Skip to main content

Optics and Optoelectronics for Parallel Image Processing

Objective

- To develop expertise in both the fabrication and the basic physics of diffractive and miniaturized refractive optical elements usable in image processors and other optical and optoelectronic cellular processors.
- To develop optoelectronic arrays combining the functions of light detection, thresholding, modulation and emission, with emphasis on the association of electronic and optical functions on the same chip. This includes, in particular work, on Asymmetric Fabry Perot Modulators implemented in III-V technology and their combination with transistors operating in an analogue mode.
- To experimentally illustrate the use of such devices in demonstrators, with emphasis on image processing applications; the devices will include those developed inside the networks as well as others made available through external collaborations.
- To link the above subjects with algorithmic research in parallel image processing and pattern recognition.
Refractive polymethyl methacrylate (PMMA) microlens arrays fitted with self aligned monomode fibres have been fabricated. Zero order diffractive optical elements have been calculated. A parallel stochastic algorithm for noise smoothing preserving discontinuities in a grey level image has been demonstrated. The design of asymmetric Fabry Perot structures has been completed. Bragg gratings with close to 100% diffraction efficiency in order 1 either in reflection or in transmission have been fabricated. 91% efficiency has been obtained for a 1 to 16 beam splitter gold coated on engraved polymide with a minimum feature width of 300 nm and a depth of nearly 1 um. The comparative study of optical correlator architectures for the implementation of optimal pattern recognition filters has been completed for the case of a shadow casting incoherent correlator and the joint Fourier transform. A working optical correlator set up has been used to produce rotation and attitude invariant pattern recognition. A combination of singular decomposition and circular harmonic filters has been devised to provide the in plane rotation invariant recognition of a pattern with indication of its orientation. 2-dimensional acoustooptic scanners have been developed.

Funding Scheme

CSC - Cost-sharing contracts

Coordinator

Centre National de la Recherche Scientifique
Address

91403 Orsay
France

Participants (5)

Heriot-Watt University
United Kingdom
Address
Riccarton
EH14 4AS Edinburgh
National Technical University of Athens
Greece
Address
42,Patission Street 42
10682 Athens
THE PROVOST, FELLOWS AND SCHOLARS OF THE COLLEGE OF THE HOLY AND UNDIVIDED TRINITY OF QUEEN ELIZABETH NEAR DUBLIN HEREINAFTER TRINITY COLLEGE DUBLIN
Ireland
Address
College Green
2 Dublin
Thomson-CSF
France
Address
Domaine De Corbeville
91404 Orsay
Université de Toulon et du Var
France
Address

83057 La Garde