FOD-A
收藏资源简介:
FOD-A数据集由具有跑道或滑行道背景的常见异物碎片 (FOD) 的图像组成。虽然主要注释样式由边界框组成,但FOD-A还包括单独的光照级别和天气分类注释 FOD-A发布了31个对象类别, 超过30,000注释实例。本文介绍了这一创作 方法,讨论公开可用的数据集扩展 流程,并展示了FOD-A的实用性,具有广泛的 用于对象检测的机器学习模型。
The FOD-A dataset comprises images of common Foreign Object Debris (FOD) against runway or taxiway backgrounds. While its primary annotation format is bounding boxes, FOD-A additionally provides separate illumination level and weather classification annotations. A total of 31 object categories and over 30,000 annotated instances have been released for this dataset. This paper introduces its development methodology, discusses the publicly available dataset expansion workflow, and validates the practical utility of FOD-A using a wide range of machine learning models for object detection.




