Abstract A high-resolution video-based technique for obtaining two-dimensional fluid acceleration field data has been developed. The algorithm uses a combination of cross-correlations and autocorrelations on doubly exposed images of particle-seeded flows. Autocorrelations of individual video frames in an image pair yield two instantaneous velocity fields from which accelerations can be computed. Cross-correlations between successive images in the pair are used to resolve directional ambiguity associated with the autocorrelation. Time intervals are made arbitrarily small through the use of a laser sheet generator circuit which is synchronized with the framing rate of the camera. The technique is validated using a fluid-filled Petri dish subject to a known periodic motion. Ongoing development, uncertainties, and limitations of the technique are discussed.