Multisensory Omnidirectional Long-term Place recognition (MOLP)

MOLP数据集是由科罗拉多矿业学院创建，旨在为多感官全方位长期地点识别提供基准。该数据集包含全方位强度和视差图像序列，捕捉了户外环境中长期外观变化和感知混叠的挑战。数据集在城市和山区两条路线收集，涵盖四季和早晚不同时间，以捕捉各种环境特征。MOLP数据集的创建过程涉及使用Occam Vision Group的3.2兆像素Omni Stereo相机和GPS模块进行数据记录。该数据集主要应用于自主移动机器人和自动驾驶车辆的导航，解决长期外观变化和感知混叠问题。

The MOLP dataset was developed by the Colorado School of Mines to serve as a benchmark for multi-sensory omnidirectional long-term place recognition. This dataset includes omnidirectional intensity and disparity image sequences, which capture the challenges of long-term appearance variations and perceptual aliasing in outdoor environments. It was collected along two routes: one urban and one mountainous, covering all four seasons and different times of day including morning and evening to encompass a wide range of environmental features. The creation of the MOLP dataset involved data recording using the 3.2-megapixel Omni Stereo camera and GPS module from the Occam Vision Group. This dataset is primarily applied to the navigation of autonomous mobile robots and autonomous vehicles, aiming to address the challenges of long-term appearance variations and perceptual aliasing.