SEARCH AND RESCUE IMAGE DATASET FOR PERSON DETECTION - SARD

For the task of detecting casualties and persons in search and rescue scenarios in drone images and videos, our database called SARD was built. The actors in the footage have simulate exhausted and injured persons as well as "classic" types of movement of people in nature, such as running, walking, standing, sitting, or lying down. Since different types of terrain and backgrounds determine possible events and scenarios in captured images and videos, the shots include persons on macadam roads, in quarries, low and high grass, forest shade, and the like. The obtained dataset comprises 1,981 manually labeled images extracted from video frames.

To increase the robustness of the SARD data, an extension of the SARD set, called Corr, was created that includes images that further simulate different weather conditions that may occur in actual search and rescue situations such as fog, snow, and ice. Also, blur images are included in the Corr set that occur in real conditions as a result of camera movement and aerial shooting in motion.