Utility scripts for loading, visualizing, and inspecting the KITTI-360 autonomous driving dataset.
KITTI360Scripts is a collection of Python utility scripts designed for working with the KITTI-360 autonomous driving dataset. It provides tools for loading, visualizing, and inspecting the dataset's 2D images and 3D point clouds along with their semantic and instance annotations. The scripts help researchers efficiently access this large-scale dataset for urban scene understanding research.
Computer vision researchers and autonomous driving engineers working with the KITTI-360 dataset who need to visualize annotations, load data programmatically, or develop algorithms for 2D/3D scene understanding.
These scripts provide officially supported, well-documented utilities that handle the complex data formats of KITTI-360, saving researchers time on data parsing and enabling immediate focus on algorithm development with proper visualization tools.
This repository contains utility scripts for the KITTI-360 dataset.
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Directly supports KITTI-360's complex data structures, saving researchers from writing custom parsers for annotations and point clouds.
Provides both 2D image and 3D point cloud viewers with fused semantic and instance annotations, enabling thorough scene inspection using tools like Open3D.
The `helpers/labels.py` file defines all semantic classes and mappings, ensuring consistent handling of annotations across 2D and 3D domains.
Designed to reduce data overhead with efficient loaders, allowing quick access to annotations for algorithm development rather than data wrangling.
Requires downloading the large KITTI-360 dataset separately (73.7km of data), which is time-consuming and storage-intensive.
Needs environment variable configuration (KITTI360_DATASET) and system packages like python-tk and python-qt5, which can be error-prone on some systems.
Scripts are tailored exclusively for KITTI-360, with no built-in adaptability to other datasets, restricting broader usability.
The README offers only basic usage examples, lacking in-depth tutorials, API references, or troubleshooting guides for advanced features.