Implement a projected gradient descent algorithm to track the 3-D orientation of a rotating body using readings from an inertial measurement unit (IMU). Using your orientation estimates, generate a panoramic image by stitching camera images obtained by the rotating body
- Install the requirement files in new virtual environmnet:
conda create --name env_ObjTrack
conda activate env_ObjTrack
git clone https://github.com/suryapilla/Object-Tracking.git
pip install requirements.txt
-
utilsCam.py and utils.py files have all the relevant functions required for the main function
-
setup the path as per the config.yml file in config directory
-
For training run the main function by giving the correct path to the data folder
python main.py
-
There are 3 sections in the main.py script:
-
Object tracking part
-
Plotting the roll, pitch and yaw for VICON, Unoptimized and Optimized data
-
Panorama generation
-
-
For testing run testMain.py script, with correct test data directory in path
python testMain.py
-
This is for testing data where VICON data is not available, the flow is similar to the main.py
-
Object tracking part
-
Plotting the roll, pitch and yaw for Unoptimized and Optimized data
-
Panorama generation
-