VSLAM-LAB数据集
收藏资源简介:
VSLAM-LAB数据集是由昆士兰大学QUT机器人中心和萨拉戈萨大学I3A共同创建的,包含12个不同的数据集,旨在为视觉同时定位与地图构建(VSLAM)研究提供统一框架。该数据集涵盖了室内外、真实与合成数据、不同难度级别和包含动态物体的场景等多种环境,能够帮助研究人员在没有地面真实数据的情况下,通过离线结构从运动技术生成伪地面真实数据来进行VSLAM方法的评估。
The VSLAM-LAB dataset was co-developed by the Queensland University of Technology (QUT) Robotics Centre and the University of Zaragoza’s I3A. It comprises 12 distinct datasets, designed to provide a unified framework for visual simultaneous localization and mapping (VSLAM) research. This dataset covers diverse environments including indoor and outdoor settings, real and synthetic data, scenes with varying difficulty levels, and scenarios containing dynamic objects. It enables researchers to evaluate VSLAM approaches by generating pseudo-ground truth data via offline structure-from-motion (SfM) techniques, even when real ground truth data is unavailable.

- 1VSLAM-LAB: A Comprehensive Framework for Visual SLAM Methods and Datasets昆士兰大学QUT机器人中心, 萨拉戈萨大学I3A · 2025年



