Directionally Weighted Wave Field Estimation Exploiting Prior Information on Source Direction
Natsuki Ueno, Shoichi Koyama, Hiroshi Saruwatari
Abstract
A wave field estimation method exploiting prior information on source direction is proposed. First, we formulate a wave field estimation problem as regularized least squares, where the norm of the wave field is used for a regularization term. The norm of the wave field is defined on the basis of the weighting function that reflects the prior information on the source direction. We derive the closed-form solution using theories on Hilbert spaces. Results of numerical experiments indicated that high estimation accuracy can be achieved by using the proposed method in comparison with other current methods that do not use any prior information.
Topics & Concepts
WeightingPrior informationRegularization (linguistics)MathematicsAlgorithmHilbert spaceField (mathematics)Mathematical optimizationWave equationComputer scienceMathematical analysisArtificial intelligenceAcousticsPhysicsPure mathematicsSeismic Imaging and Inversion TechniquesSparse and Compressive Sensing TechniquesImage and Signal Denoising Methods