KITTI is a dataset of lidar sequences of street scenes in Karlsruhe, Germany. Labels have been added over the years since its launch in 2012.

The KITTI dataset stands as a foundational resource in the world of autonomous vehicle development. Its extensive range of sensor data and real-world scenarios makes it a unique dataset for computer vision research, particularly in the areas of autonomous driving and environmental perception.

KITTI’s rich collection of images, LiDAR scans, and GPS data captures a variety of urban and rural environments. This diversity is crucial for training robust computer vision models that can adapt to different driving conditions. The dataset provides detailed annotations, including 3D object tracking, optical flow, and stereo vision.

Also, take a look at SemanticKITTI, which extends the capabilities of the original KITTI dataset by adding a semantic layer to the LiDAR point clouds.

License: Creative Commons Attribution-NonCommercial-ShareAlike 3.0
Lidar sensor: Velodyne HDL-64E

Share this dataset