In this paper an over view is given of the PCVD process as applied for the large scale production of optical fibres for telecommunication. The specific merits and potentials of the process, such as the profile independent high deposition rate and excellent controllability are discribed. The current state of the art of the process, as it is used in the Eindhoven production unit, is a deposition rate of 1 g/min., a preform size equivalent to 28 km of fibre and a drawing speed of 4 m/s. Fibre characteristics are well within the requirements imposed by the telecommunication market. The PCVD process has also proven to be suited for the production of dispersion flattened singlemode fibres and high NA graded index fibres for short distance applications. For both fibre types the high refractive index differences obtained with fluorine doping are exploited. Depending upon the market demands all fibre types can be manufactured at the same productivity. Some trends are given towards further increase of productivity and reduction of fibre costs.

Gas lasers have shown to be capable of delivering tens of terrawatt aspeak power or tens of kilowatt as average power. The efficiencies of most high power gas lasers are relatively high compared with other types of lasers. For instance molecular lasers, oscillating on low lying vibrational levels, and excimer lasers may have intrinsic efficiencies above 10%.The wavelengths of these gas lasers cover the range from the far infrared to the ultra-violet region, say from 12000 to 193 nm. The most important properties are the scalability, optical homogeneity of the excited medium, and the relatively low price per watt of output power. The disadvantages may be the large size of the systems and the relatively narrow line width with limited tunability compared with solid state systems producing the same peak power.

It is shown that the second probability density function of a stellar speckle pattern observed at the focus of a large telescope can be used for imaging double stars. For a given binary system, the number of detected photons required to establish the relative positions of the components is derived. It is shown that probability imaging can be performed in a reasonable observing time for average photon counting levels as low as 0.01 event per pixel. At such low light levels, the most significant indicator of the relative positions of the bright and weaker components is given by the simultaneous detection of a single photon event and a double photon event, with the events being spatially separated by a vector corresponding to the separation of the two stars. Under these conditions, this result has repercussions on the design requirements of a two-dimensional photon-counting detector used for probability imaging applications.

The imaging of objects through atmospheric turbulence has been studied for two particular cases: (i) when the telescope diameter D is much smaller than the atmospheric coherence scale ro and (ii) when D >> ro . Case (i) produces randomly moving images and case (ii) results in speckled images. In both cases, triple correlation techniques are investigated and compared with possible alternative methods, such as centroiding for (i) and the Knox-Thompson method for (ii). The triple correlation technique appears to be the most satisfactory one for both problems at low light levels.

To reconstruct one-dimensional near-infrared speckle-images, we recovered the visibility amplitude and phase in the Fourier domain. Comparing the Knox-Thompson and Speckle-Masking methods and generalisations thereof we obtained the best results using Speckle-Masking.

Differential Speckle Interferometry (DSI) is a super-resolution technique, based on the combination of high angular and high spectral resolution, which makes it possible to measure the displacement of the photocenter of an astronomical object, as a function of wavelength, even if it is much smaller than the telescope Airy disk. When applied to rapidly rotating stars, DSI permits the determination of the stellar angular diameter and of the position angle and, sometimes, of the inclination of the stellar rotational axis. If the star is spotted, DSI provides a one dimensional image of the star integrated in the rotational axis direction and some information related to the latitude of the starspots. The study of the temporal variation of the DSI information combined with a Doppler Imaging analysis of the broad-band light curve and of the temporal behavior of the profile of temperature unsensitive absorption lines should make it possible to reconstruct two dimensional maps of the stellar surface. This reconstruction should be easier and more unique than that based on the Doppler Imaging technique alone. This communication is a very preliminary presentation of this technique.

During the last few years it has been shown that speckle masking observations with large single-dish telescopes can yield diffraction-limited images in spite of image degradation by the atmosphere and by telescope aberrations. Much higher resolution can be obtained if the Coude beams of many telescopes are combined coherently in a central station and if the obtained long-baseline speckle interferograms are evaluated by speckle masking. For example, a 10-km array on earth can yield images with the fascinating resolution of 10 arcsec. Labeyrie has for the first time shown that it is possible to combine the Coude beams of two telescopes coherently.

