Traffic perception from aerial images using butterfly fields

Drones are nowadays considered as a valuable solution to monitor urban traffic. Object detectors face numerous challenges when dealing with high-resolution aerial images captured by drones, due to variations in altitude, viewing angle, and weather conditions. To address these challenges, the authors present an object detector called Butterfly detector that is tailored to detect objects in aerial images. It is an anchor-free method that leverages field-based representations. The authors introduce Butterfly fields, a type of composite field that describes the spatial information of output features as well as the scale of the detected objects. The authors employ a voting mechanism between related Butterfly vectors pointing to the object center. The authors highlight the benefits of their method for urban traffic monitoring by (i) evaluating the recall/precision rate of the detector on two publicly available drone datasets (UAVDT and VisDrone2019), and (ii) measuring the error rate for flow estimations on the newly released EPFL roundabout dataset. The authors outperform the performance of previous methods while remaining real-time.

Language

  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 01891729
  • Record Type: Publication
  • Files: TRIS
  • Created Date: Aug 28 2023 10:47AM