Messages

1

General / Re: 100% CPU for minutes after loading heavy project

« on: January 13, 2022, 12:21:11 PM »

Upgrading to v1.8.0 solved the problem !
I'll have at least learnt the VBO option and default loading of the point cloud rather than the heavy model (which has always been a problem for me), thanks !

2

General / Re: 100% CPU for minutes after loading heavy project

« on: January 13, 2022, 11:45:19 AM »

Same issue with v1.7.5... Metashape consumes >90% of CPU all the time, even though it is not doing anything nor ready data (disk 0%). I have a project with two chunks or 1500-2000 HD pictures, nothing spectacular compared to what I am used to process with Metashape. And once again, Metashape is supposedly not doing anything...
Any clue about this weird CPU consumption with this version ?

3

Python and Java API / Re: How to use external python libraries in PhotoScan?

« on: November 27, 2017, 11:51:30 AM »

I had the same problem and William's solution worked for me ! I had to install the exactly same Python version (3.5.2 today) and replaced the python folder in "photoscan-pro" folder by a link to the external python install.

To get external python folder, in a terminal :
python3.5 (or yours)
import sys
sys.path

and then create a link (in ubuntu)

ln -s FROM_PATH TO_PATH

4

Python and Java API / Re: Export markers

« on: October 26, 2017, 12:44:06 PM »

Ok I got it work with APS 1.3.4 (I guess the problem was due to changes in calls with last version of APS)

We have to specify "chunk.transform.matrix" and not "chunk.transform" before doing "v_t = T * v" otherwise APS does not understand ad chunk.transform contains several things beside the ".matrix".

Beware also with the "PhotoScan.Matrix.Diag" with captial for "Diag", in case needed... (old version did not take a capital)

FILEPATH = 'G:/marker_export.txt'

f = open(FILEPATH, 'w')
for item in chunk.markers:
   if item.position == None:
      continue
   v = item.position
   v.size = 4
   v.w = 1
   if not chunk.transform:
      chunk.transform.matrix = PhotoScan.Matrix.Diag( (1,1,1,1) )
   T = chunk.transform.matrix
   v_t = T * v
   v_t.size = 3
   proj = chunk.crs
   v_out = proj.project(v_t)
   f.write(item.label + ',' + str(v_out[0]) + ',' + str(v_out[1]) + ',' + str(v_out[2]) + '\n')
f.close()

Cheers

5

Python and Java API / Re: Export markers

« on: October 26, 2017, 10:29:57 AM »

Hi all,

I am trying the same export with you code but I get the same error "TypeError: unsupported types" at line "v_t = T * v".

But my project is referenced as there is a [R] beside the chunk name, and "str(type(T))" gives "<class 'PhotoScan.ChunkTransform'>".

I am using APS 1.3.4

By the way, it seems that on this version of APS, the PhotoScan.Matrix.diag( (1,1,1,1) ) is wrong and "diag" takes a capital :
PhotoScan.Matrix.Diag( (1,1,1,1) )

But not a problem in my case as my chunk.transform exists.

Thanks for your precious help !