Rolling shutter

[arafat]

Metashape 1.8.4 allows regularized and full compensation.

I understand the difference between them, but I do not know how to choose between them,

1- Is Full always better than regularized ?
2- Is there a way (manually/automatically) to detect the need for compensation in images?

[JPG_Z]

Any news regarding this?

[Alexey Pasumansky]

Any news regarding this?

https://www.agisoft.com/forum/index.php?topic=14671.msg64383#msg64383

Regularized model is suitable for the case when the drone flies without turning the camera, i.e. during the shooting to a new route, the camera does not turn and the drone moves in parallel.
When using a Regularized, only two parameters are optimized for each frame: axis shift XY in the plane of the photo. When using the Full model, 6 parameters are optimized for each frame: shift along the XYZ axes and rotation along the XYZ axes.

[JPG_Z]

Any news regarding this?

https://www.agisoft.com/forum/index.php?topic=14671.msg64383#msg64383

Regularized model is suitable for the case when the drone flies without turning the camera, i.e. during the shooting to a new route, the camera does not turn and the drone moves in parallel.
When using a Regularized, only two parameters are optimized for each frame: axis shift XY in the plane of the photo. When using the Full model, 6 parameters are optimized for each frame: shift along the XYZ axes and rotation along the XYZ axes.

Thank you for the explanation. Perfect!

[DayGeckoArt]

Any news regarding this?

https://www.agisoft.com/forum/index.php?topic=14671.msg64383#msg64383

Regularized model is suitable for the case when the drone flies without turning the camera, i.e. during the shooting to a new route, the camera does not turn and the drone moves in parallel.
When using a Regularized, only two parameters are optimized for each frame: axis shift XY in the plane of the photo. When using the Full model, 6 parameters are optimized for each frame: shift along the XYZ axes and rotation along the XYZ axes.

I've been trying to figure out if the compensation parameters are adjusted individually per photo or just averaged for all photos. From the language in this post, it appears that they are adjusted on a per-photo basis. But I've seen other statements that say the opposite.

I ask because I fly manually. I never take any pictures with the drone rotating, always moving forward, or sometimes drifting to the side due to wind. The drone speed can vary due to wind as well. If each photo is assessed by the software independently, that's good. But if not, would that mean I need to create a "group" for each photo?

[Paulo]

Hi Day,

The rolling shutter compensation parameters are applied on a per foto basis, as seen in following screen shot....

[DayGeckoArt]

Thanks Paulo, that's what I was thinking but wasn't sure if that's what all those different numbers for each picture meant!


I've been shooting a few static photos with each set, and moving them into their own group and turning off rolling shutter compensation. Is that a good practice? Presumably the software should recognize that they aren't rolling right? But is it still better to explicitly mark them as not needing any compensation?

[Paulo]

Yes,

if you are shooting a few static and kinematic photos then moving the statics into a group and turning off rolling shutter compensation would make sense...