Motivations

A critical element in the performance of any signal or image processing method is the validity and the accuracy of the underlying scene and sensor of modelling - backgrounds, contrasts, fluctuations, speckle, propagation, disturbances and jammers, sensor calibration, etc... Modern adaptive processing methods and high speed processors can indeed take full advantage of any suitable physical description of the received signals. The accuracy of this physical description and its adequation to the signal processing technique becomes an essential challenge for sensor design, in the different fields of application - from medical to military, using radar, acoustic, optronic, seismic sensors, etc...

This physical analysis of sensors will emphasize the growing similarities or communalities between signal and image processing, for instance in SAR image classification, phase conjugation, or fusion techniques.

This workshop, to be held every two years, beginning in 1999, will explore this diverse area mixing physics, signal and image processing, and sensor design, for a better adequation of signal processing to the characteristics of the sensor, and to the actual situation and operational needs.

General Information

PSIP 99 will be held in Paris, in the Ecole Nationale Supérieure des Télécommunications (ENST), which is the top school in France for graduating telecommunications engineers. ENST is located in the southern part of Paris (46 rue Barrault, Paris 13ème), near Place d'Italie, the delightful Parc Montsouris, the picturesque Butte aux Cailles district, and Paris' Chinatown. Rue Mouffetard and the Quartier Latin are just a few minutes' walk from ENST. The nearest station is " Corvisart ", on metro line 6, and is accessible from Roissy Charles de Gaulle airport through RER B, and then metro line 6, or from Orly airport, through Orly-bus towards Denfert Rochereau, and then metro line 6.The weather in Paris, in January, is generally rather cold, with an average temperature around 0° Celsius. They may be some snow. However, a " bistrot " is never far away in Paris, where shelter and human warmth are always available..

Duration: 2 days

Periodicity: 2 years, from 99

Invited Conferences (2 on each topic) : 1h including questions

TOPICS

Modelling

• Description and modelling of physical phenomena : radiation, pressure, temperature, flow, smell, motion, propagation, etc.
• Scenes and interferences modelling for signal and image processing
• Modelling of sensors in their environments
• Sensor specification and evaluation
• Physical and stochastic models for sensor processing
• Noise modelling
• Operational needs and performance criteria

Processing algorithms

• Clutter, reverberation, background elimination
• Motion estimation, motion elimination
• Noise reduction
• Separation of sources
• Multidimensional processing (space-time, space-frequency, time-frequency, space- polarization)
• Stereo and array processing of complex scenes
• Classification and identification
• Data fusion

Smart-sensors

• Integrated processing
• Calibration and compensation, self-adaptation to deformations
• Sensor adaptation to its actual environment
• Adaptive mode management
• Smart skins