Modify local coordinates to export ortho and DEM that matches 2 time points

[Carol_Coral]

Dear Metashapers,
I'm fairly new to Metashape so I hope the experts can help me solve my problem.
I'm using structure from motion to reconstruct a squared coral nursery (4m x 4m). I have imagery for 2 different time points and I would like to be able to set the exact same coordinate system so I can easily track/measure the growth/death of each of the colonies. However, I do not have GPS coordinates for GCPs because it is quite difficult to get an accurate measure for this underwater structure.
Since the imagery was taken 4 months apart, I have already built the models for the first time point. When I built the models of the second time point I found out that the Orthomosaic and DEM of the two time points are exported in a different coordinate system and I cannot fate track the colonies easily in a secondary program (Taglab).

To address this problem I've used two different approaches:

1. I've tried to align the two different time point chunks using GCPs. The resulting alignment is pretty good, but the alignment does not modify the local coordinates of the actual chunks (e.g. point 1 has different coordinates X and Y in each of the two time point chunks). Therefore, when I export the Ortho and DEM of both time points my images are exported with different coordinates and I cannot track them easily.

2. I've built a Sparse Point Cloud using the photographs of both time points, then after filtering some of the low confidence points, I duplicate the initial chunk and then I remove the cameras of each time points respectively. Then I built the Dense Point Clouds using only the cameras that correspond to each time point, respectively. The result is exactly what I want as the coordinates are consistent between the two time points and I can effectively track my colonies. However, this method has a gigantic disadvantage in the sense that for further monitoring time points I will have to re build the models from scratch  .

I'm hoping that you could instruct me on how to impose local coordinates to GCPs to each time point chunk, so in that way I can export them with exactly the same coordinates without having to rebuild the models.

Thanks in advance

[Paulo]

Hello Carol,

if you have at least 3 common GCPs from the two time series chunks then this method should work:
- in first time aligned chunk measure the 3 GCPs and export their coordinates;
- in second time chunk, import 3 GCPs coordinates from above and measure them;
- optimize with 3 GCPs checked and voila your 2 time series models will be in same CS...

Should be quite straightforward,

[Carol_Coral]

Thank Paulo,

I tried your approach and it work well for some of my models!!

Cheers,

Carol