This project is a ROS package that encompasses a pipeline to detect and estimate the 6-D pose of an object in the scene. A 3D CAD model of the object is necessary for the pipeline to function.
The ROS node can do one of two things:
- Parse a video that was recorded by the Azure Kinect (mkv format), and generate a .csv containing the pose of the object in each frame.
- Parse the live stream video of a connected Azure Kinect camera, and generate a .csv containing the pose of the object in each frame.
Please refer to the usage guide for more information on how to use the node.
Please refer to the building guide