Real-time depth-mapping three-dimension TV camera (Axi-Vision Camera)

Masahiro Kawakita, Keigo Iizuka, Yoshiki Iino, Hiroshi Kikuchi, Hideo Fujikake, Tahito Aida

Research output: Contribution to journalArticlepeer-review

Abstract

A three-dimension camera (Axi-Vision camera) has been developed that is capable of producing a color image of an object and performing high-speed mapping of depth information from the camera to the object. With this camera, intensity-modulated near-infrared light is irradiated onto the object, and distances to points on the object are computed from the image captured by a camera that has a high-speed shutter. Since a two-dimensional scanning mechanism for a laser beam or complex computational processing is unnecessary in this depth mapping system, the distance to the object can be detected at high speed for each pixel of a TV image. For the present paper, a three-dimension camera adapted to a standard television signal was constructed by development of a high-output light-emitting-diode (LED) array light source capable of high-speed intensity modulation, an image intensifier capable of high-speed shutter operation on the nanosecond order, and an optical system. The prototype camera can output a depth image with a resolution of 768 × 493 pixels at a frame rate of 15 Hz, the characteristic depth mapping resolution of the camera is 1.8 cm (when the distance from the camera to the object is 2 m), and the camera can capture an image of a person-sized object. It was also shown that the depth information thus obtained could be applied to live-action filming and to CG and other new types of image synthesis.

Original languageEnglish
Pages (from-to)77-89
Number of pages13
JournalSystems and Computers in Japan
Volume37
Issue number8
DOIs
Publication statusPublished - 2006 Jul

Keywords

  • Depth mapping
  • Image synthesis
  • Real time
  • Three-dimension camera

Fingerprint

Dive into the research topics of 'Real-time depth-mapping three-dimension TV camera (Axi-Vision Camera)'. Together they form a unique fingerprint.

Cite this