Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - JohnH1

Pages: [1]
General / main/depth_point_threshold
« on: May 04, 2022, 05:27:25 PM »
I'm trying to use the 'main/depth_point_threshold' tweak to make a dense cloud for a poorly aligned model, but Metashape is saying 'Some parameters have incorrect names and will be lost'.

Any ideas? I'm on version 1.7.4 professional.


General / Generating mesh with subset of cameras in regions of model
« on: April 13, 2022, 02:25:04 PM »
I’ve taken a transect of an area. Certain small areas are masked in most cameras - I would like to only use a small subset of cameras to build the mesh in these places. Is this possible? I’m aware of ‘Volumetric masking’, while this does exclude the cameras I don’t want, its also excludes those I do, leaving a hole in my model.
Happy to do some scripting if it might be possible through python?


Camera Calibration / Calibrate cameras only on markers
« on: December 20, 2021, 02:40:26 PM »
I would like to calibrate cameras only using markers.
Can anyone confirm that my workaround for doing this is correct?

-Align photos using standard settings
-Automatically ID markers in all images
-Delete all points
-Optimise cameras

It seems to be working - I'm getting very sensible alignments after doing this.

Did you get anywhere with with this? I'd very much like to do the same.

From what I understand, calibrating on targets only, particularly if they are placed a variety of distances from the camera could really boost the camera calibration accuracy.

General / Gradual selection without GCP's
« on: November 18, 2021, 10:23:39 PM »
I'm working on footage, typically from stereo cameras underwater, but with no GCPs.
Does anyone know if there is any value to doing a process of gradual selection and camera optimization without GCPS?

Many thanks,

General / Segment 3D model from segmented images
« on: October 26, 2021, 04:36:07 PM »
I am trying to transfer annotated segmented images (segmented version of same images that have been used to generate the model) to a 3D model. Is it possible to do this in a way that preserves the quantised labels of the segmented images?

For example, if one of the labels has a label or color (255, 0, 0), I would like this this label to be preserved when it is projected onto the model.
I understand that its possible to disable blending in ‘blending mode’, but then this only takes one camera position to generate the texture. Ideally, an option that take the most common value from all cameras that can view a point would be available.

Happy to write some python if that could be a possible solution?

Pages: [1]