Nouvelles briques de conception de systèmes intégrés pour la vision infrarouge. D'une approche minimaliste à la caméra sur puce.

Huge efforts are made in the research and industrial areas to design miniature and cheap optical systems. Indeed these new breakthroughs will contribute to spread these systems in new outlets. The aim of this thesis is to investigate new kinds of cooled infrared optical devices. I have shown that a first step in the miniaturization process can be done by integrating the optics inside the dewar, which is called dewar-level integration. However, cryogenic constraints require a more extensive miniaturization and simplification in order to be able to cool down the additional mass. At first, I have carried out a minimalist strategy, which aims to give an imagery property to the dewar without integrating a single optic. In this way, I have designed an infrared amera obscura by replacing the cooled diaphragm with a pinhole. This approach has been improved by coding a circular grating on this diaphragm, so that the angular resolution is improved while keeping the huge focal depth of the pinhole. Despite their outstanding simplicity, both devices suffer from a low throughput. This can be improved by integrating a Wollaston meniscus lens inside the cold chamber (patent pending). I have thus shown that a single cooled lens can lead to a well corrected optical system. Finally, different ways to miniaturize this optical system have been investigated. They are based on multichannel architectures inspired by the apposition compound eye of invertebrates and by the vision of Xenos Peckii. This study will lead to a further step of integration called wafer-level integration since the optical system could be integrated directly on the focal plane array