Search Results (979)

To better analyze the health and welfare of individual dairy cows, this study proposes BR-Tracker, a multi-object tracking method designed for dense monitoring environments to address missed detections, tracking failures, and frequent identity switches. In the object detection stage, a Receptive Field Attention Downsampling (RFADown) module is introduced into the neck network of YOLOv10s to effectively process the locally visible regions of occluded cows by dynamically adjusting the receptive field. An improved Partial Bi-Level Routing Attention (PBRA) module is incorporated into the backbone network to simultaneously extract global and local features, while a Spatial Pyramid Pooling with Efficient Layer Aggregation Network (SPPELAN) module is adopted to enhance multi-scale feature aggregation. In the object tracking stage, an MPDIoU-based matching algorithm is designed to improve matching accuracy and the reliability of trajectory association. The dataset contains 15,420 images for object detection and 130 independent videos for multi-object tracking, of which 40 videos are selected for the final tracking evaluation. Experimental results show that BR-YOLOv10s achieves a precision, recall, and mean average precision (mAP) of 95.7%, 90.3%, and 95.2%, respectively. Compared with the YOLOv10s-ByteTrack baseline, the proposed method improves HOTA, MOTA, MOTP, and IDF1 by 4.4, 5.7, 3.0, and 6.1 percentage points, respectively, while reducing identity switches by 23.19%. In addition, the tracker achieves a processing speed of 43.2 FPS. These results demonstrate that the proposed method can effectively perform real-time detection and tracking of densely distributed dairy cows in complex intensive farming environments. Full article

Climate change is accelerating the frequency of extreme thermal environments, threatening animal welfare, global herd productivity, and agricultural economics. Heat stress—occurring when the ambient temperature-humidity index (THI) surpasses species-specific thermoneutral thresholds—initiates physiological heat-loss cascades (panting, sweating, and peripheral vasodilation), neuroendocrine disruptions (depressed thyroid hormones, elevated glucocorticoids, and insulin dysregulation), mitochondrial oxidative stress, and gut-barrier breakdown. These systemic disruptions drive substantial performance declines: voluntary dry matter intake drops by 15–25%, average daily gain by 15–30%, milk yield by 10–25%, conception rates by 30–50%, and poultry egg output and shell integrity by 15–25%. Species-specific vulnerability is governed by intrinsic anatomical and metabolic differences, including the substantial fermentation heat increment in ruminants, the lack of functional sweat glands and heavy feather insulation in poultry, and the low cutaneous evaporative capacity of swine. Furthermore, thermal stress induces oxidative damage via free radical accumulation and alters blood composition, triggering severe immunosuppression that heightens susceptibility to mastitis, respiratory complexes, and metabolic endotoxemia. Mitigation requires an integrated solutions matrix combining environmental cooling (tunnel ventilation, pad cooling, and shade), targeted nutrition (antioxidants, organic trace minerals, osmolytes, and rumen-protected fats), genomic selection (SLICK locus and thermotolerant crossbreeding), and real-time Precision Livestock Farming (PLF) monitoring. Addressing these challenges is paramount to ensuring sustainable livestock production, animal welfare, and global food security under impending climate scenarios. Full article

Asian elephants (Elephas maximus) face substantial thermoregulatory constraints because of their large body size, low relative surface area, sparse hair, and lack of functional sweat glands. Reliable body temperature measurement is essential for assessing thermal status and evaluating welfare in both wild and managed populations, but conventional rectal thermometry requires close physical contact, animal training, and repeated manual handling, making high-frequency, continuous, large-scale monitoring impractical. This study developed a non-invasive framework for automatically detecting the outer ear and extracting auricular surface temperature from infrared thermograms. Rectal temperature, regional surface temperatures, ambient temperature, and relative humidity were measured synchronously in eight semi-captive Asian elephants, yielding 425 matched observations. The associations between rectal temperature and the surface temperatures of three anatomical regions (head, outer ear, torso and limbs) were analyzed using repeated-measures correlation accounting for the non-independence of repeated measurements. Mean outer-ear temperature showed the strongest within-individual association with rectal temperature (r_rm = 0.395, p < 0.001), identifying the outer ear as the optimal thermal window for subsequent automated monitoring. Eight lightweight YOLO models—YOLOv5n, YOLOv5s, YOLOv8n, YOLOv8s, YOLO11n, YOLO11s, YOLO26n, and YOLO26s—were trained on 2178 annotated infrared images and evaluated on an independent 194-image test set from extra elephants. Model performance was assessed using detection metrics, inference speed, Bland–Altman agreement, Taylor diagram statistics, and a weighted multi-criteria score with Monte Carlo sensitivity analysis. YOLO11n achieved the best overall performance, with an mAP50 of 0.933 and an inference speed of 164 frames per second. The proposed framework provides an efficient method for automated auricular temperature monitoring and has potential applications in elephant welfare management and remote physiological surveillance. Full article

