This article presents a method that enables the generation rate from one or /more particle sources to be estimated, using far-field concentration measurements. The method is made up of two distinct steps; a calibration phase, followed by an estimation phase. The calibration phase makes it possible to create a transfer relationship between a known source ("reference source") and the measurement of the far-field concentration. The second step consists of estimating unknown source generation rates by inverting the transfer relationship and using measurements of far-field concentrations resulting from these unknown sources. In addition, this article presents a technique to improve the positioning of the sensors in the room in which the sources are situated. A numerical study using computational fluid dynamics was first conducted to theoretically validate the estimation method and assist with choosing the sensor positions in the experimental rig. The study established that, with ideal sensors, the difference between the real and estimated generation rates can be accurate to within 0.1%. The method was then deployed on an experimental case. The results confirmed that it is possible to estimate an isolated source. However, the quality of the estimation deteriorated when the source to be estimated was significantly different from the reference source, from an aerodynamic perspective.