Messages

1

General / Re: Not all points are imported

« on: April 17, 2019, 08:35:32 PM »

Following up on this thread...

My file was simply "X Y Z" which is not the PTS format. The first line needed the number of points in the file.

2

General / Not all points are imported

« on: April 17, 2019, 01:42:14 AM »

Hello,

I'm trying to use Metashape to generate a mesh from points. I have a PTS file "x y z" format with 1.8M points. When I import it into Metashape, only 513,820 points (~25%) are imported. Is there a way to import more points?

Thanks,

   -k.

3

Python and Java API / Re: Adding photos to CameraGroup

« on: April 26, 2016, 08:58:29 PM »

Nevermind, I figured it out. To move the last photo into a group that has been created, I do the following.

Code: [Select]

    group = chunk.addCameraGroup()
    for img in image_list:
        chunk.addPhotos([img])
        camera = chunk.cameras[-1]
        camera.group = group


4

Python and Java API / Re: Adding photos to CameraGroup

« on: April 26, 2016, 08:43:26 PM »

Code: [Select]

camera = chunk.cameras[0] results in

Code: [Select]

IndexError: list index out of range

5

Python and Java API / Adding photos to CameraGroup

« on: April 26, 2016, 06:57:39 PM »

Hi,

I can add photos to a chunk with chunk.addPhotos(file_list). Is there a way to add images to a group? The group is created with group = chunk.addCameraGroup().

6

General / Add GPS coords to images based on PhotoScan results

« on: September 12, 2015, 12:44:29 AM »

I have a bunch of images of a fairly self-similar surface, w/o GPS coordinates and few tie points. Lets say ~50 images.

If I try to reconstruct the scene with all 50, it does not work. If I do it with only 25, it works much better.

Would it be possible to do it with 25, apply the derived coordinates to each image, do it again with the other 25, apply those, and then perhaps if I do it with all 50, with the help of the "known" image location, it would work, and I would get a higher quality scan than if I only use 25.

Is it possible to take the PhotoScan-derived GPS coordinates and apply them to the image EXIF data? Is this something I need to do manually w/ a script, or is there some feature in PhotoScan that supports this?

7

General / Re: why photoscan can't import video ?!

« on: January 14, 2013, 04:06:58 AM »

If you have video, it is very easy with 1 command and FFMPEG to produce a folder full of PNGs from the video, and then you can use these PNGs in PhotoScan.

8

General / Re: how to find processing options that generated a PSZ file

« on: January 03, 2013, 10:42:01 PM »

Perfect. Thank you!

   -k.

9

General / how to find processing options that generated a PSZ file

« on: January 03, 2013, 08:59:33 AM »

If I have a .PSZ file, can I determine how it was created? If I unzip it and examine the doc.xml file I see a lot of information encoded there, but it doesn't appear to record if it was reconstructed at low, medium, or high quality for example, or what options for mesh refinement.

Is there a way to determine this? If not, please consider this a feature request. It might help with debugging too.

I ask because some of my scenes reconstruct almost perfectly with the low quality option, and the photos are a random jumble if I try to do high quality. And others the reverse. If I need to re-process a segment, it would be nice to know what worked the first time, and the data file itself is a better recorder than my notes.

Thanks,

  -k.

10

Bug Reports / Flags don't have labels

« on: November 26, 2012, 07:26:29 AM »

I'm using PhotoScan Pro on Ubuntu. My flags appear as little green flags with a white dot under and a white box to the right. Some flags appear as a gray flame (?) with a white dot under and a white box to the right.

The white box is always blank and empty on Linux. It contains the marker name on OS X. I haven't tested it on Windows.

Does anyone else run this on Linux and experience the same issue? Is it a Linux bug, or specific to my setup?

Thanks,

   -k.