Handles most of the ROS-based communication. More...

#include <openni_listener.h>

Public Slots
void	getOneFrame ()
	Process a single incomming frame. Useful in pause-mode for getting one snapshot at a time.
void	toggleBagRecording ()
void	togglePause ()
	Switch between processing or ignoring new incoming data.
Signals
void	bagFinished ()
void	newDepthImage (QImage)
	Connect to this signal to get up-to-date depth images from the listener.
void	newFeatureFlowImage (QImage)
	Connect to this signal to get up-to-date featureFlow visualizations from the listener.
void	newTransformationMatrix (QString)
	Connect to this signal to get the transformation matrix from the last frame as QString.
void	newVisualImage (QImage)
	QT Stuff, to communicate with the gui.
void	setGUIInfo (QString message)
	Set the info label on the right side in the statusbar of the GUI.
void	setGUIStatus (QString message)
	Set the temporary status-message in the GUI.
Public Member Functions
void	kinectCallback (const sensor_msgs::ImageConstPtr &visual_img, const sensor_msgs::ImageConstPtr &depth_img, const sensor_msgs::PointCloud2ConstPtr &point_cloud)
	Listen to kinect data, construct nodes and feed the graph manager with it.
void	loadBag (const std::string &filename)
	Load data from bag file.
void	noCloudCallback (const sensor_msgs::ImageConstPtr &visual_img_msg, const sensor_msgs::ImageConstPtr &depth_img_msg, const sensor_msgs::CameraInfoConstPtr &cam_info_msg)
	For this callback the point cloud is not required.
	OpenNIListener (ros::NodeHandle nh, GraphManager *g_mgr)
	Ctor: setup synced listening to ros topics (kinect/stereo data) and prepare the feature handling.
void	stereoCallback (const sensor_msgs::ImageConstPtr &visual_img_msg, const sensor_msgs::PointCloud2ConstPtr &point_cloud)
	No depth image but pointcloud, e.g., for stereo cameras.
	~OpenNIListener ()
	Delete tflistener, shutdown ros publishers.
Protected Member Functions
void	callProcessing (cv::Mat gray_img, Node *node_ptr)
	Call processNode either regularly or as background thread.
void	cameraCallback (cv::Mat visual_img, pointcloud_type::Ptr point_cloud, cv::Mat depth_mono8_img)
	common processing
QImage	cvMat2QImage (const cv::Mat &image, unsigned int idx)
	Create a QImage from image. The QImage stores its data in the rgba_buffers_ indexed by idx (reused/overwritten each call)
QImage	cvMat2QImage (const cv::Mat &channel1, const cv::Mat &channel2, const cv::Mat &channel3, unsigned int idx)
	Create a QImage from image. The QImage stores its data in the rgba_buffers_ indexed by idx (reused/overwritten each call)
void	noCloudCameraCallback (cv::Mat visual_img, cv::Mat depth, cv::Mat depth_mono8_img, std_msgs::Header depth_header, const sensor_msgs::CameraInfoConstPtr &cam_info)
	as cameraCallback, but create Node without cloud
void	processNode (Node *new_node)
	processNode is called by cameraCallback in a separate thread and after finishing visualizes the results
void	retrieveTransformations (std_msgs::Header depth_header, Node *node_ptr)
	Retrieve the transform between the lens and the base-link at capturing time.
Protected Attributes
rosbag::Bag	bag
QMutex	bagfile_mutex
BagSubscriber < sensor_msgs::CameraInfo > *	cam_info_sub_
message_filters::Subscriber < sensor_msgs::CameraInfo > *	cinfo_sub_
message_filters::Subscriber < sensor_msgs::PointCloud2 > *	cloud_sub_
int	data_id_
BagSubscriber < sensor_msgs::Image > *	depth_img_sub_
cv::Mat	depth_mono8_img_
message_filters::Subscriber < sensor_msgs::Image > *	depth_sub_
cv::Ptr< cv::FeatureDetector >	detector_
cv::Ptr< cv::DescriptorExtractor >	extractor_
bool	first_frame_
QFuture< void >	future_
bool	getOneFrame_
GraphManager *	graph_mgr_
std::string	image_encoding_
message_filters::Synchronizer < KinectSyncPolicy > *	kinect_sync_
message_filters::Synchronizer < NoCloudSyncPolicy > *	no_cloud_sync_
bool	pause_
BagSubscriber < sensor_msgs::Image > *	rgb_img_sub_
std::vector< cv::Mat >	rgba_buffers_
bool	save_bag_file
message_filters::Synchronizer < StereoSyncPolicy > *	stereo_sync_
tf::TransformBroadcaster	tf_br_
	this being a pointer saves the include (using the above forward declaration)
ros::Publisher	tf_pub_
tf::TransformListener *	tflistener_
message_filters::Subscriber < sensor_msgs::Image > *	visua_sub_
cv::Mat	visualization_depth_mono8_img_
	The depth mono img is stored here for visualization purposes.
cv::Mat	visualization_img_
	The visual img is stored here for visualization purposes here.

Detailed Description

Handles most of the ROS-based communication.

The purpose of this class is to listen to synchronized image pairs from the kinect convert them to opencv, process them, convert them to a qt image and send them to the mainwindow defined in qtcv.h/.cpp

Definition at line 77 of file openni_listener.h.

Constructor & Destructor Documentation

OpenNIListener::OpenNIListener	(	ros::NodeHandle	nh,
		GraphManager *	g_mgr
	)

Ctor: setup synced listening to ros topics (kinect/stereo data) and prepare the feature handling.

Constructor: The listener needs to know the node handle and the GraphManager instance. Which topics to listen to and which feature detector/extractor to use is queried from the parameter server.

Definition at line 55 of file openni_listener.cpp.

OpenNIListener::~OpenNIListener ( )

Delete tflistener, shutdown ros publishers.

Definition at line 342 of file openni_listener.cpp.

Member Function Documentation

void OpenNIListener::bagFinished ( ) [signal]

void OpenNIListener::callProcessing	(	cv::Mat	gray_img,
		Node *	node_ptr
	)		`[protected]`

Call processNode either regularly or as background thread.

Definition at line 546 of file openni_listener.cpp.

void OpenNIListener::cameraCallback	(	cv::Mat	visual_img,
		pointcloud_type::Ptr	point_cloud,
		cv::Mat	depth_mono8_img
	)		`[protected]`

common processing

Definition at line 489 of file openni_listener.cpp.

QImage OpenNIListener::cvMat2QImage	(	const cv::Mat &	image,
		unsigned int	idx
	)		`[protected]`

Create a QImage from image. The QImage stores its data in the rgba_buffers_ indexed by idx (reused/overwritten each call)

Create a QImage from one image. The QImage stores its data in the rgba_buffers_ indexed by idx (reused/overwritten each call)

Definition at line 674 of file openni_listener.cpp.

QImage OpenNIListener::cvMat2QImage	(	const cv::Mat &	channel1,
		const cv::Mat &	channel2,
		const cv::Mat &	channel3,
		unsigned int	idx
	)		`[protected]`

Create a QImage from image. The QImage stores its data in the rgba_buffers_ indexed by idx (reused/overwritten each call)

Create a QImage from several one-channel images. The QImage stores its data in the rgba_buffers_ indexed by idx (reused/overwritten each call)

Definition at line 642 of file openni_listener.cpp.

void OpenNIListener::getOneFrame ( ) [slot]

Process a single incomming frame. Useful in pause-mode for getting one snapshot at a time.

Definition at line 638 of file openni_listener.cpp.

void OpenNIListener::kinectCallback	(	const sensor_msgs::ImageConstPtr &	visual_img,
		const sensor_msgs::ImageConstPtr &	depth_img,
		const sensor_msgs::PointCloud2ConstPtr &	point_cloud
	)

Listen to kinect data, construct nodes and feed the graph manager with it.

For each dataset from the kinect, do some data conversion, construct a node, hand it to the graph manager and do some visualization of the result in the GUI and RVIZ.

Definition at line 435 of file openni_listener.cpp.

Here is the call graph for this function:

void OpenNIListener::loadBag ( const std::string & filename )

Load data from bag file.

This function reads the sensor input from a bagfile specified in the parameter bagfile_name. It is meant for offline processing of each frame

Definition at line 152 of file openni_listener.cpp.

void OpenNIListener::newDepthImage ( QImage ) [signal]

Connect to this signal to get up-to-date depth images from the listener.

void OpenNIListener::newFeatureFlowImage ( QImage ) [signal]

Connect to this signal to get up-to-date featureFlow visualizations from the listener.

void OpenNIListener::newTransformationMatrix ( QString ) [signal]

Connect to this signal to get the transformation matrix from the last frame as QString.

void OpenNIListener::newVisualImage ( QImage ) [signal]

QT Stuff, to communicate with the gui.

Connect to this signal to get up-to-date optical images from the listener

void OpenNIListener::noCloudCallback	(	const sensor_msgs::ImageConstPtr &	visual_img_msg,
		const sensor_msgs::ImageConstPtr &	depth_img_msg,
		const sensor_msgs::CameraInfoConstPtr &	cam_info_msg
	)

For this callback the point cloud is not required.

Definition at line 346 of file openni_listener.cpp.

void OpenNIListener::noCloudCameraCallback	(	cv::Mat	visual_img,
		cv::Mat	depth,
		cv::Mat	depth_mono8_img,
		std_msgs::Header	depth_header,
		const sensor_msgs::CameraInfoConstPtr &	cam_info
	)		`[protected]`