Abstract: An emitter for emitting electromagnetic pulses includes a generator for generating an electromagnetic pulse, an optical fiber that transmits the eletromagnetic pulse, and an optical cavity. The optical cavity is placed in the path of the incident electromagnetic pulse transmitted by the optical fiber and has an input provided with a first partially reflecting mirror and an output provided with a second partially reflecting mirror. The first partially reflecting mirror is placed between first and second fiber lengths of the optical fiber and the second partially reflecting mirror is placed between the second and a third fiber length of the optical fiber. The length of the second fiber and a transmission/reflection ratio of the first and second mirrors are adjusted such that a train of electromagnetic pulses, having variable geometry characteristics, is emitted.

Abstract: A system for transmitting optical information uses several transmitters and receivers connected by optical fibers. The transmitters and receivers are arranged in individual networks in each of which all the optical fibers leaving the transmitters are coupled to one another and all the optical fibers arriving at the receivers are coupled to one another. In addition, all the optical fibers leaving the transmitters in each individual network are interconnected to the corresponding optical fibers, i.e., those leaving the transmitters, in all the other networks.

Abstract: A granulometry device embodying the invention comprises a measuring probe and a micro-computer. The micro-computer provides digital filtering on a first discrete distribution function that is calculated by statistical inversion from a plurality of radiation measurements supplied by the probe and a theoretical radiation coefficient matrix thereby determining, after an iterative calculation, an optimum relation which should be satisfied by the nodes defining a discrete distribution function so as to offer a maximum correlation with uncertainty measurements. All the discrete distribution functions satisfying this relation are then calculated by statistical inversions. The mirco-computer provides an optimum statistical evaluation of the distribution of particles in a gas or liquid fluid.