Structural inspection measurements?

[OnTheRopes]

This is not directly H520 specific but thought it might be best posted here as there are a good bunch of experienced commercial operators active in this forum.
I am just beginning to look at inspection work for buildings, towers etc.
I have no real idea of the capabilities apart from obtaining high definition close up photos of a building but have a couple of basic questions.

1) Inspecting say one side of a building and producing a series of overlapping photos, is it possible to stitch them together by using a mapping type of software?

2) What is required to produce accurate measurements from a structural drone survey? Say the length or width of a fracture or crack? Is it possible and if so what is required?

 

[javier Olalla González]

Sí, puede usar programas fotogrametricos como PIX4D y photoscan para hacer un modelo 3D de un edificio.
Después de eso, con una herramienta de escala puede producir mediciones precisas.

 

[javier Olalla González]

Here yuo have an example:

 

[LutinMalefique]

Yes it is possible with photogrammetry software. If the photos are georeferenced, you can already make measurements but with average margins of error of a few meters.
Ideally, in order to have the most precision, ground targets are needed (Ground Control Points), which are then read by GNSS GPS.
If this is not possible, the margins of error can be reduced by knowing at least 3 distances, by placing metric marks on the ground (tape meters for example)

 

[OnTheRopes]

Sí, puede usar programas fotogrametricos como PIX4D y photoscan para hacer un modelo 3D de un edificio.
Después de eso, con una herramienta de escala puede producir mediciones precisas.

Thank you, very helpful post and I will look closer at Pix4d, the video looks interesting.
So this could be done with any drone with high definition camera and preferably a longer focal length camera?

 

[OnTheRopes]

Yes it is possible with photogrammetry software. If the photos are georeferenced, you can already make measurements but with average margins of error of a few meters.
Ideally, in order to have the most precision, ground targets are needed (Ground Control Points), which are then read by GNSS GPS.
If this is not possible, the margins of error can be reduced by knowing at least 3 distances, by placing metric marks on the ground (tape meters for example)

So if I am taking images of just one profile of a large structure, let's say a residential tower block, I cannot see how ground targets can be used as this would essentially be a 2d survey.
I wonder how this can be achieved with accurate results? I admit to being ignorant of the subject and any info or reading material on the subject you can point me to would be of great help. I am in the process of persuading a company to trial the idea. Thanks

 

[LutinMalefique]

Accurate georeferenced ground targets give you scale with a least centimeter precision. For your exemple, you will, for exemple, take mesureaments of some windows, you enter the mesureaments of 3 or more windows and the software will scale your model/orthophoto.

http://www.agisoft.com/pdf/tips_and_tricks/CHI_Calibrated_Scale_Bar_Placement_and_Processing.pdf

 

[OnTheRopes]

Accurate georeferenced ground targets give you scale with a least centimeter precision. For your exemple, you will, for exemple, take mesureaments of some windows, you enter the mesureaments of 3 or more windows and the software will scale your model/orthophoto.

http://www.agisoft.com/pdf/tips_and_tricks/CHI_Calibrated_Scale_Bar_Placement_and_Processing.pdf

Brilliant. I'll have a read when i get back from work

 

[LutinMalefique]

you're welcome, you will easily find information on this subject on the Photoscan or Pix4d sites

 

[arruntus]

In a building, which is what you're asking, you do it like on the ground, but with the difference that the targets stick to the walls. Then relative scales are made between the targets you have set.

 

[OnTheRopes]

In a building, which is what you're asking, you do it like on the ground, but with the difference that the targets stick to the walls. Then relative scales are made between the targets you have set.

Yes a building, and it can be just one side of the building.
Are you saying the survey is taken with all images taken from the ground or are you saying the targets are placed on the ground and you use a drone for aerial photography and produce a map which references the scales you placed t the ground? Sorry to appear daft.

 

[arruntus]

No, hehehehehehe. The study area is the side of the building. What is done is to put targets, usually self-adhesive in the side of the building. Then you make measurements between targets, you can make several measurements. In the photogrammetry program you enter the scales, which is nothing more than visually identifying the target points and entering the previously measured distances. With these measurements the program now scales the model correctly and the measurements you make are very accurate.

On the sides of buildings you don't usually see a surveyor taking the coordinates of the targets

 

[OnTheRopes]

No, hehehehehehe. The study area is the side of the building. What is done is to put targets, usually self-adhesive in the side of the building. Then you make measurements between targets, you can make several measurements. In the photogrammetry program you enter the scales, which is nothing more than visually identifying the target points and entering the previously measured distances. With these measurements the program now scales the model correctly and the measurements you make are very accurate.

On the sides of buildings you don't usually see a surveyor taking the coordinates of the targets

Haha what I mean though, this is done by drone? If the building is 50 metres high then targets could only placed near top and bottom unless to send men down the side.
So, you place the targets, measure accurately with multi-station presumably, then photograph with drone and input to software and it can then scale it. To what accuracy if we need to measure cracks or fractures in concrete for example?

 

[arruntus]

If it is 50 meters, i.e. tall buildings then it is made with drone facing forward instead of downwards. Targets can be placed on the sides or under the windows along the length of the building. Putting them only on the ends is not a good idea. Although they are modern buildings without windows that can be opened, they often have platforms for cleaning and therefore could also be used to set the targets.

Normally, GPS georeferencing methods are not used, I mean normally, since vertically the taking of the points is more difficult and the vertical location of this façade is already determined. It is easier to establish the scale using the scaling tools of the photogrammetric programs and consists of indicating measurements between the points of the targets. But it could also be done even if it doesn't make as much sense as on a piece of land.

 

[OnTheRopes]

If it is 50 meters, i.e. tall buildings then it is made with drone facing forward instead of downwards. Targets can be placed on the sides or under the windows along the length of the building. Putting them only on the ends is not a good idea. Although they are modern buildings without windows that can be opened, they often have platforms for cleaning and therefore could also be used to set the targets.

Normally, GPS georeferencing methods are not used, I mean normally, since vertically the taking of the points is more difficult and the vertical location of this façade is already determined. It is easier to establish the scale using the scaling tools of the photogrammetric programs and consists of indicating measurements between the points of the targets. But it could also be done even if it doesn't make as much sense as on a piece of land.

That makes sense (I think) now I just need to find somewhere to study the methodology