Speckle masking is a triple correlation method that can reconstruct diffraction-limited images from astronomical speckle interferograms. The obtained resolution is about 30 times higher than the resolution of conventional astronomical photography. In this paper we describe a photon-counting version of speckle masking that can be applied to speckle interferograms consisting of a small number of photon events. We show computer simulations which illustrate the feasibility of the method. Finally, we compare photon-counting speckle masking with four-dimensional bispectrum processing, tomographic speckle masking, cross-triple correlation processing and bispectrum processing of photon-counting speckle interferograms.

To match the effect of the seeing of the object and calibrator measurements, we use the phase fluctuation of the individual specklegramms around the average speckle-masking-phase for sorting and binning of the specklegramms. Thereby the phase fluctuations are independent of the true object, hence we can directly match the obtained object and calibrator bins. This yields an improvement of the visibility-amplitude. The results were checked on real one-dimensional near-infrared speckle data.

A new method for generating an image of the un-wrapped phase is discussed which is based on real zero converting the data. The images produced by this method are of value in reducing the effects of clutter and identifying strong scatterers.

When electromagnetic radiation is incident on a fluctuating surface, it is scattered in a manner that is governed by the shape and material characteristics of the surface. Synthetic aperture radar is an imaging system that attempts to classify the nature of the earth's surface by repeatedly emitting a short pulse of micro-wave radiation and recording the back-scattered field.

Maxima in Hough space which correspond to straight line segments have a characteristic 'butterfly' shape. The intensity of regions which correspond to circles have an inverse square root singularity at their edge. Filters are thereby designed which automatically extract line and circle parameters.

Phase retrieval in the optical system is discussed. It is proved there is a finite number of phase functions which are consistent with intensity in the image plane of the diffraction-limited system. The reduction of phase ambiguity in the direction of propagation is discussed. Open problems are posed.

We consider the problem of inverting experimental data obtained in light scattering experiments described by linear theories. We discuss applications to particle sizing and we describe fast and easy-to-implement algorithms which permit the extraction, from noisy measurements, of reliable information about the particle size distribution.

For evaluating AN(mx), the essential term for calculating the Mie coefficients, we used the basic properties of continued fraction to determined the value of N* at which the backward recursion should be started. By giving the initial value of AN*(mx), we calculated AN(mx) and compared with the result of Lentz's.

We discuss two modifications to the profiled singular function method for inversion of multi-exponential data obtained in PCS experiments. The first is a rapid inversion involving look-up tables. The second incorporates the scattered intensity form factor into the profile to allow direct calculation of the particle radius-mass distribution.

In the case of monodisperse dilute systems it is possible to calculate the distance distribution function for homogeneous and inhomogeneous particles of arbitrary shape. The distance distribution function enables one to find a rough classification of the shape and to determine the size of the particle. This function can be deconvoluted to the radial polarization density profile for particles with spherical symmetry. A number, mass or intensity distribution can be calculated from the light scattering data, if the distribution can be described by a single parameter and if it is possible to calculate the shape factor of the particles as it is the case for spheres and spheroids. The range of applicability of the method depends on the experimental set-up, but is in most cases in the size range from 100 nanometers to several microns.

