Multi-Disciplinary Techniques for Understanding Time-Varying Space-Based Imagery.
Abstract
This project is intended to combine: pattern recognition, image understanding and artificial intelligence techniques for space-based image processing as well as: optical and digital processing methods. Optical feature extraction and sub-pixel target detection and tracking results are summarized. Scene representation and modeling work using: probabilistic graph matching, multiple resolution rotation-invariant operators and texture analysis are detailed. Image understanding techniques for 3D scene interpretation discussed include 2D image-level methods (using features such as edges, lines and corners) and 3D scene-level methods. New dynamic programming, stereo image and model building results are included.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 10, 1985
- Accession Number
- ADA185286
Entities
People
- Arthur Sanderson
- David P. Casasent
- Takeo Kanade
Organizations
- Carnegie Mellon University