-
-
Notifications
You must be signed in to change notification settings - Fork 4.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Supervoxel example updates #915
Supervoxel example updates #915
Conversation
if (!disable_transform) | ||
{ | ||
for (PointCloudT::iterator cloud_itr = cloud->begin (); cloud_itr != cloud->end (); ++cloud_itr) | ||
if (cloud_itr->z < 0) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Maybe we should throw out z == 0
as well?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I thought about it, but didn't include it because of those files we encountered a while back where people were putting points at (0,0,0) instead of nan. log(z) "converts" those points to nans =)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I also remembered those files! But you are right, log(z)
does the right thing for them.
On a separate note, what do you think about removing |
We definitely should, we can remove the same from LCCP segmentation as well, but I think there's one piece missing before we can. We either need a savePNGFile overload for PointXYZL that lets you colorize the output, or better yet, a general colorize function that creates a new cloud with the RGB field added. |
This is already available, though it uses
Maybe. Any suggestions on the signature, name, and location of such a function? |
Oh, I didn't see that. savePNGFile has different behavior for labeled points though, it saves them as 16-bit single channel images.
It's not really needed if there's already a write PNG function. |
Yep, this I will change. |
Just to be sure - we need to keep both available. Saving the label as a single channel is useful for benchmarking. |
Unfortunately with the current interface only one option can be conveniently exposed through But other options are still accessible, though you'll need to write some code, e.g.: PointCloudImageExtractorFromLabelField<PointT> pcie (COLOR_MONO);
pcl::PCLImage image;
pcie.extract (cloud, image);
savePNGFile (file_name, image); So which one should be the default you think? Mono or colorful? |
Oh, I didn't even see that new interface. I think Mono makes the most sense logically, but I'm split. Maybe put a note in the doxygen comments with the code snippet showing how to colorize labels? |
Agreed, makes sense to increase awareness of the various "extractors" that are now available. |
Added better messages to explain single camera transform, fixed bug where negative Z values were always abs'd, even if transform disabled.
Added ability to use normals provided within point cloud, and ability to disable this.
Adjusted visualization so voxels and supervoxels are displayed better.
Cleaned up command line parsing.