Accuracy is common concern among customers of 3D scanning services and 3D scanning systems. However, not everyone is refering to the same thing when they discuss accuracy.
The most common definition of accuracy is the deviation between the perceived value of a measurement and the actual or true value of that measurement. This is an easy concept to to grasp for one dimensional measurements such as the length of a piece of rope, or the distance between two holes. Things get a bit more complex in 3 dimensions, and significantly more complex with copious data sets as produced by 3D scanning systems.
In the scanning world, there are several characteristics of the data set that pertain to the quality of the final result:
1. Dimensional Accuracy: since hundreds of thousands of points are gathered in a single scan, the dimensional accuracy of a single data point is impossible to quantify. Rather, scans are typically performed on known surfaces or primitive geometric forms. The deviation from the data set to that true form can then be calculated and reported as an uncertainty value. For example, a flat plane is scanned. The data is fit to the flat plane, then the maximum range of deviation of the data set to the plane is .005″.
2. XY Resolution: Assuming the scanning direction of the scanner is along its own Z axis, the XY resolution will correspond to the distance between the points along the scanned surface. This becomes important in the ability of the scanner to perceive a particular feature. For example, if the XY resolution fo a scan is .5″, and we are trying to scan a hole with a .25″ diameter, we will never “see” the hole in the data set because it may lie between the scanned points. So, in order to meausure the diameter of a scanned hole of .25″, my XY resolution would need to be on the order of .005″ to ensure that we have enough data to define the edge of the hole.
3. Z Resolution: the Z resolution pertains to the resolution of the scan in a direction normal to the surface being scanned, and woudl dictate the minimum height of a surface feature or defect that could be perceived by a scanner. For example, if a scanner has a Z resolution of .001″, then it could perceive the the .010″ thickness of a piece of tape applied to a surface, but it could not perceive a .0005″ deep scratch.
4. Noise: this refers to aberations in the scan data caused by surface reflectivity and ambient lighting. Where this may not affect dimensional accuracy on feature measurements such as planes or spheres, noize level will definately impact the Z resolution if it is of similar magnitude.
In closing, when sourcing a 3d scanner, or scanner service, it is important to understand the type of accuracy that is being quoted, and how the final results will impact your project or your application.
Please feel free to contact me for more information: