tayyab's profile - activity

Hi Caroline, Yes, I managed to do it back then. You can find the respective piece of code here. http://tayyabnaseer.blogspot.de/2012/04/running-ros-openni-tracker-with.html

I managed to do it back then. You can find the respective piece of code here.

I am following this tutorial to run the simulator for the quadcopter. The only difference which i have is that i have a normal gaming joystick by SONY. Joystick is recognized by the system and also publishes axis and button info on when i run

1. rosrun joy joy_node
2. I can also see graphs of axes and buttons when i run roslaunch flyer_common bringup_sim.launch

but when i do this i get the following error and everything else mentioned in the tutorial appears as it is but I can't fly because of this error i guess.

Client [/simflyer1/manager] wants topic /simflyer1/joy to have datatype/md5sum [joy/Joy/e3ef016fcdf22397038b36036c66f7c8], but our version has [sensor_msgs/Joy/5a9ea5f83505693b71e785041e67a8bb]. Dropping connection

Any help would be appreciated thanks :)

So there is a weird behavior I am encountering. First I launch by Openni Camera driver

roslaunch camera_launch openni.launch

and then i run the openni tracker and the depth stream just gets flipped across (mirrored) .Did anyone notice this or encountered it and how its going to affect the transforms.

Before running the tracker

http://dl.dropbox.com/u/68156288/right_depth.png

After running the tracker

http://dl.dropbox.com/u/68156288/left_depth.png

Anyway to correct this ?

thanks

@GermanUser thanks for your help. The link which you gave is my own blog :) and in this post i have mentioned that this technique works as an alternative. In this post I explained how to run ros openni tracker with recorded '.oni' files not with bag files.

Thanks for the help though :)

@quimnuss you can do something like

cv::Mat image; // fill in image with your image data

CvImagePtr img_ptr;

image.copyto(img_ptr->image);

link text

Hi, I have a bag file which has /camera/depth_registered/image_rect_raw. I was thinking to receive data and create depth nodes from this topic in ros openni tracker rather than the live data from the sensor. As far as i have investigated running on the recorded data in Openni works for only '.oni' files. ROS openni tracker works well with recorded oni file but how to make it run on the raw depth data from a bag file is the question. One option could be to run the NiViewer and record the data in an .oni file in parallel but it does not seem to be a nice solution as it will make the frame rate to drop low.

Any help would be appreciated thanks.

If rviz gives you an error you can always try removing the temporary folder created by rviz. considering you are in home directory

rm -rf .rviz

Now restart the rviz.

If kinect_detect gives you a segmentation fault even after the above changes, then just re build it once again by assuming you are in ppl_detection package path

make

now do

rosrun ppl_detection kinect_detect

hope it helps :)

I have been successful in running the OpenNI ROS tracker on recorded .ONI files. This is what you need to do You need to make the following changes in your openni_tracker.cpp

Note: I had created a ros package mimicking the openni_tracker {i.e with same dependencies}. Then i created an Eclipse project for it. It makes debugging and changing the code easier.

// Declare .oni player

xn::Player g_Player;

Instead of

/* string configFilename = ros::package::getPath("openni_tracker") + "/openni_tracker.xml"; XnStatus nRetVal = g_Context.InitFromXmlFile(configFilename.c_str()); CHECK_RC(nRetVal, "InitFromXml");

*/

// added for Reading '.oni' file

XnStatus nRetVal = g_Context.Init();
CHECK_RC(nRetVal, "Init");
nRetVal = g_Context.OpenFileRecording("your .oni file", g_Player);
if (nRetVal != XN_STATUS_OK)
{
printf("Can't open recording %s\n", xnGetStatusString(nRetVal));
return 1;
}

make

run openni_tracker

rosrun rviz rviz

Hopefully you will see the skeleton.

