1.0.2.2091 won't import e57 that earlier version would
Answered
Hi,
I just today imported, in the previous version of RC, an e57 containing some number of scans. Trimble scanner, intensity only, received as Recap archive, Recap can show individual scans. Exported the whole thing from recap as e57.
RC imported that. I set registration to exact, so it wasn't trying to register, but it showed a sensible result from alignment. It didn't manage to align with any photographs, but I wasn't thinking about the scan when taking photos and that might be because the photos aren't from near enough the scanner location to match up. I've never tried using an intensity scan and colour photos together, so I don't know how reliable that is.
The new version of RC won't open the e57 file at all. Console output has
"warning: Sampling estimation failed. Probably, the imported data are not a can"
repeatedly (same number of lines as number of scans in the e57, so I assume once for each scan).
I'm intrigued by this. Does this mean the e57 was never good enough for RC to use, but RC has got better at identifying this?
I didn't try reconstruction from the scans in the old version, so I only know that it could show the scanned structure after the alignment stage (and this was with exact registration, so it wasn't trying to align the scans either), so it's quite possible that reconstruction would have failed, and this test is identifying that fact earlier.
Hamish
I just today imported, in the previous version of RC, an e57 containing some number of scans. Trimble scanner, intensity only, received as Recap archive, Recap can show individual scans. Exported the whole thing from recap as e57.
RC imported that. I set registration to exact, so it wasn't trying to register, but it showed a sensible result from alignment. It didn't manage to align with any photographs, but I wasn't thinking about the scan when taking photos and that might be because the photos aren't from near enough the scanner location to match up. I've never tried using an intensity scan and colour photos together, so I don't know how reliable that is.
The new version of RC won't open the e57 file at all. Console output has
"warning: Sampling estimation failed. Probably, the imported data are not a can"
repeatedly (same number of lines as number of scans in the e57, so I assume once for each scan).
I'm intrigued by this. Does this mean the e57 was never good enough for RC to use, but RC has got better at identifying this?
I didn't try reconstruction from the scans in the old version, so I only know that it could show the scanned structure after the alignment stage (and this was with exact registration, so it wasn't trying to align the scans either), so it's quite possible that reconstruction would have failed, and this test is identifying that fact earlier.
Hamish
-
Hi Hamish Harvey
Is it possible to get the E57 for inspection ? E57 ( Faro, Lieca, Riegel etc ) are working flawlessly, when properly exported. So it looks like not good E57 export from the Recap. Send it please to milos.lukac@capturingreality.com
As for alignment of intensity laser to images, it is nearly impossible, as INTENSITY values are not just a gray color scan, but INTENSITY values which is quite different from RGB ( or grayscale images ), so features are quite different. That is why we recommend, at least for now, scanning with COLOR ( or at least to have a few scans with color which can RC align properly.
"warning: Sampling estimation failed. Probably, the imported data are not a scan"
This means that the header of the E57 is not correct (not proper format ?!?! ), so RC cannot process it properly...
As for aligning of scans with INTENSITY values, you can use control point workflow... -
Hi wishgranter,
I've emailed a link to the E57. I wouldn't have been surprised if the Recap export wasn't processable, but I'm puzzles by the change in behaviour between the last and latest versions of RC.
Re intensity - understood. I initially assumed it wouldn't work, but since the selection was there in the import dialog I thought it worth trying.
I was hoping to do a combined scan/photogrammetry process in a pitch dark space, which will mean intensity + flash on camera. I'll make sure we have good control points.
I suppose that if we have only control points linking scan and photos, what we have really is a photogrammetric model, scaled and oriented by reference to a control network - the control network could just as well be surveyed using a total station. Or can the scan contribute to the geometry between control points in this situation?
Thanks,
Hamish
Please sign in to leave a comment.
Comments
2 comments