Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group

Distortion invariant pattern recognition using decompressed nonlinear correlation

Not Accessible

Your library or personal account may give you access

Abstract

We have shown that using compression types of nonlinearity at the Fourier plane of joint transform correlators results in higher correlation peak intensity, larger peak-to sidelobe ratio, narrower correlation width, and better discrimination sensitivity.1 For a Kth law nonlinearity (y = xk), the compression nonlinearity corresponds to K < 1. In this paper we investigate the performance of the nonlinear joint transform correlator using decompression types of nonlinearity (K > 1) at the Fourier plane. We show that decompression transformations of the joint power spectrum result in correlation signals that are less sensitive to distortion. The decompression nonlinear joint transform correlator is tested using a kth law device for rotation types of distortion. Computer simulations of the nonlinear correlator are provided using an image of a tank and a binary character.

© 1989 Optical Society of America

PDF Article
More Like This
Use of limited distortion invariant correlation filters for recognition of road signs

Masaki Taniguchi and Katsunori Matsuoka
P74 Conference on Lasers and Electro-Optics/Pacific Rim (CLEO/PR) 1995

Shift and projection invariant pattern recognition

D. Mendlovic, N. Konforti, and E. Marom
TUC2 OSA Annual Meeting (FIO) 1989

Sensitivity of a nonlinear joint-transform correlator

Bahram Javidi and Qing Tang
ThK2 OSA Annual Meeting (FIO) 1990

Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.