natural background, as well as the transmission of a hot gas signature. The spectral bandwidth for this version of the code is from 750 to 3300 cm^-1 (3 to 13 μm) with a 5 cm^-1 resolution. Gaseous absorption is computed by a Correlated K model. The spatial variability of atmospheric quantities (temperatures and mixing ratios, among others) is taken into account, using variable profiles along the line of sight. Natural backgrounds include the atmospheric background, low altitude clouds and the Earth ground. The radiation models used are designed for observation at low spatial resolution of clouds and soils, so a texture model was developed to increase the high spatial resolution rendering in the metric range. Intermediate outputs of the code deliver radiance and transmission restricted to a single line of sight, in which case atmospheric refraction effects are taken into account. Along this line of sight the transmission can also be computed using a line-by-line model, which is useful to propagate the radiation emitted by a hot gas source (fires, aircraft or missile plume). MATISSE 1.1 was released in June 2002, so this paper is devoted to a presentation of the first results obtained with the code and some validation tests.

In recent papers, the authors have described a non-imaging technique for estimating the shape of remotely illuminated objects. The approach for non-imaging target shape estimation and orientation is model-based. In the absence of pointing errors common to strategic systems, a raster scan by the transmitter would produce the convolution of the expected far-field pattern with the object, providing a starting point for distinguishing targets. The referenced papers show that it is possible to distinguish certain target shapes using a statistical confidence approach based on previous work by the authors. This paper describes a set of shape and orientation χ² estimation models for several specific targets, pointing jitters and steered boresight angle. Far-field pattern full-widths-at-half maximum comparable to those used in recent field experiments are employed.

Full order Stokes polarimeters are often composed of an analyzer consisting of a rotating quarter wave plate in front of a horizontal polarizer. A number of measurements are then made with the wave-plate oriented at different angle. The four-element Stokes vector is then computed from a linear combination of these measurements. A disadvantage of this device is that only a limited range of analyzer states can be generated. As a result a large number of measurements may be required to reduce the noise gain in the Stokes vector reconstructor. In this paper we describe a polarimeter based on a linear polarizer and two variable wave plates. It can be shown that such a device can produce an arbitrary polarization state. An active polarimeter consists of a generator stage, which transmits a laser illuminator with different polarization states and a receiver with a polarization analyzer stage. In our system both generator and analyzer stages consist of a horizontal polarizer and two variable wave-plates. A sixteen element Mueller matrix of resolved images is then formed for target characterization.

The Atmospheric Neutral Density Experiment (ANDE) is a low cost mission proposed by the Naval Research Laboratory to demonstrate a method to monitor the thermospheric neutral density at an altitude of 400 km. The primary mission objective is to provide total neutral density along the orbit for improved orbit determination of resident space objects. The ANDE mission also serves as a test platform for a new space-to-ground optical communications technique, the Modulating Retro-reflector Array in Space (MODRAS) experiment. Both are sponsored in part by the Department of Defense Space Test Program. The mission consists of two spherical spacecraft fitted with retro-reflectors for satellite laser ranging (SLR). One spacecraft is completely passive; the other carries three active instruments; a miniature Wind And Temperature Spectrometer (WATS) to measure atmospheric composition, cross-track winds and neutral temperature; a Global Positioning Sensor (GPS); and a Thermal Monitoring System (TMS) to monitor the temperature of the sphere. A design requirement of the active satellite is to telemeter the data to the ground without external protrusions from the spherical spacecraft (i.e. an antenna). The active satellite will be fitted with the MODRAS system, which is an enabling technology for the ANDE mission. The MODRAS system consists of a set of multiple quantum well (MQW) modulating retro-reflectors coupled with an electronics package, which will telemeter data to the ground by modulating the reflected light from laser interrogation beam. This paper presents a mission overview and emphasis will be placed on the design, optical layout, performance, ground station, and science capabilities of the combined missions.

