Vision meets Robotics: The KITTI Dataset
Andreas Geiger and Philip Lenz and Christoph Stiller and Raquel Urtasun

@article{geiger2013ijrr,
author= {Andreas Geiger and Philip Lenz and Christoph Stiller and Raquel Urtasun},
title= {Vision meets Robotics: The KITTI Dataset},
journal= {International Journal of Robotics Research (IJRR)},
year= {2013},
abstract= {From: http://www.cvlibs.net/datasets/kitti/raw_data.php
This is the file : 2011_09_29_drive_0071 (4.1 GB)  [synced+rectified data]

 This page contains our raw data recordings, sorted by category (see menu above). So far, we included only sequences, for which we either have 3D object labels or which occur in our odometry benchmark training set. The dataset comprises the following information, captured and synchronized at 10 Hz:

    Raw (unsynced+unrectified) and processed (synced+rectified) grayscale stereo sequences (0.5 Megapixels, stored in png format)
    Raw (unsynced+unrectified) and processed (synced+rectified) color stereo sequences (0.5 Megapixels, stored in png format)
    3D Velodyne point clouds (100k points per frame, stored as binary float matrix)
    3D GPS/IMU data (location, speed, acceleration, meta information, stored as text file)
    Calibration (Camera, Camera-to-GPS/IMU, Camera-to-Velodyne, stored as text file)
    3D object tracklet labels (cars, trucks, trams, pedestrians, cyclists, stored as xml file)

Here, "unsynced+unrectified" refers to the raw input frames where images are distorted and the frame indices do not correspond, while "synced+rectified" refers to the processed data where images have been rectified and undistorted and where the data frame numbers correspond across all sensor streams. For both settings, files with timestamps are provided. Most people require only the "synced+rectified" version of the files.
More detailed information about the sensors, data format and calibration can be found here:

    Preprint of our IJRR data paper
    Download the raw data development kit (1 MB)
    Download the raw dataset download script (1 MB) (thanks to Omid Hosseini for sharing!)
    Mark Muth has written a QT-based visualizer for point cloud and tracklet sequences.
    Yani Ioannou (University of Toronto) has put together some tools for working with KITTI raw data using the PCL
    Christian Herdtweck (MPI Tuebingen) has written a python parser for reading the object label XML files
    Lee Clement and his group (University of Toronto) have written some python tools for loading and parsing the KITTI raw and odometry datasets
    Tomáš Krejčí created a simple tool for conversion of raw kitti datasets to ROS bag files: kitti2bag
    Helen Oleynikova create several tools for working with the KITTI raw dataset using ROS: kitti_to_rosbag
    Mennatullah Siam has created the KITTI MoSeg dataset with ground truth annotations for moving object detection.

Note: We were not able to annotate all sequences and only provide those tracklet annotations that passed the 3rd human validation stage, ie, those that are of very high quality. For sequences for which tracklets are available, you will find the link [tracklets] in the download category.
},
keywords= {},
terms= {},
license= {Creative Commons Attribution-NonCommercial-ShareAlike 3.0 },
superseded= {},
url= {http://www.cvlibs.net/datasets/kitti/raw_data.php}
}