Agisoft Metashape

Agisoft Metashape => Bug Reports => Topic started by: Yoann Courtois on December 20, 2018, 06:20:41 PM

Title: Classification looks broken
Post by: Yoann Courtois on December 20, 2018, 06:20:41 PM

Ground point classification gives really bad results in Metashape (but also in PhotoScan 1.4), whereas it was really better before.
Is this normal ?

Here is a screen shot from a standard survey where we can see really random result.

Regards

Title: Re: Classification looks broken
Post by: Alexey Pasumansky on December 20, 2018, 07:29:51 PM

Hello Yoann,

Can you please specify the build number of the pre-release version that you are using and, if possible, provide the point cloud for our internal testing to support@agisoft.com in any format that is supported for the point cloud import (for example, LAS or LAZ)?

Title: Re: Classification looks broken
Post by: Yoann Courtois on December 21, 2018, 11:43:17 AM

Classification done with 1.5.0 build 7204, I'll send you the point cloud used.

Gonna try again with the very last one.

Thanks for your investigation

Title: Re: Classification looks broken
Post by: Alexey Pasumansky on December 24, 2018, 03:06:33 PM

Hello Yoann,

Which coordinate system your project is referenced in and which classification parameters you are using?

Also it seems that in the provided cloud there are some reconstruction artifacts in the middle that are related to the incorrect alignment results.

Title: Re: Classification looks broken
Post by: Yoann Courtois on December 26, 2018, 11:02:45 AM

Hello Alexey,

Even though the coordinates used for GCPs are in RGF93/CC46 system (EPSG 3946), every thing was calculated in local system. Then, no correction from projection system but il doesn't affect such small survey.

Indeed you are right. I've processed this dataset a lot of time and this one has an incorrect aligned bloc of several pictures. Nevertheless, it doesn't affect the rest of the survey either.

Even I used this example for reporting classification vagueness, I only started this topic after several experiences, using different datasets (which have been now deleted)

Regards

Title: Re: Classification looks broken
Post by: Alexey Pasumansky on December 27, 2018, 12:10:32 PM

Hello Yoann,

Seems that we have found the problem that has been causing wrong ground point classification. The fix will be implemented in the next version update.

Title: Re: Classification looks broken
Post by: Yoann Courtois on December 27, 2018, 01:22:22 PM

Thanks a lot.
May you tell me when next version update will be done ?

Regards

Title: Re: Classification looks broken
Post by: Alexey Pasumansky on December 27, 2018, 02:43:10 PM

Hello Yoann,

We are planning to release 1.5.0 by a very end of December. Not sure, yet, if before that the pre-release update will be published.