β›½Upload Dataset

Image, LiDAR and 2D & 3D Fusion LiDAR datasets format requirement

Dataset Format

Datasets can be uploaded from the local drive and throughURLs.

Image

Images data format requirement

Image files in .jpg and .png are supported.

Compressed files (in .zip, .gzip, .tar format) can also be uploaded directly..

.
β”œβ”€β”€ image_0 // Image 0
β”‚   β”œβ”€β”€ data1.jpg
β”‚   └── data2.jpg
β”œβ”€β”€ data // Data info, only for exporting, more details can reference "Data Info" section
β”‚Β Β  β”œβ”€β”€ data1.json
β”‚Β Β  └── data2.json
β”œβ”€β”€ result // Annotation result, more details can reference "Data Annotation Result" section
β”‚Β Β Β β”œβ”€β”€ data1.json
β”‚Β Β Β β”œβ”€β”€ data1_image_0_segmentation.png 
β”‚Β Β Β β”œβ”€β”€ data2.json
│   └── data2_image_0_segmentation.png
└── scene_1 // Scene, the structure is similar to the root directory
β”‚Β Β Β β”œβ”€β”€ image_0
β”‚Β Β Β β”œβ”€β”€ ...
β”‚Β Β Β β”œβ”€β”€ data
│   └── result

The image dataset sample can be downloaded from here.

This dataset is provided by PandaSet - a high-quality open-source dataset for autonomous driving.

If you don’t see a supported data or file type that you want to upload, Join our Slack community.

LiDAR (Point Cloud)

LiDAR dataset folder must be compressed and placed under the folder "point_cloud" and data must be in .pcd format.

LiDAR data format requirement

Folder name: lidar_point_cloud_0

File format: .pcd

Upload file: compressed files (in .zip, .gzip, .tar format)

The LiDAR dataset sample can be downloaded from here.

This dataset is provided by PandaSet - a high-quality open-source dataset for autonomous driving.

If you don’t see a supported data or file type that you want to upload, Join our Slack community.

2D & 3D LiDAR Fusion

2D & 3D LiDAR fusion dataset format requirement:

Folder structure

lidar_point_cloud_0: LiDAR data folder name

camera_config: camera parameter files folder name

camera_image_0: 1st set of image data folder name

camera_image_1: 2nd set of image data folder name

camera_image_2: 3rd set of image data folder name...

camera_image_N: N+1th set of image data folder name

scene_1: scene folder contains multiple lidar data structures

File format

LiDAR file format: .pcd

Image file format: .jpg and .png

Camera parameter file format: .json

Upload file: compressed files in .zip, .gzip, .tar format

All the same set of data must be placed in different folders with identical file names.

The 2D & 3D LiDAR Fusion dataset sample can be downloaded from here.

This dataset is provided by PandaSet - a high-quality open-source dataset for autonomous driving.

Structure

2D & 3D LiDAR Fusion dataset

β”œβ”€β”€ camera_config // Camera config, more details can reference "Point Cloud Camera Config" section
β”‚Β Β  β”œβ”€β”€ data1.json
β”‚Β Β  └── data2.json
β”œβ”€β”€ camera_image_0 // Camer image 0
β”‚Β Β  β”œβ”€β”€ data1.jpg
β”‚Β Β  └── data2.jpg
β”œβ”€β”€ camera_image_1 // Camer image 1
β”‚Β Β  β”œβ”€β”€ data1.jpg
β”‚Β Β  └── data2.jpg
β”œβ”€β”€ camera_image_2 // Camer image 2
β”‚Β Β  β”œβ”€β”€ data1.jpg
β”‚Β Β  └── data2.jpg
β”œβ”€β”€ lidar_config // Lidar config
β”‚Β Β  β”œβ”€β”€ data1.json
β”‚Β Β  └── data2.json
β”œβ”€β”€ lidar_point_cloud_0 // Lidar point cloud 0
β”‚Β Β  β”œβ”€β”€ data1.pcd
β”‚Β Β  └── data2.pcd
β”œβ”€β”€ result // Annotation result, more details can reference "Data Annotation Result" section
β”‚Β Β Β β”œβ”€β”€ data1.json
β”‚Β Β Β β”œβ”€β”€ data1_lidar_point_cloud_0_segmentation.pcd
β”‚Β Β Β β”œβ”€β”€ data2.json
│   └── data2_lidar_point_cloud_0_segmentation.pcd
β”œβ”€β”€ scene_1 // Scene, the structure is similar to the root directory
β”‚Β Β Β β”œβ”€β”€ camera_config
β”‚Β Β      β”œβ”€β”€ data1.json
β”‚Β Β      └── data2.json
β”‚Β Β Β β”œβ”€β”€ camera_image_0 // Camer image 0
β”‚Β Β      β”œβ”€β”€ data1.jpg
β”‚Β Β      └── data2.jpg
β”‚   β”œβ”€β”€ lidar_point_cloud_0 // Lidar point cloud 0
β”‚Β Β      β”œβ”€β”€ data1.pcd
β”‚Β Β      └── data2.pcd
│   └── result

Upload dataset with pre-annotation results

LiDAR basic and 2D & 3D LiDAR Fusion support to upload datasets with pre-annotation results.

All JSON files must be placed under theresult folder.

Tutorial

Dataset License

This dataset is provided by PandaSet - a high-quality open-source dataset for autonomous driving. These Datasets are provided to You under a Creative Commons Attribution 4.0 International Public License (β€œCC BY 4.0”).

Last updated