Messages

1

Python and Java API / Re: Is the Magic Wand masking tool available in the Python API

« on: November 20, 2019, 01:04:25 AM »

Hello! Just bumping it, wondering if anyone knows, one way or the other, if this is available? Or, perhaps, the best alternatives?

2

Python and Java API / Is the Magic Wand masking tool available in the Python API

« on: October 17, 2019, 06:16:04 PM »

Hi, this question was asked a couple years back (https://www.agisoft.com/forum/index.php?topic=8165.msg39080#msg39080), but didn't get a response then, so I thought I'd try again.

I have several tens of thousands of photogrammetric images, each with a black (or nearly-black) collar on them, and I want to mask out the collar. The collar-edge locations are not identical between each image. The Magic Wand tool works beautifully to do this in the AgiSoft Photoscan Pro UI, but I can't find any documentation or evidence that the tool exists in the Python API. What I'd ideally like to do is:

• Give it a numerical pixel location as the seed (same a clicking that pixel in the interface)
• Create my mask for each image, and
• Save the mask for each image.

...in the Python interface, without having to use the UI. If this is possible, let me know, otherwise I have to use various OpenCV tools that aren't nearly as intuitive or effective as what AgiSoft has in it. I have far too many photos to do this by hand with all of them.

My setup:

• Agisoft Photoscan Pro 1.4.5 with the Python 3.5 back-end.
• Ubuntu 18.04.3

Any help is appreciated!

- Mike

3

Python and Java API / Re: How to get Marker coordinates

« on: April 26, 2019, 02:09:26 AM »

Alexey, I haven't run it yet, but looking through your script, getting the Camera matrices and the Calibration matrices, inverting them and converting it through each of those back into image-pixel space, appears to be exactly what I need. Thank you so much!

I'll post here if I have trouble running it, your help is much appreciated.

- Mike

4

Python and Java API / Re: How to get Marker coordinates

« on: April 24, 2019, 09:48:04 PM »

Hello, Alexey or anyone else knowledgable of this: I have a follow-up question related to this post.

After generating a point cloud, I'm able to retrieve the coordinates of each tie-point, in both internal-coordinate space and projected coordinates, which is great. Beyond that, how do I retrieve the keypoint locations on each image, that the tie-points came from?  I am needing to get the original pixel coordinates on the images that each tie-point was derived from, but I haven't yet found how to get that information. I'm looking through the manual and API-tutorial but not finding it.

This is using Agisoft Photoscan Pro 1.4.5 on Ubuntu linux.

Let me know, and thanks so much!

- Mike