Forum

Author Topic: Getting vertical right  (Read 6721 times)

jrp

  • Jr. Member
  • **
  • Posts: 66
    • View Profile
Getting vertical right
« on: December 22, 2015, 02:51:36 AM »
I have asked around and scoured the internet, and am not finding what I need.

While scanning the inside of buildings (churches), I struggle to get the building the right way up. It seems that the only way to do this is to eye it in using the rotate tools in photoscan. Given that the buildings are old and have no single straight edge in them, this is fundamentally not a good a way to go.

I expect the official answer is the provide 10+ ground control points -- but I simply have no way of getting ground control points. I could get a total station, but that would defeat the point of using Photoscan.

I do not need a fully GCP located scan, all I need is vertical correct and the scale right. Scale is easy. I do not mind where north is.

I tried using 2 targets and a plumb bob to get them precisely located above one another, then measured the distance between them. This should be enough information to get the correct answer. I entered the first one as coordinates 0,0,0 and the second as the correct relative location. No other GCPs were entered, camera alignment was then done, and photoscan still insists on creating the whole building on it's ear, and complains that the GCPs are ~15 meters from where they should be.

I have just (moments ago) entered a third GCP by guessing it's location, and things have now changed in that the errors have now dropped down to something vaguely sensible, but the building is still lop sided as the guessing wasn't very good.

It seems this means Photoscan is ignoring GCPs unless there are at least 3 of them??

I could try adding a 3rd target right next to one of the others, but that doesn't seem like a good solution.

Is this a feature request?

Either way, this project is borderline unfinishable right now, despite the fact that I have all the data theoretically that I need.

Any advice on how to proceed will be gratefully received.

I'm using 1.04 build 1847 on a mac.

Kiesel

  • Sr. Member
  • ****
  • Posts: 332
    • View Profile
Re: Getting vertical right
« Reply #1 on: December 22, 2015, 03:39:55 PM »
Hello jrp,

it should work with three points!
Have you tried to construct a rectangular triangle on the floor (with the help of pythagoras, the three sides of the triangle for example with a=3, b=4, c=5 multiplied by a factor)? If the triangle is exact horizontal all verticals in your model should be exact vertical. Or if the floor isn't horizontal but you have a vertical wall draw the three points of the rectangular triangle on that. Start with a plump line= vertical side (a) and from that the horizontal (b) with two circles from the ends of (a).

Otherwise James has written a little tutorial for using the cosine rule, see:
https://dl.dropboxusercontent.com/u/28695776/Coded%20targets%20for%20alignment%20and%20scale.pdf

Karsten

PS: Why you are using the old version 1.04 ?
« Last Edit: December 22, 2015, 07:30:18 PM by Kiesel »

miami4quad

  • Newbie
  • *
  • Posts: 6
    • View Profile
Re: Getting vertical right
« Reply #2 on: December 22, 2015, 06:46:08 PM »
I would try to print some targets available in the software. Place them in a triangular pattern as the previous posted stated. When placing choose and even distance with a steel tape measure between them. Hold "0" for elevation "z" at all three points. Hold 0,0 for xy of one and figure the rest based on your measurements. This should work and your ground will be referenced to "0" and everything will be based on height from there.

Kiesel

  • Sr. Member
  • ****
  • Posts: 332
    • View Profile
Re: Getting vertical right
« Reply #3 on: December 22, 2015, 08:18:38 PM »
Thank you miami4quad! Yes this was the meaning but missing in my post.
Sometimes a z-value other than zero (for example 10) is good to avoid negative z-values.

Karsten

jrp

  • Jr. Member
  • **
  • Posts: 66
    • View Profile
Re: Getting vertical right
« Reply #4 on: December 23, 2015, 10:21:48 PM »
Some useful ideas, thank you, I will think them through.

Unfortunately the buildings I'm working in are many hundreds of years old, so literally have no surfaces that are guaranteed or are even likely to be flat, level, or vertical in them. The analysis to be done on the results (structure of arches etc) is fairly dependent on verticals being correct with respect to actual gravity.

My current strategy is going to be to use my plumb bob contraption -- basically a painters easel with a hole in it, printed target on it and a string with a weight dangling though the hole; another target is positioned on the floor under it -- creating a quite precise pair of targets one above the other.

To add the third point, I will manually add a point to one side of the bottom one ~25mm away from the centre of the target. This will provide very poor precision in terms of rotation around the z axis, but I don't care.

