Three-dimensional optical tomography techniques have been developed to reconstruct an object using its set of two-dimensional projection images. The multiplicative algebraic reconstruction technique (MART) with several formulations combining with five basis functions such as: cubic B-spline, o-Moms, keys, cosine and Gaussian for building projection matrix have been used. Investigations on the performance of each formula of the MART, effect of relaxation parameter and contribution of each basis function have been completed using numerical simulation with several different phantoms. The multiply line-of-sight (MLOS) estimation for initial object intensities and the simultaneous multiplicative algebraic reconstruction technique (SMART) for updating these intensities have been also examined in comparing with the MART which can apply on the three-dimensional tomographic particle image velocimetry (Tomo-PIV).
The numerical simulation results indicated that the MART with cubic cosine basis function is suitable to reconstruct a meaning shape of object while the MART with Gaussian basis function is appropriate for a quasi-spherical shape. This result has been applied to analyze behaviors of electrohydrydynamic jetting by reconstruction of three-dimensional view of several jetting modes from their shadowgraphic images as projections. The three cameras were used to capture the shadowgraphic images of cone-jetting mode, asymmetric-jetting mode and the multiple-jetting mode. These images’ intensities were converted to archive the projection data for tomographic reconstruction. The full view of jet using this tomographic technique could identify the jet shape and position with respect to nozzle position.
The numerical result also pointed out that the MLOS-SMART is a suitable method to prepare the volume particle contributing to the cross-correlation to obtain the three-dimensional velocity field in three-dimensional PIV due to its ability of reducing computational time. Basing on this result, a flow structure or a velocity field inside a cone shape of liquid formed under electric field has been visualized and estimated. Three cameras were also used to capture the images of illuminated particles moving inside the cone shape. These images were used to reconstruct the particles inside the cone by employing the MLOS-SMART method. This step created two volumes of particles which defines the particle intensities and their position for the cross-correlation to obtain the velocity vectors. The obtained velocity field was reasonable fit with the movie of motions of particles.