Convolutional Neural Networks for Real Time Classification of Beehive Acoustic Patterns on Constrained Devices.

Recent research has demonstrated the effectiveness of convolutional neural networks (CNN) in assessing the health status of bee colonies by classifying acoustic patterns. However, developing a monitoring system using CNNs compared to conventional machine learning models can result in higher computation costs, greater energy demand, and longer inference times. This study examines the potential of CNN architectures in developing a monitoring system based on constrained hardware. The experimentation involved testing ten CNN architectures from the PyTorch and Torchvision libraries on single-board computers: an Nvidia Jetson Nano (NJN), a Raspberry Pi 5 (RPi5), and an Orange Pi 5 (OPi5). The CNN architectures were trained using four datasets containing spectrograms of acoustic samples of different durations (30, 10, 5, or 1 s) to analyze their impact on performance. The hyperparameter search was conducted using the Optuna framework, and the CNN models were validated using k-fold cross-validation. The inference time and power consumption were measured to compare the performance of the CNN models and the SBCs. The aim is to provide a basis for developing a monitoring system for precision applications in apiculture based on constrained devices and CNNs.

Ambient Electromagnetic Radiation as a Predictor of Honey Bee (Apis mellifera) Traffic in Linear and Non-Linear Regression: Numerical Stability, Physical Time and Energy Efficiency.

Since bee traffic is a contributing factor to hive health and electromagnetic radiation has a growing presence in the urban milieu, we investigate ambient electromagnetic radiation as a predictor of bee traffic in the hive's vicinity in an urban environment. To that end, we built two multi-sensor stations and deployed them for four and a half months at a private apiary in Logan, UT, USA. to record ambient weather and electromagnetic radiation. We placed two non-invasive video loggers on two hives at the apiary to extract omnidirectional bee motion counts from videos. The time-aligned datasets were used to evaluate 200 linear and 3,703,200 non-linear (random forest and support vector machine) regressors to predict bee motion counts from time, weather, and electromagnetic radiation. In all regressors, electromagnetic radiation was as good a predictor of traffic as weather. Both weather and electromagnetic radiation were better predictors than time. On the 13,412 time-aligned weather, electromagnetic radiation, and bee traffic records, random forest regressors had higher maximum R2 scores and resulted in more energy efficient parameterized grid searches. Both types of regressors were numerically stable.

Comparative Study of Machine Learning Models for Bee Colony Acoustic Pattern Classification on Low Computational Resources.

In precision beekeeping, the automatic recognition of colony states to assess the health status of bee colonies with dedicated hardware is an important challenge for researchers, and the use of machine learning (ML) models to predict acoustic patterns has increased attention. In this work, five classification ML algorithms were compared to find a model with the best performance and the lowest computational cost for identifying colony states by analyzing acoustic patterns. Several metrics were computed to evaluate the performance of the models, and the code execution time was measured (in the training and testing process) as a CPU usage measure. Furthermore, a simple and efficient methodology for dataset prepossessing is presented; this allows the possibility to train and test the models in very short times on limited resources hardware, such as the Raspberry Pi computer, moreover, achieving a high classification performance (above 95%) in all the ML models. The aim is to reduce power consumption and improves the battery life on a monitor system for automatic recognition of bee colony states.

Model for finding the number of honey bee colonies needed for the optimal foraging process in a specific geographical location.

Finding a proper location for a bee apiary is a crucial task for beekeepers and especially for travelling beekeepers. Normally beekeepers choose an appropriate apiary location based on their previous experience and sometimes the location may not be optimal for the bee colonies. This can be explained by different flowering periods, variation of resources at the known fields, as well as other factors. In addition it is very challenging to evaluate how many bee colonies should be placed in one geographical location for an optimal nectar foraging process. This research presents a model for finding the number of honey bee colonies needed for the optimal foraging process in the specific location, taking into account several assumptions. Authors propose to take into account potential field productivity, possible chemical contamination, surroundings of the apiary. To run the model, several steps have to be completed, starting from the selection of area of interest, conversion to polygons for further calculations, defining the roads in the selected area. The outcome of the model number of colonies that should be placed is presented to the user. The Python language was used for the model development. The model can be extended to use additional factors and values to increase the precision of the evaluation. In addition, input from users (farmers, agricultural specialists, etc.) about external factors that can affect the number of bee colonies in the apiary can be taken into account. This work is conducted within the Horizon 2020 FET project HIVEOPOLIS (Nr.824069).

