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 - Kms1

Pages: [1]
General / photos not aligning past version 1.1.0 , help
« on: May 01, 2015, 08:00:52 PM »
i have previously stayed with version 1.04(ish) when aligning photos as certain images just do not align for me in the later versions. Today I upgraded to the latest version and once again photos are not alligning for a model that previously worked in my earlier version :-(

i have approx 100 photos, have chosen medium for a quick align and have tried both generic and disabled alignment and for all the settings the photos are simply projecting from the same point instead of 360 degrees.

What has changed in between versions and what other options do i have in order to allow my photos to align?

I have not masked any of them as they have a white/grey background and on my previous version i did not mask them either.

I am a bit stuck as this has happened on other machines and with other photos on new version as well...

General / laptop GPU for photoscan processing
« on: March 11, 2015, 09:24:18 PM »
Hi, a lot of the scanning i am starting to do involves being away from my workstation and  I'm really being let down by the speed my current laptop is taking to process even low quality meshes. Can anyone recommend a GPU on a laptop that would do a decent job.

Currently i am thinking either quadro - 1100m/2100m or gtx range - 965m/970m

anyone have experience of either range and can tell me which is better for agisoft as I will be primarily doing quick meshes for the cleint/myself to preview the scanning process. Heavier processing can be done on another dedicated non mobile workstation.

any benchmarks for laptop gpis?


General / system out of memory , what can i do?
« on: March 11, 2015, 08:41:41 PM »

I am using a university mac pro with 32gb ram, i am processing an ultra high dense cloud - (105,708,624 pts) with 114 cameras. When I go to process high mesh (21million) it gives me an out of memory error after around 6 per cent. I have tried on a custom count of 8 million and get the same error.

Should I reprocess the dense cloud at a high setting and go for high mesh or settle for a low mesh from an ultra high dense cloud. Bit confused as recently i have been processing everything on ultra high dense cloud and the results seem to be a lot better than high dense cloud.



I  have been using this excellent software for around 3/4 months now and have had some great tips on this site, I would like to know from others who have tried it the best positing of the cameras in a multi camera setup, I have managed to borrow between 18-20 cameras (and tripods) possibly 1 or 2 more and all remotes, usb hubs etc to try to shoot hopefully my first full facial scan. All are a mixture of full frame and entry level canon dslrs.

What would the recommended positioning be? i was thinking of 6 stands with 3 cameras on each, 30 degrees apart to allow a 180 degree scan around the face and then possibly turning the subject around and doing the back of the head with the hope of attaching the two scans together afterwards.

I will not have much time to experiment so any information for anyone who has tried it would be useful. Could i get a full 360 degree scan of the head using this number of cameras or should i stick to doing front and back seperately? Would there be any purpose or enough information to try a full body scan with this number of cameras or would it be pointless?

Any information much appreciated..

General / adding more photos to an existing project, or using chunks
« on: June 09, 2014, 07:52:23 PM »

I have a photographed an object on a turntable using approx 80 images at 3 different angles using 3 different cameras, (240 images in total). I have just processed only the middle camera due to speed/time etc and its gone well but obviously i am missing the top and underside of the object.

My question is can i import the further images from the other 2 cameras and process them into the main mesh/project without reprocessing the original images or can I add them as another chunk and then align the chunks. I tried adding another chunk, aligned the images from another camera, build dense cld,mesh etc then when i went to align chunks nothing happened and the option to merge the chunks is greyed out.

Am i missing something or just really missing the point of how photogrammetry works, should I just start again and reprocess all the images from the 3 cameras together?

thanks a lot for any info

Apologies if this has been addressed before but I have searched the forums a lot and cannot find a direct answer.

I have had great success using 1 or two cameras and now want to go for an 8 camera facial scan until i can get more cameras. All the cameras I have access to are canon EOS dslr, some full frame, some not.

My question is this and I have searched the forum but there does not seem to be a direct answer. I want to know the best software or hardware solution to trigger all the cameras together without delay.

On here I see people mention sharpshooter and I have read the FAQ its says you can trigger multiple cameras but there is a 150ms delay between triggering cameras, as other people are saying even the slightest movement from the model is not good I am not sure if 150ms delay would be satisfactory across multiple cameras as surely the subject even very still would move in a second or so, and for larger setups this would surely be troublesome. Is there other software that could trigger with less delay? I have access to phase one capture one but do not think that supports multi camAm i better off using some kind of remote control triggered system ? can anyone using a large rig suggest a particular model?

Thanks for any info

Feature Requests / image paths , moving from mac to pc ?
« on: May 27, 2014, 01:54:32 PM »
Hi maybe im missing something but I process and work across several machines, 1 mac and 1 pc, when i save a psz file and open on one or the other the image paths are missing. Normally if its just on a pc I can copy the files to the correct location but on a mac the paths cannot be set as the file structure is different, because of this I have had to do a high dense cloud and mesh all over again on numerous files and its very time consuming.

Any way we could just browse for image path if it does not find it instead of moving all the files? I know this could probably be resolved using an external Hard drive but i have faster pci x and ssd on the macs and pcs and would prefer to keep my photos on there.

Apologies if this has already been addressed.

Also is there an undo operation on mesh decimation, does not seem to work ?

General / photoscan textures in zbrush
« on: May 23, 2014, 02:21:53 AM »
Hi, ive been using zbrush and photoscan for a few months now and have been getting great results, having scanned i bring the models into zbrush for smoothing, closing, dynamesh etc

My issue is that as soon as I start changing a model in zbrush and closing holes, zremeshing etc I lose my texture, what can I do to retain the texture from photoscan and what is the best workflow with regards to maintaining texture and uv details once a mesh has been repaired or retouched in zbrush. Also any tips on retopology, some people on here are getting excellent low poly smooth meshs out of the detailed high polys that imports from photoscan

any help on workflow/textures/retoplogy in zbrush would be greatly appreciated as I see some fantastic results from people on here


Pages: [1]