Faro Photon VS Leica C10

[pollete]

Hi folks:
Here am I with some results of my in-field comparison between the two scanners. PS vs TOF

Chapter One:
Lets start saying that I like both. Faro has done huge improvements in since IQvolution times. Despite what phase-shift detractors say, datasets are clean (or noise is so well handled that is not an issue in any case), point density levels can simply fulfil (and easily surpass) any needs within its range and scantimes are near negligible among other project numbers, however it would be better to have more control on spot-spacing as C10 allows. On the other hand, the C10 scan operation takes much longer and needs to be considered for job scheduling.

The camera mount in the Photon is well done but the smoothness of its travelling should be revised as does the quality of the supplied tripod raising column (the locking mechanism and column rigidity are not tight enough to fully avoid small round travel of the origin during scan and the gear mechanism is prone to failure. The recording application integrates tethered shooting so that taking photographs can be automated and is by far a more engineered solution than the NodalNinja head provided by Leica. Better, but it is not yet possible to remotely operate the camera, that means one cannot use a light crane to raise the scanner position; instead of this, heavy duty or even dual platforms are needed to provide height and stability to both scanner and operator (or should I say camera-re-centerer?). Think about how this can affect to archaeological site ground, church wooden floors, as to any site that is not accessible to such wheeled or caterpillar machines.

I've already read some good and less than good opinion about the leica's superstar here. I would agree it is probably the best all-purpose scanner today but has it's weak points like range and the inner camera needs still to be improved. The external dslr mount and operation (Leica seems to be admitting the limitations of the on-board camera) is outlandish and arguably too manual and some embarrassing low-tech piece replacing the scanner after each scan.
But see what both have delivered here:

1.jpg

The camera/lens used with the C10 was so inferior to the one on top of the Photon that color are not comparable, lets see just points renderd with normals of two scans with very similar settings and taken from the same point:

2.jpg

zoom

3.jpg

more zoom

4.jpg

so I guess you are thinking the winner is... Faro! it delivers more detail and cleaner data!
but... back to reality.

5.jpg

Look at the pilaster!

Why the hell has the Photon "aged" its almost perfect shape? It's marble, a little glossy but not like a mirror!

I've also noticed that wall paintings that are flat, result in pseudo relieve with Faro. See the panel on top of the tympanon... I'd love to read your opinions about this. Have you ever experienced weird things like this? Would you think this problem could be attributed to a defective Faro unit?.

Thanks

[Oliver Buerkler]

Hello Jose.

Thank you very much for your elaborate testing and your report!

Let me try to answer the marble effect you are reporting:

From my experience, what you are seeing is not a problem of a defective scanner. The effect you see results from the physical properties of marble. As this is a pretty opaque material, it allows the laser light to penetrate into the surface. From within the object the light gets reflected in a diffuse and very unpredictable way. You can see in your example that even small changes in the structure of marble changes the results (this is what makes it look like a aged tree).
The effect is influenced by many uncontrollable parameters like the structure and kind of marble, the wavelength of the laser, the observation angle, surface structure and even the laser beam diameter seems to play a role.
For whatever reason it looks like TOF scanners can cope with this better than phase-shift machines. Maybe because of the way they analyse the return signal which may allow the scanner to recognize and compensate for the signal deformations.
I would not be surprised that other phase-shift scanner will show very similar results.

If you need further background information, there was a report on this topic by a team of Standford university and the Canadian National research Council held at a conference in 2001 in Vienna, Austria ("An assessment of laser range measurement on marble surfaces").

So again: what you see is not a defective scanner but most probably simply the result of the properties of (maybe phase-shift) laser scanning and it´s physical limitations.
If you repeat the same test on a different kind of marble you may get different results.

Regarding your observations on color scanning:

I am myself still sometimes surprised how difficult color scanning is. Suddenly one is depending on a passive sensor recording the environmental lighting conditions. Compared to a plain scan with its active sensor, this is a much more complex topic. Especially lighting at indoors conditions which should suffice academic demands is not an easy to solve task. This is already a challenge if you "only" want to take high quality photographs - adding the dimension of laser scanning to it is not making the task smaller.
In the first approach using a flash light mounted to the camera could solve the problem. But, to my knowledge, there is no flash light available which is able to cover the 180° field of view of our fishe eye lenses. This is unfortunately disqualifying the idea of using a flash.

I hope these explanations help to clarify the effect you observed. I am happy to discuss this further if you want.

Best regards,
Oliver Bürkler

[JGaspar]

Hello Pollete,


Congratulations on test.

Have you tried the same test with PS Leica solution HDS6100 / 6200?

Seeing the results with TOF Leica C10, I confess that I was surprise with the low details level of C10, because I work a lot with it in many applications and I believe it's possible a better result.

Can you put here, or send me in a PM with the values of distance, scan time and resolution?

Did you compare the assembly time and process time (register / aligment)?

I believe with a test like this we can extract more informations. We are gratefull if you could complete, although we know, is not the same tech principle, but is valid.

[pburrows145]

Hi there,

Thanks for posting this - it raises some interesting questions...

It would be good to see some raw meshed examples of these two datasets from the same package (Geomagic for example). Perhaps with one pass of smoothing (again equal strength in each).

What package are you using to render the image of the points. I only ask as sometimes the way certain programs do this you get strange results with the normals and data has a fuzzy effect which was not present in the raw data - yours appear to have shading applied too (?) which can obscure finer details.

What filters did you apply (if any) on import in the FARO scene software also?

Thanks in advance.

[rlasater]

Hello Pollete,


Congratulations on test.

Have you tried the same test with PS Leica solution HDS6100 / 6200?

That is a 6100 room for sure. Using the c-10 in there may be a bit unfair however it is a great comparison. Do you happen to know what firmware version the c-10 was on? We saw some issues of a thick fuzzy cloud when using the C-10 in small areas, however they seem to have fixed the problem.

Also Oliver is right, marble or any specular rock is going to cause some weird effects on any unit.

Good work!

[Matt Young]

José,

Its nice to see some practical testing going between these two very different scanners, with surprisingly similar results.

In your opinion, which scanner is better for this type of work?

Could you tell us a bit more about how the scans were processed, number of points, scan times etc. The results of the phase scanner look better than most I have seen in the past, I would be inrterested to see the scans from slightly different veiws to get a better idea of surface noise.

A meshing program would also give a good comparison between the two datasets with regards to surface deviation.

Matt

[pollete]

Chapter two is still in print. You will have much more details, (perhaps this novel will become a bit stodgy)