Livestock production must reconcile animal health, food safety, ecological limits, and farm viability. This narrative review proposes proactive health-oriented livestock production (PHLP) as a testable conceptual framework for alpine pastoral systems and as a step toward theory development, rather than as an already validated general theory. PHLP does not claim novelty for vaccination, nutrition, welfare, biosecurity, precision monitoring, or low-carbon practices considered separately. Its proposed contribution is a decision architecture that forecasts context- and stage-specific risk, detects deviation with a feasible sentinel dataset, activates a matched intervention before avoidable loss becomes clinically or economically visible, and recalibrates local triggers from animal, product, ecosystem, and livelihood outcomes. The hypothesized pathway links cold, hypoxia, forage scarcity, pathogen exposure, and management stress to sentinel change, trigger-based action, lower cumulative stress, and greater immune–metabolic, rumen–gut, and behavioral stability. We define entry points, priority rules, initial boundary conditions, measurable indicators, and three testable propositions for Tibetan sheep, yak, and other alpine ruminants. PHLP is expected to be most useful when a consequential risk has usable lead time, a sentinel variable can be observed, and a feasible action is available; it complements rather than replaces emergency clinical care and statutory disease control. Comparative longitudinal and multisite studies are needed to calibrate triggers and test whether PHLP improves decision timeliness and reduces seasonal body-condition loss, morbidity, mortality, antimicrobial treatment, product-quality variability, and emission intensity relative to reactive or calendar-based management. Full article

The You Only Look Once (YOLO) has been widely adopted in aquaculture monitoring and management due to its real-time performance and deployment flexibility. The recently introduced YOLO26 architecture incorporates Non-Maximum Suppression (NMS)-free end-to-end inference and is optimized for deployment on resource-constrained CPU-based devices, making it particularly relevant for edge deployment in commercial aquaculture applications. Nevertheless, its performance, operational efficiency, and deployment suitability compared with previous YOLO generations remain largely unvalidated in aquaculture-specific scenarios. This study benchmarks YOLO26 against three Ultralytics predecessors (YOLOv5u, YOLOv8, and YOLO11) across nano, small, and medium model scales for the detection of fish mortality, a critical indicator of fish population health and welfare, in recirculating aquaculture systems (RAS). Twelve model variants were evaluated for detection accuracy, training efficiency across seven dataset sizes, and inference performance on both high-performance NVIDIA A100 GPUs and the resource-constrained, CPU-only Raspberry Pi 5 edge device. All models achieved comparable performance on the full dataset, with mAP50 varying by only 1.25 percentage points across three independent training runs, indicating minimal influence of architectural generation on final mortality detection accuracy when sufficient training data are available. However, notable differences emerged in data efficiency and deployment performance. YOLOv8 demonstrated the strongest training efficiency, achieving 90% mAP50 with only 400 training images, whereas YOLO26 nano and small variants required 1000 images to reach comparable accuracy. In contrast, YOLO26 exhibited advantages during edge deployment, with YOLO26n achieving the highest inference speed on the Raspberry Pi 5 at 7.84 ± 0.13 FPS across three benchmark sessions, while YOLOv5mu outperformed all contemporary medium-scale architectures on CPU-based hardware. These results demonstrate that architectural novelty alone is an insufficient criterion for model selection. The findings support a deployment-oriented framework in which training data availability, target hardware, and inference requirements collectively inform model selection for aquaculture applications. Full article

Livestock production occupies a paradoxical position with respect to climate change: farm animals are highly vulnerable to heat stress, feed and water scarcity, and climate-sensitive disease, while the sector contributes an estimated 14.5% of anthropogenic greenhouse gas emissions, most of which is biogenic methane from enteric fermentation. This review evaluates the evidence for precision livestock farming (PLF)—continuous, automated, real-time monitoring of individual animals’ health, welfare, production and environmental impact—across dairy and beef cattle, small ruminants, pigs, and poultry. For mitigation, precision feeding and additive-dosing strategies have been associated with enteric methane reductions of approximately 10–25%; for adaptation, wearable and non-invasive sensors have been reported to detect heat-stress-related behavioural changes before productivity losses become apparent, and smart climate-control systems have been associated with housing energy-use reductions of roughly 5–10%. Much of this evidence derives from single-farm, small-sample or short-duration studies and should be read as indicative rather than generalisable. Adoption remains constrained by high investment costs, limited interoperability, insufficient technical support, and uneven applicability to extensive and smallholder systems. We conclude that PLF is a valuable enabling technology that, combined with genetic, nutritional, and management strategies, can strengthen the resilience and environmental sustainability of livestock systems under a changing climate. Full article

