We propose three correlation-based methods to simultaneously detect the centroids of multiple objects in an input scene. The first method is based on the modulus of the moment function, the second method is based on squaring the moment function, and the third method works with a single intensity filter. These methods are invariant to changes in the position, orientation, and scale of the object and result in good noise-smoothing performance. We use spatial light modulators (SLMs) to directly implement the input of the image and filter information for the purpose of these approaches. We present results showing simulations from different approaches and provide comparisons between optical-correlation- and digital-moment-based methods. Experimental results corresponding to an optical correlator using SLMs for the centroid detection are also presented.