Image processing methods used to identify and enhance dust devil features imaged by IMP (Imager for Mars Pathfinder) are reviewed. Spectral differences, visible red minus visible blue, were used for initial dust devil searches, driven by the observation that Martian dust has high red and low blue reflectance. The Martian sky proved to be more heavily dust-laden than pre-Pathfinder predictions, based on analysis of images from the Hubble Space Telescope. As a result, these initial spectral difference methods failed to contrast dust devils with background dust haze. Imager artifacts (dust motes on the camera lens, flat-field effects caused by imperfections in the CCD, and projection onto a flat sensor plane by a convex lens) further impeded the ability to resolve subtle dust devil features. Consequently, reference images containing sky with a minimal horizon were first subtracted from each spectral filter image to remove camera artifacts and reduce the background dust haze signal. Once the sky-flat preprocessing step was completed, the red-minus-blue spectral difference scheme was attempted again. Dust devils then were successfully identified as bright plumes. False-color ratios using calibrated IMP images were found useful for visualizing dust plumes, verifying initial discoveries as vortex-like features. Enhancement of monochromatic (especially blue filter) images revealed dust devils as silhouettes against brighter background sky. Experiments with principal components transformation identified dust devils in raw, uncalibrated IMP images and further showed relative movement of dust devils across the Martian surface. A variety of methods therefore served qualitative and quantitative goals for dust plume identification and analysis in an environment where such features are obscure.