Makrofol detectors are commonly used for long-term radon (Rn-222) measurements in houses, schools and workplaces. The use of this type of passive detectors for the determination of radon concentrations requires the counting of the nuclear tracks produced by alpha particles on the detecting material. The "image reduction and analysis facility" (IRAF) software package is a piece of software commonly used in astronomical applications. It allows detailed counting and mapping of sky sections where stars are grouped very closely, even forming clusters. In order to count the nuclear tracks in our Makrofol radon detectors, we have developed an inter-disciplinary application that takes advantage of the similitude that exist between counting stars in a dark sky and tracks in a track-etch detector. Thus, a low cost semi-automatic system has been set up in our laboratory which utilises a commercially available desktop scanner and the IRAF software package. A detailed description of the proposed semi-automatic method and its performance, in comparison to ocular counting, is described in detail here. In addition, the calibration factor for this procedure, 2.97 +/- 0.07 kBq m(-3) h track(-1) cm(2), has been calculated based on the results obtained from exposing 46 detectors to certified radon concentrations. Furthermore, the results of a preliminary radon survey carried out in 62 schools in Tenerife island (Spain), using Makrofol detectors, counted with the mentioned procedure, are briefly presented. The results reported here indicate that the developed procedure permits a fast, accurate and unbiased determination of the radon tracks in a large number of detectors. The measurements carried out in the schools showed that the radon concentrations in at least 12 schools were above 200 Bq m(-3) and, in two of them, above 400Bq m(-3). Further studies should be performed at those schools following the European Union recommendations about radon concentrations in buildings. (c) 2007 Elsevier Ltd. All rights reserved.