utiasSTARS/pykitti

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1244 stars · 240 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

pykitti is a minimal Python library for loading and working with the KITTI autonomous driving dataset, supporting raw and odometry benchmark data. It provides access to camera images, velodyne point clouds, IMU/OXTS poses, timestamps, and calibration data as convenient Python objects.

Use cases

• load KITTI dataset in python
• read KITTI calibration files
• visual odometry with KITTI data
• load velodyne point clouds python
• access KITTI stereo image pairs
• prepare KITTI data for deep learning
• parse KITTI OXTS GPS/IMU poses

When to choose

• you work with KITTI raw or odometry benchmark datasets in Python
• you need convenient access to calibration, images, point clouds, and poses
• you want generator-based sequential access for odometry or random access for training

When to avoid

• you need other KITTI benchmarks like tracking, segmentation, or flow
• you work with a different autonomous driving dataset
• you need heavy preprocessing or visualization built in

Facets

library · maturity maintenance

computer-vision data-science parser robotics computer-vision autonomous-vehicles machine-learning python cross-platform kitti-dataset dataset-loading calibration point-clouds stereo-vision odometry velodyne

Member repositories

For agents

markdown · JSON · MCP: product_card(name="utiasSTARS/pykitti")