BeeCount - cloud-based, multicriterial and multivariate hyperparameter optimization for the Pareto-optimal design of neural networks for bees tracking

The systems engineering laboratory of ISES and faculty of electrical engineering and information technology FK04 at the University of Applied Sciences in Munich developed in cooperation with Mellifera e.V. an infrared camera-based environment and bee monitoring (EnBeeMo) system, which enables beekeepers to quantify the changes of the bee population in a species-appropriate manner. For this purpose, the number of incoming and outgoing bees is counted on a hive with the help of machine learning approaches.

A so-called convolutional neural network (CNN) is used to identify the bees.

These neural networks are currently the most used common algorithms in the field of object recognition.

Click here to find more information about this project.