Jump to content
RockWare Support Forum
Frederik

problem with SOLID RESAMPLING

Recommended Posts

Although it appears that a bug fix had already been issued to solve a problem on this feature, SOLID > MATH > RESAMPLE (In RW utilities) keeps hanging endlessly, forcing me to close down the program. :( Can this be fixed please, because it's quite urgent. Thanks!

Share this post


Link to post
Share on other sites

I just tested this on my computer using the June 9th version of RockWorks14. It is working, but is very slow. For example, I am resampling a MOD file that originally had 84x92x32 nodes. I'm trying to downsize to a model with 42x46x16 nodes. This looks like its probably going to take about an hour. Have you tried running the resampling overnight?

If you need results quickly, I would recommend that you export your current solid model as an ASCII XYZ file, and then just reinterpolate the model with your new project dimensions. The Inverse Distance Anisotropic interpolation method would use the closet point within each 90 degree sector for the interpolation of the voxels in your new model. You could also reinterpolate using the Closest Point algorithm.

Our programmers will look into this speed issue to see if there is anything that we can do to speed things up.

Feel free to contact me with questions about this.

Regards,

Alison

Alison Alcott

RockWare, Inc.

alison@rockware.com

Share this post


Link to post
Share on other sites

OK, I will try your suggestions.

But I reckon there is no restriction on exceeding the current model dimensions, right? (creating new 'empty' nodes outside the original solid grid)

Share this post


Link to post
Share on other sites

No, this should be fine. You will get strange results outside of the original model area, but there are no limitations built into the program that would not allow you to do this.

Share this post


Link to post
Share on other sites

I ran the resampling overnight for more than 9 hours, but only came to 4% progress. This morning I tried using the Closest Point Interpolation, but it appears to be going the same long way. I know my model is quite big (337x185x95 resampling to 363x105x61), but I expected faster results as the spacing and node location are exactly the same for both solid grids (10x10x0.25), so the closest point is the value already lying on the respective node location.

This brings me to a another possibility: Can I in some way just remove horizontal and vertical grid slices from the solid model to 'clip' it to the right dimensions? This would be the fastest solution as spacing is the same in all directions and the nodes already overlap. (Bear in mind this is not the same as filling redundant grid slices with null values.)

Share this post


Link to post
Share on other sites

I guess I've found a simple and fast solution to the above problem by exporting my model to an ascii file and importing it again after changing the headers to fit the required model dimensions. This requires RockWorks to discard all the nodes that lie outside these dimensions without giving distorted results, but it seems to do well. Can you confirm there is no influence of these 'out of range' nodes?

Thx,

Frederik

Share this post


Link to post
Share on other sites

Hi Frederick:

I'm not convinced that this will work. Let look into this and get back to you. Until then, I would not assume that this approach will work.

Alison

Share this post


Link to post
Share on other sites

Frederick:

One of our programmers worked on this yesterday and we now have a drastically improved solid resampling tool in our test build. This should be included in the next update that we post on our website. If you need this improvement now, you can contact me via email and arrange for a build to be posted on our ftp site for you. My email address is: alison@rockware.com.

Thanks,

Alison

Share this post


Link to post
Share on other sites

Alison,

I did manage to solve the problem like I explained above. I did a thorough visual comparison between the 2 models and they apppear identical for their overlapping volume.

So I guess when resampling only implies the change of model extents, (keeping the same node spacing and locations), this method provides the fastest solution :)

Thx,

Frederik

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×