Messages

16

General / Exporitng dense point could confidence as ASCII PTS

« on: July 15, 2022, 12:46:51 AM »

Hello Agisoft support,
I am trying to explore the dense point cloud confidence values that are calculated and can be extracted in .LAS. Unfortunately "Save Point Confidence" is grayed when saving the points as ASCI PTS. Is there a slick way in Metashape around this obstacle? Alternatively, maybe with the Python API  "Save Point Confidence" can be exported as ASCI PTS? 
Thanks

17

Python and Java API / Re: Exporting TPs with RMS error in xyz

« on: June 09, 2022, 09:10:14 AM »

Paul,
Thanks for the deeper explanation. It is impressive your mastery of the Metashape API despite the limited reading content in the manual and assuming you are work/worked for Metashape.
First, is a "Track"  the line that connects two adjacent tie-points that RANSAC filters?
Second, could you give me some advice on how to better understand Metashape API functionality and be able to export the things I am interested in?
Third, I supposed a similar analysis can be done on the dense cloud? Do you have a blog/or book about your work with Metashape API?
Again thanks a million!

18

Python and Java API / Re: Exporting TPs with RMS error in xyz

« on: June 08, 2022, 02:19:19 AM »

Paul,
I don't understand what is the camera object in your code. Is camera based on :
for camera in chunk.cameras:
   camera.error()

If so, I don't understand how you are mixing camera from cameras iteration and point from chunk.point_cloud.points?

Also, you are using a norm() method, but I am also interested in the number of observations for a single tie-point (num_obs). Is the num_obs and attribute or method within point?

You are very knowledgeable about Metashape API , way outside what I have found in the Metashape API manual. Thanks for your patience to explain and provide code snippets.

19

Python and Java API / Re: Exporting TPs with RMS error in xyz

« on: June 06, 2022, 11:07:28 PM »

@Paulo,
Is num_obs how many cameras (or how many observations-camera) are associated with the tiepoint?
What is   errrep(pix) ?
How did you cacluate/extract num_obs   errrep(pix)?
I wonder if num_obs can give a sense to the population that a variance is calculated for each tie point.
Thanks

20

Python and Java API / Re: Re-opening project with chunk

« on: June 05, 2022, 09:01:18 AM »

I would like to support my interest PierreB's question!
What I have tried to do:

Code: [Select]

doc = Metashape.Document()
doc.open(project_name)
chunk=doc.findChunk(0)
I know the project has four control points, but when I run the following I get nothing.

Code: [Select]

for point in chunk.markers:
    print(point.read)
It is pretty clear I am not reading/loading the chunk correctly.
Thanks

21

Python and Java API / chunk.markers coordinate system

« on: April 25, 2022, 06:57:42 AM »

Hello,
My project is set as EPSG::26934.  the entire project is processed with the Python API.
When opening the .psx  with Metashape GUI, the coordinates make sense (UTM- as seen in the attached picture)

But when I review the coordinates in the Python interpreter the number doesn't make any sense to me:

Code: [Select]

points_dic={}
for point in chunk.markers:
    print(point, point.position)
    points_dic[point.label]=point.position
points_dic={}

for point in chunk.markers:

    print(point, point.position)

    points_dic[point.label]=point.position

<Marker 'point 1'> Vector([2.674798477808206, 6.560502854674506, 3.311077291228823])
<Marker 'point 2'> Vector([2.6818437237369803, 3.647193405560801, 0.34158157745059553])
<Marker 'point 3'> Vector([2.900154059499172, 4.059596584415228, 0.7825189294777867])
<Marker 'point 4'> Vector([8.001023537754426, -0.23136466383098236, -3.40010373078223])
<Marker 'point 5'> Vector([4.91971436138031, 3.6984685117053813, -0.2117274212552949])
<Marker '113'> Vector([8.001507569124444, -0.23004959586614865, -3.4028323038655808])

Can someone help me understand in what units are these numbers? This doesn't seem consistent with other posts in the Forum.

22

Python and Java API / Re: When cliping raster to boundary error message: Unsupported datum transformation

« on: May 06, 2021, 09:26:41 PM »

Thanks for the help Paulo.
My bad, I miss read the Metashape API manual (attached), as one can only import  CRS if the shapefiles is in CSV format.

23

Python and Java API / When cliping raster to boundary error message: Unsupported datum transformation

« on: May 05, 2021, 01:09:46 AM »

Hello,
I have processed imagery (Align, BuildDense, BuildDEM, BuildOrtho) in stateplane projection.

