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.

Topics - ashalota

Pages: 1 [2] 3
On Unix, whenever the "Align Photos" steps starts in the GUI, the window is pulled to the foreground. This is frustrating if I am working on something else while I have set up an automatic process to align many chunks, as it keeps pulling focus into the Metashape Window. Can this be changed in the next update so it doesn't do that?

Switching between Workspace and Reference tabs to see the resulting total chunk alignment error is annoying, it would be nice if it was listed alongside the label (# cameras, #points, total error)

When I switch chunks (clicking 'orthomosaic' on one, then on another), it jumps to roughly the location of the newly selected item.

I sometimes create several versions of the same chunk and want to easily compare them. The 'jump' makes it hard to notice if there was a difference or not.

Would prefer it if either by default, or as a setting option, the camera did NOT move when I switch between chunks.

General / Cloud shadow adjustments (general question)
« on: April 30, 2020, 01:11:51 AM »
Does anyone know of a good method for removing/blending cloud shadows better?

I don't have much trouble in aligning my photos to make a nice aerial orthomosaic, but I haven't found any good methods for how to blend the shadows so they look less obvious. Especially if I have gone across my area back and forth and a shadow popped up on the second pass, it looks like big stripes down my photo!

I don't think this is Metashape specific, but I thought this user base might have some ideas.


Bug Reports / Reset View - zooms in too close
« on: April 30, 2020, 12:49:40 AM »
The 'reset view' button on mesh models zooms in too close, cutting off parts of the model. Would be better if it had a bigger buffer so the entire model fits into the panel.

General / Disregard (please delete post)
« on: April 25, 2020, 07:31:27 AM »
Disregard, other issues where found.

Python and Java API / Get Shape coordinates in python
« on: April 16, 2020, 12:36:44 AM »
I'm trying to extract the coordinates of the Points/Vertices of a Shape in python. Using 1.6.2 Pro

I can access the individual shape (Polygon). From here, I'm not sure how to get useful data from this.

Code: [Select]
In [119]: chunk.shapes.shapes[0].vertex_ids
Out[119]: 2020-04-15 17:34:31 Shape.Vertices([95, 96, 97, 98, 99, 100, 101, 102])

In [120]: chunk.shapes.shapes[0].vertices

In [121]:

What do I do with the vertex id to access the data in that vertex?

My goal is to be able to draw a bounding rectangle around the shapes that I drew in. So I need to get the contents of my shape and be able to pull the values. I know there is the exportShape option that will then export it as geoJson which I can later read, but I'd rather skip that in between step.

I'm importing a DEM from a .tif file. It looks good visually in the Ortho preview window area overlaid on the map, as expected. The 'no-data' area is showing as black. The chunk has 27 images in it, not a large area.

When I try to run build orthomosaic using it DEM as the source, it crashes in the GUI, and if trying to run this from python it shows this:

Code: [Select]
BuildOrthomosaic: projection = WGS 84 / UTM zone 19N, surface = DEM, blending mode = Mosaic, refine seamlines = 0, resolution = 0
estimating tile boundaries... done in 25.7475 sec
generating 44591x35324 orthomosaic (10 levels, 0.0286791 resolution)
selected 26 cameras
saved orthomosaic data in 0.004399 sec
saved camera partition in 0.00308 sec
scheduled 2 orthophoto groups
loaded camera partition in 3.3e-05 sec
loaded orthomosaic data in 0.000608 sec
Orthorectifying 13 images
Segmentation fault

Any ideas? I want to load in my LiDAR derived DEMs in order to remove the distortion that is showing up in my orthomosaics, especially warping of buildings.

Do you have any guidelines for which General parameters to check for the Optimize Camera Alignment option?

I have been using all except "k3" and "k4" real reason why though, I just selected a lot of them.

Can someone explain why you pick certain ones or not?

Feature Requests / On-line searchable Python API
« on: March 19, 2020, 06:40:40 AM »
The pdf is clunky and hard to search. Would be great to have a python API that is easier to navigate (hiearchy for class methods, etc)

In the reference panel, I've set the accuracy (m) and (deg) for a set of cameras, along with my own reference information.

If I then select a set of cameras from this chunk, and move them into a new chunk, the accuracy information is reset (to 10m and 10deg), although the rest of my reference information is saved.

Python and Java API / 1.6 mergeChunks API is confusing
« on: February 28, 2020, 08:33:47 PM »
Hi, How can I recreate this behavior from 1.5 in 1.6 (in GUI python console)?

Code: [Select]
doc =
doc.mergeChunks([doc.chunks[0],doc.chunks[2]],merge_models = True)

As far as I can tell, in 1.6 I need to name to have chunks as a named parameter, and instead of the chunk itself, I pass its index?
If I don't name the chunk, it pulls in all my cameras and merges them.
If I name them, it will only use the first chunk in the list. I'm so confused, I really hope you can revise the python manual for 1.6 soon as it is not very clear. Sample code with each function would be great.

Python and Java API / Format for "projection" parameter in 1.6 Python API
« on: February 21, 2020, 01:11:14 AM »
What is the new format for the projection paramter used in "exportRaster" ?

Previously I had set it as: Metashape.CoordinateSystem("EPSG::"+str(epsgCode)) , when using "exportOrthomosaic" in 1.5

I can't figure out how to translate this to the 1.6 API

General / GUI Reference Checkbox for Yaw
« on: February 18, 2020, 08:31:47 PM »
Can you please explain what the checkbox is under the Reference panel and for Yaw and why it has a checkbox next to it?

Sometimes it is 'unchecked', which seems to cause the Yaw value to be ignored during photo alignment. Is there a pattern for when photos are loaded checked/unchecked, and is there a way to verify this in the python API?

I have a set of 18 aerial photos in a line with 90% overlap.

When trying to align all of them at once, I get an alignment failed error, and only the first 2/18 align. This happens even if I retry several times. I have also tried renaming and loading the cameras so that they are in a different order in the chunk, but same results.

If I remove the first 2 images and run the alignment, all 16/16 align. The first 2 images are most dissimilar to the rest, as a large cloud shadow covers them.

I need to find a way to automate this process so that I can have a rule to determine when to remove photos from the alignment.

Ideally I would be able to juggle the included photos such that I can get the best results for each set of 20 photos (lowest total chunk error, most photos included).
Do you have any idea what criteria I should base this on?

Pages: 1 [2] 3