stereo_image_proc parameter setting problem [closed]
Hi,
I am now playing with stereo camera and want to use "stereo-image-proc" package to acquire disparity image and further 3D PCL. I find there are many lost data/holes in the output point cloud from stereo-image-proc, and try to know how to make it work well. The problem is when a surface is feature-less, due to the Block-Matching algorithm the package is using, there will be holes.
I read the tutorial "ChoosingGoodStereoParameters", download their bag file, and follow the instructions step by step. However, even with the same parameters they are saying, I still cannot get the similar good results as in the tutorial. For example, in the tutorial, both the table surface and background surface are featureless, but the tutorial shows wonderful results for both surfaces. In my trial, most of the surfaces are missing, even with the same parameters.
I am using feurte version of ros. I am wondering if the stereo_image_proc package has been modified from electric version?
How can I get the similar good results, say, for the "rotating_detergent.bag" data they are using in the "ChoosingGoodStereoParameters" tutorial?
How do you guys solve the problem of missing data?
Thank you very much.
I know this is a fairly old question, however, I also have this problem. I am running Indigo on Ubuntu 14.04. I had to fix the rosbag, but this seemed to only relate to the camera info messages.
The output from the rosbag seems fine to me, I simply just cant make the stereo_image_proc perform.