Ortho projection to exact plane

[Daddazio]

Hello,
first of all congratulations for the great software!
I have the situation described in attached screenshot.
I've already written a script to export single photo in ortho mode:

Code: [Select]

import PhotoScan
import os
import sys
app = PhotoScan.Application()
doc = PhotoScan.app.document
chunk = PhotoScan.app.document.chunk
folder = PhotoScan.app.getExistingDirectory("Select folder where to export ortho photos")
folder = os.path.normpath(folder) + os.sep
app.messageBox("Photo number: " + str(len(chunk.cameras)))

for camera in chunk.cameras:
X = camera.key
camera.enabled = True
path = folder + str(X) + ".tif"
chunk.exportOrthophoto(path, format="tif", blending=PhotoScan.MosaicBlending, color_correction=False, write_kml=False, write_world=False)
camera.enabled = False

app.messageBox("All photos exported")
The question is: how to orient exactly the plane (and cameras) with the XY plane, or is it possible to directly project photo orthogonal to my imported plane in python?

Thank you for help,
Matteo.

[Alexey Pasumansky]

Hello Matteo,

You can use custom projection plane for orthophoto export. To do that you need specify "projection" argument for the .exportOrthophoto() function and specify valid 3x3 normalized matrix.

Matrices for the predefined views are given in the following thread:
http://www.agisoft.com/forum/index.php?topic=1322.msg15758#msg15758
But if you have any special case, let me know if you require some assistance for custom matrix creation.

[Daddazio]

Hello Alexey,
thank you for the reply.
I already have the plane where to project photos.
But, how to build a matrix picking 3 point from that plane?
Also, in general, how the matrix works? Is there a docs where is explained?

Thank you again.
Matteo

[Alexey Pasumansky]

Hello Matteo,

Each row of matrix is vector defining coordinate system axis orientation: horizontal, vertical and depth (normal to both of them). All vectors should be, of course, normal to each other and normalized (to avoid additional scaling).

[Daddazio]

Thank you!
The last question about the matrix: why the matrix has 4 vectors of four coordinates?

Matteo.

[Alexey Pasumansky]

Hello Matteo,

They are extended to 4x4 matrix, just zeros as fourth element in the first three rows and the last row is [0, 0, 0, 1].

[Daddazio]

Thank you!

[giancan]

Hi Alexey,
for some reason, when I use this script from certain angles it seems that I have a section of the building more then a view from "outside".
Is it something to do with the depth of field (close and far field)?
I tried also to expand the bounding box in case it was too tight, but nothing changed.

Thanks a lot,
G.

[lorso]

Dear, Alexey Pasumansky
Can you write the script for the ortho projection on a plane right?
Maybe by example?
Thanks a lot.

Lorenzo