We present a methodology for computational evaluation of camouflage effectiveness. The methodology is implemented in software at the Danish Defence Research Establishment (DDRE) under the acronym for camouflage evaluation (CAMEVA). Basic input is a single image comprising a highly resolved static target as well as a proper amount of representative background. Separate target and background images can also be handled. Target and background regions are manually selected using the computer’s standard pointing device, the mouse. From the input data, CAMEVA predicts the target detectability as a function of the target distance. The detectability estimate is based on statistical distributions of features extracted from the imagery, establishing a multidimensional feature space. In the feature space, the Bhattacharyaa distance measure is applied as an estimator of the separability between the target and background. The intention is that the extracted features should resemble those applied during the human perception process. Typically, contrast and various measures of edge strength are applied. The Bhattacharyaa distance establishes a relative separability, while the absolute detection range is obtained by deriving a relation between the Bhattacharyaa distance and the estimated target resolution, at range. Thus by introducing parameters of the sensor, typically the human unaided eye, detectability as a function of the range is obtained. The methodology does not reflect individual observer performance but is aimed at providing an estimate of the optimal detection performance, given the selected set of features. During the choice of features and of sensor parameters, perception mechanisms other than the human observer performance can be modeled with this methodology. We discuss theoretical and practical aspects of CAMEVA. Validation and application examples, including results on the Search–1 and Search–2 datasets, are presented together with other data.