Effective management of urban Asian water monitor (Varanus salvator (Laurenti, 1768)) populations requires precise individual identification, yet traditional physical-marking methods remain invasive and labor-intensive. This study developed a non-invasive, automated photographic re-identification (Re-ID) system using deep learning and computer vision to facilitate population monitoring in semi-urban environments. We evaluated seven deep learning configurations based on ResNet50 incorporating Squeeze-and-Excitation (SE), Convolutional Block Attention Module (CBAM), and Batch Normalization neck (BNNeck) optimizations and benchmarked them against traditional feature matching (HotSpotter) using an open-set evaluation dataset of 3311 images across 161 Side-IDs focusing on unique lateral head-scale patterns. HotSpotter demonstrated immediate field viability, achieving a Rank-1 accuracy of 99.88% and a mean Average Precision (mAP) of 90.40%. Among the deep learning architectures, the baseline ResNet50 achieved the highest Rank-1 accuracy of 75.39% and mAP of 55.97%. As a decision-support framework, the deep learning pipeline achieved over 87% Rank-5 accuracy, drastically reducing manual screening effort and cognitive load during capture–mark–recapture surveys. This non-invasive framework establishes a scalable, welfare-friendly protocol for long-term urban wildlife management and biodiversity monitoring. Full article

Effective welfare monitoring in laying hens is increasingly challenged by growing flock sizes, declining farm numbers, and the practical limitations of assessor-based protocols under commercial conditions. Precision Livestock Farming (PLF) technologies offer opportunities to support large-scale welfare assessment, yet their application in laying hens remains predominantly limited to small-scale or prototype systems. This study addresses three complementary objectives aimed at informing future computer-vision-based PLF research and development in commercially housed laying hens. First, a systematic literature review was conducted to identify welfare-related categories monitored in laying hens between 2005 and 2025, the methods used to assess them, and the extent to which automated monitoring approaches have been applied. Second, outcomes of a TransRegional Expert Panel (TREP) within the OMELETTE project were synthesised to rank priority welfare challenges, evaluate the feasibility of different monitoring approaches, and compare expert perspectives with trends identified in the literature. Third, two pan-tilt-zoom (PTZ) camera setups implemented in semi-commercial aviary systems were described as proof-of-concept examples illustrating how multiple welfare challenges can be monitored using a single, multipurpose camera system. The literature review revealed a pronounced imbalance in monitoring frequency, with feather pecking dominating the literature while several other welfare challenges, including piling, disturbed sleep, and toe-pecking, remain comparatively underrepresented. TREP outcomes confirmed feather pecking as the highest-priority welfare challenge but also highlighted the importance of integrated monitoring approaches that combine time-intensive assessments, shorter checklists, and automated systems rather than relying on single indicators. Experts further considered the use of digital devices during routine barn inspections to be practically feasible. The PTZ proof-of-concept demonstrates how priority welfare challenges identified through both literature and expert input can be operationalised through automated, scheduled, and location-specific monitoring under commercial conditions. Together, these findings highlight the need for future PLF research to move beyond isolated measurements and small-scale trials towards integrated, cost-effective, and farm-specific welfare-monitoring systems that support adaptive, data-informed management strategies—for example, within a Plan–Do–Check–Act framework—and enable the development of digital standard operating procedures that generate actionable insights under real-world commercial constraints. Full article

