Analysis of an ambient environment by means of perception is described. The surveillance of an object by human, who watches a scene via a monitor that shows camera sensed information, is investigated. Although the camera sensing process is a deterministic process, human perception of a scene via camera sensed information is a probabilistic process, i.e. the observer may overlook information about the scene, while this may be unlikely. The human surveillance based on camera sensed information is modelled in the present work by means of probabilistic computations. Thereafter conversion of the probability into possibility is carried out, quantifying the perception possibility attached to an object, thereby providing an assessment of the surveillance. Thereafter the suitability among alternative camera positions as to surveillance of an ambient environment is subjected to precise comparison. The method is described and applied in two computer experiments, measuring the surveillance of an ambient environment.