Again, I reiterate, because this is important!
One of the prime limitations of all forms of structured light 3D scanning is that you get interference from secondary reflections. Ideally, you want to programmatically select points on the object by shining light onto the object’s surface and only observing the results from the first bounce of such light off the object’s surface. This is not strictly possible in the real world; however, there are mitigations you can take to approach this ideal.
The primary one applied universally is to powder coat the object. The very simplest powder coats simply assure that the object has a diffuse surface. By having a diffuse surface, light shined on the object can be observed equally from all angles. Also, although there will still be interference from secondary reflections, it will be exponentially reduced due to the fact that diffuse reflections scatter incident light in all directions, hence reducing the intensity of the reflection from every angle.
To have a more extreme exponential reduction in secondary reflections, the powder coat can have a darker color. For example, rather than a white color, it can have a dark gray color. The primary problem with this is that you need a brighter light source to capture the same quality of data at the same camera exposure settings, and therefore at the same camera speed. With lasers in particular, high power lasers are often commercially restricted due to their inherent danger to human eyesight, which means that simply increasing the brightness may be impractical.