Characterisation of cough sounds to monitor respiratory infections in intensive pig farming

Cough is a symptom and central element for diagnosis of very common respiratory affection causes of death and loss of productivity in intensive pig farms. The aim of this research is the comparison between acoustic features of cough sounds originating from infectious and non infectious diseases. The acoustic parameters investigated are Peak Frequency and Duration of cough signals. The differences resulting from the sound analysis confirmed variability in acoustics parameters according to a state of health or disease. Infections change the status of respiratory system; thus infectious cough (I) sounds are different than healthy ones (H). Duration of single coughs is significant different among the classes analyzed: non infectious coughs (H), Actinobacillus (A) and Pasteurella’s (P) ones. Frequency analysis allows a more general classification between H and I. Sounds can be used in an alarm system based on an algorithm to identifies automatically cough sounds and provide early warning system for the farmer about the health status of his herd.