Opto-acoustic phenomena accompanying the propagation of high-power pulsed laser radiation in the atmosphere are used for remote determination of the energetic and geometric parameters of laser radiation. The amplitudes and shapes of acoustic pulses have been calculated for the following typical temporal behavior of Gaussian beams: a triggering pulse, short laser pulse, and harmonically modulated radiation disregarding the effect of kinetic cooling in the laser beam propagation channel. Numerical calculations allow one to determine experimentally the radiation power and the absorption coefficient in air for the chosen mode of laser radiation, because the difference between the calculated dimensionless and measured normalized sound pressure levels remains constant for each temporal behavior and is a function of the sought-after parameters. It is experimentally demonstrated that the geometric and energetic characteristics of laser beams measured by opto-acoustic methods agree well with the results of contact bolometer measurements. The standard deviation does not exceed 15% for the laser energy density W ≤ 13 J/cm². For W > 13 J/cm², the observed discrepancy of the results is explained by nonlinearity of bolometer sensors. An algorithm of reconstructing the spatial structure of laser beams from the data of wire bolometer sensors is suggested and realized. The optimal number of initial projections and readings for each projection has allowed us to reconstruct the beam spatial structure in real time. The algorithm is efficient when the noise level does not exceed 20%. The nonlinear extinction coefficient of laser radiation in the atmosphere also can be retrieved from opto-acoustic measurements without additional information about optical and meteorological situation.

Radar and infrared propagation drastically depend on the meteorological and oceanographic conditions. Concerning a joint sea trial of German research institutes at the Baltic Sea 2001, FWG was responsible for the environmental characterization of the marine boundary layer. In-situ measurements included recordings of atmospheric properties and sea surface parameters. They were studied by two multi-sensor buoys, on board a vessel and with radiosondes. Pressure, air temperature and humidity were measured from the sea surface to 1 km altitude. The free drifting buoys which have been constructed at FWG offer the opportunity to gain unperturbed, time resolved information about environmental parameters up to 5 m above sea level. Based on the in-situ measurements refractivity profiles can be calculated. With the help of the vertical refractivity gradient and the air sea temperature difference, conditions for radar and infrared propagation are determined. Further experimental results include wind speed and direction, wave height, rain rate and other important parameters. Taking advantage of the parabolic equation model radar propagation is calculated numerically. In conclusion, the experimental results and calculations underline the importance of the environmental characterization of the marine boundary layer with high temporal and spatial resolution.

The performance of sensors operating within the marine boundary layer is severely influenced by the actual atmospheric conditions and the sea surface. Propagation models are in existence, which cope with the varying environment and allow a performance prediction for sensors in different bands of the electromagnetic spectrum. Model calculations give evidence for a complementary performance of sensors operating in the IR region and at millimeterwaves (35/94 GHz). To validate existing radar propagation models like TERPEM and to compare IR and mm-wave propagation over sea under various atmospherical conditions, joint experiments were conducted over transmission ranges well beyond the horizon, assisted by a careful characterization of the environment. This paper describes the experimental approach and gives representative results for measurement and simulation. The implications on performance especially for a multispectral (IR/mmW) approach are discussed.

The propagation of IR radiation through the marine boundary layer is very much dependent on vertical temperature gradients. Due to the Air-Sea Temperature Difference (ASTD) the distance to the visible horizon for an imaging system can be shorter (ASTD < 0) or larger (ASTD > 0) than the distance to the geometric horizon. To analyse these phenomena FGAN-FOM took measurements in the mid and long wave IR. Location of the experiment was the Baltic Sea. A ship, equipped with IR point sources, was tracked while it was sailing in and out up to, and beyond, the horizon. Weather conditions during the measurement period showed interesting variations in ASTD and atmospheric turbulence (see paper 4884-11). Especially strong sub-refractive effects have been observed with ASTDs up to -5.0 °C. This paper deals with the analysis of the detection range of point targets under different meteorological conditions. Experimental results are compared with the propagation model IRBLEM (IR Boundary Layer Effects Model) which was developed by DRDC-RDDC - Valcartier, Canada.

An adaptive optics system usually has three basic elements, a wavefront sensor, a deformable element, and a feedback scheme. Typically these components are a Shack-Hartmann sensor, a bimorph or segmented mirror, and a DSP solution for performing the necessary calculations. These components are expensive, and give rise to a complex optical and computational system. In this paper a novel implementation of an adaptive optics system will be discussed. The wavefront sensor is based on an IMP grating to measure the curvature of the incoming light. This sensor has been found to be robust to scintillation, so is applicable to horizontal propagation paths. An OKO technologies deformable mirror is used, and the feedback loop calculations run on a standard Pentium III computer using Windows 2000. Results from recent trials of the system correcting for errors over various horizontal propagation lengths will be shown. Additionally results using this system for laser beam propagation will also be discussed.

