Search Results (92)

Thermoregulatory fanning behavior in honeybees is a vital indicator of colony health and environmental response. This study presents a novel dataset of 18,000 annotated video frames containing 57,597 instances capturing fanning behavior at the hive entrance across diverse conditions. Three state-of-the-art single-shot object detection models (YOLOv8, YOLO11, YOLO12) are evaluated using standard RGB input and two motion-enhanced encodings: Temporally Stacked Grayscale (TSG) and Temporally Encoded Motion (TEM). Results show that models incorporating temporal information via TSG and TEM significantly outperform RGB-only input, achieving up to 85% mAP@50 with real-time inference capability on high-performance GPUs. Deployment tests on the Jetson AGX Orin platform demonstrate feasibility for edge computing, though with accuracy–speed trade-offs in smaller models. This work advances real-time, non-invasive monitoring of hive health, with implications for precision apiculture and automated behavioral analysis. Full article

The European honey bee (Apis mellifera) significantly contributes to Australian agriculture, especially in honey production and the pollination of key crops. However, managed bee populations are declining due to pathogens, agrochemicals, poor forage, climate change, and habitat loss. Major threats include bacteria, fungi, mites, and pests. With the increasing demand for pollination and the movement of bee colonies, monitoring these threats is essential. It has been demonstrated that honey constitutes an easily accessible source of environmental DNA. Environmental DNA in honey comes from all organisms that either directly or indirectly aid in its production and those within the hive environments. In this study, we extracted eDNA from 135 honey samples and tested for the presence of DNA for seven key honey bee pathogens and pests—Paenibacillus larvae, Melissococcus plutonius (bacterial pathogens), Nosema apis, Nosema ceranae (microsporidian fungi), Ascosphaera apis (fungal pathogen), Aethina tumida, and Galleria mellonella (arthropod pests) by using end-point singleplex and multiplex PCR assays. N. ceranae emerged as the most prevalent pathogen, present in 57% of the samples. This was followed by the pests A. tumida (40%) and G. mellonella (37%), and the pathogens P. larvae (21%), N. apis (19%), and M. plutonius (18%). A. apis was detected in a smaller proportion of the samples, with a prevalence of 5%. Additionally, 19% of the samples tested negative for all pathogens and pests analysed. The data outlines essential information about the prevalence of significant arthropod, fungal, and bacterial pathogens and pests affecting honey bees in Australia, which is crucial for protecting the nation’s beekeeping industry. Full article

Honey bees play a vital role in ecosystem stability, and the need to monitor colony health has driven the development of IoT-based systems in beekeeping, with recent studies exploring both empirical and machine learning approaches to detect and analyze key hive conditions. In this study, we present an IoT-based system that leverages sensors to record and analyze the acoustic signals produced within a beehive. The captured audio data is transmitted to the cloud, where it is converted into mel-spectrogram representations for analysis. We explore multiple data pre-processing strategies and machine learning (ML) models, assessing their effectiveness in classifying queenless states. To evaluate model generalization, we apply transfer learning (TL) techniques across datasets collected from different hives. Additionally, we implement the feature extraction process and deploy the pre-trained ML model on a deep edge IoT device (Arduino Zero). We examine both memory consumption and execution time. The results indicate that the selected feature extraction method and ML model, which were identified through extensive experimentation, are sufficiently lightweight to operate within the device’s memory constraints. Furthermore, the execution time confirms the feasibility of real-time queenless state detection in edge-based applications. Full article

This study investigates the impact of microphone placement on honey bee audio monitoring for time-of-day classification, a key step toward automated activity monitoring and anomaly detection. Recognizing the time-dependent nature of bee behavior, we aimed to establish a baseline diurnal pattern recognition method. A custom apparatus enabled simultaneous audio acquisition from internal (brood frame, protected from propolization) and external hive locations. Sound signals were preprocessed using Power Spectral Density (PSD). Extra Trees and Convolutional Neural Network (CNN) classifiers were trained to identify diurnal activity patterns. Analysis focused on feature importance, particularly spectral characteristics. Interestingly, Extra Trees performance varied significantly. While achieving near-perfect accuracy (98–99%) with internal recordings, its accuracy was considerably lower (61–72%) with external recordings, even lower than CNNs trained on the same data (76–87%). Further investigation using Extra Trees and feature selection methods using Mean Decrease Impurity (MDI) and Recursive Feature Elimination with Cross-Validation (RFECV) revealed the importance of the 100–600 Hz band, with peaks around 100 Hz and 300 Hz. These findings inform future monitoring setups, suggesting potential for reduced sampling frequencies and underlining the need for monitoring of sound inside the beehive in order to validate methods being tested. Full article