The answers above have certainly confirmed that I'm not missing anything fundamental.

bisenberger

  • Sr. Member
  • ****
  • Posts: 329
    • View Profile
    • Digital Mapping & Graphics
Re: Getting vertical right
« Reply #5 on: December 25, 2015, 11:30:10 PM »
You might try using a laser level to set your targets on a horizontal line. Cross-line lasers features Vertical, Horizontal and Cross Line Modes.

see examples google images:
https://www.google.com/search?q=cross+line+laser+level&client=ubuntu&channel=fs&biw=1920&bih=919&source=lnms&tbm=isch&sa=X&ved=0ahUKEwiRofGE7vfJAhWJQyYKHUJ5BtsQ_AUICCgD
« Last Edit: December 25, 2015, 11:36:11 PM by bisenberger »
Digital Mapping & Graphics LLC
https://digital-mapping.net/

Kiesel

  • Sr. Member
  • ****
  • Posts: 332
    • View Profile
Re: Getting vertical right
« Reply #6 on: December 28, 2015, 02:12:07 PM »
@jrp So you have found your way.  :)
Just a little suggestion:
Quote
My current strategy is going to be to use my plumb bob contraption -- basically a painters easel with a hole in it, printed target on it and a string with a weight dangling though the hole; another target is positioned on the floor under it -- creating a quite precise pair of targets one above the other.
Perhaps you can use a level judge instead of plump bob like this one:
http://goecke.de/Products/Surveying-accessories/Levels--plumb-bobs-and-accessories/Levels.html
and a ranging pole support:
http://goecke.de/Products/Surveying-material/Ranging-poles-and-accessories/Tripods-for-ranging-poles/ to fix your painters easel for easier work.

Perhaps this publication is of any help for you too:
https://content.historicengland.org.uk/images-books/publications/measured-and-drawn/measured-and-drawn.pdf/

@bisenberger
Quote
You might try using a laser level to set your targets on a horizontal line. Cross-line lasers features Vertical, Horizontal and Cross Line Modes.

This will work for a horizontal but not for a vertical line on an uneven or slanted wall, because the vertical laser line is only vertical in one direction (from projection direction side) and not from any other.



Karsten

bisenberger

  • Sr. Member
  • ****
  • Posts: 329
    • View Profile
    • Digital Mapping & Graphics
Re: Getting vertical right
« Reply #7 on: December 29, 2015, 05:34:51 AM »
Karsten, your statement about the vertical laser makes no sense.
Digital Mapping & Graphics LLC
https://digital-mapping.net/

Kiesel

  • Sr. Member
  • ****
  • Posts: 332
    • View Profile
Re: Getting vertical right
« Reply #8 on: December 29, 2015, 12:23:30 PM »
Bisenberger, if a vertical laser line is projected on an uneven object and you look from the side you can see the profil of your object as when it was cut by the laser line (think of a cut of a rock). This principle is for example used in David Laserscanner.
Hope this make more sense now.  :)

Karsten
« Last Edit: December 29, 2015, 02:55:10 PM by Kiesel »

jrp

  • Jr. Member
  • **
  • Posts: 66
    • View Profile
Re: Getting vertical right
« Reply #9 on: December 29, 2015, 01:06:33 PM »
There are certainly some good ideas there.

A ranging pole with a bubble level would provide a good equivalent to the plumb bob option, I'm not sure it would be as accurate, but it may be ideal for some people.

@Kiesel: I have looked at laser levels before, but I could not work out how to set one up in a way that would be useful; photoscan is very specific and limited over what it will accept.

Kiesel

  • Sr. Member
  • ****
  • Posts: 332
    • View Profile
Re: Getting vertical right
« Reply #10 on: December 29, 2015, 02:04:20 PM »
jrp, the suggestion to use laser levels came from bisenberger  ;). But I think they can be used to make a perfect horizontal system, in which you could measure three corner points of a horizontal triangle (see the link to the tutorial by James above in first post how to use them) and use that for GCPs in PhotoScan. Nevertheles I think your solution is simpler to work with.

Quote
A ranging pole with a bubble level would provide a good equivalent to the plumb bob option, I'm not sure it would be as accurate, but it may be ideal for some people.

The accuracy of a judge level depends on the ground joint/ polish/ (? - in german "Schliff") it has and how accurate you use it. The drawback of a plump bob is it is predisposed to wind and touch and it could take a while until it hanging absolut still.

Btw. there exists also optical and (expensive) laser plummets.

Karsten
« Last Edit: January 06, 2016, 03:26:27 PM by Kiesel »