Search Results (42)

Unmanned aerial vehicles (UAVs) are increasingly used to monitor wildlife, but training robust detectors still requires large, consistently annotated datasets collected across seasons, habitats and flight altitudes. In practice, such data are scarce and expensive to label, especially when animals occupy only a few pixels in high-altitude imagery. We present a behaviour-aware synthetic data pipeline, implemented in Unreal Engine 5, that combines parameterised animal agents, procedurally varied environments and UAV-accurate camera trajectories to generate large volumes of labelled UAV imagery without manual annotation. Each frame is exported with instance masks, YOLO-format bounding boxes and tracking metadata, enabling both object detection and downstream behavioural analysis. Using this pipeline, we study YOLOv8s trained under six regimes that vary by data source (synthetic versus real) and input resolution, including a fractional fine-tuning sweep on a public deer dataset. High-resolution synthetic pre-training at 1280 px substantially improves small-object detection and, after fine-tuning on only 50% of the real images, recovers nearly all performance achieved with the fully labelled real set. At lower resolution (640 px), synthetic initialisation matches real-only training after fine-tuning, indicating that synthetic data do not harm and can accelerate convergence. These results show that behaviour-aware synthetic data can make UAV wildlife monitoring more sample-efficient while reducing annotation cost. Full article

Animal-borne tags are widely used for tracking and monitoring the movements, behaviour, and ecology of marine animals. Tagging can, however, adversely affect the hydrodynamic force balance and welfare of tagged animals, and consequently, the reliability and accuracy of data, such as by increasing drag, altering swimming characteristics, and reducing the survival rate of tagged animals. Therefore, it is important to understand and quantify the impact of tagging on marine animal hydrodynamics and to optimize the choice of tag and attachment position. In this study, computational fluid dynamics (CFD) modelling is used to simulate the flow around tagged and untagged mako sharks (Lamnidae) across their swim speed range for two dominant tag shapes, tagging sites, and body sizes. The results indicate that fin mounted tags can have a significant impact on shark hydrodynamics and energetic balance, increasing drag between 17.6% and 31.2% for a mako shark (2.95 m fork length) across the range of flow velocities tested (0.5 to 9.1 m/s). In comparison, the optimal tagging site for archival tags attached to the dorsal musculature leads to a minimal increase in drag for the larger sharks (>1.5 m), which becomes considerable for small sharks (1 m fork length; 5.1% to 7.6% increase) and leads to an average energetic cost equivalent to 7% of the daily energetic requirement of an untagged animal. Other aspects of the force balance are considered, which reveal a range of varied and complex effects. Recommendations for animal size thresholds (>1.5 m FL) and refinements of tagging practice are suggested. Full article

Stress, being an inherent trait, is a major driver of farm animal disease, leading to significant antimicrobial use (AMU). AMU is the recognized source of antimicrobial resistance (AMR). Among other ways, AMR spread can be controlled by selective breeding. To address this, stress susceptibility among daughters (gilts) of stressed mothers (sows) is tracked using AI and computer vision techniques. A deep learning-based model is trained and validated on the ground truth labels (through behaviour testing during recording of the videos) of stress susceptibility (SS) of mothers (sows). Then, this trained model is employed as a feature extractor for clustering techniques, such as K-means, Agglomerative, etc. This leads to the conclusion that more than 90% of stress-susceptible (SS) mothers had SS daughters. This result is crucial, as it can ease the process of selective breeding, where low stress-susceptible (LS) mothers could further be used for the breeding of the new generation. Full article

Marine mammal monitoring, a growing field of research, is critical to cetacean conservation. Traditional ‘tagging’ attaches sensors such as GPS to such animals, though these are intrusive and susceptible to infection and, ultimately, death. A less intrusive approach exploits UUV commanded by a human operator above ground. The development of AI for autonomous underwater vehicle navigation models training environments in simulation, providing visual and physical fidelity suitable for sim-to-real transfer. Previous solutions, including UVMS and L2D, provide only satisfactory results, due to poor environment generalisation while sensors including sonar create environmental disturbances. Though rich in features, image data suffer from high dimensionality, providing a state space too great for many machine learning tasks. Underwater environments, susceptible to image noise, further complicate this issue. We propose SWiMM_2.0, coupling a Unity simulation modelling of a BLUEROV UUV with a DRL backend. A pre-processing step exploits a state-of-the-art CMVAE, reducing dimensionality while minimising data loss. Sim-to-real generalisation is validated by prior research. Custom behaviour metrics, unbiased to the naked eye and unprecedented in current ROV simulators, link our objectives ensuring successful ROV behaviour while tracking targets. Our experiments show that SAC maximises the former, achieving near-perfect behaviour while exploiting image data alone. Full article