We explored the stationarity, trend, and seasonality of the hive weight and in-hive temperature of ten managed honey bee (Apis mellifera) colonies at a research apiary of the Carl Hayden Bee Research Center in Tucson, Arizona, USA. The hives were monitored with electronic scales and in-hive temperature sensors from June to October 2022. The weight and temperature were recorded every five minutes around the clock. The collected data were curated into 2160 timestamped weight and 2160 timestamped temperature observations. We performed a systematic autoregressive integrated moving average (ARIMA) time series analysis to answer three fundamental questions: (a) Does seasonality matter in the ARIMA forecasting of hive weight and in-hive temperature? (b) To what extent do the best forecasters of one hive generalize to other hives? and (c) Which time series type (i.e., hive weight or in-hive temperature) is better predictable? Our principal findings were as follows: (1) The hive weight and in-hive temperature series were not white noise, were not normally distributed, and, for most hives, were not difference- or trend-stationary; (2) Seasonality matters, in that seasonal ARIMA (SARIMA) forecasters outperformed their ARIMA counterparts on the curated dataset; (3) The best hive weight and in-hive temperature forecasters of the ten monitored colonies appeared to be colony-specific; (4) The accuracy of the hive weight forecasts was consistently higher than that of the in-hive temperature forecasts; (5) The weight and temperature forecasts exhibited common qualitative patterns. Full article

Traditional methods for assessing honey storage in beehives predominantly rely on manual visual inspection, which often leads to inconsistencies and inefficiencies. This study presents an automated deep learning approach utilizing the YOLOv11 model to detect, classify, and quantify honey cells within Apis mellifera frames across monthly sampling periods. The model’s performance varied depending on image resolution and dataset partitioning. Using the free version of YOLOv11 with high-resolution images (960 × 960 resolution) and a dataset split of 90:5:5 for training, validating, and testing, the model achieved a mean average precision at IoU threshold of 0.5 (mAP@0.5) of 83.4% for uncapped honey cells and 80.5% for capped honey cells. A strong correlation (r = 0.94) was observed between the 90:5:5 and 80:10:10 dataset splits, indicating that increasing the volume of training data enhances classification accuracy. In parallel, the study investigated the relationship between the physical properties of honey and image-based honey storage detection. Of the four tested properties, electrical conductivity (R² = 0.19) and color (R² = 0.21) showed weak predictive power for honey storage area estimation, with even weaker associations found for pH and moisture content. The honey storage areas via 90:5:5 and 80:10:10 datasets moderately correlated (r = 0.44–0.46) with increasing electrical conductivity and color. Especially, electrical conductivity exhibited statistically significant correlations with dataset performance across different dataset splits (p < 0.05), suggesting some potential influence of chemical composition on model accuracy. Our findings demonstrate the viability of image-based honey classification as a reliable technique for monitoring beehive productivity. Additionally, the research on image-based honey detection can be a non-invasive solution for improved honey production, beehive productivity, and optimized beekeeping practices. Full article

Monitoring insect activity at hive entrances is essential for advancing precision beekeeping practices by enabling non-invasive, real-time assessment of the colony’s health and early detection of potential threats. This study evaluates deep learning models for detecting worker bees, pollen-bearing bees, drones, and wasps, comparing different YOLO-based architectures optimized for real-time inference on an RTX 4080 Super and Jetson AGX Orin. A new publicly available dataset with diverse environmental conditions was used for training and validation. Performance comparisons showed that modified YOLOv8 models achieved a better precision–speed trade-off relative to other YOLO-based architectures, enabling efficient deployment on embedded platforms. Results indicate that model modifications enhance detection accuracy while reducing inference time, particularly for small object classes such as pollen. The study explores the impact of different annotation strategies on classification performance and tracking consistency. The findings demonstrate the feasibility of deploying AI-powered hive monitoring systems on embedded platforms, with potential applications in precision beekeeping and pollination surveillance. Full article

In the context of increasing urbanisation, the question arises as to whether urban environments can provide honeybee colonies with floral resources comparable to those available in rural areas. The present study sought to evaluate the impact of land cover on nectar collection by bees in urban and rural apiaries. To this end, changes in the mass of 10 hives located in five urban–rural site pairs were monitored over two years (2021–2022) to assess nectar yield, weight loss, and the number of foraging days. The 3 km surroundings of each apiary were analysed using Sentinel-2 satellite imagery from the S2GLC-PL (National Satellite Information System 2025). The analysis identified eight distinct land cover classes: anthropogenic, agricultural, broad-leaved forest, coniferous forest, grassland, shrubs, wetlands, and water bodies. The findings revealed no statistically significant variation in the total nectar collected between urban and rural colonies (72.9 kg vs. 64.5 kg; p > 0.6). However, urban colonies exhibited a significantly higher number of foraging days (67 vs. 56). No significant correlations were identified between specific land cover types and nectar yield. Principal component analysis (PCA) and clustering revealed distinct landscape gradients, yet these did not influence nectar collection. The findings of this study indicate that diverse urban environments have the capacity to support beekeeping to a similar extent as rural areas and may even have superior conditions, provided that the continuity and diversity of nectar plants are maintained. Full article

