Cambridge Landmarks

该数据集描绘了城市环境，并代表了户外定位任务。它包含了多种定位挑战，如遮挡、反射、运动模糊、光照条件、重复纹理以及视角和轨迹的变化。其规模大约在900到5500平方米之间，任务主要是相机姿态回归。

This dataset depicts urban environments and is intended for outdoor localization tasks. It encompasses various localization challenges, including occlusions, reflections, motion blur, lighting conditions, repetitive textures, as well as variations in viewing angles and trajectories. It covers an area of approximately 900 to 5500 square meters, with the core task being camera pose regression.

搜集汇总

背景与挑战

背景概述

Cambridge Landmarks数据集中的Kings College场景是一个大型户外视觉重定位数据集，包含原始视频、带有6-DOF相机姿态标签的图像帧和场景重建，主要用于支持PoseNet论文中实时6-DOF相机重定位的研究。