Projektvoorstel Active Vision (Project Proposal Active Vision).
Abstract
The goal of computer vision is to construct automatically by analysing 2-dimensional images a symbolic interpretation that describes the environment from which the images are captured. Via numerous intermediate descriptions at a symbolic level, the vision system finally generates a semantic world model that provides the interpretation of the scene. This interpretation can serve as an input for a higher level process as path planning or collision avoidance. In this respect computer vision can be considered as a feature extractor process, like is the case for identification and classification systems. Though the feature space, which is in the computer vision context denoted as world model, is very different from nature for all these systems, they have the same need to identify objects. This implies that vision systems must have access to a knowledge base containing descriptions of objects or generalized object classes. Many concepts in the field of computer vision are generic within the context of autonomous systems and can be transformed to other applications: the development of a powerful computer vision system depends on the successful integration of sensor, sensor preprocessing, signal analysis and artificial intelligence techniques. Therefore the main challenge in this project is to bridge the analytic world of signal processing and the symbolic world of artificial intelligence.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1992
- Accession Number
- ADA256490
Entities
People
- P. F. Krekel
Organizations
- Netherlands Organisation for Applied Scientific Research