Automated monitoring of welfare-relevant behavior in laying hens (Gallus gallus domesticus) has advanced with developments in computer vision, deep learning, and precision livestock farming (PLF). This narrative review integrates the ethological basis of welfare indicators with the development and readiness of monitoring technologies. It distinguishes routinely expressed diagnostic behaviors, including preening, locomotion, dustbathing, feeding, drinking, and nesting, from high-priority welfare risks such as aggression, piling, feather pecking, and inactivity or prostration. The review traces the progression from manual ethograms to semi-automated tools and recent artificial intelligence (AI) applications. These include You Only Look Once (YOLO)-based detection, multi-object tracking with BoT-SORT (a robust association-based tracking algorithm), pose estimation, and multimodal sensor fusion. Application-specific Technology Readiness Levels (TRL 1–9) indicate that most behavior-analysis systems remain at TRL 4–6. Several technologies extend into TRL 6–8, whereas few established systems reach TRL 9. Persistent barriers include domain shift under production conditions, annotation costs, inconsistent validation protocols, and limited economic accessibility. By linking welfare relevance, evaluation level, and deployment evidence, this review identifies priorities for scalable and actionable monitoring in commercial laying-hen production. Full article

Pasture quality monitoring is essential for optimizing grazing management and reducing the incidence of nutritional and metabolic disorders in ruminants, yet conventional field-based measurements remain labor-intensive and limited in spatial coverage. This review examines how remote sensing (RS) technologies can support pasture-based livestock systems by providing timely, scalable assessments of biomass, botanical composition, and nutritive attributes. Data from multispectral, hyperspectral, radio detection and ranging (RADAR), and light detection and ranging (LiDAR) sensors, acquired via satellite, unmanned aerial vehicle (UAV), and proximal platforms, are combined with machine learning (ML) methods and radiative transfer models to derive pasture biophysical and quality indicators. The reviewed evidence shows that RS reliably estimates pasture biomass and structural traits, while advances in spectral unmixing, data fusion, and artificial intelligence (AI) improve the characterization of heterogeneous swards and support emerging indicators related to forage quality. Integrating these remotely sensed metrics into grassland decision-support frameworks can enhance grazing allocation, inform fertilization and irrigation decisions, and help detect conditions associated with nutritional imbalances. Overall, the synthesis demonstrates that RS, particularly when combined with advanced modelling and cloud-based processing, offers a robust pathway for improving pasture monitoring and strengthening the nutritional management of ruminants, thereby supporting more sustainable and animal welfare-focused grazing systems. Full article

Extensive small-ruminant grazing sustains food production on the Mediterranean land with few alternative uses, yet supervision is intermittent and welfare oversight weak. Digital monitoring is expected to close that gap, but detection alone does not produce intervention. In remote rangelands the link is unreliable, so an inference may reach the farmer late, degraded or not at all, and an alert arriving after the animal has moved is not a weaker alert but a different decision. This study develops a human-in-the-loop decision framework in which the delivery state of the link becomes a decision variable rather than a transport detail, coupling technology, farmer behaviour and governance. Constructed through design science research and structured analysis of an EU-funded project’s specifications and evaluation plans, it comprises a seven-layer architecture, a decision-episode schema, a connectivity-aware alert lifecycle, a decision-provenance matrix, and an evidence-maturity classification constraining what may be claimed. Five decision classes from a Cretan sheep and goat pilot instantiate it. After applying it, specification inconsistencies invisible to component-level review surfaced, among which is an alert payload lacking the coordinates, confidence and severity that its own indicators require. The framework is an instantiated specification, not a validated intervention: no detection, delivery, acknowledgement, usability or sustainability outcome is measured. Full article

Precision livestock farming (PLF) integrates sensing technologies, data acquisition (DAQ) systems, and machine learning (ML) frameworks to continuously monitor individual animals and support welfare assessment through physiological and behavioral observations. Advances in infrared thermography, radar sensing, vision-based systems, acoustic monitoring, and wearable technologies have substantially expanded the ability to collect high-resolution data describing animal responses to internal and external stimuli. However, despite considerable technological progress, a persistent gap remains between sensing performance demonstrated under controlled experimental conditions and reliable deployment within commercial livestock environments. This gap is characterized by environmental variability, unrestricted animal movement, and operational constraints within commercial environments. Using a structured review methodology, this review examines sensing modalities, embedded DAQ architectures, communication strategies, ML methodologies, data privacy, farmer adoption, and an illustrative engineering workflow through the lens of welfare-relevant physiological characteristics. Emphasis placed on the distinction between direct sensor measurements and the biological processes they represent. Sensor outputs do not directly quantify welfare, stressors, or management outcomes; rather, they provide measurements of physiological and behavioral responses that require appropriate biological context for meaningful interpretation. As a result, welfare assessment does not depend solely on the ability to acquire data, but also on the ability to accurately relate those data to underlying physiological mechanisms. Within this framework, ML serves as a critical bridge between measurement and interpretation by enabling the analysis of complex, multimodal datasets. Future advancement of welfare-oriented PLF systems will require stronger alignment among sensing methodologies, physiological understanding, and practical deployment realities to generate meaningful, scalable, and biologically grounded welfare assessments. Full article

