Multiple Airport Scene Dataset

The MASD consists of many actual and online airport images. The experiment involves image acquisition equipment at aircraft training bases to record aircraft takeoff and landing videos. We also collect online airport videos. This article collects typical airport data from more than 200 airport videos to ensure the diversity of data, including rain, snow, fog, night and other environments. Finally, the MASD collects images every 5-20 seconds and manually filters and removes similar images from airport scene videos. The MASD has 5736 images, including 2,548 actual images and 3,188 online images.We uses the Labelling tool to manually label the images and draw bounding boxes in three categories: aircraft, person and vehicle. A total of 13576 objects are labelled in the MASD. The distribution plot shows that small objects occupy the majority of the data, which is consistent with the characteristics of airport scene images. If the aspect ratio of the object to the original image is less than 0.1, it is considered a small object. The proportion of small objects in the MASD is 61.1%.