Høye, T.T., Mann, H.M.R. & Bjerge, K. 2020. Kamerabaseret overvågning af insekter på grønne bytage. Aarhus Universitet, DCE – Nationalt Center for Miljø og Energi, 18 s. - Videnskabelig rapport nr. 371. http://dce2.au.dk/pub/SR371.pdf
During the summer of 2019, we studied insects on a green roof in the city centre of Aarhus, Denmark and developed and tested a system for automated monitoring of insects. In particular, flower visitors were studied by three different methods; by direct observations of insects on sunny days, by collection in blue, yellow, and white pan traps, and by recording images of flowers and insects with cameras mounted right above the flowers of the green roof. Crops of honeybee, a few species of bumblebees and hoverflies, as well as seven-spot ladybird from a subset of the images were used to train a deep neural network for detection of these species in the entire collection of images. In this way, >100,000 insects were found among the 2,500,000 images recorded. We found a remarkably high frequency of insects in the images. About half of the insects in the images were honeybees, whereas the other half was mostly flower-visiting insects. Our study shows that flower resources (pollen and nectar) from green roofs makes up a resource for wild pollinators. The results also demonstrate that image-based monitoring of insects has a large potential in ecological monitoring
