So, you want to measure the shape of an object, perform a 3D surface scan or conduct surface metrology tests. Google tells us there are many measurement methods to achieve these goals. Problem is, which one’s best for you?
Object-measurement techniques can be divided into two basic categories: contact and non-contact. In the former, the object has to touch the sensor, which usually complicates the measurement process, slows it, and can possibly damage the object we examine. Moreover, if the measured object is a soft material, like rubber, textile or tissue, it can’t be measured at all. So here we’ll only discuss non-contact measurement sensors, in which the sensor and the object to be measured are separated from each other.
It all comes down to determining the precise distance between the two and moving one with respect to the other, in order to reconstruct the object’s shape. This goal can be achieved using various optical techniques, each best suited for a different type of application.
Time of Flight
Perhaps the simplest measurement concept is the time of flight method – if we shoot a pulse of light that reflects from an object and measure the time it takes it to return to the sensor we can calculate the distance to the object, since light’s velocity in air is constant. Problem is that it moves rather quickly: light can circle the earth roughly seven times in a single second… If you need to measure large objects and the accuracy is not that important, that’s perfect. But if you want to measure short distances accurately you’ll need ridiculously-fast and extremely-expensive electronics to create the light pulse and to time the measurement, making it impractical if not impossible. So what else can we do?
The most common way to go is using the triangulation method. A sensor emits a laser light that hits the object at some incidence angle, reflects from it and detected. Since light moves in straight lines, a triangle is formed between the laser source, the measured object, and the detector. By measuring the exact location the laser hits the detector we can calculate the distance to the object using simple geometry (see Fig. 1). Sounds great, but what happens if we want to measure inside holes or steep-angled objects? In these cases the light path will be obstructed and it will not return into the sensor, rendering the measurement useless. So, if the shape of the object is complex or you need a large angular coverage (e.g., measuring a sphere), triangulation-based sensors are not the solution you need.
Figure 1: Triangulation distance sensor (source: Wikipedia)
Other measurement methods that suffer from the same problems are vision-based ones: stereoscopic imaging uses two slightly-separated cameras to measure distances. A huge advantage of using vision-based methods is that they’re typically able to measure an area (or a volume) without the need for moving the object or the camera. This is due to the fact that distance information is extracted from the two-dimensional images taken by the sensor.
But again, problems arise when measuring grooves, holes, etc. It’s just impossible to measure them!
So we need to find a collinear solution, in which the light travels to and back from the object on the same path.
Confocal sensing is one such method. Here light is focused using a lens on a measured object, reflects from it and returns into the sensor (see Fig. 2). In a monochromatic confocal sensor a single-color light is used and the sensor and object have to be mechanically moved with respect to each other to keep the object at the focal point of the lens. This makes the technique very slow.
In contrast, white-light confocal sensors use light composed of a many colors. The lens focuses each color at a slightly different location, and by measuring the returned light’s exact color we can evaluate the distance with up to nanometric precision.
However, these sensors are very expensive and cumbersome and they are limited in both their stand-off, which is the distance between the sensor and the object, and their working range, which is the distance span in which the measurement is possible. This means they can only measure relatively small objects located very close to them. Another drawback is their relatively-high temperature dependency, which means that if the measurement is taken at a different temperature than the ambient one during calibration, their accuracy is reduced.
Figure 2: Confocal microscope diagram (source: Wikipedia)
Interferometry-based sensors have slightly different drawbacks: in these types of sensors the light splits into two distinct paths, with one hitting the object under investigation and the other serves as a reference (see Fig. 3). Again, a mechanical scan to change one arm with respect to the other is required, which complicates and slows the process significantly. Moreover, only relative distances can be measured, and the system has to be carefully monitored so both interferometer arms will experience the exact same environmental conditions.
A slightly different solution is called optical coherence tomography, in which we do not change the optical path length, but rather the light source’s color is being scanned. Though this is much faster than mechanically changing the length of one of the arms, the accuracy is reduced, and it is mainly used for medical applications.
Figure 3: Interferometer-based distance sensor (source: Wikipedia)
So what else can be done, you ask? Here’s where Optimet enters the story: Optimet distance measurement sensors use a unique patented technology called conoscopic holography which allows fast and accurate measurement of extremely complex features like holes and grooves, with angular coverage of ±85°.
This measurement technique is both collinear and allows large stand-offs and extended measurement ranges. How? Let us briefly explain: the laser displacement sensor emits a beam that reflects off the measured object and returns to the sensor collinearly (see Fig. 4). The cone of light collected by the lens then passes through various optical elements, in which it is divided into two beams that move at the same path but have orthogonal polarizations. The two differently-polarized beams travel through an optical crystal that slows the velocity of one of them with respect to the other, so when they’re measured at the detector a relative delay is created between them. Since light can be classically described by its wave behavior (no need for quantum physics here!), this creates a measurable interference pattern that is calibrated to yield the distance between the object and the sensor.
There are times when other technologies are better suited: if you need to measure large objects and accuracy is not crucial for you, a time-of-flight sensor is a better solution. When you have very simple geometric features (i.e., no deep holes and steep angles), triangulation sensors are generally cheaper. If you need to scan a large volume quickly, vision-based technologies are faster. Finally, if you need nanometric resolution, then confocal or interferometry-based sensors are probably best suited for you.
More details about Optimet’s point and line sensors and unique advantages will be available in our next posts. Stay tuned!
Figure 4: Light path inside Optimet’s sensor, based on conoscopic holography