stereo_image_proc origin of stereo camera system? [closed]

asked 2013-06-28 00:56:33 -0500

Gaviria R gravatar image

updated 2014-11-18 11:02:12 -0500

130s gravatar image

I’m currently working on a stereo vision system. I have successfully implemented the stereo system using the stereo_image_proc node.

I indexed the sensor_msgs::Pointcloud2 message and got the average value for a 5 pixel window, in order to perform range measurements of a target placed in space. However from the results it seems that the origin of the stereo system is approximately 10cm in front of the stereo cameras. But I believe it should be close to the left sensors origin?

Im pretty confused as to why that is? When I perform the same indexing process with the kinect point clouds I get more accurate range measurements.

Where or what do you think is the source of this offset? does it have to do with tf?

Im confused....

edit retag flag offensive reopen merge delete

Closed for the following reason question is not relevant or outdated by tfoote
close date 2017-11-30 10:26:07.296348


Do you remember the pixel error you obtained at the calibration step? Knowing the error, the pixel size and the camera focal may give you some insight on what's happening.

Miquel Massot gravatar image Miquel Massot  ( 2014-11-24 10:43:23 -0500 )edit

To add to @Miquel Massot, make sure you've gone through the calibration process.

psammut gravatar image psammut  ( 2017-09-26 07:55:45 -0500 )edit