Radiometric compensation techniques based on projector-camera system

Feng Chen, Yue Liu

Research output: Contribution to journalReview articlepeer-review

Abstract

With the increasing capabilities of the projectors and the development of computer graphics, projection techniques have been widespread. More and more everyday surfaces could be used for projection. Using real-time correction, images projected onto textured surfaces with arbitrary geometry can be as good as those projected on the optimized canvases. In this paper, the radiometric compensation techniques based on projector-camera systems developed in recent years are reviewed. Compensations based on color mixing matrix use only a few calibrating patterns for recovering the light properties of the system. For applications with simple light effects, the surfaces could be compensated efficiently. Using fixed co-axial alignment or embedded structure light would further reduce the calibration time and make real-time mobile projection systems practical. Researches on adaptive compensation make the best use of light projected by the system so as to extend the dynamic range while keeping the imagery compensated accurately. Light transport matrix based algorithm takes global illuminations into account, compensating images are calculated in real-time after solving the inverse light-transport out using constrained optimization. Tiled projections are used to extend the dynamic range and increase the resolution, which makes projections onto radiometrically complex surfaces possible.

Original languageEnglish
Pages (from-to)992-1004
Number of pages13
JournalJisuanji Fuzhu Sheji Yu Tuxingxue Xuebao/Journal of Computer-Aided Design and Computer Graphics
Volume27
Issue number6
Publication statusPublished - 1 Jun 2015

Keywords

  • Color mixing property
  • Light transport matrix
  • Projector-camera system
  • Radiometric compensation

Fingerprint

Dive into the research topics of 'Radiometric compensation techniques based on projector-camera system'. Together they form a unique fingerprint.

Cite this