DroneVehicle Large-scale Drone Aerial Vehicle Detection Dataset

* This dataset supports online use.Click here to jump.

This dataset was collected and annotated by a research team from Tianjin University during their research on vehicle detection and counting in drone aerial images. The research team published the paper in 2020.Drone-based RGB-Infrared Cross-Modality Vehicle Detection via Uncertainty-Aware Learning".

The DroneVehicle dataset consists of 56,878 images collected by drones, half of which are RGB images and the rest are infrared images. The research team has rich annotations with directional bounding boxes for five categories. Among them, car has 389,779 annotations in RGB images and 428,086 annotations in infrared images; truck has 22,123 annotations in RGB images and 25,960 annotations in infrared images; bus has 15,333 annotations in RGB images and 16,590 annotations in infrared images; van has 11,935 annotations in RGB images and 12,708 annotations in infrared images; freight car has 13,400 annotations in RGB images and 17,173 annotations in infrared images.

In DroneVehicle, in order to mark objects on the borders of the image, the research team set a white border with a width of 100 pixels on the top, bottom, left, and right sides of each image, so that the size of the downloaded image is 840 x 712. When training the detection network, preprocessing can be performed to remove the surrounding white borders and change the image size to 640 x 512.