VisDrone 2021

VisDrone2021 数据集由天津大学机器学习与数据挖掘实验室 AISKYEYE 团队收集。基准数据集由 400 个视频片段组成，由 265,228 帧和 10,209 张静态图像组成，由各种无人机摄像头拍摄，涵盖了广泛的方面，包括位置（取自中国相隔数千公里的 14 个不同城市）、环境（城市和乡村）、物体（行人、车辆、自行车等）和密度（稀疏和拥挤的场景）。请注意，数据集是使用各种无人机平台（即具有不同型号的无人机）、在不同场景以及各种天气和照明条件下收集的。这些框架使用超过 260 万个边界框或经常感兴趣的目标点进行手动注释，例如行人、汽车、自行车和三轮车。为了更好地利用数据，还提供了一些重要的属性，包括场景可见性、对象类别和遮挡。

The VisDrone2021 dataset was collected by the AISKYEYE Team from the Machine Learning and Data Mining Laboratory of Tianjin University. The benchmark dataset comprises 400 video clips, totaling 265,228 frames and 10,209 static images captured by various drone cameras. It covers a wide range of aspects, including locations (collected from 14 different cities in China that are thousands of kilometers apart), environments (urban and rural areas), objects (pedestrians, vehicles, bicycles, etc.), and scene densities (sparse and crowded scenarios). It is noteworthy that the dataset was collected using various drone platforms (i.e., drones with different models), in diverse scenarios, and under varying weather and lighting conditions. Over 2.6 million bounding boxes or target points of interest, such as pedestrians, cars, bicycles, and tricycles, were manually annotated for these frames. To facilitate better utilization of the data, several important attributes are also provided, including scene visibility, object categories, and occlusion status.

搜集汇总

背景与挑战

背景概述

VisDrone 2021是一个用于目标检测的无人机视觉数据集，由天津大学团队收集，包含400个视频片段和10,209张图像，总计超过260万个边界框注释，覆盖多种场景和目标。数据集在多样化的城市、环境和天气条件下采集，适用于计算机视觉任务如对象检测和跟踪。