This paper presents a method for multi target detection and localization based on heterogeneous uncertain information produced by multiple agents. Information association is based on graph analysis which considers georeferences, spatial precision and pre-existing knowledge. The objective of the scoring fusion is to suggest as quickly and as precisely as possible the hypothetic positions of trapped persons by increasing the quality of uncertain information. The overall aim is to ameliorate the search efficiency by increasing the detection capabilities while reducing risks, false alarms and oversight.