Average bit error rates in a free-space laser communication system can be reduced by the use of a partially (spatially) coherent transmitted signal beam. When the spatial coherence of the transmitted signal beam is reduced intensity fluctuations (scintillations) decrease, leading to a reduction in the bit error rate of the optical communication link. A model that offers a method for calculating optimal system performance with full consideration of the lognormal turbulent channel and source beam characteristics is described. The impact of atmospheric turbulence strength and degree of source coherence on the average bit error rate is examined.

An experimental validation of the differential image motion (DIM) lidar concept for measuring C_n² is reviewed. The field validation was performed by building a hard-target analog of the DIM lidar and testing it against a conventional scintillometers on a 300 m horizontal path, throughout a range of turbulence conditions. The test results supported the concept and confirmed that the structure characteristic C_n² can be accurately measured with this method. A practical method is described for extending the validation technique to vertical profiles of C_n².

An infrared (IR) or optical signal propagating along a 'line-of-sight' horizontal or slant path near the earth's surface can encounter substantial perturbations. These perturbations result in refractive distortions (low-frequency modulations that can amplify or reduce a signal) and scintillation (a higher frequency fluctuation in signal intensity). In an effort to elucidate the above issues a field test was conducted at the Naval Air Warfare Center (NAWC) at China Lake, CA, during July 2001. Transmission and scintillation measurements were made along slant, near-surface paths over land at 1280 m and 3850 m with a SSC (SPAWAR Systems Center) San Diego transmissometer that operated in the IR regime and in an almost aerosol-free environment. The field test has revealed that slow-scale refractive effects can create pronounced changes in the recorded one-minute average intensity of the IR source. Scintillation can also generate signal changes by a factor of 5 to 10 over very short time scales. In this paper we explore the relation between the refractive changes and scintillation, as well as models developed to describe and predict transmission and scintillation effects. Models include the exploitation of the propagation factor (a multiplicative factor defined from the local refractive field and geometry of the measuring system) and the use of wavelets as IR signal processors.

This study represents an integrated research capability based on (1) data from a scanning water vapor lidar, (2) a hydrodynamic model (HIGRAD) with a observing routine (VIEWER) that simulates the lidar scanning, and (3) an extended Kalman filter (EKF) algorithm for data assimilation which merges data into a model for the best estimate of the system under study. The purpose is to understand the degree to which the lidar measurements represent faithfully the atmospheric boundary layer's spatial and temporal features and to extend this utility in studying other remote sensing capabilities employed in both field and laboratory experiments. Raman lidar water vapor data collected over the Pacific warm pool and the HIGRAD simulations were first compared with each other. Potential aliasing effects of the measurements are identified due to the relatively long duration of the lidar scanning. The problem is being handled by the EKF data assimilation technique which incorporates measurements, that are unevenly distributed in space and time, into a model that simulates the flow being observed. The results of this study in terms of assimilated data will help to resolve and describe the scales and mechanisms that govern the surface evaporation.

Fourier telescopy offers the capability to obtain images of geosynchronous satellites by trading the need for a high power laser and a very large high optical quality receiver for a very large collecting aperture of lower optical quality. The GLINT Program is managed by the US Air Force Research Laboratory and is funded by a congressional add and completion requires continued support. The proposed collector consists of 40 heliostats, each 10m². The optical quality of the heliostats is not a major issue as the images are obtained by demodulating a received time-series signal. Even with a 4,000m² collector the energy requirements dictate that as many as 200 pulses must be averaged for each triple product to be obtained. With a laser operating at several Hertz, the collection of data for a single triple product will take minutes and the collection of the full set of triplets for a modest system, will take several hours. This paper studies strategies for dealing with dynamic solar panels. During an engagement, the body of the satellite will remain fixed relative to the ground site; however the solar panels will turn to point at the sun. Data collected for a full night will be corrupted. One approach is to take data for a short period over many nights at the same solar time. Examples are presented of image quality versus length of time of data collection. The analysis makes uses of the Air Force Research Laboratory’s Time-Domain Analysis and Simulation for Active Tracking (TASAT) code. The amount of time that data may be collected on a given night is compared to the quality of the image recovered. A successful laboratory demonstration of Fourier telescopy is also described.

