Does anyone have experiences with scanning rubber surface profile (like conveyor) with the use of laser scanners (profiler or 3D), such as SICK, Riegel, Leica? What is the practical achievable accuracy of the profile scanning? I've tried to scan it with Leica P20, and there is a noise +/-3mm at short distance (3m) - much bigger than in specification of the scanner...
I worked on laser scanner accuracy and noise removal during my phd at Politecnico di Milano and Torino. Usually laser scanners have an indicated accuracy that is calculated considering the standard deviation of the rangefinder, the 1 sigma value. This means that about 70% of measurements have the indicated accuracy. For a real accuracy value you have to consider 3 sigma. From a statistic point of view 3 sigma means that 99% of measured points have that accuracy.
In addition the surface material can change the laser scanner spot reflection. The rubber usually is black and laser scanner do not like black materials so much. This could decrease a little bit the accuracy.
I hope my answer could help you.
You are right, but the decrease of accuracy is not "a little bit".
For black rubber noise is increased x10, for black metalic car there is almost no useful signal!
I am using Leicaa Scanstation P20 and so far made a lot of experiments with different type of spherical targets white surface. What I obtained from my preliminary results is that the instrument works better far away from the surfaces. For the spherical targets a reasonable distance is 6 meters. Of course the type of surface is important too. As said by Leandro Bornaz black is not the best, and if it is shiny you can get a lot of noise.
I've used white paper as a reflective surface at distance 15m , it is white and not shiny, so theoretically the noise should be low, and in fact it is.
I'm tried as well to make measurement with laser tachimeter Leica TRDA6000, potentially more exact device.
I've found that measurement on retroreflective tape at distance 15m is not better than on white paper (in the noise level).
The question is, what is the best surface to minimize the noise in laser measurement at distances, lets say, 5m, 15m and 100m?
Hello.
You will always find noise .
When you use a 3D scanner usually the accuracy of the instrument is indicated in this way: accuracy = e.g. 3 mm @ 1 sigma in company test conditions. This means that the ~70% of points have 3 mm accuracy.
If you want to consider the 99% of points you have to consider 3 sigma as accuracy, this means 9 mm.
In addition it is very difficult to establish the best acquisition distance. Usually is the minimum distance because the accuracy decrease with the distance (e.g. accuracy =3 mm @ 10 m + 1 ppm).
If you move the laser away from the object you have to consider the divergence value of the laser beam . Increasing the acquisition distance the laser beam spot increase. A bigger value of the spot means a more smoothed surface that could be considered more accurate, but you lost detail.
Hope this can help you
- Badges
- Science topic
- Similar topics
- Psychometrics
- Measurement