nuscenes-devkit 1.1.2

The official devkit of the nuScenes dataset (www.nuscenes.org).

nuScenes devkit

Welcome to the devkit of the nuScenes and nuImages datasets.

Overview

• Changelog
• Devkit setup
• nuImages
  • nuImages setup
  • Getting started with nuImages
• nuScenes
  • nuScenes setup
  • nuScenes-lidarseg
  • Prediction challenge
  • CAN bus expansion
  • Map expansion
  • Map versions
  • Getting started with nuScenes
• Known issues
• Citation

Changelog

• Nov. 23, 2020: Devkit v1.1.2: Release map-expansion v1.3 with lidar basemap.
• Nov. 9, 2020: Devkit v1.1.1: Lidarseg evaluation code, NeurIPS challenge announcement.
• Aug. 31, 2020: Devkit v1.1.0: nuImages v1.0 and nuScenes-lidarseg v1.0 code release.
• Jul. 7, 2020: Devkit v1.0.9: Misc updates on map and prediction code.
• Apr. 30, 2020: nuImages v0.1 code release.
• Apr. 1, 2020: Devkit v1.0.8: Relax pip requirements and reorganize prediction code.
• Mar. 24, 2020: Devkit v1.0.7: nuScenes prediction challenge code released.
• Feb. 12, 2020: Devkit v1.0.6: CAN bus expansion released.
• Dec. 11, 2019: Devkit v1.0.5: Remove weight factor from AMOTA tracking metrics.
• Nov. 1, 2019: Tracking eval code released and detection eval code reorganized.
• Jul. 1, 2019: Map expansion released.
• Apr. 30, 2019: Devkit v1.0.1: loosen PIP requirements, refine detection challenge, export 2d annotation script.
• Mar. 26, 2019: Full dataset, paper, & devkit v1.0.0 released. Support dropped for teaser data.
• Dec. 20, 2018: Initial evaluation code released. Devkit folders restructured, which breaks backward compatibility.
• Nov. 21, 2018: RADAR filtering and multi sweep aggregation.
• Oct. 4, 2018: Code to parse RADAR data released.
• Sep. 12, 2018: Devkit for teaser dataset released.

Devkit setup

We use a common devkit for nuScenes and nuImages. The devkit is tested for Python 3.6 and Python 3.7. To install Python, please check here.

Our devkit is available and can be installed via pip :

pip install nuscenes-devkit

For an advanced installation, see installation for detailed instructions.

nuImages

nuImages is a stand-alone large-scale image dataset. It uses the same sensor setup as the 3d nuScenes dataset. The structure is similar to nuScenes and both use the same devkit, which make the installation process simple.

nuImages setup

To download nuImages you need to go to the Download page, create an account and agree to the nuScenes Terms of Use. For the devkit to work you will need to download at least the metadata and samples, the sweeps are optional. Please unpack the archives to the /data/sets/nuimages folder *without* overwriting folders that occur in multiple archives. Eventually you should have the following folder structure:

/data/sets/nuimages
    samples	-	Sensor data for keyframes (annotated images).
    sweeps  -   Sensor data for intermediate frames (unannotated images).
    v1.0-*	-	JSON tables that include all the meta data and annotations. Each split (train, val, test, mini) is provided in a separate folder.

If you want to use another folder, specify the dataroot parameter of the NuImages class (see tutorial).

Getting started with nuImages

Please follow these steps to make yourself familiar with the nuImages dataset:

• Get the nuscenes-devkit code.
• Run the tutorial using:

jupyter notebook $HOME/nuscenes-devkit/python-sdk/tutorials/nuimages_tutorial.ipynb

• See the database schema and annotator instructions.

nuScenes

nuScenes setup

To download nuScenes you need to go to the Download page, create an account and agree to the nuScenes Terms of Use. After logging in you will see multiple archives. For the devkit to work you will need to download all archives. Please unpack the archives to the /data/sets/nuscenes folder *without* overwriting folders that occur in multiple archives. Eventually you should have the following folder structure:

/data/sets/nuscenes
    samples	-	Sensor data for keyframes.
    sweeps	-	Sensor data for intermediate frames.
    maps	-	Folder for all map files: rasterized .png images and vectorized .json files.
    v1.0-*	-	JSON tables that include all the meta data and annotations. Each split (trainval, test, mini) is provided in a separate folder.

