# Import Pointclouds PCD

Overview • Demo Data

[](https://ecosystem.supervise.ly/apps/supervisely-ecosystem/import-pointcloud-pcd) [](https://supervise.ly/slack)  [](https://supervise.ly) [](https://supervise.ly)

Overview

This app allows you to upload pointcloud PCD files with (or without) related images to Supervisely. The new pointloud project will be created.

Be aware that "remove files after successful import" flag is enabled by default, it will automatically remove source directory after import.

🏋️ Starting from version v1.1.1 application supports import from special directory on your local computer. It is made for Enterprise Edition customers who need to upload tens or even hundreds of gigabytes of data without using drag-ang-drop mechanism:

1. Run agent on your computer where data is stored. Watch how-to video.
2. Copy your data to special folder on your computer that was created by agent. Agent mounts this directory to your Supervisely instance and it becomes accessible in Team Files. Learn more in documentation. Watch how-to video.
3. Go to Team Files -> Supervisely Agent and find your folder there.
4. Right click to open context menu and start app. Now app will upload data directly from your computer to the platform.

Input files structure

Directory name defines project name, subdirectories define dataset names. Files in root directory will be moved to dataset with name "ds0".
ℹ️ You can download the archive with data example here.

Example 1. Import structure:

my_project
├── xxx.pcd
├── ds_ny
│   └── frame.pcd
└── ds_sf
    └── kitti_0000000001.pcd

In this case the following datasets will be created:

• ds_0 with a single file xxx.pcd
• ds_ny with a single file frame.pcd
• ds_sf with a single file kitti_0000000001.pcd

Example 2. Import structure:

my_project
└── dataset_01
    ├── xxx.pcd
    ├── ds_ny
    │   └── frame.pcd
    └── ds_sf
        └── kitti_0000000001.pcd

In this case only the one dataset dataset_01 will be created with all pointcloud files.

Example 3. PCD files with photo context:

my_project
└── dataset_01
    ├── frame.pcd
    ├── kitti_0000000001.pcd
    └── related_images
        └── kitti_0000000001_pcd
            ├── kitti_0000000001.png
            └── kitti_0000000001.png.json

if you want to attach photo context to pcd file just create a directory related_images near the file. Then create directory (in this example we name directory kitti_0000000001pcd - it's a filename + extension + all symbols . are replaced to ) and put there images and corresponding json files with projection matrix. See example for more info.

As a result we will get project my_project with 1 dataset dataset_01. Dataset will contain 2 pointcloud files, kitti_0000000001.pcd with related image, and frame.pcd without related image.

Demo Data

Download zip archive with demo pointcloud project with related image (0.79 MB)