Wearable collar technologies have become integral to the advancement of precision livestock farming, revolutionizing how dairy cattle are monitored in terms of their behaviour, health status, and productivity. These devices leverage cutting-edge sensors, including accelerometers, RFID tags, GPS receivers, microphones, gyroscopes, and magnetometers, to provide non-invasive, real-time insights that enhance animal welfare, optimize resource use, and support decision-making processes in livestock management. This systematized review focuses on analyzing the sensors integrated into collar-based systems, detailing their functionalities and applications. However, significant challenges remain, including the high energy consumption of some sensors, the need for frequent recharging, and limited parameter coverage by individual devices. Future developments must focus on integrating multiple sensor types into unified systems to provide comprehensive data on animal behaviour, health, and environmental interactions. Additionally, advancements in energy-efficient designs, longer battery life, and cost-reduction strategies are essential to enhance the practicality and accessibility of these technologies. By addressing these challenges, wearable collar systems can play a pivotal role in promoting sustainable, efficient, and responsible livestock farming, aligning with global goals for environmental and economic sustainability. This paper underscores the transformative potential of wearable collar technologies in reshaping the livestock industry and driving the adoption of innovative farming practices worldwide. Full article

The agricultural industries have embraced the use of technologies as they improve efficiency and food security. The pork industry is no exception to this, as monitoring techniques and artificial intelligence allow for unprecedented capacity to track the physiological and behavioural condition of individual animals. This article reviews a range of those technologies in reference to the detection of oestrus in sows, a time when the ability to precisely ascertain physiological and behavioural changes associated with fluctuating hormone levels can have an immense impact on the economic profitability of the farm. The strengths and weaknesses of each technique from a practical application perspective are discussed, followed by considerations for further research and refinement. Full article

UWB positioning systems offer innovative solutions for precision monitoring dairy cow behaviour and social dynamics, yet their performance in complex commercial barn environments requires thorough validation. This study evaluated the TrackLab 2.13 (Noldus) UWB system in a dairy barn housing 44–49 cows. We assessed stationary tag positioning using ten fixed tags over seven days, proximity detection between eight cows and ten stationary tags, and moving tag positioning using three tags on a stick to simulate cow movement. System performance varied by tag location, with reliability ranging from 4.09% to 96.73% and an overall mean accuracy of 0.126 ± 0.278 m for stationary tags. After the provider updated the software, only 0.62% of measures exceeded the declared accuracy of 0.30 m. Proximity detection between moving cows and stationary tags showed 81.42% accuracy within a 2-m range. While generally meeting specifications, spatial variations in accuracy and reliability were observed, particularly near barn perimeters. These findings highlight UWB technology’s potential for precision livestock farming, welfare assessment, and behaviour research, including social interactions and space use patterns. Results emphasise the need for careful system setup, regular updates, and context-aware data interpretation in commercial settings to maximise benefits in animal welfare monitoring. Full article

Significant research investment into tropical rock lobster (TRL) aquaculture production methods has led to a rapidly developing industry in Vietnam and, more recently, in Australia. The need for an effective formulated feed has been highlighted for both industries; however, feed intake has been a consistent limitation. Visual and chemical cues regulating feed recognition and consumption are expected to yield valuable data, leading to increased feed intake. Lobsters were placed in white- and grey-coloured enclosures to examine the effect of background colour on their feeding behaviour in terms of feeding occurrence and response time. The impact of background colour on TRL feeding behaviour found no statistically significant differences between TRL in white and grey enclosures, suggesting grayscale contrast does not directly affect feeding behaviour. Experiment 2 studied the effects of coloured feeding zones on feeding response in white enclosures. Yellow feeding zones led to a decreased feeding time (473 ± 443 s) and increased time spent feeding (168 ± 1832 s) compared to other colours, possibly due to the yellow-blue chromaticity (b*) contrast. Experiment 3 examined chemo-attractants (glycine, taurine and inositol) and their influence on the feeding behaviour of TRL, but no increased responses were observed. Experiments two and three assessed TRL feeding activity in morning and evening periods, highlighting their nocturnal behaviour, with more feeding occurring in the evening. This project enhances our understanding of photoreceptive and chemoreceptive factors affecting TRL feeding behaviour with formulated feed. It also reveals the potential for background colour changes to enhance marketable colours in commercial settings. Additionally, the study confirmed the effective use of animal tracking software (EthoVision XT) for lobster species tracking in future behavioural trials. Full article

