Model-independent light field reconstruction using a generic camera calibration

Abstract Sophisticated and highly specialized optical measuring devices are becoming increasingly important for high-precision manufacturing environment perception. In particular, light field cameras experiencing an ever-increasing interest in research industry as they enable a variety of new measurement methods. Unfortunately, due to their complex structure, calibration is very difficult usually precisely tailored the particular type camera. To overcome these difficulties, we present method that decodes from raw data any imaging system without knowing modeling internal elements. We calibrate camera using precise generic transform obtained ray set into equivalent representation. Finally, reconstruct rectified irregularly sampled addition derive geometric properties intrinsic parameters. Experimental results validate by showing both information observed scene structure preserved adequate rectification calibration.