Interactions of conduit geometry with gas-liquid flows control volcanic activity, implying that the evaluation of volcanic hazards requires quantitative understanding of the inner structure of the volcano. The more established geophysical imaging techniques suffer from inherent ambiguity, may require spatially dense measurements in active areas and may not provide sufficient spatial resolution in the uppermost part of the conduit system. It is thus desirable to develop new imaging techniques allowing a better spatial resolution of a volcano's upper feeding system, with reduced ambiguity and a low level of risk for operators. Muon particles can be utilized to image the internal density distribution of volcanic structures. The principle of muon radiography is essentially the same as X-ray radiography, except for substituting penetrating particles in place of photons.Muons are more attenuated by higher density parts inside the target and thus information about its inner structure are obtained from the differential muon absorption. We report on a muon-imaging experiment that was conducted at Mt Etna in 2010. The target structure was one of the summit craters of the volcano. This experiment was performed using a muon telescope suitably designed to withstand the harsh conditions in the summit zone of a high volcano. We found a marked difference between synthetic and observed attenuation of muons through the target. This discrepancy is likely due to the bias on the observed flux, arising from false muon tracks. They are caused by low-energy particles that, by chance, hit simultaneously the two matrixes of the telescope, leading to detection of a false positive. We separated the useful from the unwanted signal through a first-order model of the background noise. The resulting signal is compared with the corresponding synthetic flux. Eventually, we found regions of higher- and lower-than-expected muon flux, that are possibly related to inner features of the target crater.