Light Field Reconstruction using a Generic Imaging Model
David Uhlig, Michael Heizmann · 2020
Light field cameras play an increasingly important role in computer vision and optical metrology. However, due to their complex design, their calibration is very difficult and usually precisely adapted to the respective light field camera type. We present a method that extracts a light field from an arbitrary light field imaging system without knowing and without modelling the internal optical elements. We calibrate the camera using a generic calibration procedure, transform the obtained set of rays into an equivalent light field representation and finally, reconstruct a rectified light field from the irregularly sampled data. Experimental results validate the method and demonstrate that the geometrical structure of the light field is preserved by an adequate rectification.