The performances of Electro-Optical (EO) systems such as visible or infrared cameras, lasers, operating within the Marine Surface Boundary Layer (MSBL), i.e. at heights up to a few tens of meters above the sea surface, are disturbed by various propagation mechanisms: molecular and aerosol extinctions, refraction and turbulence. By using meteorological measurements collected in various coastal areas (French Atlantic coast, Mediterranean Sea, North Sea), we present in this paper, a statistical study of the distributions of some of the relevant EO propagation parameters.

The Rough Evaporation Duct Experiment (RED) assessed the effects of the air-sea boundary layer on microwave and infrared (IR) signal propagation near the sea surface. The experiment was designed around the Floating Instrument Platform (FLIP) research platform, which was moored 10 kilometers off the northeast shore of Oahu, Hawaii. A 10-kilometer infrared propagation path was created from FLIP to a shore-based receiver and both scintillation and transmission measurements were made around the clock for a two-week period. An accurate model for the propagation of infrared signals in the marine atmospheric surface layer remains an elusive goal. Within the first tens of meters of elevation above the sea surface there are substantial vertical gradients of mass and temperature, and this has a strong effect on the prediction of extinction of the infrared signal. The effectiveness of the propagation models will be investigated and the results from the infrared signal propagation study during RED will be shown.

An extensive data set of aerosol vertical concentration profiles (height range 0.5-15 meters, diameter range 10-75 micron), acquired with the Rotorod³ device at various geographical locations has been statistically analyzed. The analysis supplied a parameterization for the 4^th (large particle) lognormal mode in the Advanced Navy Aerosol Model (ANAM). The analysis revealed a positive correlation between concentration and wind speed, and a negative correlation between concentration and height. No clear dependence of radius and width on meteorological parameters has been found. Numerical simulations with the Dutch-French SeaCluse model supported the results of the statistical analysis. The present version of ANAM predicts the experimental aerosol concentration to within a factor of 3. This result was confirmed using an independent data set. Initial extinction calculations with the ANAM indicate that the addition of the 4^th mode changes aerosol extinction values by about 20% as compared to NAM calculations. Closer to the surface, this number is even higher due to the height dependence of the 4^th mode.

The cost of adaptive optics technology is dominated by the cost of current deformable mirror technology which has a range of price going from $2K to $15K per channel. Liquid crystal technology promises to be at least a couple of orders of magnitude cheaper. Liquid crystals offer other advantages such as reliability, low power consumption and with a huge technological momentum based on a wide variety of industrial applications. In this paper we present some preliminary characterizations of a new, large format liquid crystal device. Such devices have the potential for extremely high-resolution wave-front control due to the over 10,000 corrective elements.

A comparison of the wavefronts recorded simultaneously by a Shack-Hartmann and a Distorted Grating Wavefront Sensor (DGWFS) has been completed. The DGWFS is a phase diversity/wavefront curvature type of sensor using a grating to generate the multiple image planes. Data were collected under simulated propagation conditions using the Advanced Concept Laboratory at Lincoln Laboratory. The sensors were arranged such that both recorded a time varying sample of the wavefront at exactly the same time. Dynamic and static comparisons of the two sensors under conditions that ranged from a benign path, D/r₀ = 2, to a propagation condition with significant scintillation, D/r₀ =50, were completed. The data show that the two techniques measure static, low amplitude, on the order of a few waves, aberrations with little difference. Under conditions where there are significant aberrations, the wavefronts measured by the two sensors show notable differences with the DGWFS exhibiting a smoothed, low passed, rendition of the wavefront. As the aberrations increase to produce scintillated condition the differences become more pronounced.

