Lionel Thomas

Optimization of the volume reconstruction for classical Tomo-PIV algorithms (MART, BIMART and SMART): synthetic and experimental studies

Optimization of multiplicative algebraic reconstruction technique (MART), simultaneous MART and block iterative MART reconstruction techniques was carried out on synthetic and experimental data. Different criteria were defined to improve the preprocessing of the initial images. Knowledge of how each reconstruction parameter influences the quality of particle volume reconstruction and computing time is the key in Tomo-PIV. These criteria were applied to a real case, a jet in cross flow, and were validated.

Particle volume reconstruction based on a marked point process and application to TOMO-PIV

In this paper, we propose a new tomographic reconstruction method, called IOD-PVRMPP, to reconstruct 3D particle volumes from 2D particle images provided by the Tomographic Particle Image Ve-locimetry (Tomo-PIV) technique. Our method, based on marked point processes (or object processes), allows to solve the problem in a parsimonious way. It facilitates the introduction of prior knowledge and solves memory problem which is inherent to voxel based approaches used by classical tomographic reconstruction methods. The reconstruction of a 3D particle set is obtained by minimizing an energy function which defines the marked point process. To this aim, we use a simulated annealing algorithm based on Reversible Jump Markov Chain Monte Carlo (RJMCMC) method. To speed up the convergence of the simulated annealing, we develop an initialization method which provides the initial distribution of 3D particles. To do that, we proceed by detecting 2D particles located in projection images. Using synthetic data, we show that IOD-PVRMPP method gives better results than MinLOS-MART method for different seeding densities.

3D particle volume tomographic reconstruction based on marked point process: Application to Tomo-PIV in fluid mechanics

In recent years, marked point processes have received a great deal of attention. They were applied with success to extract objects in large data sets as those obtained in remote sensing frameworks or biological studies. We propose in this paper a method based on marked point processes to reconstruct volumes of 3D particles from images of 2D particles provided by the Tomographic Particle Image Velocimetry (Tomo-PIV) technique. Unlike other reconstruction methods, our approach allows us to solve the problem in a parsimonious way. It facilitates the introduction of prior knowledge and naturally solves the memory problem which is inherent to pixel based approach used by classical tomographic reconstruction methods. The best reconstruction is found by minimizing an energy function which defines the marked point process. In order to avoid local minima, we use a simulated annealing algorithm. Results are presented on simulated data.

Tomographic Reconstruction of 3D Objects Using Marked Point Process Framework

For reconstructing sparse volumes of 3D objects from projection images taken from different viewing directions, several volumetric reconstruction techniques are available. Most popular volume reconstruction methods are algebraic algorithms (e.g. the multiplicative algebraic reconstruction technique, MART). These methods which belong to voxel-oriented class allow volume to be reconstructed by computing each voxel intensity. A new class of tomographic reconstruction methods, called “object-oriented” approach, has recently emerged and was used in the Tomographic Particle Image Velocimetry technique (Tomo-PIV). In this paper, we propose an object-oriented approach, called Iterative Object Detection—Object Volume Reconstruction based on Marked Point Process (IOD-OVRMPP), to reconstruct the volume of 3D objects from projection images of 2D objects. Our approach allows the problem to be solved in a parsimonious way by minimizing an energy function based on a least squares criterion. Each object belonging to 2D or 3D space is identified by its continuous position and a set of features (marks). In order to optimize the population of objects, we use a simulated annealing algorithm which provides a “Maximum A Posteriori” estimation. To test our approach, we apply it to the field of Tomo-PIV where the volume reconstruction process is one of the most important steps in the analysis of volumetric flow. Finally, using synthetic data, we show that the proposed approach is able to reconstruct densely seeded flows.

Volume reconstruction optimization for tomo-PIV algorithms applied to experimental data

Tomographic PIV is a three-component volumetric velocity measurement technique based on the tomographic reconstruction of a particle distribution imaged by multiple camera views. In essence, the performance and accuracy of this technique is highly dependent on the parametric adjustment and the reconstruction algorithm used. Although synthetic data have been widely employed to optimize experiments, the resulting reconstructed volumes might not have optimal quality. The purpose of the present study is to offer quality indicators that can be applied to data samples in order to improve the quality of velocity results obtained by the tomo-PIV technique. The methodology proposed can potentially lead to significantly reduction in the time required to optimize a tomo-PIV reconstruction, also leading to better quality velocity results. Tomo-PIV data provided by a six-camera turbulent boundary-layer experiment were used to optimize the reconstruction algorithms according to this methodology. Velocity statistics measurements obtained by optimized BIMART, SMART and MART algorithms were compared with hot-wire anemometer data and velocity measurement uncertainties were computed. Results indicated that BIMART and SMART algorithms produced reconstructed volumes with equivalent quality as the standard MART with the benefit of reduced computational time.