Hyperspectral imaging differs from conventional techniques by exploiting the spectral dimensionality of remote scenes. This additional information promotes discrimination of image elements, especially anomalies that are dissimilar with respect to global features. Algorithms for anomaly detection are designed to overcome the inherent difficulty of analysing hypercubes, which are the higher-dimensional analogues of conventional broadband images. Such algorithms are prolific in their variety and design, which could become an obstacle in choice or application for the neophyte researcher in this field. This paper seeks to consolidate this plethora of algorithms into succinct categories for clarity of rudimentary decision making. A duplicate of article 012048 Snapshot hyperspectral imaging and practical applications was originally published here, in error, as article 012051. The present article replaced the duplicate and was published on 18 August 2009. © 2009 IOP Publishing Ltd.