1
Bug Reports / Re: 1.4.2 Mesh Generation Problem
« on: May 12, 2018, 03:27:37 PM »
Mesh generation has been succesful in less than 5 minutes :-). So it's something with the new experimental method ;-)
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.
Pages: [1] 2
2
Bug Reports / Re: 1.4.2 Mesh Generation Problem
« on: May 12, 2018, 02:02:50 PM »
Oh maybe I got the problem. I used the new experimental method. I will try again with the old one!
3
Bug Reports / Re: 1.4.2 Mesh Generation Problem
« on: May 12, 2018, 01:55:56 PM »
Hi Alexey,
thats very strange because with earlier version I calculated similar models with more Photos and points with the same system. I will try again and monitor the RAM.
I have system controlled disk swap, but I will try with one manual set swap space (I have two SSDs with > 200GB free space).
Is it right, that I can not anymore create a arbitrary model from sparse cloud and cannot set the filtering method? It's greyed out and set to mild automatically? Maybe it is the reason I used everytime aggressive, so this could be a lot faster.
thats very strange because with earlier version I calculated similar models with more Photos and points with the same system. I will try again and monitor the RAM.
I have system controlled disk swap, but I will try with one manual set swap space (I have two SSDs with > 200GB free space).
Is it right, that I can not anymore create a arbitrary model from sparse cloud and cannot set the filtering method? It's greyed out and set to mild automatically? Maybe it is the reason I used everytime aggressive, so this could be a lot faster.
4
Bug Reports / Re: 1.4.2 Mesh Generation Problem
« on: May 10, 2018, 12:49:17 PM »
Hi Alexey,
Build Mesh
Arbitrary (3D)
Medium
High
Mild
Calculate vertex colors
Processing Log
https://www.dropbox.com/s/bdxd00omw4ud2nn/log_badallocation.txt?dl=0
Build Mesh
Arbitrary (3D)
Medium
High
Mild
Calculate vertex colors
Processing Log
https://www.dropbox.com/s/bdxd00omw4ud2nn/log_badallocation.txt?dl=0
5
Bug Reports / 1.4.2 Mesh Generation Problem
« on: May 10, 2018, 11:34:00 AM »
Hello,
it's now my 4th time to generate a mesh out of my Point Cloud. It takes extreme long and it aborts with a bad allocation error.
Project Data:
1251 / 1426 Cameras aligned
Dens Cloud 17.905.712 Pts low quality (8 hours 38 min)
i7 5820k
64 GB Ram
GTX 970 4GB
GTX 640 4GB
It takes unusual long almost 11 hours on low quality and ends with bad allocation. I remember on earlier projects which were bigger on old Photoscan release... the creation of the 3D model was much faster (around 2 hours).
So can be wrong?
Thank you!
it's now my 4th time to generate a mesh out of my Point Cloud. It takes extreme long and it aborts with a bad allocation error.
Project Data:
1251 / 1426 Cameras aligned
Dens Cloud 17.905.712 Pts low quality (8 hours 38 min)
i7 5820k
64 GB Ram
GTX 970 4GB
GTX 640 4GB
It takes unusual long almost 11 hours on low quality and ends with bad allocation. I remember on earlier projects which were bigger on old Photoscan release... the creation of the 3D model was much faster (around 2 hours).
So can be wrong?
Thank you!
6
General / AWS Experiences
« on: February 23, 2016, 09:33:13 PM »
Hello everybody,
I need to calculating my upcoming project and I think about using AWS because project is somehow timecritical. Does anybody have experiences using AWS vs. local build pc? The performance interests me really much. It is a big project, so I am planning around 10k Budget for AWS.
But for 10k i also get a quiet decent local machine without struggling network things and to buy multiple licenses. Or is there a license model for monthly paying for network licenses?
I need to calculating my upcoming project and I think about using AWS because project is somehow timecritical. Does anybody have experiences using AWS vs. local build pc? The performance interests me really much. It is a big project, so I am planning around 10k Budget for AWS.
