Average Tie point multiplicity parameter

[Paulo]

Hi all,

I am looking at Average Tie point multiplicity which is printed in Parameters page of report:

General   
   Cameras                       163
   Aligned cameras       162
   Markers                       11
   Coordinate system        ITRF2008 / UTM zone 14N + GGM2010 geoid height
   Rotation angles               Yaw, Pitch, Roll
Point Cloud   
   Points                                            226,305 of 478,570
   RMS reprojection error            0.130199 (0.363276 pix)
   Max reprojection error            0.503363 (1.66923 pix)
   Mean key point size                    2.75397 pix
   Point colors                            3 bands, uint8
   Key points                                   No
   Average tie point multiplicity   3.21228

Software                                         Version 1.5.3 build 8469

So average tie pt multiplicity is 3.21

But from Survey data page
 
Number of images:    163                   Camera stations:            162
Flying altitude:            137 m              Tie points:                       226 305
Ground resolution:   4.18 cm/pix   Projections:                      647 136
Coverage area:             0.307 km²   Reprojection error:    0.363 pix

I think the average point multiplicity should be :
No. Projections / No. Tie points = 647 136 / 226 305 = 2.86

Why are they different^?

[Alexey Pasumansky]

Hello Paul,

Can you send this PSZ project with only alignment results to support@agisoft.com?

Also please specify, if you are observing similar behavior in every project or not?

[Paulo]

Hello Alexey,

I think, I found the reason for this disparity.... in Metashape the average multiplicity is calculated using total original Tie points not final number after discarding those that do not meet some criteria.

That is using number of tracks (initial number of Tie points, 478 570 in provided example), and all projections based on these tracks we get the number 3.21 of example....

[Alexey Pasumansky]

Hello Paul,

Do you mean that the number is not updated after you remove some points from the tie point cloud?

[Paulo]

No they are not, Alexey

[Paulo]

Other example where I print final number of points and projections and initial number of points and projections, which explains difference in average point multiplicity.....