Added more ROS logging messages to document the stages of setting up the data feed, visualizer, and publisher. Helps with debugging.

Added more ROS logging messages to document the stages of setting up the data feed, visualizer, and publisher. Helps with debugging.

Found some possible memory leaks and cleared them up. Updated the constructor to include an argument to let the user decide if they want to publish teh incoming data via a rOS publisher or if they would also like to visualize that data in real time with a cloud viewer object.

Changed around some formatting

Changed around the way ROS interacts with the program. The CloudGrabber class now has it''s own spin() call inside to keep the thread going, so no longer is a ros::spin() call needed inside the file where the class is instantiated. Also, we removed the separability of the publishing and data grabbing/visualization functionality. Now the user just calls CloudGrabber::start() and this will call startFeed() and startPublishing(). The program is now set up to publish at user defined intervals, instead of just when they push a button. The publishing can be toggled on or off by pressing the ''p'' button with focus on the visualization window.

Formatting updates

Took out some un-needed ro::spin() calls and added to the code documentation in CloudGrabber.cpp file

Publisher functionality fixed to publish at a user defined interval (or a default 10 times per second), this can be toggled by pressing the ''p'' key. The default state is don''t publish so press ''p'' to start the publishing after calling startPublisher

Once again updated documentation, no code logic changes

Updated CloudGrabber class to accept a parameter to the CloudGrabber::startPublisher function that specifies the millisecond delay for publishing the incoming Point Cloud data. This allows a continuous stream of data to be published without the user having to press the ''p'' key continuously as was previous. I think next I will let the user use the ''p'' key to toggle th publishing of data, instead of using it as a ''send'' button for the publisher

Once again upgrades code documentation to describe pre-conditions for using this class and what the various functions accomplish

Extensively updated inline code documentation. Explained exactly what the CloudGrabber class needs the user to do before it can gather and publish data from a kinect camera. Changed the ros::Node variable from having to be declared by the user and passed into the class to being a member field that the user can get() if they need it. It is currently only publishing data when the user presses the ''p'' key, next step is to change this to having it constantly output data at a rate defined by the user. Program is working quite well to be honest

Extensively updated inline code documentation. Explained exactly what the CloudGrabber class needs the user to do before it can gather and publish data from a kinect camera. Changed the ros::Node variable from having to be declared by the user and passed into the class to being a member field that the user can get() if they need it. It is currently only publishing data when the user presses the ''p'' key, next step is to change this to having it constantly output data at a rate defined by the user. Program is working quite well to be honest

Final commit for the day. Cleaned up the code formatting and extending the in code documentation so people understand exactly what part of the code does what.

finally, after much fucking headache I got it to successfully publish pcl data through a ROS messanger. The main fristration came from the fact that sensor_msgs::Pointcloud2 is depricated, and toROSMsg is depricated, so we instead have to use pcl::toPCLPointCloud2, and we have to convert our point clouds into a PCLPointcloud2 type, which is a type of message that ROS can use. Also, we had to include . Also we had to changet he type of data we are publishing from Pointcloud2 to PCLPointCloud2 inside the advertise call in startPublishing

Changed around some thing. 1.) The key to save is now ''s'', press it to save the currently displayed point cloud. Press ''v'' to toggle the visualize flag. It starts out as true, which means the program actively visualizes incoming PCL data from the kinect. If you press ''v'' to toggle it off, the visualizer will pause on the current point cloud and allow you to analyze it visually, where you can then save it. Updated some of the order of the ROS publishing, all ROS stuff is inside the class except for the ros::start() and the creation of the ros::Node, which are both in the main function

Added ROS publishing traits into the CloudGrabber class with the function CloudGrabber::startPublishing(). This will run in it''s own loop for as long as ros::spin() is still running in main(). This abstracts the publishing of data away from the main file where we can concentrate on more important tasks. Currently it is only publishing string data to another program (test_sub) so the next thing we need to do is make it publish PCL::PointCloud2 data types, then we can really get started working with point cloud filtering and registering

Fully integrated with ROS now. The class CloudGrabber can now be created, and when CloudGrabber::startFeed() the class will start a data feed on its own thread, converting kinect data to PCD data and visualizaing it. The class is now ready to add functions to publish the data so that other programs can subscribe to the data feed and do with it as they wish. Cod formatting and documentation has also been updated and improved

Got ROS working inside of the program. ROS is now handling the main loop and the program is gathering data through the kinnect camera while ros keeps the thread going. Camera data is coming in smoothly and without much lag, we are ready to move on to publishing the PCL data through ROS to another program that will run filters and tests on it

Initial commit