A self-made electronic nose consisting in a sensor array of six commercial tin oxide gas sensors is used to monitor the odour emission from a compost facility. Supervised data processing tools, such as discriminant analysis, are able to recognize, in real time, the odour of compost with respect to other possible sources in the hall. The paper shows that with unsupervised methods, such as principal component analysis, it is not essential to identify all the possible odour sources during the learning phase. The closeness to the compost group centroid could be used as an indicator of the compost odour level. Alternatively, by a suitable calibration from olfactory measurements, the signals generated by the sensor array can be used to estimate the odour emission rate from the compost hall. Such real time monitoring should allow to assess and to anticipate the annoyance in the surrounding.
展开▼
