visualodometry's Introduction

Visual Odometry Scripts

Overview

This repository contains two Python scripts for visual odometry using different feature detection algorithms:

VO.py: Visual odometry using ORB. Please note that it is currently not fully operational.
VO2.py: Visual odometry using SIFT. This implementation is functional.

Usage

Before running the scripts, make sure to configure the necessary paths by providing the following information:

Path to Data Sequence: <path_to_data_sequence>
Path to Calibration File: <path_to_calib_file>
Path to Ground Truth: <path_to_ground_truth>
Path to Store Generated Data: <path_to_generated_data>

Example Usage for SIFT Odometry

# Visualize trajectory
evo_traj kitti VOPoses/SIFT_02.txt --ref=GTPoses/02.txt -p --plot_mode=xz

# Compute Absolute Pose Error (APE) and visualize results
evo_ape kitti GTPoses/02.txt VOPoses/SIFT_02.txt -va --plot --plot_mode xz --save_results results/SIFT02.zip

Replace the placeholder paths (<...>) with your specific file and directory paths.

Important Note

VO.py Status: The ORB-based visual odometry in VO.py is currently not fully functional.
VO2.py Status: The SIFT-based visual odometry in VO2.py is functional and ready for use.

Feel free to experiment and contribute to the improvement of the ORB-based visual odometry implementation. If you encounter any issues or have suggestions, please submit them through the repository's issue tracker.

Happy coding!

Recommend Projects