Sound analysis for respiratory health monitoring in commercial piggeries

Respiratory pathologies have a high prevalence in intensive pig farming and cough is their principal symptom. It is well-known that, under intensive breeding conditions, it is very unlikely for a pig to reach the slaughter weight without having shownany kind of respiratory infection. Higher mortality and drop in production for reduced feed conversion and growth rate, are often related to expensive veterinarian intervention costs. The economic side of this problem is particularly relevant for farmers and they face diseases administrating large spectrum antibiotics mass therapy which helps for antibiotic resistance in both animals and meat consumers. In nowadays intensive livestock farming the high density of breeded animals helps a rapid spread of the disease which is not parallely followed by punctual observation of animals. Cough is a clear marker in case of respiratory disease warning the health status. It is common practice by veterinarians to assess cough sounds, by audio monitoring, in pig houses for diagnostic purposes. A limitation to this technique stands in the short observational period in both time and space. To achieve this goal, there have been attempts to identify the characteristics of coughing in animals (Van Hirtum and Berckmans, 2002, 2002; Ferrari et al., 2008, 2009, 2010) and automatically identify and localize cough sounds in field recordings (Aerts et al., 2005; Van Hirtum and Berckmans, 2001, 2003, 2003, Exadaktylos et al., 2009, Silva et al., 2008, 2009). This paper wants to show an overview on the highlights of these ten years research in the field of sound analysis for health monitoring in piggeries focusing especially on the last two years results.