Constrained optimization for color pattern recognition with a nonzero order joint transform correlator

Abstract

Abstract In this article, we apply a novel method on color pattern recognition with the nonzero order joint transform correlator. By the special arrangement of separated RGB channels of objects and reference images at the input plane, we can predict the location of correlation outputs. Furthermore, by the use of power‐spectrum subtraction strategy, the zero‐order part can be removed. The test pattern is a frog. The desired correlation peak is found to be quite distinctive and sharp. © 2002 Wiley Periodicals, Inc. Microwave Opt Technol Lett 33: 385–388, 2002; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mop.10330