obtaining Kinect registered images (X,Y and depth) [closed]

asked 2014-08-28 06:00:22 -0600

altella gravatar image

Hello all;

I need to obtain the organized depth information from the Kinect sensor. The topic I haven seen that provides the organized point cloud is: /kinect/depth/image_raw, with height = 480, width = 640 and encoding = 32FC1. For each pixel in the image, I need the calibrated measure (preferably in m) for X, Y and Z. For example , for pixel (320, 240) -> X=1.3 m Y=2.1 m and Z = 3.2 m

THe problem I see is that in the topic, I obtain the depth information, but, in what format?. In addition, how can I obtain the information for X and Y? This topic does not seem to be providing it....

I need this information because I want to use MvTec Halcon machine vision libraries from inside ROS, so I would like to avoid the use of OpenCV, or other machine vision libraries.

Thank you very much in advance.

edit retag flag offensive reopen merge delete

Closed for the following reason question is not relevant or outdated by tfoote
close date 2016-07-08 14:03:38.151001


Assuming you are running Hydro or Indigo, have you checked the documentation? There you can see which topic publishes depth images, registered or not, in metres or millimetres. It might be a good starting point.

Murilo F. M. gravatar image Murilo F. M.  ( 2014-08-28 09:14:58 -0600 )edit

The /kinect/depth/image_raw is not really a pointcloud it's a depth image. If possble, you can instead use the /kinect/depth/points topic, you can work with the pointcloud that has already been converted into real-world distances.

jarvisschultz gravatar image jarvisschultz  ( 2014-08-28 09:45:13 -0600 )edit

I just noticed that you want to use the MvTec lib, so my suggestion is probably not very helpful!

jarvisschultz gravatar image jarvisschultz  ( 2014-08-28 09:46:17 -0600 )edit