Automatic Pattern Recognition in a Multi-Sensor Environment
Abstract
During this period a number of investigations were undertaken directed at the general goals of improving the accuracy, reliability and speed of visual recognition systems. Work was carried out on basically three different problems. The development of a multiple look sequential recognition system to improve accuracy and reliability, the development of an image array processor for application to real time image identification, and the analysis of various features and authority file generation algorithms to identify properties which may reduce the training period. In addition, an investigation was also carried out on the application of 'fuzzy' automata to the scene segmentation problem. Finally, an application of the work carried out here was developed in the use of visual feedback for work piece positioning and control.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 25, 1977
- Accession Number
- ADA037968
Entities
People
- Kenneth J. Breeding
Organizations
- Ohio State University