A digital watermark is a visible, or preferably invisible, identification code that is permanently embedded in some digital data to prove owner authentication and provide protection of that document. In this paper we utilize a watermark generation technique based on the use of chaotic functions and the motivation for using these functions is presented. The technique used for watermark embedding is also described, together with a watermark detection scheme based on an optical Matched Filter correlator. We provide results of optical simulations of the watermark detection scheme and show that correlation-based detection is an excellent method for detecting chaotically-generated watermarks embedded in the Fourier domain using multiplicative embedding. We also show that it is possible to detect chaotically-generated watermarks in images that have been subjected to noise.