Julio Galli

Resource heterogeneity and foraging behaviour of cattle across spatial scales

BackgroundUnderstanding the mechanisms that influence grazing selectivity in patchy environments is vital to promote sustainable production and conservation of cultivated and natural grasslands. To better understand how patch size and spatial dynamics influence selectivity in cattle, we examined grazing selectivity under 9 different treatments by offering alfalfa and fescue in patches of 3 sizes spaced with 1, 4, and 8 m between patches along an alley. We hypothesized that (1) selectivity is driven by preference for the forage species that maximizes forage intake over feeding scales ranging from single bites to patches along grazing paths, (2) that increasing patch size enhances selectivity for the preferred species, and that (3) increasing distances between patches restricts selectivity because of the aggregation of scale-specific behaviours across foraging scales.ResultsCows preferred and selected alfalfa, the species that yielded greater short-term intake rates (P < 0.0001) and greater daily intake potential. Selectivity was not affected by patch arrangement, but it was scale dependent. Selectivity tended to emerge at the scale of feeding stations and became strongly significant at the bite scale, because of differences in bite mass between plant species. Greater distance between patches resulted in longer patch residence time and faster speed of travel but lower overall intake rate, consistent with maximization of intake rate. Larger patches resulted in greater residence time and higher intake rate.ConclusionWe conclude that patch size and spacing affect components of intake rate and, to a lesser extent, the selectivity of livestock at lower hierarchies of the grazing process, particularly by enticing livestock to make more even use of the available species as patches are spaced further apart. Thus, modifications in the spatial pattern of plant patches along with reductions in the temporal and spatial allocation of grazing may offer opportunities to improve uniformity of grazing by livestock and help sustain biodiversity and stability of plant communities.

A real-time algorithm for acoustic monitoring of ingestive behavior of grazing cattle

A novel algorithm for monitoring the livestock grazing behavior is proposed.The three basic grazing events are detected and classified using acoustic signals.The algorithm shows robustness to different operational conditions.It has linear computational complexity and works fully automatically in real-time. Assessment of both grazing behavior and herbage intake are two very difficult tasks that can be concurrently accomplished by means of accurate detection, classification and measurement of grazing events such as chews, bites and chew-bites. It is well known that acoustic monitoring is among the best methods to automatically quantify and classify ingestive and rumination events in grazing animals. However, most existing methods of signal analysis appear to be computationally complex and costly, and are therefore difficult to implement. In this work, we present and test a novel analysis system called Chew-Bite Real-Time Algorithm (CBRTA) that works fully automatically in real-time to detect and classify ingestive events of grazing cattle. The system employs a directional wide-frequency microphone facing inwards on the forehead of animals, and a coupled signal analysis and decision logic algorithm that measures shape, amplitude, duration and energy of sound signals to iteratively detect and classify ingestive events. Performance and validation of the CBRTA was determined using two databases of grazing signals. Signals were recorded on dairy cows offered either, natural pasture ( N = 25 ), or experimental micro-swards in indoor controlled environment ( N = 50 ). The CBRTA exhibited a simple linear complexity capable to execute 50 times faster than real-time and without undermining overall recognition rate and accuracy when signals were processed at 4kHz sampling frequency and 8bits quantization. Furthermore, CBRTA was capable to detect ingestive events with a 97.4% success rate, while achieving up to 84.0% success for their classification as exclusive chews, bites or composite chew-bites. The methodology proposed with CBRTA has promising application in embedded microcomputer systems that necessarily depend on fast real-time execution to minimize computational load, power source and storage memory. Such a system can readily facilitate the transmission of processed data through wireless network or the storage in an onboard device.

