LaMAR数据集下载和使用流程

下载地址：https://lamar.ethz.ch/

github地址：microsoft/lamar-benchmark: Source code for the ECCV 2022 paper "Benchmarking Localization and Mapping for Augmented Reality". (github.com)

数据集的介绍可以看这个：LaMAR: Benchmarking Localization and Mapping for Augmented Reality（论文阅读笔记） - 知乎 (zhihu.com)

1、进入官网，点击链接下载：

2、进入之后会给你一个表单，填写基本信息后就可以获得下载地址：

3、地址点进去后显示如下：

4、一个是处理后的基准信息，为benchmark

benchmark的格式可以看这里，写的很清晰：https://github.com/microsoft/lamar-benchmark/blob/main/CAPTURE.md

location1/                                  # a Capture directory
├── sessions/                               # a collection of Sessions
│   ├── navvis1/                            # NavVis Session #1
│   │   ├── bt.txt                          # list of bluetooth measurements
│   │   ├── depths.txt                      # list of (rendered) depth maps, one per image
│   │   ├── images.txt                      # list of images with their paths
│   │   ├── pointclouds.txt                 # list of point clouds with their paths
│   │   ├── rigs.txt                        # rigid geometric relationship between sensors
│   │   ├── sensors.txt                     # list of all sensors with specs
│   │   ├── trajectories.txt                # pose for each (timestamp, sensor)
│   │   ├── wifi.txt                        # list of wifi measurements
│   │   ├── raw_data/                       # root path of images, point clouds, etc.
│   │   │   ├── images_undistorted/
│   │   │   ├── render/                     # root path for the rgb and depth maps renderings
│   │   │   └── pointcloud.ply
│   │   └── proc/                           # root path of processed assets
│   │       ├── meshes/                     # a collections of meshes
│   │       ├── alignment_global.txt        # global transforms between sessions
│   │       ├── alignment_trajectories.txt  # transform of each pose to a global reference
│   │       └── overlaps.h5                 # overlap matrix from this session to others
│   ├── hololens1/
│   │   ├── bt.txt
│   │   ├── depths.txt                      # list of depth maps with their paths
│   │   ├── images.txt
│   │   ├── rigs.txt
│   │   ├── sensors.txt
│   │   ├── trajectories.txt
│   │   ├── wifi.txt
│   │   ├── raw_data/
│   │   │   ├── images/
│   │   │   └── depths/
│   │   └── proc/
│   │       └── alignment/
│   └── phone1/
│       └── ...
├── registration/                           # the data generated during alignment
│   ├── navvis2/
│   │   └── navvis1/                        # alignment of navvis2 w.r.t navvis1
│   │       └─ ...                          # intermediate data for matching/registration
│   └── hololens1/
│   │   └── navvis1/
│   └── phone1/
│       └── navvis2/
└── visualization/                          # root path of visualization dumps
    └─ ...                                  # all the data dumped during processing (TBD)

另外里面的位姿没有直接给出，需要用superpoint+superglue进行处理

5、另一个是原始数据，为raw

我们在 3 个 Capture 目录中发布原始数据，其中每个会话对应于其中一台设备的记录。该数据包括未用于基准测试的其他传感器模式，例如深度、IMU 或 GPS。我们还发布了 NavVis 扫描仪获得的 3D 激光扫描（作为点云和网格），用于估计地面真实姿态。每个场景目录都包含文件“metadata_{phone,hololens}.json”，该文件指示每个记录的持续时间（以秒为单位）以及是否缺少某些传感器模式。

原始数据的一些处理可以看这个：lamar-benchmark/RAW-DATA.md at main · microsoft/lamar-benchmark (github.com)