Behaviour is often the fundamental driver of disease transmission, where behaviours of individuals can be seen to scale up to epidemiological patterns seen at the population level. Here we focus on animal behaviour, and its role in parasite transmission to track its knock-on consequences for parasitism, production and pollution. Livestock face a nutrition versus parasitism trade-off in grazing environments where faeces creates both a nutritional benefit, fertilizing the surrounding sward, but also a parasite risk from infective nematode larvae contaminating the sward. The grazing decisions of ruminants depend on the perceived costs and benefits of the trade-off, which depend on the variations in both environmental (e.g., amounts of faeces) and animal factors (e.g., physiological state). Such grazing decisions determine the intake of both nutrients and parasites, affecting livestock growth rates and production efficiency. This impacts on the greenhouse gas costs of ruminant livestock production via two main mechanisms: (1) slower growth results in longer durations on-farm and (2) parasitised animals produce more methane per unit food intake. However, the sensitivity of behaviour to host parasite state offers opportunities for early detection of parasitism and control. Remote monitoring technology such as accelerometers can detect parasite-induced sickness behaviours soon after exposure, before impacts on growth, and thus may be used for targeting individuals for early treatment. We conclude that livestock host x parasite interactions are at the centre of the global challenges of food security and climate change, and that understanding livestock behaviour can contribute to solving both. Full article

The ability to overcome an opposition in team sports is reliant upon an understanding of the tactical behaviour of the opposing team members. Recent research is limited to a performance analysts’ own playing team members, as the required opposing team athletes’ geolocation (GPS) data are unavailable. However, in professional Australian rules Football (AF), animations of athlete GPS data from all teams are commercially available. The purpose of this technical study was to obtain the on-field location of AF athletes from animations of the 2019 Australian Football League season to enable the examination of the tactical behaviour of any team. The pre-trained object detection model YOLOv4 was fine-tuned to detect players, and a custom convolutional neural network was trained to track numbers in the animations. The object detection and the athlete tracking achieved an accuracy of 0.94 and 0.98, respectively. Subsequent scaling and translation coefficients were determined through solving an optimisation problem to transform the pixel coordinate positions of a tracked player number to field-relative Cartesian coordinates. The derived equations achieved an average Euclidean distance from the athletes’ raw GPS data of 2.63 m. The proposed athlete detection and tracking approach is a novel methodology to obtain the on-field positions of AF athletes in the absence of direct measures, which may be used for the analysis of opposition collective team behaviour and in the development of interactive play sketching AF tools. Full article

Triops longicaudatus is a crustacean typically inhabiting temporary freshwater bodies in regions with a Mediterranean climate. These crustaceans are easily maintained in the laboratory and show a set of biological features that make them good candidates for diagnosing environmental quality and health. However, information about their responses to environmental contamination is scarce. This study characterised the locomotor responses of juvenile and adult/mature T. longicaudatus to low concentrations of five model toxicants upon a very short 1.5 h exposure: tributyltin, mercury, lindane, sodium hypochlorite and formaldehyde. A video-tracking system was used to record the locomotor behaviour. The data were analysed with an artificial neural network to identify distinct behaviours, followed by Chi-square and Correspondence analysis to characterise the response to each toxicant. The results showed that T. longicaudatus is sensitive to aquatic contamination, particularly sodium hypochlorite. Six behaviour types were defined, which allowed for the characterisation and discrimination of the test toxicants. The results support the need for more investigation into this species and its behaviour types as an alternative to animal testing and the more apical and often invasive endpoints commonly recommended in standard guidelines. Full article

