A Fast and Sparsity-Aware Generalization of SMART for Tomographic Particle Image Velocimetry

Ioana Barbu (1) , Cédric Herzet (1)

1 FLUMINANCE - Fluid Flow Analysis, Description and Control from Image Sequences

Understanding the tridimensional (3D) motion of turbulent flows is one of the most challenging problems in fluid dynamics. Because the Navier-Stokes equations, describing the spatio-temporal evolution of a fluid, are intractable both at a theoretical and numerical level when dealing with turbulent flows, the idea has soon been proposed to resort to experimental measures. In this line of search, the most common approach is the so-called " Particle Image Velocimetry" (PIV) [1] which consists in seeding the fluid with passive particles 1 and accessing to motion measures by processing images of the spatio-temporal evolution of the latter. Recently, researchers have moved their attention to the tridimensional setup, where the 3D motion of a fluid must be inferred from the observation of a set of images captured at each time instant. The most-advanced experimental scheme addressing this problem is the so-called " Tomographic PIV " system, introduced by Elsinga et al. in [2], [3]. The work presented in this paper takes place in this particular applicative context. We focus hereafter on an intermediate but important step arising in Tomographic PIV: the estimation of the 3D position of the particles from the set of collected images, i.e., the so-called " Volume reconstruction" problem [4]. Formally, the task consists in inverting a model of the form y = Dx + noise, (1)

Domaines: Traitement du signal et de l'image [eess.SP]

Ioana Barbu, Cédric Herzet. A Fast and Sparsity-Aware Generalization of SMART for Tomographic Particle Image Velocimetry. Signal Processing with Adaptive Sparse Structured Representations, Jul 2015, Cambridge, United Kingdom. ⟨hal-01245022v1⟩