This repository focuses on Orientation tracking based on IMU measurements with optimization of quaternions and gradient decent. It will go on to show how accurate these tracking estimates are by creating a panoramic image using images taken from the IMU body frame while rotating. The method discussed in this project can be adapted for environment mapping and live orientation estimations of a moving ridged body.
Check Project PDF for thorough explanation and results.
Orientation Estimation: Estimates the orientation of the body frame using utilizing values from the IMU.
Panoramic construction: Creates a panoramic image from images taken in the IMU body frame while rotating by utilizing the orientation estimations.
Vicon is actual orientation and optimized is the orientation estimated from IMU measurements.
Panoramic image constructed from images taken in the IMU body frame while rotating.
numpy
Data not included but shape description can be found in project pdf
pip install -r requirements.txt
Data Preparation:
Ensure your dataset is structured as described in the project report.
Run the Code:
Execute the main script: python main.ipynb.
Adjust dataset selection and other parameters as needed within the script.
main.py: Main script to run the odometry estimation and panoramic image construction.
Project1.pdf: Project documentation.
../data/: Directory containing datasets.