Provident Vehicle Detection at Night (PVDN) 夜间公车检测（PVDN）

通过检测前照灯引起的光反射，直观地假设即将迎面驶来的车辆在实际物理上是可见的。这个数据集包含夜间农村环境中346个不同场景中的59746个带注释的灰度图像。在这些图像中，标记了所有迎面驶来的车辆，其相应的照明对象（例如前大灯）及其各自的光反射（例如护栏上的光反射）。

By detecting light reflections induced by vehicle headlights, one can intuitively assume that an oncoming vehicle is physically visible. This dataset includes 59,746 annotated grayscale images captured from 346 distinct nighttime rural scenarios. Within these images, all oncoming vehicles, their corresponding illuminated objects (e.g., headlights), and their respective light reflections (e.g., light reflections on guardrails) are annotated.

搜集汇总

背景与挑战

背景概述

该数据集名为Provident Vehicle Detection at Night (PVDN)，专注于夜间迎面车辆的检测，通过分析车辆前照灯产生的光反射来实现早期识别。它包含59746张灰度图像，采集自346个不同的夜间农村场景，并标注了车辆、光源对象及其光反射信息。