If there are any experienced users out there I am wondering how well it works?
Can you say... model the inside of a storm sewer structure with images?
Please shoot me an email I think this may be work trying..
> If there are any experienced users out there I am wondering how well it works?
>
> Can you say... model the inside of a storm sewer structure with images?
>
> Please shoot me an email I think this may be work trying..
I like where your head is at.
Keep us posted.
Can you model it, sure.. the real question is can you get results that are useable. That is less likely.
The basics of the software rely on different perspectives to model the features... with a manhole you really only get one perspective. The differences aren't enough to compute the geometry probably. Plus you have other challenges.. a dark environment, a low variety textures, required high iso.. all negatively effect the results.
If you were to climb down in the manhole and rotate around with lots of over lap between the images, you could get decent results.. but that kind of defeats the purpose doesn't it?
This is not a magic bullet... but you have me interested now, I'll pop a manhole and report back the results.
What if you used one of those "selfie stick thingys" and a light with continuous video?
I am also trying to determine the viability of the technology but for looking up not down. I am curious to know if it could be used to determine attachment points and wire sag on powerline surveys. I know there needs to be control in the photos but how much? How many pictures would be required? What would the accuracy be?
As far as the manholes, a surveyor with a Faro described how he lowered his F360 down into a manhole to do what you are describing. They rigged up a system for that.
Alan
Manhole Measurements
Here is a recent conversation I had with a similar firm:
Them:
Comments: the sewer structure has 3 pipes, (diameter, pipe angles apart, depth to manhole cover, cement shoulders adjacent the pipes). Just what features are the most important, inside the sewer hole cavity?
I would assume you want to do this with photogrammetry to avoid contact w/ the nasty fecal matter. There are a few different ways of doing this, but it depends on whether the features are “targeted” (could even be chalk marks) or if it is entirely non-contact measurement (i.e. measurement of natural features). CR-Photogrammetry requires convergent images, so the same features need to be measured in at least 2 images, but more (say min. of 3) are preferred.
Interesting application, and we’d like to work with you, but we’d still need a better understanding of the pipe run area, for what features (and accuracy) is key for the project use.
Me:
Excellent response. I think the e-mail approach so far is “getting there”. We can talk on the phone too anytime. I appreciate your willingness to discuss this. I honestly believe there is an untapped market for this application.
For me the most critical information needed is the vertical distance from the top of the rim of the structure to the invert of each pipe. That might not be as simple as we would like with your product, but maybe CR-Photogrammetry could get us the vertical measurement from the invert of the pipe to the inside top of the concrete manhole structure, then it would be a simple (and clean) tape measurement from the top of rim to the inside top of concrete.
Also, we want the pipe material type (PVC, concrete, corrugated plastic,…) This would be relatively simple by just viewing the image (hopefully).
Finally we need the inside pipe diameter size and direction.
So one of my thoughts is we just put a camera on the end of a pole of some sort (most structures will be much deeper than the sample I sent you) then simply run video while rotating and moving up and down. Then later at the office we could hand pick overlapping images?
Thanks for the open conversation so far.
Them:
Sorry Brad, I don’t think this is going to work with photogrammetry and my brief explanations below.
1. We need good strong angle perspective overlap in the images. The shallow depth image you sent would work, but say e.g., a depth of 8’, is not going to work easily as one cannot acquire the necessary camera separation (angles) to image this “sewer depth” from a manhole cover opening of say 3 to 3.5’ cover dia.
2. In order to measure the pipe dia., the project would need a scale distance (not in all the images, but at least 3 images).
3. “Rotating” the camera on-axis, will not work… That is not perspective image overlap.
From what you are (now) describing with depth, I really don’t see a solution. There are some things that would work for this “depth”, but that is only when physically targeting the Depth of the sewer, down to the invert of the pipe(s). So, the fact that photogrammetric targeting is required over the depth of the pipe, it starts to get into man hours doing the work, and still having to go down there… so if you have to go down there, might as well do the measurements manually.
Further, even if there was a way to do it with photogrammetry, I don’t see this working in all cases, as if there is sewage debris (crap, toilet paper, etc) occluding the pipe at its invert level, one still has to clear it, to even get to measuring it.
Finally, putting a camera down in fecal matter, is sure to get it in contact with it, or at the very least, “banged around” on the concrete base, or walls… just not a good thing for the camera.
If there was a simple way to do this, (photogrammetry) it would have probably been put to use, a long time ago.
Manhole Measurements
An encouraging update:
Let me think about this for awhile…. I have a decent idea of what you are after, and with 35 years of metrology experience, I might be able to come up with something for your application.
