Article Preview
TopIntroduction
Market demands have raised the level of requirements concerning quality of food products, and specifically, the quality of animal products (milk and meat). Such demands encourage research on factors that influence behaviour parameters that are directly or indirectly related to animal husbandry, such as animal welfare.
The behaviour of an animal can be a clear indicator of its physiological state (Frost et al., 1997). Under thermal stress, cattle change their behaviour as a way to regulate body temperature and to maintain homothermia (Volpi et. al., 2020). These mechanisms can be physiological and/or behavioural, including changes in ingestive behaviour as a way to overcome adverse situations, whether related to environment temperature or to the quality and quantity of food available. The ingestive activities (grazing and rumination) can be indicators the animal health, welfare or nutrition status (De Boever et al., 1990), once that an animal changes its eating habits when it has its routine modified due to physical changes in the body or the environment. Monitoring ingestive behaviour is a procedure that helps early detection of diseases, improving welfare and reducing economic losses.
One commonly adopted procedure to monitor ingestive behaviour in cattle is the visual method (Figure 1, item 2). This method allows experts to identify, while cattle grazing, activities and jaw movements related to ingestive behaviour. This process requires the observers to be at a significant distance from the animals in order not to disrupt their natural routine. Experts are responsible for taking notes on ingestive behaviour and animal’s environment and positioning data (weather, temperature, animal position, etc). Despite being a well-known and simple procedure, the visual approach lacks scale and accuracy, since it is a time-consuming, and error-prone method, requiring an expert to be on site, monitoring the animal, and taking notes of its activities (Volpi et. al., 2020).
This work presents a non-invasive device to monitor ingestive behaviour based on the bioacoustic technique (Figure 1, item 1). The bioacoustics technique is a tool to characterize animal behaviour from the sounds they produce. The no-vocal sounds produced by the jaw when performing activities related to ingestive behaviour are the focus of many studies (Volpi et. al., 2020), (Chelotti et. Al., 2016). Those studies have used the bioacoustics technique, visual method, and labelling method for validation purposes.
The Labeling method (Figure 1, item 3) is commonly used in bioacoustics data to characterize events and compare results as well as it works as an important tool in the dataset training for machine learning algorithms (Tullo et. al., 2017), (Chelotti, et. al., 2016). It consists in the observation, extraction, identification, and classification of individual animal sounds based on amplitude or frequency of the sound signal (Tullo, et. Al., 2017). Figure 1 presents BPMN (Business Process Management Notation) diagrams correlating the bioacoustics technique to the visual method and labeling method. The bioacoustic technique uses the visual method and labelling method to perform the validation process and to serve as a basis for sound data capture.
Figure 1. BPMN diagram of bioacoustic technique applied in grazing cattle
We have designed and developed a low-cost, non-invasive computational and electronic device for the acquisition, processing, storage, and transmission of bioacoustics signals in grazing cattle. We have evaluated the device in grazing Nelore (Bos indicus) cattle and compared the results to the visual technique to identify ingestive behaviour activities. In addition to the flexibility and automatic data acquisition brought by the device, our tests showed that the proposed solution achieved high accuracy compared to the visual technique on detecting activities such as ruminating, idle, and grazing.