|Ph.D Student||Swirski Yohay|
|Subject||Three-Dimensional Reconstruction Using Natural Flickering|
|Department||Department of Electrical Engineering||Supervisor||Professor Yoav Schechner|
|Full Thesis text|
The underwater environment is challenging for computer vision tasks. Poor visibility, geometrical distortions and spatio-temporal varying illumination are some of the physical sources of these challenges. Spatio-temporal varying illumination is created by refraction of light through the wavy water surface. The resulting underwater illumination field forms a caustic network and is known as flicker. In past studies, flicker has often been considered to be an undesired effect, which degrades the quality of images. In contrast, this thesis shows that flicker can actually be useful for vision in the underwater domain. Specifically, it solves very simply, accurately, and densely the stereo correspondence problem, irrespective of the object's texture. The temporal radiance variations due to flicker are unique to each object point, thus disambiguating the correspondence, with very simple calculations. Theoretical limitations of the method are analyzed using a ray-tracing simulation. This process is further enhanced by compounding the spatial variability in the flicker field and a smoothness constraint. This is done using a variational formulation for multi-frame stereo.
Furthermore, this approach is generalized for a free-moving stereo camera rig. The 3D scene structure is illumination invariant. Thus, as a reference for motion estimation, we use projections of stereoscopic range maps, rather than object radiance. Consequently, each object point can be tracked and then filtered in time, yielding deflickered videos. Since objects are viewed from different distances as the stereo rig moves, scattering effects on the images are modulated. This modulation, the recovered camera poses, 3D structure and deflickered images yield inversion of scattering and recovery of the water attenuation coefficient. Thus, coupled difficult problems are solved in a single framework. The entire research is demonstrated by underwater in-situ field experiments and in a lab.