For an upcoming application we’re going to be researching how the accuracy of planar measurements might be improved through the use of 3D imaging – particularly a camera pair. The concept is that point imaging in stereo may enhance general single-plane measurement accuracy in one or more ways:
Depending upon the library or component, multi-camera calibration may be more precise
The location of a feature, even if only being considered in a single plane could be over-determined and possibly more accurate.
Visualizing the feature from two views may overcome some inconsistencies introduced by imaging or lighting.
When part presentation is inconsistent (it always is), the extraction of the measurement points of the features using 3D points might reduce error.
I speculate that this technique might help even simple measurements, and I’m anxious to test that theory. With camera prices at very reasonable levels, even a little improvement in accuracy could justify the cost of the additional camera.
Any input? I will add to this discussion as we try this out.