Three-Dimensional Motion Estimation Using Shading Information in Multiple Frames

Abstract

Traditionally motion and shading have been treated as two disjoint problems. On the one hand, researchers studying motion or structure from motion often assume uniform lighting conditions over the whole surface and good contrast at high spatial frequencies to minimize the effects of variations of the image irradiance of the patch as the surface moves. On the other hand, researchers primarily concerned with the shape from shading problem only consider static brightness data in order to recover the shape without considering the change of brightness induced by motion. A new formulation for recovering the structure and motion parameters of a moving patch is presented in this thesis. It is based on using the spatiotemporal derivatives of irradiance that are computed from a time-varying irradiance sequence and combined into a differential constraint equation. The new approach determines the rigid body motion and the structure of the patch directly from the irradiance sequence using both motion and shading information. A new constraint equation, the full irradiance constraint equation (FICE), is derived. It links the spatiotemporal gradients of irradiance to the motion and structure parameters and the temporal variations of the surface shading. This equation separates the contribution to the irradiance spatiotemporal gradients of the gradients due to texture from those due to shading and allows the FICE to be used for textured and textureless surface.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 1989
Accession Number
ADA216804

Entities

People

  • Jean-pierre Schott

Organizations

  • Massachusetts Institute of Technology

Tags

Communities of Interest

  • Air Platforms

DTIC Thesaurus Topics

  • Algorithms
  • Artificial Intelligence
  • Cameras
  • Computational Science
  • Computer Graphics
  • Computer Vision
  • Coordinate Systems
  • Equations
  • Equations Of Motion
  • Geometric Forms
  • Geometry
  • Image Processing
  • Kalman Filtering
  • Lines (Geometry)
  • Motion Planning
  • Three Dimensional
  • Two Dimensional

Readers

  • Computer Vision.
  • Fluid Dynamics.