omnigrid

Course project for CMU 16-825 - Occupancy Grid prediction using Multi-View Fisheye cameras

Preprocessing

ros2bag setup

These instructions are for processing the bag files in ROS2. If not needed, these steps can be skipped.

sudo apt-get install ros-$ROS_DISTRO-rosbag2

Next, get the rosbags python library by installing them from requirements.txt or by running the following command:

pip install rosbags
pip install rosbags-image

For future reference, the rosbags library can be found here. This library is convenient as we don't need to setup a ROS2 environment to process the bag files.

Usage

First convert the ROS1 bag files to ROS2 using the following command:

rosbags-convert lidar_points-007.bag

This will generate a folder with a sqlite database and metadata. Copy the folder to the data directory (create if it doesn't exist) and update the path in config.py.

Then run the following script to extract the lidar data.

# From the root directory
mkdir -p data/lidar
python scripts/lidar_extractor.py

This will take a couple of minutes and will extract the lidar data to the data directory.