LISA research project Several projects are developped for the purpose of direct detection of gravitational waves (distorsions of spacetime predicted by Einstein's general relativity). Just as in the famous experience of Michelson, the method is laser telemetry. It relies on measuring by interferometry the changing separations (caused by the passage of gravitational waves) between test masses set in different directions. There are several long ground based interferometers (VIRGO, LIGO, GEO600, TAMA) in activity or under construction in the world. The goal of LISA (Laser Interferometer Space Antenna), project ESA/NASA is to avoid some sources of noise and to have long optical paths by being in orbit around the sun. LISA will open a frequency window between 10-4 and 10-1 Hz in which it is expected to observe the gravitational radiations produced by a wide variety of astrophysics phenomena. One of them is a large range of coalescence of white dwarf binaries [WDB] (final stage of stellar evolution) in our galaxy. Each of those coalescences emitts a gravitational radiation [1] wich can be described as an oscillating signal (chirp). The frequency of such a signal grows up as a non-integer power of time. The [WDB] are so numerous that they cannot be resolved at frequencies below 10-3 Hz, leading to a confusion-limited background. Another interesting source will be provided by extragalactic merging massive black holes (105 times the solar mass). Their emission is also a chirp with the same law of variation but in a different frequency range. One problem for the LISA's data analysis is the detection and separation of the two types of sources. We propose a wavelet method to solve this problem. Such a method has been proving to be efficient in the chirp characterization ([2] et [3]). Our goal is to construct different detection algorithms, to test it numerically with a simple model by Monte Carlo type simulations. [1] K. S. Thorne, Gravitational radiation in _300 Years of Gravitation_, Èds. S. W. Hawking et W. Israel, pp. 330--458, 1987. [2] Y. Meyer, _Oscillating Patterns in Image Processing and Nonlinear Evolution_, University Lecture Series, AMS, 2001. [3] Ph.Tchamitchian et B.Torresani, Ridge and skeleton extraction from the wavelet transform, Wavelets and their applications, ed. par M.Ruskai, Jones et Bartlett, Publ.Boston (1992) Participants: P. Flandrin, C. Melot, A. Fraysse, M. Ben Slimane.