188宝金博页面版

Real-time holographic video images with commodity PC hardwareV. Michael Bove, Jr. *a , Wendy J. Plesniak a,b , Tyeler Quentmeyer a , James Barabas aa MIT Media Laboratory, 20 Ames St., Room E15-368B, Cambridge MA 02139 USAb Harvard Center for Neurodegeneration and Repair, 1249 Boylston St., 2nd Floor Room 425,Boston MA 02115 USAABSTRACTThe MIT second-generation holographic video system is a real-time electro-holographic display. The system produces asingle-color horizontal parallax only (HPO) holographic image. To reconstruct a three-dimensional image, the displayuses a computed fringe pattern with an effective resolution of 256K samples wide by 144 lines high by 8 bits persample. In this paper we first describe the implementation of a new computational subsystem for the display, replacingcustom computing hardware with commodity PC graphics chips, and using OpenGL. We also report the implementationof stereogram computing techniques that employ the PC hardware acceleration to generate and update holographicimages at rates of up to two frames per second. These innovations shrink the system’s physical footprint to fit on thetable-top and mark the fastest rate at which full computation and update have been achieved on this system to date.Finally we present first results of implementing the Reconfigurable Image Projection (RIP) method of computing high-quality holograms on this new system.Keywords: holographic video, autostereoscopic displays, synthetic holography1. INTRODUCTIONThe MIT second-generation (“Mark II”) holographic video system 1 (Fig. 1) is a real-time display system that diffractslight by means of eighteen parallel cross-fired shear mode TeO 2 acousto-optic modulators (AOMs) and passes the resultto a chain of optics and scanning mirrors to produce a monochromatic horizontal-parallax-only (HPO) image volume150mm wide, 75mm high, and 160mm deep, visible over a range of 30 degrees, refreshed 30 times per second, with avertical resolution of 144 lines. As the optical design of this system has been described at length elsewhere, in this paperwe will concentrate only on the aspects of the system that set requirements for the video signals driving it.The Mark II system is optically very similar to the Scophony 2 television display of the 1930s, differing mostly in the useof multiple parallel AOMs in place of the latter’s single AOM; the signals differ in that the Scophony system (whichwas displaying a single 2D raster) amplitude-modulated a fixed-frequency fringe pattern with the intensity of the videoimage, while in Mark II both the amplitude and instantaneous phase of the fringes are varied. Also, Mark II, through itsuse of cross-fired AOMs, is able to use both the forward and retrace horizontal scans for active video, a techniquesometimes called a boustrophedonic scanning pattern. This latter feature will have implications for the generation ofvideo signals, as will be discussed in a later section of this paper.Since its construction in the 1990s, computation for the Mark II system has been performed by a combination of an SGIOnyx workstation and a Cheops Imaging System, 3 a compact, block data-flow computing system optimized for real-time parallel computations on streams of data. Because the Cheops system is capable of driving up to six genlockedvideo output cards whose video timing parameters can be freely configured to meet unusual requirements, because it iseasily interfaced to a host workstation by SCSI (and HIPPI, if higher speed is required), and because it contains (for itstime, and to some degree even now) a large amount of random-access memory, it was a good match to the needs of theMark II display. The modularity of Cheops also permitted the development and installation of specialized basis-functionsuperposition processors (“Splotch Engines”) which were optimized for the computational needs of the holo-stereogramalgorithm, below. 4 * vmb@media.mit.edu, http://www.media.mit.edu/~vmbZhejiang University of Media and Communications (218.75.124.130) - 2014/10/8 Download