jhallard/PointCloudProcessor

This project serves to create a group of classes that simplify and abstract the overall process of using a Kinect camera to load, process, visualize, save, and publish PointCloud objects. These classes are meant to be used by other PointCloud processing programs to simplify the setup process and increase productivity. This with a few object instantiations and member function calls one can obtain XYZ-RGBA data from a Kinect camera in a continuous stream, converting this raw data into Point Cloud objects, allow the user to visualize this incoming data stream, and allow them to save and publish the data for other programs to use. These classes rely on ROS-Hydro, OpenNI, and the pcl_ros libraries includes in the ROS-Hydro desktop build.

PointCloudProcessor

This program is a real attempt to make a program that combined the Point Cloud Library, OpenNI, OpenCV, and ROS. It's goal is to design modular and reusable classes and files to import data from a kinect or other range camera, then convert this data into a point cloud. This point cloud is displayed in real time to the user, who can then choose to save the current point cloud for editing. When the user saves the current PC, the data gets piped to another program via the ROS publish and subscription functions. This next program will import the PCL data and allow the user to run this data through various filters and statistical outlier removers.