Code: [Select]

Metashape.CoordinateSystem('EPSG::26934')Then I imported a shapefile with the same projection and datum:

Code: [Select]

#import shape
chunk.importShapes(path=r'125_bounds.shp',boundary_type=Metashape.Shape.OuterBoundary)Then when I try to clip the DEM and ortho during export:

Code: [Select]

chunk.exportRaster(path=r'new_file.tif',source_data=Metashape.DataSource.ElevationData,white_background=False,clip_to_boundary=True,nodata_value=-99)I get the following error:

Code: [Select]

RuntimeError: Unsupported datum transformation
Is this workflow limited to only latlong?

Maybe relevant, when importing the shapefile in the API I cannot see it- shapefile was created in ArcGIS. Rasterio reads the shapefile just fine.

24

Python and Java API / Re: Removing low confidence dense cloud points

« on: May 04, 2021, 03:48:59 AM »

@Pauloa,
I am confused.  If I want to use the cropSelectedPoints, like described below, how would I got about it selecting the points to be cropped?

Code: [Select]

chunk.dense_cloud.setConfidenceFilter(4,255)
chunk.dense_cloud.cropSelectedPoints(list(range(128)))thanks

25

Python and Java API / Re: Removing low confidence dense cloud points

« on: April 30, 2021, 09:02:07 PM »

Hello Paulo,
Thanks for your comment. I am now confused, but Alexey wrote above:
"dense_cloud.removeSelectedPoints command, or select points that you would like to keep (10-255 confidence) and use dense_cloud.cropSelectedPoints.". How do use the crop.SelectedPoints then?

26

Python and Java API / Re: Removing low confidence dense cloud points

« on: April 29, 2021, 08:47:45 PM »

Makes sense.
But I am getting a Null dense cloud selection error

Code: [Select]

   207     chunk.dense_cloud.setConfidenceFilter(4,255)
--> 208     chunk.dense_cloud.cropSelectedPoints(list(range(128)))
    209     chunk.dense_cloud.resetFilters()
    210

Exception: Null dense cloud selection
My guess is that for some reason

Code: [Select]

chunk.dense_cloud.setConfidenceFilter(4,255) is not capturing points and therefore I cannot crop them.

Thoughts!?

27

General / Re: Missmatch between DetectMarkers none-coded and reference

« on: April 29, 2021, 09:54:53 AM »

Hello Alexey,
 I submitted the ticket on April 1st. Ticket #154074. Despite providing supporting data I haven't heard back for two weeks.  Can someone help me?
thanks

28

Python and Java API / Re: Removing low confidence dense cloud points

« on: April 29, 2021, 01:28:02 AM »

Alexey, why did you use list(range(128)) in your example?
I want to use dense_cloud.cropSelectedPoint, should I also use list(range(128))?

29

General / Missmatch between DetectMarkers none-coded and reference

« on: March 25, 2021, 10:22:38 PM »

Hello,
I would like to use the same GCP list for all projects and not make a subset for each.
When using DetectMarkers, I often get a bunch of false-positive detections. As in one case, I have 3 correct detections + 9 false-positive = 12 points.
When I import Reference 29 out of 45 of possible GCPS (29) with 0.2m threshold, there is a match for the 3 correct detections.
But if I try to import a larger GCP list 30 or larger, there are no detections. It doesn't make sense but I have been testing it all morning.

Agisoft, do you have some threshold that is causing this behavior?

thanks in advance

30

Python and Java API / remove and clear chunk.markers dosen't seem to work

« on: March 25, 2021, 05:55:39 AM »

Hello,
I cannot remove specific markers from chunks.
Here is what I tried:

Code: [Select]

chunk.markers.remove(chunk.markers[18])But when I run again:

Code: [Select]

chunk.markers

[<Marker '113'>,
 <Marker '111'>,
 <Marker 'point 3'>,
 <Marker 'point 4'>,
 <Marker '110'>,
 <Marker 'point 6'>,
 <Marker 'point 7'>,
 <Marker 'point 8'>,
 <Marker 'point 9'>,
 <Marker 'point 10'>,
 <Marker 'point 11'>,
 <Marker 'point 12'>,
 <Marker 'point 13'>,
 <Marker 'point 14'>,
 <Marker 'point 15'>,
 <Marker 'point 16'>,
 <Marker 'point 17'>,
 <Marker 'point 18'>,
 <Marker 'point 19'>][/code
As can be seen  <Marker 'point 19'>  18'th on the list was not deleted.

[code]chunk.markers.clear()Dosen't clear the list.

I assume I am doing something wrong, what?