Forum

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - an198317

Pages: [1] 2 3
1
General / Python script can't load image GPS info in EXIF?
« on: July 04, 2015, 11:35:33 PM »
Hi all,

I just realized when I used Python script to do the batch processing, the image GPS info in EXIF doesn't loaded in PhotoScan Pro. But when I do manual "Add Photos", I can see images' GPS info in Ground Control panel. I am using PhotoScan Pro 1.0.4 for a number of reasons.

Do you guys think this is something wrong with my code or just the version issue.

Thanks!
Nan

2
General / Re: GPU fail changing to CPU
« on: October 01, 2014, 08:46:33 PM »
Hi Alexey,

Thank you very much for your information. I switched back to 337.88 version, and it worked perfectly with GPU! I will look forward to trying the PS 1.1.

Thanks,


Hello Nan,

Please check driver 337 or older.

PhotoScan version 1.1 should work correctly with latest nVidia drivers.

3
General / Re: GPU fail changing to CPU
« on: September 30, 2014, 10:34:49 PM »
Thanks for putting my original post under this topic. It seems to be a driver issue. I will see if I can reinstall the older driver to solve the problem.

4
General / Dell XPS 8700 w/ NVIDIA GeForce GTX 745 GPU fails?
« on: September 30, 2014, 10:16:37 PM »
Hi,

I just got a new Dell XPS 8700 desktop with a NVIDIA GeForce GTX 745 OEM graphic card, and I have upgraded to the latest graphic card driver.

But when I run my python code to run the model, the console showed that the GPU fails and switched to CPU. I think this NVIDIA GeForce GTX 745 OEM card supports OpenCL 1.1. So I am curious that if someone had the similar problem?

Thanks,

5
General / Re: Hardware/Processing optimization
« on: July 29, 2014, 12:40:24 AM »
I did multiple test and found the same thing David found. I was thinking why Generating Mesh can't use GPU, which is the slowest part of the processing.

6
General / Re: GPU processing during the modeling and settings
« on: July 29, 2014, 12:35:42 AM »
That's what I would hope PhotoScan Pro can change....

7
General / GPU processing during the modeling and settings
« on: July 08, 2014, 01:04:37 AM »
Hi all,
I am curious about how much the GPU is involved in the modeling process. I can see during the dense point cloud step the GPU is used for computing. But during the mesh generation which is the most heavy computing process, I didn't see a lot of GPU involvement. So I am curious about during what the step(s) GPU is really used for processing?

And also what should I set up the OpenCL appropriately to optimize the speed? My workstation has 12 CPU cores, and Quadro 4000 has 8 cores based on PhotoScan OpenCL interface. Based on the OpenCL interface suggestion, deactivate one CPU core (using 11 of them) will use all Quadro 4000 8 cores?

Thanks,

8
Hi Alexey,

Thanks for your response. As I talked to you before, I have already turned off the "switch to the model view by default". So the PhotoScan is actually showing nothing but running my python code to process the model.

And I used .remove() method to remove the current chunk. And after this step, I used gc.collect() to collect the garbage.

Thanks again!
Nan

9
Hi,

I created a Python workflow to loop through the different folders for making a series models.

As I can see the first one or two Photoscan run is fairly fast, but PhotoScan get slower and slower. And I can tell the mesh creating step is really taking a long time not processing but "waiting" when it's already showing mesh creating is 100% finished.

In my python workflow, I did remove chunks after each model made, and also did garbage collector run. But it seems not work. The only way can speed PhotoScan is closing it and reopen PhotoScan, then the process will be back to fast again.

Any thoughts or inputs are really appreciated.

10
General / PhotoScan Pro is getting slower with Python workflow
« on: June 25, 2014, 08:04:54 PM »
Hi,

I created a Python workflow to loop through my different folders for making a series models.

As I can see the first one or two Python run is fairly fast, but PhotoScan get slower and slower. And I can tell the mesh creating step is really taking a long time not processing but "waiting" when it's showing mesh creating is 100% finished.

In my python workflow, I did remove chunks after each model made, and also did garbage collector run. But it seems not work. The only way can speed PhotoScan is closing it and reopen PhotoScan, then the process will be back to fast again.

Any thoughts or inputs are really appreciated.

11
General / Any possibility of using Leap Motion to control Agisoft
« on: February 19, 2014, 08:55:01 PM »
Hi everyone,

I had this crazy idea couple of days ago, and I figured I should post here to ask.

I will give a big presentation at a conference about our project of using Agisoft and will do some cool demos. I want to make my presentation even cooler, so I thought about using Leap Motion sensor.

I bought Leap Motion since it was first released, but I didn't really use it for work than fun stuff. So I thought is there any possible that Agisoft can talk to Leap Motion sensor that we can make the 3D model fly-over demo very cool?? We can zoom in and zoom out, or do model rotation, like Tony Stark with Jarvis? ;D

I hope this is not too crazy

Nan

12
General / Nvidia Tesla K20 Graphic card vs Geforce Series
« on: February 07, 2014, 09:06:35 PM »
Hi everyone,

We are going to put together a workstation for Agisoft PhotoScan Pro processing. We were looking at Nvidia Tesla K20 card. But I found a 3-yr old thread talking about this very topic. But that was about Tesla C2075.

And I also found Alexey's response as following:
"PhotoScan uses single precision arithmetics in the GPU processing part (actually the most time consuming calculations are done even using integer arithmetics). So the double precision performance limitations of the consumer video cards doesn't affect the processing speed at all. Consumer GeForce video cards should be as fast as professional (Quadro or Tesla) ones for PhotoScan."

Question 1: Does the new Photoscan Pro 1.0 version still use single precision arithmetic for the GPU processing?

Question 2: I can't find any Geforece GTX series specs on peak single precision floating point performance. But the Tesla C2075 is 1.03 Tflops, and the K20 is 3.52 Tflops, and K20 is significantly faster than C2075. So will K20 be faster than consumer level Geforce GTX series?

Thanks!
Nan

13
Python and Java API / Re: loading images from a dir
« on: December 10, 2013, 09:34:45 AM »
Hi Alexey,

This is the very topic I am looking for. How can I disable adding the TIFF thumbnail to the chunk in Python API??

Thanks a lot!
Nan


Hello MCSQRD,

Could you please specify how these files are generated? Looks like the jpg-file contain both the image and thumbnail. If it is so, please disable adding the thumbnail to the file.

The other option is to create multiframe chunk (with two frames) and remove the one with the thumbnails.

14
I hope this is the last question, Alexey,

What does your code "camera.label = camera.path.rsplit("/",1)[1]" mean?

Nan

15
Hi Alexey,

Yes, we only have one calibration group for now, so where I should change the code you showed?

And I assume the sensor.width and sensor.heigh are just the image dimension, right?

Thanks,
Nan

Pages: [1] 2 3