| Issue 4: | RunCMVS loops! | |
| 1 person starred this issue and may be notified of changes. | Back to list |
What steps will reproduce the problem? 1. Use a photoset of 30 photos 2. RunBundler.py 3. RunCMVS.py What is the expected output? What do you see instead? Expected is a finite process during RunCMVS, however RunCMVS does loop during: "subprocess.call([cmvsExecutable, "./", clustersize])" (/osmcmvs/__init__.py, line 109) What version of the product are you using? On what operating system? svn r52, both Linux and Windows Please provide any additional information below. It seems that the clustersize is way too small. Setting a clustersize of 100 (or omitting the clustersize) fixes the problem. Default is 4.
Oct 16, 2010
Project Member
#1
pmou...@gmail.com
Dec 29, 2010
Know you could specify as a parameter in how many cluster you want to divide the initial point cloud. https://code.google.com/p/osm-bundler/source/detail?r=54 Since we do not have vocation to fix the existing bug in CMVS. You could now try different value of cluster number... and find a not infinite loop solution.
Jan 10, 2012
It appears to me that osm-bundler misunderstands the call signature for CMVS. A quick fix is to use the commandline option --ClusterToCompute as the maximum number of images per cluster. The code sets this to 1 by default (line 34 in osmcmvs/__init__.py). A quick fix is to run it as RunCMVS.py --ClusterToCompute=100 (This was originally the CMVS default which is changed by osm-cmvs). I will attach a patch or code correction later. |