Our 3D scanner is a device capable of reconstructing the 3D surface of real objects into 3D models with extreme accuracy (up to 0.1mm). It can be used to bypass 3D object modeling by hand or provide great reference for such tasks. It uses a laser beam for measurements (non-contact measurement), can be simply transported and used even for scientific purposes (e.g. archaeological).

We offer 3D scanning and limited processing of the scanned data. Please read details below.

Typical processing pipeline

Recommended objects

• Acceptable scale of the object: we recommend something around 80x80x80cm for a cube or 100x30cm for a cylinder at most. This is not a hard limit as more scans can be still used to cover larger surfaces, however many scans are much harder to post process.
• Acceptable and recommended usage: human face with ears, palm and fingers, statues (esp. useful for transferring traditionally modelled clay based models to the game environment), rare artifact / statue scanning (eg. archaeological purposes, molding of copies for museums..)
• We especially discourage from scanning full human body - the actor will be unable to stand still during long time period needed to take all the scans necessary, which will result in misaligned data. Different types of scanners must be used for this, these are very large and fast devices capable of scanning whole body in ~2 seconds.
• And metallic objects are also often hard/impossible to scan with laser scanners generally because these are shiny and reflect laser rays away! It is possible to cover their surface with some powder making them suitable for scanning, however this must be considered in advance and approved by the client.

Prepare an item for scanning - must not be shiny nor dark and reasonably complex

• The device insists on laser beam reflected off a scanned item back to the lens of it - completely shiny surface bounces the beam away, too dark surface absorbs the beam at all and too complex surface with great depth distances over small surface area (e.g. fine grid) causes erratic interferences. Scanning of such surfaces may result in completely unusable data!
• Our best experience is the use of gypsum castings - the surface is white enough and the process of casting naturally smears all excessively fine structures.
• This must be prepared by the customer in advance. We only can provide fine white powder, that can be used to cover the object's surface, however it may be impossible to go this way as the item might degrade by the powder and so on..

Scan the model

• The scanning is much like taking photos of the object from all sides. The scanner is, in fact, a camera, which stores common RGB color photo, however additionally it also stores Z depth of the pixel (so each single pixel has RGBZ information). The number of scans needed depends on the item being scanned, for common 3D object <10 scans may be enough.
• Each single scan takes just few seconds, so that the process is very fast. It is essential, however, to ensure scanned object stays still, imagine that like common camera taking shot with shutter time of roughly 3 seconds.
• Also, the object should remain identical among scans - that is a problem of scanning the faces as men tend to change mimic over time, which results in misaligned data - much like taking photos of a panorama in a wind, where overlapping areas change significanly between consecutive shots.
• We provide the scanning. In case of expensive items, client is asked to rotate the item themselves to our needs to provide all the sides of the item available to the scanner over time. We may refuse to touch the item at all (risk of damage, fall, corrosion from perspiration..) - this applied to very rare items 2500 years old we scanned for the Czech National Museum, one can imagine that any damage we would cause, though unintentionally, might result in serious problems.
• The scanner is very successful in scanning displacement data. The data can be used for painting onto final 3D model's surface and later transformed onto bump/parallax textures! This is one of very powerful uses of the technology.

Take high res photos of the item

• We normally use our high-res digital camera and take photos of the model in time of scanning. The scanner saves RGB photos, however the quality of RGB data is limited - CCD chip of the scanner is optimised for laser beam registration and color texture data is overall low quality. We overcome this taking photos in parallel.

Process the data

As mentioned above, the scanner works much like common digital camera. To build one model of the item needs to align separate 3D scans (so called shells) together so that they form a complete model - this is much like aligning photos to create panorama of them, it only needs to be done in 3D, of course. It is not too hard to learn the aligning process and we have good tools for this task, however it may take really significant amout of time to finish aligning. We are normally unable to provide aligning service, however upon agreement, we may e.g. lend the PC with tools installed and client's operator will be introduced to the processing pipeline and may align the data using it.

• It must be mentioned, that this type of laser scanners generally suffers from 3D noise. It generally appears as a sort of "orange peel" look on places where flat surface was expected and it is mostly impossible to get rid of that completely, though smoothing certainly helps.
• If any part of the object is made of material impossible to scan at all, eg. glass, this must be completely built manually later.
• Resulting model is critically high poly. It may result in so extremely large models, that 3D rendering packages might have serious problems manipulating it at all! (the tools allow for polycount reduction)
• That is why we often recommend to simply take 3D scans of the object and then to use these scans as a reference for modelling of new human created model only. Good is proportins of the new model would perfectly fit to the original, however the mesh would be perfectly prepared eg. for skinning to the bone structure and texture mapping, making work far more convenient.
  Simply said - load 3D scan in the wievport and start modelling new lower poly model inside of the scanned one. Easy, isn't it? And believe, if making real time things such as games, forget about using direct scanned data in these projects, it is mostly impossible to create such a low poly model by automatic polycount decimation..

Mapping the high-res photos on the model

• It is possible to use RGB data created by the scanner itself - the processing tools keep the information intact and create mapping coordinates automatically. However, it must be expected that the UV map of the model will not be "nice" as the software routines lack intelligence of human operator. Also, as mentioned above, RGB data of the scanner is of a limited quality and resolution.
• That is why we recommend to use RGB data of the scanner just for reference and map hires digital photos taken by digital camera onto the model manually - most probably using Zbrush software.
• The client is provided with set of digital images - either uncompressed 8 or 16-bit bitmaps or camera RAW data (preferred, the client may develop the images to their needs, eg. adjust color temperature, exposure..), the mapping of the photos onto actual 3D surface must be done by client's operators.

We are always open to share our knowledge on how to process and the data quickly to our clients. Please don't hesitate to ask!