Understanding the differences in feature inspection with tactile probing versus laser scanning
Traditionally, laser line scanners are being used for creating digital copies of (generally freeform) objects. The high data rate of optical systems, compared to traditional touch probing, makes this technology extremely suitable for collecting many measurement points required to sample complex (non-prismatic) geometrical shapes.
Laser line scanners have evolved over the last few years to a point where they become a valid alternative for tactile inspection of geometrical primitives. Improvements in resolution, quality of the optics, image processing and data analysis have turned laser line scanning into a sufficiently accurate but much more productive substitute for tactile measurements, even for feature inspection. There are several real-life situations where the higher measurement density of optical systems leads to more representative (closer to reality) results than tactile measurements.
In the current situation where the technology of laser line scanners has evolved to measurement accuracies that are very close to the accuracy of tactile probes, the greater data density of laser scanning provides a major benefit for feature fitting. The accuracy of tactile probing technology relies heavily on the assumption of perfect geometry for good feature fitting. Due to the low sampling rate, tactile data is very sensitive to imperfections of the geometrical feature.
The out-of-roundness of a stamped hole in a sheet metal part is typically much larger than the measurement accuracy of a tactile CMM. When identifying the hole with just three to five measurement points, the roundness error may result in an uncertainty that is many times larger than the accuracy of the CMM.
Optical measurement systems collect much more data in the same or less time, resulting in more correct handling of geometrical imperfections. The individual measurement points in an optical point cloud may be noisier, but the higher point density results in improved representation of the real geometrical feature.
A laser scanner captures the entire part geometry for dimensional inspection or reverse engineering purposes, covering both freeform surfaces and geometric features. Laser scanners are used with CMMs, articulated arms, optical CMMs or even robots. The resulting 3D copy forms the backbone of a digital inspection process that streamlines quality control, reduces time to market and cuts development costs. Ultimately, a laser scanner captures thousands of points in the time it takes a touch probe to measure one point, making it a more productive and cost effective inspection solution.
Laser scanning technology works by projecting a precision laser stripe onto a workpiece while its built-in digital camera captures the projected laser line under a fixed angle. Today’s digital line scanners with advanced CMOS camera technology offer impressive point resolution, high accuracies and capture tens to hundreds of thousands of non-interpolated points per second. As they achieve high point cloud density with tremendous scanning speed, they accurately digitize freeform surfaces and geometric features at high speed.
Limitations of tactile inspection
This chapter explains a number of issues that occur during tactile feature inspection. They typically occur on imperfect parts that have flatness and roundness errors, edge rollover, burrs, etc. which could be generally termed as “near” geometrical features. The ‘undersampling’ of tactile inspection makes the results very sensitive to such influences. Laser scanners on the other hand collect much more data, allowing elimination or reduction of such influences.
1. Flatness errors around the feature
Material deformation around a punched hole results in a flatness error in the surrounding plane. In such cases, the locations where the few tactile points are taken (to identify the relative plane of the hole) may influence the result strongly.
2. Non-orthogonally between feature and the plane of the feature
When the measured relative plane is not orthogonal to the side wall of the hole, the points measured on the side wall are projected along a direction that is not parallel to the centre of the hole. This results in a position error of the hole centre. This problem typically occurs when the hole is located on a (slightly) curved surface.
In case of laser scanning, enough points are measured on the inner side wall of the hole in order to identify its position and orientation accurately. Also, the relative plane is measured with much more points, resulting in much better conditioning of the fitting of the plane.
3. Hole roundness errors
The roundness error of a hole has a large influence on the identified centre of the hole. There is a significant difference between the location of the yellow and blue circle in the sketches above, as a result of different touch points in a hole with out-of-roundness issues.
4. Rolled-over, burnished and fractured edges causing uncertainty on where the tactile probe is making contact (probing problem)
When measuring a hole with a curved edge, the position of the probe on this edge determines the position of the edge. In the above right picture, the black dashed lines show the identified edge, depending on how deep the tactile probe takes measurement points. Especially on thin material, it is very difficult to assure that the tactile probe makes contact with the material in the most extreme point of the curved profile. Measuring chamfered features requires very dense sampling in order to determine in what direction to calculate radius compensation for tactile probes.
5. Radius compensation error (calculation problem because of uncertainty of the surface normal).
When a tactile probe makes contact with material as indicated in the above picture, the radius compensation may result in unexpected measurement points if a neighbouring surface is touched first. For example, the green point may be considered for determining the horizontal plane instead of the blue point.
Measuring results on realistic sheet metal parts
A real sheet metal part (A-pillar) is was measured with a tactile probe taking three points on the surrounding plane and five points on the inner mantle of all holes. The same holes were also scanned with a Nikon Metrology LC15Dx scanner. The measurement uncertainty of the LC15Dx laser scanner for feature inspection is not very different from the tactile measurements, while there is a significant (35%) reduction in measurement time. Diameter and position errors for both measurement techniques fall well within the required tolerances for sheet metal applications. Results of these measurement are presented at various conferences and can be obtained on request.
Analysis and practical tests reveal that laser line probes achieve a feature inspection accuracy that cannot be distinguished from tactile accuracy. The higher data density of optical systems enables elimination of the local imperfections on real parts that mislead tactile inspection. Comparable accuracy but a much higher productivity makes optical laser line scanners an attractive way to significantly increase the productivity of traditional CMMs.
The work reported in this publication is supported by IDIFIX