BEELIVINGSENSOR

THE IDEA

The 10 to 60 thousand honey bees in a colony are used as living sensors to determine the biodiversity in their environment.

With video cameras we observe how much pollen the bees carry in and what colors the pollen has. With this non-invasive observation method, the bees are not disturbed in their natural behavior. This also allows us to recognize behavior patterns at the flight hole.

Our formula for determining biodiversity is simple.

The more different pollen colors are added, the greater the plant diversity in the area and the more balanced the food for the bee brood.

COLLECT VIDEO DATA

With a video camera we observe the bees 7x24 at the flight hole. In addition, we record temperature and humidity data in the bee housing and in the direct vicinity. With a balance the weight of the beehive is recorded during the observation phase. All data are transmitted to a server.

LABELING OF THE VIDEO DATA

Before artificial intelligence can work, it must be trained. For this purpose we have a program to manually mark bees and pollen. Over time, the software recognizes the bees and pollen by itself.

EVALUATION OF THE VIDEOS WITH ARTIFICIAL INTELLIGENC

The trained artificial intelligence analyzes the individual images of the videos. That is 25 frames per second. The bee traffic at the flight hole is recorded and the pollen brought in is counted and its color determined.

IDENTIFICATION OF PLANTS, POLLEN AND INSECTS IN THE AREA

With a Citizen Science project, the flora and fauna in the environment of the bee colony is being 

recorded with a smartphone App. The plant species, their pollen colors and their positions are determined. This allows us to correlate flora and fauna with the pollen colors brought in by the bees.

DRONE FLIGHTS OVER THE AREA

Drone flights over the area provide information about the amount of flora present. The resolution of the drone images is not sufficient to identify the individual plants. But in comparison with the data from the smartphone app, conclusions can be drawn.

COMPARISON WITH PUBLIC WEATHER DATA

In addition to the microclimate data in and directly around the bee dwelling, public weather data is also collected. Location and exposure of the bee housing have a decisive influence on the behavior of the bees.

INTEGRATION OF AGRONOMIC DATA

Wherever possible, agronomic data are also collected. Location, extent and treatment of the plantation.

AGGREGATION OF ALL DATA ON AZURE

All data is stored on a server. As environment, we use Azure from Microsoft. Thanks to an AI for Earth grant we can collect and evaluate the large amount of data and make it available to other interested parties. Basically we want to make all data publicly available.