Image processing algorithms for a light-field camera and their application for optical flow diagnostics

Application of modern optoelectronic devices extends research in the field of experimental fluid mechanics. The methods of computational photography gradually penetrate into the various fields of science and technology due to using devices based on these methods. A light-field camera can be used to register a three-dimensional velocity distribution in fluid flows where the location of several panoramic optical sensors is difficult because of restrictions in an optical access and vibrations. in this paper we study the possibility of using a plenoptic system consisting of an industrial light-field camera to diagnose liquid and gas flows. A new software algorithm for computing a depth field of a measurement area is proposed. According to the obtained results, the spatial resolution of the method by depth reaches 1/40th of the depth of the field of the optical system when using 11 MP sensor. This method was used to measure 3D velocity fields of a turbulent jet inside a slot channel throughout its depth. In the future, the number of applications will grow for the cases where the use of plenoptic devices with high spatial resolution is appropriate.