Prospecting in areas characterized by highly fractured reflectors is afflicted by a low seismic response. In these cases, a reliable earth's model for seismic imaging can hardly be defined. The Common Reflection Surface (CRS) theory supplies a proper framework to approximate, without the knowledge of the medium velocity model, the time of flight of all reflected events moving close to a normal incident ray. The CRS kinematic description of each normal ray is specified by eight geometric attributes. They can be computed by means of a data-driven optimization method which is penalized by the unavailability of any gradient. In areas of poor seismic response, the knowledge of these attributes is essential for high resolution time imaging. The Imaging and Numerical Geophysics group was the first one to develop a 3D CRS software for seismic exploration projects. This application immediately became strategic for industrial processing and is often cited as an example of technological innovation. Recently, the ING team devised a new strategy for the simultaneous estimate of the eight CRS parameters, solving a global non-linear optimization problem by combining a conjugate-direction method with an effective line-search scheme. The resulting algorithm has remarkable convergence properties even in the presence of multimodal objective functions. The solution to this mathematical problem has led to a great enhancement of the final image resolution and represents a crucial step for the industrial use of CRS analysis tools.
The cookie settings on this website are set to "allow cookies" to give you the best browsing experience possible. If you continue to use this website without changing your cookie settings or you click "Accept" below then you are consenting to this.