Real-time machine learning detection algorithms are often found within autonomous vehicle technology and depend on quality datasets. It is essential that these algorithms work correctly in everyday conditions as well as under strong sun glare. Reports indicate glare is one of the two most prominent environment-related reasons for crashes. However, existing datasets, such as LISA and the German Traffic Sign Recognition Benchmark, do not reflect the existence of sun glare at all. This paper presents the GLARE traffic sign dataset: a collection of images with U.S based traffic signs under heavy visual interference by sunlight. GLARE contains 2,157 images of traffic signs with sun glare, pulled from 33 videos of dashcam footage of roads in the United States. It provides an essential enrichment to the widely used LISA Traffic Sign dataset. Our experimental study shows that although several state-of-the-art baseline methods demonstrate superior performance when trained and tested against traffic sign datasets without sun glare, they greatly suffer when tested against GLARE (e.g., ranging from 9% to 21% mean mAP, which is significantly lower than the performances on LISA dataset). We also notice that current architectures have better detection accuracy (e.g., on average 42% mean mAP gain for mainstream algorithms) when trained on images of traffic signs in sun glare.
翻译:实时机器学习检测算法往往是在自主车辆技术中找到的,并且取决于高质量的数据集。这些算法在日常条件下和在强烈的太阳光光下运行是正确的。报告显示,光线是撞车的两个最突出的环境相关原因之一。然而,现有的数据集,如LISA和德国交通信号识别基准,根本不反映太阳光亮的存在。本文展示GLARE交通信号数据集:在阳光的强烈视觉干扰下,以美国为基地的交通信号集成。GLARE包含2 157个太阳光线线的交通信号图像,从33个美国道路破胶片视频中提取。它为广泛使用的LISA交通信号数据集提供了必要的浓缩。我们的实验研究表明,尽管一些最先进的基线方法在对没有太阳光光光线的交通信号进行训练和测试时表现优异。在GLARE(例如,从9%到21%的平均太阳光线图示)测试时,它们受到极大伤害。在LISAAP中,我们所训练的正常的交通信号也大大低于目前主流图像的精确度。