Abstract

We have developed a technique for pattern recognition in large-format images based on the use of two-dimensional Fast Fourier Transforms (2-D FFTs). The technique was developed at the Image Processing Laboratory of the Jet Propulsion Laboratory. Unlike existing techniques, it fully utilizes the information content of the 2-D FFT. Past studies (e.g. Bunting and Fournier, 1980; Toldalagi and Lebow, 1982) dealt with information extraction from 1-D power spectra as derived by radially averaging 2-D FFTs. Our technique utilizes the 2-D FFT to extract pattern size both parallel and perpendicular to the major axis, pattern shape and pattern orientation.

© 1983 Optical Society of America