MVSEC 交通工具事件相机数据集

MVSEC 全称 Multi Vehicle Stereo Event Camera，是用于开发基于事件相机的新型 3D 感知算法的数据集（事件相机具有新型异步传感方式，可用于测量图像强度的变化）。数据来自汽车、摩托车、六轴飞行器和手持式数据，并与激光雷达、 IMU 、运动捕捉和 GPS 相融合，以提供 ground truth pose 和深度图像。该数据集还提供基于标准立体框架相机拍摄的图像，可与传统技术进行比较。

MVSEC, whose full name is Multi Vehicle Stereo Event Camera, is a dataset developed for advancing novel 3D perception algorithms based on event cameras. Event cameras adopt a novel asynchronous sensing modality that measures changes in image intensity. The dataset includes data collected from automobiles, motorcycles, hexacopters, and handheld setups, and is fused with LiDAR, IMU, motion capture systems and GPS to provide ground truth pose and depth images. Additionally, it provides images captured by standard stereo frame cameras, enabling comparison with traditional computer vision techniques.

搜集汇总

背景与挑战

背景概述

MVSEC是一个用于开发基于事件相机的3D感知算法的数据集，其数据来自汽车、摩托车等交通工具，并融合了激光雷达、IMU等多种传感器以提供姿态和深度真值。该数据集还包含传统立体相机图像，便于与传统技术进行比较。