LARD
收藏资源简介:
LARD数据集是由法国空中客车公司等机构创建,专注于视觉基础的着陆跑道检测。该数据集包含14433张合成图像和2315张真实着陆视频帧,总计16748条数据。数据集主要由合成图像构成,但也包括手动标注的真实着陆视频帧,以提高检测任务的真实性。创建过程中,使用了Google Earth Studio生成高质量的跑道图像,并通过自动标注技术丰富数据集。LARD数据集旨在支持自主着陆系统的研究,特别是在跑道检测方面的算法开发和数据质量分析。
The LARD Dataset was developed by institutions including Airbus France, focusing on vision-based landing runway detection. It contains 14,433 synthetic images and 2,315 real landing video frames, totaling 16,748 data samples. Although the dataset is primarily composed of synthetic images, it also includes manually annotated real landing video frames to improve the realism of the detection task. During the creation process, Google Earth Studio was used to generate high-quality runway images, and automatic annotation techniques were employed to enrich the dataset. The LARD Dataset aims to support research on autonomous landing systems, particularly algorithm development and data quality analysis for runway detection.

- 1LARD -- Landing Approach Runway Detection -- Dataset for Vision Based Landing空中客车公司, 法国 · 2023年



