Florian Thürkow

Professur für Informationssysteme im Tenure-Track




Performance evaluation of ML- object-detection architectures for drone-based hamster hole detection

Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam nonumy eirmod tempor invidunt ut labore et dolore magna.


Agricultural food production is critical for food security, but overabundant rodent populations, especially field mice, can destroy farmers‘ yields. However, interfering with the population is not allowed if protected species, such as the hamster, are present. To address this challenge, a German research group is developing a new technology that uses a drone to capture images of affected farmland and AI-based image classification to count rodent borrows. In a recent study, researchers compared 32 neural network architectures from TensorFlow’s object-detection model zoo to find the best machine learning model for the job. While three models were recommended, the study’s limitations included the need for more diverse data and hardware limitations that affected the model’s quality.


Share on linkedin
Share on xing
Share on facebook
Share on twitter
Consent Management Platform von Real Cookie Banner