Application of Structured light projection

Structured light projection (SLP) is a non-contact optical metrology method which can measure to the micron scale. The technique involves projecting a known light irradiance pattern onto the surface under test, which then scatters the light creating a 3D image on the surface. The scattered light is recorded by one or more cameras and the scattered pattern is compared to the modeled projected pattern to determine the surface profile. For a test surface to be measurable with the SLP technique, it must be a scattering surface over the spectrum of light that is projected on to it. Typically, visible light is used for a SLP device, as the light sources and projection mechanisms are more straightforward to achieve, although another wavelength can be used when necessary. Figure below demonstrates a projected fringe pattern that was captured by a camera, and the corresponding surface that was calculated.

Surface height accuracy

The surface height features that are measurable by the SLP technique are correlated to the resolution of the pattern projected onto the test surface. The standard methods have produced surface height resolution of 2 to 10 microns RMS over several square millimeters. Also, the measurable area can often be increased at the expense of surface measurement accuracy. However, submicron accuracy has been achieved by utilizing a cube beam splitter to create a coherent fringe projection, which resulted in a 150-nm RMS surface accuracy.

High-speed measurements

Resolution is not where SLP offers its largest benefit; instead, the technique allows for measuring larger snapshots of test optical surfaces and its surrounding mechanical structures. Particularly in scenarios where the test surface may be changing rapidly, SLP can offer metrology snapshots of the surface on the millisecond scale, depending on the capture rate of the camera(s) used. In order to capture such fast data sets, a stroboscopic fringepattern method has been developed that ‘freezes’ the test surface in place. This stroboscopic pattern was able to measure a precision flat to 6.2-mm RMS in surface error.

Freeform surface metrology

An additional benefit of the SLP method is the ability to measure highly freeform surfaces. Multiple projectors may even be used to fully illuminate a test surface that otherwise may be impossible to fully project onto due to the test surface itself blocking light ray paths [21]. Using this multi-projector technique, a precision ball was able to be fully measured with only several microns of RMS error. In some scenarios the freeform surface may instead return interreflections, where the projected light onto one surface reflects from another surface and returns to the camera. Epipolar imaging and regional fringe projection can help to accommodate these issues, although accuracy may suffer.

Practical tips

A phrase often muttered in optics is ‘you can never have too much light’, and this is particularly true when it comes to structured light applications. Take care to reduce stray or ambient light in any test setup, and don’t be afraid to use a powerful light source if you find it difficult to get a good measurement.

Further, if you have time, repetition can be your friend. Multiple measurements can help to average out temporal random noise.
Also, if you can, measure a standard that you know well to calibrate your system. The projected light will be imperfect, and your camera may also have aberrations such as distortions. Properly accounting for them is essential in getting an accurate geometry of the measurement configuration. If it is possible, you can try a comparative measurement (e.g., quality check application), although this is not always feasible. Additionally, if you have a rotationally symmetric optic under test, you can perform a rotation calibration to reduce systematic errors.

Lastly, lock down your parts for this heavily geometry-dependent optical metrology solution. It can take quite a lot of time to properly configure and position your system and all the components; however, it only takes a moment to bump an optical table and move something. Take the time to properly secure all components in place, not only so that they remain secure during testing but also so that you can repeatably position your optic under test if  necessary.