7 factors affecting
3d scan precision
firstly determine if precision is necessary for your project.
if it’s a decorative display, precision is likely not needed.
precision is needed for: (i) 3D data that will be analyzed for preservation & research purposes, (2) nested mechanical components.
consider the maxim “more is more”. mobile phone camera manufacturers have raced to outdo one another with higher resolution offerings. however, higher resolution is rarely needed if used purely for social media. resulting tradeoffs being larger storage, compression cost & time, longer loading time.
it goes without saying that scanner calibration is a critical process that cannot be skipped. any worthy precision tool must be calibrated:- even analog vernier calipers need calibration. see video here.
7 factors affecting 3D scan precision
1) 3D scanner accuracy thresholds
it is extremely important to know the scanner’s range; different scanners are manufactured to cater to different scales & purposes. know what is the expected per meter deviation. stacking errors will result if this information is ignored & not compensated for.
consider kitchen knives that come in a mind-boggling variety (size, shape, material). when cutting a premium 180kg fish, compare a fruit knife’s performance VS a purpose-built fish saw.
2) temperature & humidity.
what is the scanner’s optimal working temperature & humidity VS scan location’s conditions? Soaring worldwide temperatures make this question even more pertinent. How do those conditions affect the material properties of the scanned object?
3) object movement
can the scanner handle object movement with precision?
4) data collection & processing
good data tracking, alignment, & registration can only be achieved through sound preparation & technique.
remember that any toddler can take a scanner, push the start button & start scanning.
5) automated hole-filling
for precision scans, the use of automated hole-filling is best avoided.
there are several ways to achieve a hole-free 3d scan result, but bear in mind that the longer the process, the higher the cost. also when reverse engineering is involved, having strategically-planned holes does not affect the precision of the final deliverable.
6) smoothing
it can be really tempting to smooth 3D data in order to make it look “more presentable.” However, smoothing is also a (flaming) red flag because it casts a huge doubt on the quality of captured data as well as processed data.
7) addition of external “noise”
there are many ways this can happen:- introduction of dirt, inadequate handling of 3D data, & improper application of matting material.