Noise in optical measurements

Optics has undergone spectacular development over the last few decades, and remains today an evolving science linked to advanced technology.

There are several reasons for this revival:

• the advent of lasers in the 60s, which made coherent light sources available, and the miniaturization of these sources thanks to advances in integrated optics;

• the development of new techniques, such as holography and optical information processing, which have developed with coherent sources;

• the use of powerful computational resources, supported by computer systems with enhanced performance and reduced footprint, to model a wide range of optical components: ray-tracing calculations to optimize imaging lenses, modeling of light propagation in confined waveguides (optical fibers or planar guides, etc.);

• the development of optoelectronics and fiber optic telecommunications;

• the ever-increasing use of optics in metrology applications: fiber optic or integrated optics sensors, laser triangulation, lidars, etc.

The formalism, applications and miniaturization of optical components have led to optics becoming increasingly similar to electronics, and optical measurement has become standard practice both in research laboratories and on industrial production lines. As a result, the recognition of noise and its sources with the aim to eliminate them, as well as the extraction of the signal embedded in the noise, are now essential metrological concerns.

For a proper understanding of the rest of the presentation, we invite the reader to familiarize himself with a number of concepts concerning signal processing and light phenomena. In particular, we recommend prior reading of the articles "Random processes" [R 210] and "Random functions" [R 220] , in the present treatise Mesures et Contrôle, and of the articles "Characteristic parameters of a signal", in the present treatise Mesures et Contrôle. and...