Hello,
i am a new user (currently using the trial) and doing some tests using gogle earth (all this stuff is for personal testing so i don't think there are license issues) to rebuild part of my hometown (actually, to test how/if it would work for KAP and how many photos i'd need).
Now i have an issue when building the dense cloud, and hopefully some of you can help with some tips: in attempt i took about 1000 screenshots flying around town in GE (in my previous tests i used smaller 200 photos chunks). After aligment (took 48+ hrs) this is how caeras and tie points look:
https://dl.dropboxusercontent.com/u/218516/PSissue_tiepoints-cameras.JPG..and this are the tiepoints:
https://dl.dropboxusercontent.com/u/218516/PSissue_tiepoints.JPGNow, usually i would just go to the next workflow step, and build the dense cloud and have no issue at all.
But this time when i try to build it, no matter which parameters i set, it just selects
49 cameras out of the 1013 aligned cameras, and outputs this incomplete cloud:
https://dl.dropboxusercontent.com/u/218516/PSissue_dense.JPGI am relly clueless, as i just used the same workflow used in other tests (here i just have a larger workspace and more cameras).. and never got such a problem.
Is it maybe due to too many cameras in a single chunk? I don't get any errors inthe console (i have now 32gb of ram so the pc should handle it), and i would like to use a single chunk so i don't need to re-align stuff later. Here's the last part of the console log:
Total performance: 841.835 million samples/sec
Generating dense point cloud...
selected 49 cameras in 0.047 sec ----> WHY ONLY 49?!?!
working volume: 5303x4035x732
tiles: 1x1x1
selected 49 cameras
preloading data... done in 0.047 sec
filtering depth maps... done in 1.827 sec
preloading data... done in 1.139 sec
accumulating data... done in 0.265 sec
building point cloud... done in 0.25 sec
2105903 points extracted
Finished processing in 8.553 sec (exit code 1)I think the issue is leaving the key point limit at 40000 and the tie point limit at the defsult 1000, i *think* setting the tie point limit to "0" could solve the issue. Before i run another 48hrs test, i'd really like some feedback, ideas/tips.
Thanks a lot!!!