Back

Print

DLR Multi-class Vehicle Detection and Orientation in Aerial Imagery (DLR-MVDA)

The automatic detection of vehicles from aerial imagery is of interest for various applications such as traffic management, parking surveillance, and urban planning. The airborne optical sensors enable a fast wide-area recording of traffic data and parking information. In comparison to the ground-based sensors, using aerial images for monitoring the traffic in an area is much more time and cost efficient. In order to support research in this field, we provide an open aerial image dataset with annotated vehicles, so-called DLR-MVDA.

Example from the dataset: The annotated vehicles are illustrated by bounding boxes, where the green and blue-green colors indicate different vehicle types.