In the paper is presented a localization system composed of video cameras and infrared detectors. First, a localization is done by using the information obtained from video cameras, using a proper set of attributes. An uncertainty measurement is performed based on the similarity with the nearest prototype and the dissimilarity from the rest of prototypes. The best granularity is also determined. The information delivered by the infrared detectors establish an estimation of the object's localization and the heating surface. These reinforce the video information, by using an information fusion.