I am a Computer Science student that has a personal project to do. I decided to try and take on a project for a company, and one of the things that I need to measure is how perpendicular a laser is to a surface. The surface is non-reflective, and I need to adjust it until it is perpendicular to a 5 degree error. The adjustment is not an issue, but I am having trouble figuring out how I could check for how perpendicular it is. The surface is spherical in nature, but it is not a perfect sphere. Any alternative methods of checking it without having equipment too close to the surface (due to the destructive nature of what happens later). I'm sorry if I'm not very accurate, I have never taken any engineering courses and I'm not sure of the terminology. Any help would be appreciated
-
$\begingroup$ That's an interesting problem. Is the distance fixed or variable? What is the surface material / texture / finish? What is the minimum radius of curvature of the surface and what is the measurement distance? $\endgroup$– TransistorDec 28, 2022 at 0:07
-
$\begingroup$ @Transistor The distance doesn't change, however the material/texture/finish will be variable from item to item $\endgroup$– Stefan ConovaliDec 28, 2022 at 11:25
2 Answers
There's some irregularities in the question that might make the answer easier or more difficult. With a non-spherical object, one can and must allow for discrepancies.
Consider the following: There exist measuring devices which use a laser beam "bounce" to determine distance. Mounting this laser aimed at the approximate center of the "sphere" will provide a reference distance. My DME reads to 0.1 mm but I'm not sure of the distance capacity, as I'm unable to locate the device.
Once the reference distance is determined, angling the measuring device to four cardinal directions by an equal amount will provide an indication of squareness.
The image above, created by myself, shows C as the centerline of the laser, with A and B as equal angular displacement from the centerline. Measuring distance A will return a value greater than that of distance C, while both will be values greater than B.
The view can be considered both a top/bottom and a front/back representation. When A and A' and B and B' are all as equal as possible (and within accepted tolerance), the spherical shape will be as close as possible to the desired perpendicular.
Even though the question presents as a non-reflective surface, one is not using a requirement of reflectivity for the laser. It need only provide a visible laser dot for the measuring device to function.
If the final distance exceeds the maximum distance of the measuring device, one could create a couple of intermediate points within the measuring distance and extend a line using other methods to reach the required station.
-
$\begingroup$ The problem needs to be fully automated, as the reason for creating the program is to cut a lot of time from the process. While the items are spherical in nature, they aren't ever perfect spheres I'm pretty sure. I understand the premise behind the solution, and I've thought of something similar. But if the surface is not a perfect sphere, then this no longer applies, correct? $\endgroup$ Dec 28, 2022 at 11:34
-
$\begingroup$ My thought on a possible solution would be a small reflective sticker (or something of the sort) and shining a laser into it to get a bounce back. Though I'm not sure how I would read that, as the return would be going back to the origin (or as close a possible) which leads me to the problem that I cannot have a sensor or something to read the return laser $\endgroup$ Dec 28, 2022 at 11:34
-
$\begingroup$ Non-spherical objects would not be a problem, depending on the precision required of the results. The laser measuring equipment includes a display presenting distance. If two or three lasers were placed close together, one could determine a variation across the three. The difference would then be providing your precision for the non-spherical object. Some of the information added in the comments suggest that a diagram or drawing would be helpful. $\endgroup$ Dec 28, 2022 at 16:18
If, as you mention in a comment, you can apply a small reflective sticker, and the laser isn't too powerful (at least during this alignment procedure), and the radius of curvature of the surface isn't too small, then I think it should be doable. (You don't specify the range of possible radius-of-curvatures of the surface, but if they are comparable to or smaller than the laser beam diameter, "perpendicular" becomes ill defined.)
When aligning mirrors, I often put a transparent plastic ruler between the laser and the mirror so I can see a bright spot on the ruler where the laser passes through the plastic on the way to the mirror, and a fainter (but still very visible) spot where the beam reflected from the mirror hits the ruler. I adjust the mirror until the two spots are on top of each other, which tells me the mirror is perpendicular to the beam. With a camera observing the ruler, some image processing software to determine the position of the spots, and control over either the object or the laser position/angle, you should be able to automate this.
A possible method that does not require specular reflection is to use photometric stereo imaging to determine the normal to the surface, and then use that information to set the angle of the laser. There is a large research literature, public code (e.g. on github), and commercial photometric stereo systems are available.
