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 - badger

Pages: [1] 2
from the site you linked:

"Phase One iXU and iXU-RS series camera systems, now with full support for DJI M600 and M600 Pro drones, are priced from $29,000 (complete with integration kit and lens)."

have you checked how many Canons or Nikons you can buy with 29,000$ ?

Face and Body Scanning / Re: Issue with Alignment in Photogrammetry rig
« on: November 11, 2020, 02:53:06 PM »

it's hard to say, from the picture you are posting. However yes, the photographs should overlap more than a bit one with the other, for Metashape (or any photogrammetry application) to be able to align.

You should have two adjacent cameras overlap by possibly 40/50% of the image. Try not go below 30% of overlap.

If you don't have enough cameras to complete the full cylinder with your favourite radius, try enlarging the radius so that each camera shoots a bigger portion of the subject, since you can't unzoom.

50mm is quite long focal length, for instance my scanner has 1.3 meters of radius and I use 24mm fl in order to cover the people in an optimal way. With 50mm you need a much larger radius than what I have.


Face and Body Scanning / Re: Choosing a single camera for photogrammetry
« on: September 16, 2020, 01:49:26 PM »
Take the cheapest (*) but take 3 of them (at least two). So you can take images from 3 different levels instead of just one, which is definetely not enough, in spite of how many shots you take on a complete revolution.

(*) assuming all the models you listed have a decent lens.

Face and Body Scanning / Re: Best Camera Choices for a rig - budget cameras?
« on: September 16, 2020, 01:43:26 PM »
What type of cameras do you use with raspberries?
I've only found cameras with a few megapixels, but I didn't mess around a lot with rasp until now...
Thank you!  :)

I use the standard raspberry camera module, which is 8 megapixels. Actually 2 thirds of my cameras are still equipped with the older 5 megapixel module.

The models come out quite acceptable for low resolution applications (printing, AR) but not for high res applications.

I'm adding DSLRs to the rig and now have the front of the persons shot by 8 DSLRs, growing to 16 soon. The detail is of course far better, as the cost of the rig.

By experience I can tell that the megapixels do not count that much. I did not notice any improvement from movig to 5 MP to 8 MP. A friend of mine uses 40 old DSLRs, just 11  megapixels each, to shoot only the head and he get stunning models.

The difference to quality is not given by the individual sensor megapixel but by the lens (the raspberry's sucks) and by the number of sensors.

Face and Body Scanning / Re: Best Camera Choices for a rig - budget cameras?
« on: September 10, 2019, 06:56:57 PM »
I recently did this
Used Canon EOS 1500D, 24MPIxels, working great.

very a good result with respect to textures. However the shape still comes out more crisp, when done with 96 Raspberries.

Certainly the advantage of not needing projection and double shooting is immense, but you might want to grow a little more the number of points of view in order to capture finer details.

Face and Body Scanning / Re: The Business of 3D Scanning
« on: June 01, 2019, 12:44:22 PM »
This is a very interesting thread for me. It was started in 2014 and it has been stated, that technological advance of handheld scanners will render complex and expensive photogrammetry systems with many DSLR cameras obsolete.

Now 5 years later how do things look like? I am very interested in this as I am looking into getting into commercial 3D scanning to produce 3D printed figurines. I am considering what system should I get to start this kind of business. There are different systems available on the market right now. One can custom build a photogrammetry system (which is what I was looking into originally), or buy 'off the shelf' system like Twinstant or Artec Shapify or get a portable scanner like Artec Leo.

Of course every system has it's own strengths and weaknesses, however I would like to know where the market is going these days. Does it still make sense to build a big and complex photogrammetry rig with a lot of postprocessing involved or one could get faster, better results with one of the latest hand held scanners, in particular Artec Leo (big plus, it is mobile)?

Hi Volteco.

the type of scanner you choose depends on what type of business you want to set up. If you aim at lowish quality, low price, mass market then a handheld might fit your needs. As long as you can actually complete the scan. It takes long minutes and if the subjects moves odds are that the scan has to be restarded from scratch.

Therefore you cannot scan dogs, kids and therefore family groups.  Then the printing costs will probably increase the figurine price above the mass-market threshold anyway.

A photogrammetric scanner can scan any kind of subject, no market accessibility restrictions.

Another disadvantage of handhelds is the quality/resolution of the scan. Sometime ago I scanned a local popstar and his staff asked a textureless print. They explained me that they wanted to compare my scan with that produced by one competitor of mine (operating handheld or even ipad based scanner) and did not want to be deceived by the textures. Result: they admitted that you could recognize the guy in my print but you could not in the other one. I only have a raspberry based photogrammetric scanner, not even a DSLR...

