Accuracy is important in 3D scanning, but maybe not why you think
In our last post about 3D scanning, we talked about resolution – and why pixel count shouldn’t be your only consideration.
Resolution and accuracy are commonly thought to be one and the same, but they are not. Neither is it true to say that a higher resolution scanner will give a more accurate result.
Resolution only determines the quantity and spacing of data points you capture, but accuracy gauges how well it relates to the object you scanned in the first place. That’s why it should always be the most important thing you consider when choosing a new 3D scanner.
What should you buy?
To replicate all sides of an object, you may need to scan it many times to capture every angle, then register and merge the parts into one. Any minor imperfections in each scan will accumulate as the parts are combined, and could lead to a face or edge misaligned – or worse, a gap in the finished data.
These gaps and errors are difficult and very time-consuming to fix. Because you can only do that by manipulating the data, you’ll end up with a model that’s no longer true to the original.
That’s why you should always buy the most accurate scanner you can afford – and choose accuracy over resolution almost every time. With a high degree of accuracy, the data that you capture will be true to the object you’re scanning, and you’ll need to spend much less time processing the result. A more accurate scanner may cost you more, true, but will save you processing time and money. Over its useful life, you would expect to recoup any extra cost many times over. It will also give you data that is actually useful!
Structured light scanners project a pattern onto the object, which is captured by one or two cameras placed either side. Because they are offset, these cameras can see how the geometry of the part distorts the pattern and, because the distortion is a function of range, the scanner can work out the shape and dimensions of your object.
Some of the scanners we offer use blue light, rather than white light, to produce the pattern on the subject. Why? Because blue light from LEDs has a very narrow band and a specific, short wavelength. Detecting only this wavelength helps you gather data with less noise, which is useful if you’re working with awkward surface finishes, or in difficult scanning environments. In short, blue light produces better data quality which also means savings on processing time.
Naturally, most objects can be scanned with blue light, but for those rare occasions where the object colour absorbs high levels of blue, the scanners give you the option to switch to red or green instead. Other scanners in our range use white light, which could in some cases produce slightly noisier data, but can be used on any object, regardless of its colour. Whichever you choose, the price will be roughly the same for equivalent scanners in either class.
How accurate is “accurate enough”?
If you’re scanning something for use in animation, small errors aren’t particularly worrisome, even flaws in data don’t cause too many problems. When scanning people, the subject’s shape can change by millimetres with every heart beat so micron accuracy is pointless.
That’s not true in engineering, though. When you’re scanning an object for physical reproduction, accuracy and time saving are the main reasons for using a scanner at all, rather than manually recreating a model in software alone.
With a more accurate scanner, and by using blue light, it means less time and frustration spent processing data, and an end result that is true to the original object.