For further details you could check my blog link text

I have been successful in running the OpenNI ROS tracker on recorded .ONI files. This is what you need to do You need to make the following changes in your openni_tracker.cpp

Note: I had created a ros package mimicking the openni_tracker {i.e with same dependencies}. Then i created an Eclipse project for it. It makes debugging and changing the code easier.

// Declare .oni player

xn::Player g_Player;

Instead of

/* string configFilename = ros::package::getPath("openni_tracker") + "/openni_tracker.xml"; XnStatus nRetVal = g_Context.InitFromXmlFile(configFilename.c_str()); CHECK_RC(nRetVal, "InitFromXml");

*/

// added for Reading '.oni' file

XnStatus nRetVal = g_Context.Init();
CHECK_RC(nRetVal, "Init");
nRetVal = g_Context.OpenFileRecording("your .oni file", g_Player);
if (nRetVal != XN_STATUS_OK)
{
printf("Can't open recording %s\n", xnGetStatusString(nRetVal));
return 1;
}

make

run openni_tracker

rosrun rviz rviz

Hopefully you will see the skeleton.

For further details you could check my blog link text

Ideally you should not be required to run any node or launch file to get openni tracker started. ROS openni tracker is a stand alone node. As far as /tf topics are concerned they will only be published once the node has started tracking the user. You can monitor the console output of the program and when it says 'started tracking user #' then /tf topics will be published.

No change is required in 'camera_frame_id' , keep it to 'openni_depth_frame' and while visualizing in rviz keep your fixed/reference frame to 'openni_depth_frame'.

Just in case you want to visualize the pointcloud/depth/rgb images along with /tf skeleton tree , just change the 'camera_frame_id'

string frame_id("openni_depth_frame");

pnh.getParam("camera_frame_id", frame_id);

to

string frame_id("openni_rgb_optical_frame"); // what ever you want here

pnh.getParam("camera_frame_id", frame_id);

Hope it helps, you might want to checkout my posts about other issues while get these things running [http://tayyabnaseer.blogspot.de/2012/04/running-ros-openni-tracker-with.html] :-)

What you can do is you can modify the openni_tracker.cpp. I have modified it as follows

string frame_id("openni_depth_frame");

pnh.getParam("camera_frame_id", frame_id);

to

string frame_id("openni_rgb_optical_frame");

pnh.getParam("camera_frame_id", frame_id);

so that you can run camera node and openni tracker together. Hope it helps.

/tf topics will be published when it has started started tracking the User. Before 'tracking' you cannot visualize anything in the rviz. Just make sure you have added the TF and select /openni_depth_frame as Fixed/Reference frame .

Second option is to use the ros openni_tracker on offline data. Record a '.oni' file with Openni Sample NiViewer and then run the ros openni_tracker on this recorded data. For implementation details please checkout my post.

[http://tayyabnaseer.blogspot.de/2012/04/running-ros-openni-tracker-with.html]

/tf topics will be published when it has started tracking the User. Before 'tracking' you cannot visualize anything in the rviz. Just make sure you have added the TF and select /openni_depth_frame as Fixed/Reference frame .

Second option is to use the ros openni_tracker on offline data. Record a '.oni' file with Openni Sample NiViewer and then run the ros openni_tracker on this recorded data. For implementation details please checkout my post.

[http://tayyabnaseer.blogspot.de/2012/04/running-ros-openni-tracker-with.html]

Try add your OS in row

rep111version_map = {'lucid':'10.04', 'maverick':'10.10', 'natty':'11.04','oneiric':'11.10'}

and in debian.py add '12':'oneiric' *

self.version_map = { '12':'oneiric', '11':'11.04', '10':'10.10', '9':'10.04', '8':'9.10', '7':'9.04', '6':'8.10', '5':'8.04'}

Worked for me, hope it helps :)

Make sure you have set the path variable and add these two lines to the top of your main python script to avoid any dependency errors.

import roslib; roslib.load_manifest('your_package_name')

then write your rest of the script. Please keep in mind that this does not need to be in every python script but just in the main entry point.