In short, a photogrammetric scanner gives far better models and opens you interesting side businesses, such as producing models for VR/AR industry.

Pricewise the investment is higher, for a DSLR scanner, but definetly on par for a Raspberry scanner, compared to professional handhelds. Kinetcs / ipads and the likes are not professional, are toys. May fit your goals but remain toys.

The only real advantages of handhelds are portability and that they return a finished scan right away, while with photogrammetry you still need to process the photos.

Face and Body Scanning / Re: Adding dots to a subject
« on: February 22, 2017, 08:52:18 PM »
This is related to my last post but different enough that it deserves its own thread. I've been having trouble with PhotoScan capturing smooth undifferentiated areas on the subject. Would it help PhotoScan if I added some dots to the subject. I've thought of either little 1/16" or 1mm sized circles of sticky note paper or even tiny dots of a playdough type clay that I could easily remove afterwards. In either case I could then remove them in the final mesh using Meshmixer or another similar tool.

Would these approaches help PhotoScan to stay locked in on otherwise undifferentiated areas of a model?


yes, it might help. Put many and tiny and possibly of irregular shape, not all the same. Or you can consider dark powder to dust the thing. You can wipe it afterwards

Face and Body Scanning / Re: Scanning a ceramic bust with limited success
« on: February 22, 2017, 05:58:47 PM »
Thanks for the tip. Where can I learn more about using projected patterns? I looked in the PhotoScan manual and didn't find a reference. A Google search turned up very little as well.

It would be very complicated if this is the onlt project that requires this. You could consider dusting the subject with darkish powder, graphite or so, which should be way easier to clean out afterwards.

Face and Body Scanning / Re: How to import a Reality Capture scene?
« on: February 04, 2017, 02:49:38 PM »
Hello everyone,

is it possible to import a mesh from RC so that i can texture it in Photoscan? Everyway i have tried it did not work.

Kind regards

I guess you also agree that RC texturing badly sucks ?  :) :)

However, it is not possible unfortunately, unless you have the pro version.

You need to export from RC as "bundler output with negative Z", but then you can't import it into Photoscan because of a photoscan bug. You import the cameras, but Photoscan translates all their positions, while the imported mesh mesh will not be translated (hence texturing in Photoscan of an RC generated mesh would not work).

There's a workaround, that was kindly provided to me by agisoft customer support, which works through python scripting and therefore only on the professional version of Photoscan.

Bug Reports / Re: Issues with importing bundler.out alignment
« on: December 16, 2016, 12:52:28 PM »
Hi Alexey,

I think it's fair to thank you guys also publicly, on the forum. A very top level customer support.

keep it  ;)

Bug Reports / Re: Issues with importing bundler.out alignment
« on: December 14, 2016, 08:40:26 PM »
Hello badger,

Is it Standard or Professional edition?

I think we have found what's the problem with bundler.out import. To avoid coordinate overflow (internal coordinates of the chunk) PhotoScan is translating the point of origin for the imported cameras.

I'll try to send you back the project where the camera positions are aligned with mesh (planning to do that with Python script).

It's standard edition, for the time being. And thanks for the explanation. Would you be so kind to give details about the translation of the PoO, so that would be possible to compensate it ?

Bug Reports / Re: Issues with importing bundler.out alignment
« on: December 14, 2016, 02:25:11 PM »
Hello badger,

Can you also send the list of files that you are using for Meshlab import? We are trying to open the mesh with the bundler file there, but it keeps crashing.

Also please specify which edition of PhotoScan you are using.


I'm using Photoscan 1.2.6

I sent the list of photos yesterday, but pathnames need to be adjusted to your testing environment, otherwise meshlab crashes because it does not find the photos.

Bug Reports / Re: Issues with importing bundler.out alignment
« on: December 12, 2016, 04:41:00 PM »
by eyeballing the alignment it looks a bit offset also along the X and Y axes. Not as much as the Z axis.

Bug Reports / Re: Issues with importing bundler.out alignment
« on: December 12, 2016, 04:33:43 PM »
thank you. I'll sent everything via WeTransfer.

Do you also need the photographs ?

Bug Reports / Issues with importing bundler.out alignment
« on: December 12, 2016, 04:01:08 PM »

I'm trying to import scene and mesh from RealityCapture, to take advantage of the great texturing capabilities of Photoscan. The scene is exported from RC in bundler format but when importing in Photoscan, the mesh is translated on the z axis with respect to the cameras (see attached screenshot).

I imported the same mesh and the same bundler.out with meshlab and the alignment shows correct there, therefore it is not a problem on the export side, but Photoscan is rendering a translated mesh, or translated camera positions.

Any suggestion how this could be fixed on my side, or any hope to get a patch for this ?

thank you in advance

Pages: [1] 2