White Light Scanning: How It Works & Benefits

AI Thread Summary
White light scanning utilizes a white light source and cameras positioned at known angles to measure dimensions of parts with high speed and efficiency. The system triangulates the position of features by determining the center of energy of the white light dot, allowing for measurements even at steep angles. While it offers advantages like faster scanning compared to traditional CMMs and the ability to generate data compatible with CAD models, accuracy can be a concern, particularly for parts requiring tight tolerances. Recent advancements, particularly from German manufacturers, have improved precision to +/- 8 microns, enhancing the technology's applicability. Overall, white light scanning is a valuable tool in manufacturing, though it requires careful calibration and may involve significant costs.
Ranger Mike
Science Advisor
Messages
2,446
Reaction score
432
How does white light scanning work?
I know the system is used in industry to measure parts for dimensions like diameter, size.

I know there is a white light source. There is a camera mounted at a known angle relative to the white light source. I was told the signal is obtained from the center of energy of the white light source which means the system can measure features smaller than the size of t he white light dot. Also the white light source can be targeted at surfaces up to 85 degrees form perpendicular and still yield credible data.

does anyone have more insight regarding this ?
 
Last edited:
Engineering news on Phys.org
Center of energy? White light scanning is really just glorified pictures. The idea is that if you set up a part in a fixture of known dimension, then you can use stereo cameras to triangulate position of the part inside of the fixture.

White light scanning is pretty cool and it has some really nice applications. Firstly, the scanning is typically much faster than a CMM (depending on what you're doing exactly). Secondly, some features are much easier to measure (think of the profile of an airfoil). Lastly, the data is output in such a way that it can be read into a modeler.

The last point is particularly neat as you can overlay the geometry and see where you have thinning or out of tolerancing. This can greatly help with manufacturing.

The downside is the accuracy. We recently spend some time talking to some white light guys. We really couldn't get the precision that we needed. IIRC, we had a radial diffuser, and no matter what, we couldn't get better than +-0.002" or 0.004. While that sounds nice, we typically have parts toleranced to +-0.0005, so the precision was unacceptable.

Aside from that, there is significant costs associated not only with the hardware, but the post-processing software.
 
Thanks for the reply..
The Germans have taken it to the next level. They developed a cmm with white light sensor mounted ion two axis rotating head. They claim volumetric accuracy of +/- 8 micron certified with the same ISO 10360 standard used to certify CMMs. I was wondering about the camera and how it gathers its data. The stand off distance and triangulation is pretty straight forward.
you are correct about speed and ability to generate a cloud of points that cam be exported to software that can compare the data to the CAD model.,,
 
Those crazy Germans; +/- 8 micron is pretty impressive for a system like that.
 
a si understand it, the white light scan head has a light source and two cameras mounted at known angles from the source. from what i read, the cameras pick up the XYZ coordinate of the center of the white light dot. since these cameras ate at different locations from the white light source we have triangulation to determine the exact location of the white light dot. the whole scan probe mounted on a two axis motorized probe head which is mounted on XY and Z precision guideways..these have hi resolution scales and read heads for positioning the probe head. it is old cmm techonolgy with new age sensor..trick huh??
i am trying to find out exactly how the camera determines the white light spots center of energy??
 
From what I understand, it doesn't necessarily read a single point in space per "snapshot", it reads the entire "viewable" surface. In order to do this, you must prep the part with markers. You typically place these markers (they are just like clay dots) on features or corners.

When the camera scans the image, it uses the markers to get the orientation of the park. It then uses the image from that camera, and a secondary (or tertiary) camera to locate the depth of each individual point in the dataset it just created.

What you're describing sounds a little more like laser scanning.
 
close...the machine is calibrated prior to taking measurements. a datum sphere is used to this. the part to be inspected is placed on the table. to write a part program ( used to automatically measure the part ) the operator must joy stick the white light probe over the production part and manually take measurement of a plane. lien and a point to set up a part coordinate system. ( the machine was moving in machine coordinates before this.) once this is done the operator can import a CAD model of the production part and any data points take on the production part will be compared to the CAD Nominal points and we can see manufacturing trends from this. ( drift out of tolerance, etc..)
so markers no longer are required. I think the camera picks up the white light dot and reads the brightest area which would be the center regardless of point shape, ( ellipse, oval) so the white light can scan at very steep angles to normal and get good results..
i am trying to find info for a graph of this center of brightness ( energy) ..should be bell curve distribution i think
 
Back
Top