Light-scattering techniques are potentially very important for the low-level detection and identification of particulate species such as asbestos in aerosol and liquid suspensions. Low-level detection is essential because asbestos is a known carcinogen, even at very-low exposure levels. At present, most asbestos particulate monitoring is used on optical microscopy. If detailed analysis is required, then electron microscopy is employed. Both of these methods are labour intensive. Furthermore, the optical microscopy method is not very reliable. Although the light-scattering techniques described here have general applicability, the emphasis is on asbestos measurements. Ordinary measurements of Mie scattering from asbestos suspensions can provide only limited information on asbestos content. owever, a more sophisticated technique can be employed which relies on the fact that asbestos particulate is fibrous rather than spherical in shape, and that the fibres align in a strong magnetic field (approximately 0.5 T). Particulate other than asbestos is generally non-fibrous in shape. Measurements have been carried out on liquid suspensions of asbestos contained in a small cell placed between the poles of a rotating magnet. The aligned fibres, which rotate about their centre of mass as they follow the field, are illuminated using a laser source. The Mie-scattering intensity is measured as a function of rotation angle, and the resulting data is then analysed with the aid of a microcomputer. Intensity maxima and minima provide reliable information on asbestos concentration, even in the presence of strong scattering from other particulate. In addition, the angular location of the intensity peaks provides information on the type of asbestos present. Each type has a characteristic alignment behaviour in a strong magnetic field. Using relatively-simple equipment, chrysotile asbestos (the most commonly-used type) has been detected at levels below 30 ng/l.

This paper first examines theoretically the effect of coating of microbubbles on Mie Scattering as compared with clean bubbles. The effect is relatively small under certain conditions. This analysis leads to propose a new inversion algorithm for determining bubble size by using scattered intensities collected by just a few detectors. Finally, a new laser system specifically designed and constructed for bubble detection is described.

The use of a laser Doppler microscope, developed to measure flow and/or diffusion in living cells, is described. It analyses Doppler shifts given to the frequency of laser light scattered at a known angle from particles moving in a volume of less than 200 cubic micrometres. This scattering volume is defined in the specimen by an aperture in an image-plane and can be placed within a single cell. Specimens can be viewed and recorded continuously by video-enhanced differential interference contrast microscopy during experiments so that, independently of laser-Doppler measurements, sizes of individual particles in images may be measured and their displacements timed with the aid of a video-micrometer. This paper describes how ways are now needed to constrain and improve computation and analysis of the laser Doppler measurements by incorporating information from the images.

An investigation of Laser Doppler Anemometry (L.D.A.) for the measurement of acoustic velocity fields using the photon correlation method of signal analysis has been made. It is shown how estimates of velocity amplitudes can be obtained for the cases of single tone and band limited noise fields. Measurements made in a travelling wave tube are presented and are shown to compare well with microphone measurements. A discussion is also made of the measurement of complex impedances and possible industrial uses of the technique are mentioned.

The objective of this report is the presentation of the experience gained in the experimental determination of the spatial correlations in a turbulent velocity field by means of an especially designed Laser-Doppler (LDV) holographic optics. The determination of the spatial velocity correlations requires the generation of two closely spaced LDV measuring volumes at specified positions. In conventional LDV optics this spacing is limited by the size of the optical components. The holographic optical elements (HOE) used in this work facilitate the positioning of the measuring volume at a spacing of 200 micrometers or better. Basic design considerations, the manufacturing of the holographic optics and the data acquisition are discussed.

The scattering of electromagnetic radiation from a dielectric zone-plate of finite thickness is considered. It is shown that the scattering at X-ray wavelengths is weak, so enabling the Born approximation to be used in formulating the diffraction theory. For the usual design of zone plate, the scattered field is a separable function of the thickness and the separation of the Fresnel zones, enabling the variations of either parameter to be studied independently. It is shown that the effects of finite thickness on volume scattering from dielectric material are small.

The optical phase retrieval problem occurs in imaging, microscopy and several types of interferometry. This paper is concerned with retrieving an optimal estimate for the Fourier phase from a restricted number of noisy Fourier magnitude (or power spectrum) data samples. Of particular interest is the case when the direct Fourier transform of such limited data provides a poor object estimate even when the correct complex data are available. A method to achieve this is described based on the minimisation of a phase dependent cost function and the results of an approximate but sub-optimal iterative implementation are discussed.