Understanding how marine animals behave around tidal turbines is essential if we are to quantify how individuals and populations may be affected by the installation of these devices in the coming decades. Our particular interest is in collision risk, and how this may be affected by the fine-scale behaviour of seals and small cetacean species around devices. We report on a study in which multibeam sonar data were collected close to an operational tidal turbine in Scotland continuously over a twelve-month period. The sonars provide high-resolution (a few cm) data over a 120° angle out to a range of 55 m at a rate of 10 frames per second. We describe a system which uses automatic computer algorithms to detect potential targets of interest, verified by human analysts using a sophisticated computer user interface to confirm detections and assign target species. To date, we have identified 359 tracks of marine mammals in the data, as well as several thousand tracks from fish and diving birds. These are currently being parameterised to study how these species react to the moving turbine rotors, and the data are now being used to explore the development of improved automated detection and classification algorithms. Full article

With rising interest in marine renewable energy (MRE) associated with offshore wind, waves, and tidal flows, the effects of device placement on changes in animal behaviour require proper assessment to minimise environmental impacts and inform decision making. High-frequency multibeam echosounders, or imaging sonars, can be used to observe and record the underwater movement and behaviour of animals at a fine scale (tens of metres). However, robust target detection and tracking of closely spaced animals are required for assessing animal–device and predator–prey interactions. Dual-frequency multibeam echosounders combine longer detection ranges (low frequency) with greater detail (high frequency) while maintaining a wide field of view and a full water column range compared to acoustic or optical cameras. This study evaluates the performance of the Tritech Gemini 1200ik imaging sonar at 720 kHz (low frequency) and 1200 kHz (high frequency) for small target detection with increasing range and the ability of the two frequency modes to discriminate between two closely spaced targets using a 38.1 mm tungsten carbide acoustic calibration sphere under controlled conditions. The quality of target detection decreases for both modes with increasing range, with a 25 m limit of detection at high frequency and a low-frequency mode able to detect the target up to 30 m under test conditions in shallow water. We quantified the enhanced performance of the high-frequency mode in discriminating targets at short ranges and improved target detection and discrimination at high ranges in the low-frequency mode. Full article

Cetaceans are known for their intelligence and display of complex behaviours including object use. For example, bowhead whales (Balaena mysticetus) are known to rub on rocks and some humpback whale (Megaptera novaeangliae) populations undertake lateral bottom feeding. Such underwater behaviour is difficult to observe but can play a critical role in the whales’ survival and well-being. Distinguishing social behaviours from those which serve a specific function remains challenging due to a lack of direct observations and detailed descriptions of such behaviours. A CATS (Customized Animal Tracking Solutions) suction cup tag with on board video and a 3D inertial measurement unit was deployed on three different humpback whales to assess their behaviour in the Gold Coast bay, Australia. Here, we present evidence of humpback whales (tagged and untagged individuals) performing bottom contact with prolonged rolling on sandy substrate. In addition, we showed that fish were actively feeding from the whales’ skin during this behaviour. We detail the behaviour and discuss possible drivers, with a focus on cetacean innovation, possible ectoparasite removal, and habitat preferences. Full article

Large grazers are essential for nature conservation. In order to prevent grazers from moving to unintended areas, it may be necessary to keep them inside enclosures. Physical fences present a number of problems, such as fragmenting the landscape. Virtual fencing, however, is a possible replacement for physical fencing, making it possible to enclose grazers without physical boundaries. Virtual fencing systems utilise collars with GPS technology to track animals and deliver auditory warnings and electric impulses to keep animals within predefined boundaries. This study examines how effective the virtual fencing system Nofence© is at enclosing calves in a holistically managed setting. Holistic management is a rotational grazing technique where an enclosure is grazed in small strips at a time. It is investigated whether the calves become habituated to the virtual fence and whether there is a correlation between the number of warnings received by every two calves in order to explore potential herd behaviour. Finally, this study examines which calves interact the most with the virtual fence by investigating the relationship between physical activity and number of interactions. Seventeen calves were fitted with a GPS collar from the company Nofence© and placed in a holistically managed enclosure. Data were gathered from 4 July to 30 September 2022. The study found that virtual fence was able to contain calves inside the designated enclosure, and over time the calves received notably fewer electrical impulses compared to auditory warnings. The results of Pearson’s correlation between auditory warnings received by two random calves were inconclusive, but the use of a sliding window analysis should be further explored. Lastly, the most physically active animals were the ones who received the most auditory warnings, but they did not receive more impulses. No significant correlation was found between the number of electric impulses received and the physical activity of the animals. Full article