If you want to use another folder, specify the dataroot parameter of the NuScenes class (see tutorial).

nuScenes-lidarseg

In August 2020 we published nuScenes-lidarseg which contains the semantic labels of the point clouds for the approximately 40,000 keyframes in nuScenes. To install nuScenes-lidarseg, please follow these steps:

• Download the dataset from the Download page,
• Extract the lidarseg and v1.0-* folders to your nuScenes root directory (e.g. /data/sets/nuscenes/lidarseg, /data/sets/nuscenes/v1.0-*).
• Get the latest version of the nuscenes-devkit.
• If you already have a previous version of the devkit, update the pip requirements (see details): pip install -r setup/requirements.txt
• Get started with the tutorial.

Prediction challenge

In March 2020 we released code for the nuScenes prediction challenge. To get started:

• Download the version 1.2 of the map expansion (see below).
• Download the trajectory sets for CoverNet from here.
• Go through the prediction tutorial.
• For information on how submissions will be scored, visit the challenge website.

CAN bus expansion

In February 2020 we published the CAN bus expansion. It contains low-level vehicle data about the vehicle route, IMU, pose, steering angle feedback, battery, brakes, gear position, signals, wheel speeds, throttle, torque, solar sensors, odometry and more. To install this expansion, please follow these steps:

• Download the expansion from the Download page,
• Extract the can_bus folder to your nuScenes root directory (e.g. /data/sets/nuscenes/can_bus).
• Get the latest version of the nuscenes-devkit.
• If you already have a previous version of the devkit, update the pip requirements (see details): pip install -r setup/requirements.txt
• Get started with the CAN bus readme or tutorial.

Map expansion

In July 2019 we published a map expansion with 11 semantic layers (crosswalk, sidewalk, traffic lights, stop lines, lanes, etc.). To install this expansion, please follow these steps:

• Download the expansion from the Download page,
• Extract the contents (folders basemap, expansion and prediction) to your nuScenes maps folder.
• Get the latest version of the nuscenes-devkit.
• If you already have a previous version of the devkit, update the pip requirements (see details): pip install -r setup/requirements.txt
• Get started with the map expansion tutorial. For more information, see the map versions below.

Map versions

Here we give a brief overview of the different map versions:

• v1.3: Add BitMap class that supports new lidar basemap and legacy semantic prior map. Remove one broken lane.
• v1.2: Expand devkit and maps to include arcline paths and lane connectivity for the prediction challenge.
• v1.1: Resolved issues with ego poses being off the drivable surface.
• v1.0: Initial map expansion release from July 2019. Supports 11 semantic layers.
• nuScenes v1.0: Came with a bitmap for the semantic prior. All code is contained in nuscenes.py.

Getting started with nuScenes

Please follow these steps to make yourself familiar with the nuScenes dataset:

• Read the dataset description.
• Explore the lidar viewer and videos.
• Download the dataset.
• Get the nuscenes-devkit code.
• Read the online tutorial or run it yourself using:

jupyter notebook $HOME/nuscenes-devkit/python-sdk/tutorials/nuscenes_tutorial.ipynb

• Read the nuScenes paper for a detailed analysis of the dataset.
• Run the map expansion tutorial.
• Take a look at the experimental scripts.
• For instructions related to the object detection task (results format, classes and evaluation metrics), please refer to this readme.
• See the database schema and annotator instructions.
• See the FAQs.

Known issues

Great care has been taken to collate the nuScenes dataset and many users have praised the quality of the data and annotations. However, some minor issues remain:

Maps:

• For singapore-hollandvillage and singapore-queenstown the traffic light 3d poses are all 0 (except for tz).
• For boston-seaport, the ego poses of 3 scenes (499, 515, 517) are slightly incorrect and 2 scenes (501, 502) are outside the annotated area.
• For singapore-onenorth, the ego poses of about 10 scenes were off the drivable surface. This has been resolved in map v1.1.
• Some lanes are disconnected from the rest of the lanes. We chose to keep these as they still provide valuable information.

Annotations:

• A small number of 3d bounding boxes is annotated despite the object being temporarily occluded. For this reason we make sure to filter objects without lidar or radar points in the nuScenes benchmarks. See issue 366.

Citation

Please use the following citation when referencing nuScenes or nuImages:

@article{nuscenes2019,
  title={nuScenes: A multimodal dataset for autonomous driving},
  author={Holger Caesar and Varun Bankiti and Alex H. Lang and Sourabh Vora and 
          Venice Erin Liong and Qiang Xu and Anush Krishnan and Yu Pan and 
          Giancarlo Baldan and Oscar Beijbom},
  journal={arXiv preprint arXiv:1903.11027},
  year={2019}
}