We present here results using two novel active optic elements, an electro-static membrane mirror, and a dual frequency nematic liquid crystal. These devices have the advantage of low cost, low power consumption, and compact size. Possible applications of the devices are astronomical adaptive optics, laser beam control, laser cavity mode control, and real time holography. Field experiments were performed on the Air Force Research Laboratory 3.6 meter telescope on Maui, Hawaii.

Application of liquid crystal (LC) modulators as phase modulators for wavefront control is important for the development of inexpensive adaptive optics systems. Currently only piston-type LC correctors are available, however the implementation of a modal approach promises much higher optical performances. Besides, implementation of silicon technology allows integration of this device and part of its control electronics in a single chip. It was found that the optical performance of a modal LC corrector is comparable to that of existing deformable mirrors; besides, adjustment of frequency and phase of driving AC voltages can further improve it. We illustrate it by numerical simulation results obtained in the framework of the Kolmogorov statistical theory of atmosphere. Development of a compact integrated device imposes some additional requirements - using of low voltages (units of volts), small amount of energy consumed by LC and integrated circuitry, preferably digital design of electronics. We report the results of feedback loop operations obtained with the device manufactured using different technology. We also discuss the technology and present the design of a modal LC corrector with silicon backplane. Several control techniques are discussed in terms of optical performance, energy consumption, cost of manufacturing and possibility for integration.

An inexpensive large aperture (10 m class) receiver for optical wavelength imaging and remote sensing applications is discussed. The design was developed for active (laser illumination) imaging of remote objects using pupil plane measurement techniques, where relatively low optical quality collecting elements can be used. The approach is also well suited for conventional imaging at lower resolutions when light collection capability is of primary importance. The approach relies on a large aperture heliostat consisting of an array of flat mirror segments, like those used in solar collector systems, to collect light from the region of interest. The heliostat segments are tilted in a manner to concentrate the light, by making the light from all segments overlap at a common point, resulting in a region of higher intensity about the size of a segment at the heliostat “focus”. A smaller secondary collector, consisting of a concave mirror located at the overlap point, further concentrates the light and forms a pupil image of the heliostat. Additional optics near the pupil image collimate the light for efficient transmission though a narrow band interference filter used to reduce sky background, and focus the light onto a PMT, or other sensor, for detection. Several design approaches for the collimating optics are discussed as well as system performance and limitations.

A Shack-Hartmann wave-front sensor has been used to characterize non-isotropic turbulence simulated in a transonic wind-tunnel. Wavefront measurements have been obtained for a large number of turbulent conditions. The phase 2-D power spectra exhibit standard Kolmogorov -11/3 power law but also -17/3 power law in the transverse direction, which appears to be a new characteristic for such turbulent flows. Results are further discussed in terms of the various simulated turbulent parameters.

An increasing number of both civilian and military applications require the motion description of translating targets from a sequence of frames acquired through long atmospheric paths. These images are randomly distorted, due to atmospheric turbulence, although adaptive optics systems can partially compensate for this distortion in real time. In these adverse conditions, a velocimetry technique that is based on the spatio-temporal Fourier transform of a series of images presents several advantageous features. In those cases where the target is very dim or an additional processing time reduction is needed, low-light-level images are recorded. Consequently, we have developed a simulation algorithm that generates atmospherically distorted low-light-level images corresponding to different atmospheric conditions and different degrees of compensation. In this paper, simulated low-light-level images are used to analyze the technique accuracy for estimating the object velocity for several atmospheric conditions and for different correction degrees.

Space-based, high resolution, Earth remote sensing systems, that employ large, flexible, lightweight primary mirrors, will require active wavefront correction, in the form of active and adaptive optics, to correct for thermally and vibrationally induced deformations in the optics. These remote sensing systems typically have a large field-of-view. Unlike the adaptive optics on ground-based astronomical telescopes, which have a negligible field-of-view, the adaptive optics on these space-based remote sensing systems will be required to correct the wavefront over the entire field-of-view, which can be several degrees. The error functions for astronomical adaptive optics have been developed for the narrow field-of-view correction of atmospheric turbulence and do not address the needs of wide field space-based systems. To address these needs, a new wide field adaptive optics theory and a new error function are developed. Modeling and experimental results demonstrate the validity of the wide field adaptive optics theory and new error function. This new error function, which is a new extension of conventional adaptive optics, lead to the development of three new types of imaging systems: wide field-of-view, selectable field-of-view, and steerable field-of-view. These new systems can have nearly diffraction-limited performance across the entire field-of-view or a narrow movable region of high-resolution imaging. The factors limiting system performance will be shown. The range of applicability of the wide field adaptive optics theory is shown. The range of applicability is used to avoid limitations in system performance and to estimate the optical systems parameters, which will meet the system’s performance requirements.