But for 10k i also get a quiet decent local machine without struggling network things and to buy multiple licenses. Or is there a license model for monthly paying for network licenses?
7
General / Re: Advice for Workflow on big area in city needed
« on: February 11, 2016, 03:32:42 AM »
Thank you! This sounds like a good plan :-)
8
General / Re: Advice for Workflow on big area in city needed
« on: February 10, 2016, 04:36:25 PM »
Hi Wishgranter,
I think that an airplane will be to expensive and a lot of afford (also for equipment, the right technique for good quality images, etc).
I called the state office for geoinformation, and the they can give me georeferenced laserscandata pointcloud with an resolution of 4pts / meter for 84€ / km². This will be a good quality for mid distant objects. So I could focus on near objects for drone flights maybe. The question is, how can I map texture data on this pointcloud / meshed model. I am afraid that Satellite data resolution is not high enough? Do I need special software or can I achieve this with Photoscan?
I think that an airplane will be to expensive and a lot of afford (also for equipment, the right technique for good quality images, etc).
I called the state office for geoinformation, and the they can give me georeferenced laserscandata pointcloud with an resolution of 4pts / meter for 84€ / km². This will be a good quality for mid distant objects. So I could focus on near objects for drone flights maybe. The question is, how can I map texture data on this pointcloud / meshed model. I am afraid that Satellite data resolution is not high enough? Do I need special software or can I achieve this with Photoscan?
9
General / Advice for Workflow on big area in city needed
« on: February 10, 2016, 03:05:38 AM »
Hello everyone,
I got a new project for VR architectural visualisation and I need a 3D representation of a big part of the inner city (19ha, located in germany). Because of the VR experience there is a need of a high resolution. Quality of Google Earth is Minimum (Buildings, not Terrain). The mainfocus on the project is emotion and a consitency. A patchwork of different quality methods would break this.
Important places will be definitly covered with a drone and get a single photogrammetry each of them. The other places should maybe covered also with a drone (on higher altitude) and maybe satellite images. My problem is, that it is very hard to access high res satelliteimages which are usable for Photogrammetry. And it seems to cover the whole region with drones will be very expensive in money, time and datasets. There are a lot of high and complicated buildings and there will also be a ground view from egoperspective. My last project suffered a lot of low quality on walls from buildings because the droneflights covered a lot of roofs in high quality but the facades where underrepresentated.
Does anybody of you have experiences with similar projects? What is your advice? Is there a golden way how drones should fly? Do you know a workflow for working with satellite images?
Thank you very much for every advice you can offer to me :-)
I got a new project for VR architectural visualisation and I need a 3D representation of a big part of the inner city (19ha, located in germany). Because of the VR experience there is a need of a high resolution. Quality of Google Earth is Minimum (Buildings, not Terrain). The mainfocus on the project is emotion and a consitency. A patchwork of different quality methods would break this.
Important places will be definitly covered with a drone and get a single photogrammetry each of them. The other places should maybe covered also with a drone (on higher altitude) and maybe satellite images. My problem is, that it is very hard to access high res satelliteimages which are usable for Photogrammetry. And it seems to cover the whole region with drones will be very expensive in money, time and datasets. There are a lot of high and complicated buildings and there will also be a ground view from egoperspective. My last project suffered a lot of low quality on walls from buildings because the droneflights covered a lot of roofs in high quality but the facades where underrepresentated.
Does anybody of you have experiences with similar projects? What is your advice? Is there a golden way how drones should fly? Do you know a workflow for working with satellite images?
Thank you very much for every advice you can offer to me :-)
10
Feature Requests / Re: Smart Grid for exactly fitting regions
« on: September 02, 2015, 03:36:23 PM »
Oh great, I will check this. Thank you very much!
11
Feature Requests / Smart Grid for exactly fitting regions
« on: September 02, 2015, 03:04:03 PM »
If I am working on big grounds, it is to hard to work with one giant scene. So I cut them in seperate parts for model generation. But my problem is, that the region tool cant be used for exactly fitting parts.