Climate change has intensified heat stress in poultry production systems, raising concerns about productivity, animal welfare, and system sustainability. This study aimed to analyze the evolution, structure, and emerging trends of scientific research on climate change and heat stress in industrial poultry production using a bibliometric approach. Records were retrieved from the Scopus and Web of Science Core Collection databases. After deduplication and eligibility screening, 342 documents published between 1974 and 2025 were retained. Bibliometric analyses were performed using Bibliometrix/Biblioshiny in RStudio, including publication dynamics, international collaboration, author co-citation, conceptual structure, thematic evolution, trend topics, and keyword co-occurrence. Scientific production increased markedly in recent years, with an annual growth rate of 8.36% and international co-authorship in 22.51% of the publications. China and the United States were the leading contributors. At the same time, the international research structure showed interconnected collaboration networks and complementary intellectual communities focused on physiological and neuroimmune responses, cellular and oxidative mechanisms, and nutritional and management-based mitigation. Heat stress was the main conceptual hub, connecting thermoregulation, oxidative stress, immunity, welfare, productive performance, and product quality. Heat tolerance, thermotolerance, chronic heat stress, and genetic adaptation showed recent or developing bibliometric visibility, although their structural and temporal patterns differed across analyses. However, heat waves, long-term resilience, validation under commercial conditions, climate-vulnerable regions, environmental engineering, and digital monitoring exhibited lower bibliometric representation and network centrality within the analyzed corpus. Future research should combine nutritional, physiological, genetic, environmental, and precision-monitoring approaches to support economically viable and climate-resilient poultry production. Full article

Organic dairy farming has expanded rapidly over the past three decades, driven by regulatory reforms, consumer demand, and growing recognition of its environmental, animal welfare, and potential human health benefits. Despite this growth, evidence comparing organic and conventional dairy systems remains fragmented across genetics, phenomics, animal health, and human health outcomes. This review synthesizes current knowledge through the lens of genotype × environment interactions, integrating evidence from four complementary domains: (1) genomic architecture and breeding strategies; (2) phenotypic performance, including milk production and composition, meat quality, nutrition, and reproductive traits; (3) animal health, disease resistance, antimicrobial use, and welfare; and (4) implications for human health. Holstein–Friesian cattle remain the predominant breed in both systems; however, organic production favors animals with greater robustness, longevity, grazing efficiency, and disease resilience. Genetic studies further demonstrate that highly heritable production traits share similar genetic architecture across production systems, whereas health, fertility, longevity, and other low-heritability functional traits exhibit stronger genotype × environment interactions and more system-specific genomic signatures. These findings suggest that breeding strategies developed for high-input conventional systems are unlikely to maximize performance under organic management. Collectively, the evidence supports a shift from selection focused primarily on milk yield toward genomic improvement of robustness, disease resistance, reproductive resilience, grazing adaptation, and lifetime productivity. We propose Genomically Optimized Organic Dairy (GOOD) as an emerging framework that integrates genomic selection, precision phenotyping, health monitoring, and environmental adaptation to develop dairy cattle better suited to organic production. Full article

Lameness remains one of the most significant welfare and economic challenges in modern dairy production. Traditional detection methods, particularly visual locomotion scoring, are limited by subjectivity, inconsistent application, and infrequent monitoring, which often delays the recognition of painful lesions. This review synthesizes recent validation studies of automated lameness detection (ALD) technologies and evaluates their diagnostic performance, validation design, and practical implementation across dairy systems. Studies were screened for relevance by a single reviewer based on title, abstract, and full-text content. Included studies represented sensor-based, pressure-based, vision-based, and multimodal detection platforms, with reported accuracies ranging from approximately 70% to 98% depending on modality and environmental setting. Vision-based systems demonstrated strong performance in controlled conditions, whereas field-validated systems showed more moderate but potentially more generalizable accuracy. Pressure-based platforms reported high diagnostic discrimination via area under the curve (AUC) analysis but face infrastructural limitations in commercial settings. Risk-of-bias assessment indicated that controlled experimental studies without external validation may overestimate deployment performance. Despite technological advances, variability in validation protocols, lesion thresholds, and environmental robustness limits direct comparison across systems. Future research should prioritize multi-farm external validation, standardized benchmarking frameworks, and multimodal integration within precision livestock farming ecosystems to improve reliability and adoption. Full article