Abstract
A motorized rotating retarder, a linear polarizer, and a digital camcorder are used to obtain polarization profiles of natural daylight scenes. A correlation technique synchronizes real-time video frames with respect to the rotating retarder's fast axis. Polarization parameters are calculated for each red-green-blue (RGB) channel and scene pixel. Extending the quasimonochromatic approximation to larger bandwidths is solved by expressing the Stokes parameters explicitly as functions of wavelength. A detailed numerical analysis examines errors due to ill-conditioning effects that are associated with specific relative angular orientations of the optical components. A comparison of the theoretical results with empirical data from the digital camcorder provides a validation process for the results in this paper. A RGB pseudocolor encoding algorithm provides a means to visualize the polarization imagery. An ellipticity study was conducted during the early morning hours when scene lighting conditions change rapidly. A comparison of the results from a digital still camera, using a 30-s data acquisition time period, was made with respect to a 0.6-s time period for the rotating retarder configuration. The results clearly show that errors due to temporal registration effects are dramatically reduced as the time interval between images goes to zero.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call