Automatic Recognition and Tracking of Objects.
Abstract
The broad program of research consisted of multisensor image understanding including integration of information from multiple sources, the tracking of objects in a sequence of images in real time, together with the estimation of motion parameters, the characterization of the descriptions and invariances of objects, and the registration of objects. Intensity, color, and range were used in the segmentation of scenes, and the extraction and identification of general areas of interest in the scenes. Shape descriptor provide structural descriptions of objects and lead to recognition of objects and background components. Motion analysis is instrumental in tracking and prediction of the movement of objects. The analysis and understanding of the structure and motion of three-dimensional space from from a sequence of two-dimensional images in real time is the fundamental goal of the present investigation.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1981
- Accession Number
- ADA121678
Entities
People
- J. K. Aggarwal
Organizations
- University of Texas at Austin