OOVD (ORNL Overhead Vehicle Dataset)

创建此数据集是为了了解机器学习，计算机视觉和HPC通过利用GRIDSMART交通摄像机作为自适应交通控制的传感器来提高交通控制的能效方面的潜力，该传感器对交通的油耗特性具有敏感性。摄像机视野中的交通。GRIDSMART摄像机-一种现有的，现场的商业产品-感知交叉路口车辆的存在，并取代更传统的传感器 (如感应回路) 来向交通控制发出呼叫。这些摄像机具有从地平线到地平线的视图，提供了改善交通环境视图的潜力，可用于生成更好的控制算法。

This dataset was developed to investigate the potential of machine learning, computer vision, and high-performance computing (HPC) to improve the energy efficiency of traffic control. Specifically, it leverages GRIDSMART traffic cameras as sensors for adaptive traffic control, which are sensitive to the fuel consumption characteristics of traffic within their respective fields of view. GRIDSMART cameras are existing, field-deployed commercial products that detect the presence of vehicles at intersections, replacing more traditional sensors such as inductive loops to issue requests to traffic control systems. These cameras boast a horizon-to-horizon field of view, offering the potential to provide an enhanced perception of the traffic environment and facilitating the development of more effective traffic control algorithms.

搜集汇总

背景与挑战

背景概述

OOVD数据集由橡树岭国家实验室等机构于2020年发布，旨在利用GRIDSMART交通摄像机作为传感器，研究机器学习、计算机视觉和HPC在提高交通控制能效方面的潜力，以取代传统感应回路。