So i would like to have some kind of smartregion tool or multi-region-tool / grid where I can choose 2x3 or 3x3 or 5x4 or similar and automatically creates chunks with the splittet point clouds exactly fitting to each other. So i can put them nicley together after model generation / texture generation and export in my 3d app of choice.
Thank you for your interest :-)
So i would like to have some kind of smartregion tool or multi-region-tool / grid where I can choose 2x3 or 3x3 or 5x4 or similar and automatically creates chunks with the splittet point clouds exactly fitting to each other. So i can put them nicley together after model generation / texture generation and export in my 3d app of choice.
Thank you for your interest :-)
12
General / Re: good preview settings
« on: July 17, 2015, 04:01:24 PM »
I used bounding box already and its really cool. But how can I see more in preview? Its recommended to have a higher sparse cloud or its cheaper (time and ressources) to render dense cloud?
13
General / Re: weired textures?
« on: July 17, 2015, 03:53:48 PM »
I've done the whole process from the beginning and it works now. strange.
14
General / good preview settings
« on: July 17, 2015, 03:51:37 PM »
Hello everybody. I'm still working on my big project - Realtime 3D Village scanned by UAV - 3300 Images
I got problems to receive a model with good textures and form with this huge scale. So I found a good compromise to have something to work with. It was a heavy lifting from a 2 mio polygon model with lot of textures to a 50k model with one 8k texture. I tried all programs to get a good lowpoly uv setup and used 3D Max to project multitexture highpoly to singletexture lowpoly. But it worked.
Now I and my client are not yet satisfied with the quality. But getting good quality from one piece is almost impossible within the budget and with my hardware (I upgraded to xeon 6-core with 24 GB Ram).
So i tried to puzzle out the right images for a small piece of the scene and just reconstructed a single building. And voila good quality, good texture, small polygon size. The qualityfactor now is 1:1000. But it was really complicated to find the right images. Because the drone flight is systematically around the village, not the building.
So I thought its maybe a good idea to raise the sparse cloud value, to be able to slice my parts through preview rather than scanning through 3000 images by myself to find the right ones for specific buildings. Dense Cloud for whole scene takes too much time and is to heavy to work with. So i want to avoid this step and create dense cloud only for my specific area. I want to save as much time as possible.
What would you do? Are there any recommendations or expieriences? Just to compare a image with the old church (from the whole set) and the new one.
I got problems to receive a model with good textures and form with this huge scale. So I found a good compromise to have something to work with. It was a heavy lifting from a 2 mio polygon model with lot of textures to a 50k model with one 8k texture. I tried all programs to get a good lowpoly uv setup and used 3D Max to project multitexture highpoly to singletexture lowpoly. But it worked.
Now I and my client are not yet satisfied with the quality. But getting good quality from one piece is almost impossible within the budget and with my hardware (I upgraded to xeon 6-core with 24 GB Ram).
So i tried to puzzle out the right images for a small piece of the scene and just reconstructed a single building. And voila good quality, good texture, small polygon size. The qualityfactor now is 1:1000. But it was really complicated to find the right images. Because the drone flight is systematically around the village, not the building.
So I thought its maybe a good idea to raise the sparse cloud value, to be able to slice my parts through preview rather than scanning through 3000 images by myself to find the right ones for specific buildings. Dense Cloud for whole scene takes too much time and is to heavy to work with. So i want to avoid this step and create dense cloud only for my specific area. I want to save as much time as possible.
What would you do? Are there any recommendations or expieriences? Just to compare a image with the old church (from the whole set) and the new one.
15
General / Re: weired textures?
« on: July 15, 2015, 03:10:07 PM »
But it's a merged chunk. I created the model from this. I dont use the smaller chunks above the big one. The model was build correctly from this, so why should the texture be build wrong?
Pages: [1] 2
