How to convert pointcloud as well as depth image of Kinect to laser scan?

Question

I am using Kinect to get the depth image and pointclouds. I actually want to convert either pointcloud or depth image to laserscan. I launched openni.launch file and then tried converting depth image to laser scan using: rosrun depthimage_to_laserscan depthimage_to_laserscan image:/camera/depth/image_raw. But when I subscribe to /scan topic I do not get any message being published. I tried the same with pointcloud to laser scan conversion but the problem is same. However, if I only subscribe to those topic messages without converting them to laser scan they seem to be published. I could even visualize them in Rviz. Do we need to do something more to convert the data to laserscan and get the converted data? I am using odroid to get openni data and subscribing to those topics in my virtual machine (that has Ubuntu trusty and ROS Indigo).

---

Originally posted by Ros User on ROS Answers with karma: 40 on 2016-07-29

Post score: 0

---

Original comments

Comment by Chrissi on 2016-07-29:
You did make sure that the depthimage_to_laserscan actually subscribes to the depth image topic, right? E.g. by looking at rqt_graph.

Comment by Ros User on 2016-07-31:
Actually, I had missed out a syntax in rosrun ("=" sign after image). Now, it works perfectly. Thanks.

Comment by Chrissi on 2016-08-01:
Glad to hear.

Comment by pallavbakshi on 2017-01-23:
hey! @rosuser I am trying to create laser scan data from Kinect v2. I am having trouble finding the right libraries. Can you help? http://answers.ros.org/question/252780/pointcloud-to-laserscanner-using-kinectv2/

Comment by Ros User on 2017-01-23:
@pallavbakshi The kinect I am using is version 1 which works with two libraries namely "openni-launch" and "freenect-launch". I suggest you to try running one of these libraries and see if that works. If these libraries do not work, you may want to try the library named "openni2-launch".

Answer 1

Hi,Ros User, I am ROS beginner, now I want to convert pointcloud to laserscan,the pointclouds recived by Kinect . I refer to the web page :http://wiki.ros.org/pointcloud_to_laserscan ,but I don't know how to build my program ? Can you help me ?

I launched openni.launch file and using: rosrun pointcloud_to_laserscan pointcloud_to_laserscan_node pointcloud:=/camera/depth_registered/points ,and rosrun rviz rviz ,but when I subscribe to /scan topic, I don't get anythings. Do you know why ?

---

Originally posted by chengwei with karma: 51 on 2016-08-01

This answer was ACCEPTED on the original site

Post score: 0

---

Original comments

Comment by Chrissi on 2016-08-01:
Depth registration is not enabled by default when launching openni_launch so there might not be a topic like /camera/depth_registered/points (try to visualise it in rviz). Instead just use /camera/depth/points. You don't need the colour information from the registered points anyway.

Comment by chengwei on 2016-08-01:
Thank you for your reply. I have modified the code: rosrun pointcloud_to_laserscan pointcloud_to_laserscan_node pointcloud:=/camera/depth/points , and I could get the /scan topic , but there is nothing in RVIZ, I could't visualize them in Rviz

Comment by chengwei on 2016-08-01:

Comment by chengwei on 2016-08-01:
This is my launch file, I don't know why there is little point in RVIZ? I know the result is wrong ,but I don't know why? can you help me ? Thank you!

Comment by Chrissi on 2016-08-01:
So you have the scan topic, which you remap to something else <remap from="scan" to="$(arg camera)/scan"/> so I guess you tried to visualise that in rviz? But you can echo it, right? Does rviz complain about anything? Most likely the tf frame you are using as a global frame is not there.

Comment by Chrissi on 2016-08-01:
By default, the global frame is /map so if you do not have a map then change the global frame to something else from the drop down menu. Have a look at the rviz user guide

Comment by Ros User on 2016-08-01:
If you get error when Fixed frame is map, you can change Fixed frame to "camera_link". It worked for me.

Comment by chengwei on 2016-08-01:
Hi, Chrissi and Ros User, Thank you for your help! I can get the laserscan points after modified the target_frame:"target_frame: camera_link". In view of the laserscan points number too little, I adjusted the parameters "angle_increment". Thanks again.

Comment by Chrissi on 2016-08-02:
You're welcome!