Common eastern bumble bees (Bombus impatiens Cresson) play an essential role in pollinating lowbush blueberries (LB) in northern Quebec, but their costs and the suboptimal weather conditions during pollination highlight the need to find appropriate hive management strategies. A study was conducted in a LB field in Saguenay (Québec, Canada) focusing on the effects of hive insulation (I+ and I−), heating (H+ and H−), and placement in a single-row tree line windbreak. High-definition time-lapse cameras monitored hive activities and bumble bee foraging behaviors. We found that the conventional management of placing hives in full sun without insulation (I−) resulted in the lowest levels of bumble bee foraging activity and overall hive traffic. Placing bumble bee hives against a windbreak resulted in the highest numbers of bees entering hives with pollen (+156%), leaving hives (+69%), and overall hive traffic (+76%). Insulating hives with extruded polystyrene foam gave intermediate results, with a 105% increase in foraging activity compared to the conventional management method (I−H−). Interestingly, placing hives on seedling mats to maintain colony temperatures above 15 °C (H+) tended to decrease foraging activity and overall hive traffic. Our results show that strategic placement of bumble bee hives against windbreaks can significantly increase the activity of Bombus workers from those hives and can be used as a simple, low-cost, and efficient bumble bee hive management method by LB growers. Full article

The aim of our study was to compare the composition of polyphenolic compounds between the Croatian acacia (Robinia pseudoacacia L.) and chestnut (Castanea sativa Mill.) honey from several aspects: production season, hive position (on the edge and in the middle of a series of hives), part of the hive (small or normal extension), and honey extraction method (centrifuging or draining honey combs). Additionally, in acacia honey, we also monitored the influence of different storage temperatures (room temperature (RT) and 4 °C) on the content of phenolic compounds. To separate, identify and quantify individual flavonoids and phenolic acids from the honey, we used the HPLC method. The total polyphenols and antioxidant activity of the samples, their antimicrobial activity and their elemental content were also measured. The significant influence of the season, hive position, and extraction method on the total identified phenolic compounds, phenolic acids, flavonoids, total phenols and antioxidant activity was detected in almost all the acacia and chestnut honey samples. Chestnut honey from 2013 had more total phenolics (TPs) and antioxidant capacity (FRAP) than chestnut from 2014 and 2015. Honey collected from smaller extensions of hives had significantly higher TPs and FRAP compared to normal hive extensions. Centrifugation reduced the TPs and FRAP in most cases, but not always uniformly. Storage at RT led to the predominance of gallic, p-coumaric and benzoic acid in acacia honey, while storage at 4 °C maintained p-coumaric acid as the dominant phenolic acid. Flavonoids, particularly pinobanksin in acacia honey and hesperetin/pinobanksin in chestnut honey, were less affected by the storage conditions compared to phenolic acids. The non-centrifuged chestnut sample from 2015 showed the lowest MIC values against the most tested pathogenic bacteria. All the honey samples showed an extremely low concentration of heavy metals and relatively high concentrations of potassium and calcium. Full article

This study investigates the impact of Nosema infection in beehives on the physico-chemical and biochemical properties and spectral characteristics of honey as indicators of honey quality. Comprehensive analyses were performed on honey samples from hives with varying levels of Nosema infection, examining water content, free acidity, optical rotation, electrical conductivity, sugar composition, catalase activity, and pollen content. Honey from highly infected hives showed higher water content (up to 17.3%), lower optical rotation, reduced electrical conductivity, decreased glucose levels, and increased sucrose levels. Principal component analysis (PCA) identified distinct clustering of samples based on infection levels, with changes in the sugar profile, particularly higher phenolic compounds, correlating with increased infection levels. Fluorescence spectroscopy combined with PARAFAC modeling identified proteins and phenolic compounds as key discriminators of honey from infected hives. Correlation and PLS modeling further demonstrated strong relationships between spectral features and honey properties, including catalase activity and pollen content. This research presents a novel approach to evaluating the impact of Nosema infection on honey quality by integrating physico-chemical and biochemical analyses and sugar composition profiling with advanced spectroscopic techniques. These insights are invaluable for improving bee health monitoring practices and advancing sustainability in the beekeeping and honey production industries. Full article

