Hierarchical multisensor image understanding
Abstract
This report describes the research results on Honeywell's Hierarchical Multisensor Image Understanding program. Honeywell is developing a unified framework for the different hierarchical levels of image processing such as segmentation, detection, classification, and identification of outdoor scenes and across different sensor modalities such as millimeter wave, infrared, and visible. Current activities on the project are reviewed under the following headings: (1) artificial-intelligence-based generic image segmentation and object recognition; (2) evidence-confidence paradigms for image understanding; (3) hierarchical systems theory for control structures; and (4) invariant methods in image understanding. Also discussed are scene analysis and attributed graphs.
- Publication:
-
NASA STI/Recon Technical Report N
- Pub Date:
- July 1984
- Bibcode:
- 1984STIN...8434688A
- Keywords:
-
- Hierarchies;
- Image Processing;
- Artificial Intelligence;
- Classifications;
- Confidence Limits;
- Detection;
- Graph Theory;
- Identifying;
- Image Analysis;
- Images;
- Information Theory;
- Infrared Spectra;
- Millimeter Waves;
- Pattern Recognition;
- Scene Analysis;
- Segments;
- Target Recognition;
- Visible Spectrum;
- Electronics and Electrical Engineering