In general case it is appropriate to represent the phase gradient fields g(r) measured by wave front sensor as a sum of potential g_p and .g = g_p + g_s. The phase is separable into two parts. One of them, the potential phase corresponds to the gradient component g_p and is successfully reconstructed by the least squares technique. To reconstruct the other part, so called hidden phase or slope discrepancy component of the phase, the two approach can be used. One of them is associated with detection of branch points and consequent calculation of the hidden phase. The other is based on reconstruction of the hidden phase directly from the measured phase gradient. In the paper we compare both approach to the reconstruction of the phase of optical wave propagating in a turbulent atmosphere under conditions of strong scintillation. We study their effectiveness in dependence on strength of optical turbulence for two different type of adaptive systems.

Correction for turbulent distortions with an adaptive optics system including Shack-Hartmann sensor is considered in the paper. Efficiency of control was analyzed under conditions of singular points development in a wavefront of a laser beam. It was shown that the presence of dislocations causes the discrepancy of wavefront detection by the sensor and as a consequence leads to the control instabilities. It was also shown that dislocations appear mainly in the peripheral regions of the beam. Using this feature and optimizing the size of the sensor the stable adaptive control could be achieved in an adaptive optics system.

It was shown that there exist two sources of errors in an adaptive optics system. The first source appears due to limitations induced by the elements of the system such as a Shack Hartmann sensor and deformable mirror. The second associated with the violation of the optical reciprocity principle in algorithm of phase conjugation, namely, with substitute of beacon amplitude distribution by distribution of a Gaussian beam generated by a laser. Absolute correction of turbulent aberration is possible only in case of strict maintenance of a principle, i.e. in case of phase reversal. In the paper the possibility is considered to realize phase reversal in a linear system and only with the use of phase control of the beam. The system should include two mirrors separated by the vacuum gap of a finite size. Estimations were obtained by correction efficiency on the base of phase conjugation and with the use of two mirror adaptive system.

Statistical properties of a turbulent flow have been investigated experimentally in a shock tube with a turbulence grid. The fluctuations of the refractive index were detected with the aid of a laser-schlieren technique. It was found that the structure function D_n(r) has the form: D_n(r)=c(1-exp(-br)sin(ar)/(ar)),a,b,c being constants, in contrast with Kolmogorov's two-thirds law. These data were used in order to determine the PSF and MTF. Numerical results for the image of an object in a turbulent flow are also presented.

A genetic algorithm (GA) is employed to determine the structure of measured, Shack-Hartmann data and its optimum artificial neural-network (ANN) predictor. In the GA approach there are no preordained architectures imposed. The NN architecture that evolves out of many generations of adaptation can also be interpreted as a mapping of the signal complexity. The GA approach inherently addresses the problems of generalization, over fitting of data, and the trade-off between ANN complexity and performance. One objective was to establish how much improvement could ideally be expected from NNs compared to linear techniques. The principal conclusions are: (i) The main input-output relationship is linear with only a small contribution from the nonlinear elements. (ii) The improvement achievable with ANNs compared to optimal linear predictors was less than a 10% reduction in predictor error. (iii) The optimum temporal input window of tip-tilt data corresponds to the time constant introduced by aperture averaging.

We discuss several schemes and methods, providing the possibility to extend the method of dynamic holographic correction of distortions by means of optically addressed liquid crystal spatial light modulators (OA LC SLMs) to IR spectral range. These are first of all the schemes with TV (computer) relay of interferogram to modulator and double-wavelength dynamic holography schemes. We consider also the performance of the first OA LC SLMs, specially developed for correction in 2.5-4 μm spectral range, providing diffraction efficiency of ~10% in traditional scheme and about 75% in the case of grating with asymmetric profile.