PR 254922 Why does filtering this scan data induce so much more error?

huangyhg

Why does filtering this scan data induce so much more error?
<< Don Ruggieri  --  05/13/08  10:31:48>>
data is in 05132008_02a.zip.  Why does filtering this scan data induce so much more error?
Steps to reproduce:
This is a prototype ISO10360-4 program I've used in the past to perform some tests of various scan functions.
In this instance I am taking the 4 original scans and performing a Gauss filter on them.  Before this filtering the profile deviations are in the range 0.008mm but as high as 0.800 mm.
Results:
After the Gauss filter is applied the deviations are typcailly overr 4mm!  This is huge!  Why does th Gauss filter induce so much more rrror?
<<END>>

<< Don Ruggieri  --  05/30/08  09:41:00>>
Using the v43b from the May 28 build, this works nicely.  I will go ahead and close this report.
<<END>>
<< Changes made by Don Ruggieri -- 05/30/08  09:42:49>>
Status:  RESOLVED to CLOSED
<<END>>
<< Dan Zwick  --  05/22/08  13:09:58>>
This has been fixed along with PR 255080. Files uploaded:
V43B\CONSFEAT\FILTEROB.CPP
V43B\BASESCAN\BASICSCN.CPP
I'll upload to 4.4B when I am able to verify the build (debug and release).
<<END>>
<< Changes made by Dan Zwick -- 05/22/08  13:11:47>>
Action:  Dan Zwick to Don Ruggieri, Status:  OPEN to RESOLVED
<<END>>
<< Dan Zwick  --  05/22/08  08:22:00>>
The data coming into the filter routine is ballcenter data, but the probe radii are all 0. The filtered data is not much different from the input data but the filtered scan has hits with probe radius 0, so when they get displayed as centroids they are about 16.49 mm from the sphere center instead of about 12.49 mm. I'm not sure why the probe radius is 0 coming into the filter routine, but it's probably a scan issue. I'll look into it a bit more.
<<END>>
<< Don Ruggieri  --  05/14/08  11:19:53>>
This is apparently because of probe compensation being disabled for this proces, but why and how?  I could not find any option to ensure it remained enabled.
<<END>>
<< Changes made by Tim Wernicke -- 05/13/08  09:34:32>>
Action:  Andy Roberts to Dan Zwick, Assigned:  Andy Roberts to Dan Zwick, Priority:  to Critical
<<END>>