This repository contains the code and tools associated with our paper, "USCILab3D Dataset: A Large-scale, Long-term Outdoor Dataset."
This repository provides the code necessary to process and utilize the USCILab3D Dataset, a large-scale, long-term outdoor dataset. The dataset and associated tools are designed for research and development in the fields of computer vision, SLAM (Simultaneous Localization and Mapping), and AI.
The repository is organized into two main directories, each with its own README:
- vision_toolkit: Contains code for processing raw bagfiles, creating SLAM info, dividing sectors, and running COLMAP.
- 3d2d_ann: Contains code related to GPT-4, Grounded-SAM, and projection.
To get started, clone the repository and install the necessary dependencies:
git clone https://github.com/yourusername/USCILab3D-Dataset.git
cd USCILab3D-Dataset
# Follow the installation instructions in each subdirectory's READMEInstructions for processing raw bagfiles can be found in the vision_toolkit directory. Refer to the vision_toolkit README for detailed steps.
Instructions for creating SLAM info are also located in the vision_toolkit directory. Refer to the vision_toolkit README for detailed steps.
To divide sectors, follow the instructions provided in the vision_toolkit directory. Refer to the vision_toolkit README for detailed steps.
Instructions for running COLMAP are available in the vision_toolkit directory. Refer to the vision_toolkit README for detailed steps.
The code for GPT-4 and Grounded-SAM is located in the 3d2d_ann directory. Refer to the 3d2d_ann README for detailed steps.
Projection-related code and instructions can be found in the 3d2d_ann directory. Refer to the 3d2d_ann README for detailed steps.
This project is licensed under the MIT License.