Problems of information extraction associated with the laser light scattering studies of fluid interfaces are discussed. Such experiments may be affected by several interfacial properties. The uniqueness and stability of the estimates of these properties determined by various data analysis schemes will be examined. In particular the direct fitting of experimental data by the theoretical spectrum is discussed. This approach appears optimal, those problems of resolution which remain being fundamental to the problem.

Low angle elastic light scattering techniques have been successfully applied to the study of solid particle size distributions. In this work we evaluate the possibility of extending this method to the study of fractal objects, like aggregating colloids. We present computer simulations on a peculiar iterative inversion scheme to recover the cluster gyration radiuses distribution. The computer simulations are made assuming realistic estimates of the noise associated to the measurements and the results of the data inversion procedure will be compared with the input data. The rate of convergence with the number of iterations will also be discussed.

We consider the problem of restoring a signal degraded by a band-pass filter, in the case where this signal is modulated by a given profile function. We study the properties of the associated singular system and derive analytic expressions for the singular functions in some particular cases.

I use information theoretic techniques to derive schemes for the Bayesian analysis of images with spatially homogeneous statistical properties. In any particular case the scheme is equivalent to deducing the structure of the Markov random field which models the data. This scheme may also be viewed as a generalised sampling technique where the data is reduced by a set of sampling functions to a more compact set of data, which nevertheless retains all the information content of the original data.

The usual regularization theory for posed ill-problem may be called the regularization of Weierstrass type in the context of approximation theory. On the contrary, this paper proposes the regularization of Chebyshev type, named the "projection filter regulariza-tion". In the absence of noise, the projection filter provides the best approximation Pf to the real solution f, where P is the orthogonal projection operator onto a subspace. In the presence of noise, it provides the best approximation to the individual Pf in the average sense with respect to noise. A general form of projection filter is obtained. Properties of projection filter are discussed in detail.

A non destructive optical technique for testing transparent media is applied to the study of concentration change in a supersaturated solution near a growing crystal.

A telescope with a long thin pupil, called SAT appears to be a promising alternative to the classical telescope for high resolution astronomical imagery. For such application a SAT must be able to rotate around its optical axis. The problem is to faithfully reconstruct the two-dimensional image of an observed object, from the one-dimensional data set (Radon transform) provided by the SAT during its rotation. This problem of image reconstruction from projections has been investigated in the field of medical imagery. The proposed algorithm, illustrated in this paper, for the reconstruction of astronomical images, is based upon the use of Fast Fourier Transforms (FFT). Although the reconstruction is accurate for unperturbed images, ground based observations are strongly affected by atmospheric turbulence ; under such conditions, we propose the use of phase closure, thus enabling unperturbed astronomical images to be retrieved. Image reconstruction entered radioastronomy in the 1950's in response to a need for high angular resolution in the mapping of radio emission from celestial objects. The technique used was that of aperture synthesis.

In this paper we consider the back-propagation algorithm of Devaney which is based on the Rytov approximation. To examine its validity we first generate scattered data using exact solutions for acoustic scattering by a penetrable elliptic cylinder. We then run these data through the inversion algorithm to find the reconstructed object. In this manner we determine how the accuracy of the algorithm depends on the wavelength of the incident wave and the number of views employed, as well as on the size, geometry and contrast of the object.

A new adaptive technique for heat images reconstruction from integral measurements of coherence functions is presented. By defining the energetical norms of special Sobolev's structure on a compact solution space we formulate a generalized adaptive least-square problem for brightness estimation. The results presented here follow our previous work on the restoration of the random fields by the decision theory methods.

The estimation problem of the frequency-wavenumber spectral density function of the heat radiation field of the spatially-spread objects is suggested to be considered as a multi-dimensional spectral analysis boundary-value problem for an initial field in the stochastic measurement channel. Defining the regularization constraints, incorporating the boundary conditions, we formulate a generalized maximum-likelihood problem for spectral density function estimation and then apply standard optimization techniques to derive an optimal spectral analysis algorithm.