Fourier Pattern Recognition Techniques Applied to Titanium Alloy Photomicrographs.
Abstract
The properties of the Fourier Transforms of Titanium alloy photomicrographs are investigated. The characteristics of the discrete Fourier Transforms are outlined, and the principles of pattern recognition by spatial filtering and matched filter correlation are discussed. The technique of data preparation is described, and the results from both spatial filtering and correlation methods are presented and discussed. The conclusion is reached that texture patterns such as metal microstructure are too noisy for successful pattern recognition by spatial filtering, but matched filter correlation techniques show some promise for accurate measurement of defined object populations in noisy samples. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1972
- Accession Number
- AD0746720
Entities
People
- Graham L. Perry
Organizations
- Air Force Institute of Technology