Smart apiculture management services for developing countries-the case of SAMS project in Ethiopia and Indonesia.

The European Union funded project SAMS (Smart Apiculture Management Services) enhances international cooperation of ICT (Information and Communication Technologies) and sustainable agriculture between EU and developing countries in pursuit of the EU commitment to the UN Sustainable Development Goal "End hunger, achieve food security and improved nutrition and promote sustainable agriculture". The project consortium comprises four partners from Europe (two from Germany, Austria, and Latvia) and two partners each from Ethiopia and Indonesia. Beekeeping with small-scale operations provides suitable innovation labs for the demonstration and dissemination of cost-effective and easy-to-use open source ICT applications in developing countries. SAMS allows active monitoring and remote sensing of bee colonies and beekeeping by developing an ICT solution supporting the management of bee health and bee productivity as well as a role model for effective international cooperation. By following the user centered design (UCD) approach, SAMS addresses requirements of end-user communities on beekeeping in developing countries, and includes findings in its technological improvements and adaptation as well as in innovative services and business creation based on advanced ICT and remote sensing technologies. SAMS enhances the production of bee products, creates jobs (particularly youths/women), triggers investments, and establishes knowledge exchange through networks and initiated partnerships.

Dynamics of Weight Change and Temperature of Apis mellifera (Hymenoptera: Apidae) Colonies in a Wintering Building With Controlled Temperature.

Honey bee wintering in a wintering building (indoors) with controlled microclimate is used in some cold regions to minimize colony losses due to the hard weather conditions. The behavior and possible state of bee colonies in a dark room, isolated from natural environment during winter season, was studied by indirect temperature measurements to analyze the expression of their annual rhythm when it is not affected by ambient temperature, rain, snow, wind, and daylight. Thus, the observed behavior in the wintering building is initiated solely by bee colony internal processes. Experiments were carried out to determine the dynamics of temperature above the upper hive body and weight dynamics of indoors and outdoors wintered honey bee colonies and their brood-rearing performance in spring. We found significantly lower honey consumption-related weight loss of indoor wintered colonies compared with outdoor colonies, while no significant difference in the amount of open or sealed brood was found, suggesting that wintering building saves food and physiological resources without an impact on colony activity in spring. Indoor wintered colonies, with or without thermal insulation, did not have significant differences in food consumption and brood rearing in spring. The thermal behavior and weight dynamics of all experimental groups has changed in the middle of February possibly due to increased brood-rearing activity. Temperature measurement above the upper hive body is a convenient remote monitoring method of wintering process. Predictability of food consumption in a wintering building, with constant temperature, enables wintering without oversupply of wintering honey.

Honey Bee Colonies Remote Monitoring System.

Bees are very important for terrestrial ecosystems and, above all, for the subsistence of many crops, due to their ability to pollinate flowers. Currently, the honey bee populations are decreasing due to colony collapse disorder (CCD). The reasons for CCD are not fully known, and as a result, it is essential to obtain all possible information on the environmental conditions surrounding the beehives. On the other hand, it is important to carry out such information gathering as non-intrusively as possible to avoid modifying the bees' work conditions and to obtain more reliable data. We designed a wireless-sensor networks meet these requirements. We designed a remote monitoring system (called WBee) based on a hierarchical three-level model formed by the wireless node, a local data server, and a cloud data server. WBee is a low-cost, fully scalable, easily deployable system with regard to the number and types of sensors and the number of hives and their geographical distribution. WBee saves the data in each of the levels if there are failures in communication. In addition, the nodes include a backup battery, which allows for further data acquisition and storage in the event of a power outage. Unlike other systems that monitor a single point of a hive, the system we present monitors and stores the temperature and relative humidity of the beehive in three different spots. Additionally, the hive is continuously weighed on a weighing scale. Real-time weight measurement is an innovation in wireless beehive-monitoring systems. We designed an adaptation board to facilitate the connection of the sensors to the node. Through the Internet, researchers and beekeepers can access the cloud data server to find out the condition of their hives in real time.