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.


Topics - adrien_gbx

Pages: [1]
1
General / (sun)Light angle on lens for every pixel
« on: May 15, 2018, 05:39:30 PM »
Hi everyone,

We are working with fisheye lens (90° lens) and we would like to extract for every pixel of the individual images of a 3D reconstruction the ray light angle regarding the principal point / focal point segment. The desired answer is a image providing for each pixel such angle (ranging from 0 to 90).

I know all the information's needed are in the photoscan project  notably from the internal orientation but I don't know how to get what we want within photoscan...

Has someone already tried such project?

Kind regards

Adrien


2
Python and Java API / Run python script on cluster
« on: September 18, 2017, 05:43:21 PM »
Hi,

We'd like to run an already working python script on a cluster for consecutive actions, eg. dense cloud -> DEM /-> ortho

We work with an already aligned chunk and make a moving bounding to process all the further steps. So the tasks have to be repeated againg and again, following or moving bounding box.

We are looking for a way to run a working python pipeline directly to already configured photoscan cluster.

As I understand from this previous post ( http://www.agisoft.com/forum/index.php?topic=4966.0 ), we should not simply create several tasks and run it through the script.

Alexey Pasumansky mentionned that users should create a photoscan network task running the script, which will launch the whole pipeline on the cluster. There is no function in python API to run a python script.

How can we create a network task running a script ? Someone in the community has experience about that ?

Kind regards,

Adrien

3
Camera Calibration / Direct use of already oriented cameras
« on: June 23, 2016, 02:58:59 PM »
Dear phootoscan users,

We are working on already oriented camera's (camera ID , X, Y, Z, Omega, phi, kappa). We also have got calibration information (focal length and pixel size). We use vexcel camera, with very low distortion. Our project is based on aerial imagery. We re-use an aerial survey which was initially acquired for orthophoto coverage.

Currently, we leverage the orientation information with the import tool in the 'reference' toolbar. Photoscan re-compute new orientation and as far as i know, PS does not use the omega phi kappa information.

We would like to keep the provided camera orientation 'as it is' to run the entire photogrammetric process.

Since PS is now 1.2 , somebody knows a work-around to proceed ?

Regards,

Adrien

4
Python and Java API / list of camera used to build the dense cloud
« on: April 21, 2015, 10:57:08 AM »
Hi erveyone,

Does someone know how to extract the list of the camera name used to compute the dense cloud ?

Regards,

Adrien

5
Hi everyone,

I'm using photoscan to handle an heavy dataset of  more than 5000 Vexcel aerial images (700 Mo). The alignment step is performed within a single project and the dense cloud step is performed through the use of a moving bounding box of 2*2 km.

After the dense cloud generation (at the scale of an individual tile of 2*2 km), I want to save the dense cloud and the tiled project. This allows me to perform some postprocess checks or process (orthophoto generation, troubleshooting, ...)

My problem is that I can't find a method to save only the tie point which are located within the tiled bounding box.
I have more 5000 tiled bounding box and I can't save the entire sparse cloud because the amount will represent more than 5 To !

I want to "physically" remove tie point using the python scripting. I do know the .valid command but this one doesn't really remove them. Tie points are just disabled but not really removed.

Someone has an idea ?

Regards,

Adrien


Pages: [1]