The honey bee is an important insect pollinator that provides critical pollination services for natural and agricultural systems worldwide. However, inadequate food weakens honey bee colonies, making them vulnerable to various biotic and abiotic factors. In this study, we examined the impact of supplementary feeding on bees’ genes for antioxidative enzymes and vitellogenin, oxidative stress parameters, and the hygienic and grooming behavior. The colonies were divided into two experimental groups (with ten hives each): a treatment group that received the plant-based supplement and a control group. The experiment was conducted in two seasons, spring and summer. After the treatment, in both seasons, all the monitored parameters in the treatment group differed from those in the control group. The expression levels of genes for antioxidative enzymes were significantly lower, but the vitellogenin gene transcript level was significantly higher. Values of oxidative stress parameters were significantly lower. The levels of hygienic and grooming behavior were significantly higher. Therefore, our field study indicates that the tested supplement exerted beneficial effects on bees, reflected in reduced oxidative stress and enhanced hygienic and grooming behavior. Full article

Beeswax is one of the most important products for the well-being of bee colonies. The wax glands of young worker bees produce beeswax, which serves as a building material for honeycomb construction. Beekeepers using hives with mobile frames mainly utilize local beeswax to make foundations. Any paraffin addition represents adulteration, resulting in a high degree of contamination. During the preparation of re-used beeswax, losses during the process may instigate producers to add cheaper, wax-like substances like paraffin and tallow. This article presents a systematic investigation of the quality of beeswax foundation from six major producers in Vojvodina, Serbia, by applying the classic analytical procedure for the determination of selected physicochemical parameters and instrumental gas chromatography coupled with mass spectrometry (GC–MS) and Fourier transform infrared attenuated total reflection (FTIR–ATR) spectroscopy techniques. FTIR–ATR detected possible paraffin and beef tallow in 72 foundation sheet samples. This technique was complemented with GC–MS. This analysis revealed that paraffin content ranged between 19.75 and 85.68%, while no tallow was detected over the two-year period. Two sheets from each manufacturer were placed into wired Langstroth–Ruth frames and placed in beehives. The construction, based on built cells, was monitored every 24 h. Evaluating newly inserted sheets proved that without quality nectar, there is no intensive building, regardless of adulteration. Full article

Objective: This qualitative study sought to explore patient experiences with technologies used in the Community Health in a Virtual Environment (Co-HIVE) pilot trial. Technology is becoming increasingly prevalent in mental healthcare, and user acceptance is critical for successful adoption and therefore clinical impact. The Co-HIVE pilot trialled a model of care whereby community-dwelling patients with symptoms of depression utilised virtual appointments and remote monitoring for the assessment and management of their condition, as an adjunct to routine care. Methods: Using a qualitative descriptive design, participants for this study were patients with symptoms of moderate to severe depression (based on the 9-item Patient Health Questionnaire, PHQ-9), who had completed the Co-HIVE pilot. Data was collected via semi-structured interviews that were audio-recorded, transcribed clean-verbatim, and thematically analysed using the Framework Method. Results: Ten participants completed the semi-structured interviews. Participants reported experiencing more personalised care, improved health knowledge and understanding, and greater self-care, enabled by the remote monitoring technology. Additionally, participants reported virtual appointments supported the clinician–patient relationship and improved access to mental health services. Conclusions: This experience of participants with the Co-HIVE pilot indicates there is a degree of acceptance of health technologies for use with community mental healthcare. This acceptance demonstrates opportunities to innovate existing mental health services by leveraging technology. Full article

From June to October, 2022, we recorded the weight, the internal temperature, and the hive entrance video traffic of ten managed honey bee (Apis mellifera) colonies at a research apiary of the Carl Hayden Bee Research Center in Tucson, AZ, USA. The weight and temperature were recorded every five minutes around the clock. The 30 s videos were recorded every five minutes daily from 7:00 to 20:55. We curated the collected data into a dataset of 758,703 records (280,760–weight; 322,570–temperature; 155,373–video). A principal objective of Part I of our investigation was to use the curated dataset to investigate the discrete univariate time series forecasting of hive weight, in-hive temperature, and hive entrance traffic with shallow artificial, convolutional, and long short-term memory networks and to compare their predictive performance with traditional autoregressive integrated moving average models. We trained and tested all models with a 70/30 train/test split. We varied the intake and the predicted horizon of each model from 6 to 24 hourly means. Each artificial, convolutional, and long short-term memory network was trained for 500 epochs. We evaluated 24,840 trained models on the test data with the mean squared error. The autoregressive integrated moving average models performed on par with their machine learning counterparts, and all model types were able to predict falling, rising, and unchanging trends over all predicted horizons. We made the curated dataset public for replication. Full article