Embedded system for real-time monitoring of foraging behavior of grazing cattle using acoustic signals

An embedded system for monitoring the livestock grazing behavior is presented.Grazing events are detected and classified in real-time using acoustic signals.Date, time and position information is obtained using a GPS receiver.Only the results from processing of the sound signal and GPS information are stored. Estimating forage intake and monitoring behavior of grazing livestock are difficult tasks. Real-time detection and classification of events like chew, bite and chew-bite are necessary to estimate that information. It is well-known that acoustic monitoring is one of the best ways to characterize feeding behavior in ruminants. Although several methods have been developed to detect and classify events, their implementation is restricted to desktop computers, fact that confines their application to off-line analysis of a reduced number of animals. In this work, we present the design and implementation of an electronic system specifically developed for real-time monitoring of feeding patterns in dairy cows. The system is based on an embedded circuit to process the sound produced by the animal in order to detect, classify and quantify events of ruminant feeding behavior. The system implements an algorithm recently developed, which was adapted to be executed on a microcontroller-based electronic system. Only the results of sound analysis are stored in flash memory units. In addition to sound information, data from a GPS receiver is also stored, thus building a package of information. A microcontroller with power management technology, combined with a high-efficiency harvesting power supply and power management firmware, enables long operational time (more than five days of continuous operation). The system was evaluated using audio signals derived from the feeding activity of dairy cows that were acquired under normal operational conditions. The system correctly detected 92% of the events (i.e. considering them as possible events without making a classification). When the three types of events (i.e. chew, bite and chew-bite) were considered for classification, the recognition rate was about 78%. These results were obtained using reference labels provided by experts in ruminant ingestive behavior. The technology presented within this publication is protected under the international patent application PCT/IB2015/053721.

A pattern recognition approach for detecting and classifying jaw movements in grazing cattle

Abstract Precision livestock farming is a multidisciplinary science that aims to manage individual animals by continuous real-time monitoring their health and welfare. Estimation of forage intake and monitoring the feeding behavior are key activities to evaluate the health and welfare state of animals. Acoustic monitoring is a practical way of performing these tasks, however it is a difficult task because masticatory events (bite, chew and chew-bite) must be detected and classified in real-time from signals acquired in noisy environments. Acoustic-based algorithms have shown promising results, however they were limited by the effects of noises, the simplicity of classification rules, or the computational cost. In this work, a new algorithm called Chew-Bite Intelligent Algorithm (CBIA) is proposed using concepts and tools derived from pattern recognition and machine learning areas. It includes (i) a signal conditioning stage to attenuate the effects of noises and trends, (ii) a pre-processing stage to reduce the overall computational cost, (iii) an improved set of features to characterize jaw-movements, and (iv) a machine learning model to improve the discrimination capabilities of the algorithm. Three signal conditioning techniques and six machine learning models are evaluated. The overall performance is assessed on two independent data sets, using metrics like recognition rate, recall, precision and computational cost. The results demonstrate that CBIA achieves a 90% recognition rate with a marginal increment of computational cost. Compared with state-of-the-art algorithms, CBIA improves the recognition rate by 10%, even in difficult scenarios.

A regularity-based algorithm for identifying grazing and rumination bouts from acoustic signals in grazing cattle

Abstract Continuous monitoring of cattle foraging behavior is a major requirement for precision livestock farming applications. Several strategies have been proposed for this task but monitoring of free-ranging cattle for a long period of time has not been fully achieved yet. In this study, an algorithm is proposed for long-term analysis of foraging behavior that uses the regularity of this behavior to recognize grazing and rumination bouts. Acoustic signals are analyzed offline in two main stages: segmentation and classification. In segmentation, a complete recording is analyzed to detect regular masticatory events and to define the time boundaries of foraging activity blocks. This stage also defines blocks that correspond to no foraging activity (resting bouts). The detection of event regularity is based on the autocorrelation of the sound envelope. For classification, the energy of sound signals within a block is analyzed to detect pauses and to characterize their regularity. Rumination blocks present regular pauses, whereas grazing blocks do not. The evaluation of the proposed algorithm showed very good results for the segmentation task and activity classification. Both tasks were extensively analyzed with a new set of multidimensional metrics. Frame-based F1-score was up to 0.962, 0.891 and 0.935 for segmentation, rumination classification, and grazing classification, respectively. The average time estimation error was below 0.5 min for classification of rumination and grazing on recordings of several hours in length. In addition, a comparison for rumination time estimation was done between the proposed system and a commercial one (Hi-Tag; SCR Engineers Ltd., Netanya, Israel). The proposed algorithm showed a narrower error distribution, with a median of −2.56 min compared to −13.55 min in the commercial system. These results suggest that the proposed system can be used in practical applications. Web demo available at: http://sinc.unl.edu.ar/web-demo/rafar/ .